CP-CML Seems Related to Specific HLA Subtypes and the HLA DRB1 07 Allele Might Identify a Subgroup of Newly Diagnosed CP-CML Patients on Imatinib First-Line of Better Prognosis

Introduction

Since recently, HLA typing of CP-CML patients (pts) ≤65 years old was still recommended in at diagnosis and treated by Imatinib (IM) first-line, especially in case of baseline warnings (Baccarani et al. 2006/2009/2013), in the aim of performing allo-transplant if TKI failure occurs. Additionally, HLA molecules are processing BCR-ABL peptides at the surface of leukemic cells and trigger immune response, exploitable in autologous (Bocchia Lancet 2005) and allogeneic (Fefer NEJM 1979; Kolb Blood 1995) settings, including in residual disease conditions. The performance of antigen processing might differ according to the type of HLA molecules involved, but this remains largely debatable.

Aims

1) To analyse the frequency of HLA subtypes in CP-CML pts as compared to those of the general population.

2) To investigate if IM first-line response differ according to the HLA subtypes, suggesting an involvement of the autologous immune system in the response observed.

Methods

We retrospectively analyzed a cohort of newly diagnosed CP-CML pts treated with IM first-line 400 mg alone and HLA typed (generic and/or allelic) in our center between years 2000 and2018. All pts were followed according to the ELN recommendations 2006/2010/2013. Clinical data were extracted from medical files, and responses were analysed with standard methods. Molecular results were ELN standardised since 2003, and were expressed as BCR-ABL^IS in %. Overall survival (OS), progression-free survival (PFS), failure-free survival (FFS: progression to advanced phases death, loss of CHR, CCyR, or MMR, discontinuation of IM for toxicity, primary cytogenetic resistance) were analysed since IM initiation in intention-to-treat.

Results

Eighty-three pts have been included, with a median age of 47.5 (11-64.8) years at diagnosis, with 58% males and 42% females. Sokal score (n=80) was low in 44%, intermediate in 42% and high in 14%. ACA were present at diagnosis in 7% of pts (NA in 3), 2 pts had a cryptic Ph, and 2 a variant Ph. Major BCR transcripts were found in 96.4% pts, and atypical transcripts in 3.6%. We compared the frequencies of HLA A, B, C and DR in our 83 CP-CML to those of 5,225 unrelated bone marrow donors (UBMD) typed in the registry of the Rhône-Alpes region (35 million inhabitants). We could demonstrate that the HLA B13, B28, B51, B53 were significantly over-represented in the CP-CML population vs UBMD (4.24 vs 2.02%; 3.63 vs 2.48%, 11.5 vs 8.53%, 1.82 vs 0.88, overall p value = 0.007, Monte-Carlo test) as well as C05 (13.75 vs 8.39%, p=0.044) respectively. In addition, HLA B27, B45 and B55 were underrepresented in CP-CML pts vs UBMD (0.60 vs 3.43%, 0 vs 0.6%, 0.6 vs 1.66%, p=0.007, Monte-Carlo test). There was no difference for HLA A, and DR antigens.

All pts got IM (Glivec®) 400 mg daily since diagnosis. The median follow-up after IM initiation was 111.5 (7.6-203) months. CHR was reached after a median of 1 (0-3.5) month, <10% BCR-ABL transcript (IS) level at 3 months was found in 72.3% of pts (11 unknown). Thirty (36%) pts switched to TKI2 for IM failure. Overall, in ITT, 51.8% (n=43) pts gained MMR by 12 months (optimal response according to ELN recommendations), 38.55% (n=32) did not (8 patients ND or NA) and when we performed a linear regression analysis adjusted on the optimal response at 12 months, the HLA DRB1 07⁺ (at least one allele) pts (n=23 patients) were over represented, with an Odds Ratio of 3.63 [95%CI: 1.33-9.91], p=0.012. There was no difference between the DRB1 07⁺ and 07^- groups for age, sex, Sokal, ACA+, follow-up. We observed 13/60 (21.6%) failures in the DRB1 07^- vs 3/23 (13%) in the DRB1 07⁺ group, but the FFS was not different (p=ns). We observed 5/60 (8.3%) blast crises in the DRB1 07^- group vs 0/23 (0%) in the DRB1 07⁺ group although non- significant (see PFS survival in Figure 1A. p=0.11). There was 5 CML-related deaths and 1 unrelated in the DRB1 07^- group, vs 0 death in the DRB1 07⁺ group. Overall the OS was close to significance in the favour of DRB1 07⁺ pts (see Figure 1B) with a p=0.112. The multivariate analysis will be presented.

Conclusions

CP-CML seems overrepresented in some particular HLA subtypes, and the sensitivity of the disease to IM seems improved in the HLA DRB1 07 allele although not significant. Whether this is related to the biological function of HLA molecules at the surface of CML cells or to a specific genetic background of these pts needs to be determined. These findings need to be confirmed in larger ongoing series.

View largeDownload slide

View largeDownload slide

Close modal