The State of Chromatin Condensation, the Expression of Genes Involved in DNA Damage Response and the DNA Repair Capacity Affect the Drug Sensitivity of PBMCs of Myeloma Patients Treated with Melphalan

Melphalan is an interstrand cross-link (ICL)-inducing agent and one of the most active chemotherapeutic drugs in the treatment of multiple myeloma (MM). However, the molecular mechanisms contributing to differential response of MM patients to melphalan are poorly understood. Herein, we investigated the underlying mechanisms in processing and repair of melphalan-induced DNA lesions and their potent contribution to the outcome of anti-myeloma therapy.

We studied peripheral blood mononuclear cells (PBMCs) from 15 newly diagnosed multiple myeloma (MM) patients (8M/7F; median age 61 years) responders (≥PR, n=9) and non-responders (<PR, n=6) to subsequent melphalan therapy. PBMCs from 25 healthy controls were also included in this study (HC; 14M/11F, median age 58 years). PBMCs were ex vivo treated with melphalan and the extent of the DNA damage formation/repair (monoadducts by using Southern blot; ICLs by using a quantitative PCR-based assay; double strand breaks (DSBs) using immunofluorescence quantification of γH2AX foci), the induction of the apoptotic pathway by using a photometric enzyme-immunoassay and the local chromatin condensation by using micrococcal nuclease digestion were examined. Finally, the expression of a focused panel of 84 genes involved in DNA damage response (DDR) pathways (ATM/ATR signaling, DNA repair pathways, cell cycle regulation, apoptosis) was also evaluated.

Following ex vivo treatment of PBMCs with melphalan, in all individuals examined biphasic DNA repair kinetics were observed, including a fast first-phase of repair and a much slower second phase. Interestingly, the accumulation of monoadducts was inversely correlated with the first-phase repair capacity of the PBMCs, being significantly higher in HC than in responders and lowest in non-responders (all P<0.001). The second phase repair capacity showed no differences in all individuals analyzed. Also, although ICLs “unhooking” rates were similar in all individuals, accumulation of ICLs was significantly higher in HC compared to responders’ PBMCs (P<0.01), due to higher levels of monoadducts  that are precursors of ICLs left unrepaired in these cells, resulting in higher formation of ICLs. Minimal amounts of ICLs were observed in non-responders. Moreover, DSBs burden was significantly higher in HC than in responders, due to higher accumulation of ICL, which are precursors of DSBs and lower rates of DSB repair in these cells (P<0.01). Again, minimal amounts of DSBs were observed in non-responders.

Interestingly, apoptosis rate of PBMCs was inversely correlated with the repair efficiency of both the first-phase monoadducts and the DSBs, with the apoptosis  being significantly higher in HC compared to responders and lowest in non-responders (all P<0.05). In untreated PBMCs, we found an inverse correlation between the local chromatin condensation and the repair capacity. Therefore,  we observed a progressive, significant increase in the looseness of the local chromatin structure, from HC to MM patients, with responders showing more condensed chromatin structure compared to non-responders (all P<0.05). Interestingly, by using a-amanitin, an inducer of chromatin condensation, PBMCs from non-responders showed the DNA repair capacity and the melphalan sensitivity similar to PBMCs from responders, suggesting that the state of the chromatin structure contributes to the response to melphalan therapy.

Finally, microarray analyses of untreated PBMCs consistently point to an altered expression of several DNA damage response-related genes in MM patients compared to HC. Particularly, responders’ PBMCs showed upregulation of 4 genes (ATR, CHEK2, XPA, XRCC1) and downregulation of 5 genes (ATM, MPG, UNG, CDKN1A, CDC25C) compared to non-responders (in all cases, fold changes between groups were >2, P<0.001), suggesting that changes in the molecular components of the DDR pathways correlate with the outcome of melphalan therapy.

We conclude that the state of chromatin condensation, the expression of genes involved in DDR pathways and the repair capacity of monoadducts and DSBs affect the drug sensitivity of PBMCs and maybe used for the prediction of response of myeloma patients to melphalan therapy.