Introduction: In a context of allogeneic hematopoietic stem cell transplantation (Allo-HSCT) for acute myeloid leukemia (AML), the impact of peripheral cytokine levels has been poorly studied so far.

Methods: The aim of this prospective study (NCT02693899) was to analyze peripheral levels of 8 cytokines in 35 AML patients (pts) before conditioning (BC) then at Day (D) 0, 30, 60 and 100 post-Allo-HSCT by comparison to median levels from 20 healthy controls (16 males, median age 48.5 yo). ELISA (Bio-Techne, Minneapolis, MN) was used to assay the levels of FLt3-ligand (FL), TNF-alpha, stem-cell factor (SCF), c-kit, IL-1 beta, IL-6, IL-10 and granulocyte-monocyte colony-stimulating factor (GM-CSF). A total of 1313 and 157 samples were analyzed in pts and controls, respectively. The impact of BC, D0 and D30 levels and secretion (secretors = pts with detectable cytokine levels) on overall (OS), disease free (DFS) and GVHD and relapse free survival (GRFS) was also assessed.

Results: Pts characteristics and cytokine levels are given in Tables 1 and 2.

At D0, IL-6 (p=0.0006), FL (p<0.001) and SCF (p=0.003) levels were significantly higher in pts than controls. This remained true for SCF levels on D30 (p=0.0009), D60 (p=0.001) and D100 (p=0.01). Median GM-CSF and c-kit levels were significantly lower in pts at any time (p<0.001). Two-by-two paired cytokine comparisons showed no correlation, especially considering c-kit and its ligand SCF.

ELN2017 classification and disease-risk index (DRI) were significantly associated with OS and LFS by univariate analysis (UA), only the former remaining associated with both OS (HR 8.51; 95%CI 1.18-61.54, p=0.034) and LFS (HR 8.10; 95%CI 1.20-54.82, p=0.032) in multivariate analysis (MA). Only SCF, IL-6 and c-kit levels impacted survivals.

Higher pre-transplant SCF levels were associated with lower OS (p=0.02), DFS (p=0.01) and GRFS (P<0.001) by UA and remained associated with lower GRFS (HR 3.19; 95%CI 1.48-6.88, p=0.003) in MA, with a trend for lower LFS (HR 1.96; 95%CI 0.94-4.08, p=0.07). There was also an association between the absence of SCF secretion and better OS, LFS and GRFS (p=0.02, p=0.02, p=0.03) while c-kit secretion was associated with better OS (p=0.03) in UA. This remained true in MA for OS (HR 0.30; 95%CI 0.10-0.92, p=0.03) and there was a trend between no SCF secretion and GRFS (HR 0.45; 95%CI 0.18-1.14, p=0.09) by MA.

Higher IL-6 levels at D0 were associated with lower OS (p=0.01) by UA and lower LFS both by UA (p=0.01) and MA (HR: 2.83; 95%CI: 1.04-7.70, p=0.04), with a trend for lower OS (HR: 2.55; 95%CI: 0.90-7.29, p=0.07) in MA. ROC curve analysis identified 7pg/mL as the best cut-off for D0 IL-6 levels which resulted, for pts under this level, in significantly better 4-year OS (80% (62-1) vs 35.7% (18-69), p=0.02) and LFS (73.3% (54-99) vs 28.6% (13-63) p=0.01; Figure).

For alive patients at D30, higher c-kit and lower SCF levels were associated with better OS (p=0.003 and p=0.01) and LFS (p=0.009 and p=0.04) by UA while c-kit levels remained associated with LFS (HR: 0.37; 95%CI: 0.17-0.80, p=0.01) by MA. The absence of SCF secretion was associated with better OS (p=0.03) and a trend for better LFS (p=0.07). c-kit secretion was associated with better OS (p=0.008), LFS (p=0.01) and a trend for better GRFS (p=0.07). Only c-kit secretion remained associated with better OS (HR: 0.15; 95%CI: 0.04-0.57, p=0.005), LFS (HR: 0.19; 95%CI: 0.06-0.62, p=0.006) and GRFS (HR: 0.40.95%CI: 0.16-1.00, p=0.05) by MA.

Conclusion: In this cohort, early higher IL-6 and SCF and lower c-kit levels appeared to be significantly associated with lower survivals in AML pts after Allo-HSCT. These high-risk cytokine profiles might allow to propose therapeutic intervention to decrease relapses in selected pts, either by modulating cytokine levels or by starting early maintenance treatment post-Allo-HSCT. These results have to be confirmed on larger studies.

Chevallier:Jazz Pharmaceuticals: Honoraria; Takeda: Honoraria; Incyte: Research Funding; Pfizer: Research Funding; Abbvie: Honoraria.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution