Epstein-Barr Virus Status Correlates with Composition and Prognostic Impact of the Tumor Microenvironment in Classical Hodgkin Lymphoma

Classical Hodgkin lymphoma (cHL) is often characterized by a minority of neoplastic cells surrounded by a heterogeneous background of reactive non-neoplastic cells. An increased amount of certain cell subsets, such as T-regulatory lymphocytes and macrophages, in the microenvironment of cHL tumor lesions has been found to correlate with an adverse prognosis, probably as a result of enhanced immune tolerance towards tumor cells. Furthermore, it has also been suggested that the presence of Epstein-Barr virus (EBV) infection in the Hodgkin Reed-Sternberg (HRS) cells, may modulate the composition of the tumor microenvironment.

In the present study, we have analyzed the possible correlation between EBV-status, a number of tumor microenvironment parameters and outcome in a large retrospective series of newly diagnosed cHL patients.

A tissue microarray was constructed from paraffin embedded pre-therapeutic tumor tissue biopsies obtained from 288 newly diagnosed cHL cases. The expression in the tumor microenvironment of the macrophage markers CD68 and CD163, the regulatory T-cell marker FoxP3 and the cytotoxic T-cell marker Granzyme-B (GrB) was assessed by immunohistochemistry (IHC) using a previously described semi-automated stereological counting approach (Haematologica 2011;96:269–276). The presence of EBV in HRS cells was investigated using 'in situ' hybridization for EBV-encoded RNAs 1 and 2 and LMP-1 IHC. Clinical data were obtained from clinical records. The correlation between clinic-pathological features and EBV was assessed using the rank-sum or Kruskal-Wallis test. The impact of clinico-pathological parameters on event-free (EFS) and overall survival (OS) was evaluated using the log rank test.

The 288 patients had a median age of 37 yrs (range: 6–86 yrs). The M:F ratio was 1.3. One third (33%) of the patients were positive for EBV in the HRS cells. EBV-positive cases exhibited higher numbers of CD68 (p=0.001), CD163 (p=0.0002), GrB (p<0.0001), and FoxP3 (0.0009)-positive cells. Excluding cases of mixed cellularity from the analysis, the significant correlation between EBV and CD163 (p=0.03), GrB (p=0.003), FoxP3 (p=0.006) remained, whereas the correlation for CD68 was slightly weakened (p=0.06).

In the entire cohort (n=288), a high expression of CD68, CD 163 and GrB were found to correlate with significantly lower OS and EFS (high vs. low CD68: 5-year OS, 73% vs. 87% p=0.002, 5-year EFS, 58% vs. 70% p=0.03; high vs. low CD163: 5-year OS, 78% vs. 87%, p=0.03, 5-year EFS, 62% vs. 69%, p= 0.04) and high GrB: 5-year OS, 77% vs 88 %, p=0.004, 5-year EFS, 61% vs. 69%, p= 0.02). Interestingly, the influence of tumor microenviromental parameters on outcome was more pronounced in EBV-negative cases (n=193) than in EBV-positive ones (n=95). In the former, significantly lower OS and EFS values were associated with a high expression of CD68 (high vs. low CD68: 5-year OS, 60% vs. 92%, 5-year EFS, 43% vs. 71%, both p<0.001), high CD163 (5-year OS, 72% vs. 89%, p<0.001, 5-year EFS, 58% vs. 69%, p= 0.03) and high GrB (5-year OS, 61% vs. 90%, 5-year EFS, 43%vs. 71%, both p<0.001). Among the EBV-positive cohort, the corresponding OS and EFS values were high CD68 (5-year OS, 85% vs. 82%, p=0.69, 5-year EFS, 72%vs. 66%, p=0.43), high CD163 (5-year OS, 86% vs. 84%, p=0.34, 5-year EFS, 67%vs. 71%, p=0.48) and high GrB (5-year OS, 88% vs. 81%, p=0.34, 5-year EFS, 73%vs. 65%, p= 0.63).The number of FoxP3-cells was not found to affect the prognosis in neither EBV-negative nor EBV-positive cases.

The present study confirms that the EBV-status in cHL is associated with distinct features of the tumor microenvironment. As a novel finding, our results suggest that the prognostic impact of intratumoral reactive non-neoplastic cell subsets is EBV-status dependent, i.e. a significantly adverse impact of an increased amount of certain bystander cell subsets on outcome was only found in EBV-negative cases.

No relevant conflicts of interest to declare.