Integration of microbiology, molecular pathology, and epidemiology: a new paradigm to explore the pathogenesis of microbiome-driven neoplasms

Abstract

Molecular pathological epidemiology (MPE) is an integrative transdisciplinary field that addresses heterogeneous effects of exogenous and endogenous factors (collectively termed “exposures”), including microorganisms, on disease occurrence and consequence utilising molecular pathological signatures of the disease. In parallel with the paradigm of precision medicine, findings from MPE research can provide aetiological insights into tailored strategies of disease prevention and treatment. Due to the availability of molecular pathological tests on tumours, the MPE approach has been utilised predominantly in research on cancers including breast, lung, prostate, and colorectal carcinomas. Mounting evidence indicates that the microbiome (inclusive of viruses, bacteria, fungi, and parasites) plays an important role in a variety of human diseases including neoplasms. An alteration of the microbiome may be not only a cause of neoplasia but also an informative biomarker that indicates or mediates the association of an epidemiological exposure with health conditions and outcomes. To adequately educate and train investigators in this emerging area, we herein propose the integration of microbiology into the MPE model (termed “microbiology-MPE”), which can improve our understanding of the complex interactions of environment, tumour cells, the immune system, and microbes in the tumour microenvironment during the carcinogenic process. Using this approach, we can examine how lifestyle factors, dietary patterns, medications, environmental exposures, and germline genetics influence cancer development and progression through impacting the microbial communities in the human body. Further integration of other disciplines (e.g. pharmacology, immunology, nutrition) into microbiology-MPE would expand this developing research frontier. With the advent of high-throughput next-generation sequencing technologies, researchers now have increasing access to large-scale metagenomics as well as other omics data (e.g. genomics, epigenomics, proteomics, and metabolomics) in population-based research. The integrative field of microbiology-MPE will open new opportunities for personalised medicine and public health.

The role of the microbiome in carcinogenesis

Microbiology characterises small organisms such as viruses, bacteria, fungi, archaea, and protozoa, and links them to the pathogenesis of human diseases. The human microbiome represents an interactive ecosystem consisting of numerous microorganisms, which continuously interacts with the environment and host, especially the immune system [1–4]. Accumulating evidence points to the role of endogenous and exogenous microorganisms in the pathogenesis of various neoplasms [5–8] as well as non-neoplastic diseases [4,9–13]. Microorganisms with established or possible carcinogenic effects include Helicobacter pylori and Epstein-Barr virus for gastric carcinoma [14]; hepatitis B and C viruses (HBV and HCV) for hepatocellular carcinoma (HCC) [15,16]; human herpesvirus-8 for Kaposi’s sarcoma [17]; human immunodeficiency virus for Kaposi’s sarcoma, aggressive B-cell non-Hodgkin lymphoma, and cervical carcinoma [17]; HPV for uterine cervical, anal, and oropharyngeal carcinomas [18,19]; human T-lymphotropic virus type 1 for adult T-cell leukaemia / lymphoma [20]; and Fusobacterium nucleatum (F. nucleatum) for colorectal carcinoma [21–23]. Furthermore, tumours arising among carriers of these pathogenic microorganisms may have distinct molecular pathological features compared to tumours arising among non-carriers [14,16,19,24–26]. Accumulating evidence indicates that alterations of the microbial ecosystem also play a major role in the pathogenesis of various neoplasms. Studies have associated altered microbial communities not only with tumours arising in the affected organs, but also with tumours arising in distant organs (e.g. the colorectal microbiome and non-colorectal gastrointestinal cancers [5,27–31]). In the tumour microenvironment, there is a dynamic interactive network that includes neoplastic cells, microorganisms, and immune cells, all of which are affected by the genetic architecture and epidemiological factors including aging, diet, nutrition, smoking, alcohol, adiposity, diabetes mellitus, physical exercise, and medication (Figure 1) [1–3,7,32–36]. Therefore, an integrative approach is required to elucidate the role of microbial communities in the pathogenic process of human cancers.

Figure 1.

Interactions of the microbiome, host immune system, and tumour in the microenvironment. The imbalance of microbial communities may cause neoplasms as well as non-neoplastic diseases.

Molecular pathological epidemiology (MPE) in the era of precision medicine

Epidemiology provides conceptual and analytical frameworks for investigations of an association of an endogenous or exogenous factor (termed “exposure”) with incidence of a disease or its consequence. Epidemiology aims to identify determinants of diseases and ultimately contribute to human health. Conventional epidemiological research has utilised organ- and/or function-based disease classifications [e.g. the International Statistical Classification of Diseases and Related Health Problems (ICD)] and addressed the relationship of an exposure with a specific disease entity or a collective group of similar disease entities. Owing to the availability of molecular diagnostic tests [37], molecular pathology has become a major subfield of pathology. Molecular pathological epidemiology (MPE) has been proposed as an integrative research field, which utilises molecular pathological signatures for disease sub-classification and addresses inter-individual differences in disease phenotypes in relation to specific epidemiological factors in human populations [35,38,39]. MPE studies examine occurrence of disease subtype or its consequence as an outcome, and assess a difference between subtypes of a single disease entity (Figure 2) [40]. “The unique disease (tumour) principle [41–43]” has paved the way for the MPE paradigm by highlighting the uniqueness of disease pathogenesis within each individual. Of note, MPE analyses can provide evidence for links between aetiological factors and pathogenic signatures, which can augment causal inference [39,44]. MPE analyses compute risk estimates for specific disease subtypes, thereby contributing to the advancement of personalised management strategies [45]. This echoes the aim of precision medicine for developing personalised prevention and treatment strategies to optimise the risk-benefit balance [46–50]. Due to the wide availability of established molecular pathology diagnostics and tumour tissue specimens for research, the MPE approach has been most commonly utilised in epidemiological studies of breast, colorectal, lung, and prostate cancers; however, this MPE model can be readily applied to research on any neoplastic and non-neoplastic diseases that represent substantial interpersonal heterogeneity [51,52]. The emerging MPE framework has been recognised and discussed in international conferences [53–56] and publications [6,7,57–98].

Figure 2.

Disease incidence and consequence analyses in molecular pathological epidemiology (MPE) research. Considering different characteristics of disease subtypes classified by molecular pathological signatures, incidence and consequence analyses assess the heterogeneity in associations of an exposure of interest with disease incidence and consequence, respectively. For a simple illustration, a disease with two subtypes is exemplified, but more than two subtypes can be modelled in MPE studies. Arrows indicate the time course of disease subtypes.

Integration of microbiology and MPE

Emerging evidence attests to not only the carcinogenic effects of the altered microbiome but also the distinctive phenotypes of tumours arising in the presence of specific microorganisms [14,16,19,24–26]. Microorganisms, the immune system, and tumour cells interact each other in complex manners. Hence, to better understand cancer aetiologies and their consequences in populations, analyses of the microbiome in various body sites including pathologically-altered tissue (such as tumour) should be adequately integrated into MPE (referred to as “microbiology-MPE” [40]). Microbiology-MPE provides a promising approach to explore the interpersonal heterogeneity of the carcinogenic process in relation to the altered microbial composition and to generate evidence for the role of microorganisms in specific processes of tumour initiation and progression [40]. While mechanistic studies have been a major part of microbiology research on carcinogenesis, any in vitro or in vivo model can never exactly replicate the complex molecular and cellular network in the tumour microenvironment in the human [99]. Therefore, insights from microbiology-MPE research would serve as particularly valuable evidence for the microbial aetiologies and pathogenesis of human neoplasms.

Study design

The epidemiological term “exposure” indicates a factor that may (or may not) cause, prevent, or influence an outcome such as incidence or consequence of a disease of interest. In microbiology-MPE research, the microbial profile is incorporated as an exposure or outcome variable. Namely, the microbial profile in non-neoplastic tissue or biospecimen obtained before cancer diagnosis can be examined as an exposure in relation to disease incidence or consequence as an outcome. The microbial profile in non-neoplastic or pre-malignant biospecimens can also serve as an intermediary outcome variable, particularly under a hypothesis that a certain exposure can cause the disease through altering the microbiome. Notably, the microbial profile in neoplastic tissue can serve as an outcome variable used to subclassify a tumour of interest in incidence analyses, or an exposure variable in consequence analyses. Nonetheless, we should be aware that the detection of a certain microorganism may suggest its causal relationship with the tumour or may merely be a consequence of tumour development.

Designs of microbiology-MPE studies are illustrated in Figures 3 and 4. In disease incidence analyses, the case-control study design has often been adopted because of the limited availability of prospective cohort studies with microbial data. Although there are advantages in the case-control study design including no requirement for follow-up of participants, lower costs compared to prospective cohort studies, and the ability to investigate rare tumours, special considerations are needed for potential biases such as recall and selection biases. In particular, selection bias may substantially affect the generalisability of study findings. In case-control studies examining tumour subtypes classified by the microbial profile, we compare proportions or levels of an exposure (which can be a microbial exposure) between patients with each subtype and controls, and assess a differential association between the exposure and subtypes (Figure 3A). In case-control studies examining the microbial profile as an exposure, we examine differences in the microbial communities between molecularly- or pathologically-classified subtypes and controls, and assess a differential association between the microbial profile and subtypes (Figure 3B). In contrast, the strengths of prospective cohort studies are that the source population that has given rise to tumours is well-defined, which diminishes issues due to selection bias. Recall bias can also be reduced by prospective data collection. When examining tumour subtypes classified by the microbial profile in a prospective cohort design, we can examine the association of an exposure to a specific factor (which can be microbes) and tumour subtypes (Figure 3C). If the microbial data before cancer diagnosis are available, we can link a microbe with the incidence of tumour subtypes classified by molecular pathological signatures of the tumour (Figure 3D).

Figure 3.

Disease incidence analyses in microbiology-molecular pathological epidemiology (microbiology-MPE) research. Arrows indicate the time course of disease subtypes. A. Case-control study examining a between-group difference in an exposure (which can be microbes) according to the microbial profile of the tumour (as an outcome variable). B. Case-control study examining a between-group difference in the microbial profile (as an exposure variable) according to molecular pathological signatures of the tumour. C. Prospective study examining the association of an exposure (which can be microbes) with incidence of tumours characterised by the microbial profile (as an outcome variable). D. Prospective study examining the association of the microbial profile (as an exposure variable) with incidence of tumours characterised by molecular pathological signatures of the tumour.

Figure 4.

Disease consequence analyses in microbiology-molecular pathological epidemiology (microbiology-MPE) research. Arrows indicate the time course of disease subtypes. A. Study examining the association of an exposure (which can be microbes) with disease consequence according to the microbial profile. B. Study examining the association of the microbial profile with disease consequence according to molecular pathological signatures of the tumour.

In disease consequence analyses, the microbial data can be analysed as an exposure and/or be used to sub-classify a tumour if a microbe of interest is a constituent of the tumour. Effects of epidemiological factors including microbial exposures on tumour progression may differ by the presence of other microbes that are associated with altered anti-tumour immunity and/or specific tumour molecular pathological features (Figure 4A). Microbes may exert promoting or suppressing effects on tumour growth through modulating specific signalling pathways; therefore, the association of specific microbes with tumour behaviour after clinical diagnosis may differ by molecular pathological signatures of the tumour (Figure 4B). As an experimental design, interventional trials using microbial manipulation strategies such as antibiotics, probiotics, prebiotics, and synbiotics targeting a specific tumour subtype can be conceived [100] though there exist concerns including emergence of antibiotic resistance.

In microbiology-MPE studies, specimen types for acquisition of microbial data should be discussed. In addition to tissue specimens, we can utilise various biospecimens including stool. There is an advantage of microbial analyses utilising tissue specimens. Compared to any microbe in stool, a microbe detected in tumour tissue is more likely to have a causal association with tumour development. Fresh frozen tissue specimens are amenable to various laboratory assays and metagenomic analyses for high resolution mapping of tissue microorganisms; however, their collection is not part of routine clinical practice. In contrast, formalin-fixed paraffin-embedded (FFPE) tissue blocks are often utilised, especially in population-based studies. Limitations of FFPE tissue specimens include contamination with non-pathogenic microorganisms and alteration of microbial composition due to fixation, processing, and storage of the tissue.

Study examples and proposals

In this section, we discuss microbiology-MPE studies that have examined intratumoural microorganisms in relation to various exposures, and successfully provided novel insights into cancer aetiologies in the human. There have been many epidemiological studies that examined microorganisms as exposures; most of those studies dealt with a disease entity of interest (e.g. organ-specific cancer) as a single outcome variable. Here, we focus on studies that investigated biological heterogeneity of tumour subtypes.

Human papillomavirus (HPV) is a sexually-transmitted DNA virus and has been involved in carcinogenesis of uterine cervical and anal cancers, and head and neck squamous cell carcinomas (particularly high-risk oncogenic types 16 and 18) [18,19]. In addition to distinctive characteristics of HPV-related carcinomas [19,101,102], evidence suggests differential responses to cancer treatment by HPV viral load [102–106]. HPV-positive subtypes may have a reduced ability to repair DNA damage, potentially representing better response to chemoradiotherapy which induces high-degree DNA damage and promotes apoptosis in neoplastic cells [102,106–108]. Intriguingly, studies have reported that epidemiological factors (e.g. caffeine) may exert radio-sensitizing effects in cancer cells through inhibition of DNA repair mechanisms (e.g. ATM, ATR) [109,110]; therefore, high HPV load may enhance the effects of those epidemiological exposures on clinical outcomes of cancer patients. In addition, viral pathogens including HPV contribute to high levels of epitopes in the tumour microenvironment [111,112], which potentially lead to an immunogenic tumour microenvironment and better response to immune checkpoint blockade. Accordingly, immune checkpoint blockade may synergise with HPV-16 vaccination to enhance their anti-tumour effects [113]. In HPV-negative head and neck squamous cell carcinomas, lifestyle factors such as smoking and heavy alcohol consumption play a major role in carcinogenesis [114,115], whereas HPV carriers are more common in current smokers than in non-current smokers [116]. In addition, genetic variants at the HLA region may be differentially associated with incidence of oropharyngeal carcinomas by HPV positivity [117]. Therefore, an integrative analysis of host and epidemiological factors incorporating tumour subtyping based on HPV infection status is mandatory when examining risk factors for head and neck carcinomas [115,118,119]. In aggregate, there are expected to be open opportunities for microbiology-MPE research in HPV-related malignancies.

Hepatocellular carcinoma (HCC) is the most common primary liver cancer and a collection of pathogenically heterogeneous carcinomas [15,16,120]. HBV and HCV are DNA and RNA viruses, respectively, which have been established as pathogens involved in chronic inflammation, cirrhotic changes, and carcinogenesis in the liver [15,16]. Heavy alcohol drinking is a major risk factor for incidence of non-viral HCC [15,121]. Recently, steatohepatitis among individuals with non-alcoholic fatty liver disease has gained attention as an alternative pathogenic mechanism of HCC [122–124]. It is evident that the incidence of HCC derived from non-alcoholic steatohepatitis (NASH) increases by the presence of metabolic risk factors including type 2 diabetes, obesity and, collectively, the metabolic syndrome. Emerging data indicate that dysbiosis of the gut microbiome may be associated with NASH and HCC [27–30,125,126]. Therefore, intervention strategies modulating lifestyle and the microbiota for prevention of HCC can be indicated for individuals with NASH [121,127–129]. Interestingly, among HBV carriers, the association of metabolic risk factors or insulin resistance with HCC incidence may be stronger in individuals with lower HBV load than in those with higher HBV load [130]. On the other hand, studies suggest that the association of epidemiological factors (e.g. smoking) with patient survival may differ by viral subtypes of HCC [131]. Taken together, the paradigm of microbiology-MPE would enrich cohort studies investigating incidence and mortality of HCC [71].

The colorectum is an organ that hosts the most abundant and diverse microorganisms in the human body, and dysregulation of the gut microbial ecosystem may result in colorectal carcinogenesis through impairing intestinal immune status, provoking a chronic proinflammatory reaction and affecting host metabolism [132–139]. Influences of the gut microbiome on carcinogenesis likely underlie the continuum of changes in colorectal tumour characteristics [such as microsatellite instability (MSI) status, CpG island methylator phenotype (CIMP), BRAF and PIK3CA mutations, and abundant intratumour F. nucleatum] along the detailed sublocations from caecum or ascending colon to rectum [140–143]. These findings led to the “colorectal continuum” model, implicating the pathogenic influences of the gut microbiome on colorectal tumours [144] (Figure 5). Compelling evidence indicates considerable heterogeneity between colorectal cancer subtypes due to underlying genetic, epigenetic, and microbial statuses in each tumour [39,145–148]. Accordingly, colon and rectal carcinomas have served as a practical disease model for microbiology-MPE research. F. nucleatum is a microbial pathogen that has been implicated in tumourigenesis and progression of colorectal cancer [21–23,26,149–151]. In two studies using a sample size of at least 200 cases of colorectal cancer (with available FFPE tissue specimens), intratumour F. nucleatum was detected in 9–13% patients [26,152]. Mechanistic studies indicate that F. nucleatum may have carcinogenic properties through up-regulating signalling pathways such as the CTNNB1 (beta-catenin)-WNT pathway [153] and also confer a metastatic potential to colorectal cancer [154]. In U.S. population-based studies, researchers explored the heterogeneity in associations of dietary patterns with colorectal cancer sub-classified by the presence of F. nucleatum in tumour tissue. A so-called prudent dietary pattern (rich in vegetables, whole grains, fish, fruits, and poultry) has been associated with a lower risk of F. nucleatum-positive colorectal cancer, but not with a risk of F. nucleatum-negative cancer [155]. In contrast, an inflammatory dietary pattern (rich in red and processed meat, refined grains, and sugar) has been associated with a higher risk of F. nucleatum-positive colorectal cancer, but not with a risk of F. nucleatum-negative cancer, and this differential association is pronounced for proximal colon cancer [156]. These studies provide population-based evidence for the potential role of the microbiome in mediating the relationship between diet and colorectal carcinogenesis. Studies support the roles of viruses and other bacteria, such as enterotoxigenic Bacteroides fragilis, pks-positive Escherichia coli, Enterococcus faecalis, Lactobacillus, and Bifidobacterium in colorectal cancer [5,157]. Therefore, investigation of these bacteria and a microbial community as a whole in relation to incidence and progression of colorectal cancer is also warranted.

Figure 5.

The colorectal continuum theory that highlights the interaction between the gut microbiome and immune cells in the colorectal microenvironment. Certain characteristics of colorectal cancer [e.g. abundant intratumour Fusobacterium nucleatum (F. nucleatum)] represent a gradual transition along the colorectal axis without a clear cut-off at the splenic flexure [140,143]. CIMP, CpG island methylator phenotype; MSI, microsatellite instability.

The immune system functions as the body’s defence mechanism against non-self, including microorganisms, foreign objects, and tumours with immunogenic peptides (“neoantigens”) [111,158–160]. Therefore, immunology has close ties to both microbiology and oncology [161–164]. Anti-tumour immune response is modified by not only immune cells but also tumour and microbial factors [3,33–36,165–168]. Immunology-MPE has been derived from an integration of immunology into MPE research to examine inter-individual heterogeneity in disease process by host immune status [3,169]. The immunology-MPE approach has been successfully utilised to investigate immune-enhancing or -suppressing effects of epidemiological factors in relation to the incidence of colorectal cancer and patient survival [170–180]. The data supporting dietary components and medications as immunomodulators can inform immunoprevention strategies against cancers [181–184]. In addition to the carcinogenic effects of F. nucleatum, evidence suggests molecular and pathological characteristics of F. nucleatum-positive colorectal cancer, including high levels of MSI and CIMP [26,152,185]. In MSI-high colorectal cancer, abundant neoantigens due to numerous frameshift mutations result in vigorous immune response to the tumour, potentially contributing to the enhanced efficacy of the immune checkpoint inhibitors [159,186–190]. On the other hand, immunosuppressive effects of F. nucleatum have been implicated as an alternative mechanism through which this microbe can exert carcinogenic effects [26,191–194]. Evidence suggests that the virulence factor FAP2 of F. nucleatum binds to the host factor Gal-GalNAc and thereby interacts with the inhibitory receptor TIGIT on T lymphocytes [193,195,196]. In our previous study, the enrichment of F. nucleatum in colorectal carcinoma tissue was negatively associated with the density of CD3⁺ pan-T cells infiltrating the tumour [26]. To address these seemingly conflicting findings (Figure 6A), a microbiology-MPE study was conducted based on large U.S. prospective cohort studies [197]. The presence of F. nucleatum was inversely associated with levels of tumour-infiltrating lymphocytes (TIL) in MSI-high colorectal tumours, but positively associated with TIL levels in non-MSI-high tumours. Based on these findings, it is speculated that the immunosuppressive effects of F. nucleatum may dominate in MSI-high colorectal cancer, whereas its proinflammatory effects may dominate in non-MSI-high cancer (Figure 6B) [192,194,198]. The data support interactive effects of tumour microbiota and molecular features on anti-tumour immunity and the potential role of F. nucleatum as a new immunotherapeutic target.

Figure 6.

The association between Fusobacterium nucleatum (F. nucleatum) and immune response to colorectal carcinoma depends on tumour microsatellite instability (MSI) status. A. The seemingly paradoxical relationship of F. nucleatum, tumour MSI status, and T lymphocyte infiltrates. B. The relationship of F. nucleatum with lymphocytic reaction according to tumour MSI status. F. nucleatum may exert immunosuppressive effects in MSI-high colorectal cancer, whereas F. nucleatum is associated with high-level lymphocytic reaction in non-MSI-high cancer in which immune reaction is generally inactive and the proinflammatory properties of F. nucleatum are thought to be dominant. MSI, microsatellite instability.

Of note, survival benefits from chemotherapy, radiotherapy, and immunotherapy may be modulated by the gut microbiome [8,93,199,200]. Using clinical faecal specimens of patients receiving anti-PDCD1 (PD-1) immunotherapy, studies have identified several candidate bacteria that may play a particular role in modulating benefits from the therapy [201–204]. The enrichment of specific bacteria in the gut (e.g. Faecalibacterium genus, Bifidobacterium genus) has been correlated with response to anti-PDCD1 blockade therapy. Compared to mice receiving faecal microbiota transplantation from non-responders, mice receiving transplantation from responders were associated with higher levels of tumour CD8⁺ cytotoxic T cells [201–203]. Therefore, immunomodulatory factors may affect the survival of cancer patients differentially by the microbial repertoire in the host.

Challenges in microbiology-MPE research

There exist challenges in the field of microbiology-MPE, some of which are attributed to epidemiological and MPE research in general as previously discussed [39]. To address the general issue of small sample sizes due to limited availability of specimens, MPE researchers should make efforts to maximise the number of cases available for research. The recent trend in data sharing and world-wide collaborative consortia may help collect large-scale data from different settings to increase statistical power and generalizability of study findings [205–208]. In order to facilitate collaborative projects on microbiology-MPE, we need to standardise procedures for collecting, processing, and storing biospecimens, because microbial compositions change due to those preanalytical factors. It should also be noted that certain preanalytical variables are uncontrollable especially when archival tissue specimens are used. For enhancement of robust data analyses, statistical methods specific for MPE analyses have been developed to assess disease heterogeneity in various study designs [80,209–216] and to address a bias due to missing data [217,218]. In addition, statistical methods such as inverse probability weighting and multiple imputation can be used to mitigate selection bias due to the availability of tissue specimens [218–222]. We need to establish certain general scientific standards across many fields for overall scientific rigour and reproducibility, while there is also a great need for standardisation of methodologies in specific fields.

There are also challenges specific to microbiological research. When a number of microorganisms or microbial communities as a whole are examined, microbiological studies may be prone to false positive findings if multiple hypothesis testing is not taken into account. In microbiology-MPE research, FFPE tissue specimens are often utilised due to relatively easy accessibility; however, the microbial compositions in those specimens might be different from those in vivo or in fresh tissue specimens. Finally, the necessity for transdisciplinary education to cover all component fields of microbiology-MPE is even more difficult to address [223]. Due to the nature of microbiology-MPE, it is of considerable importance to set up transdisciplinary research teams. Nonetheless, researchers in microbiology-MPE should at least obtain fundamental knowledge of the principles of the component fields. It should be noted that the new conceptual and methodological developments in MPE have only been accelerated after 2010. Integrative expertise and viewpoints have been essential in these developments. Essentially, collaboration of an expert in pathology and another expert in epidemiology (which in fact has been ongoing since the 1990s) cannot match a single expert who has obtained adequate knowledge in MPE resulting from appropriate training. A reformation of school curricula may be considered. For example, lectures on pathology and training at laboratories of clinical pathology and microbiology would be useful for students at schools of public health. We may implement training programmes of molecular pathology, epidemiology, microbiology, and immunology for physicians and researchers (e.g. lectures and/or hands-on training of epidemiology and biostatistics in departments of pathology). In addition, the International MPE Meeting series has successfully provided diverse types of researchers with the latest information on the methodology and findings of interdisciplinary MPE research [224,225], and can continue to be an educational resource. To address all of these challenges, we should develop educational programmes that integrate training for microbiology, molecular pathology, and epidemiology.

Future perspectives and conclusions

The integrative approach of microbiology-MPE can be a powerful tool that potentially expands our knowledge of the aetiologies and pathogenesis of carcinomas evolving through dysregulated microbiota [40,226]. While our discussion primarily focuses on neoplastic diseases, this methodology can be readily extrapolated to non-neoplastic diseases that have substantial inter-personal heterogeneity in the disease process and a potential link to the microbiome. It should be noted that the purpose of using the name of a particular scientific discipline is to clarify the need for adequate education, training, knowledge, and expertise in order to properly conduct high-quality research in the discipline. That is, to conduct microbiology-MPE research, one should be well trained in all of the component fields, i.e. microbiology, molecular pathology, and epidemiology.

Due to the nature of MPE as a method-based discipline (but not a disease- or organ-specific discipline), it would be possible to further integrate other important disciplines, including immunology, pharmacology, and nutritional science, and optimise the potential of microbiology-MPE. Preclinical studies suggest the modulatory effects of common medications (e.g. aspirin, statins, and metformin) [227–231] and dietary and nutritional factors (e.g. vitamins, omega-3 polyunsaturated fatty acids, red and processed meat, coffee, alcohol, fibre, and sugar) [231–238] on the microbiota and host immune response. Owing to the advance and cost reduction in high-throughput sequencing technologies and analytical platforms for large-scale data, multi-omics data are increasingly available for epidemiology studies [239]. Sequencing-based analyses including RNA transcriptomic sequencing, metagenomic analyses [240–243] and single cell sequencing [244–246] will allow us to examine the complex relationship of the microbiota, immune cells, and tumour cells in a more comprehensive fashion.

In conclusion, microbiology-MPE can provide a novel methodological framework to gain insights into the tumour-immune-microbiome interaction from human tissue and population data, thereby informing targeted microbiome-modulating strategies for cancer prevention and treatment. Given the increasing availability of omics data on host and tumour with microbial and immune profiles, this new approach should further promote the global trend of precision medicine. Nonetheless, analytical frameworks and educational programmes should be urgently refined specifically for MPE and microbiology-MPE research.

Why do we need names for scientific fields?

The primary purpose of this article is to discuss an integration of microbiology, molecular pathology, and epidemiology. Why do these fields need names? The existence of the name of a scientific field implies that a specific set of education, training, knowledge, and expertise is needed to conduct research in that field. For example, if one investigates the epidemiology of human papillomavirus (HPV), one needs to gain adequate knowledge and expertise in both microbiology and epidemiology through appropriate education and training. However, not all researchers who study the epidemiology of HPV may have proper expertise in both microbiology and epidemiology. There are substantial pitfalls in non-experts conducting research studies. Hence, the integration of microbiology, molecular pathology, and epidemiology necessitates knowledge and expertise in all of these fields. The broader goal of this article is to illustrate an increasing need for transdisciplinary education and training systems for future science.

Abbreviations:

CIMP

CpG island methylator phenotype

FFPE

formalin-fixed paraffin-embedded

HBV

hepatitis B virus

HCC

hepatocellular carcinoma

HCV

hepatitis C virus

HPV

human papillomavirus

MPE

molecular pathological epidemiology

MSI

microsatellite instability

NASH

non-alcoholic steatohepatitis

TIL

tumour-infiltrating lymphocytes