Abstract
While immuno-oncology has made significant advances in activating local tumor immune responses, leading to improved outcomes, the role of systemic immunity in cancer incidence remains poorly understood. Le Cornet and colleagues prospectively studied circulating immune cells quantified by DNA methylation markers in relation to incidence of breast, colorectal, lung, and prostate cancer among initially healthy individuals. A positive association with cancer risk was observed for higher FOXP3+ T-cell–mediated immune tolerance and lower CD8+ T-cell–mediated cytotoxicity. Further studies of systemic immunity in cancer development are crucial to identify novel prediction markers and interventional targets for cancer immunoprevention.
See related article by Le Cornet et al., p. 1885
Harnessing the immune system to improve cancer prevention and treatment represents a major frontier of cancer research. The now recognized role of local immune response in prognosis and the impact that tumors have on modulating the tumor immune milieu has led to significant breakthroughs in immunotherapy, which has become a standard regimen for patients with certain advanced cancers. Despite this, immunoprevention research and the role of immune function in tumor development and early progression remains in its infancy.
Systemic immunity may influence cancer risk. For example, patients with HIV infection or pharmacologic immunosuppression have a higher cancer risk than the general population. Furthermore, germline genetic variants affecting immunosuppressive cells [e.g., regulatory T cells (Treg)] and immune regulators predict the abundance of tumor-infiltrating immune cells and cancer risk, and compared with cancer-free individuals, patients with cancer have higher circulating functional Tregs.
Despite this, direct evidence linking prediagnosis systemic immunity to cancer risk is limited. This is due, in part, to the lack of tools to deeply characterize systemic immunity using archival blood samples in large populations necessary to study cancer incidence. The current gold standard, flow cytometry–based cell counting, requires intact leukocytes in fresh or well-preserved blood and its performance is significantly compromised by cell deterioration within a few hours after blood collection or lysing due to freezing and thawing. Therefore, development of a robust quantitative method for assessment of immune cells in banked samples from large-scale prospective studies is critical to better understand the role of systemic immunity in cancer risk. Recently, epigenetic assays have leveraged unique methylation markers of specific immune cell types to characterize immune cell composition using stored DNA from buffy coat. One such assay measures circulating cellular ratio of immune tolerance (“immunoCRIT”) based on the ratio of FOXP3+ Treg and CD3+ T cells using the quantitative PCR (qPCR) for FOXP3 and CD3 cell type–specific demethylated loci. This assay has been validated against flow cytometry in DNA-spiking experiments (1). Also, the assay showed good stability (Spearman correlation coefficient = 0.53) within individuals over 15 years (2). Applying this assay to a prospective study in the European Prospective Investigation into Cancer and Nutrition (EPIC)-Heidelberg cohort, higher immunoCRIT prior to diagnosis was associated with increased risk of lung, colorectal, and estrogen receptor (ER)-negative breast cancer, independent of lifestyle risk factors (2). That study provided the first well-powered, prospective evidence for a role of increased peripheral immune tolerance in cancer development.
In this issue, Le Cornet and colleagues extend this work by further quantifying different lineages of leukocytes in archival blood using novel cell type–specific DNA methylation qPCR assays (3), including total (CD3+), cytotoxic (CD8+), and regulatory (FOXP3+) and nonregulatory (FOXP3−) Th lymphocytes, as well as neutrophils (CD15+), monocytes (CD14+), natural killer cells (CD56+), and B lymphocytes (CD19+). Relative counts of these immune cell types were calculated after normalization to ensure that the sum of all major cell lineages equaled 100%. The measurements showed a remarkably high correlation (0.70–0.90) with flow cytometry and good within-person reproducibility (Spearman correlation coefficient: 0.48–0.67) over 14–15 years. The assay was conducted in a case–cohort study of incident breast (n = 181), colorectal (n = 96), lung (n = 63), and prostate cancer (n = 201) and a randomly selected subcohort (n = 254). In age- and sex-adjusted analyses, a HR of 0.80–0.90 was noted for CD8+ T cells for lung and breast cancer, whereas a HR of 1.50–2.30 was observed for FOXP3+ T cells. Similar results were found after adjustment for lifestyle factors correlated with immune cell profiles.
The study has multiple strengths that enhance its impact, including its prospective design with a median follow-up of 6.7 years, use of rigorous immune assays that have good validity and reproducibility, and assessment of multiple cancer types that have both shared and distinct risk factors. Furthermore, the authors examined multiple, inter-related immune cell subtypes that have both cytotoxic and immunosuppressive roles. Importantly, their findings highlight a potential role in cancer incidence of systemic immune homeostasis between CD8+ cytotoxic effector T cells and FOXP3+ Tregs. This balance has been well-characterized in the tumor microenvironment and is established as a determinant for cancer survival outcomes. The study by Le Cornet and colleagues indicates that the cancer–immunity cycle (4) may start years before cancer diagnosis and play a role across the entire continuum of cancer development, from initiation, progression, to ultimately evasion. However, this initial study had relatively small sample sizes of individual cancer types and, given the observational design, the associations could be due to residual confounding by cancer risk factors that have a strong immunomodulatory effect, such as adiposity, physical activity, and diet. Indeed, a dietary and lifestyle pattern associated with higher circulating inflammatory markers has been linked to higher risk of various cancers, including colorectal and breast cancer (5). In contrast, no strong associations have been found between these inflammatory markers/exposures and incidence of prostate cancer. This, together with the Le Cornet and colleagues' findings of a significant association of systemic immune cells with breast and colorectal cancer, but not prostate cancer, raise the question to what extent the altered immune cell composition in prediagnosis blood represents a true cause or a marker of inflammation or immunomodulatory-related environmental exposures.
Despite this, the Le Cornet and colleagues' study (6) provides a strong rationale for further studies to better understand the role of systemic immunity in cancer susceptibility. First, the findings need to be confirmed in larger and more diverse populations. While there currently are no strong data suggesting that the immune system operates differently across populations, consistent results observed across studies will be crucial to enhance a case for causality. Second, further methodologic improvements in quantification of the immune cell composition are needed. While the method used by Le Cornet and colleagues had good performance, it measures a limited number of cell types and requires separate qPCR runs for each cell marker. Other methods have been developed to computationally predict cell-type proportions using blood-derived DNA methylation data in epigenome studies (7). This deconvolution approach has been well-validated and represents an attractive, and more cost-effective option for large-scale studies with already available genome-wide methylation data. Methylation deconvolution also assesses a limited number of immune cell subsets, thus continued development is needed to identify methylation markers unique to more specialized immune cell types.
Third, further studies are needed to better assess the relative importance of different immune cell subsets, individually or in combination, for cancer risk. In the Le Cornet and colleagues' study, a decomposition analysis was performed, in which the relative count of neutrophils was not included in the model, with mutual adjustments for other cells. Assuming that the total immune cell counts are constant, the HR reflects the effect of replacing 1% of neutrophils with 1% of other specific cell subtypes. Although the decomposition approach allowed for simultaneous assessment of several immune cell types on the same scale, the dramatically different distributions of these cells made it difficult to compare the risk estimates. For example, the mean relative count of FOXP3+ T cells was less than the one fifth of that for CD8+ T cells and the one tenth of that for FOXP3− T cells (1.4%, 7.4%, and 16.6%, respectively). Therefore, the HR per 1% increment as calculated in the study can have dramatically different meaning for different cell subsets, necessitating future studies to account for the differences in abundance. This becomes more important as the methodologies allow for assessment of more specific immune cell types. Moreover, to ensure that the risk estimates in the decomposition model have a proper interpretation for a substitutional effect, total cell counts should be adjusted for in the model, similar to the total energy adjustment in the nutrient density model in nutritional epidemiology (8). Fourth, while increasing evidence indicates that alterations in the immune response emerge early in the precancer phase (9), it remains unknown to what extent preclinical lesions and early invasive cancer contribute to the systemic immune landscape. Therefore, work is needed to characterize the relationship of circulating immune cell profiles with the tumor immune microenvironment at different times in the carcinogenic process. Finally, the implications of the Le Cornet and colleagues' findings for cancer prevention require further investigation. If systemic immune profiles are confirmed to be an important risk factor for cancer incidence, it will be crucial to assess whether these measures can improve risk prediction models to identify those at high risk of cancer or enhance early detection. Such data could be used to identify individuals who may benefit from immunoprevention interventions, such as prophylactic vaccines, targeted immunomodulators, and dietary approaches (10). Additional understanding of environmental, lifestyle, medication, and genetic modulators of system immune function will be critical to inform prevention recommendations in this space.
We believe that the Le Cornet and colleagues' study will stimulate further research in probing the role of systemic immune patterns in cancer development. These investigations will fill an important and yet missing piece in the area of immuno-oncology and ultimately lead to full realization of the promise of cancer immunoprevention.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Acknowledgments
This work was supported by the American Cancer Society Mentored Research Scholar Grant (MRSG-17-220-01 – NEC to M. Song) and by the NIH grant (R00 CA215314 to M. Song).