Tumor-Infiltrating T Cells and Their Importance in Cancer

Immunotherapies have changed the management of advanced cancers and the course of oncology. Nevertheless, not all patients derive long-term clinical benefit from these agents. T cells are the primary mediators of antitumor immunity and play a central role in responses to immunotherapies. However, despite major therapeutic advances, across different tumor types, tumor-infiltrating / invading T cells are still not fully understood in terms of their characteristics, function and how they influence clinical outcomes.

In recent years, single-cell RNA sequencing analyses (scRNA-seq) have provided an unprecedented view of the cellular composition and characteristics of various tumors and their infiltrating immune cells.

Studies of chronic viral infection and melanoma in mouse models have revealed highly dysfunctional, "terminally exhausted" CD8+ T cells, as well as precursor and intermediate T-cell populations capable of proliferation and differentiation.

Studies of human tumor types have also revealed, for example, the role of precursor CD8+ T cells in responses to immunotherapy in patients with melanoma and T-cell development in patients with renal cell carcinoma.

However, T-cell phenotypes and recurring populations across various tumor types have remained largely undefined. An up-to-date, comprehensive characterization of the heterogeneous immune structures of tumors by Liangtao Zheng and colleagues from Peking University, published in NEJM was therefore both a welcome and a substantial advance.

A Deep Look into Tumor Immunity

Zheng and colleagues integrated data from 316 patients covering 397 thousand tumor-infiltrating or circulating T cells, combining scRNA-seq data (and T-cell receptor data where available) to create, across 21 cancer types, a pan-cancer atlas.

The researchers found highly diverse CD8+ and CD4+ T-cell phenotypes in tumors. Among CD8+ T cells, 4 cell populations with an exhausted phenotype were abundant in tumors.

Using computational inference of cell differentiation trajectories, from naive T-cell activation to end-stage or terminal CD8+ T-cell exhaustion in most cancer types, the researchers identified two potential pathways as follows;

• The first pathway passed through a GZMK (gran-zyme K) messenger RNA (mRNA)-positive effector memory T-cell population, whereas
• the second pathway passed through a ZNF683 (zinc finger 683) mRNA-positive tissue-resident memory T (Tm)-cell population.

These broad pathways were found in most cancer types, but individual tumors favored both routes. As the study team expected, additional cancer-type-specific pathways leading to T-cell exhaustion were identified.

Zheng and colleagues also computationally inferred potentially tumor-reactive CD4+ T cells that were abundant in tumor tissue – INFG (interferon gamma) mRNA-positive follicular helper T/type 1 helper T (Th1) cells and TNFRSF9 (tumor necrosis factor receptor superfamily member 9) mRNA-positive regulatory cells. These CD4+ T-cell populations co-occurred with CD8+ T-cell populations in the tumor, and this T-cell module appeared to respond to type 1 interferons and transforming growth factor β. This finding suggests that, in shaping the phenotype of infiltrating T cells, the tumor microenvironment has a role. Building on this finding, the researchers identified clinical features (such as body mass index) and somatic tumor features (tumor mutational burden and FAT1 mutation) that had affected the composition of infiltrating tumor cells.

Finally, by analyzing T-cell composition across all tumors, they identified 2 main tumor groups;

• one had high levels of memory T cells, and
• the other was enriched in CD8+ T cells.

In external datasets, tumors enriched in CD8+ T cells were associated with poor patient survival and a reduced response to PD-1 blockade (immunotherapy) in patients with melanoma.

Figure 1: Mapping the Cancer Immune Microenvironment

A Column: Scientific studies in immuno-oncology use biological samples obtained from patients to improve mechanistic understanding of human tumor immunobiology, with the aim of developing new therapeutics. Recently, Zheng and colleagues published a pan-cancer map of tumor-infiltrating T cells.

B Column: Using single-cell RNA sequencing analyses, they identified multiple conserved pathways leading to terminally exhausted CD8+ T cells, a link between somatic alterations and other microenvironmental features and specific T-cell populations, and an association between high infiltration by terminally exhausted CD8+ T cells and poorer patient survival across multiple cancer types. Descriptions of the tumor immune microenvironment, including "archetypes" of specific immunity, will be needed to provide immunotherapies tailored to the genetic and molecular state of tumors. These include identifying tumor antigens, defining the specificity of cognate T cells and mapping the spatial locations of individual cellular populations and T-cell types.

C Column: The resulting data should be used to develop diagnostic methods, predictive indicators and interventions for cancer patients, such as rational combinations of immunotherapies and new antigen-specific therapeutic approaches.

ADCs, antibody-drug conjugates; CAR, chimeric antigen receptor; mABs, monoclonal antibodies; mRNA, messenger RNA; TCR, T-cell receptor; Tem, effector memory T cell; Tfh, follicular helper T cell; Th1, type 1 helper T cell; Treg, regulatory T cell; and Trm, tissue-resident memory T cell.

This hard-won computational achievement represents a major first step toward defining tumors not only by their histological features but also by their immune-cell composition.

Ultimately, describing conserved immunological features (known as immune-cell archetypes) may enable new approaches leading to precision immunotherapies. These could guide the selection of individual or combination treatments targeting specific immunopathological mechanisms in concert with antigen-directed therapies – particularly target tumor epitopes (Figure 1).

However, bridging the gap from this computational characterization of tumors to identifying actionable features of the tumor microenvironment will require functional validation of the proposed interactions and differentiation pathways, and integration of cellular phenotypes with additional layers of information.

At a biochemical level, RNA expression is one of several factors that functionally influence protein. In addition, to create an antitumor effect, T cells must recognize specific tumor antigens and overcome immunosuppressive cell-cell interactions. Developing a framework of functional information on the specific targets, cellular interactions and spatial locations of antitumor T cells will therefore be critical to understanding how T-cell subgroups influence the course of disease in patients and to developing effective cancer immunotherapies.

The pan-cancer framework of tumor-infiltrating T cells described by Zheng and colleagues provides an important foundational resource for future characterizations of tumors according to their immunological features. Such analyses will certainly contribute to future research into responses to current immunotherapies and emerging resistance mechanisms, and to the development of new therapeutic approaches.