Comprehensive Cell Specific Transcriptome Profiling of a Pediatric Hodgkin Lymphoma Cohort

Introduction: Pathogenic Hodgkin Reed-Sternberg (HRS) cells constitute approximately 1% of Hodgkin lymphoma (HL) tumor cells. Studies characterizing genomic lesions and gene expression of HRS gene cells have been limited due to technical challenges of studying these rare cells, and the majority of existing data has focused on adult HL. We therefore developed a multi-parameter flow sorting strategy to isolate viable cells from pediatric HL tumors and to define the transcriptomes of HRS cells and infiltrating lymphocytes in order to inform underlying mechanisms of HL pathogenesis and also create an opportunity to identify cell-specific biomarkers to predict disease risk and response to therapy.

Methods: Flow cytometry was used to sort HRS cells, CD4⁺ T cells, CD8⁺ T cells, and CD20⁺/30⁺B cells from pediatric subjects' HL lesions and control tonsils. Purity was confirmed by quantitative reverse transcriptase polymerase chain reaction (RT-PCR) and immunohistochemistry (IHC). Affymetrix GeneChip HTA 2.0 was used to assess the gene expression profiles (GEPs) for 16 HRS primary tumor cell samples, 14 HL CD4⁺ and CD8⁺ T cell samples, 6 control tonsillar CD20⁺, CD30⁺, CD4⁺, and CD8⁺ cell samples, and 6 HL cell lines. Unsupervised hierarchical clustering and principal component analysis (PCA) were used to determine relatedness, and Cibersort was performed to confirm the phenotype of the sorted cell types. GEPs of HRS, HL CD4⁺, and HL CD8⁺ cells were compared to respective controls using a univariate t-test. Significance was determined using a multivariate permutation test with the confidence level of FDR assessment at 80 percent and the maximum allowed proportion of false-positive proteins at 0.1. Gene set enrichment analysis (GSEA) and ingenuity pathway analysis (IPA) were performed to analyze DEGs.

Results: Effectiveness of the sorting strategy of HRS cells was confirmed by quantitative RT-PCR and IHC that demonstrated significant enrichment of CD30expression and CD30+ cells in the sorted HRS cell fraction. GEP comparisons were performed for 13 HL samples with matched HRS/CD4⁺/CD8⁺ cells: HRS vs. control tonsil CD20⁺/CD30⁺ (1934 and 3846 DEGs, respectively), HL CD4⁺ vs. control CD4⁺ (635 DEGs), HL CD8⁺ vs. control CD8⁺ (2 DEGs). We carried out a transcriptomic analysis of HRS cells, and a set of multifunctional genes were more than 2-fold downregulated (P < .001), involved in telomere maintenance and packaging (TERF2, RFC3, DNA2 and a group of HIST1) when compared to healthy lymph node CD30⁺ cells. A set of genes related to cytokine/chemokine dysregulation was also upregulated in HRS cells, including IL6, CCL18, and CXCL9. IPA and GSEA of specific HRS genes were also performed and demonstrated pathways associated with HL pathogenesis, including NFĸB activation and T cell exhaustion. Over-expression of genes associated with T cell pathways was demonstrated in HRS cells. While this may be a result of T cell rosetting and contamination, it may also reflect innate T cell signature within HRS cells, as HRS cells clustered separately from T cells in both unsupervised hierarchical clustering and PCA. Cibersort analysis of HRS cells revealed a heterogeneous phenotype that may reflect aberrant differentiation. In comparing clinical characteristics within HRS cells, TCEAL1 was elevated in slow vs. rapid early responders and 3 DEGs were identified when comparing EBV+/- samples. Within HL CD8 cells, KLF2 was elevated in EBV- samples.

Conclusions: This study was the first to successfully isolate highly purified HRS cell populations from whole HL lesions in a pediatric HL cohort. Transcriptomic analysis of pediatric HRS cells identified mechanisms previously associated with HL pathogenesis, and also identified potential novel mechanisms, including telomere maintenance. Additional analyses demonstrated significant heterogeneity of HRS trasncriptomes across specimens that may reflect distinct differentiation pathways and differences in HRS-immune cell interactions. Finally, this study identified increased expression of some genes associated with EBV status and response to therapy. Future studies in an expanded cohort will validate these findings, compare pediatric and adult GEPs, and test these cell-specific biomarkers into the current risk stratification strategies of prospective clinical trials.