While cytotoxic CD8+ T cells have been emphasized as the primary cell type responsible for the beneficial effects of immunotherapy, there are other components of the tumor microenvironment (TME) that could enhance long term responses in cancer patients. Several recently published papers in Nature (January 2020) have demonstrated that B cells within tertiary lymphoid structures (TLS) are important for antitumor immunity and response to immunotherapy1–3. B cells are instrumental in antigen-specific adaptive immunity and TLS are like secondary lymphoid structures that respond to inflammation or infection within lymph nodes or tonsils. These exciting findings in skin cancer and adult sarcoma raised additional questions, such as what are the transcriptional signatures of B cells in TLS, how does the composition of TLS vary across tumors, and what markers are reflective of an effective TLS? Many of these questions are multifaceted, and to add to the complexity, there are very few animal or in vitro model systems that are reflective of the in vivo biology. Leveraging our mutual strengths, I teamed up with Dr. Anthony Cillo (postdoc in the Vignali Lab) to deeply interrogate B cells and TLS in HNSCC patients with the hopes of addressing some of these important questions about the role of B cells in antitumor immunity.
Why did we choose to interrogate B cells in HNSCC?
My mentor, Dr. Tullia Bruno, has had a longstanding interest in understanding the role of B cells and TLS in the TME. Her previous work in non-small cell lung cancer (NSCLC) revealed that distinct B cell phenotypes are present in the TME and showed that B cells in NSCLC can present tumor-associated antigens to CD4+ tumor infiltrating T cells4. As a part of her research program, she was interested in validating these findings in other tumor indices. HNSCC was a logical next choice because: (1) HNSCC is unique as it is caused by two different etiologies: either through virally driven carcinogenesis via human papillomavirus (HPV) infection, or through mutational carcinogenesis from tobacco and alcohol use. We hypothesized that HPV-driven cancer would lead to enhanced antitumor immunity due to the presence of immunogenic viral antigens that could be recognized and presented by intratumoral B cells, (2) previous RNA-seq and immunohistochemistry (IHC) data demonstrated that CD20+ B cells were significantly increased in HPV+ HNSCC patients compared to HPV- patients and correlated with improved survival while infiltration of other immune cell subsets was comparable between HPV+ and HPV- patients5,6. We were fortunate to have excellent collaborators in Dr. Robert Ferris and Dr. Dario Vignali for our study, who both enthusiastically supported our investigation of B cells in HNSCC and were interested in gaining a deeper understanding of the immune landscape between the two etiologies of HNSCC to identify new targets for immunotherapy. Dr. Cillo, using his expertise in systems immunology, captained an initial study that compared transcriptional profiles of immune cells by single-cell RNA seq in HPV+ and HPV- HNSCC. This study revealed that B cells, CD14+ myeloid cells, and conventional CD4+ T cells were transcriptionally different between the two etiologies, while CD8+ T cells and T regulatory cells (Tregs) were similar7. For our study, we leveraged Dr. Cillo’s initial dataset to perform a deeper scRNA seq analysis of the differences in B cell signatures between HPV+ and HPV- HNSCC. We identified key genes in the scRNA seq analysis to help design high-dimensional flow cytometry panels that could interrogate cell surface and intracellular proteins in B cells and used immunohistochemistry (IHC) and multispectral immunofluorescence to uncover spatial localization and morphology of TLS within the HNSCC TME.
“It was the best of times...”: Lessons learned from comparing B cells and TLS in HNSCC
Our multifaceted analysis in HNSCC yielded several unique findings (Figure 1). First, we found that B cell signatures were favorably associated with survival; more B cells in the tumor meant longer time to tumor progression. We also discovered by both flow cytometry and single-cell RNA seq, that some patients with virally driven HNSCC had higher numbers of germinal center (GC) B cells within their tumors. GC B cells are normally restricted to specific regions of lymph nodes, so we were surprised to find them within the TME. Our flow cytometry panels that were informed by unique bioinformatic analysis of the scRNA seq data allowed us to identify additional B cell subsets in the TME of HNSCC patients such as memory B cells and plasma cells and uncover novel GC B cell subsets in patients. We also queried what markers (if any) were associated with an early transition from activated B cells to GC B cells, and to our surprise we observed that semaphorin 4a (SEMA4a) was expressed on early GC B cells, and within TLS. Finally, we determined that the composition of TLS was also important: if TLS had clear B and T cell zones with a bona fide GC, patients had improved survival (regardless of HPV status). These findings are exciting to the field of cancer immunotherapy and were uniquely possible in the context of our study because of the interdisciplinary and multifaceted approached that we employed.
“The age of wisdom”: The next generation of immunotherapy…Targeting B cells and TLS
As we know, great scientific findings always lead to more questions. The most common question we get asked is: Are there HPV-specific B cells present in the HPV+ HNSCC TME? A recent study in Nature showed that HPV-specific B cells and plasma cells are present in HPV+ tumors and produce HPV-specific IgG antibodies8. It’s still unclear as to whether these HPV-specific antibodies have any effector function and can activate NK or macrophage mediated killing of HPV+ tumors. Another common question is: Do B cells in HNSCC patients also recognize tumor associated antigens? Addressing these questions will help determine if patient B cells could be leveraged in novel immunotherapies like therapeutic monoclonal antibodies, adoptive B cell therapy, and/or cancer vaccines.
Our study demonstrated that regardless of HPV status, having TLS with a bona fide GC is associated with favorable outcomes. In our minds the next logical question to address is how to therapeutically induce TLS with GCs in patients who don’t have them. Additionally, are naturally forming TLS with GC within tumors indicative of tumor-specific B and T cell responses? We can begin to address these big picture questions by understanding what key factors (cytokine/chemokines and immune cell interactions) involved in TLS and GC formation are present in patients and how they may be impacted by different TMEs as well as developing physiologically relevant in vivo murine models that can reflect what we observe in patients. While there is still much to learn about B cells and TLS in the TME, we believe that future studies will shed more light on the biology underlying B cells and TLS within tumors and how both can be leveraged to enhance antitumor immunity.