A Peptide Nucleic Acid Targeting Nuclear Rad51 Sensitizes Myeloma Cells to Melphalan Chemotoxicity Both in Vitro and in Vivo

Multiple myeloma (MM) cells are characterized by extensive genomic heterogeneity, which contributes to patient differences in prognosis and response to treatment. We previously reported that MM cells have elevated homologous recombination (HR) rates and expression of RAD51 and its paralogs, promoting genomic instability and disease progression that are reversed by RAD51 siRNA. We now examine the roles of HR and RAD51 in resistance to melphalan, one of the most widely used drugs for MM chemotherapy. The drug induces a variety of DNA lesions, with DNA interstrand crosslinks (ICL) accounting for most of the drug’s cytotoxicity. RAD51 is a central protein in the HR pathway and its overexpression may contribute to chemoresistance by enabling repair of DNA lesions induced by DNA damaging agents such as melphalan. MM cell sensitivity to melphalan correlates directly with melphalan-induced RAD51 foci, and high RAD51 expression predicts poor event-free and overall survival of MM patients. Activity of the Rad51 promoter increases >850-fold in cancer cells compared to normal cells, and tumor cells are selectively killed by a construct in which P_Rad51 drives expression of diphtheria toxin. In this study, we tested whether inhibiting RAD51 expression with a peptide nucleic acid (PNA) would inhibit MM cell growth and/or sensitize MM cells to melphalan. PNAs are DNA or RNA mimics in which a polymer of (2-amino­ethyl) glycine replaces the nucleic acid’s sugar-phosphate backbone. PNAs are highly specific, binding DNA with higher affinity than RNA or DNA, and they are quite stable to degradation both in vitro and in vivo.

We designed a PNA to target the promoter region of the RAD51 gene (PNArad51), encompassing the transcription start site. To enhance cellular uptake and nuclear delivery without transfection, we conjugated the PNA to a nuclear localization signal rich in basic residues (PKKKRKVR). As a control we employed a scrambled PNA (PNAmt) with the same nucleotide composition but not targeting any genomic sequences. We used qRT-PCR to assess the effect of PNA on RAD51 mRNA expression and that of melphalan on mRNA levels of RAD51 and its paralogs (RAD51B, RAD51B, RAD51C, RAD51D, XRCC2 and XRCC3) and BRCA1. Propidium iodide staining and flow cytometry were used to examine the cell-cycle effects of melphalan. γH2AX and RAD51 foci were quantitated using confocal immuno­fluorescence microscopy, and MM cell viability was assessed with the WST-1 assay. To examine the in vivo consequences of PNA ± melphalan for tumor growth, we injected H929 MM cells expressing luciferase into rabbit bone fragments implanted in SCID-rab mice, as previously described by us,. Total RNA extracted from cells recovered from the rabbit bones was analyzed by qRT-PCR to determine the in vivo effect of PNA on expression of RAD51.

Melphalan treatment (10 µM) significantly induced expression of RAD51 and its paralogs, particularly RAD51 and XRCC3 (p≤0.01). Melphalan caused cell-cycle arrest, predominantly in the S-phase (55%, significantly elevated over vehicle alone, 17%; p<0.0001), the period in which HR is most active, and during which ICLs are converted into double strand breaks (DSBs) on encountering DNA replication forks. PNArad51 (10 µM) significantly reduced expression of RAD51 (~60%, p<0.001) relative to PNAmt. Pretreatment with PNArad51 inhibited melphalan-induced RAD51 focus formation, far more than PNAmt pretreatment (21% compared to 66%, p<0.0001) whereas the number of γH2AX foci increased (66%) relative to PNAmt (39%; p<0.0001). Consequently, pretreatment with PNArad51 produced synergistic synthetic lethality with melphalan, reducing the IC₅₀ of melphalan by 4.5-fold. PNArad51 alone caused significant cytotoxicity compared to PNAmt (p<0.05). In the SCID-rab mouse model, a two-week treatment with PNArad51 alone or in combination with melphalan resulted in significant inhibition of tumor volume (p≈0.01 and p<0.05, respectively) compared to PNAmt, although the combination of PNAmt plus melphalan was ineffectual. Prolonged treatment (4 weeks) with PNArad51 ± melphalan (but not PNAmt + melphalan) reduced tumor growth compared to PNAmt treatment, although this was not statistically significant (p>0.05).

These results highlight the importance of RAD51 in the response of MM cells to melphalan, and indicate for the first time the potential for RAD51-targeted PNA in tumor chemosensitization.