Addicted to PAR?

Keywords: :

• Poly(ADP-ribose)
• PARP inhibitor
• DNA damage response
• synthetic lethality
• predictive biomarkers
• cancer

This article refers to:

Evaluation of candidate biomarkers to predict cancer cell sensitivity or resistance to PARP-1 inhibitor treatment

Defects in genome maintenance and repair pathways are common features of human cancers. In some cases, specific DNA repair defects can render cancer cells dependent on back-up pathways for their survival, and targeting these back-up mechanisms is a promising strategy for cancer therapy.¹ One example for such synthetic lethality is the pronounced sensitivity of BRCA-deficient cancer cells, defective in DNA repair by homologous recombination (HR), to inhibition of poly(ADP-ribose) polymerases (PARPs).² Consequently, PARP inhibitors have entered clinical trials as single agents, but also in combination therapy as chemo- or radiation-sensitizing drugs.³ Following exciting proof-of-concept results, however, PARP inhibitors recently encountered first difficulties, and the current challenge is to understand why certain cancers respond better to these compounds than others.

In a study now published in Cell Cycle, Oplustilova and colleagues used a panel of human cancer cell lines derived from carcinomas of breast, prostate, colon, pancreas and ovary to study the response to PARP inhibition and analyze cellular determinants of sensitivity or resistance.⁴

Acquired resistance by drug efflux can be a major barrier to therapeutic efficacy, and Oplustilova et al. demonstrate that P-glycoprotein drug efflux pumps contribute to inhibitor resistance in colon cancer cells, thereby extending previous findings obtained in a mouse model for BRCA1-associated breast cancer.⁵ Importantly, inhibition of P-glycoprotein by verapamil resulted in elevated intracellular levels of the PARP inhibitor KU 58948 and restored inhibitor sensitivity, consolidating the notion that acquired resistance by enhanced drug efflux can, in principle, be overcome.

The authors also investigated a second mechanism of alleviated PARP inhibitor sensitivity based on the recent discovery that loss of the genome caretaker 53BP1 restores HR in BRCA1-deleted cells.^6,7 Consistent with the proposed function of 53BP1 to restrict DNA end resection, exacerbate HR deficiency and enhance PARP inhibitor sensitivity, Oplustilova et al. show that shRNA-mediated depletion of 53BP1 in a human BRCA1-defective breast cancer cell line reduced sensitivity to PARP inhibition. Given the aberrant reduction of 53BP1 in subsets of BRCA-associated breast carcinomas and sporadic triple-negative breast cancers,⁷ immunohistochemical analyses of 53BP1 expression, as also performed in the present study, might thus have predictive value when assessing PARP inhibitor sensitivity.

Intrigued by the emerging notion that the concept of synthetic lethality could have wider applicability to other defects in the DNA damage response network, and in line with previous studies demonstrating PARP inhibitor sensitivity associated with BRCA-independent HR defects,⁸ Oplustilova et al. show that even partial depletion of the MRN components MRE11 or NBS1 sensitizes to PARP inhibition, whereas ectopic expression in mutant cells had the opposite effect.

Further, arguing that ongoing PARP activity must be a prerequisite for PARP inhibitors to work, Oplustilova et al. went on to assess steady-state poly(ADP-ribose) (PAR) levels in different cell lines and report a correlation between detection of PAR and inhibitor sensitivity. While these results generally support PAR levels as potential candidate biomarker, additional studies are needed to determine whether PAR detection is technically feasible in tissue biopsies, and whether PARP inhibitor sensitivity is inevitably associated with detectable amounts of PAR, especially in light of the limited sensitivity of the available antibodies for shorter PAR chains. Likewise, when Oplustilova et al. assessed RAD51 foci numbers as a surrogate marker for HR, several cell lines were sensitive to PARP inhibition, yet they had normal RAD51 foci levels, suggesting that either specific HR defects do not entail reduced RAD51 loading, or that synthetic lethality can be achieved through HR-independent mechanisms. Together, these results raise the concern that use of single biomarkers could indeed be misleading, and that a combination of markers to assess which cancer cells are likely “addicted to PAR” might be more reliable. The study by Oplustilova et al. evaluates several of such candidate biomarkers and thus contributes to the collective effort to guide targeted cancer therapy to those patients who might benefit most from PARP inhibitor treatment. However, it also illustrates once again that such guidance crucially relies on a more detailed understanding of the complex DNA damage repair network and how the interacting pathways may be rewired in cancers with unstable genomes, as well as on deeper insights into the diverse functions of PARPs and how they contribute to synthetic lethality.