Information

BIS 105: Biomolecules and Metabolism (Murphy) - Biology

BIS 105: Biomolecules and Metabolism (Murphy) - Biology



We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

BIS 105: Biomolecules and Metabolism (Murphy)

Transacylase and Phospholipases in the Synthesis of Bis(monoacylglycero)phosphate†

Article Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.

Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.

The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated.


Abstract

The outer membrane of Gram-negative bacteria constitutes a permeability barrier that protects the cell from exterior hazards, but also complicates the uptake of nutrients. In the case of iron, the challenge is even greater, because of the scarcity of this indispensable element in the cell's surroundings. To solve this dilemma, bacteria have evolved sophisticated mechanisms whereby the concerted actions of receptor, transporter and energy-transducing proteins ensure that there is a sufficient supply of iron-containing compounds, such as siderophores.


Regulation of intercellular biomolecule transfer–driven tumor angiogenesis and responses to anticancer therapies

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

Find articles by Ortiz, A. in: JCI | PubMed | Google Scholar | />

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

Find articles by Verginadis, I. in: JCI | PubMed | Google Scholar

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

Find articles by Cho, C. in: JCI | PubMed | Google Scholar | />

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

Find articles by Kubanoff, R. in: JCI | PubMed | Google Scholar

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

Find articles by Sun, Y. in: JCI | PubMed | Google Scholar | />

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

Find articles by Krespan, E. in: JCI | PubMed | Google Scholar

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

Find articles by Beiting, D. in: JCI | PubMed | Google Scholar

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

Find articles by Radaelli, E. in: JCI | PubMed | Google Scholar | />

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

Find articles by Rui, H. in: JCI | PubMed | Google Scholar | />

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

Find articles by Koumenis, C. in: JCI | PubMed | Google Scholar

1 Department of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

2 Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

3 Department of Chemistry, School of Arts and Sciences, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

4 Department of Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA.

5 Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

6 Department of Cancer Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.

7 Department of Biochemistry, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA.

Address correspondence to: Serge Y. Fuchs, Dept. of Biomedical Sciences, School of Veterinary Medicine, University of Pennsylvania, 380 S. University Ave, Hill 316, Philadelphia, Pennsylvania 19104, USA. Phone: 1.215.573.6949 Email: [email protected]

Find articles by Fuchs, S. in: JCI | PubMed | Google Scholar | />

Intercellular biomolecule transfer (ICBT) between malignant and benign cells is a major driver of tumor growth, resistance to anticancer therapies, and therapy-triggered metastatic disease. Here we characterized cholesterol 25-hydroxylase (CH25H) as a key genetic suppressor of ICBT between malignant and endothelial cells (ECs) and of ICBT-driven angiopoietin-2–dependent activation of ECs, stimulation of intratumoral angiogenesis, and tumor growth. Human CH25H was downregulated in the ECs from patients with colorectal cancer and the low levels of stromal CH25H were associated with a poor disease outcome. Knockout of endothelial CH25H stimulated angiogenesis and tumor growth in mice. Pharmacologic inhibition of ICBT by reserpine compensated for CH25H loss, elicited angiostatic effects (alone or combined with sunitinib), augmented the therapeutic effect of radio-/chemotherapy, and prevented metastatic disease induced by these regimens. We propose inhibiting ICBT to improve the overall efficacy of anticancer therapies and limit their prometastatic side effects.

Within a multicellular organism, the horizontal exchange of biomolecules such as nucleic acids, polypeptides, lipids, and others between cells has emerged as an important mode of communication that encourages collective cell behavior ( 1 ). In a healthy organism, biomolecule exchange between diverse types of normal cells helps to maintain homeostatic balance and organize normal biologic processes (e.g., antigen presentation [refs. 2 , 3 ] or normal angiogenesis [ref. 4 ]). However, in an organism that bears a malignant tumor, this intercellular biomolecule transfer (ICBT) from cancer cells to normal cells often stimulates tumor growth, progression, and metastasis ( 5 – 8 ).

ICBT can occur through a multitude of mechanisms, including uptake of extracellular vesicles and apoptotic bodies, cell fusion, trogocytosis, trans-endocytosis, cell junction and tunneling nanotube formation, and others (reviewed in ref. 9 ). In the context of tumorigenesis, ICBT mediated by tumor-derived extracellular vesicles (TEVs) has emerged as a pivotal driver of pathogenesis and outcome of oncologic diseases ( 5 , 7 ). Biomolecules delivered by TEVs reprogram normal cells to contribute to many processes that promote tumor growth and progression, including modulation of metabolic activities, formation of a metastatic niche, suppression of immune responses, stimulation of angiogenesis, etc. ( 5 , 7 , 10 – 12 ).

Among other types of normal cells, endothelial cells (ECs) can become targets for ICBT. Intratumoral ECs in proximity to malignant cells were found to harbor genetic alterations similar to those found in the malignant cells, and ICBT between these cell types during cell fusion or efferocytosis has been proposed as a putative mechanism underlying this phenomenon ( 13 ). Subsequent studies confirmed transfer of tumoral DNA from malignant cells to ECs ( 14 ), demonstrated activation of ECs by TEVs ( 15 ), and established the paradigm supporting the important role of TEV-mediated ICBT in angiogenesis within tumor microenvironments (reviewed in refs. 5 , 8 , 16 ).

Importantly, TEV-mediated ICBT often undermines the efficacy of anticancer therapies ( 17 – 19 ). Furthermore, radio- and chemotherapeutic agents increase production and/or release of TEVs by malignant cells this increase is implicated in the iatrogenic metastatic disease arising from the treatment of primary tumors ( 20 – 25 ). Addressing these challenges requires a better understanding of genetic regulators of mechanisms underlying the protumorigenic ICBT. In addition, development of pharmacologic means to restrict ICBT should offer a novel approach to curtail tumor growth and progression and to improve the efficacy of existing anticancer therapies.

The biological barriers against ICBT are expected to preserve functional integrity of normal cells and restrict their cooperation with malignant cells. The type I interferon (IFN) pathway acts as one of those barriers that prevents generation of the prometastatic niche and pulmonary metastases ( 26 – 28 ). Cholesterol 25-hydroxylase (CH25H), an enzyme that is induced by IFN ( 29 – 31 ), acts to catalyze the formation of 25-hydroxycholesterol (25HC). This oxysterol inhibits lipid membrane fusion ( 32 ) and, accordingly, uptake of TEVs ( 26 ). Uptake of TEVs by normal cells was also shown to be inhibited by the antihypertensive agent reserpine, which in turn could restore CH25H expression otherwise downregulated by TEVs ( 26 ).

Here we characterize endothelial CH25H as a key genetic suppressor of ICBT and of ICBT-driven activation of ECs, as well as of intratumoral angiogenesis and tumor growth. Our data also demonstrate that the suppressive effect of CH25H on ICBT and ICBT-driven angiogenesis could be reenacted pharmacologically by administering reserpine, which, upon combination with different types of anticancer therapies, increases their efficacy and abolishes therapy-stimulated metastatic disease.

CH25H and reserpine inhibit ICBT between malignant cells and ECs. We previously reported that uptake of DiD dye–labeled TEVs is decreased in normal, wild-type (WT) splenocytes pretreated with reserpine or in splenocytes from knockin mice expressing a stabilized mutant of IFN receptor 1 (IFNAR1) ( 33 ). Importantly, the latter phenotype was lost upon ablation of Ch25h ( 26 ). Given that, in addition to TEVs, there are other mechanisms of biomolecule exchange such as uptake of apoptotic bodies, cell fusion, trans-endocytosis, etc. ( 9 ), we sought to determine the role of CH25H expression in benign cells in regulating ICBT within solid tumors in vivo.

To this end, we grew subcutaneous tumors from B16F10-TdTomato melanoma cells in the flanks of WT or Ch25h –/– mice that constitutively expressed green fluorescent protein (GFP) (Figure 1A). Tumors were harvested, dissociated, and analyzed for the numbers of TdTomato + GFP + double-positive cells among immune CD45 + immune and CD45 – nonimmune populations (Supplemental Figure 1, A and B supplemental material available online with this article https://doi.org/10.1172/JCI144225DS1). We found that tumors from the Ch25h –/– mice contained a greater number of TdTomato + GFP + cells in the CD45 – nonimmune stromal population (Figure 1B). Subsequent analysis of double-positive cells specifically in fibroblastic (PDGFRα + , Supplemental Figure 1C) and endothelial (CD31 + , Supplemental Figure 1D) compartments revealed that ECs are the main target for the CH25H-regulated exchange of biomolecules with malignant cells (Figure 1C). The extent of ICBT in the ECs from B16F10-TdTomato tumors growing in GFP + mice was notably decreased by in vivo administration of reserpine (Figure 1D).

CH25H and reserpine control ICBT between malignant cells and endothelial cells. (A) A schematic of experiments for assessing intratumoral ICBT in vivo. (B) Flow cytometric analysis of percentage of TdTomato + CD45 – and TdTomato + CD45 + cells in the tumor microenvironment (n = 5 for each group). (C) Flow cytometric analysis of the percentage (left) and absolute number (right) of TdTomato + CD31 + cells in tumors from GFP + WT and GFP + Ch25h –/– mice (n = 4–5 for each group). (D) Flow cytometric analysis of percentage (left) and absolute number (right) of TdTomato + CD31 + cells (n = 4 for each group) in tumors from GFP + WT and GFP + Ch25h –/– mice administered i.p. vehicle or reserpine (1 mg/kg given every other day for 4 days). (E) qPCR analysis of Gfp mRNA in primary WT and Ch25h –/– ECs after in vitro treatment with vehicle or reserpine (10 μM for 8 hours) followed by a 12-hour exposure to TEVs (20 μg/mL) isolated from GFP + B16F10 cells (n = 4 for each group). Data are presented as mean ± SEM. Statistical analysis was performed using 1-way ANOVA with Tukey’s multiple-comparison test (B, D, and E) or 2-tailed Student’s t test (C). NS, not significant. Experiments were performed independently at least 3 times.

We next examined ICBT in vitro mediated by isolated TEVs characterized in Supplemental Figure 1, E–G. To safeguard against possible peculiarities of the intratumoral ECs in the GFP-expressing mice, we isolated ECs from the lungs of naive mice (Supplemental Figure 1, H and I) and treated them with TEVs isolated from B16F10 cells stably expressing GFP. A greater amount of Gfp mRNA was transferred into CH25H-deficient ECs compared with WT ECs however, this phenotype was partially reversed upon treatment with reserpine (Figure 1E). Similar results were obtained when ICBT was assessed by transfer of DiD dye from TEVs into ECs (Supplemental Figure 1, J and K). Collectively, these results suggest that CH25H acts as a genetic suppressor of ICBT between malignant cells and ECs and characterize reserpine as a pharmacologic agent capable of inhibiting ICBT in vitro and in vivo.

Inactivation of stromal CH25H promotes tumor growth and angiogenesis. CH25H levels have been found to be decreased in leukocytes from tumor-bearing mice and melanoma patients compared with tumor-free control groups ( 26 ) however, the importance of CH25H downregulation in normal cells for tumor growth is not completely understood. Intriguingly, during the course of experiments described in Figure 1A, we noticed that subcutaneous TdTomato-expressing B16F10 melanoma tumors grew faster in Ch25h –/– compared with WT mice (Figure 2, A and B). Similar results were obtained with tumors formed by parental B16F10 cells (Supplemental Figure 2, A and B). This observation prompted us to examine the importance of CH25H expression in the tumor microenvironment for growth of other types of cancer cells.

Stromal CH25H restricts growth of solid tumors. (A) Growth of B16F10-TdTomato tumors (inoculated s.c. at 1 × 10 6 cells/mouse) in GFP + WT and GFP + Ch25h –/– mice. n = 4–5 for each group. (B) Representative images and quantification of tumor mass on day 15 from the experiment described in panel A. (C) Growth of MH6499c4 pancreatic ductal adenocarcinoma tumors (inoculated s.c. at 1 × 10 5 cells/mouse) in WT and Ch25h –/– mice. n = 12–13 for each group. (D) Growth of MC38 colon adenocarcinoma tumors (inoculated s.c. at 1 × 10 6 cells/mouse) in WT (n = 8) and Ch25h –/– (n = 13) mice. (E) Representative images and quantification of mass of MC38 colon adenocarcinoma tumors at 40 days after orthotopic inoculation of 5 × 10 5 cells into the cecum of WT (n = 5) or Ch25h –/– (n = 4) mice. (F) Representative images and quantification for in vivo luciferase analysis in male WT (n = 8) and Ch25h –/– (n = 9) mice, which were inoculated into prostatic glands with TRAMP-C2-luc prostate cells (1 × 10 6 ). Data are presented as mean ± SEM. Statistical analysis was performed using 2-way ANOVA with Tukey’s multiple-comparison test (A, C, D, and F) or 2-tailed Student’s t test (B and E). Experiments were performed independently at least 3 times.

An accelerated growth of syngeneic tumors in Ch25h –/– (compared with WT) mice was observed for transplanted MH6499c4 pancreatic ductal adenocarcinoma (Figure 2C) and MC38 colon adenocarcinoma (Figure 2D and Supplemental Figure 2C). Furthermore, this phenotype was not limited to subcutaneous tumors, as Ch25h –/– (compared with WT) mice exhibited a significantly faster growth of orthotopically transplanted MC38 colon tumors (Figure 2E) and TRAMP-C2-luciferase prostate neuroendocrine tumors (Figure 2F and Supplemental Figure 2D). Collectively, these data indicate that stromal CH25H inhibits growth of different types of solid tumors.

Visual appearance of tumors growing in Ch25h –/– mice was suggestive of a greater extent of vascularization (as compared with WT Figure 3A and Supplemental Figure 3A). This phenotype, as well as an increased ICBT involving ECs in Ch25h –/– mice (Figure 1), prompted us to examine the status of CH25H and angiogenesis in the stromal compartment of human colorectal cancers (CRCs).

The angiostatic role of CH25H in the tumor microenvironment. (A) A representative image of B16F10-TdTomato tumors and surrounding blood vessels in GFP + WT and GFP + Ch25h –/– mice. (B) Representative image of blood vessels (green) and CH25H (red) in normal stroma and colorectal cancer stroma. Scale bar: 100 μm. (C) Scatterplot of quantitative stromal CH25H protein expression levels in normal adjacent stroma and tumor stroma (Cohort 1). (D) Kaplan-Meier survival analysis of CRC stromal CH25H protein levels, dichotomized into high and low CH25H expression, indicating increased risk of recurrence with loss of CH25H protein levels (Cohort 2). (E) Scatterplot of quantitative CH25H protein levels within the endothelium of normal adjacent stroma and CRC tumor stroma (Cohort 3, left panel) and the validation of endothelial CH25H expression levels between paired samples of normal adjacent stroma and CRC stroma (Cohort 4, right panel). (F) Analysis of CD31 + ECs in B16F10 tumors (s.c., 1 × 10 6 cells/mouse) of comparable volume grown for approximately 2 weeks in WT (n = 4) and Ch25h –/– (n = 5) mice. Scale bar: 100 μm. (G) Quantification of data from experiment described in panel F. Data averaged from 5 random fields in sections from each of 4 or 5 animals are shown. (H) Representative image (left) of colocalization of blood vessels (red) and lectin + (green) area in tumor from WT and Ch25h –/– mice after injection with FITC-lectin (i.v., 100 μg/mouse). Quantification (right) of FITC-positive area (n = 5 for each group) of the images. Scale bar: 50 μm. (I) qPCR analysis of relative expression indicated genes in B16F10 (n = 5) and MC38 (n = 6) tumor tissues from WT and Ch25h –/– mice. For each gene, mRNA levels in WT tumors were defined as 1.0. Data are presented as mean ± SEM. Statistical significance was determined by 2-tailed Student’s t test (C, E, and GI) or log-rank (Mantel-Cox) test (D). Experiments were performed independently at least 3 times.

Analysis of tumors from an initial cohort of CRC patients (Cohort 1) revealed that levels of CH25H were significantly downregulated in the stromal compartment of malignant colorectal tumors compared with stroma of benign adjacent colon tissue (Figure 3, B and C). Importantly, survival analysis in a cohort of CRC patients with available clinical follow-up data (Cohort 2) revealed that low levels of CH25H within the tumor stroma were significantly associated with highly unfavorable prognosis (HR, 4.3 95% CI, 1.58–11.57 P = 0.004 Figure 3D). Moreover, costaining of CH25H and CD31 revealed that CH25H levels specifically in the CD31 + ECs were notably downregulated in ECs within tumor stroma compared with ECs from healthy colon stroma, as shown first in an analysis of unmatched cases (Cohort 3, Figure 3E). These findings were further validated in an independent set of CRC cases with matched cancer stroma and nearby normal colon stroma (Cohort 4, Figure 3E). These results provide strong correlative support for the notion that downregulation of CH25H occurs in human CRC stroma and, particularly, in the intratumoral ECs, and that this inactivation promotes tumor progression in human CRCs.

To validate these data from human patients in mouse models we examined the role of CH25H in development of the intratumoral vasculature. A greater number of CD31 + ECs was found in the B16F10 melanoma tumors grown in Ch25h –/– compared with WT mice (Figure 3, F and G). Furthermore, ablation of CH25H in the tumor microenvironment resulted in a greater number and increased length of blood vessels within these tumors (Figure 3G). Similar results were obtained in subcutaneous tumors of equal size formed by pancreatic adenocarcinoma cells (Supplemental Figure 3B) or MC38 colon adenocarcinoma cells (Supplemental Figure 3C), as well as in MC38 tumors transplanted orthotopically (Supplemental Figure 3D). Analysis of functionality of these blood vessels by injection of FITC-lectin revealed that knockout of CH25H increased perfusion in these tumors (Figure 3H). In all, these results suggest that inactivation of CH25H in the tumor microenvironment stimulates angiogenesis and increases tumor vascularization.

A limited screen for genes known to control tumor angiogenesis revealed a comparable level of Vegfa, Vegfr2, Tie2, Glut1, Mmp9, and Fgf1 mRNA in B16F10 tumors grown in WT and Ch25h –/– mice (Figure 3I). Intriguingly, we detected a relatively greater expression of angiopoietin-2 (Angpt2) in tumors from Ch25h –/– mice compared with those from WT animals. Angpt2 was also increased in MC38 tumors from Ch25h –/– mice (Figure 3I), further suggesting that inactivation of stromal CH25H stimulates the expression of Angpt2 in solid tumors.

Importantly, an increase in Angpt2 induced by TEVs and reversed by reserpine was also independently detected in an RNA sequencing–based profiling of gene expression in CH25H-deficient primary mouse lung ECs (Figure 4A). In addition to Angpt2, 812 out of 2998 differentially expressed genes were induced by TEVs unless pretreated with reserpine (Figure 4B). Previously implicated in transactivation of ANGPT2, transcription factors such as SP1, EGR1, GATA2, and ELF1 also exhibited altered expression after these treatments (Figure 4, A and B).

The ICBT-driven activation of endothelial cells is controlled by CH25H. (A) Volcano plot (upper) and Gene Ontology (GO, bottom) analyses of gene expression in Ch25h –/– mouse lung ECs treated as indicated. BP, Biological Process MF, Molecular Function CC, Cellular Component TF, transcription factors. (B) Heatmap analysis of gene expression from panel A. (C) Western blot analysis of TIE2 levels/phosphorylation in indicated ECs treated with MC38 TEVs (20 μg/mL for 12 hours). (D) qPCR analysis of Angpt2 expression (n = 3 ) in indicated ECs pretreated with vehicle or reserpine (10 μM for 8 hours) or cyclosporin A (0.25 μM for 24 hours) followed by PBS or TEVs (20 μg/mL for 12 hours). (E) ELISA analysis of ANGPT2 in supernatant of indicated ECs from panel D. (F) Tube formation by indicated ECs treated with VEGF165 (20 ng/mL) or MC38 tumor cell–conditioned media (TCM) or TCM –TEV (TEV-free tumor cell–conditioned media) or TCM with addition of anti-ANGPT2 neutralizing antibody as in Supplemental Figure 4C. Data averaged from 3 random fields in each of 5 wells were quantified. (G) Proliferation of indicated ECs exposed to MC38-derived TEVs (20 μg/mL) for 9 days. (H) Tube formation by indicated ECs treated (or not) with MC38-derived TEVs (20 μg/mL for 12 hours) in the presence or absence of anti-ANGPT2 antibody (60 ng/mL). Representative images (left) and quantified data (n = 5 for each group) averaged from 3 random fields in each of the 5 wells are shown. Scale bar: 100 μm. (I) Tube formation by Ch25h –/– ECs transduced with empty (Ctrl) or CH25H-expressing lentivirus (for 48 hours) or treated with vehicle or 25-hydroxycholesterol (25HC, 4 μM for 4 hours) and then exposed or not to MC38 TEVs (20 μg/mL for 12 hours). Data are presented as mean ± SEM. Statistical analysis was performed by 1-way ANOVA with Tukey’s multiple-comparison test (DF, H, and I) or 2-way ANOVA with Tukey’s multiple-comparison test (G). NS, not significant. Experiments were performed independently at least 3 times.

The TEV-induced increase in ANGPT2 expression and its reversal by reserpine was further validated by in vitro studies in primary mouse lung ECs. Under these conditions, treatment of Ch25h –/– cells with TEVs increased Angpt2 mRNA and protein to a greater extent compared with WT ECs (Supplemental Figure 4A), whereas Angpt1 and Tie2 were not induced (Supplemental Figure 4B). Furthermore, TEVs induced a more robust activation of TIE2 in Ch25h –/– ECs (Figure 4C). Importantly, pretreatment with reserpine reversed the TEV-induced increase in Angpt2 mRNA (Figure 4D) and protein (Figure 4E) in Ch25h –/– ECs. These results link ICBT in ECs with induction of ANGPT2.

The angiostatic and antitumorigenic roles of endothelial CH25H. ANGPT2 is produced by intratumoral ECs to facilitate angiogenesis (reviewed in refs. 34 – 36 ). Given that robust induction of Angpt2 in tumors from CH25H-deficient mice in vivo (Figure 3I) can be faithfully recapitulated in the cultures of primary mouse lung ECs (Figure 4, D and E), we next used this in vitro system to further interrogate the importance of CH25H in regulating angiogenic activity. Treatment with MC38 tumor cell–conditioned media elicited a greater increase in tube formation by the Ch25h –/– ECs, as compared with WT cells (Figure 4F and Supplemental Figure 4C). Importantly, we detected no difference in activities of WT and Ch25h –/– ECs treated with vehicle, tumor cell–conditioned media lacking TEVs, or with recombinant VEGF (Figure 4F and Supplemental Figure 4C). These results suggest that CH25H does not indiscriminately suppress activation of ECs but specifically limits their activation by ICBT.

Indeed, treatment with purified TEVs triggered a more robust in vitro activation of CH25H-deficient ECs compared with their WT counterparts. These phenotypes included an increased TEV-induced proliferation manifested by either cell numbers (Figure 4G) or percentage of Ki67-positive cells (Supplemental Figure 4D) and a greater migration (Supplemental Figure 4E) in Ch25h –/– ECs. Furthermore, compared with WT, Ch25h –/– ECs were more adept in forming endothelial tubes upon treatment with TEVs from MC38 (Figure 4H) or B16F10 (Supplemental Figure 4F) cancer cells. Importantly, neutralization of ANGPT2 prevented an increase in tube formation in response to tumor-conditioned media (Figure 4F) or purified TEVs (Figure 4H), indicating that ANGPT2 is required for the ICBT-induced activation of ECs. Furthermore, ICBT-driven hyperactivation of Ch25h –/– ECs could be effectively reversed by either reexpressing CH25H or by treating these cells with the end-product of the CH25H enzymatic activity — 25HC (Figure 4I). Collectively, these results suggest that CH25H acts as a biological barrier that restricts the ICBT-induced ANGPT2-dependent activation of ECs.

These results, together with the protumorigenic phenotypes observed in Ch25h –/– mice (Figure 2), prompted us to concentrate on specific roles of endothelial CH25H in vivo. To this end, we isolated ECs from the lungs of naive GFP + WT or Ch25h –/– mice and coinjected these cells with B16F10 cells into the flanks of WT mice. Under these conditions, at least some of the transferred GFP + ECs were incorporated into intratumoral blood vessels (Figure 5A). Whereas ECs of both genotypes accelerated tumor growth, Ch25h –/– ECs were significantly more active in these settings (Figure 5B and Supplemental Figure 5A), indicating that endothelial CH25H interferes with the ability of ECs to support tumor growth.

Endothelial expression of CH25H drives its angiostatic and antitumorigenic functions in vivo. (A) Representative images of colocalization of GFP-expressing Ch25h –/– ECs (green) with blood vessels (red). Scale bar: 20 μm. (B) Analysis of growth of tumors formed in WT host mice by B16F10 malignant cells (3 × 10 5 /mouse) coinjected s.c. with vehicle (n = 4) or with primary lung ECs (6 × 10 4 /mouse) isolated from WT or Ch25h –/– mice (n = 5 for both groups). (C) Analysis of growth of B16F10 tumors (s.c., 1 × 10 6 cells/mouse) in VE-cadherin–Cre + WT and VE-cadherin-Cre + Ch25h fl/fl mice (n = 5 for each group). (D) Representative images and B16F10 tumor mass analysis from experiment described in panel C (tumors harvested on day 17 after inoculation). (E) Representative immunofluorescence images of CD31 staining of B16F10 tumors from VE-cadherin–Cre + WT and VE-cadherin–Cre + Ch25h fl/fl mice (left) and quantification of CD31-positive areas (right). Quantification averaged from 5 random fields in sections from each of 5 animals is shown. Scale bar: 100 μm. (F) Quantification of average length and number of blood vessels (>50 μm) from experiment shown in panel E. Data averaged from 5 random fields in sections from each of 5 animals are shown. Data are shown as mean ± SEM. Statistical analysis was performed using 2-way ANOVA with Tukey’s multiple-comparison test (B and C) or 2-tailed Student’s t test (DF). Experiments were performed independently at least 3 times.

To further test this hypothesis, we generated a conditional knockout allele of Ch25h by flanking the sole exon of this gene with loxP sites (Supplemental Figure 5B). We crossed these mice with animals that express Cre recombinase under the EC-specific VE-cadherin promoter ( 37 ) the resulting animals lacked CH25H in primary lung ECs but not fibroblasts (Supplemental Figure 5C). Whereas no obvious vascular alterations or other abnormalities were detected in these naive animals, they displayed a notable phenotype when challenged with subcutaneous tumors. Under these conditions, efficient ablation of CH25H was achieved in the intratumoral ECs (Supplemental Figure 5D). Importantly, growth of B16F10 melanoma tumors was significantly accelerated in mice lacking CH25H in the ECs (Figure 5, C and D). Furthermore, ablation of CH25H in ECs notably increased intratumoral angiogenesis, as evident from an increased CD31 + area as well as a greater number of blood vessels and an increase in their length (Figure 5, E and F). Similar observations were made in experiments involving MC38 colon adenocarcinoma tumors (Supplemental Figure 5, E–H). In all, these results suggest that endothelial CH25H plays important angiostatic and antitumorigenic roles.

Reserpine inhibits tumor angiogenesis and improves the outcome of radio- and chemotherapies. We have previously reported that treatment with reserpine increases the expression of CH25H in TEV-treated cells ( 26 ). Thus, we utilized this agent as a complementary pharmacologic tool to ascertain the importance of ICBT-driven endothelial activation and tumor angiogenesis. Used at previously described low doses ( 26 ), reserpine was well tolerated and did not cause sleepiness or decrease animal weight while inhibiting the intratumoral ICBT (Figure 1, D and E). Furthermore, in vitro pretreatment of ECs with reserpine suppressed TEV-induced expression of Angpt2 (Figure 4, B, D, and E) and significantly inhibited endothelial tube formation (Supplemental Figure 6A). These results prompted us to examine the effect of reserpine on tumor angiogenesis in vivo.

Administration of reserpine did not decrease expression of Angpt1 or Tie2 in the B16F10 tumors growing in WT or Ch25h –/– mice (Supplemental Figure 6B). However, this treatment abolished an increase in expression of Angpt2 (Figure 6A) and notably suppressed angiogenesis (Figure 6B) in tumors from Ch25h –/– mice. Similar results were also obtained in animals bearing MC38 colon tumors (Supplemental Figure 6C), indicating that reserpine exhibits a robust angiostatic effect in solid tumors.

Angiostatic and antitumorigenic effects of reserpine in solid tumors. (A) qPCR analysis of relative Angpt2 mRNA levels in B16F10 tumors from WT and Ch25h –/– mice treated with vehicle or reserpine (1 mg/kg, i.p. every other day for 4 days). n = 5 for each group. (B) Representative immunofluorescence images and quantification of CD31-positive areas in B16F10 tumors from WT and Ch25h –/– mice (n = 5 for each group) treated with vehicle or reserpine as described in panel A. Scale bar: 100 μm. (C) Growth of human HCT116 tumors (inoculated s.c. at 5 × 10 6 cells/mouse) in NSG mice treated with vehicle or reserpine (1 mg/kg) every other day. n = 5 for each group. (D) Representative immunofluorescence image (upper) of CD31 staining of HCT116 tumors from NSG mice treated with vehicle or reserpine. Quantification (bottom) of CD31-positive areas and average distance of blood vessels. Quantification averaged from 5 random fields in sections from each of 5 animals is shown (n = 5 for each group). Scale bar: 100 μm. (E) Analysis of B16F10 tumor growth (inoculated s.c. at 1 × 10 6 cells/mouse) in WT and Ch25h –/– mice (n = 5 for each group) followed by vehicle or reserpine treatment (1 mg/kg, i.p. every other day). (F) Analysis of B16F10 tumor mass on day 15 of the experiment described in panel E. (G) Analysis of B16F10 tumor mass on day 15 after inoculation (s.c. at 1 × 10 6 cells/mouse) into indicated mice (n = 5 for each group) followed by vehicle or reserpine treatment (1 mg/kg, i.p. every other day). Data are presented as mean ± SEM. Statistical analysis was performed using 1-way ANOVA with Tukey’s multiple-comparison test (A, B, F, and G), 2-way ANOVA with Tukey’s multiple-comparison test (C and E), or 2-tailed Student’s t test (D). NS, not significant. Experiments were performed independently at least 3 times.

Suppressive effects of reserpine on angiogenesis and growth were seen in human HCT116 colon tumors growing in the immune-deficient mice (Figure 6, C and D). Likewise in the immune-competent settings, administration of reserpine inhibited B16F10 tumor growth in both WT and Ch25h –/– mice (Figure 6, E and F). This effect was also observed in mice harboring the VE-cadherin–Cre allele in the Ch25h +/+ or Ch25h fl/fl background (Figure 6G) as well as in Ch25h –/– mice inoculated with MC38 tumors (Supplemental Figure 6, D and E).

Expression of Angpt2 is thought to undermine the efficacy of antiangiogenic therapies targeting VEGF (bevacizumab) or its receptor (sunitinib) ( 35 , 38 – 41 ). Given that reserpine suppresses the ICBT-induced production of ANGPT2 (Figures 4 and 6A), we hoped to optimize the antiangiogenic therapy by combining these agents (Figure 7A). Indeed, adding reserpine to the sunitinib regimen decreased plasma levels of AngpT2 (Supplemental Figure 7A), maximized the suppression of angiogenesis (Figure 7B), and augmented the antitumorigenic effects in mice bearing MC38 (Figure 7, C and D) or B16F10 tumors (Supplemental Figure 7B). These effects of reserpine were similar to those of rebastinib, a bona fide inhibitor of the ANGPT2 pathway (Supplemental Figure 7, C–G). Collectively, these results indicate that reserpine alone or in combination with agents targeting VEGF can be used as a pharmacologic agent to suppress the intratumoral angiogenesis and inhibit growth of primary solid tumors.

Combination of reserpine and antiangiogenic therapy. (A) Schematic of treatment of MC38 tumor–bearing mice with sunitinib, reserpine, or their combination. (B) Representative immunofluorescence images and quantification of angiogenesis parameters in MC38 tumors from WT mice treated as in panel A (n = 5 for each group). Scale bar: 100 μm. (C) Analysis of MC38 tumor volume in WT mice treated as in panel A (n = 5 for each group). (D) Analysis of mass of MC38 tumors in WT mice on day 25 of the experiment described in panel C (n = 5 for each group). Data are presented as mean ± SEM. Statistical analysis was performed using 1-way ANOVA with Tukey’s multiple-comparison test (B and D) or 2-way ANOVA with Tukey’s multiple-comparison test (C). NS, not significant. Experiments were performed independently at least 3 times.

Reserpine was previously included in the screen for agents affecting TEV uptake ( 26 ) based on its activity as a vesicular reuptake inhibitor (reviewed in ref. 42 ). Intriguingly, in addition to inhibiting TEV uptake by splenocytes or ECs (ref. 26 and Figure 1E and Supplemental Figure 1, J and K), we noticed a significant suppressive effect of reserpine on both protein content and numbers of TEVs produced by B16F10 or MC38 cancer cells in vitro (Figure 8A). Furthermore, treatment of MC38 cells with reserpine decreased the expression of several genes including Rab11b, Rab27a, Ykt6, and Snap23 (Figure 8B), which play an important role in formation and/or uptake of extracellular vesicles ( 43 , 44 ).

Mechanism of reserpine-mediated TEV uptake inhibition. (A) Quantification of numbers (upper panel) and total protein content (bottom panel) of TEVs released by the indicated cells following treatment with vehicle or reserpine (10 μM for 48 hours) in vitro. (B) qPCR analysis of Rab7, Rab11b, Rab27a, Sdcbp, Arf6, Ykt6, Snap23, Hgs, and Pdcd6ip relative levels (n = 4 for each group) in MC38 cancer cells treated with vehicle or reserpine (10 μM for 12 hours). (C) Analysis of plasma membrane polarization in WT and Ch25h –/– ECs treated with reserpine (10 μM) for 12 hours. (D) Flow cytometric analysis of percentage of DiD + CD31 + cells upon incubation of indicated ECs with DiD-labeled liposomes (1 μg/mL) in the presence or absence of reserpine (10 μM) for 8 hours (n = 5 for each group). Data are presented as mean ± SEM. Statistical analysis was performed using 1-way ANOVA with Tukey’s multiple-comparison test (A, C, and D) or 2-tailed Student’s t test (B). NS, not significant. Experiments were performed independently at least 3 times.

Preincubation of DiD-labeled TEVs with reserpine did not affect transfer of DiD into ECs (Supplemental Figure 8A), suggesting that ECs themselves are targets for reserpine-elicited uptake inhibition. Importantly, given that reserpine upregulates CH25H expression ( 26 ) and following an analogy with 25HC, which alters membrane fluidity and inhibits lipid membrane fusion ( 32 ), we analyzed the effects of reserpine on membrane fluidity using 1,6-diphenyl-1,3,5-hexatriene as a sensor of the bilayer membrane structure alterations. Polarization of this biosensor was increased by reserpine (Figure 8C), indicating that reserpine can increase the rigidity of EC membranes. This possibility was further tested by studies involving transfer of DiD dye to ECs from liposomes that lack any protein molecules on their surface reserpine suppressed this transfer (Figure 8D and Supplemental Figure 8B). These results collectively suggest that mechanisms of ICBT inhibition by reserpine include its effect on the EC lipid membranes.

A yet greater impetus for the use of reserpine to suppress ICBT came from the observations that chemo- and radiotherapies activate the formation and release of prometastatic TEVs by primary tumors ( 24 , 25 ) this activation contributes to resistance ( 5 , 7 , 19 , 45 ) as well as to therapy-triggered stimulation of metastatic disease ( 22 – 24 ). We next sought to determine whether reserpine addition can improve efficacies of chemo-/radiotherapies while preventing their negative prometastatic effects.

Consistent with suppression of TEV production in vitro (Figure 8A), in vivo administration of reserpine decreased the amount of the exosomal marker CD63 in the plasma from the B16F10 or MC38 tumor–bearing mice subjected or not to ionizing radiation or chemotherapy treatment (Figure 9A and Supplemental Figure 9A). Likewise, reserpine treatment abolished an increase in the numbers of TEVs found in plasma of mice subjected to chemotherapy (Supplemental Figure 9A). Given that reserpine can also inhibit TEV uptake ( 26 ), these data collectively suggest that reserpine can suppress ICBT through multiple mechanisms.

Administration of reserpine improves the outcomes of radio-/chemotherapies. (A) ELISA analysis of CD63 levels in the plasma from mice bearing B16F10 tumors of similar volume exposed or not to ionizing radiation (IR, 12 Gy), reserpine (1 mg/kg), or both. n = 4 for all groups. (B) Analysis of tumor volume in B16F10 tumors (inoculated s.c. at 1 × 10 5 cells/mouse) in WT mice treated with vehicle or reserpine (1 mg/kg) upon reaching 30 to 50 mm 3 . Three days later, vehicle- and reserpine-treated mice with similar tumor volumes underwent 12-Gy irradiation and continued their assigned vehicle or reserpine treatments 3 times per week (n = 4). (C) Kaplan-Meier survival analysis of B16F10 tumor–bearing mice treated as described in panel B until tumors reached 2000 mm 3 (n = 4). (D) Quantification of total area of B16F10 metastatic load in lungs from mice described in panel B (n = 4). (E) Schematic of the experiments combining reserpine with FOLFOX treatment of orthotopically inoculated MC38 colon tumors. (F) Representative images and the mass of MC38 tumors from animals treated as in panel E. n = 5 for each group. (G) Representative images of livers from MC38 tumor–bearing mice described in panel E. Arrowheads show macroscopic metastatic lesions found in 8% of vehicle-treated animals, 57% of FOLFOX only–treated animals, and none of the animals that received reserpine (with or without FOLFOX). Data are shown as mean ± SEM. Statistical analysis was carried out using 1-way ANOVA with Tukey’s multiple-comparison test (A, D, and F), 2-way ANOVA with Tukey’s multiple-comparison test (B), or log-rank (Mantel-Cox) test (C). NS, not significant. Experiments were performed independently at least 3 times.

Combining reserpine treatment with radiation therapy augmented the inhibition of subcutaneous primary B16F10 tumor growth and prolonged animal survival (Figure 9, B and C). Importantly, reserpine also dramatically suppressed the appearance of distant metastatic melanoma lesions in the lungs and prevented an increase in the number of these lesions in response to irradiation of primary tumors (Figure 9D and Supplemental Figure 9B).

The effects of reserpine were next determined in the chemotherapy setting, in which orthotopically inoculated MC38 colon adenocarcinomas were treated with the FOLFOX regimen (Figure 9E). This regimen, which includes oxaliplatin, folinic acid, and 5-fluorouracil, is standard for treatment of patients with metastatic colorectal cancers and is often combined with antiangiogenic agents ( 46 ). Intriguingly, 5-fluorouracil was also implicated in stimulation of postliminary metastatic disease ( 47 ) and in the induction of genes associated with poor prognosis in a cohort of patients that received FOLFOX ( 48 ). In our preclinical model, the inclusion of reserpine notably decreased the intratumoral angiogenesis otherwise stimulated by the FOLFOX regimen (Supplemental Figure 9, C and D). Importantly, while being well tolerated (ref. 26 and Supplemental Figure 9E), reserpine also potentiated therapeutic effects of FOLFOX on primary tumors (Figure 9F), robustly suppressed metastases into liver (Figure 9G), and decreased the number of local lesions in the gut (Supplemental Figure 9F). Collectively, these data provide the rationale for the inclusion of reserpine into the radio-/chemotherapy regimens to increase their overall efficacy and, most importantly, to prevent the prometastatic side effects of such treatments.

Endothelial CH25H as a genetic suppressor of ICBT-driven intratumoral angiogenesis. Given the paramount role of ICBT in tumor growth, progression, and resistance to therapies ( 5 – 7 ), it is of critical importance to understand how ICBT is regulated and could be pharmacologically controlled in vivo. An increased ICBT between malignant cells and CH25H-deficient ECs (Figure 1) might be reflective of ECs representing the first barrier encountered by the circulating tumor-derived matter. CH25H is induced by IFN, and downregulation of its receptor often occurs in many types of cancer ( 28 , 49 – 51 ) and contributes to angiogenic activation by VEGF ( 52 ). Importantly, the loss of CH25H per se does not stimulate ECs however, CH25H-deficient ECs are more sensitive to TEV-induced activation of ECs. Although the importance of CH25H in regulating ICBT in other cell types should not be ruled out, our data specifically characterize endothelial CH25H as a key suppressor of the ICBT-induced upregulation of ANGPT2, activation of ECs, and intratumoral angiogenesis.

CH25H catalyzes monooxygenation of cholesterol into 25HC. Its absence may promote formation of other types of oxycholesterols such as proangiogenic 27-hydroxycholesterol ( 53 ). Nevertheless, evidence demonstrating the ability of 25HC to directly suppress tube formation (Figure 4I) and EC proliferation ( 54 ) suggests a key role for 25HC in the angiostatic phenotype. It is important to note that 25HC inhibits lipid membrane fusion ( 32 ), which is essential for cell fusion, uptake of TEVs and apoptotic bodies, cell junction and tunneling nanotube formation, and other events enabling ICBT (reviewed in ref. 9 ). Whereas the importance of CH25H and 25HC in restricting the uptake of TEVs has been demonstrated ( 26 ), future studies will focus on determining specific roles of these regulators in other mechanisms of ICBT.

Numerous mediators of angiogenesis delivered to ECs by extracellular vesicles from either normal or malignant cells include VEGF-A and -D, WNT4, IL-8, carbonic anhydrase 9, diverse types of noncoding RNAs, and others ( 55 – 62 ). It is likely that more than one type of these diverse biomolecules is redundantly responsible for increased ANGPT2 and resulting angiogenesis in the absence of endothelial CH25H. Under these conditions, induction of ANGPT2 upon inactivation of CH25H is likely triggered by an increased extent of ICBT as well as by additional CH25H-dependent mechanisms that affect TEV-sensitive and/or reserpine-sensitive transcription factors such as SP1, EGR1, ELF1, and GATA2 and related pathways uncovered in our study (Figure 4, A and B). Intriguingly, cyclosporin A, an inhibitor of the calcineurin/NFAT/ANGPT2 signaling axis ( 63 ), also had a modest effect on ANGPT2 induction by TEVs (Figure 4D), suggesting that diverse and likely redundant pathways mediate the effects of ICBT.

It is likely that the importance of endothelial CH25H for suppressing tumor angiogenesis and growth gleaned from experiments using knockout models is seriously underestimated because of notable decreases in CH25H levels in the intratumoral ECs, as detected in human CRC patients (Figure 3, B–D). This downregulation of CH25H in the tumor microenvironment was significantly associated with poor prognosis (Figure 3D) and could be mediated by downregulation of CH25H expression in response to TEVs ( 26 ) as well as by additional, yet to be delineated mechanisms.

Reserpine as a pharmacologic inhibitor of ICBT and, potentially, a component of anticancer therapies. Importantly, the ICBT-driven ANGPT2 expression, EC activation, and intratumoral angiogenesis can be virtually nullified by treatment with reserpine (Figures 4 and 6). Given that production of ANGPT2 is implicated in clinical refractoriness to the antiangiogenic therapies involving bevacizumab ( 38 , 39 ), reserpine can be employed to specifically suppress the ICBT-stimulated ANGPT2-dependent intratumoral angiogenesis and complement agents targeting the VEGF pathway, similarly to TIE2 pathway inhibitors such as rebastinib (Figure 6 and Supplemental Figure 6).

Perhaps even more exciting is the potential for clinical use of reserpine to improve the efficacy of other types of anticancer treatment, including chemo- and radiotherapies supported by the preclinical data presented here (Figure 9). Reserpine has been widely used as a drug for treatment of hypertension and has also been chosen for its ability to inhibit vesicular monoamine transporter-2 and vesicular reuptake ( 42 ). We have previously reported that treatment with reserpine increases the expression of CH25H in TEV-treated cells ( 26 ). Here we show that reserpine limits the ICBT between malignant and benign cells in vivo (Figure 1) by likely more than one mechanism, including suppression of TEV uptake ( 26 ) and a decrease in circulating TEVs (Figure 9A and Supplemental Figure 9A), which can be plausibly attributed to altered TEV production/loading in vitro (Figure 8A). Molecular mechanisms underlying the inhibition of ICBT by reserpine are likely to involve effects on expression of genes involved in production of extracellular vesicles and on plasma membrane fluidity and its ability to fuse with lipid membrane vesicles (Figure 8, B–D). Potential effects of reserpine on additional modes of ICBT (e.g., cell fusion or formation of tunneling nanotubes) require further studies.

Although reserpine elicited only a modest effect on proliferation of malignant cells in vitro ( 26 ), additional ICBT-independent mechanisms by which this agent may help to suppress angiogenesis and growth of solid tumors cannot be excluded. Regardless of these specific mechanisms, it is important to note that reserpine robustly suppressed the increase in metastatic disease triggered by therapeutic regimens mostly designed to efficiently eradicate primary tumors (Figure 9).

The relatively low cost of reserpine therapy adds to its benefits, which could be especially important to battle metastatic disease in economically disadvantaged patients worldwide. Given that reserpine is well tolerated and has already been approved for use in human patients, this drug can be immediately tested in clinical trials for inclusion in standard regimens used in the treatment of solid tumors. Furthermore, future development of novel means to control ICBT and improve the outcome of anticancer therapies hold additional promise.

Detailed methods can be found in the supplemental material.

Human CRC specimens and their analyses. Human CRC tissue microarrays, consisting of formalin-fixed, paraffin-embedded tissue cores, were stained for CH25H. The detailed description of Cohorts 1–4 is provided in the supplemental material. CH25H, CD31, and cytokeratin detection was performed using immunofluorescence and immunohistochemistry, as previously described ( 64 ). Quantitative biomarker analysis was performed using Tissue Studio image analysis software (Definiens) to identify epithelial or stromal regions, facilitated by DAPI-stained cell nuclei and cytokeratin-stained cancer cells.

Animal studies. Besides NSG, all other mouse strains (including WT, Ch25h –/– , TgN(ActbEGFP)1Osb/J, and VE-cadherin–Cre ( 65 ) purchased from The Jackson Laboratory were on the C57BL/6J background. The conditional Ch25h allele was created by flanking the single exon of the Ch25h gene with loxP sites inserted into the nonconservative regions (

1.8 kb upstream of exon 1 and

0.5 kb downstream of exon 1). Targeting vector, homology arms, and the conditional knockout region were generated by PCR using BAC clones RP23-392N3 and RP24-61K11 from the C57BL/6J library as the templates. In the targeting vector (Supplemental Figure 5B), the Neo cassette was flanked by the self-deletion anchor sites and diphtheria toxin A was used for negative selection. C57BL/6J ES cells were used for gene targeting. Other mouse strains were generated by intercrossing male and female 6- to 8-week-old littermates were used in all experiments.

Cell culture. Human 293T and HCT116 cells and mouse B16F10 cells were purchased from ATCC. Mouse MC38 colon adenocarcinoma (from S. Ostrand-Rosenberg, University of Maryland, Baltimore, Maryland, USA), TRAMP-C2-luc prostate neuroendocrine tumor cells (from L. Languino, Thomas Jefferson University, Philadelphia, Pennsylvania, USA), and MH6499c4 pancreatic ductal adenocarcinoma (from B. Stanger, University of Pennsylvania, Philadelphia, Pennsylvania, USA) were gifted. All cells were maintained at 37°C with 5% CO2 in DMEM supplemented with 10% heat-inactivated fetal bovine serum (FBS), 100 U/mL penicillin-streptomycin, and L-glutamine.

TEV isolation, characterization, and assessment of uptake and production were carried out as previously described ( 10 , 26 ). Briefly, TEVs were collected from (extracellular vesicle–free) media (Gibco, 11965-084) supernatants by ultracentrifugation and additional purification was carried out using discontinuous iodixanol gradients. TEVs were spun over 10%/20%/30% iodixanol layers at 350,000g (52,000 rpm with Beckman SW 55 Ti rotor), 4°C, for 120 minutes. Then, 10 fractions of 260 μL each were collected starting from the top of the tube, diluted with 1 mL PBS, and resedimented at 100,000g and 4°C for 70 minutes (Beckman Optima Ultracentrifuge and TLA-100.2 rotor at 53,000 rpm). Pellets in each fraction were resuspended in PBS and characterized for density by measuring the weight of each fraction (g/mL, indicated in Supplemental Table 1). Fractions 1 and 2 (1.11–1.12 g/mL) were the only ones containing TEVs.

Quantitative real-time PCR. Total RNA from ECs or tumors was extracted using TRIzol reagent (Life Technologies, 15596018) and analyzed by quantitative real-time PCR using SYBR Green Master Mix (Applied Biosystems, 4367659). Primers of indicated genes are listed in Supplemental Table 2.

TEV and liposome uptake in vitro was examined in the WT and Ch25h –/– ECs cultured with TEV-free media and pretreated with vehicle or reserpine (diluted as previously described in ref. 26 ), followed by either DiD-labeled liposomes (FormuMax, F60103F-DD, 1 μg/mL) or TEVs that were either labeled with DiD as described previously ( 26 ) or derived from B16F10 cells stably expressing GFP. Uptake of DiD was monitored by flow cytometry as described previously ( 26 ). Uptake of Gfp mRNA was analyzed after total RNA was isolated from ECs using TRIzol reagent and chloroform. RNA concentration and purity were determined by using a NanoDrop spectrophotometer (Thermo Fisher Scientific). An Applied Biosystems High-Capacity RNA-to-cDNA Kit was used to make cDNA. The Gfp mRNA level was measured by quantitative real-time PCR.

For assessment of TEV production by tumor cells in vitro, 6 × 10 6 cells were plated in 15-cm dishes. Upon attachment of all cells, the media were removed and replaced with fresh media that contained 10% extracellular vesicle–free FBS. The cells were treated with vehicle (DMSO) or reserpine (10 μM). After 2 days, the conditioned media were collected for extracellular vesicle isolation and the total number of cells were counted for each condition. Ten microliters of isolated extracellular vesicles was submitted for nanoparticle tracking analysis (NTA), and 10 μL was used for protein concentration. The number of vesicles per microliter and the amount of protein per vesicle were calculated correspondingly.

For assessment of TEV absolute number in plasma from tumor-bearing mice undergoing chemotherapy, 100 μL of plasma was harvested and TEVs were isolated with an exosome isolation kit (Invitrogen, 4485229). Pellets were resuspended in 50–80 μL PBS and samples were assessed by NTA.

Flow cytometric analysis of ICBT and other immunological techniques. Volume of tumors measured by caliper was calculated as width × width × length × 0.5. Tumor tissues were dissected and digested with 1 mg/mL Collagenase D (Roche, 11088882001) with 100 μg/mL DNase I (Roche, 10104159001) in RPMI medium with 2% FBS for 1 hour with continuous agitation at 37°C. The digestion mixture was passed through a 70-μm cell strainer to prepare a single-cell suspension and washed with PBS supplemented with 2 mM EDTA and 1% FBS. Single cells were stained with cell-surface antibodies: anti-CD45–APC-Cy7 (BioLegend, catalog 103115), anti-CD31–PE-Cy7 (BioLegend, catalog 102417), and anti-PDGFRα–APC (BioLegend, catalog 135907). Data were acquired using an LSRFortessa flow cytometer (BD Biosciences) and analyzed with FlowJo software (Tree Star).

Enzyme-linked immunosorbent assay–based kits from Boster Bio, LLC (for ANGPT1 and ANGPT2) or Cusabio Technology, LLC (for CD63) were used to analyze levels of respective proteins in serum, tumor homogenates, or supernatants of EC cultures according to the manufacturers’ instructions.

For the immunofluorescence analyses, tumor tissues were harvested, embedded in frozen OCT, and cryosectioned into 7-μm sections using a Leica CM3050 S cryostat. Tumor sections were fixed in acetone, washed with PBS, and then blocked with PBS containing 5% goat serum and 1% BSA. Then, the sections were incubated for 12 hours at 4°C with anti-CD31 primary antibody (BD Biosciences, catalog 553370) diluted 1:200 in PBS, and followed by washing 3 times with PBS. Next, samples were incubated at room temperature for 1 hour with goat anti-rabbit secondary antibody (Invitrogen, catalog A-11006 or A-11007) diluted 1:500 in PBS. ProLong Gold Antifade reagent with DAPI (Invitrogen, P36935) was added after washing. Immunofluorescence images were captured with an Olympus BX51 microscope. CD31 staining area, distance, and number of blood vessels were analyzed with Metaphor software (Molecular Devices).

Isolation, culture, and proliferation/migration analyses of primary lung ECs was carried out as previously described ( 63 ). For Western blotting analysis of TIE2 phosphorylation, WT and Ch25h –/– ECs were isolated from respective mice and treated with PBS or MC38-derived vesicles (20 μg/mL) for 12 hours. Western blotting was carried out using a rabbit anti–mouse TIE2 polyclonal antibody (Thermo Fisher Scientific, catalog PA5-80103) and p-TIE2 was detected with rabbit anti–p-TIE2 antibody (Cell Signaling Technology, catalog 4221S), as described previously ( 66 ).

Analysis of plasma membrane fluidity was carried out as previously described ( 67 ). ECs treated as indicated were suspended in PBS at 4 × 10 5 /mL and incubated with the fluorescent probe 1,6-diphenyl-1,3,5-hexatriene (Sigma-Aldrich, D208000 3 μM) at 37°C for 20 minutes and held at 25.0 ± 0.5°C for intensity polarization measurements using a Tecan Infinite F200 Fluorescence Microplate Reader System (λex = 313 nm λem = 460 nm). The degree of cell membrane polarization was calculated using P = (F1F2)/(F1 + F2), in which F1 and F2 refer to the fluorescence intensity of vertically and horizontally polarized components, respectively, with excitation vertically polarized.

Tumorigenesis studies. For the syngeneic subcutaneous tumor model, B16F10, MC38, or MH6499c4 tumor cells were inoculated into the right flank of indicated C57BL/6J mice. For xenograft studies, human HCT116 cells (5 × 10 6 ) were injected into NSG mice (The Jackson Laboratory). Studies using the orthotopic colon tumor growth cancer model were carried out as previously described ( 68 ). Briefly, after anesthesia, a 1.5-cm nick was made in the abdominal wall (around) and the cecum was exteriorized and kept moist using PBS. Twenty-five microliters of the MC38 cell suspension (2 × 10 7 /mL) was injected into the cecal wall using a 30-G needle and the injection site was covered with a cotton swab for 3 minutes to monitor for leakage. The cecum was gently returned to the abdominal wall, and then the abdominal wall and skin were sutured carefully. For the orthotopic prostatic cancer model, 1 × 10 6 TRAMP-C2-luc/GFP cells were injected into the prostate of WT or Ch25h –/– mice. Tumor volumes were tracked via detecting bioluminescence intensity weekly.

FITC-lectin perfusion. MC38 cells (1 × 10 6 ) were injected into the right flank of WT and Ch25h –/– mice. Fourteen days after injection, mice were anesthetized, injected with FITC-conjugated Lycopersicon esculentum (tomato) lectin (FITC-lectin) (Thermo Fisher Scientific, L32478 100 μg/mouse, i.v.) and allowed to circulate for 10 minutes. After that, the chest was opened rapidly and the vasculature was perfused with 30 mL of 4% paraformaldehyde (PFA) for 5 minutes. Tumor tissues were harvested and stored in PFA overnight before being frozen in OCT. Cryosections of tumors were stained with anti-CD31 for the whole blood vessels and FITC-positive areas were calculated with Metaphor software.

Combination therapies. Reserpine (Sigma-Aldrich, 83580) was administered as previously described ( 26 ). Briefly, reserpine (dissolved in 0.1% ascorbic acid and diluted in ddH2O) or vehicle (0.1% ascorbic acid diluted in ddH2O) was administered to B16F10 tumor–bearing mice when the tumors reached 75 mm 3 at the dose of 1 mg/kg (i.p. 3 times per week). Matched vehicle and reserpine mice with similar (100–130 mm 3 ) tumor volumes were chosen to undergo irradiation using the Small Animal Radiation Research Platform (SARRP, Xstrahl Medical & Life Sciences) 2 days later. Mice were anesthetized using inhaled 2.5% isoflurane and placed on the stage of the SARRP. Once the tumor isocenter was determined, delivery of the single 12-Gy dose was made using a 1 × 1 cm collimated beam operating at 175 kV, 15 mA, with copper filtration and the dose rate at 1.65 Gy/min. The beam was delivered at such an angle as to avoid the spine. Dosimetry was performed using EBT2 gafchromic films.

The ingredients for the FOLFOX regimen (oxaliplatin, PHR1528 5-fluorouracil, F6627 folinic acid calcium salt hydrate, F7878 — all Sigma-Aldrich) were dissolved in PBS and administered (150 mg/kg folinic acid, 5 mg/kg 5-fluorouracil, 1.4 mg/kg oxaliplatin all i.p. every other day, with or without 1 mg/kg reserpine i.p.) into mice 10 days after the animals were inoculated with 5 × 10 5 MC38 cells injected into the cecum. All mice were sacrificed 45 days after tumor inoculation and tumor, liver, and intestine were harvested for the histopathologic analysis. Sunitinib (BioVision, 1611) was dissolved in a vehicle (composed of carboxymethylcellulose sodium [0.5% w/v], NaCl [1.8% w/v], Tween 80 [0.4% w/v], and benzyl alcohol [0.9% w/v] in water the whole formulation was adjusted to pH 6.0). Rebastinib was dissolved in 0.4% hydroxypropyl methylcellulose. MC38 or B16F10 cells (5 × 10 5 ) were inoculated into the right flank of WT mice. Nine days after tumor inoculation, mice were treated with sunitinib (40 mg/kg, gavage) 3 times per week with or without reserpine (1 mg/kg, i.p.) every other day or rebastinib (20 mg/kg, gavage) twice per week.

RNA sequencing. Primary lung ECs from Ch25h –/– mice were pretreated with vehicle or reserpine (10 μM for 8 hours) followed by treatment with MC38 TEVs (20 μg/mL) or PBS for 12 hours in vitro and total RNA was isolated using the RNeasy Plus Mini Kit (QIAGEN) and analyzed for Angpt2 mRNA levels by qPCR. These samples were then used for RNA sequencing (carried out as previously described, ref. 69 ). Raw reads were mapped to the mouse reference transcriptome (Ensembl) using Kallisto version 0.46.0. Raw data are available in the NCBI’s Gene Expression Omnibus database (GEO GSE163941). All subsequent analyses were carried out using the statistical computing environment R version 4.0.0 in RStudio and Bioconductor version 3.11.1 as described in the supplemental material.

Statistics. All experiments described here are representative of at least 3 independent experiments (n > 5 mice for each group unless otherwise specified). For in vitro experiments, cells or tissues from each animal were incorporated in triplicate. All data are shown as mean ± SEM. Statistical analyses were conducted using GraphPad Prism 7 software. Comparisons between 2 groups were conducted with a 2-tailed Student’s t test and multiple comparisons were performed using 1-way ANOVA or 1-way ANOVA followed by Bonferroni’s post hoc test. Tumor growth curve analysis was conducted with repeated-measures 2-way ANOVA (mixed model) followed by Bonferroni’s post hoc test. Kaplan-Meier curves were used to analyze the survival data, and Cox regression was used to compute hazard ratios. P values of less than 0.05 were considered significant.

Study approval. Use of preexisting human archival decodified and deidentified CRC tissue arrays, previously collected under informed consent, and samples that could not be directly or indirectly linked to individual human subjects was exempt from institutional review or approved by the IRB of the Medical College of Wisconsin.

All animal experiments were approved by the Institutional Animal Care and Use Committee of the University of Pennsylvania and were carried out in accordance with the IACUC guidelines. All mice had water ad libitum and were fed regular chow. Mice were maintained in a specific pathogen–free facility in accordance with American Association for Laboratory Animal Science guidelines. Littermates from different cages were randomly assigned to the experimental groups. These randomized experimental cohorts were either cohoused or systematically exposed to the bedding of other groups to ensure equal exposure to the microbiota of all groups.

SYF, ZL, HR, JAD, CK, and SWR designed the research. ZL, AO, IIV, ARP, FZ, CC, PY, RMD, HZ, RK, YS, ER, ATY, EK, and DPB performed the experiments and interpreted the data. SYF, ZL, HR, JAD, CK, and SWR wrote the manuscript with the help of all authors.

This work was supported by NIH/NCI grants R01 CA247803 (to SYF and DPB), R01 CA240814 (to SYF and HR), and P01 CA165997 (to JAD, CK, and SYF). We are thankful for support for RMD and equipment for characterizing TEVs from The Extracellular Vesicle Core Facility, School of Veterinary Medicine, University of Pennsylvania. We are grateful to Elena Voronov (Ben Gurion University of the Negev, Beer Sheva, Israel) and Bang-Jin Kim (SWR lab) for valuable advice. We thank Susan Ostrand-Rosenberg (University of Maryland, Baltimore, Maryland, USA), Ben Stanger (University of Pennsylvania), and Lucia Languino (Thomas Jefferson University) for reagents. We also thank Ze’ev Ronai (Sanford Burnham Prebys Medical Discovery Institute, San Diego, California, USA), Gennady Belitsky (N.N. Blokhin National Medical Research Center of Oncology, Moscow, Russia), Ana Gamero (Temple University, Philadelphia, Pennsylvania, USA), Igor Astsaturov (Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA), and the members of the Fuchs, Koumenis, and Ryeom labs for critical suggestions.

Conflict of interest: The authors have declared that no conflict of interest exists.


Potential applications of engineered nanoparticles in medicine and biology: an update

Nanotechnology advancements have led to the development of its allied fields, such as nanoparticle synthesis and their applications in the field of biomedicine. Nanotechnology driven innovations have given a hope to the patients as well as physicians in solving the complex medical problems. Nanoparticles with a size ranging from 0.2 to 100 nm are associated with an increased surface to volume ratio. Moreover, the physico-chemical and biological properties of nanoparticles can be modified depending on the applications. Different nanoparticles have been documented with a wide range of applications in various fields of medicine and biology including cancer therapy, drug delivery, tissue engineering, regenerative medicine, biomolecules detection, and also as antimicrobial agents. However, the development of stable and effective nanoparticles requires a profound knowledge on both physico-chemical features of nanomaterials and their intended applications. Further, the health risks associated with the use of engineered nanoparticles needs a serious attention.

Graphical Abstract

This is a preview of subscription content, access via your institution.


Mitochondria—hubs for regulating cellular biochemistry: emerging concepts and networks

Mitochondria are iconic structures in biochemistry and cell biology, traditionally referred to as the powerhouse of the cell due to a central role in energy production. However, modern-day mitochondria are recognized as key players in eukaryotic cell biology and are known to regulate crucial cellular processes, including calcium signalling, cell metabolism and cell death, to name a few. In this review, we will discuss foundational knowledge in mitochondrial biology and provide snapshots of recent advances that showcase how mitochondrial function regulates other cellular responses.

1. Introduction

All modern-day eukaryotes are believed to have arisen from a primordial ancestor that engulfed an α-protobacterium with the capacity for respiration [1]. This event gave rise to modern-day mitochondria, an event that is now deeply integrated in eukaryotic cell homeostasis and survival. Mitochondria are dynamic networks capable of remodelling their morphology and activity. They provide energy and biomolecules for the cell, in addition to contributing to pathways of cell stress, immune responses, intra- and intercellular signalling, cell-cycle control and cell death. The unique biology of mitochondria underpins their influence on the cell and the ability to calibrate their structure and proteome is an efficacious means of adapting their function. As such, we will begin with a brief outline of three fundamental concepts in mitochondrial biology: (i) mitochondrial ultrastructure (ii) mitochondrial protein import and (iii) mitochondrial dynamics. This will inform subsequent discussion of mitochondria as key players in broad and diverse roles, including metabolism, signal transduction, immunity, cell cycle, cell differentiation, cell death and stress.

2. Mitochondrial ultrastructure, dynamics and protein import

2.1. Mitochondrial ultrastructure

Mitochondria have a double membrane that defines four compartments: the outer membrane, the intermembrane space, the inner membrane and the matrix. The architecture of the inner membrane is malleable and typically convoluted into folded invaginations, called cristae, that dictate the spatial arrangement of proteins [2]. Remodelling cristae structure of cristae can also alter enzymatic flux between the compartments, consistent with the diverse cristae structures observed across cell types with different metabolic demands [2]. The recently described MICOS complex (mitochondrial contact site and cristae organizing system) is required to maintain cristae morphology [3] (figure 1). Loss of MICOS assembly ablates cristae junctions and manifests severe defects in energy metabolism, calcium handling and lipid trafficking [4]. However, it remains unclear how MICOS is regulated by cellular conditions to produce diverse cristae morphologies. Interestingly, disruption of MICOS alters the activity and/or abundance of mitochondrial morphology proteins [5,6]. Perturbations to organelle function have long been associated with gross morphological changes in the mitochondrial network, therefore cristae reorganization by MICOS assembly/disassembly may be an intermediary between function and dynamics. Recently identified associations between MICOS and protein import complexes point to the broad influence of MICOS on mitochondrial function [7,8].

Figure 1. Nuclear-encoded mitochondrial proteins are imported by multi-subunit translocases. Mitochondrial proteins synthesized in the cytosol are imported into mitochondria post-translationally. The TOM complex at the outer membrane serves as a general protein entry gate. hTom40 forms the pore of the translocase, while hTom20, hTom22 and hTom70 function as receptors. hTom22 plays an additional role in the assembly of the complex. hTom5, hTom6 and hTom7, collectively called the small TOMs, regulate the dynamics and assembly of the complex. The TIM22 complex at the inner membrane mediates the import of multi-pass transmembrane proteins into the inner membrane. hTim22 forms the pore through which proteins are inserted, while AGK and hTim29 function as receptors and in complex assembly. The TIM23 complex can translocate precursor proteins into the matrix or the inner membrane. hTim23 and hTim17 form the channel pore, and hTim50 functions as a receptor for precursors. The core complex associates with an import motor that helps to translocate proteins into the matrix in an ATP-dependent manner. The MIA complex mediates the import of soluble intermembrane space proteins by catalysing the formation of disulfide bonds. hMia40 carries out the disulfide bond formation and is anchored to the inner membrane through an interaction with AIF. ALR removes electrons from hMia40 so that it can undergo further rounds of catalysis. The SAM complex of the outer membrane mediates insertion of β-barrel proteins into the outer membrane. hSam50 associates with MTX1 and MTX2. Cristae, the large invaginations of the inner mitochondrial membrane, are stabilized by a multi-subunit complex called MICOS. Mic60 is the core subunit of MICOS, which additionally contains Mic10, Mic13, Mic14, Mic19, Mic25, Mic26 and Mic27. MICOS also associates with the SAM complex at the outer membrane to form a structure known as the mitochondrial intermembrane space bridging complex (MIB).

(Recommended further reading on cristae, MICOS and ultrastructure: [2,9,10].)

2.2. Mitochondrial protein import

From their endosymbiont origins, human mitochondria have retained only 37 genes in a small circular genome known as the mitochondrial DNA (mtDNA), which encodes 13 polypeptides, 22 tRNAs and 2 rRNAs. The remaining 1000–1500 mitochondrial proteins are nuclear encoded and must be imported and sorted to the relevant mitochondrial compartment following synthesis in the cytosol. Fundamentally, mitochondrial protein import is mediated by multimeric protein complexes known as translocases, which are located at mitochondria (figure 1). Briefly, two major translocases reside in the outer membrane of mitochondria: the Translocase of the Outer Membrane (TOM) complex and the Sorting and Assembly Machinery (SAM). The TOM complex is the initial point of contact for almost all mitochondrial precursors and provides a means of entry into the organelle. Following translocation through TOM, precursor import pathways diverge based on their targeting information and ultimate location within the organelle. β-barrel proteins of the outer membrane are sorted to the SAM complex for integration into the membrane. There are two translocases embedded in the inner membrane of mitochondria: the Translocase of the Inner Membrane (TIM) 22 and 23 (TIM22 and TIM23) complexes. TIM22 mediates the insertion of non-cleavable polytopic membrane proteins into the inner membrane, while the TIM23 complex is responsible for importing precursors across the inner membrane into the matrix or in some instances can laterally release transmembrane precursors into the inner membrane. Finally, the Mitochondrial Intermembrane space Assembly (MIA) machinery mediates the import of small cysteine-rich intermembrane space proteins and couples their import to their oxidation [11]. These import pathways and machines have been predominately characterized in fungal organisms however, in more recent years, analysis in higher eukaryotes has uncovered important physiological consequences due to perturbations in protein import. Specifically, mutations in genes encoding protein import subunits cause distinct mitochondrial diseases with phenotypes ranging from severe muscular defects to neurodegeneration and congenital growth defects [12].

(Recommended further reading on mitochondrial protein import: [13–15].)

2.3. Mitochondrial dynamics: fission, fusion and organelle contact sites

As an organellar network, mitochondria undergo fission and fusion to replicate, be recycled, and alter their bioenergetics. Fusion of the outer membrane is mediated by homotypic interactions between the GTPases Mfn1 and Mfn2 on adjacent mitochondria (figure 2) [16], but the domains involved and stepwise mechanism of fusion are still debated. Fusion of the inner membrane is controlled by Opa1, which exists as five isoforms generated by mRNA splicing and proteolytic cleavage (figure 2) [17]. It is believed that the stoichiometry of these isoforms governs Opa1 interactions with the mitochondrial-specific lipid cardiolipin and, subsequently, fusion events [18]. Mitochondrial fusion is associated with increased ATP production by oxidative phosphorylation and protects against oxidative and proteostatic stress [19]. Conversely, mitochondrial fission is concomitant with a reliance on glycolysis and precedes mitochondrial turnover. Fission is largely dependent on the dynamin-related and cytosolic protein Drp1, which oligomerizes around and constricts mitochondrial tubules (figure 2). The recruitment of Drp1 from the cytosol requires adaptor proteins on the mitochondrial outer membrane, including Mff, Mid49 and Mid51 [20,21], although human Fis1 can promote Drp1-independent mitochondrial fragmentation through inhibition of fusion proteins [22] (figure 2). While conflicting models of Drp1 recruitment have been proposed, its localization and activity are known to be regulated by numerous post-translational modifications [23]. The scission ability of Drp1 oligomers is sterically limited to tubules up to 250 nm diameter, indicating pre-constriction is required for larger mitochondria [24]. This is achieved by the endoplasmic reticulum (ER), which wraps around and constrict tubules to mark future fission sites and aid correct partitioning of mitochondrial contents [25,26].

Figure 2. Cellular machineries mediating mitochondrial fission, fusion and formation of contact sites with the endoplasmic reticulum. Mitochondria continuously undergo fission and fusion. Fission is mediated by the GTPase Drp1, which can be recruited to the outer mitochondrial membrane by a variety of receptors, including Mff, Fis1, Mid49 and Mid51. Drp1 at the outer membrane can oligomerize into fibrils that constrict mitochondria to initiate fission. Mitochondrial fusion is initiated by tethering of mitochondria through homotypic interactions between Mfn1 and Mfn2 on opposing mitochondria. Inner membrane fusion is mediated by OPA1, which exists as long and short forms generated through proteolysis. Contact sites between the mitochondria and the endoplasmic reticulum (ER) are established and maintained through protein–protein interactions. Interactions occur between Mfn2 molecules on the ER membrane and the outer mitochondrial membrane, and between VAPB on the ER membrane and RMDN3 on the mitochondrial outer membrane. Interactions also occur between IP3R3, a calcium channel on the ER membrane, and VDAC1 and hTom70 on the mitochondrial outer membrane.

Mitochondria also engage in extensive dynamic inter-organelle contacts that coordinate functional exchanges between mitochondria and other cellular components [27]. In particular, ER–mitochondria contact sites (ERMCs) facilitate a multitude of functions including mitochondrial fission, coenzyme Q biosynthesis, lipid transfer, Ca 2+ transfer, mtDNA replication and autophagy [25,26,28–31]. The ER–mitochondria encounter structure (ERMES) has been well characterized in Saccharomyces cerevisiae [32], however no human equivalent has been identified [33]. Preliminary work in humans suggests that metazoan ERMCs are tethered by interactions between hTom70 and IP3R3, VDAC1 and IP3R3, RMDN3 and VAPB, Mfn2 homodimers, Vps13a and Pdzd8 with an unknown partner (figure 2) [34–39]. Furthermore, acetylated microtubule ‘tracks’ have been proposed to maintain these contacts despite the movements and remodelling of the two organellar networks [40]. Other inter-organelle contacts have been described between mitochondria and Golgi [27,41], peroxisomes [42], lysosomes [43], lipid droplets [44] and the plasma membrane [45]. The interconnectivity of mitochondria with cellular components enables significant interplay across various pathways, examples of which we will highlight throughout this review (table 1).

Table 1. Full names and identifiers of proteins discussed in this review.

(Recommended further reading on mitochondrial dynamics: [46–48] on organelle contacts: [49–51].)

3. Mitochondria and metabolism

Mitochondria are well known for providing energy to the cell, predominantly by coupling the tricarboxylic acid (TCA) cycle with oxidative phosphorylation. The TCA cycle is a series of eight enzymatic reactions that occur in the matrix to harvest electrons from citrate and its catabolic intermediates (figure 3a). The typical input to the cycle is acetyl-CoA, which can be produced from glucose (via glycolysis), fatty acids (via β-oxidation) and amino acids (via deamination) (figure 3a). The electrons scavenged throughout the cycle are transferred by NADH and FADH2 to the complexes of the electron transport chain. Complexes I–IV of the electron transport chain shuttle electrons, using their energy to pump protons into the intermembrane space and establish an electrochemical gradient across the inner membrane. Complex V (ATP synthase) releases the protons back into the matrix, using the energy of the electrochemical gradient to produce ATP, the cell's energy currency (figure 3a) [52]. Although normally efficient, oxidative phosphorylation is negatively regulated by the accumulation of its toxic by-product, reactive oxygen species (ROS). If unchecked, ROS can cause damage to mitochondria, induce protein aggregation and introduce mutations in DNA [53–55]. Recent advances in cryoelectron microscopy have revealed Complexes I, III and IV can assemble to form supercomplexes thought to reduce the amount of ROS produced during electron transport, as well as enhance respiration rates [56].

Figure 3. Mitochondria coordinate essential metabolic processes. (a) Mitochondria are best known for housing the protein machinery required for generating ATP. When oxygen is available, most cells will generate ATP through oxidative phosphorylation, where electrons harvested through catabolic reactions are used to power ATP synthase. Electrons are obtained through the TCA cycle, which occurs in the matrix and consists of eight enzymatic reactions. Acetyl-CoA is the primary input for the TCA cycle, and can be obtained through metabolism of glucose, fatty acids and amino acids. Electrons extracted during the TCA cycle are loaded onto NAD + and FAD 2+ . Electrons are subsequently transferred from NADH and FADH2 onto Complexes I and II of the electron transport chain. Electrons are passed through Complexes III and IV, which transport protons into the intermembrane space. Protons are allowed to flow back into the matrix through ATP synthase (Complex V), which uses the energy of the proton gradient to convert ADP to ATP. (b) Mitochondrial one-carbon (1C) metabolism comprises a series of parallel and reversible reactions which occur in the cytosol and mitochondrial matrix. In proliferating cells, the reaction normally proceeds in a specific direction such that formate produced within mitochondria can be used for biosynthetic processes in the cytosol. Within the mitochondria, THF and serine imported from the cytosol are acted upon sequentially by SHMT2, MTHFD2 and MTHFD1 L to produce formate, which is exported back into the cytosol. Cytosolic MTHFD1 loads formate onto THF to form charged folate intermediates that can be used to synthesize purine and pyrimidine nucleotides. Mitochondrial 1C metabolism is also an important source of glycine. (c) The mitochondrial matrix functions as an important storage site for calcium ions. Mitochondrial calcium uptake often occurs at ER contact sites, where large volumes of Ca 2+ can be released through IP3R3. Calcium can pass freely through the outer membrane via VDAC channels and is transported across the intermembrane space and inner membrane through the coordinated function of a MICU1/MICU2 dimer docking to MCU in the inner membrane. Calcium can exit the mitochondrial matrix through LETM1 or SLC8B1 (in exchange for H + or Na + , respectively) and can cross the outer membrane through VDACs or NCX3.

Mitochondria also produce fatty acids, amino acids, nucleotides and haem groups for the cell through biosynthetic pathways [57–59]. One such process, one-carbon (1C) metabolism, produces glycine, methionine, nucleotides, phosphatidylcholine and 1C units (methyl-like groups) from serine catabolism through the redox chemistry of folate and its derivatives (figure 3b) [60]. These 1C units charge the universal methyl donor S-adenosylmethionine required for the methylation of proteins and chromatin [61]. There is now significant evidence of metabolic enzymes and metabolites altering gene expression as reporters of environmental conditions (nutrient availability, hypoxia, oxidative stress) or mitochondrial dysfunction. This has been shown for acetyl-CoA, TCA intermediates, ketones, lactate, fatty acids and amino acids [62–68]. Emerging studies also indicate cellular nutrient and energy sensing by mTOR kinase regulates mitochondrial biogenesis and protein synthesis [69]. Through downstream effectors of transcription and translation, mTORC1 stimulates mitochondrial biogenesis and oxidative metabolism to meet the energy demand of anabolism [70–72]. Interestingly, the tumour suppressor p53 inhibits mTOR-mediated growth and proliferation to prevent oncogenesis [73,74]. p53 activity increases electron transport chain efficacy [75], mtDNA stability [76,77] and reduced glutathione (GSH) levels [78] to limit ROS production as well as inhibiting glycolysis [79,80], which contributes to the replicative potential of tumour cells [79,81,82]. Thus, metabolism is intimately integrated with other cellular pathways, but is not the sole contribution of mitochondria to signalling mechanisms.

(Recommended further reading on metabolism: [60,83] on metabolite signalling: [68,84] on mTOR/p53: [85,86].)

4. Signalling

4.1. Mitochondria control calcium homeostasis

Calcium ions are common to diverse signalling pathways. The outer mitochondrial membrane is permeable to Ca 2+ , in part due to channel-forming VDAC proteins [87] and export via SLC8A3 [88]. The mitochondrial inner membrane calcium uniporter (MCU) complex regulates transport into the matrix (figure 3c). Permeability of the MCU complex is calibrated by two regulatory subunits, MICU1 and MICU2, that are linked by an intermolecular disulfide bond introduced by hMia40 [89,90]. The ability of mitochondria to accumulate Ca 2+ up to 20-fold higher concentrations than the cytosol allows them to function as buffering systems and re-establish homeostasis following Ca 2+ bursts [91,92]. Bursts of Ca 2+ into the cytosol, from across the plasma membrane or intracellular stores, can initiate neurotransmitter release, muscle fibre contraction and transcriptional regulation. In neurons, mitochondrial Ca 2+ buffering modulates both the propensity and duration of neurotransmitter release [93,94]. In cardiac muscle, contraction is coupled to enhanced mitochondrial ATP production via Ca 2+ -increased activities of TCA cycle enzymes, Complex V and the ADP/ATP transporter [95–98] an effect maximized by local Ca 2+ concentrations at ERMCs [29,99] (figure 3c). Additionally, mitochondrial Ca 2+ regulation influences hormone secretion [100], tissue regeneration [101] and interferon-β signalling via the mitochondrial antiviral signalling protein, MAVS [102].

(Recommended further reading on mitochondrial Ca 2+ signalling: [92,103,104].)

4.2. Roles of mitochondria in immune responses

The contribution of mitochondria to immune responses is a growing area of research. Cell-autonomous immune signalling is driven by MAVS at the outer membrane, which acts as a relay point for immune signal transduction. Rig-like receptors in the cytosol undergo conformational changes upon detecting viral RNA or DNA and are recruited to MAVS, particularly at ERMCs [105]. MAVS then dimerizes to enable the binding of multiple downstream signalling adaptors including TRADD, TRAF3 and STING to activate NF-κB and IRF-3/7 transcription of interleukins and pro-inflammatory cytokines [106–108] (figure 4a). Interestingly, MAVS dimers and many of its adaptors co-immunoprecipitate with hTom70 of the TOM complex, the overexpression of which increases the signalling response [109]. MAVS signalling is also affected by ROS and negatively regulated by Nlrx1, a binding partner of Complex III and MAVS [110,111] (figure 4a). As mitochondrial protein import and oxidative metabolism can be hijacked by virulence factors [112], these interactions may make MAVS sensitive to consequences of infection. Finally, if mitochondria are compromised by infection, the increased ROS and release of mtDNA into the cytosol can activate the NLRP3 inflammasome to evoke an inflammatory response [113,114] (figure 4a).

Figure 4. Mitochondria make crucial contributions to diverse cellular processes. (a) The mitochondrial outer membrane is the site of important signalling events during the innate immune response. Detection of viral nucleic acids by Rig-like receptors (RLRs) induces dimerization of MAVS, a protein of the mitochondrial outer membrane. Dimerized MAVS recruits signalling adaptors that initiate downstream activation of IRF3/7 and NF-κB, transcription factors that induce expression of type I interferons and pro-inflammatory cytokines. MAVS is regulated by NLRX1, a protein which downregulates MAVS when localized to the outer membrane, but activates MAVS when at the inner membrane by interacting with Complex III to induce ROS production. Release of mtDNA during infection can also activate the NLRP3 inflammasome. (b) Mitophagy is a process that allows damaged mitochondria to be identified and destroyed. Under normal conditions, PINK1 is imported into mitochondria and degraded by PARL. When mitochondria are damaged, import is impaired and PINK1 accumulates in the TOM complex at the outer membrane. Autophosphorylated and active PINK1 at the outer membrane phosphorylates monoubiquitin molecules on outer membrane proteins, recruiting and activating the E3 ubiquitin ligase Parkin. Activated Parkin synthesizes polyubiquitin chains that recruit autophagy receptors to initiate mitophagy. (c) Mitochondrial proteostatic stress is sensed through the partitioning of the transcription factor ATF5 between the mitochondria and the nucleus. Under normal conditions, ATF5 is imported into and sequestered within mitochondria. If mitochondrial protein import becomes compromised, ATF5 is trafficked into the nucleus, where it upregulates expression of genes that enhance proteostasis. (d) Mitochondria play crucial roles in the initiation of apoptosis. In response to pro-apoptotic stimuli, Bax and Bak oligomerize in the outer membrane to form pores that allow for efflux of apoptogenic proteins (Cytochrome c, Diablo, AIF and Endonuclease G) from the intermembrane space into the cytosol. Cytochrome c binds to Apaf-1 to induce formation of the apoptosome and activation of caspases. Diablo blocks inhibitors of apoptosis (IAPs) which would otherwise mitigate the effect of caspases. AIF and Endonuclease G translocate into the nucleus where they contribute to destruction of the genome.

Mitochondrial metabolism also directs rapid changes to specialized immune cells during infection. Changes in membrane potential can activate or supress M2 macrophages [115,116] and M1 macrophages shunt intermediates from the TCA cycle to generate nitrous oxide, IL-1β and the antibacterial itaconic acid [117,118]. Furthermore, the phagocytic abilities of macrophages depend on mitochondrial ROS production to destroy internalized pathogens [119]. Naive T-cells display increases in mitochondrial mass, mtDNA copy number, glycolysis, and glutamine metabolism during differentiation for rapid proliferation and to escape quiescence [120,121]. Metabolic remodelling then also decides the T-cells' mature fate [122,123], by altering cristae architecture [124] or by direct effect of metabolites on epigenetic transcription regulation [125].

(Recommended further reading on mitochondrial immune signalling: [106,118,126].)

5. Cell cycle, differentiation and death

Mitochondria are implicitly tied to cell-cycle control as providers of energy and nucleotides however, they also coordinate checkpoints and respond to signals of proliferation. To meet the metabolic demand of mitosis, mitochondrial mass and membrane potential increase from G1/S until late mitotic stage [127]. Indeed, hyperpolarization and increased ATP production inhibit AMP kinase to allow cyclinE-mediated entry to S-phase [128]. In the late G2 stage of dividing S. cerevisiae, the cyclinB1/Cdk1 complex traffics to mitochondria to phosphorylate Complex I subunits and Tom6, stimulating oxidative metabolism both directly and indirectly via increased protein import [129,130]. During mitosis, a highly fused and reticular mitochondrial network progressively fragments to small tubular organelles that segregate in anticipation of cytokinesis [127,131]. Mitochondria can also delay cell-cycle progression to increase their biogenesis [132], because of insufficient nucleotide production [133], or because of ROS accumulation [134]. Moreover, the fusion mediator Mfn2 can sequester both Ras and Raf to inhibit proliferative signalling [135].

Stem cell differentiation also relies on mitochondria as a ‘metabolic switch’. Human embryonic stem cells are glycolytic however, they develop mature cristae, rapidly replicate mtDNA and increase ATP production upon differentiation [136]. In haematopoietic stem cell differentiation, the downregulation of Pdk2, an inhibitor of pyruvate dehydrogenase, releases suppression of acetyl-coA production and enables oxidative phosphorylation [137]. The subsequent increase in ROS production and oxidative phosphorylation during differentiation drives upregulation of mitochondrial antioxidant proteins by the transcription factors Oct4, Sox2 and Nanog [138]. Mitochondrial fusion is believed to facilitate these metabolic changes, although the importance of specific proteins and fission/fusion balance may be cell-type specific [139–141]. This is supported by somatic cell reprogramming studies showing deletion of Mfn2 permits pluripotency as glycolysis becomes predominant over oxidative phosphorylation [142] the same effect being achieved by the pluripotency factor ZFP42 activation of Drp1 [143].

If cellular conditions or external insults are too harsh, mitochondria can trigger multiple forms of cell death. Apoptosis, or programmed cell death, can be elicited from extrinsic signalling via the Fas, TRAIL and TNFα receptors or intrinsic insults such as DNA damage, Ca 2+ overload, ROS and ER stress [144]. Mitochondria contribute to the extrinsic pathway but are the nexus of the intrinsic apoptotic pathway. In the latter pathway, cytosolic pro-apoptotic Bax oligomerizes with Bak at the outer membrane to permeabilize mitochondria and release pro-apoptotic proteins, including cytochrome c, Diablo, Htra2, Endonuclease G and AIF (figure 4d) [145]. In the cytosol, cytochrome c nucleates the formation of the apoptosome and activation of the caspases that dismantle the cell in an immunologically silent manner. Cytosolic Diablo and Htra2 block inhibitors of caspase activation, which would otherwise protect the cell from basal cytochrome c leakage [146,147]. Endonuclease G and AIF translocate to the nucleus to fragment DNA (figure 4d), AIF first requiring proteolytic cleavage of its transmembrane domain [148–150]. AIF is normally part of the intermembrane space import machinery, or MIA complex, anchoring the oxidoreductase hMia40 to the inner membrane. The outer membrane protein VDAC2 protects against apoptosis by sequestering Bak [151,152], yet new evidence suggests it may be required for Bax-mediated apoptosis [153]. Emerging research also implicates mitochondria in alternate and less-studied cell-death pathways such as ROS-induced necrosis [154], immune-activated necroptosis [155], ferroptosis [156,157] and parthanotosis [158].

(Recommended further reading on mitochondria in the cell cycle: [159,160] on differentiation [161–163] on cell death: [164,165].)

6. Mitochondrial quality control

The loss of mitochondrial function has profound negative effects on cellular health therefore, multiple quality control and stress response mechanisms have evolved. The mitochondrial unfolded protein response (mtUPR) detects proteostatic stress within mitochondria [166]. Central to the mtUPR is the transcription factor ATF5. When stress causes protein import and/or electron transport chain dysfunction ATF5 accumulates in the nucleus to transcribe mitochondrial chaperones and protease genes (figure 4c) [167,168]. The Caenorhabditis elegans homologue ATFS-1 has also been shown to repress translation of the electron transport chain subunit and assembly proteins from both mitochondrial and nuclear genomes [169]. Translation of ATF5 is partly controlled by its homologue ATF4, both of which are upregulated in the integrated stress response (ISR) [170,171]. The ISR can be triggered by ER stress, amino acid starvation or degradation of hTim17A, a TIM23 complex subunit [172,173]. The ISR is characterized by phosphorylation of eIF2α, leading to global reduction of translation and selective induction of cytoprotective genes including pro-survival MCL1 and autophagy proteins. This illustrates the preference for clearance of defective organelles over controlled cell death although the response may alter with cell type or insult [174].

The selective autophagic clearance of mitochondria is termed mitophagy and is controlled by the mitochondrial serine/threonine protein kinase PINK1 and the E3 ubiquitin ligase Parkin. PINK1 is constitutively imported into healthy mitochondria through the TOM complex and laterally released into the inner membrane by TIM23 [175] before cleavage by the PARL protease (figure 4b) [176]. Depolarization of the inner membrane in defective mitochondria prevents import of PINK1, causing it to oligomerize at the outer membrane TOM complex [177], where it becomes auto-phosphorylated [178]. This triggers phospho-PINK1 phosphorylation of basal outer membrane monoubiquitin and recruits Parkin to rapidly poly-ubiquitinate outer membrane proteins for the recruitment of autophagosome factors (figure 4b) [179,180]. Recent data suggest that mitochondria can identify and initiate mitophagy of specific tubules [181], while mitophagy induced by CSNK2/CK2 phosphorylation of hTom22, FUNDC1 and BCL2L13 suggests a potential cytoplasmic influence or pathway [182–185]. Additionally, observations of transcellular mitophagy in astrocytes illustrate much is still unknown in these processes [186].

New stress responses are emerging that demonstrate the reciprocal communication between mitochondria and cytoplasm. Ablation of MIA import pathways in S. cerevisiae activates the proteasome to mitigate mitochondrial precursor accumulation in the cytosol [187]. This correlates with the mammalian, intermembrane space-specific mtUPR (mtUPRIMS) where ERRα transcriptional activity upregulates intermembrane space proteases and activates the proteasome [188,189] the proteasome being previously shown to degrade unfolded intermembrane space proteins that retrotranslocate to the cytosol [190]. In S. cerevisiae, the proteasome is also engaged by Ubx2 to clear mitochondrial protein precursors arrested during translocation, blocking the TOM complex [191]. Reciprocally, mitochondria can degrade defective proteins to aid cytosolic proteostasis. In S. cerevisiae, cytosolic Vms1 can remove mistranslated mitochondrial precursors from stalled ribosomes and direct their import for intra-mitochondrial degradation [192] and aggregation-prone cytosolic proteins may be imported for intra-mitochondrial degradation if cytosolic Hsp70s fail [193]. Intriguing for further research are reports of lysosomal fusion of mitochondria-derived vesicles enriched for non-natively oxidized proteins [194,195] and the extracellular jettison of aggregates by neurons of C. elegans [196].

(Recommended further reading on mitochondrial quality control: [197–199] on mitophagy: [200,201].)

7. Concluding remarks

This review illustrates the importance of mitochondria to eukaryotic cellular functions. As mitochondrial biologists we are frequently surprised by novel pathways or protein networks that involve mitochondria and/or mitochondrial proteins. Mitochondrial protein import and structural dynamics provide the means for rapid alterations in activity to facilitate biological responses to signalling molecules, nutrient availability and pathogenic insult. The temporal coordination of mitochondrial energetics and their biosynthetic capacity drives cell proliferation and differentiation. However, the highly reactive biochemistry compartmentalized in the organelle makes it capable of inducing cell death and necessitates quality control mechanisms. An understanding of this interplay between mitochondrial functions and their diverse cellular implications is therefore critical to a comprehensive holistic model of cellular homeostasis and biochemistry. The importance of this is evident in the escalating occurrence of mitochondria in post-genomic medical research [202]. Although mitochondria are undeniably hubs of cellular biochemistry, further fundamental research is required. In particular, elucidating how the mitochondrion regulates and integrates the various pathways it is associated with, in specialized cells/tissue types and in the context of health and in disease, will help uncover the true depth of influence this amazing organelle has on eukaryotic cells.


PGC-1α and Type 2 Diabetes

Type 2 diabetes is the most common metabolic disease in the world, reaching epidemic proportions. Although there are multiple complicated genetic elements involved in the pathogenesis of Type 2 diabetes, lifestyle and age seem to be the two major risk factors triggering the disease. A sedentary lifestyle and overeating can contribute to excessive weight gain and obesity, with the latter being closely associated with insulin resistance. Insulin resistance is considered to be present when the biological effects of insulin are less than expected for both glucose disposal in skeletal muscle and suppression of glucose production by the liver (11). When the insulin level is not sufficient to overcome this insulin-resistant state, Type 2 diabetes ensues. In this regard, there is there is increasing evidence suggesting that mitochondrial dysfunction plays a role in the pathogenesis of insulin resistance and Type 2 diabetes. For example, it has been reported that the activities of both mitochondrial oxidative enzymes and mitochondria complex I are decreased in Type 2 diabetic patients (21, 63). In keeping with the observations that both obesity and mitochondrial dysfunction are risk factors for the development of insulin resistance, obese individuals have been reported to have smaller mitochondria that exhibit a compromised bioenergetic capacity (21). Insulin resistance also develops with age, and, in this case, a potentially important defect has been found in the mitochondrial fatty acid oxidation pathway. There appears to be a reduction in the rate of fatty acid oxidation, leading to the accumulation of intracellular triglycerides in elderly patients compared with matched young controls (43). Furthermore, the accumulation of triglycerides has been directly correlated with the development of insulin resistance in skeletal muscle and the liver. Thus, a specific type of mitochondrial dysfunction, i.e., decreased fatty acid oxidation, may play a significant role in the development of insulin resistance in both aging and obesity (43). It is interesting to note that triglycerides can also accumulate in pancreatic β-cells, resulting in decreased rates of insulin secretion, which would further exacerbate the problem of glucose disposal (32, 60). As noted above, the accumulation of intracellular triglycerides could be the result of a deficiency in fatty acid oxidation related to mitochondrial dysfunction. As a result, the cause of insulin resistance may well be due to mitochondrial dysfunction rather than the accumulation of triglycerides, a mechanistic concept that will be important to establish in gaining an understanding of the etiology of Type 2 diabetes.

It has been hypothesized that a close relationship exists among PGC-1α function, insulin sensitivity, and Type 2 diabetes, which is most likely related to the essential roles of PGC-1α in mitochondria biogenesis and glucose/fatty acid metabolism. Evidence supporting this hypothesis has been found in a number of studies. For example, it has been observed that expression of PGC-1α is downregulated in muscles of Type 2 diabetic subjects (36, 42, 44). In addition, a common polymorphism of the PGC-1α gene (Gly482Ser), expressing reduced PGC-1α activity, has been linked to an increased risk of Type 2 diabetes (18, 40). These observations would suggest that either reduced levels or compromised activity of PGC-1α can be associated with the development of insulin resistance and Type 2 diabetes (Fig. 1). In this context, it is interesting to note that TZDs, an important class of antidiabetic drugs, also act to increase insulin sensitivity coincident with the activation of PGC-1α (66). The effects of TZDs are likely mediated through the ability of PGC-1α to activate mitochondria biogenesis and increase mitochondrial function. As described above, PGC-1α stimulates the conversion of muscle fiber type toward more oxidative type I and IIa fibers, favoring fatty acid oxidative metabolism (28). The promotion of fatty acid oxidative metabolism would be expected to lead to a reduction of fat accumulation in muscle, which, in turn, could increase insulin sensitivity because there is a close relationship between triglyceride accumulation and insulin resistance (43). One manifestation of increased insulin sensitivity would be increased glucose uptake by insulin-sensitive tissues. PGC-1α has, as previously noted, been reported to activate the expression of insulin-sensitive GLUT4 in skeletal muscle (5, 33). Taken together, the evidence presented supports a role for PGC-1α in preventing insulin resistance and Type 2 diabetes mellitus.

Fig. 1.Potential role of decreased peroxisome proliferator-activated receptor-γ coactivator (PGC)-1α expression in skeletal muscle as it relates to the development of the metabolic phenotype of insulin resistance and Type 2 diabetes mellitus. [Revised from Ref. 34.]

However, it should be pointed out that PGC-1α expression has been reported to be increased in the liver of both Type 1 and Type 2 diabetic mouse models, in contrast to the reported decrease in PGC-1α expression observed in the muscle of human Type 2 diabetic subjects described above (48). Furthermore, PGC-1α has been shown to inhibit the insulin signaling pathway in the liver (22), inhibit glucose utilization in cultured myotubes (65), and suppress β-cell energy metabolism and insulin release in mice (70). Increased hepatic PGC-1α expression could be expected to stimulate hepatic glucose output, and, when coupled with the reported inhibitory effect of PGC-1α on insulin signaling and secretion, the combined effects would be to contribute to the hyperglycemic state, a prodiabetes effect. This prodiabetic effect of PGC-1α has been further supported by the manifestation of improved insulin sensitivity in PGC-1α-deficient mouse models (22, 26, 29). Moreover, another recent report (38) has indicated that PGC-1α expression is normal in insulin-resistant subjects despite severe impairments in mitochondrial function. These apparently paradoxical actions of PGC-1α in different tissues are intriguing. The underlining mechanisms for the interactions between these tissues are currently unknown and deserve further intensive investigations.


Molecular and cellular bases of iron metabolism in humans

Iron is a microelement with the most completely studied biological functions. Its wide dissemination in nature and involvement in key metabolic pathways determine the great importance of this metal for uniand multicellular organisms. The biological role of iron is characterized by its indispensability in cell respiration and various biochemical processes providing normal functioning of cells and organs of the human body. Iron also plays an important role in the generation of free radicals, which under different conditions can be useful or damaging to biomolecules and cells. In the literature, there are many reviews devoted to iron metabolism and its regulation in proand eukaryotes. Significant progress has been achieved recently in understanding molecular bases of iron metabolism. The purpose of this review is to systematize available data on mechanisms of iron assimilation, distribution, and elimination from the human body, as well as on its biological importance and on the major iron-containing proteins. The review summarizes recent ideas about iron metabolism. Special attention is paid to mechanisms of iron absorption in the small intestine and to interrelationships of cellular and extracellular pools of this metal in the human body.


The Biological and Metabolic Fates of Endogenous DNA Damage Products

DNA and other biomolecules are subjected to damaging chemical reactions during normal physiological processes and in states of pathophysiology caused by endogenous and exogenous mechanisms. In DNA, this damage affects both the nucleobases and 2-deoxyribose, with a host of damage products that reflect the local chemical pathology such as oxidative stress and inflammation. These damaged molecules represent a potential source of biomarkers for defining mechanisms of pathology, quantifying the risk of human disease and studying interindividual variations in cellular repair pathways. Toward the goal of developing biomarkers, significant effort has been made to detect and quantify damage biomolecules in clinically accessible compartments such as blood and and urine. However, there has been little effort to define the biotransformational fate of damaged biomolecules as they move from the site of formation to excretion in clinically accessible compartments. This paper highlights examples of this important problem with DNA damage products.

1. Introduction

Endogenous processes of oxidative stress and inflammation cause DNA damage that is mechanistically linked to the pathophysiology of cancer and other human diseases [1]. The DNA damage is comprised of dozens of mutagenic and cytotoxic products [2–4] reflecting the full spectrum of chemical mechanisms, including oxidation, nitrosation, halogenation, and alkylation, as described in numerous published reviews [5–15]. There has been significant interest in developing DNA damage products as biomarkers of disease risk given the strong association between DNA damage and disease pathology [12, 14, 16–22]. However, there has been little consideration given to the biological fate of DNA damage products, such as release from DNA as a result of instability, repair, and reaction with local nucleophiles, and the effect of this fate on the steady-state level of DNA lesions in cells and tissues. Further, the use of tissue-derived DNA for biomarker development poses the problem of accessibility and limits clinical studies, so researchers are exploring the presence of DNA damage products in other sampling compartments, such as urine (e.g., [16, 23]). These efforts have presumed that DNA repair or cell death leads to dissemination of DNA damage products in blood, with subsequent excretion of specific molecular forms predicted to arise from the various DNA repair or other enzymatic processes. However, one of the major drawbacks to the use of blood or urine as a sampling compartment for development of DNA damage products as biomarkers is the lack of mechanistic information about the fates of the damage products in terms of metabolism and distribution. While information about the metabolic fate and pharmacokinetics of drugs based on nucleobases has been well defined (e.g., [24, 25]), studies of the metabolism of DNA damage products have been limited to a few products such as adducts of ethylene dibromide [26], the pyrimidopurinone adduct of dG, M1dG [27–29], and the base propenal and butenedialdehyde species arising from 2-deoxyribose oxidation in DNA [30–32].

The mechanisms governing the fate of endogenous DNA damage products can be viewed from two perspectives, the first being local reactions that lead to the release of the damage product, such as chemical instability or DNA repair, or the reaction of electrophilic damage products with local nucleophiles. The second perspective is that of drug and xenobiotic metabolism and distribution. In both cases, the release of the damage products from DNA results in their diffusion or transport into extracellular space for subsequent distribution in the blood circulation to the liver and excretory organs. Chemical stability governs the extent and form of distribution of the damage product, with electrophilic species reacting with local nucleophiles and more stable products circulating throughout the body. The damage products may be recognized as substrates for the variety of local or distant metabolic enzymes that cause oxidation, reduction, hydrolysis, and conjugation (e.g., glucuronic acid, sulphate, or glutathione), with metabolites excreted in either urine or bile [33, 34]. We can also view DNA damage products from the perspective of metabolic toxification and detoxification. Metabolic reactions are well known to either reduce the activity of reactive and toxic xenobiotics or to convert unreactive molecules to reactive intermediates that are genotoxic, hepatotoxic, or nephrotoxic [33, 34]. This paradigm applies to DNA damage products that range from relatively stable (e.g., nucleobase deamination products) to highly electrophilic (e.g., base propenals from 2-deoxyribose oxidation in DNA), with metabolic reactions occurring in cells in which the DNA damage occurs or in the liver or other metabolic tissues.

This review addresses the current state of understanding of the metabolic and biological fates of DNA damage products, with an eye on the implications of these fates for mechanisms of toxicity and for development of biomarkers of oxidative stress and inflammation.

2. The Spectrum of Nucleic Acid Damage Products

As a prelude to understanding the biological fate of damaged nucleic acids, we must first consider the spectrum of damage products. Nucleobases in DNA, RNA, and the nucleotide pool are subject to damage by a variety of chemical mechanisms related to normal and pathological processes. The superoxide (

) and hydrogen peroxide (H2O2) generated during aerobic respiration participate in Fenton chemistry to produce hydroxyl radical (HO • ), while the activated macrophages and neutrophils of chronic inflammation generate a host of chemically reactive species, including the oxidants peroxynitrite (

), hypohalous acids (HOCl, HOBr), and nitrosating agents (N2O3) [8]. Damage to nucleic acids and nucleotides can occur by direct reaction with these agents or indirectly by reaction with electrophiles generated during oxidation of lipids, carbohydrates, and proteins. Both the nucleobase and sugar moieties are susceptible to attack, with examples of nucleobase damage products shown in Figure 1 and 2-deoxyribose oxidation products shown in Figure 2. The biological and metabolic fates of nucleobase damage products will be addressed first and that of 2-deoxyribose oxidation products later in this chapter.



3. The Biological and Metabolic Fates of Damaged Nucleobases

The biological fates of damaged nucleotides and nucleic acids can be viewed from the perspective of either the site of initial damage or from the final sampling compartment used for analysis of the damage products. Among the issues that arise are (1) the reactivity of a damage product and the chemical form of the lesion that is released from the site of generation (2) the mechanism by which the released damage product reaches the systemic circulation (3) the potential for the damage product to be chemically modified between the steps of formation and excretion (4) the mechanism of excretion (5) the potential for further chemical modification in the excretory compartment. The first of these issues, that of reactivity, is best illustrated by the susceptibility of 8-oxoguanine to further oxidation, as will be discussed shortly, and the deglycosylation of many damaged purines, such as 8-nitroguanine [8], and of purines subjected to N 7 -nitrosation or alkylation [8], both of which have been addressed in detail in the literature. Here we will focus on the metabolic fates of nucleobase damage products.

3.1. 8-Oxoguanine

The first consideration of the metabolic fate of a nucleobase damage product is the well-studied 7,8-dihydro-8-oxoguanine (8-oxo-G Figure 1) [35]. Perhaps the most comprehensive consideration of the biological fate of 8-oxo-G in terms of sources of 8-oxo-G-containing species excreted in the urine is the recent review by Cooke et al. [36], with a very recent review of the utility of 8-oxo-dG as a urinary biomarker [23]. Among the nucleobases in DNA, RNA, and the nucleotide pool, guanine is the most readily oxidized due to its favorable redox potential [35, 37–39] with the spectrum of oxidation products depending on the nature of the oxidant [8, 35] (Figure 1). 8-Oxo-G is one of the major products common to oxidation of guanine by most oxidizing agents, and it has thus been touted as a biomarker of oxidative stress (e.g., [23, 36, 40, 41]. While oxidation of G in DNA is one source of 8-oxo-G, another involves polymerase incorporation of 8-oxo-dGTP formed by oxidation of dGTP in the nucleotide pool [42]. Prokaryotes and eukaryotes are equipped with oxidized purine nucleotide di- and triphosphatases (e.g., E. coli MutT, 8-oxo-dGTP triphosphatase) to remove damaged nucleotides from the pool [43].

There are four fates of 8-oxoG in cellular DNA and nucleotides: further oxidation to more stable products, which will be discussed shortly, removal from DNA by repair mechanisms, removal from the nucleotide pool by nucleotide di- and triphosphatases, and eventual release from DNA following cell death. Like many nucleobase oxidation products, 8-oxo-G in DNA is removed by the base excision repair (BER) pathway [44–47], with the ultimate release of free 8-oxo-G nucleobase by N-glycosylase activity. On the other hand, dephosphorylation of 8-oxo-dGTP and –dGDP ultimately releases 8-oxo-dGMP and 8-oxodG, which are also the likely forms released from DNA following cell death.

So, we are faced with the choice of quantifying either 8-oxo-G, 8-oxo-dG, or 8-oxo-dGMP in sampling compartments such as blood and urine. The most abundant of these species appears to be 8-oxo-dG, which is present in human urine at concentrations in the micromolar range. 2-Deoxynucleosides are chromatographically well behaved, and this concentration is amenable to precise and accurate quantification by liquid chromatography-coupled with mass spectrometric methods. While the excretion of 8-oxo-dG may correlate well with conditions of oxidative stress and inflammation [23], the source of this 8-oxodG has yet to be established.

Another fate of 8-oxoG in DNA, RNA, and the nucleotide pool, as well as the fate of 8-oxo-G-containing species released from cells, is further oxidation to form a variety of stable end products, as shown in Figure 1. 8-Oxo-G is significantly more susceptible to further oxidation than G itself (0.74 V versus 1.29 V relative to NHE [39]) and is thus susceptible to reaction with oxidants less potent than hydroxyl radical (2 V versus NHE), such as (1.04 V versus NHE [48]) and alkyl hydroperoxides (

0.9 V versus NHE [49]). The oxidation of 8-oxo-dG results in the formation of several new products (Figure 1), most of which are more stable than 8-oxo-dG itself and thus potentially better candidates for biomarkers of inflammation and oxidative stress. One must again consider the roles of DNA repair, nucleotide pool cleaning activities, and excretory pathways in finalizing the fate of 8-oxo-G oxidation products.

Finally, recent studies suggest two other confounding factors in the biological fate of 8-oxo-G. The first relates to alternate sources. A study by Tannenbaum and coworkers reveals that 8-oxo-G can arise by further oxidation of species such as 8-nitro-G, which arises from nitrative oxidation of G by and [50]. This and other analogous chemistries further confound the assignment of the source of 8-oxo-G-containing species as mechanistic biomarkers. The second confounder involves an alternative fate for 8-oxo-G: deamination to uric acid. Hall et al. have described 8-oxo-G deaminase activity in bacteria [51], which raises the possibility of similar activities in human cells. While we have not observed adventitious deamination of G in our studies of DNA deamination in vitro and in vivo [52–55], a G deaminase activity cannot be ruled out.

3.2. Etheno Adducts

Another major group of DNA lesions with a well-established association with oxidative stress and inflammation involves adducts formed in the reaction of DNA with electrophiles generated by lipid peroxidation [56–58]. This group includes the substituted and unsubstituted etheno nucleobase adducts [58–63] (Figure 1). Extensive study of the urinary excretion of unsubstituted etheno adducts has revealed a strong correlation of excretion with host of human diseases, pathologies, and environmental exposures related to oxidative stress (e.g., see recent studies in [16–21, 64]). Nonetheless, there have been few if any studies aimed at defining the source of the etheno 2-deoxynucleosides measured in these studies.

By analogy to 8-oxo-G, the fate of etheno adducts can be viewed from the perspectives of DNA repair and metabolism. Etheno adducts in DNA are presumed to be repaired by the BER pathway [65], with the release of the free-base adducts. However, biomarker studies again focus on the 2-deoxynucleoside form of the adducts [16–21, 64], which must arise from pathways other than DNA repair. The current focus on quantifying etheno adducts as 2-deoxynucleosides has recently been called into question by the Marnett group’s pioneering studies of the metabolism of endogenous DNA adducts [27–29, 66]. With regard to etheno adducts, they incubated 2-deoxynucleoside forms of substituted and unsubstituted etheno adducts in rat liver cytosol and observed an initial deglycosylation of G-derived etheno adducts followed by oxidation of 1,

-G to 2-oxo-1, - -G and of the corresponding substituted adduct, heptanone-1, - -G, to 2-oxoheptanone-1, - -G (Figure 3) [66]. This raises the possibility that urinary biomarker studies may be underestimating the true level of etheno adducts as a result of loss of the 2-deoxynucleoside forms. Further, the oxidized free-base forms may also be useful as biomarkers if they are excreted at high enough levels.


3.3. M1dG

This mutagenic pyrimidopurinone adduct of dG (Figure 1) forms in reactions of DNA with the lipid peroxidation product, malondialdehyde, and with base propenals derived from 4′-oxidation of 2-deoxyribose in DNA [56, 67–72]. As an endogenous DNA adduct, M1dG has been detected at levels ranging from 1 to 1000 lesions per 10 8 nucleotides in a variety of organisms, including humans [67, 71, 73–79]. Recent studies suggest that the major source of M1dG in vivo is base propenals from DNA oxidation [67], which is consistent with the higher reactivity of base propenals than malondialdehyde [68, 69] and the proximity of base propenals to dG in DNA. However, contributions from both malondialdehyde and base propenals are likely to occur in an oxidant-, cell-, and tissue-dependent manner [72].

In terms of the biological fate of M1dG, the adduct has been demonstrated to be a substrate for nucleotide excision repair (NER) [80, 81], which may explain the appearance of M1dG in human and rodent urine [27–29, 79]. However, M1dG was detectable in the human urine at levels of 10–20 fmol per kg per 24 h [79], which is a significantly lower excretion rate than other DNA lesions such as 8-oxo-dG (400 pmol per kg per 24 h) [82]. To explore the basis for this low rate of excretion, Marnett and coworkers undertook metabolic and pharmacokinetic studies of M1dG in rats [27]. When intravenously administrated to rats, M1dG was rapidly eliminated from the plasma with a half-life of 10 min [27]. In contrast to the rapid clearance from blood, M1dG was found in the urine for more than 24 hr after dosing, which suggested a rapid distribution to tissue followed by slower phase of excretion. Analysis of the urine revealed a metabolite of M1dG, 6-oxo-M1dG, likely derived from hepatic xanthine oxidase activity [27]. Studies in rat liver extracts revealed further oxidation of 6-oxo-M1dG on the imidazole ring to give 2,6-dioxo-M1G (Figure 4) [28]. While most of the M1dG was excreted unchanged in the urine and the problem of low levels of excretion remains unsolved, these studies point to the importance of defining the biological and metabolic fate of damaged biomolecules in efforts to develop biomarkers of inflammation and oxidative stress.


4. The Biological and Metabolic Fates of 2-Deoxyribose Oxidation Products

In addition to the nucleobases in DNA, the 2-deoxyribose moiety is also subjected to oxidative damage that merits consideration of biological fate and metabolism [9]. As opposed to the concept of simple “strand breaks,” growing evidence points to 2-deoxyribose oxidation in DNA as a critical determinant of the toxicity of oxidative stress [9]. Oxidation of each of the five positions in 2-deoxyribose in DNA occurs with an initial hydrogen atom abstraction to form a carbon-centered radical that rapidly adds molecular oxygen to form an unstable peroxyl radical. The resulting product spectra for 2-deoxyribose oxidation under aerobic conditions are shown in Figure 2 [9]. Many of these oxidation products are highly electrophilic, with α,β-unsaturated carbonyl motifs, and are thus capable of reacting with proximate nucleophilic sites in DNA, RNA, and proteins to form adducts [9]. This section of the paper will focus on the biological and metabolic or, more broadly, biotransformational fates of 2-deoxyribose oxidation products.

4.1. DNA Adducts of 2-Deoxyribose Oxidation Products

One fate of DNA oxidation products is reaction with local electrophiles to form protein and nucleic acids adducts. In this regard, oxidation of 2-deoxyribose in DNA produces a variety of reactive electrophilic species (Figure 2) that readily form adducts with neighboring DNA bases. Oxidation of both the 2′- and 3′-positions of 2-deoxyribose can lead to the formation of the 2-phosphoglycolaldehyde residue (Figure 2), the latter directly from the oxidation [83, 84] and the former by an induced and indirect oxidation mechanism involving an erythrose intermediate [85, 86]. By either mechanism, 2-phosphoglycolaldehyde undergoes a relatively slow phosphate-phosphonate rearrangement to generate the ubiquitous lipid and carbohydrate oxidation product, glyoxal, that reacts with dG and DNA to form diastereomeric 1,N 2 -glyoxal adducts of dG (Figure 5) [83].


At the 4′- and 5′-positions, 4′-oxidation generates base propenals that readily react with neighboring dG to form the pyrimidopurinone adduct, M1dG, as described earlier [67–69]. Oxidation of the 5′-position leads to formation of a 2-phosphoryldioxobutane residue that, possibly following β-elimination to form an α,β-unsaturated trans-dioxobutene species, reacts with dC

dG>dA to form bicyclic oxadiazabicyclo-(3.3.0)octaimine adducts (Figure 6) [87–91].


4.2. Protein Adducts of 2-Deoxyribose Oxidation Products

In addition to DNA adducts, the electrophiles derived from 2-deoxyribose oxidation react with amino acid side chains in proteins to form a variety of adducts, some with functional consequences. One of the earliest examples of protein adducts from 2-deoxyribose oxidation involves the 1′-position. The 2-deoxyribonolactone abasic site resulting from 1′-oxidation in DNA reacts with DNA repair proteins to form stable protein-DNA cross-links [92, 93]. This phenomenon was first demonstrated by Hashimoto et al. with the E. coli DNA BER enzyme endonuclease III [92]. This enzyme normally functions in base excision repair pathways with both an initial N-glycosylase activity against oxidized pyrimidines and a subsequent incision of the resulting abasic site by a lyase activity [94]. Upon binding to the 2-deoxyribonolactone abasic site, however, the active site (lysine 120), which normally forms a Schiff base with the 1′-aldehyde in the ring-opened form of the native abasic site, performs a nucleophilic attack on the carbonyl group of the lactone ring (Figure 7). Unlike a Schiff base, the resulting cross-link is irreversible and complicates the DNA repair process [92]. DeMott et al. observed similar results in which a covalent amide bond was formed by the 1′-carbon of the lactone and the lysine 72 in human polymerase β [93]. Additionally, the 2-deoxyribonolactone undergoes a rate-limiting β-elimination reaction to form a butenolide species with a half-life of 20 h in single-stranded DNA (32–54 h in duplex DNA), followed by a rapid δ-elimination to release 5-methylene-2(5H)-furanone [95]. Both the intermediate butenolide and the product methylenefuranone are electrophilic species capable of reaction with nucleophilic sites in DNA and protein, and possibly subject to metabolic reactions such as glutathione conjugation.


Another potential source of protein adducts arises from the variety of α,β-unsaturated carbonyl and dicarbonyl products of 2-deoxyribose oxidation in DNA. The potential here lies in the high concentration of nucleophilic lysine and arginine residues in histone proteins proximate to the sites of DNA damage and in the well-established reactivity of α,β-unsaturated carbonyl and dicarbonyl species with nucleophilic amino acids, which is perhaps best illustrated by lipid peroxidation products (e.g., [96–103]. Several recent studies have identified specific lysine and histidine adducts of well-defined lipid peroxidation products such as malondialdehyde [100], 4-hydroxynonenal [99], and its oxidation product, 4-oxononenal [97] (Figure 8). The reactions forming these adducts are highly analogous to reactions that could occur with 2-deoxyribose oxidation products, as illustrated in Figure 8. For example, the unsaturated β-elimination product of the 2-deoxypentose-4-ulose product of 4′-oxidation of deoxyribose is a chemical analog of 4-oxononenal derived from lipid peroxidation. It would thus be expected to react with lysines and histidines in histone and other chromatin proteins to form the bis-adduct or cross-link observed by observed by Sayre and coworkers [104] and the stable furan derivative observed by observed by Blair and coworkers [97], respectively (Figure 8). Indeed, histones 2A, 2B, and 3 contain 3–5 histidines that have been exploited to cross-link histones to DNA in the classic studies of Mirzabekov and coworkers [105, 106].


The malondialdehyde adducts of lysine, arginine, and histidine represent another protein adduct chemistry with potential parallels between 2-deoxyribose oxidation and lipid peroxidation. The reaction of lysine by nucleophilic substitution yields a moderately stable N-propenal-lysine species (Figure 8) that can react with another lysine to form a propyl-bridged cross-link [107], while the reaction of malondialdehyde with arginine has been shown to produce a stable pyrimidyl-ornithine species (Figure 8) [107]. In both cases, the proportions of modified amino acids are high [108]. Given the analogous reactions of malondialdehyde and base propenals from 4′-oxidation, it is reasonable to expect the formation of propyl-bridged cross-links and pyrimidyl-ornithine species in histone proteins in cells subjected to oxidative stresses.

A final example of protein adducts derived from 2-deoxyribose oxidation products involves N-formylation of lysine by transfer of formyl groups from 3′-formylphosphate residues (Figure 9) [109], among other possible sources such as oxidation of formaldehyde adducts of lysine. N 6 -formyllysine was detected in histone proteins from a variety of sources to the extent of 0.04%–0.1% of all lysines in acid-soluble chromatin proteins including histones, which suggests that the adduct represents an endogenous secondary modification of histones [109]. The chemical analogy of the N-formyl modification to the physiologically important lysine N-acetylation and N-methylation suggests that lysine N-formylation may interfere with signaling mediated by histone and other chromatin protein modifications (e.g., [110, 111]).


In all of these cases, the adducted proteins are subject to degradation, with the potential for the release and excretion of adducted peptides or amino acids. Their potential as biomarkers warrants further study of DNA-derived protein adducts.

4.3. Metabolism of 2-Deoxyribose Oxidation Products

As in the case of nucleobase lesions, the products of 2-deoxyribose oxidation of DNA must also be considered as substrates for metabolic enzymes and biotransformational reactions. This is all the more apparent given the electrophilic nature of the products, which points to glutathione (GSH) adduct formation, and the α,β-unsaturated carbonyl structure of many of the products, which makes them ideal substrates for glutathione S-transferases (GST) [34]. Indeed, GSTs have been shown to react with α,β-unsaturated aldehyde-containing lipid peroxidation products, many of which are chemical analogues of 2-deoxyribose oxidation products [9, 68]. Two examples of GST reactions with 2-deoxyribose oxidation products illustrate this biotransformation concept.

The first example involves GSH conjugation of base propenals. One of the classic definitions of GST substrates is that they must react directly with GSH to a measurable extent [34]. This is indeed the case with base propenals, as demonstrated in studies by Berhane et al. in which GSH added to give a Michael adduct and a substitution product with loss of the nucleobase (Figure 10) [30]. In addition, base propenals were found to be among the best substrates for the Pi class of GSTs, producing a single GSH conjugate (Figure 10).


GSH conjugates have also been identified for furan metabolite cis-1,4-dioxo-2-butene [31, 32], the conformational isomer of the trans-1,4-dioxo-2-butene product of 5′-oxidation (Figure 2). Given the similarity in the reactivity of cis- and trans-1,4-dioxo-2-butene toward DNA adduct formation [9], it would not be surprising to identify GSH adducts of the trans-isomer product of 2-deoxyribose oxidation, as has been observed in vitro and in vivo with the cis-isomer derivative of furan metabolism [31, 32, 112].

5. Prospects

Molecules damaged during normal physiological processes and in states of pathology represent a large source of biomarkers with potential clinical utility in defining etiological mechanisms, quantifying the risk of human disease and studying interindividual variations in cellular repair pathways. In spite of this potential, there has been little effort to define the biotransformational fate of damaged biomolecules as they move from the site of formation to excretion in clinically accessible compartments. This paper has highlighted examples of this important problem with DNA damage products. Coupled with the development of more sensitive and specific analytical technologies, there are likely to be major advancements in defining the metabolism of DNA damage products and other damaged biomolecules in the coming years.

References

  1. L. A. Loeb and C. C. Harris, “Advances in chemical carcinogenesis: a historical review and prospective,” Cancer Research, vol. 68, no. 17, pp. 6863–6872, 2008. View at: Publisher Site | Google Scholar
  2. D. E. Barnes and T. Lindahl, “Repair and genetic consequences of endogenous DNA base damage in mammalian cells,” Annual Review of Genetics, vol. 38, pp. 445–476, 2004. View at: Publisher Site | Google Scholar
  3. J. C. Delaney and J. M. Essigmann, “Biological properties of single chemical-DNA adducts: a twenty year perspective,” Chemical Research in Toxicology, vol. 21, no. 1, pp. 232–252, 2008. View at: Publisher Site | Google Scholar
  4. G. P. Pfeifer, “Mutagenesis at methylated CpG sequences,” Current Topics in Microbiology and Immunology, vol. 301, pp. 259–281, 2006. View at: Publisher Site | Google Scholar
  5. R. De Bont and N. van Larebeke, “Endogenous DNA damage in humans: a review of quantitative data,” Mutagenesis, vol. 19, no. 3, pp. 169–185, 2004. View at: Publisher Site | Google Scholar
  6. B. Epe, “Role of endogenous oxidative DNA damage in carcinogenesis: what can we learn from repair-deficient mice?” Biological Chemistry, vol. 383, no. 3-4, pp. 467–475, 2002. View at: Publisher Site | Google Scholar
  7. L. J. Marnett, “Oxyradicals and DNA damage,” Carcinogenesis, vol. 21, no. 3, pp. 361–370, 2000. View at: Google Scholar
  8. P. C. Dedon and S. R. Tannenbaum, “Reactive nitrogen species in the chemical biology of inflammation,” Archives of Biochemistry and Biophysics, vol. 423, no. 1, pp. 12–22, 2004. View at: Publisher Site | Google Scholar
  9. P. C. Dedon, “The chemical toxicology of 2-deoxyribose oxidation in DNA,” Chemical Research in Toxicology, vol. 21, no. 1, pp. 206–219, 2008. View at: Publisher Site | Google Scholar
  10. D. Pluskota-Karwatka, “Modifications of nucleosides by endogenous mutagens-DNA adducts arising from cellular processes,” Bioorganic Chemistry, vol. 36, no. 4, pp. 198–213, 2008. View at: Publisher Site | Google Scholar
  11. Y. Wang, “Bulky DNA lesions induced by reactive oxygen species,” Chemical Research in Toxicology, vol. 21, no. 2, pp. 276–281, 2008. View at: Publisher Site | Google Scholar
  12. U. Nair, H. Bartsch, and J. Nair, “Lipid peroxidation-induced DNA damage in cancer-prone inflammatory diseases: a review of published adduct types and levels in humans,” Free Radical Biology and Medicine, vol. 43, no. 8, pp. 1109–1120, 2007. View at: Publisher Site | Google Scholar
  13. S. J. Sturla, “DNA adduct profiles: chemical approaches to addressing the biological impact of DNA damage from small molecules,” Current Opinion in Chemical Biology, vol. 11, no. 3, pp. 293–299, 2007. View at: Publisher Site | Google Scholar
  14. M. S. Cooke, R. Olinski, and S. Loft, “Measurement and meaning of oxidatively modified DNA lesions in urine,” Cancer Epidemiology Biomarkers and Prevention, vol. 17, no. 1, pp. 3–14, 2008. View at: Publisher Site | Google Scholar
  15. M. D. Evans, M. Dizdaroglu, and M. S. Cooke, “Oxidative DNA damage and disease: induction, repair and significance,” Mutation Research, vol. 567, no. 1, pp. 1–61, 2004. View at: Publisher Site | Google Scholar
  16. J. Nair, P. Srivatanakul, C. Haas et al., “High urinary excretion of lipid peroxidation-derived DNA damage in patients with cancer-prone liver diseases,” Mutation Research, vol. 683, pp. 23–28, 2009. View at: Publisher Site | Google Scholar
  17. S. Dechakhamphu, S. Pinlaor, P. Sitthithaworn, J. Nair, H. Bartsch, and P. Yongvanit, “Lipid peroxidation and etheno DNA adducts in white blood cells of liver fluke-infected patients: protection by plasma α-tocopherol and praziquantel,” Cancer Epidemiology Biomarkers and Prevention, vol. 19, no. 1, pp. 310–318, 2010. View at: Publisher Site | Google Scholar
  18. S. Dechakhamphu, P. Yongvanit, J. Nair, S. Pinlaor, P. Sitthithaworn, and H. Bartsch, “High excretion of etheno adducts in liver fluke-infected patients: protection by Praziquantel against DNA damage,” Cancer Epidemiology Biomarkers and Prevention, vol. 17, no. 7, pp. 1658–1664, 2008. View at: Publisher Site | Google Scholar
  19. M. Meerang, J. Nair, P. Sirankapracha et al., “Increased urinary 1, N 6 -ethenodeoxyadenosine and 3, N 4 -ethenodeoxycytidine excretion in thalassemia patients: markers for lipid peroxidation-induced DNA damage,” Free Radical Biology and Medicine, vol. 44, no. 10, pp. 1863–1868, 2008. View at: Publisher Site | Google Scholar
  20. H.-J. C. Chen and C.-F. Kao, “Effect of gender and cigarette smoking on urinary excretion of etheno DNA adducts in humans measured by isotope dilution gas chromatography/mass spectrometry,” Toxicology Letters, vol. 169, no. 1, pp. 72–81, 2007. View at: Publisher Site | Google Scholar
  21. P. R. Hillestrøm, M.-I. Covas, and H. E. Poulsen, “Effect of dietary virgin olive oil on urinary excretion of etheno-DNA adducts,” Free Radical Biology and Medicine, vol. 41, no. 7, pp. 1133–1138, 2006. View at: Publisher Site | Google Scholar
  22. J. Son, B. Pang, J. L. McFaline, K. Taghizadeh, and P. C. Dedon, “Surveying the damage: the challenges of developing nucleic acid biomarkers of inflammation,” Molecular BioSystems, vol. 4, no. 9, pp. 902–908, 2008. View at: Publisher Site | Google Scholar
  23. M. D. Evans, R. Olinski, S. Loft et al., “Toward consensus in the analysis of urinary 8-oxo-7,8-dihydro- 2 ′ - deoxyguanosine as a noninvasive biomarker of oxidative stress,” FASEB Journal, vol. 24, no. 4, pp. 1249–1260, 2010. View at: Publisher Site | Google Scholar
  24. G. B. Elion, A. Kovensky, G. H. Hitchings, E. Metz, and R. W. Rundles, “Metabolic studies of allopurinol, an inhibitor of xanthine oxidase,” Biochemical Pharmacology, vol. 15, no. 7, pp. 863–880, 1966. View at: Google Scholar
  25. S. K. Roy, K. R. Korzekwa, F. J. Gonzalez, R. C. Mosche, and M. E. Dolan, “Human liver oxidative metabolism of O 6 -benzylguanine,” Biochemical Pharmacology, vol. 50, no. 9, pp. 1385–1389, 1995. View at: Publisher Site | Google Scholar
  26. D.-H. Kim and F. P. Guengerich, “Excretion of the mercapturic acid S-[2-( N 7 -guanyl)ethyl]N-acetylcysteine in urine following administration of ethylene dibromide to rats,” Cancer Research, vol. 49, no. 21, pp. 5843–5847, 1989. View at: Google Scholar
  27. M. B. Otteneder, C. G. Knutson, J. S. Daniels et al., “In vivo oxidative metabolism of a major peroxidation-derived DNA adduct, M1dG,” Proceedings of the National Academy of Sciences of the United States of America, vol. 103, no. 17, pp. 6665–6669, 2006. View at: Publisher Site | Google Scholar
  28. C. G. Knutson, D. Akingbade, B. C. Crews, M. Voehler, D. F. Stec, and L. J. Marnett, “Metabolism in vitro and in vivo of the DNA base adduct, M1G,” Chemical Research in Toxicology, vol. 20, no. 3, pp. 550–557, 2007. View at: Publisher Site | Google Scholar
  29. C. G. Knutson, P. L. Skipper, R. G. Liberman, S. R. Tannenbaum, and L. J. Marnett, “Monitoring in vivo metabolism and elimination of the endogenous DNA adduct, M1dG <3-(2-Deoxy-β-D-erythropentofuranosyl)pyrimido[1,2- α]purin-10(3H)-one>, by accelerator mass spectrometry,” Chemical Research in Toxicology, vol. 21, no. 6, pp. 1290–1294, 2008. View at: Publisher Site | Google Scholar
  30. K. Berhane, M. Widersten, Å. Engström, J. W. Kozarich, and B. Mannervik, “Detoxication of base propenals and other α , β -unsaturated aldehyde products of radical reactions and lipid peroxidation by human glutathione transferases,” Proceedings of the National Academy of Sciences of the United States of America, vol. 91, no. 4, pp. 1480–1484, 1994. View at: Google Scholar
  31. L.-J. Chen, S. S. Hecht, and L. A. Peterson, “Characterization of amino acid and glutathione adducts of cis-2-butene- 1,4-dial, a reactive metabolite of furan,” Chemical Research in Toxicology, vol. 10, no. 8, pp. 866–874, 1997. View at: Publisher Site | Google Scholar
  32. L. A. Peterson, M. E. Cummings, C. C. Vu, and B. A. Matter, “Glutathione trapping to measure microsomal oxidation of furan to cis-2-butene-1,4-dial,” Drug Metabolism and Disposition, vol. 33, no. 10, pp. 1453–1458, 2005. View at: Publisher Site | Google Scholar
  33. A. F. Nassar and P. F. H. JoAnn Scatina, Drug Metabolism Handbook, John Wiley & Sons, River Street Hoboken, NJ, USA, 2009.
  34. L. J. Casarett, J. Doull, and C. D. Klaassen, Casarett and Doull's Toxicology: The Basic Science of Poisons, McGraw-Hill, New York, NY, USA, 7th edition, 2008.
  35. J. Cadet, T. Douki, and J.-L. Ravanat, “Oxidatively generated damage to the guanine moiety of DNA: mechanistic aspects and formation in cells,” Accounts of Chemical Research, vol. 41, no. 8, pp. 1075–1083, 2008. View at: Publisher Site | Google Scholar
  36. M. S. Cooke, P. T. Henderson, and M. D. Evans, “Sources of extracellular, oxidatively-modified DNA lesions: implications for their measurement in urine,” Journal of Clinical Biochemistry and Nutrition, vol. 45, no. 3, pp. 255–270, 2009. View at: Publisher Site | Google Scholar
  37. J. C. Niles, J. S. Wishnok, and S. R. Tannenbaum, “Peroxynitrite-induced oxidation and nitration products of guanine and 8-oxoguanine: structures and mechanisms of product formation,” Nitric Oxide, vol. 14, no. 2, pp. 109–121, 2006. View at: Publisher Site | Google Scholar
  38. J. Cadet, T. Douki, and J.-L. Ravanat, “One-electron oxidation of DNA and inflammation processes,” Nature Chemical Biology, vol. 2, no. 7, pp. 348–349, 2006. View at: Publisher Site | Google Scholar
  39. S. Steenken and S. V. Jovanovic, “How easily oxidizable is DNA? One-electron reduction potentials of adenosine and guanosine radicals in aqueous solution,” Journal of the American Chemical Society, vol. 119, no. 3, pp. 617–618, 1997. View at: Publisher Site | Google Scholar
  40. M. K. Shigenaga, E. N. Aboujaoude, Q. Chen, and B. N. Ames, “Assays of oxidative DNA damage biomarkers 8-oxo- 2 ′ -deoxyguanosine and 8- oxoguanine in nuclear DNA and biological fluids by high-performance liquid chromatography with electrochemical detection,” Methods in Enzymology, vol. 234, pp. 16–33, 1994. View at: Publisher Site | Google Scholar
  41. K. Broedbaek, H. E. Poulsen, A. Weimann et al., “Urinary excretion of biomarkers of oxidatively damaged DNA and RNA in hereditary hemochromatosis,” Free Radical Biology and Medicine, vol. 47, no. 8, pp. 1230–1233, 2009. View at: Publisher Site | Google Scholar
  42. Y. Nakabeppu, S. Oka, Z. Sheng, D. Tsuchimoto, and K. Sakumi, “Programmed cell death triggered by nucleotide pool damage and its prevention by MutT homolog-1 (MTH1) with oxidized purine nucleoside triphosphatase,” Mutation Research, vol. 703, no. 1, pp. 51–58, 2010. View at: Publisher Site | Google Scholar
  43. A. G. McLennan, “The Nudix hydrolase superfamily,” Cellular and Molecular Life Sciences, vol. 63, no. 2, pp. 123–143, 2006. View at: Publisher Site | Google Scholar
  44. T. Paz-Elizur, Z. Sevilya, Y. Leitner-Dagan, D. Elinger, L. C. Roisman, and Z. Livneh, “DNA repair of oxidative DNA damage in human carcinogenesis: potential application for cancer risk assessment and prevention,” Cancer Letters, vol. 266, no. 1, pp. 60–72, 2008. View at: Publisher Site | Google Scholar
  45. S. Maynard, S. H. Schurman, C. Harboe, N. C. de Souza-Pinto, and V. A. Bohr, “Base excision repair of oxidative DNA damage and association with cancer and aging,” Carcinogenesis, vol. 30, no. 1, pp. 2–10, 2009. View at: Publisher Site | Google Scholar
  46. J. I. Friedman and J. T. Stivers, “Detection of damaged DNA bases by DNA glycosylase enzymes,” Biochemistry, vol. 49, no. 24, pp. 4957–4967, 2010. View at: Publisher Site | Google Scholar
  47. S. S. David, V. L. O'Shea, and S. Kundu, “Base-excision repair of oxidative DNA damage,” Nature, vol. 447, no. 7147, pp. 941–950, 2007. View at: Publisher Site | Google Scholar
  48. D. M. Stanbury, “Reduction potentials involving inorganic free radicals in aqueous solution,” Advances in Inorganic Chemistry, vol. 33, no. C, pp. 69–138, 1989. View at: Publisher Site | Google Scholar
  49. T. N. Das, T. Dhanasekaran, Z. B. Alfassi, and P. Neta, “Reduction potential of the tert-butylperoxyl radical in aqueous solutions,” in Proceedings of the 4th International Conference on Chemical Kinetics, National Institute of Standards and Technology, Gaithersburg, Md, USA, 1997. View at: Google Scholar
  50. J. M. Lee, J. C. Niles, J. S. Wishnok, and S. R. Tannenbaum, “Peroxynitrite reacts with 8-oitropurines to yield 8-oxopurines,” Chemical Research in Toxicology, vol. 15, no. 1, pp. 7–14, 2002. View at: Publisher Site | Google Scholar
  51. R. S. Hall, A. A. Fedorov, R. Marti-Arbona et al., “The hunt for 8-oxoguanine deaminase,” Journal of the American Chemical Society, vol. 132, no. 6, pp. 1762–1763, 2010. View at: Publisher Site | Google Scholar
  52. M. Dong and P. C. Dedon, “Relatively small increases in the steady-state levels of nucleobase deamination products in DNA from human TK6 cells exposed to toxic levels of nitric oxide,” Chemical Research in Toxicology, vol. 19, no. 1, pp. 50–57, 2006. View at: Publisher Site | Google Scholar
  53. M. Dong, C. Wang, W. M. Deen, and P. C. Dedon, “Absence of 2 ′ -deoxyoxanosine and presence of abasic sites in DNA exposed to nitric oxide at controlled physiological concentrations,” Chemical Research in Toxicology, vol. 16, no. 9, pp. 1044–1055, 2003. View at: Publisher Site | Google Scholar
  54. B. Pang, X. Zhou, H. Yu et al., “Lipid peroxidation dominates the chemistry of DNA adduct formation in a mouse model of inflammation,” Carcinogenesis, vol. 28, no. 8, pp. 1807–1813, 2007. View at: Publisher Site | Google Scholar
  55. K. Taghizadeh, J. L. McFaline, B. Pang et al., “Quantification of DNA damage products resulting from deamination, oxidation and reaction with products of lipid peroxidation by liquid chromatography isotope dilution tandem mass spectrometry,” Nature Protocols, vol. 3, no. 8, pp. 1287–1298, 2008. View at: Publisher Site | Google Scholar
  56. L. J. Marnett, “Oxy radicals, lipid peroxidation and DNA damage,” Toxicology, vol. 181-182, pp. 219–222, 2002. View at: Publisher Site | Google Scholar
  57. T. J. Montine, M. D. Neely, J. F. Quinn et al., “Lipid peroxidation in aging brain and Alzheimer's disease,” Free Radical Biology and Medicine, vol. 33, no. 5, pp. 620–626, 2002. View at: Publisher Site | Google Scholar
  58. H. Bartsch and J. Nair, “Oxidative stress and lipid peroxidation-derived DNA-lesions in inflammation driven carcinogenesis,” Cancer Detection and Prevention, vol. 28, no. 6, pp. 385–391, 2004. View at: Publisher Site | Google Scholar
  59. M. V. Williams, H. L. Seon, M. Pollack, and I. A. Blair, “Endogenous lipid hydroperoxide-mediated DNA-adduct formation in min mice,” The Journal of Biological Chemistry, vol. 281, no. 15, pp. 10127–10133, 2006. View at: Publisher Site | Google Scholar
  60. S. H. Lee, M. V. Williams, R. N. DuBois, and I. A. Blair, “Cyclooxygenase-2-mediated DNA damage,” The Journal of Biological Chemistry, vol. 280, no. 31, pp. 28337–28346, 2005. View at: Publisher Site | Google Scholar
  61. K. Schmid, J. Nair, G. Winde, I. Velic, and H. Bartsch, “Increased levels of promutagenic etheno-DNA adducts in colonic polyps of FAP patients,” International Journal of Cancer, vol. 87, no. 1, pp. 1–4, 2000. View at: Google Scholar
  62. J. Nair, A. Barbin, and H. Bartsch, “Etheno DNA-base adducts from endogenous reactive species,” Mutation Research, vol. 424, no. 1-2, pp. 59–69, 1999. View at: Publisher Site | Google Scholar
  63. S. H. Lee, J. A. Arora, T. Oe, and I. A. Blair, “4-Hydroperoxy-2-nonenal-induced formation of 1,N 2 -etheno- 2 ′ -deoxyguanosine adducts,” Chemical Research in Toxicology, vol. 18, no. 4, pp. 780–786, 2005. View at: Publisher Site | Google Scholar
  64. H.-J.C. Chen, G.-J. Lin, and W.-P. Lin, “Simultaneous quantification of three lipid peroxidation-derived etheno adducts in human DNA by stable isotope dilution nanoflow liquid chromatography nanospray ionization tandem mass spectrometry,” Analytical Chemistry, vol. 82, no. 11, pp. 4486–4493, 2010. View at: Publisher Site | Google Scholar
  65. L. Gros, A. A. Ishchenko, and M. Saparbaev, “Enzymology of repair of etheno-adducts,” Mutation Research, vol. 531, no. 1-2, pp. 219–229, 2003. View at: Publisher Site | Google Scholar
  66. C. G. Knutson, E. H. Rubinson, D. Akingbade et al., “Oxidation and glycolytic cleavage of etheno and propano DNA base adducts,” Biochemistry, vol. 48, no. 4, pp. 800–809, 2009. View at: Publisher Site | Google Scholar
  67. X. Zhou, K. Taghizadeh, and P. C. Dedon, “Chemical and biological evidence for base propenals as the major source of the endogenous M1dG adduct in cellular DNA,” The Journal of Biological Chemistry, vol. 280, no. 27, pp. 25377–25382, 2005. View at: Publisher Site | Google Scholar
  68. P. C. Dedon, J. P. Plastaras, C. A. Rouzer, and L. J. Marnett, “Indirect mutagenesis by oxidative DNA damage: formation of the pyrimidopurinone adduct of deoxyguanosine by base propenal,” Proceedings of the National Academy of Sciences of the United States of America, vol. 95, no. 19, pp. 11113–11116, 1998. View at: Publisher Site | Google Scholar
  69. J. P. Plastaras, J. N. Riggins, M. Otteneder, and L. J. Marnett, “Reactivity and mutagenicity of endogenous DNA oxopropenylating agents: base propenals, malondialdehyde, and N ε -oxopropenyllysine,” Chemical Research in Toxicology, vol. 13, no. 12, pp. 1235–1242, 2000. View at: Publisher Site | Google Scholar
  70. A. K. Basu, S. M. O'Hara, P. Valladier, K. Stone, O. Mols, and L. J. Marnett, “Identification of adducts formed by reaction of guanine nucleosides with malondialdehyde and structurally related aldehydes,” Chemical Research in Toxicology, vol. 1, no. 1, pp. 53–59, 1988. View at: Google Scholar
  71. A. K. Chaudhary, M. Nokubo, G. R. Reddy et al., “Detection of endogenous malondialdehyde-deoxyguanosine adducts in human liver,” Science, vol. 265, no. 5178, pp. 1580–1582, 1994. View at: Google Scholar
  72. Y.-C. Jeong and J. A. Swenberg, “Formation of M1G-dR from endogenous and exogenous ROS-inducing chemicals,” Free Radical Biology and Medicine, vol. 39, no. 8, pp. 1021–1029, 2005. View at: Publisher Site | Google Scholar
  73. C. A. Rouzer, A. K. Chaudhary, M. Nokubo et al., “Analysis of the malondialdehyde- 2 ′ -deoxyguanosine adduct pyrimidopurinone in human leukocyte DNA by gas chromatography/electron capture/negative chemical ionization/mass spectrometry,” Chemical Research in Toxicology, vol. 10, no. 2, pp. 181–188, 1997. View at: Publisher Site | Google Scholar
  74. A. K. Chaudhary, G. R. Reddy, I. A. Blair, and L. J. Marnett, “Characterization of an N 6 -oxopropenyl- 2 ′ -deoxyadenosine adduct in malondialdehyde-modified DNA using liquid chromatography/electrospray ionization tandem mass spectrometry,” Carcinogenesis, vol. 17, no. 5, pp. 1167–1170, 1996. View at: Google Scholar
  75. Y.-C. Jeong, R. Sangaiah, J. Nakamura et al., “Analysis of M1G-dR in DNA by aldehyde reactive probe labeling and liquid chromatography tandem mass spectrometry,” Chemical Research in Toxicology, vol. 18, no. 1, pp. 51–60, 2005. View at: Publisher Site | Google Scholar
  76. F. F. Kadlubar, K. E. Anderson, S. Häussermann et al., “Comparison of DNA adduct levels associated with oxidative stress in human pancreas,” Mutation Research, vol. 405, no. 2, pp. 125–133, 1998. View at: Publisher Site | Google Scholar
  77. C. Leuratti, R. Singh, C. Lagneau et al., “Determination of malondialdehyde-induced DNA damage in human tissues using an immunoslot blot assay,” Carcinogenesis, vol. 19, no. 11, pp. 1919–1924, 1998. View at: Google Scholar
  78. P. Yi, X. Sun, D. R. Doerge, and P. P. Fu, “An improved 32P-postlabeling/high-performance liquid chromatography method for the analysis of the malondialdehye-derived 1, N 6 - propanodeoxyguanosine DNA adduct in animal and human tissues,” Chemical Research in Toxicology, vol. 11, no. 9, pp. 1032–1041, 1998. View at: Publisher Site | Google Scholar
  79. A.-M. Hoberg, M. Otteneder, L. J. Marnett, and H. E. Poulsen, “Measurement of the malondialdehyde- 2 ′ -deoxy-guanosine adduct in human urine by immuno-extraction and liquid chromatography/atmospheric pressure chemical ionization tandem mass spectrometry,” Journal of Mass Spectrometry, vol. 39, no. 1, pp. 38–42, 2004. View at: Publisher Site | Google Scholar
  80. S. P. Fink, G. R. Reddy, and L. J. Marnett, “Mutagenicity in Escherichia coli of the major DNA adduct derived from the endogenous mutagen malondialdehyde,” Proceedings of the National Academy of Sciences of the United States of America, vol. 94, no. 16, pp. 8652–8657, 1997. View at: Publisher Site | Google Scholar
  81. L. J. Marnett, “Lipid peroxidation𠅍NA damage by malondialdehyde,” Mutation Research, vol. 424, no. 1-2, pp. 83–95, 1999. View at: Publisher Site | Google Scholar
  82. A. Weimann, D. Belling, and H. E. Poulsen, “Quantification of 8-oxo-guanine and guanine as the nucleobase, nucleoside and deoxynucleoside forms in human urine by high-performance liquid chromatography-electrospray tandem mass spectrometry,” Nucleic acids research, vol. 30, no. 2, p. E7, 2002. View at: Google Scholar
  83. M. Awada and P. C. Dedon, “Formation of the 1, N 2 -glyoxal adduct of deoxyguanosine by phosphoglycolaldehyde, a product of 3 ′ -deoxyribose oxidation in DNA,” Chemical Research in Toxicology, vol. 14, no. 9, pp. 1247–1253, 2001. View at: Publisher Site | Google Scholar
  84. G. A. Lahoud, A. L. Hitt, and A. C. Bryant-Friedrich, “Aerobic fate of the C- 3 ′ -thymidinyl radical in single-stranded DNA,” Chemical Research in Toxicology, vol. 19, no. 12, pp. 1630–1636, 2006. View at: Publisher Site | Google Scholar
  85. H. Sugiyama, Y. Tsutsumi, K. Fujimoto, and I. Saito, “Photoinduced deoxyribose C 2 ′ oxidation in DNA. Alkali-dependent cleavage of erythrose-containing sites via a retroaldol reaction,” Journal of the American Chemical Society, vol. 115, no. 11, pp. 4443–4448, 1993. View at: Google Scholar
  86. J. Kim, Y. N. Weledji, and M. M. Greenberg, “Independent generation and characterization of a C 2 ′ -oxidized abasic site in chemically synthesized oligonucleotides,” Journal of Organic Chemistry, vol. 69, no. 18, pp. 6100–6104, 2004. View at: Publisher Site | Google Scholar
  87. T. Bohnert, L. Gingipalli, and P. C. Dedon, “Reaction of 2 ′ -deoxyribonucleosides with cis- and trans-1,4-dioxo-2-butene,” Biochemical and Biophysical Research Communications, vol. 323, no. 3, pp. 838–844, 2004. View at: Publisher Site | Google Scholar
  88. B. Chen, T. Bohnert, X. Zhou, and P. C. Dedon, “ 5 ′ -(2-phosphoryl-1,4-dioxobutane) as a product of 5 ′ -oxidation of deoxyribose in DNA: elimination as trans-1,4-dioxo-2-butene and approaches to analysis,” Chemical Research in Toxicology, vol. 17, no. 11, pp. 1406–1413, 2004. View at: Publisher Site | Google Scholar
  89. B. Chen, C. C. Vu, M. C. Byrns, P. C. Dedon, and L. A. Peterson, “Formation of 1,4-dioxo-2-butene-derived adducts of 2 ′ -deoxyadenosine and 2 ′ -deoxycytidine in oxidized DNA,” Chemical Research in Toxicology, vol. 19, no. 8, pp. 982–985, 2006. View at: Publisher Site | Google Scholar
  90. M. C. Byrns, D. P. Predecki, and L. A. Peterson, “Characterization of nucleoside adducts of cis-2-butene-1,4-dial, a reactive metabolite of furan,” Chemical Research in Toxicology, vol. 15, no. 3, pp. 373–379, 2002. View at: Publisher Site | Google Scholar
  91. M. C. Byrns, C. C. Vu, J. W. Neidigh, J.-L. Abad, R. A. Jones, and L. A. Peterson, “Detection of DNA adducts derived from the reactive metabolite of furan, cis-2-butene-1,4-dial,” Chemical Research in Toxicology, vol. 19, no. 3, pp. 414–420, 2006. View at: Publisher Site | Google Scholar
  92. M. Hashimoto, M. M. Greenberg, Y. W. Kow, J.-T. Hwang, and R. P. Cunningham, “The 2-deoxyribonolactone lesion produced in DNA by neocarzinostatin and other damaging agents forms cross-links with the base-excision repair enzyme endonuclease III,” Journal of the American Chemical Society, vol. 123, no. 13, pp. 3161–3162, 2001. View at: Publisher Site | Google Scholar
  93. M. S. Demott, E. Beyret, D. Wong et al., “Covalent trapping of human DNA polymerase β by the oxidative DNA lesion 2-deoxyribonolactone,” The Journal of Biological Chemistry, vol. 277, no. 10, pp. 7637–7640, 2002. View at: Publisher Site | Google Scholar
  94. S. Boiteux, “Properties and biological functions of the NTH and FPG proteins of Escherichia coli: two DNA glycosylases that repair oxidative damage in DNA,” Journal of Photochemistry and Photobiology B, vol. 19, no. 2, pp. 87–96, 1993. View at: Publisher Site | Google Scholar
  95. Y. Zheng and T. L. Sheppard, “Half-life and DNA strand scission products of 2-deoxyribonolactone oxidative DNA damage lesions,” Chemical Research in Toxicology, vol. 17, no. 2, pp. 197–207, 2004. View at: Publisher Site | Google Scholar
  96. R. M. LoPachin, T. Gavin, D. R. Petersen, and D. S. Barber, “Molecular mechanisms of 4-hydroxy-2-nonenal and acrolein toxicity: nucleophilic targets and adduct formation,” Chemical Research in Toxicology, vol. 22, no. 9, pp. 1499–1508, 2009. View at: Publisher Site | Google Scholar
  97. A. K. Yocum, T. Oe, A. L. Yergey, and I. A. Blair, “Novel lipid hydroperoxide-derived hemoglobin histidine adducts as biomarkers of oxidative stress,” Journal of Mass Spectrometry, vol. 40, no. 6, pp. 754–764, 2005. View at: Publisher Site | Google Scholar
  98. L. M. Sayre, D. A. Zelasko, P. L. R. Harris, G. Perry, R. G. Salomon, and M. A. Smith, “4-Hydroxynonenal-derived advanced lipid peroxidation end products are increased in Alzheimer's disease,” Journal of Neurochemistry, vol. 68, no. 5, pp. 2092–2097, 1997. View at: Google Scholar
  99. M. Hashimoto, T. Sibata, H. Wasada, S. Toyokuni, and K. Uchida, “Structural basis of protein-bound endogenous aldehydes: chemical and immunochemical characterizations of configurational isomers of a 4-hydroxy-2-nonenal-histidine adduct,” The Journal of Biological Chemistry, vol. 278, no. 7, pp. 5044–5051, 2003. View at: Publisher Site | Google Scholar
  100. T. Ishii, S. Kumazawa, T. Sakurai, T. Nakayama, and K. Uchida, “Mass spectroscopic characterization of protein modification by malondialdehyde,” Chemical Research in Toxicology, vol. 19, no. 1, pp. 122–129, 2006. View at: Publisher Site | Google Scholar
  101. W. Siems and T. Grune, “Intracellular metabolism of 4-hydroxynonenal,” Molecular Aspects of Medicine, vol. 24, no. 4-5, pp. 167–175, 2003. View at: Publisher Site | Google Scholar
  102. D. L. Carbone, J. A. Doorn, Z. Kiebler, B. R. Ickes, and D. R. Petersen, “Modification of heat shock protein 90 by 4-hydroxynonenal in a rat model of chronic alcoholic liver disease,” Journal of Pharmacology and Experimental Therapeutics, vol. 315, no. 1, pp. 8–15, 2005. View at: Publisher Site | Google Scholar
  103. S. Yamada, T. Funada, N. Shibata et al., “Protein-bound 4-hydroxy-2-hexenal as a marker of oxidized n-3 polyunsaturated fatty acids,” Journal of Lipid Research, vol. 45, no. 4, pp. 626–634, 2004. View at: Publisher Site | Google Scholar
  104. W.-H. Zhang, J. Liu, G. Xu, Q. Yuan, and L. M. Sayre, “Model studies on protein side chain modification by 4-Oxo-2-nonenal,” Chemical Research in Toxicology, vol. 16, no. 4, pp. 512–523, 2003. View at: Publisher Site | Google Scholar
  105. G. A. Nacheva, D. Y. Guschin, O. V. Preobrazhenskaya, V. L. Karpov, K. K. Ebralidse, and A. D. Mirzabekov, “Change in the pattern of histone binding to DNA upon transcriptional activation,” Cell, vol. 58, no. 1, pp. 27–36, 1989. View at: Google Scholar
  106. K. K. Ebralidse, S. A. Grachev, and A. D. Mirzabekov, “A highly basic histone H4 domain bound to the sharply bent region of nucleosomal DNA,” Nature, vol. 331, no. 6154, pp. 365–367, 1988. View at: Google Scholar
  107. D. A. Slatter, C. H. Bolton, and A. J. Bailey, “The importance of lipid-derived malondialdehyde in diabetes mellitus,” Diabetologia, vol. 43, no. 5, pp. 550–557, 2000. View at: Publisher Site | Google Scholar
  108. D. A. Slatter, N. C. Avery, and A. J. Bailey, “Identification of a new cross-link and unique histidine adduct from bovine serum albumin incubated with malondialdehyde,” The Journal of Biological Chemistry, vol. 279, no. 1, pp. 61–69, 2004. View at: Publisher Site | Google Scholar
  109. T. Jiang, X. Zhou, K. Taghizadeh, M. Dong, and P. C. Dedon, “N-formylation of lysine in histone proteins as a secondary modification arising from oxidative DNA damage,” Proceedings of the National Academy of Sciences of the United States of America, vol. 104, no. 1, pp. 60–65, 2007. View at: Publisher Site | Google Scholar
  110. P. Close, C. Creppe, M. Gillard et al., “The emerging role of lysine acetylation of non-nuclear proteins,” Cellular and Molecular Life Sciences, vol. 67, no. 8, pp. 1255–1264, 2010. View at: Publisher Site | Google Scholar
  111. X.-J. Yang and E. Seto, “Lysine acetylation: codified crosstalk with other posttranslational modifications,” Molecular Cell, vol. 31, no. 4, pp. 449–461, 2008. View at: Publisher Site | Google Scholar
  112. L. A. Peterson, K. C. Naruko, and D. P. Predecki, “A reactive metabolite of furan, cis-2-butene-1,4-dial, is mutagenic in the Ames assay,” Chemical Research in Toxicology, vol. 13, no. 7, pp. 531–534, 2000. View at: Publisher Site | Google Scholar

Copyright

Copyright © 2010 Simon Wan Chan and Peter C. Dedon. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


Materials and Methods

Plasmids

DNA fragments encoding wild-type human IKKɛ (UniProt: Q14164) were amplified separately by PCR using primers containing Kpn1 (5′) and EcoR1 (3′) restriction sites.

  • For 5′-ttggtaccagccagctcagggcaggagatgcagagcacagccaatta-3′
  • Rev 5′-gatggatatctgcagaattcaggaggtgctgggactc-3′

The PCR products were double digested by these two enzymes and ligated to vector pcDNA5.5 (a kind gift from Dr Tencho Tenev), which provides a 2xHA tag at the c-terminal, to generate the pcDNA5.5-wt-IKKε plasmid. For mutant variants of IKKε with disrupted functional domains, the kinase domain mutant (KD-m) was created by site-directed mutagenesis, using primers to introduce a K38A mutation to the wild-type IKKε sequence.

UbLD-M-IKKɛ plasmid, encoding the ubiquitin-like domain mutant (UbLD-m) variant of IKKε (containing L353A and F354A mutations) was a kind gift from Prof Ivan Dikic. Both KD-m and UbLD-m IKKε variants were amplified and ligated into the pcDNA5.5 vector using the same Kpn1 and EcoR1 double digestion and ligation method as the wild-type kinase, generating the pcDNA5.5-KD-M-IKKε and pcDNA5.5-UbLD-M-IKKε plasmids.

Cells

To generate Flp-In 293 cells expressing either wild type, kinase domain mutant (KD-m) or ubiquitin-like domain mutant (UbLD-m) IKKε, Flp-In 293 cells (Invitrogen) were transfected with either pcDNA5.5-wt-IKKɛ, pcDNA5.5-KD-M-IKKɛ (K38A), pcDNA5.5-UbLD-M-IKKɛ (L353A F354A) or pcDNA5.5-GFP, together with a pOG44 plasmid at a molar ratio of 1:9. cDNA plasmids were mixed with Lipofectamine LTX (15338100, Thermo Fisher Scientific) or Fugene HD (E2311, Promega) according to the manufacturer's instruction and transfected into the different cell lines for 48 h.

Stable cell lines and single cell clones expressing wild-type IKKε (wt) or mutant IKKε, with disruption of either kinase domain or ubiquitin-like domain function (KD-m or UbLD-m), in a doxycycline-dependent manner were selected with 300 μg/ml hygromycin (Calbiochem). All Flp-In 293 cells were cultured in DMEM (Sigma-Aldrich). The panel of breast cancer cell lines were kindly provided by Dr. Alice Shia and Prof. Peter Schmid. MDA-MB-231, MDA-MB-468, MDA-MB-175, ZR75.1, T47D, HCC1143, MCF7 were cultured in RPMI-1640 (Sigma-Aldrich), Cal120 and MDA-MB-453 were cultured in DMEM (Sigma-Aldrich) and Sum44 in DMEM (Sigma-Aldrich) and 1 nM estrogen (Sigma-Aldrich). For all cell lines, medium was supplemented with 10% FBS, penicillin–streptomycin and Normocin (InvivoGeN). Serine-free medium was custom made DMEM without serine, with 10% dialysed FBS and penicillin–streptomycin. All cells were cultured with environmental conditions of 37°C, 5% CO2.

Drugs

The following drugs were used: 6-Diazo-5-oxo- l -norleucine (Don, D2141, Sigma-Aldrich) Sodium dichloroacetate (DCA, 347795, Sigma-Aldrich) NCT-502 and PHGDH inactive (19716 and 19717, Cayman) Doxycycline (Dox, D9891, Sigma-Aldrich), Oligomycin (sc-203342, Santa Cruz Biotechnology) FCCP (sc-203578, Santa Cruz Biotechnology) Antimycin (sc-202467, Santa Cruz Biotechnology) Rotenone (sc-203242, Santa Cruz Biotechnology) Cyt.C (C2037, Sigma-Aldrich) CB-839 (10-4556, Focus Biomolecules) Adenosine diphosphate (ADP, A2754, Sigma-Aldrich).

SiRNA transfection

The following oligos were transfected for siRNA-mediated knockdown: AllStars Negative Control siRNA (1027281, Qiagen) Hs_ATF4_9 FlexiTube siRNA (SI04236337, Qiagen) Hs_IKBKE_6 FlexiTube siRNA, (S102622319, Qiagen) Hs_IKBKE_7 FlexiTube siRNA (S102622326, Qiagen) Hs_IKBKE_8 FlexiTube siRNA (S102655317, Qiagen) Hs_IKBKE_9 FlexiTube siRNA (s102655324, Qiagen) Hs_IRF3_4 FlexiTube siRNA (SI02626526, Qiagen) Hs_PSAT1_10 FlexiTube siRNA (SI03019709, Qiagen, UK) Hs_PSAT1_12 FlexiTube siRNA (SI03222142, Qiagen, UK) Hs_PSAT1_14 FlexiTube siRNA (SI04265625, Qiagen, UK) Hs_PSAT1_15 FlexiTube siRNA (SI04272212, Qiagen, UK) Hs_RELA_5 FlexiTube siRNA (SI00301672, Qiagen).

For transfection, siRNA was mixed with Dharmafect 4 (T200402, Dharmacon), and cells were transfected according to the transfection reagent manufacturer's instruction for 48 h or 72 h prior to measurements. Cells were transfected with a final concentration of 50 nM siRNA, and a pool of all 4 IKBKE-targeting oligos was used for suppression of IKKε, a pool of all 4 PSAT1-targeting oligos was used for suppression of PSAT1, and single targeting oligos were used for the suppression of ATF4, p65 and IRF3.

Oxygen consumption and extracellular acidification rate measurements

An XF24 Extracellular or XF96e Extracellular Flux analyser (Seahorse Biosciences, Agilent Technologies) was used to determine the bioenergetic profiles in breast cancer cell lines. Cells were plated in six-well corning dishes first and then transfected with siRNA 24 h after plating. Twenty-four hours after transfection, cells were trypsinised, counted and plated into a 24 or 96-well Seahorse plate. Oxygen consumption rates (OCR) and extracellular acidification rates (ECAR) were assessed in Seahorse medium according to the manufacturer protocols. Respiratory parameters were assessed as described in Fig EV1B. Oxygen consumption rate (OCR) of Flp-In 293 cells was measured using an Oroboros high-resolution respirometer (Oroboros) at 37°C, in Seahorse XF assay medium containing 4.5 g/l glucose, 1 mM pyruvate and 25 mM Hepes, and the assay was performed as in Fig EV1B.

For measurements in isolated mitochondria, Flp-In 293 cells were first washed with PBS and collected in homogenisation buffer (250 mM sucrose, 5 mM Hepes, pH 7.4, 0.5 mM EGTA), and Protease inhibitor cocktail (1187358001, Roche) and then homogenised in a glass/glass, tight potter by 100 strokes on ice, followed by centrifugation for 5 min at 800 g at 4°C. The supernatant, containing mitochondria, was centrifuged again at 9,000 g. The pellet was resuspended and adjusted to a protein concentration of 0.8 mg/ml in OCB buffer (125 mM KCl, 20 mM MOPS, 10 mM Tris ph7.2–7.3, 0.2 mM EGTA, 2.5 mM KH2PO4, 2.5 mM MgCl2). 10 mM glutamate and 5 mM malate were added to the mitochondrial suspension before the experiment, and OCR was measured in OCB buffer using the Oroboros high-resolution respirometer. ADP (final concentration 0.25 mM), Cyt.C (10 μM), oligomycin (5 μM) were injected step by step, and 50 μM FCCP was added in 1 μl steps until maximum respiratory capacity was detected. At the end of the run, antimycin (5 μM final concentration) was injected. Data were then analysed by the Datalab 5.5 (Oroboros) software.

Cell proliferation assay

Cells were plated in Corning 96-well plates at a density between 2,000 and 10,000 cells per well for different cell lines. Cell proliferation rate was then measured using the IncuCyte ZOOM instrument (Essen Biosciences) for 3–7 days, and proliferation rate was analysed with the Incucyte Zoom 2015A software (Essen Biosciences).

Metabolic labelling and metabolome analysis

Flp-In 293 cells and breast cancer cell lines (T47D and MDA-MB-468) were first plated separately in six-well plates in five technical replicas per each condition. IKKε expression in Flp-In 293 cells was then induced by 50 ng/ml doxycycline, and breast cancer cells were transfected with siRNA to suppress IKKɛ. Two hours after induction for the Flp-In 293 cells, and 48 h after siRNA transfection for the breast cancer cell lines, cells were incubated with either 13 C6-glucose (CLM-1396-5, Cambridge Isotope Laboratories) medium or 15 N2-glutamine (NLM-1328-0.25, Cambridge Isotope Laboratories) medium for 14 h. Cells were then washed three times with PBS, and metabolites were extracted using cold extraction buffer (50% methanol, 30% acetonitrile, 20% ultrapure water, 100 ng/ml HEPES) at a ratio of 1 ml extraction buffer/10 6 cells. After 15-min incubation time on methanol and dry ice, cells were placed on a shaker for 15 min using a thermal mixer at 4°C and incubated for 1 h at −20°C. Cell lysates were centrifuged, and the supernatant was collected and transferred into autosampler glass vials, which were stored at −80°C until further analysis.

Samples were randomised in order to avoid bias due to machine drift and processed blindly. LC–MS analysis was performed using a Q Exactive Hybrid Quadrupole-Orbitrap mass spectrometer coupled to a Dionex U3000 UHPLC system (Thermo Fisher Scientific). The liquid chromatography system was fitted with a Sequant ZIC-pHILIC column (150 mm × 2.1 mm) and guard column (20 mm × 2.1 mm) from Merck Millipore and temperature maintained at 45°C. The mobile phase was composed of 20 mM ammonium carbonate and 0.1% ammonium hydroxide in water (solvent A) and acetonitrile (solvent B). The flow rate was set at 200 μl/min with the gradient described previously (Mackay et al, 2015 ). The mass spectrometer was operated in full MS and polarity switching mode. The acquired spectra were analysed using Xcalibur Qual Browser and Xcalibur Quan Browser software (Thermo Fisher Scientific).

Phosphoproteomics

Sample preparation

Flp-In 293 single cell clones for IKKɛ (Clones 1,2 and 3) and GFP (Clones 1,2 and 3) were seeded in 6-well plate in three replica for each condition. After 24 h of seeding, cells were induced with doxycycline for 16 h. Cells were first washed with ice-cold PBS containing 1 mM Na3VO4 and 1 mM NaF and then lysed in a lysis buffer containing 8M Urea, 20 mM HEPES, 1 mM Na3VO4, 1 mM NaF, 1 mM B-Glycerol phosphate and 0.25 mM Na2H2P2O7. After incubation on ice for 5 min, the cells were then scraped and collected in Eppendorf tubes and stored at −80°C. For sample analysis, cell lysates were thawed, protein digested with trypsin, and phosphopeptides were enriched using TiO2 as described in (Wilkes & Cutillas, 2017 ).

Nanoflow-liquid chromatography tandem mass spectrometry (LC–MS/MS)

Dried samples were dissolved in 0.1% TFA (0.5 μg/μl) and run in a LTQ-Orbitrap XL mass spectrometer (Thermo Fisher Scientific) connected to a nanoflow ultra-high pressure liquid chromatography (UPLC, NanoAcquity, Waters). Peptides were separated using a 75 μm × 150 mm column (BEH130 C18, 1.7 μm Waters) using solvent A (0.1% FA in LC–MS grade water) and solvent B (0.1% FA in LC–MS grade ACN) as mobile phases. The UPLC settings consisted of a sample loading flow rate of 2 μl/min for 8 min followed by a gradient elution with starting with 5% of solvent B and ramping up to 35% over 100 min followed by a 10-min wash at 85% B and a 15-min equilibration step at 1% B. The flow rate for the sample run was 300 nL/min with an operating back pressure of about 3,800 psi. Full scan survey spectra (m/z 375–1,800) were acquired in the Orbitrap with a resolution of 30,000 at m/z 400. A data-dependent analysis (DDA) was employed in which the five most abundant multiply charged ions present in the survey spectrum were automatically mass-selected, fragmented by collision-induced dissociation (normalised collision energy 35%) and analysed in the LTQ. Dynamic exclusion was enabled with the exclusion list restricted to 500 entries, exclusion duration of 30 s and mass window of 10 ppm.

Database search for peptide/protein identification and MS data analysis

Peptide identification was by searchers against the Swiss-Prot database (version 2013-2014) restricted to human entries using the Mascot search engine (v 2.5.0, Matrix Science). The parameters included trypsin as digestion enzyme with up to two missed cleavages permitted, carbamidomethyl (C) as a fixed modification and Pyro-glu (N-term), Oxidation (M) and Phospho (STY) as variable modifications. Datasets were searched with a mass tolerance of ± 5 ppm and a fragment mass tolerance of ± 0.8 Da.

The automated programme Pescal (Cutillas & Vanhaesebroeck, 2007 ) was used to calculate the peak areas of the peptides identified by the mascot search engine. Proteins were identified with at least two peptides matched to the protein and a mascot score cut-off of 50 was used to filter false-positive detection. The resulting quantitative data were parsed into Excel files for normalisation and statistical analysis. Significance was assessed by t-test of log2 transformed data. When required, P-values were adjusted using the Benjamini–Hochberg method. Results are shown as log2 fold IKKε over control.

Western blot

Protein levels were assessed using Western blotting. Cells were lysed in a lysis buffer (20 mM Tris–HCl, pH 7.4, 135 mM NaCl, 1.5 mM MgCl2, 1% Triton, 10% glycerol) containing cOmplete protease inhibitor cocktail (Roche) and, where necessary, HALT phosphatase inhibitor cocktail (78428, Thermo Fisher Scientific). Samples were quantified using DC protein assay kit (Bio-Rad), and equal concentration samples were then prepared for SDS-PAGE in loading buffer (40% Glycerol, 30% β-Mercaptoethanol, 6% SDS, bromophenol blue). SDS-PAGE was performed using either 10 or 4–12% NuPAGE™ Bis-Tris Protein gels (Invitrogen) and resolved protein was transferred to Immobilon-P PVDF 0.45 μm Membrane (Merck). For immunoblotting, membranes were blocked for 1 h at room temperature in 5% w/v skimmed milk powder (Sigma-Aldrich) diluted in TBS-T solution (1× tris-buffered saline (TBS) (Severn Biotech) containing 0.1% v/v TWEEN ® 20 (P1379, Sigma-Aldrich)) and then incubated overnight with primary antibodies diluted in 5% w/v milk in TBS-T at 4°C with constant agitation. Membranes were washed a minimum of three times over 15 min in TBS-T at room temperature before incubation with secondary antibodies diluted in 5% w/v milk in TBS-T at room temperature with constant agitation. Membranes were washed again prior to development with Pierce™ Enhanced Chemiluminescence Western Blotting Substrate (32106, Thermo Fisher Scientific), SuperSignal™ West Pico PLUS Chemiluminescence Substrate (34577, Thermo Fisher Scientific) or SuperSignal™ West Femto Maximum Sensitivity Substrate (34094, Thermo Fisher Scientific). Chemiluminescent signal was detected using either Fuji Medical X-Ray Film (Fujifilm) or an Amersham Imager 600UV chemiDoc system (GE Healthcare).

Primary antibodies used were as follows: Actin (sc-1615, Santa Cruz Biotechnology) ATF4 (ab1371, Abcam) c-Myc (Y69 clone, Abcam) HA-tag (11867423001, Roche) IKKɛ (14907, Sigma-Aldrich) IRF3 (11904, Cell Signaling) p-IRF3 Ser396 (4947, Cell Signaling) OAS1 (sc-98424, Santa Cruz Biotechnology) PHGDH (HPA021241, Sigma) PSAT1 (20180-1-AP, Proteintech Europe) PSPH (14513-1-AP, Proteintech Europe) SHMT2 (12762, Cell Signaling) P65 (8242, Cell Signaling) p-P65 (3039, Cell Signaling) STAT1 (9172, Cell Signaling) p-STAT1 Tyr701 (9167, Cell Signaling) Vinculin (66305-1-Ig, Proteintech Europe). Secondary antibodies used were as follows: Anti-mouse IgG, HRP-linked Antibody (7076, Cell Signaling) chicken anti-rat IgG-HRP (sc-2956, Santa Cruz Biotechnology) donkey anti-goat IgG-HRP (sc-2020, Santa Cruz Biotechnology) goat anti-mouse IgG-HRP (sc-2005, Santa Cruz Biotechnology) goat anti-rabbit IgG-HRP (sc-2004, Santa Cruz Biotechnology) Rabbit IgG-HRP Linked Whole Ab (NA934, GE Healthcare).

All Western blots were performed with a minimum of three independent biological replicates, unless otherwise indicated in specific figure legends.

For densitometry analysis of Western blots in Fig 5C–E and Fig EV2D, relative protein band density was quantified using NIH's ImageJ software (Schneider et al, 2012 ). Vinculin protein band density was initially calculated for each sample. Then, within each cell line, the percentage of total density for control siRNA and IKKε siRNA transfected samples was calculated. This process was repeated to calculate relative densities for each protein of interest. Finally, the protein of interest percentage density was divided by the corresponding vinculin percentage density for each sample to generate normalised relative density values.

High-content imaging and measurement of mitochondrial membrane potential (Δψm)

Cells were seeded in thin, clear bottom black 96-well plates (BD Falcon) at medium density (4,000 cells/well) 24 h before the experiments. Prior to imaging cells were loaded with 1 μg/ml Hoechst 33342 (Sigma-Aldrich) and 30 nM tetramethyl-rhodamine-methylester (TMRM) for 30 min. TMRM was present during imaging in the solution (DMEM w/o phenol red). Images were acquired with the ImageXpress Micro XL (Molecular Devices) high-content wide field digital imaging system using a Lumencor SOLA light engine illumination, ex377/50 nm em447/60 nm (Hoechst) or ex562/40 nm and ex624/40 nm (TMRM) filters, and a 60X, S PlanFluor ELWD 0.70 NA air objective, using laser-based autofocusing. Sixteen fields/well were acquired. Images were analysed with the granularity analysis module in the MetaXpress 6.2 software (Molecular Devices) to find mitochondrial (TMRM) and nuclear (Hoechst) objects with local thresholding. Average TMRM intensities per cell were measured and averaged for each well. The mean of wells was then used as individual data for statistical analysis to compare each condition.

PDH activity measurement

PDH activity was measured on whole cell lysates using the pyruvate dehydrogenase (PDH) Enzyme Activity Microplate Assay Kit (ab109902, Abcam).

QRT–PCR

mRNA levels were assessed using quantitative real-time PCR (qRT–PCR). Total RNA was extracted from cells using the RNeasy Mini Kit (Qiagen) as per the manufacturer's protocol. RNA yield was quantified using the NanoDrop ND-1000 (Thermo Fisher Scientific), and 1 mg of RNA was reverse transcribed to cDNA using the Omniscript RT Kit (Qiagen). qPCR was performed using the TaqMan™ assay system.

The following TaqMan™ gene expression probes were used: PHGDH (Hs00198333_m1, Thermo Fisher Scientific) PSAT1 (Hs00795278_mH, Thermo Fisher Scientific) PSPH (Hs00190154_m1, Thermo Fisher Scientific) ACTB (β-Actin, 4310881E, Applied Biosystems).

Assay mixtures were prepared consisting of 10 μl TaqMan™ Master Mix (Applied Biosystems), 1 μl TaqMan™ gene probe & 1 μl cDNA, topped up to 20 μl with 8 μl RNase free H2O. The qPCR reaction was carried out using either the 7500 Real Time or the QuantStudio 5 Real-Time PCR systems (Applied Biosystems), and the process was 2 min at 50°C, followed by 10 min holding at 95°C, then 40 cycles of 15 seconds at 95°C and 1 min at 60°C. Relative mRNA quantifications were obtained using the comparative Ct method, and data were analysed using either the 7500 software v2.3 or QuantStudio Design & Analysis Software (Applied Biosystems).

Generation of conditioned medium

Flp-In 293 HA-GFP or HA-IKKε cells were treated for 16 h with 50 ng/ml doxycycline in 1 ml of medium per well of a 6-well plate, allowing secretion of potential signalling factors into the medium. Following induction, medium was collected and filtered using a 0.22 μM pore size filter and stored at 4°C till use.

Gene expression analysis of clinical samples

The METABRIC dataset (Curtis et al, 2012 ) was obtained from Synapse: https://www.synapse.org/#!Synapse:syn1688369 (METABRIC Data for Use in Independent Research). All analysis was carried out using Bioconductor R packages. Overexpression of all genes was determined by fitting a Gaussian distribution to the central subpopulation shifted to zero and then determining samples which had expressions greater than 1.96 times the standard deviation from zero.

Data availability

Datasets generated as part of this study through labelled metabolite analysis and phosphoproteomic analysis are both provided in full as part of this manuscript as Datasets EV1 and EV2, respectively.

Statistical analysis

Data are presented as mean ± either standard deviation (SD) or standard error of the mean (SEM) as indicated in the figure legends. Statistical analysis tests were performed using GraphPad Prism (version 8), and specific tests were performed as indicated in the figure legends. Statistical significance was assumed at P < 0.05 and is noted on figures using *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 where appropriate.


General education writing emphasis

This department incorporates a significant amount of writing through the required courses instead of identifying particular courses as writing emphasis courses. Students who complete a major in this department will fulfill the general education writing emphasis requirement.

The following is the department's faculty and staff as of the publication date of this catalog. This list will not be updated again until the next catalog is published in July.


Watch the video: Biomolecules Updated (August 2022).