HLA-G and MHC Class II Protein Expression in Diffuse Large B-Cell Lymphoma

Major histocompatibility complex (MHC) class I and class II are similar in function since they present peptides at the cell surface to CD8+ and CD4+ T cells, respectively. There is mounting evidence indicating that MHC genes play a role in the etiology and clinical course of non-Hodgkin lymphomas (NHL), especially in diffuse large B-cell lymphoma (DLBCL). We investigated the expression of human leukocyte antigen-G (HLA-G) and HLA class II protein in DLBCL among 148 patients and related their expression to the clinical course of the disease. Negative HLA-G expression was associated with a lower probability of achieving a complete remission (P= 0.04). There was no impact of HLA-G or HLA class II expression on the probability of progression-free survival (PFS). Patients with positive HLA-G expression presented a trend towards a higher 3-year overall survival (OS) rate compared to those with negative expression of HLA-G (P= 0.08). The estimated 3-year OS rate of patients with positive HLA class II expression was 59.4% (95%CI 49-69) in comparison to the 37.4% (95%CI 22-56; P= 0.04) in subjects with negative expression. In a multivariate Cox analysis adjusted for the IPI factors, we found that both the intermediate high/high IPI risk group (P= 0.001) and the loss of HLA class II expression (P= 0.05) independently increased the risk of death in the study group. We also investigated whether the impact of HLA class II expression on OS may be related to the subtypes of DLBCL. In the subgroup of 58 patients (39%) with GCB-type pattern, the patients with the loss of HLA class II expression presented a significantly lower 3-year OS rate than those with its positive expression (26% [95% CI 10-53] vs 68.2% [95% CI 51-81], P= 0.02). In contrast, in the subgroup of 90 non-GCB patients (61%), HLA class II expression did not influence OS. To further explore the unexpected favorable effect of positive HLA-G expression on the clinical course of DLBCL, we performed an additional analysis that considered the type of treatment (chemotherapy with or without rituximab). In the group of patients treated with immunochemotherapy, a more pronounced effect of the positive HLA-G expression on OS was revealed. The estimated 3-year OS rate of patients with the positive HLA-G expression was 73.3 % (95% CI 49-88) compared to 47.5% (95% CI 35-60, P= 0.03) in subjects with the negative HLA-G expression. In contrast, in the group treated with CHOP-like regimens, no significant impact of HLA-G expression on OS was observed: the 3-year OS rate for HLA-G positivity was 20% (95% CI 4-62) vs 55.3% (95% CI 37-72; P= 0.08) for the absence of HLA-G. Additionally, the prognostic value of HLA class II expression was also shown to depend on the use of rituximab as a part of first line treatment. In the patients receiving immunochemotherapy, those that had positive HLA class II expression demonstrated a 3-year OS rate of 65.3% (95% CI 52-76) compared to 29.6% (95% CI 13-53, P= 0.04) in subjects with the loss of HLA class II expression. However, HLA class II expression did not have a prognostic impact on OS in the patients treated with chemotherapy alone: the 3-year OS rate was 49.5% (95% CI 32- 67) in the subjects with positive expression in comparison to 50% (95% CI 15-85, P= 0.8) in those with the loss of HLA class II expression.

In conclusion, we demonstrated for the first time expression of HLA-G protein in DLBCL and its association with the clinical course of the disease as well as we confirming the association of the loss of HLA class II protein expression with poor survival in patients treated with immunochemotherapy. Although the clinical significance of the loss of HLA class II protein expression seems to be well understood, the contribution of HLA-G to the prognosis of B-cell malignancies deserves further study, especially its immunoregulatory functions in relation to treatment with rituximab, which remains an open question.