In this issue of Blood, Veldman et al present a novel in vitro coculture model that enables the molecular study of microarchitectural rosette formation of CD4+ T cells surrounding the malignant Hodgkin and Reed-Sternberg (HRS) cells in Hodgkin lymphoma (HL).1
HL displays very distinctive histomorphology features, most notably the extreme rarity of the malignant HRS cells (on average ∼1% of all cells) and the surrounding immune-cell rich tumor microenvironment. Current pathogenesis models assume “education” of the tumor microenvironment and particularly emphasize the bidirectional cross talk of HRS with highly abundant CD4+ T cells for the survival advantage of malignant cells.2 Among many histological and microanatomical patterns recognizable as features of specific HL subtypes, T-cell rosettes are of particular interest due to the presumed direct contact between HRS and T cells, and these rosettes are well referenced in the literature for decades.3 However, clear molecular correlates and the biological importance of these structures remain largely elusive. To study rosette formation in more detail, Veldman and colleagues established an elegant coculture model using HL-derived cell lines and HLA-matched peripheral blood mononuclear cells for quantitative assessment of rosette formation in the context of experimental disruption (via knockout) of key molecules involved in immunological synapse formation. The studied interactions between CD4+ T cells and the B-cell–derived HRS cells in this model include T-cell receptor (TCR)-major histocompatibility complex (MHC)-II, CD2-CD58, and CD11a-CD54. A key output of this study is the dissection and reported relevance of each of these interaction pairs for cellular adhesion/rosette formation and T-cell activation (see figure). Most strikingly, CD58 was identified as indispensable for both rosette formation and T-cell activation, whereas HLA-II expression was found to be only involved in T-cell activation. Based on their findings, the authors suggest a 2-step model, with step 1 being engagement of the TCR with antigen-HLA complexes, and step 2 being signal amplification and contact stabilization through the CD2-CD58 axis. In contrast, the results did not suggest a prominent role for CD54 in either rosette formation or T-cell activation in the model systems studied.
The presented findings emphasize the importance of CD58 in the pathogenesis of HL and solidify MHC class II expression as a major parameter that drives a biological dichotomy into HLA-II–positive vs HLA-II–negative disease with implications for treatment considerations. Beyond MHC-II being a prognostic biomarker for standard-of-care first-line therapy,4 a recent study strongly suggested that MHC-II expression also has predictive value for checkpoint inhibitor therapy in the relapsed/refractory setting.5 Of note, both loss of CD58 function and MHC-II expression are underpinned by somatic gene mutations in CD586 and CIITA (the master transcriptional activator of MHC-II),7 providing evidence for somatically acquired immune privilege in response to selective pressure by immune surveillance. Hence, and with further support from the current study, characterization of MHC-II expression is now considered standard workup of HL biology, in addition to and synergy with latent Epstein-Barr virus infection status.
In the past, it has proven difficult to robustly study cellular interactions in primary samples with so far only a limited number of studies that have attempted to correlate spatial findings, such as rosette patterns or synapse formation, to clinical parameters and treatment outcome. This difficulty is also evidenced in the present study where a number of clinical samples were involved for validation of synapse formation and CD2-CD58 interaction in primary cases, but the relatively low case number has limited the ability to uncover clinical correlates with statistical power. Toward this goal, the right assay method to effectively read out cellular cross talk will need to be determined, and in particular, multiparametric imaging and sequencing technologies will have to reach technical maturity to be applied to highly annotated biospecimens from patients treated in clinical trials or with standard of care. It will be of specific interest, if spatial patterns, synapse formation and the involvement of CD58 and MHC-II change dynamically over time (eg, before and after treatment and disease recurrence) and will be predictive of treatment outcome using checkpoint inhibitors and/or brentuximab vedotin.
Overall, providing important functional insight, the presented work is highly complementary to recently published studies interrogating malignant cell-tumor microenvironment interactions using single-cell technologies, including multicolor immunohistochemistry,8 flow-based mass cytometry,9 imaging mass cytometry, single-cell RNAseq,10 and, as employed in this study, proximity ligation assays. The new opportunities linked to all these studies include unprecedented insight into the architecture of the tumor microenvironment and into spatial relationships of different cell types at single-cell resolution. These relationships have relevance not only to HL but also to other lymphoma entities that rely on a complex cellular ecosystem of malignant cells with reactive immune cells. However, the molecular underpinnings (the “interior design”) in the form of somatic gene mutations, receptor-ligand interactions, and involved soluble factors (cytokines/ chemokines) are still only partially understood and are rarely combined with the characteristic morphology features (“architecture”) of HL. Thus, the present study can be seen as a paradigm for an accelerating process deciphering the relatedness of “architecture” and “interior design.”
Conflict-of-interest disclosure: The author receives research support from Bristol-Myers Squibb, Trillium, consulted for Seattle Genetics, Bayer, AbbVie, and Curis Inc, and is inventor of a patent owned by BC Cancer for a subtyping assay for aggressive lymphomas (NanoString).