Diverse origin and nature of tumor-infiltrating lymphocytes: the infection hypothesis

Abstract

In the 19th century, Rudolf Virchow observed lymphocyte infiltration in tumors and suggested a potential link between cancer and inflammation. Current explanations for the origin of tumor-infiltrating lymphocytes are based on the interaction between cancer cells and the immune system. This model was proposed before it was recognized that tumors are frequently infected or colonized by microbial elements, suggesting the presence of bacteria-specific and virus-specific T cells within tumors. While several aspects of this microbial experience remain under investigation, their impact on the tumor microenvironment and antitumor immune response may be significant, as the pathogens may elicit the direct or indirect recruitment of tumor-infiltrating lymphocytes. This “Infection Hypothesis” regarding the origin of tumor-infiltrating lymphocytes complements the classical tumor immunoediting hypothesis, offers new avenues for a comprehensive understanding of the tumor ecosystem, and highlights the need for more nuanced cancer treatment approaches that account for the potential interconnected roles of lymphocytes and pathogens in the tumor milieu.

For the graphical abstract of this paper see figure 1.

Figure 1. Graphical abstract.

Introduction

In this hypothesis, we propose the infection hypothesis as a complementary framework to the classical theory regarding the origin and function of tumor-infiltrating lymphocytes (TILs). While TILs have traditionally been attributed to antitumor immune responses, recent studies have demonstrated that cancer recognition is often restricted to a small subset of TILs.¹ By incorporating microbes into the current cancer-immune system paradigm, the infection hypothesis suggests that the recently described intratumoral microbes may also contribute to the recruitment, localization, and function of TILs. Considering microbial influences on TIL biology and frequency provides a potentially more comprehensive view of tumor-immune interactions and may have potential therapeutic implications.

Immunogenic tumors and the cancer immunosurveillance and immunoediting theories

In the 19th century, Rudolf Virchow noted leukocyte infiltration in tumors and suggested a potential link between cancer and inflammation. Current models on the origin of TIL emphasize the dynamic interplay between cancer cells and the immune system, grounded in the concepts that circulating T cells can recognize tumor antigens and that tumors undergo immunoediting, progressively evading immunosurveillance. The cancer-immunity cycle provides a conceptual framework for understanding this process, describing how tumor antigen release, presentation, and T-cell activation occur iteratively, allowing the immune system to adapt to tumor evolution.¹

Within individual cancer types, tumors may exhibit distinct immunological phenotypes, commonly referred to as “immunotypes”. The three classical immunotypes—immune-inflamed, immune-excluded, and immune-desert—are defined by the spatial distribution and abundance of immune cell infiltration. Based on this paradigm, several immunotherapeutic approaches aim to enhance TIL presence and function. These strategies include, but are not limited to, adoptive cell therapy, which involves the ex vivo expansion and reinfusion of TILs; immune checkpoint blockade, which reactivates suppressed antitumor lymphocytes; and viroimmunotherapy and cancer vaccines, which seek to convert immune-desert tumors into immune-inflamed phenotypes.

Limitations of tumor-infiltrating lymphocytes in cancer eradication

Despite the presence of TILs, effective tumor eradication is often not achieved, a paradox central to tumor immunology. According to the cancer-immunity cycle framework, several dysfunctional mechanisms contribute to this outcome, such as anergy, insufficient co-stimulation, and exhaustion from chronic antigen exposure.¹ The tumor microenvironment further impairs T-cell function through hypoxia, nutrient deprivation, immunosuppressive mediators, and regulatory T cells. Tumor cells may also evade immune detection by downregulating HLA expression, a phenomenon frequently reported, even prior to therapeutic intervention.¹ Another factor potentially contributing to the limited antitumor efficacy of TILs is the high prevalence—ranging from 20% to 80%—of “bystander T cells”, lymphocytes that recognize non-cancer peptides or are cancer ignorant. These observations may explain why the mere presence of TILs does not always translate into effective antitumor immunity across most solid tumors.

Adoptive expansion of TILs has emerged as a promising therapeutic strategy in selected malignancies, particularly melanoma, where ex vivo expanded TILs exhibit enhanced cytotoxic activity compared with peripheral T cells.² However, the therapeutic benefit of TIL-based therapy in other solid tumors remains limited. One major challenge is the low frequency of tumor-specific lymphocytes within the bulk TIL population. Consequently, effective TIL-based therapies require the infusion of tens to hundreds of billions of ex vivo-expanded TILs to induce meaningful improvements in patient outcomes, underscoring both the promise and logistical complexity of this therapeutic approach.

The emerging interface of intratumoral microbes and cancer immunoresponse

The relationship between inflammation, infection, and cancer has been extensively studied. Bacteria and viruses, such as Helicobacter pylori, human papillomavirus (HPV), hepatitis B virus, hepatitis C virus, Epstein-Barr virus (EBV), and the Merkel cell polyomavirus, act as drivers of cancer initiation in approximately 15% of cancers.³ Advances in deep sequencing and spatial transcriptomic technologies have enabled the characterization of microbes within the tumors. In a landmark study,⁴ comprehensive analyses of 1,526 tumor samples and their paired adjacent normal tissues across seven cancer types revealed distinct intratumoral bacterial signatures. Notably, microorganisms have been detected not only in malignant cells but also in immune cells,^{4 5} suggesting a role in modulating TIL recruitment.

Growing evidence supports the presence of active bacterial communities, an intratumoral microbiota, in tumor types in close proximity to the gut and oral microbiota, such as colorectal and pancreatic cancer, as well as oral squamous cell carcinoma, describing an association between intratumoral bacteria and immunosuppressive niches, and the influence on their response to immunotherapy.

Notably, recent work investigating large cohorts of primary and metastatic brain tumors identified bacterial elements within the tumor microenvironment, with evidence supporting their intracellular localization.⁵ Spatial investigation of tumor niches associated with bacterial elements in brain tumors demonstrated activation of toll-like receptor pathways and enrichment of neutrophils and antigen-presentation proteins.⁵ Importantly, extensive immunopeptidome analyses of melanoma and glioblastoma tumors demonstrated that bacteria-specific peptides are presented by tumor cells and can be recognized by TILs alongside other tumor antigens.

In addition to bacteria, results from the International Cancer Genome Consortium (ICGC) project Pan-Cancer Analysis of Whole Genomes Consortium showed a high prevalence of tumor-associated viruses. The study of whole-genome data from 5,354 tumor-normal samples across 38 tumor types and 1,057 RNA-sequencing data from 25 cancer types identified a total of 23 viral genera associated with cancer. Notably, a higher abundance of T cells and M1-phenotype macrophages was observed in virus-positive versus virus-negative head-and-neck cancer specimens, suggesting an association between viral infections and inflammatory characteristics in cancer.³

Collectively, this body of literature strongly supports a role for intratumoral microbes in shaping tumor immunity. Although emerging evidence implicates microbe-derived epitopes, along with pathogen-associated molecular patterns, as potential mechanisms underlying microbe-driven tumor immunomodulation, further investigation is required to determine the extent to which these processes influence TILs and antitumor immune responses.

On the origin of tumor-infiltrating lymphocytes: the infection hypothesis

Recent studies aimed at understanding the antigen specificity of tumor-infiltrating lymphocytes have revealed abundant bystander CD8+ T cells. Simoni and colleagues⁶ analyzed the antigen specificity of CD8+ TILs in tumors from 144 patients (lung and colorectal cancers). CD8+ TILs were screened for 1,091 putative neoantigens, 123 tumor-associated antigens, and 46 cancer-unrelated epitopes. The resulting data showed the presence of bystander CD8+ TILs specific for tumor-unrelated epitopes, such as EBV, human cytomegalovirus (HCMV), or influenza epitopes. Using coupled single-cell RNA sequencing and T-cell receptor (TCR) sequencing, a study using non-small cell lung cancer specimens, also detected clonotypes against EBV, influenza virus, HCMV, and HPV.⁷ Similarly, Oliveira et al integrated single-cell RNA/TCR sequencing with functional studies to detect non-tumor-reactive cells enriched for viral specificities, such as influenza A, in human melanoma.⁸ Of note, since screening for anti-viral TCRs is limited to publicly available data, the fraction of viral-specific TILs may be underestimated.

Virotherapy represents an unintended experiment in TIL recruitment. Direct infection of a tumor with oncolytic viruses frequently leads to increased T-cell infiltration.^{9 10} From the perspective of immunodominance, most of these T-cell responses are thought to be directed primarily against viral rather than tumor-derived antigens, with only a minor subset recognizing tumor-associated targets. As a result, tumors treated with oncolytic viruses often exhibit a paradoxical pattern in which both lymphocyte infiltration and tumor progression occur simultaneously. This suggests that viral infection transforms tumors into a “pseudo-hot” phenotype: a microenvironment enriched with T cells, yet functionally dominated by antiviral activity. In this regard, while lymphocyte exhaustion remains a widely accepted explanation for the limited efficacy of TILs, markers such as PD-1 and TIM-3 (T cell immunoglobulin and mucin-domain containing-3) are also observed in chronic infections, complicating the accurate interpretation of T-cell function. Extending this framework, the “infection hypothesis” suggests that some tumors may have undergone prior microbial infections, leaving behind persistent populations of pathogen-specific TILs. These cells may not contribute to tumor eradication but instead coexist with tumor-reactive lymphocytes, diluting the overall effectiveness of the antitumor immune response.

Functional implications of the infection hypothesis for cancer therapy

Advances in next-generation sequencing and machine learning enable unprecedented resolution in characterizing tumor-microbe interactions. Rather than passively coexisting, intratumoral microbes can actively reprogram tumor behavior and immune infiltration. Despite growing recognition of the role of gut microbiota in cancer therapy, the direct impact of tumor-resident microbial communities remains underexplored, particularly in the context of advanced immunotherapies, where microbial regulation of TILs may represent a critical yet overlooked determinant of therapeutic efficacy.

This emerging understanding is expected to refine adoptive cell therapy. Specifically, the identification and selective isolation of tumor-specific T cells could enable a more targeted approach while reducing the need for large cell infusions. In the context of virotherapy, the infection hypothesis suggests that lowering the intrinsic immunogenicity of oncolytic viruses may enhance the likelihood of eliciting antitumor immunity across a broader range of treated tumors.

In summary, the infection hypothesis complements existing paradigms regarding the origin, nature, and function of TILs. Recognizing the interplay between tumors, immune cells, and microbes may refine our understanding of cancer immunology and support the development of more effective and nuanced therapeutic approaches