Among the platinum family of anti-cancer compounds, CP has been a very effective agent against a host of tumors such as those of the ovary and testis. However, its utility is severely restricted by dose limiting side effects and cellular resistance, either intrinsic or acquired during CP administration. We identified NSC109268 as a compound increasing cellular sensitivity to CP in Saccharomyces cerevisiae
]. Enhancement of cellular sensitization to CP by NSC109268 was confirmed for CP-sensitive ovarian carcinoma 2008 cells and the derived CP-resistant line 2008/C13 [12
]. Shown in both yeast and 2008/C13 cells, the increase in cellular sensitivity to CP by NSC109268 correlated with inhibition of S phase progression [12
]. In in vivo
studies with xenografted human head and neck carcinomas treated with CP alone, others had shown that a hindrance of S phase traversal correlated with CP responsiveness better than G2/M arrest [51
In this study, we explored the possible targets of NSC109268 responsible for enhancing CP sensitivity using the systematic yeast gene deletion collection. We were guided by the concept that absence of the target should cause CP sensitivity but preclude additional sensitization by NSC109268. A similar strategy resulted in the identification of HR as the target of synergism between CP and camptothecin in yeast and human cancer cells [52
]. Alterations in uptake or efflux of CP appeared unlikely to be the cause of sensitization since synergism was not diminished if CP was administered first, prior to treating with NSC109268 in CP-free buffer. This notion was further confirmed by mutant analysis.
After investigating various DNA repair and tolerance pathways such as NER, DNA translesion synthesis or interstrand crosslink repair, we identified a DNA damage tolerance pathway, the Rad5 pathway, as critical determinant of the synergism between CP and NSC109268 in budding yeast. This pathway is believed to bypass DNA damage at replication forks, that have been stalled in S phase due to DNA adducts, by using the newly synthesized strand of the sister chromatid as a template (Figure S1
]. Fork regression or recombinative sister strand junction formation have been discussed as the underlying mechanisms in this process [29
]. CP primarily causes bulky DNA adducts which lead to a slow down of the replication fork, eventually leaving behind single-stranded gaps to be filled by Rad5 mediated TS pathway, involving certain HR components [28
]. Inhibition of Rad5 mediated lesion bypass activity by NSC109268 may also result in fork collapse leading to genetic instability – as shown here with reporter strain D7 – and finally cell death, thus mediating cellular sensitization to CP. It has been debated if the Rad5 mechanism is critical for S-phase progression in the presence of DNA damage, as suggested here, or acts predominantly in G2-phase to fill remaining single-stranded gaps [30
]. The discrepancies in timing found in the literature may very well depend on level and kind of DNA damage.
Our findings seem to fit particularly well with the model of filling of single-stranded gaps by template switch using a subset of recombination functions (Figure S1
]. We found proteins involved in initial PCNA monoubiquitination, such a Rad6 or Rad18, and in subsequent polyubiquitination, such as Rad5, Ubc13 or Mms2, to be required for CP+NSC109268 synergism. But we also identified HR proteins such as Rad51 or the junction-resolving enzyme Sgs1 as participants of the targeted mechanism(s). This includes Psy3, a member of the Shu complex proposed to participate in TS following strand invasion [44
]. All of these proteins that are required for synergism of CP and NSC109268 are exclusively or non-exclusively involved in the TS pathway.
Interestingly, the mutant group showing no or very reduced synergism did not include Rad59 which has been shown by others to play a role in canonical HR but not in the TS pathway [28
]. We also demonstrated that CP-mediated and HR-dependent gene conversion events [46
] are increased by NSC109268 and not decreased. (Current models of the Rad5 pathway do not involve the homologous chromosome in a template switch mechanism, so this result does not contradict the inhibition of the Rad5 pathway.) Based on these observations, we do not favor that NSC109268 acts as a general HR inhibitor to exert its function on CP sensitivity, however, a more rigorous demonstration is still required.
A screen for single gene deletion mutants of budding yeast conferring exquisite CP sensitivity had previously revealed RAD5
and other key players of the same pathway as top hits [39
]. In humans, this pathway was initially regarded as a tumor suppressor pathway since the essentially error-free bypass should counteract genetic instability [29
]. On the other hand, the human ortholog of the yeast Rad5, HLTF, was reported to be overexpressed in radiation resistant recurrent human cervical carcinoma and a knockdown of HLTF in HeLa cells lead to a decrease in cellular proliferation [58
]. Conceivably, targeting HLTF in chemotherapy-resistant cancers may lead to increase in cellular CP sensitivity.
Upon investigating the known activities of NSC109268 as a 20S proteasome and phosphatase 2C and 2A inhibitor [13
], we found no evidence for the importance of proteasome inhibition in mediating CP sensitivity in yeast. In contrast, our survival studies indicated 2C and 2A phosphatase inhibition by NSC109268 to be responsible for mediating CP sensitization. Analysis of the yeast single mutants and double mutants ptc2Δ, ptc3Δ
(each one defective for functionally redundant PP2C phosphatases), pph3Δ
(defective for PP2A-like phosphatase) compared to the triple phosphatase mutant ptc2Δptc3Δpph3Δ
revealed a marked defect in the viability of ptc2Δptc3Δpph3Δ
cells in the presence of CP, while neither of the single mutants were similarly sensitive to CP. These results suggested redundancy of phosphatase PP2C and phosphatase PP2A-like mediated pathways in CP damage responses, as was already reported for other agents such as hydroxyurea or methyl methanesulfonate [32
]. If phosphatase redundancy is removed in the triple mutant, sensitization of CP-treated cells by NSC109268 is greatly reduced suggesting phosphatases as the relevant target.
Phosphatases Psy2 or Ptc2 and, in a redundant fashion, Ptc3 have been shown to be required for turning off the DNA damage-induced cell cycle checkpoint by dephosphorylating kinase Rad53 [59
]. However, upon treatment of rad53Δ
background), Rad53 was excluded as a major target of NSC109268 in mediating CP sensitization. Thus, a reduced ability to dephosphorylate Rad53 checkpoint kinase and to release checkpoint arrest is not responsible for the cellular sensitization to CP by NSC109268.
Although the actual protein target(s) remain to be determined, this study underlines the critical importance of protein phosphatases in DNA damage responses that goes beyond recovery from checkpoint arrest. Even in a model organism such as yeast, significant gaps in our knowledge will need to be filled. For the first time, this study shows an influence of phosphatases on the Rad5 tolerance pathway. Additionally, the notion of intertwined HR and TS pathways is supported [28
]. Interestingly, the triple phosphatase mutant pph3Δptc2Δptc3Δ
used by us was reported to be defective in early steps of HR [32
] that may also be necessary for sister chromatid junction formation and processing within the Rad5 TS pathway [28
]. Interestingly, Rpa2, a subunit of the single-strand DNA binding complex Replication Protein A, represents another target candidate since its phosphorylation status, which may be enhanced by NSC109268-mediated phosphatase inhibition, is inversely correlated with RAD51 foci formation [61
Taken together, our studies with NSC109268 in budding yeast demonstrate its potential as a cellular chemotherapy sensitizer, most likely by acting through the inhibition of dephosphorylation of critical protein(s) of the Rad5 pathway mediating CP damage tolerance during S phase. NSC109268 and related compounds may thus be especially valuable in CP combination therapy of the subset of cancers that developed CP resistance due to alterations within the Rad5 pathway. It remains to be seen if this is a mechanism of resistance that is widespread among cancer patients.