Introduction. Granzyme B (GZMB) is a serine protease commonly found in the granules of cytotoxic lymphocytes and natural killer cells. It is secreted with the pore forming protein perforin and mediates apoptosis in target cells. Previously we have already reported about impact of GZMB expression in T-regulatory cells on acute graft-versus-host disease (aGVHD) development (Drokov et al. 2017). Recent publication of W. Du et al. shown that GZMB in conventional CD4+CD25- T-cells (Tcon) contributes to the optimal graft-versus-tumor (GVT) effect in mouse models. Here we reported data about GZMB expression in Tcon on day +30 and its influence on relapse rate during first 6 months in allo-HSCT patients.

Patients and methods. The study included 65 patients with hematological malignancies who underwent allo-HSCT. Detailed patients' characteristics are given in Table 1. Peripheral blood samples were collected in EDTA-tubes at day +30 after allo-HSCT. We use PBMC obtained by density gradient media. This method was used due to lymphopenia in this group of patients. The anti-CD4-APC-Cy7, anti-CD25-APC and anti-Granzyme B-PE (Becton Dickinson, USA) antibodies were used to determine Tcon cells as CD4+CD25- population by flow cytometry. At least 30000 of CD4+ cells were analyzed on a BD FACSCanto II to achieve sufficient statistical power (Becton Dickinson, USA). Our data was not normally distributed (Shapiro-Wilk test, p <0.05). For this reason, the Mann-Whitney U test was used for nonparametric data analysis. Fisher's exact test was used for 2×2 tables. A p-value less than 0.05 was considered significant.

Results. As we can see on Chart 1A level of GZMB+ Tcon on day +30 was higher in patients who developed relapse during 6 months after allo-HSCT. GZMB+ Tcon in group of patients who developed relapse was 16.77±4.19% in comparison with 7.95±1.74% in group who did not relapse during this period (p=0.011*). As Du W. et al. we assume that Tcon cells use GZMB for optimal GVT. Due to that fact GZMB+ Tcon fraction is low. On the other hand, when Tcon is not involved in GVT or work not properly, Tcon cells do not use GZMB. So fraction of GZMB+ Tcon is high.Further ROC-curve analysis was used to estimate the cut-off value as 4.4% (Area under curve - 0.71; sensitivity - 80.00%; specificity - 52.94%). According to that all patients were divided in two groups with >4.4%("high") and <4.4% ("low") of GZMB+ Tcon respectively. Study groups were balanced (similarly distributed) on disease status before allo-HSCT, HLA-disparity, aGVHD incidence and conditioning regimen (p>0.05). According to Kaplan-Meier analysis (with Log-rank test) probability of relapse was 36.7% vs 10% in patients with >4.4% ("high") and <4.4%("low") of GZMB+ Tcon respectively (p=0.021*) (see Chart 1B). Hazard ratio (logrank method) for relapse during 6 months after allo-HSCT in group with "high" level of GZMB+ Tcon (>4.4%) was 3.9 (95% CI, 1.4 - 10.8).

Conclusion. Here we report that group with "high" level relapsed 3.9 times higher than group with "low" level of GZMB+ Tcon. Also, our data shows that level of GZMB+ Tcon cells on day +30 after allo-HSCT may help to predict relapse during 6 months from allo-HSCT with satisfactory test results (AUC=0.71; sensitivity - 80%; specificity - 52.94%).


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.