Expression and Function of the P-Glycoprotein (P-gp) In Myelodysplastic Syndromes (MDS) Treated with Azacytidine (AZA)

In MDS and AML, P-gp expression is associated with low response rate to conventional anthracycline-AraC chemotherapy (Lepelley et al., Leukaemia, 1994; Mahadevan and List, Blood, 2004). Since hypomethylating agents, notably AZA, can improve survival and have become a reference first line treatment in high and int 2 risk MDS (Lancet Oncol, 2009), we assessed whether P-gp expression and function in ex vivo MDS cells from patients treated with AZA was associated to clinical response to that drug.

Bone marrow (BM) cells from 30 patients with MDS or AML post MDS treated with AZA were studied, at onset of AZA. All patients received AZA for at least one cycle (75 mg/m2/d during 7 days every 28 days), and response was evaluated according to IWG 2006 criteria.

Mononuclear cells from samples were isolated using a Ficoll-Paque PLUS density gradient. The evaluation of P-gp expression and functionality was made on CD45dim cells (blast population) and CD45dim CD34+ population, gated by flow cytometry.

Quantification of P-gp expression was evaluated by FACS, using a phycoeythrin(PE)-conjugated antibody specific for ABCB1(P-gp)(clone UIC2).

P-gp functionality was quantified by flow cytometry evaluation of Rhodamine123 (Rhoda: a specific substrate of P-gp) efflux in the presence or the absence of ciclosporine A (CSA, an inhibitor of P-gp). Patient cells were considered Rhoda pos when the RFI (Relative Fluorescence Intensity) CSA+/CSA- was >1.5.

WHO diagnosis at onset of AZA was MDS in 19 pts (1 RA, 1 RCMD, 1 LMMC2, 5 RAEB-1, 10 RAEB-2, 1 RAEB-t) and AML post MDS in 11 pts. Median age was 76, M/F: 19/11. Karyotype (IPSS) was fav (n=10), int (n=3) unfav (n=14) and a failure (n=3). In MDS, IPSS was int-1 in 3pts, int-2 in 9 pts and high in 5 pts (and not evaluable in 2). The median number of cycles of AZA administered was 6 [2–19]. Fourteen pts (47%) responded including 6 (20%) CR and 8 hematological improvements (HI). Median OS from initiation of AZA was 418 d. No correlation between response and karyotype was found.

P-gp expression was found in 64% of pts, and 65% of the pts had Rhoda pos cells, with some discordant cases for P-gp expression and functionality (including 23% P-gp pos Rhoda neg and 10% P-gp neg Rhoda pos). Karyotype was not correlated to P-gp expression or Rhoda RFI. The response rate was 39% in P-gp pos and 71% in P-gp neg pts (p=0.2) and the CR rate 17% in P-gp pos pts vs 28% in P-gp neg pts (P=0.6) The response rate was 47% in Rhoda pos and 33% in Rhoda neg pts (p=0.6) and the CR rate 23% in Rhoda pos vs 0% in Rhoda neg pts (P=0.26).Median OS was 373 d. in P-gp pos and not reached in P-gp neg pts (p= 0.18), and 428 d in Rhoda pos pts vs 292 d. in Rhoda neg pts (P=0.07).

Contrary to what is observed with anthracycline based chemotherapy, P-gp expression and functionality did not negatively affect clinical response to AZA in this cohort of MDS and AML post MDS, with even a trend for longer OS in patient with Rhoda pos cells. This difference may further explain the possible efficacy of AZA in chemoresistant cases of MDS and AML post MDS

Fenaux:Celgene: Honoraria, Research Funding.