To the Editor:
A News story in the May issue by Malini Guha discusses recent clinical setbacks of poly(ADP-ribose) polymerase 1 (PARP1) inhibitors for the treatment of certain cancers1. Agreeing with the article's conclusion that it is too early to discount the effectiveness of PARP1 inhibitors in cancer treatments, I outline below three principal problems regarding our understanding of PARP1 function and the role of PARP1 inhibitors in limiting tumor growth that could hamper their progress in the clinic.
The first problem proceeds from the underlying assumption that the main role of PARP proteins in cancer cells is controlling the DNA repair pathway. The direct evidence supporting this view is the ability of PARP1 to bind damaged DNA and to become activated upon binding2. PARP1 also interacts with a subset of DNA repair enzymes2 and Parp1 null mutant mice show a substantial level of genomic instability3. On the basis of these data, a widely accepted model1 postulates that cells accumulate DNA breaks upon inhibition of PARP1. Such a model, however, does not accord with the observation that no increase in DNA breaks is detected in cancer cells treated with a PARP1 inhibitor4. Therefore, the purported role of PARP1 in DNA repair seems tangential to tumor proliferation. After DNA-dependent activation, PARP1 immediately becomes automodified and loses contact with DNA5. A likely purpose of this chain of molecular events is to remove PARP1 from chromatin and terminate PARP1-dependent processes, thereby facilitating access of DNA repair enzymes to DNA.
The specific distribution of PARP1 protein in chromatin presents additional evidence against the notion that the main PARP1 function is to repair DNA. PARP1 is predominantly localized within the promoter regions of protein-encoding genes6 and accumulates in nucleoli7 as well as in telomeres8,9. These are very peculiar sites of localization for a protein that purportedly controls a general DNA repair pathway. Furthermore, the PARP1 protein has been shown to regulate transcription as well as chromatin decondensation coupled with transcription10,11,12 and in regulating nuclear factor kB (NF-kB)-dependent cellular responses11,13. As NF-kB signaling is important for tumor growth14, PARP1 inhibitors might be effective in preventing this stage of progression, particularly when combined with drugs that act at other steps in the overall process. PARP1 controls expression of a chaperone protein heat shock protein 70 (HSP70)11,15, which also contributes to tumor cell survival and resistance to therapy16. These functions correspond well with sites of PARP1 accumulation and may be crucial for survival and proliferation of tumor cells. Reexamining PARP1 functions in tumor cells and targeting pathways other than DNA repair should improve the theoretical framework surrounding PARP1 inhibitors in cancer research.
The second obstacle to effective PARP1-targeted cancer treatments is the assumption that the sensitivity of cancer cells to inhibitors is specifically linked to the presence of BRCA mutations17. It is now clear that not all tumor cells carrying BRCA1/2 mutations are sensitive to PARP1 inhibitors, whereas some cancer cells that have mutations other than BRCA1/2 are sensitive to them. It appears that the best biomarker of PARP1 inhibitor sensitivity is not a particular mutation but a pre-existing high level of poly(ADP-ribose) in cancer cells4. Previously, it was shown that the main target of poly(ADP-ribose) in a living cell is PARP1 protein itself18,19. This auto-poly(ADP-ribosyl) ation leads to PARP1 auto-inactivation and removal of PARP1 from chromatin5. Therefore, the accumulation of poly(ADP-ribose) reflects not only a high level of PARP1 protein activity, but also a high level of PARP1 protein inactivation and the deficit of PARP1 protein in a cell18. The inactivation of PARP1 by auto-poly(ADP-ribosyl)ation may explain the sensitivity of cancer cells with a high level of poly(ADP-ribose) to PARP1 inhibitors. Such cells simply have a very small pool of functional PARP1, thereby allowing an inhibitor to quickly titrate away the remainder of PARP1 and block its activity (Fig. 1), abolishing PARP1-dependent transcription and other pathways. The mechanism of this high level of poly(ADP-ribose) accumulation is still to be discovered, and may open new gateways to PARP1-based cancer therapy.
The third pitfall to PARP1-based cancer treatments concerns the strategies deployed for designing PARP1 inhibitors. The majority of PARP1 inhibitors have been designed to compete with NAD for a binding site on the PARP1 molecule. This strategy resulted mainly in discovery of nucleotide-like PARP1 inhibitors that may target not only PARPs, but also other enzymatic pathways involving NAD and nucleotides as co-factors. Using such inhibitors affects multiple NAD/nucleotide-dependent enzymatic pathways, which results in secondary toxic effects proceeding from the inactivation of other pathways, whereas the efficiency against the PARP1 pathway specifically is diminished. A possible strategy to bypass this pitfall is to design inhibitors by targeting other binding sites on the PARP1 protein. Our research20,21 has shown that interaction with core histones plays a crucial role in PARP1 protein regulation in vivo. Moreover, histones control PARP1 activation in both transcriptional and DNA-repair pathways21. Therefore, identifying small molecule inhibitors that attack PARP1 interaction with histones22 should yield compounds inhibiting PARP1 pathways with high specificity.
The recent disappointments by the first anti-PARP1–based treatments of cancer require a re-evaluation of the approach. I believe the field should step back and reassess which mechanisms are really affected in PARP inhibitor–sensitive cells and then develop novel PARP inhibitors based on our knowledge of the mechanism of PARP1 protein enzymatic activation in chromatin in vivo and specifically in tumorigenic cells.
Guha, M. Nat. Biotechnol. 29, 373–374 (2011).
D'Amours, D., Desnoyers, S., D'Silva, I. & Poirier, G.G. Biochem. J. 342, 249–268 (1999).
Menisser-de Murcia, J., Mark, M., Wendling, O., Wynshaw-Boris, A. & de Murcia, G. Mol. Cell. Biol. 21, 1828–1832 (2001).
Gottipati, P. et al. Cancer Res. 70, 5389–5398 (2010).
Mendoza-Alvarez, H. & Alvarez-Gonzalez, R. J. Biol. Chem. 268, 22575–22580 (1993).
Krishnakumar, R. et al. Science 319, 819–821 (2008).
Tulin, A., Stewart, D. & Spradling, A.C. Genes Dev. 16, 2108–2119 (2002).
Beneke, S. et al. Nucleic Acids Res. 36, 6309–6317 (2008).
Salvati, E. et al. Oncogene 29, 6280–6293 (2010).
Aubin, R.J. et al. EMBO J. 2, 1685–1693 (1983).
Tulin, A. & Spradling, A. Science 299, 560–562 (2003).
Kim, M.Y., Mauro, S., Gevry, N., Lis, J.T. & Kraus, W.L. Cell 119, 803–814 (2004).
Oliver, F.J. et al. EMBO J. 18, 4446–4454 (1999).
Dajee, M. et al. Nature 421, 639–643 (2003).
Petesch, S.J. & Lis, J.T. Cell 134, 74–84 (2008).
Leu, J.I., Pimkina, J., Frank, A., Murphy, M.E. & George, D.L. Mol. Cell 36, 15–27 (2009).
Bryant, H.E. et al. Nature 434, 913–917 (2005).
Tulin, A., Naumova, N.M., Menon, A.K. & Spradling, A.C. Genetics 172, 363–371 (2006).
Kotova, E., Jarnik, M. & Tulin, A.V. PLoS Genet. 5, e1000387 (2009).
Pinnola, A.D., Naumova, N., Shah, M. & Tulin, A.V. J. Biol. Chem. 282, 32511–32519 (2007).
Kotova, E. et al. Proc. Natl. Acad. Sci. USA 108, 6205–6210 (2011).
Kotova, E., Pinnola, A.D. & Tulin, A.V. Methods Mol. Biol. 780, 491–516 (2011).
The author declares no competing financial interests.
About this article
Biochemical Pharmacology (2019)
Cell Chemical Biology (2016)
Scientific Reports (2015)
GATA3 cooperates with PARP1 to regulate CCND1 transcription through modulating histone H1 incorporation