Proteome-wide Profiling of Clinical PARP Inhibitors Reveals Compound-Specific Secondary Targets

Abstract

Poly(ADP-ribose) polymerase (PARP) inhibitors (PARPi) are a promising class of targeted cancer drugs, but their individual target profiles beyond the PARP family, which could result in differential clinical use or toxicity, are unknown. Using an unbiased, mass spectrometry-based chemical proteomics approach, we generated a comparative proteome-wide target map of the four clinical PARPi, olaparib, veliparib, niraparib, and rucaparib. PARPi as a class displayed high target selectivity. However, in addition to the canonical targets PARP1, PARP2, and several of their binding partners, we also identified hexose-6-phosphate dehydrogenase (H6PD) and deoxycytidine kinase (DCK) as previously unrecognized targets of rucaparib and niraparib, respectively. Subsequent functional validation suggested that inhibition of DCK by niraparib could have detrimental effects when combined with nucleoside analog pro-drugs. H6PD silencing can cause apoptosis and further sensitize cells to PARPi, suggesting that H6PD may be, in addition to its established role in metabolic disorders, a new anticancer target.

Keywords: H6PD; PARP inhibitor; chemical proteomics; polypharmacology; target identification; target selectivity.

Figure 1. Synthesis and validation of linker-modified PARPi

(A) Structures of clinical PARP inhibitors and their modified coupleable derivatives (denoted by “c-” prefix) used for covalent attachment to NHS-sepharose beads. (B) In vitro inhibition of PARP1 activity by unmodified and coupleable versions of each PARPi, n = 3, s.e.m. (C) Inhibition of CAL-51 viability by PARPi, n = 5, s.d. (D) Immunoblots of eluates from PARPi-modified beads incubated with CAL-51 lysate ± 20 μM of the corresponding free PARPi (e.g. 20 μM niraparib added to c-niraparib matrix and lysate). Multiple bands arise from different isoforms of each protein. Blots are representative of three independent experiments. TCL: total CAL-51 cell lysate.

Figure 2. PARPi affinity matrices identify PARP1/2 protein complexes

(A) PARPs and other proteins identified as interacting with PARPi. Color indicates NSAF values for individual drug-protein interactions averaged across three replicates. (B) Drug-protein interaction network for olaparib, veliparib, rucaparib and niraparib in CAL-51 cells. Shown are all NAD(P)⁺ binding proteins and their interactors that pass all 6 filtering criteria as specified in Table S3. PARP family proteins are depicted in red, NAD(P)⁺-binding proteins in light red. Direct and indirect interaction partners of known or putative drug targets are shown in dark and light grey, respectively. Known and putative drug-protein interactions are represented by black edges, PPIs are shown in grey. PPIs between observed proteins were retrieved from public databases using the Cytoscape app BisoGenet. For drug-protein interactions, edge width indicates the respective NSAF value. Dashed edges indicate observed drug-protein interactions that do not pass all filtering criteria, but do so for interactions with other PARPi. IMPDH2 is depicted as interacting with all PARPi based on the proteomics data; its newly identified interaction with PARP1 is indicated by a solid grey edge. (C) Anti-IMPDH2 immunoblot of eluates from c-PARPi-modified beads incubated with CAL-51 lysate ± 20 μM corresponding free PARPi (e.g. 20 μM niraparib added to c-niraparib matrix and lysate). TCL: total CAL-51 cell lysate. (D) Immunoblots of PARP1, PARP2 and rabbit IgG immunoprecipitates from CAL-51 cell lysate. Blots are representative of three independent experiments.

Figure 3. SAINTexpress analyses of all four PARPi

Fold change and SAINT score were calculated with SAINTexpress and all proteins with SAINT scores above 0.9 (A-D) or 0.5 (E-H) were plotted. Bubble size is proportional to the NSAF value and bubble color corresponds to CRAPome probability (scale 0-1, cutoff 0.5; low value: high probability for specific interaction; high value: high probability for non-specific interaction). No proteins yielded SAINT scores above 0.5 for the comparison of either olaparib or veliparib to the other three PARPi (F and H).

Figure 4. DCK is a new target of niraparib

(A) Schematic of DCK enzymatic activity and biosynthesis of nucleotide and nucleotide analog triphosphates. (B) Immunoblot of eluates from PARPi-modified beads incubated with CAL-51 lysate ± 20 μM of the corresponding free PARPi; blot is representative of two independent experiments. (C) Immunoblots of eluates from c-niraparib beads incubated with CAL-51 lysate ± 10 μM DCK inhibitor DI-39 or 1 mM DCK substrate deoxycytidine (dC). (D) Protection from cytarabine-induced toxicity. A549 cells were incubated for 3 days with 2 μM cytarabine ± the indicated concentrations of DCK inhibitor DI-39, niraparib, rucaparib, veliparib, or olaparib. Cell viability was measured by CellTiterGlo. Rescue was calculated by normalizing to the toxicity of cytarabine alone (0% rescue) and to cells incubated with PARPi or DI-39 only (100% rescue). n = 3, s.e.m.

Figure 5. H6PD is a specific target of rucaparib

(A) Schematic of H6PD enzymatic activity for conversion of the natural substrate glucose-6-phosphate and the alternative H6PD-specific substrate galactose-6-phosphate. (B) Immunoblot of eluates from PARPi-modified beads incubated with CAL-51 lysate ± 20 μM of the corresponding free PARPi; blot is representative of two independent experiments. (C) H6PD activity in H6PD overexpressing (HEK293 H6PD-OE) cell lysate in the presence of increasing concentrations of PARP inhibitors. n = 5, s.d. Representative immunoblot for H6PD in both parental and overexpressing HEK293 cells is shown (bottom panel). (D) Overlay of rucaparib and niraparib structures. Shared features are shown in black, features unique to niraparib are shown in grey, and features unique to rucaparib are highlighted in purple. (E) IC₅₀ values for PARPi inhibition of cell viability were determined in parental CAL-51 cells and H6PD overexpressing (CAL-51 H6PD-OE) cells as measured by CellTiterGlo after 5 d, n = 5, s.e.m. (F) IC₅₀ values for PARPi inhibition of cell viability were determined in CAL-51 shNT and CAL-51 shH6PD cell lines as measured by CellTiterGlo after 5 d, n = 5, s.e.m. Representative immunoblot for H6PD in CAL-51 mock, shNT, shH6PD and H6PD-OE cells is shown.

Figure 6. Loss of H6PD function causes apoptosis in CAL-51 cells

(A) Immunoblot analysis of transient shRNA-mediated H6PD knockdown in CAL-51 cells after 48 and 72 h. (B) Immunoblot analysis of siRNA-mediated H6PD knockdown in CAL-51 cells after 48 and 72 h. Blots are representative of three independent experiments (A, B). (C) Flow cytometry analysis of APC Annexin V/DAPI staining of CAL-51 cells 48 and 72 h after transfection with siNT or siH6PD. Right panel: quantification of cells in early apoptosis (Q3) and late apoptosis/necrosis (Q2). Data shown in density blots and their respective bar graphs correspond to one representative of three independent replicates. (D) Number of viable CAL-51 cells after transfection with siNT or siH6PD at 48 and 72 h, n=4, s.e.m. (E) Immunoblot analysis of CAL-51 cells after 48 h treatment with 40 μM PARPi. Blots are representative of four independent experiments.

Figure 7. Loss of H6PD function causes apoptosis in some but not all cancer cells

(A) Immunoblot of relative H6PD levels in 11 cancer cell lines and HEK293 cells (included for comparison). (B) Immunoblot analysis of transient siRNA-mediated H6PD knockdown in H322 cells after 48 and 72 h. (C) Immunoblot analysis of transient siRNA-mediated H6PD knockdown in MDA-MB-468 cells after 48 and 72 h. Mock: transfection without siRNA; siNT: non-targeting siRNA; blots are representative of 2 independent experiments. (D, E) Inhibition of H322 (D) and MDA-MB-468 (E) cell viability by PARPi as determined by CellTiterGlo after 5d, n = 3, s.e.m.