Cancer Microbiology

Abstract

Microbes play important roles in cancer from direct carcinogenic effects to their use in treatment. Cancers caused by microorganisms account for approximately 15% of cancers, primarily in low- and middle-income countries. Unique features of infectious carcinogens include their transmissibility, mutability, and specific immune interactions, which provide challenges and opportunities for cancer prevention and treatment. For these agents, infection control through exposure reduction, antivirals, antibiotics, and vaccines is cancer control. In addition, developing evidence suggests that microorganisms including the human microbiome can indirectly modulate cancer formation and influence the effectiveness and toxicity of cancer treatments. Finally, microorganisms themselves can be used to prevent or treat cancer. The convergence of these factors signals the emergence of a new field, cancer microbiology. Recognition of cancer microbiology will spur research, stimulate cross-disciplinary training, inform drug development, and improve public health.

Since the discovery more than a century ago that certain viruses cause cancer in animals, studies of tumor viruses have had a profound impact on our understanding of carcinogenesis. These studies led to the discovery of oncogenes such as ras and myc and tumor suppressor proteins such as p53. Importantly, some viruses and other microorganisms trigger the development of human cancer. Twelve microbes (1 bacterium, 8 viruses, and 3 parasites) are currently recognized as human carcinogens by the International Agency for Research on Cancer or the National Toxicology Program. These microbes are thought to be responsible for more than 2.2 million incident cancers annually and account for at least 15% of the global cancer burden (Table  1) (1,2). This is an underestimate because it does not include all cancers associated with HIV infection or cancers whose infectious basis is not yet recognized. Some cancers with known microbial causes, most notably cervical cancer, which is strictly dependent on human papillomavirus (HPV) infection, are more common in resource-limited regions, where they cause substantial mortality (Figure  1). With increased international travel, migration, and climate change, individuals will be exposed to an increasingly diverse array of microorganisms, including some that are carcinogenic. The human microbiome, the trillions of bacteria and other microbes that colonize our tissues, appears to modulate cancer risk or affect the response to therapy. In addition, cancers associated with microorganisms present unique opportunities for prevention and treatment, and in some situations, modified microorganisms can be used therapeutically. In this review, we summarize this and other information that herald the emergence of a new field, cancer microbiology, encompassing the extensive and diverse relationships between microorganisms and cancer (Figure  2).

Table 1.

Cancers associated with known infectious carcinogens^a

Infectious carcinogen	Major associated cancers	Annual casesb
Bacteria
Helicobacter pylori	Gastric carcinoma (stomach cancer), MALT lymphoma	770 000
Viruses
Human papillomavirus (HPV)	Cervical carcinoma, other anogenital cancers, oropharyngeal and other head-and-neck carcinomasc	640 000
Hepatitis B virus (HBV)	Hepatocellular carcinoma (liver cancer)	420 000
Hepatitis C virus (HCV)	Hepatocellular carcinoma, non-Hodgkin lymphoma	170 000
Epstein Barr virus (EBV)	Burkitt lymphoma, nasopharyngeal carcinoma, non-Hodgkin lymphoma, gastric carcinoma	120 000d
Kaposi sarcoma herpesvirus (KSHV)	Kaposi sarcoma, primary effusion lymphoma, multicentric Castleman disease	44 000
Human T-cell leukemia virus type 1 (HTLV-1)	Adult T-cell leukemia/lymphoma	3000
Merkel cell polyomavirus (MCPyV)	Merkel cell carcinoma (neuroendocrine skin cancer)	—e
Human immunodeficiency virus (HIV)	Various cancers	—f
Parasites
Schistosoma haematobium	Bladder cancer	7000
Clonorchis sinensis and Opisthorchis viverrini	Cholangiocarcinoma (bile duct cancer)	1300

^aInfectious agents classified as carcinogens by the International Agency for Cancer Research of the World Health Organization or the National Toxicology Program of the United States Department of Health and Human Services (https://www.cancer.org/cancer/cancer-causes/general-info/known-and-probable-human-carcinogens.html). MALT = mucosa-associated lymphoid tissue.

^bInfection-associated cases worldwide in 2012 as estimated by Plummer et al. (1).

^cOnly certain types of high-risk HPV (eg, HPV16 and HPV18) are associated with the cancers listed here. Other HPV types are associated with skin cancer, which is not included in this tabulation.

^dThis total does not include EBV-associated gastric carcinoma, which number in excess of 50 000 cases annually.

^e— indicates that no estimate was made in Plummer et al. (1). MCPyV is thought to be responsible for several thousand deaths annually worldwide.

^f— indicates that no esimate was made in Plummer et al. (1) because HIV is classified as a human carcinogen but is not generally regarded as a tumor virus. Many cancers associated with HIV infection are caused by the tumor viruses listed in this table.

Figure 1.

Global distribution of cancers associated with infectious agents. The figure shows a world map and the fraction of cancers in 2012 attributable to infections (1). This figure is reprinted from Plummer M, de Martel C, Vignat J, Ferlay J, Bray F, Franceschi S. Global burden of cancers attributable to infections in 2012: a synthetic analysis. Lancet Glob Health. 2016;4(9):e609-16, with permission from Elsevier.

Figure 2.

Infectious agents and cancer. The figure summarizes in schematic form the many ways infectious microorganisms may enhance cancer formation and how various interventions can suppress cancer formation.

Microorganisms With Known Carcinogenic Activity

There are several unique features of infectious carcinogens (Figure  3). Unlike classic carcinogens like chemicals or radiation, infectious carcinogens can be transmitted from one host to another (or to many others), often after the first host has traveled far from the site of the initial exposure. Infectious carcinogens can mutate and evolve, which could affect their carcinogenicity or susceptibility to immune control or treatment. They can specifically modulate the immune system, and the immune response in turn can affect the ability of the microorganism to replicate, spread, or cause disease. Initial exposure to the carcinogen can generate an immune response that inhibits infection on subsequent exposures, the basis for prophylactic vaccination. Most importantly, prevention or treatment of the underlying infection can reduce cancer risk. These features greatly influence the carcinogenic potential of infectious agents and the design and success of preventive and therapeutic strategies.

Figure 3.

Unique features of carcinogenic microorganisms. Schematic diagram of key ways infectious carcinogens differ from classic physical or chemical carcinogens: they can be transmitted between hosts; they can expand and spread; they can mutate and evolve; they can specifically modulate the immune and inflammatory response and vice versa; and infections can be prevented or cured by public health measures including vaccination or medical treatments, resulting in reduced cancer risk.

Classic Tumor Viruses

Worldwide, tumor viruses are responsible for approximately 1.5 million cancer cases in humans annually (Table  1) (1,2), typically following persistent or latent infection and often in the setting of immune suppression or chronic inflammation. A recent comprehensive sequence analysis of cancer cell genomes and transcriptomes suggests that few cancers in humans are likely to be caused by as-yet-unrecognized tumor viruses (3), but this conclusion may be premature because this analysis could not detect viral genes or transcripts that do not persist in cancer cells.

Tumor viruses employ diverse mechanisms of carcinogenesis (4). HPV, Epstein Barr virus (EBV), Kaposi sarcoma-associated herpesvirus (KSHV), and Merkel cell polyomavirus (MCPyV) express strong, well-characterized oncoproteins that stimulate cell proliferation, inhibit cell death, and/or cause genetic instability. The immune evasion strategies commonly employed by these viruses may also impair immune surveillance against tumor development.

Even for viruses with strong oncogenes, the vast majority of infected people do not develop cancer. This likely reflects at least in part a requirement for the accumulation of additional cell mutations for tumor formation. Consistent with this explanation, cancers induced by viruses with strong oncogenes contain many mutations (albeit often fewer and distinct mutations than in nonviral cancers) (3,5-7). Notably, viral oncogenes and the host immune responses to infection, including chronic inflammation (which generates mutagenic reactive oxygen species), tissue injury, or DNA damage, can increase mutagenesis in infected and possibly even in nearby uninfected cells. Thus, a major theme in cancer microbiology is that the host immune and inflammatory response to infection is a key factor in carcinogenesis.

Despite the presence of cellular mutations, some cancers induced by viruses remain dependent on expression of the viral oncogenes. In the case of HPV and cervical cancer, repression of HPV oncogene expression in cervical cancer cells or tumors in genetically engineered mice reactivates dormant tumor suppressor pathways and causes cellular senescence, apoptosis, or tumor regression (8–10). Thus, continuous expression of the HPV oncogenes is required to maintain ongoing proliferation of the cancer cells, an early example of oncogene addiction. HPV infection also plays a causal role in some oropharyngeal, anal, vaginal, vulvar, penile, and skin cancers. Different HPV types are associated with different cancers: cervical cancer is caused by a variety of high-risk genital HPV types of the α class including HPV16, whereas HPV16 is the only HPV type commonly found in the other HPV-associated anogenital cancers, and the genetically distinct β-type HPVs are implicated in nonmelanoma skin cancer (11,12). It is not known if the same carcinogenic mechanisms are at play in each of these HPV-associated cancers or if ongoing expression of the viral oncogenes is required, as it is in cervical cancer.

Some human tumor viruses, such as hepatitis B virus (HBV) and hepatitis C virus (HCV), lack clearly defined oncogenes. Chronic infection by these viruses leads to the development of approximately 80% of hepatocellular carcinoma (HCC), the third leading cause of cancer mortality worldwide, most often in the setting of liver cirrhosis (1). Repeated cycles of tissue injury due to virus replication and the host immune and inflammatory response to infection result in liver regeneration and mutagenesis, which are thought to be key in hepatic carcinogenesis (13). Viral gene products such as HBV HBx may also play a role in HCC formation, as might integration of HBV DNA into the liver cell genome near growth-promoting genes (3,14). The mechanism of oncogenesis by the retrovirus human T-cell leukemia virus type 1 (HTLV-1) may involve the ability of viral accessory proteins including TAX to induce genomic instability and regulate expression of cellular genes that affect cell growth (15).

Although viral carcinogenesis in humans has been intensely studied for decades, many unanswered questions remain. For example, why does only a small proportion of infected individuals develop cancer, and why do years or decades often lapse before tumors arise? As noninfectious carcinogens (eg, cigarette smoke) often display similar properties, studies of tumor viruses with relatively well-defined carcinogenic mechanisms may help elucidate general principles of carcinogenesis. Why is cervical cancer strictly dependent on HPV infection, whereas HPV accounts for only a fraction of other HPV-associated cancers? Why does EBV-induced nasopharyngeal carcinoma have striking geographical restriction, even though approximately 90% of the worldwide adult population has been infected with EBV? Host genetics and environmental factors such as a diet rich in fermented and preserved foods have been implicated in this pattern (16,17), but the underlying mechanisms are not understood. How do viruses without strong oncogenes promote cancer, and what is the contribution of the chronic inflammation and the immune response? These and other questions must be addressed to understand disease pathogenesis and rationally develop optimal preventive and therapeutic approaches to cancers associated with virus infection.

Human Immunodeficiency Virus

Even though HIV is not considered a tumor virus, HIV infection confers substantial excess risk for many cancers, and cancer prognosis in HIV-infected individuals is often worse than for people without HIV, even for cancers whose incidence is not increased (18,19). Certain cancers in the context of an active HIV infection establish the diagnosis of AIDS (20). These AIDS-defining cancers (ADCs) are caused by known tumor viruses—HPV (cervical cancer), KSHV (Kaposi sarcoma), and EBV (non-Hodgkin lymphoma)—and these cancers are thought to be primarily driven by HIV-induced destruction of lymphocytes that normally suppress tumor virus replication (21–23). The use of combination antiretroviral therapy (ART) leading to HIV suppression and at least partial immune reconstitution has reduced the incidence of ADCs (Figure  4), demonstrating that antiviral therapy can prevent cancer (24,25).

Figure 4.

Historic and projected cancer burden in the HIV-infected population. The graph shows marked decrease in AIDS-defining cancers (ADCs) in the United States with the widespread adoption of antiretroviral therapy, with little change in the number of non-AIDS-defining cancers (NADCs), including lung cancer, the most common NADC. Data were extracted from Shiels et al. (24).

Despite these successes, people with treated HIV remain at increased risk for ADCs and other malignancies known as non-AIDS-defining cancers (NADCs), which are now the most common tumors in this population (Figure  4) (24,26). After adjustment for known risk factors, the risk of developing both ADCs and NADCs is highest for individuals who failed to achieve virologic suppression by ART (27). The greatest excess risk is seen for NADCs associated with co-infections, including Hodgkin lymphoma (EBV), HCC (HBV and HCV), and anal and head-and-neck cancer (HPV), as well as lung cancer associated with prior bacterial pneumonia (28). The pathogenesis of NADCs is unclear but may reflect residual immune dysfunction, inflammation due to persistent viral infection, metabolic dysfunction, translocation of bacteria from the intestinal tract to extra-intestinal sites, or mutagenic or other effects of the antiretroviral drugs. However, it should also be noted that some studies suggest that the incidence of prostate cancer and some forms of breast cancer is decreased in HIV-infected individuals on ART (29,30), observations with as yet no clear explanation.

Helicobacter pylori. The bacterium Helicobacter pylori is also a recognized human carcinogen. H. pylori causes chronic gastritis, which progresses to gastric carcinoma in a small percent of cases (31,32). Because approximately half of the human population is infected with H. pylori, gastric cancer is the most common cancer caused by a microorganism (Table  1). H. pylori has also been implicated in other cancers including mucosa-associated lymphoid tissue (MALT)  lymphoma and colorectal and pancreatic cancer (33), but there is also evidence that H. pylori infection confers a protective effect against esophageal adenocarcinoma (34).

Chronic inflammation induced by H. pylori proteins such as CagA and VacA is thought to play an important role in carcinogenesis (35). These effector proteins are required for virulence during normal infection and may enhance tumor formation (36). Thus, as is the case for viruses, host responses to infection are likely to be a key factor in bacterial carcinogenesis.

Eucaryotic Multicellular Parasites

Three multicellular parasites are recognized as human carcinogens (37). The blood fluke, Schistosoma haematobium, is found in the water in parts of the Middle East, Africa, and Asia and causes schistosomiasis, including urogenital schistosomiasis, which can progress to bladder cancer. Liver flukes (flatworms), Opisthorchis viverrini and Clonorchis sinensis, induce the development of cholangiocarcinoma, cancer of the biliary tree (38,39). Adult flukes reside in the human bile ducts for many years, feeding on epithelial cells and excreting bioactive molecules, causing mechanical damage and immunopathology due to inflammation and exerting direct toxic effects.

Microbial Links to Cancer Beyond Recognized Infectious Carcinogens

Although H. pylori is the only bacterium currently recognized as a human carcinogen, there are epidemiological associations between other bacteria and cancer. For example, Streptococcus bovis endocarditis is associated with a 14.5-fold increased risk of colorectal cancer (40), and serological studies have implicated Chlamydia trachomatis and possibly mycoplasma infection in the genesis of epithelial ovarian cancer (41,42). However, there is no definitive evidence to date that these bacteria play a causative role in cancer.

Much recent attention has been focused on the microbiome, a complex collection of many microbial species, which is increasingly implicated in cancer formation and response to treatment (Figure  5) (44–47). Different subtypes of breast and prostate cancer display characteristic populations of bacteria and viruses (48,49), and particular bacteria that reside in the colon are associated with the progression of adenomas to carcinoma (50,51). Although many types of microorganisms constitute the microbiome, here we focus on bacteria in the human microbiome.

Figure 5.

Effect of the microbiome on the hallmarks of cancer. The outer circle shows the 10 hallmarks of cancer defined by Hanahan and Weinberg (43). The center shows the effects of specific bacteria or their toxins on these hallmarks (44). This figure is reprinted from Fulbright LE, Ellermann M, Arthur JC. The microbiome and the hallmarks of cancer. PLoS Pathog 2017;13(9): e1006480. doi.org/10.1371/journal.ppat.1006480. BFT = Bacteroides fragilis toxin; FadA = Fusobacterium nucleatum adhesion protein; Fap2 = Fusobacterium nucleatum surface protein; pks+ E. coli = colibactin-producing E. coli; TLR = Toll-like receptor.

Plausible mechanisms by which bacteria may be carcinogenic have been identified. Salmonella typhi, implicated in gall bladder cancer, and Campylobacter jejuni and some pathogenic types of Escherichia coli, implicated in colon cancer, secrete toxins that damage DNA (52–56). In a mouse model, C. jejuni can promote colorectal tumorigenesis in a process dependent on cytolethal-distending toxin, a deoxyribonuclease (57). DNA damage, together with chronic inflammatory responses and tissue injury, is thought to result in mutations in cellular genes, thereby promoting carcinogenesis (58–60). Enterotoxigenic Bacteroides fragilis toxin cleaves E-cadherin, resulting in Wnt/β-catenin signaling and altered gene expression in colonic epithelial cells (61). Certain strains of E. coli produce colibactin, a toxin that induces mutations characteristically found in colon tumors (55,62). Notably, compared with healthy individuals, colonic biofilms from patients with familial adenomatous polyposis show enrichment for genes encoding B. fragilis toxin and colibactin, and the combination of these toxins accelerates tumor formation in mice (63). In other mouse experiments, a dysbiotic gut microbiome (one with imbalance of bacterial species) can cause colitis and thus increase the risk for the development of colorectal cancer (64). There is also evidence to suggest that the gut microbiome can promote HCC (65).

It has been proposed that the skin microbiome plays a role in leukemias and lymphomas by inducing the proliferation of lymphocytes directed against epitopes of commensal skin microorganisms (66). The microbiome can also affect the secretion of hormones, which could influence hormone-response tumors (67), and it plays a general role in the tumor microenvironment by modulating and pro- and antitumorigenic host immune responses (68,69), possibly contributing to tumor formation (70).

The microbiome is also a metabolic powerhouse and synthesizes molecules that can act as carcinogens, such as bile acids including deoxycholic acid, which have been implicated in colorectal cancer (71,72). In addition, the bacterial population composition in the human microbiome can vary markedly over time or between individuals, exposing hosts to a complex and varying constellation of carcinogens and other bioactive molecules. Finally, the gut microbiome is implicated in obesity (73), which will soon overtake smoking as the leading preventable cancer risk factor (74). It should be emphasized, however, that despite many tantalizing clues, the causal role of bacterial species other than H. pylori in cancer has not been established.

In addition to bacteria and viruses, fungi in the microbiome (the mycobiome) may also contribute to human cancer, with different fungal genera in the fecal or oral flora associated with colorectal, oral cavity, and pancreatic cancer (75). For example, certain fungal infections are associated with pancreatic ductal adenocarcinoma cancer because of activation of mannose-binding lectin (76).

Hit-and-Run Carcinogenesis?

The detection of viral genes in cancer cells is often the first laboratory evidence implicating a virus with a specific cancer. Indeed, high-risk HPV, KSHV, and Merkel cell polyomavirus were discovered because of the presence of viral DNA or RNA in tumors (77–80). A virus or other carcinogenic microorganism could also trigger carcinogenic progression (for example, by inducing genomic instability or epigenetic reprogramming) and then depart without leaving residual microbial genes or gene products in the cancer cells, so-called hit-and-run carcinogenesis (81). For example, skin carcinogenesis triggered by cutaneous HPV has been proposed to involve a hit-and-run mechanism in which the virus potentiates mutagenesis by ultraviolet radiation but is not needed to maintain the cancerous state (82). Similar principles may apply in the context of the microbiome, whose composition is dynamic and impacted by changes in diet, environment, or other factors. Bacterial toxins may trigger oncogenic events but not persist for the time needed for tumors to develop. If hit-and-run microbial carcinogenesis occurs in humans, the proportion of cancer with an infectious contribution may be substantially higher than is currently recognized.

Synergistic and Antagonistic Effects of Infectious and Noninfectious Carcinogens

In addition to the impact of individual infections, there are important interactions between microbes that increase cancer risk, including the interactions between HIV and tumor viruses. Another interesting example is Burkitt lymphoma, the most common pediatric tumor in malaria-endemic areas of sub-Saharan Africa. EBV infection is required for tumor development, but co-infection with the malarial parasite Plasmodium falciparum plays a synergistic role by causing lytic reactivation of EBV and chromosomal translocation involving the c-myc oncogene, resulting in tumor formation (37). Another example of synergy between microorganisms is the increased risk of HCC in individuals co-infected with HBV and either HCV or hepatitis delta virus, a satellite RNA virus (13).

The microbiome can affect the susceptibility of host cells to infection by viruses by affecting virus attachment or replication or by inducing innate or adaptive immune signaling (83–85). For example, the production of short- and medium-chain fatty acids by the microbiota might affect lytic reactivation of oncogenic herpesviruses (86).

An individual’s prior personal infection history can affect cancer risk by inducing a specific immune response that facilitates or inhibits infection with an infectious carcinogen or affects development or growth of cancers (87). Infection with a microorganism that bears epitopes similar to those expressed by a carcinogenic microbe or cancer cells may generate an immune response that inhibits subsequent infection by the carcinogen or attacks the cancer itself. The latter mechanism has been proposed to explain the negative correlation between mumps and ovarian cancer (88,89). Conversely, a history of influenza virus infection is associated with increased lung cancer risk (90). In addition, infection-mediated or microbiome-induced skewing of immune cell populations or differentiation may indirectly affect cancer risk, possibly accounting for the alleged linkage between the immune response to some common childhood infections and pediatric leukemia (87,91).

Classic carcinogens such as sunlight, smoking by-products, alcohol, and other factors can also suppress immunity or induce mutations that cooperate with microorganisms to promote cancer development. For example, in addition to the necessary role of HPV, cervical cancer is also linked to smoking, estrogen, and other infections. Hepatitis virus infections can synergize with aflatoxins to promote HCC (92,93). Aflatoxins are potent liver carcinogens produced by Aspergillus infection of crops such as corn and peanuts and are ingested when contaminated foods are consumed (94).

Genetic and Evolutionary Considerations

The genetics of an infected individual can influence the risk of developing some microbe-associated cancers. Certain ABO blood groups affect the likelihood of H. pylori infection and gastric cancer, possibly by affecting H. pylori attachment to the gastric epithelium, and human leukocyte antigen (HLA) genotype affects the risk of nasopharyngeal cancer induced by EBV (16,95,96). In addition, genome-wide association studies have identified germline sequence variants that are associated with increased risk of developing cancer in response to infection. For example, distinct genetic loci have been identified as increasing the likelihood of developing viral-associated HCC (97,98). Although common germline genetic variants can modulate the ability of dendritic cells to sense pathogens (99), in most cases, the mechanistic basis for these associations is unknown. It should also be pointed out that genetic associations may reflect the action of a gene linked to the studied polymorphism.

Host-microbe interactions relevant to cancer can also act on a population scale over evolutionary time. A large fraction of our genome is derived from ancient viruses and related genetic elements, such as endogenous retroviruses and retrotransposons, and infections have sculpted our genomes for millions of years. The higher incidence of cancer in immunosuppressed people and patients with certain immunodeficiency syndromes including AIDS also suggests that immune defenses provide surveillance against cancer (100). Conversely, the proinflammatory response to infections may contribute to cancer risk by increasing mutagenesis.

In many cases, our cells and microorganisms are locked in an ongoing evolutionary arms race, where mutations in the pathogen are countered by mutations in the host, and vice versa, with these mutations becoming fixed in the respective genomes (101). Cellular mutations can affect cancer predisposition or susceptibility to or control of infection, whereas mutations in the microbe can affect its replication, transmissibility, cellular tropism, ability to evade immune surveillance, or carcinogenicity. For example, many human cancers exhibit a pattern of mutations consistent with the action of antiviral apolipoprotein B mRNA editing enzyme, catalytic polypeptide-like (APOBEC) cytidine deaminases, suggesting that this antiviral mechanism can confer increased cancer risk (102,103).

Some infectious diseases, including smallpox, malaria, and plague, have imposed intense selective pressures on human populations for millennia, resulting in genetic bottlenecks that may influence susceptibility to infection by carcinogenic microbes, cell and tissue response to stress, emergence of cancer, and response to treatment (87). For example, the severe epidemics caused by plague (Yersinia pestis) appear to have driven evolution of the chemokine receptor CCR5, which mediates transmission of HIV, and toll-like receptors, which defend against infections by many microorganisms (104,105).

Implications for Cancer Prevention and Therapy

The panoply of antimicrobial and public health measures used to combat infectious disease can be marshalled to prevent or treat cancer caused by microbes. A large fraction of the cancers listed in Table  1 are preventable or treatable by currently available antibiotics, vaccines, and specific antiviral and antiparasitic agents. In addition, the blood supply in much of the world has been cleansed of HBV, HCV, and HIV, and public health measures such as needle exchange programs and encouragement of safe sex practices have reduced transmission of some tumor viruses and HIV.

More than half of all infants worldwide are vaccinated against HBV, reducing the incidence of HCC (Figure  6, A) (106,108,109). HPV vaccination reduces the risk of cervical cancer, generating high confidence that cervical carcinoma rates will eventually plummet in vaccinated populations (110). Antiviral agents can also prevent and treat cancer. ART prevents HIV transmission and, as noted above, reduces cancer incidence. In addition, HIV treatment can cause resolution of Kaposi sarcoma lesions, presumably by enhancing the immune response to tumor antigens expressed by sarcoma cells (111,112), and anti-herpesvirus drugs can inhibit KSHV-mediated tumorigenesis (113). Direct-acting antiviral drugs can cure HCV infections in a large fraction of patients. Sustained virologic response to anti-HCV agents reduces the incidence of HCC by three- to fourfold, with better protection if HCV cure precedes the onset of cirrhosis (Figure  6, B) (107,114,115). It may also be possible to design novel therapeutic approaches that exploit infection-induced reprogramming of cellular metabolic or signaling pathways.

Figure 6.

Prevention of hepatocellular cancer by vaccination and antiviral treatments. A) Hepatitis B virus (HBV) vaccination reduces the incidence of hepatocellular carcinoma. Graph shows reduction in the incidence of hepatocellular carcinoma (HCC) in individuals vaccinated against HBV compared with unvaccinated (control) individuals (106). This figure is modified from Qu C, Chen T, Fan C, Zhan Q, Wang Y, et al. Efficacy of neonatal HBV vaccination on liver cancer and other liver diseases over 30-year follow-up of the Qidong Hepatitis B intervention study: a cluster randomized controlled trial. PLoS Med. 2014;11(12): e1001774. 10.1371/journal.pmed.1001774. B) Direct acting antiviral drugs that cure hepatitis C virus (HCV) infection reduce the incidence of hepatocellular carcinoma. Graph shows reduction in the incidence of HCC in treated individuals displaying a sustained anti-HCV response to antiviral therapy (SVR) compared with treated individuals who did not display SVR. This figure is modified from Ioannou et al. (107), with permission from Elsevier.

Viral proteins expressed in cancer cells can serve as targets for immune-based therapeutic approaches. Adoptive immunotherapy with EBV-specific donor T cells can control EBV-related lymphomas in bone marrow transplant recipients (116), and expression of viral tumor antigens may contribute to the efficacy of immune checkpoint inhibitors against cervical cancer, HPV-positive oropharyngeal cancer, and virus-associated HCC (117,118). Viral proteins will also presumably be good antigens for therapeutic vaccination to generate cytotoxic T cells that specifically kill cancer cells. Indeed, a recent small clinical trial showed that therapeutic vaccination with HPV E6 and E7 peptides cooperate with chemotherapy in treating cervical cancer (119), and it is possible to generate chimeric T-cell receptors that target viral proteins in cancer cells (120). In addition, the oncogene addiction of some virally induced tumors suggests that inhibition of the expression or activity of viral oncogene products may provide novel cancer treatment opportunities (121).

Most H. pylori infections can be cured by a short treatment with antibiotics, usually in combination with medicines that inhibit gastric acid secretion, lowering the risk of gastric cancer (122). However, development of antibiotic resistance is likely to prevent the eradication of H. pylori–associated cancers at the population level, and spread of antibiotic resistance could impair our ability to treat other serious bacterial diseases.

The human microbiome also presents therapeutic opportunities and challenges. The gut microbiome synthesizes fatty acids, which can affect the likelihood of chronic graft-vs-host disease following allogeneic stem cell transplant (123). Bacteria can also metabolize important chemotherapeutic drugs, either attenuating or enhancing their activity (124,125), and the person-to-person variability in the response to cancer treatment may be due, in part, to the variability in the composition of the human microbiome (46). Thus, profiling the composition of the microbiome in cancer patients is likely to inform the optimal use of cancer therapeutics. In addition, high bacterial diversity and the presence of particular bacterial species in the gut microbiome is associated with improved response to immunotherapy or success of allogeneic stem cell transplantation (45,126–128), and the human microbiome can modulate the risk of immune-related adverse events during immunotherapy (129,130). The fungal mycobiome can also influence treatment response (131).

In the future, it may be possible to use selective antibiotics, fecal transplantation, prebiotics, or next-generation probiotics to modulate the composition of the microbiome to reduce cancer risk in susceptible individuals or to improve response to therapy [eg (132)]. Early experience suggests that these opportunities may arise soon. For example, fecal microbiota transplantation from donor patients who responded to checkpoint blockade immunotherapy can increase the effectiveness of immunotherapy against melanoma and other cancers in previously refractory patients [eg (133–136)]. However, a careful balance must be struck between targeted antibacterial treatments and disruption of the gut microbiome, lest these treatments have untoward health effects (137,138).

Microorganisms as Anticancer Therapeutics

Oncolytic Agents

There has been considerable activity identifying or constructing viruses that preferentially kill cancer cells by exploiting biochemical differences in tumor cells compared with normal cells. In addition to direct cytotoxic effects of the virus, oncolytic viruses might facilitate immune-mediated cancer cell killing by causing the release of tumor antigens or by modulating the tumor microenvironment into one that is more conducive to immune attack (139). Consistent with this idea, oncolytic viruses are more effective in immune-competent animals compared with immunosuppressed ones (140). Oncolytic viruses can also be armed for greater activity by inserting genes encoding toxic or immunostimulatory products, and their capsids can be modified for better tumor cell entry or selectivity. A theoretical advantage of oncolytic viral vectors is their potential to replicate and spread within the host to additional tumor cells at neighboring or distant sites, amplifying the therapeutic effect and possibly targeting metastatic lesions. Candidate oncolytic viruses have been derived from diverse families of viruses, and many are in clinical trials or in clinical use (141), but a potential limitation to oncolytic therapy is preexisting immunity to the oncolytic agent.

Attenuated bacteria are also being explored as possible antitumor agents (142). These bacteria might elicit a beneficial cell- or cytokine-mediated immune response or produce proteins, metabolites, or natural products with anticancer activity. In fact, the first widely used immunotherapy was the Bacillus Calmette-Guérin vaccine, which is standard therapy for early stage bladder cancer by mobilizing T cells to the tumor (143–145). We also note that oncolytic agents can be combined with other treatment modalities. For example, combining an oncolytic herpes simplex virus with immune checkpoint inhibitors showed promising results in clinical trials for melanoma (146,147).

Viral Vectors and Vaccines

Gene expression vectors based on viruses can be used as vaccines to express recombinant antigens including viral oncogene products and proteins preferentially expressed in cancer cells to elicit an immune response that attacks tumor cells or enhances response to therapy. It may also be possible to vaccinate against neo-epitopes, new antigenic determinants that are generated by somatic mutations in cancer cells. Viral vectors can also be used to express specific antibodies with broad or high-affinity reactivity against proteins expressed by tumor or immune cells. There is a recent report that intratumoral injection of influenza virus vaccine can convert immunologically cold tumors into ones that respond to immune checkpoint blockade (148).

Viral vectors also are being used to transduce chimeric antigen receptors into T-lymphocytes to generate cytotoxic T cells with heightened activity toward tumors. Such CAR-T cell approaches are effective with some leukemias and lymphomas and are being tested against solid tumors (149,150). Viral vectors could similarly be used to transduce tumor cells with genes encoding proteins with antitumor activity.

Microbial Products

Finally, we note that microbes are a rich source of molecules used to treat cancer. For example, the chemotherapeutic drugs doxorubicin and bleomycin are synthesized by Streptomyces species, and fungal extracts are being systematically evaluated for additional compounds with anticancer activity (151).

Cancer Microbiology as a Global Imperative

It is estimated that 75% of the world’s cancer burden will lie in low- and middle-income countries (LMICs) by 2040 (152,153), with a preponderance of infection-associated cancers being particularly skewed toward that part of the world (Figure  1). Many factors may contribute to this disparity including regional differences in circulating pathogens, host genetic susceptibilities, and access to prevention and treatment. The high rates of HIV infection in sub-Saharan Africa and of HBV and H. pylori infection in Asia are particularly striking. Similarly, other infections that contribute to carcinogenesis are more common in LMICs, such as malaria, which cooperates with EBV infection to cause Burkitt lymphoma. In other cases, the inadvertent spread of HCV and HIV in Egypt and Romania, respectively, was caused by contaminated medical equipment (154,155). Certain microbial strains with increased virulence or transmissibility may show geographic preferences. For example, non-European strains of oncogenic HPV16 appear to confer higher risk of cervical cancer than European strains (156). In addition to the genetics of the microbe, host genetics or environmental factors may make certain population groups more susceptible to infection-associated cancers than others.

In addition to these biological differences, cancer-related outcomes in LMICs are driven in large part by environmental and behavioral risk factors, inadequate preventative and early detection initiatives, and poor access to vaccines and treatment. For example, higher-income countries but not most LMICs have successfully conducted pap smear cervical cancer screening programs, resulting in the identification and treatment of cervical precancerous lesions and a dramatic reduction in the incidence of this cancer (157), and as of 2015, only 3% of low-income countries have introduced the HPV vaccine (158). Finally, poor nutrition and co-infections may impair immune status and defense against infections.

Future Outlook: Cancer Microbiology as an Emerging Field

There are pervasive connections between microbiology and cancer (Figure  2). Some microbes have direct carcinogenic potential, HIV increases risk for many cancer types, and the microbiome impacts cancer formation and treatment. There are likely to be additional microbes that directly or indirectly affect cancer risk or the response to therapy. Cancers caused by microorganisms are uniquely susceptible to prevention and treatment efforts, and microorganisms themselves can be used to prevent and treat cancer. Although not discussed here, cancer and its treatment often put patients at higher risk for many infectious diseases. The convergence of these factors warrants the recognition of a broad new field of inquiry, cancer microbiology. Although there is much research ongoing independently in cancer and microbiology, the explicit recognition of cancer microbiology as a field encompassing these topics will have critical implications for training, research, funding, public health, clinical care, and outreach.

The microbial universe influences oncogenesis and cancer treatment in many ways, not merely by acting as direct carcinogens. Stress responses to infection, such as the induction or suppression of a DNA damage response, an innate or adaptive immune response, inflammation, or the unfolded protein response, could affect cancer development or response to treatment. Thus, as was the case for the discovery of oncogenes, the value of continued studies of infectious agents and cancer will not be restricted to cancer caused by microorganisms but will have general importance for many types of cancer. Similarly, concepts developed to study and treat infections may be applicable to cancer. As one example, the phenomenon of immune exhaustion and checkpoint controls, much of it first studied in chronic viral infection (159,160), has emerged as an important factor in tumor immunotherapy.

Students should be educated to appreciate the connections between microbiology and cancer. Graduate programs or courses that highlight these connections will help eliminate the silos between microbiology, infectious disease, cancer research, and oncology and help establish cancer microbiology as a field. Funding agencies should encourage and support research in cancer microbiology. Despite the importance of microorganisms in human carcinogenesis, most contemporary cancer research focuses on studies of the occurrence and consequences of somatic mutations in cancer cell DNA, whereas most infectious disease funding focuses on acute infections. Cancer funding agencies should take care to support research on replication of tumor viruses and HIV and the development of drugs and vaccines that inhibit infectious carcinogens, because as is the case for noninfectious carcinogens, understanding the mechanisms of infectious carcinogenesis and developing effective countermeasures has important implications for cancer control. However, despite the value of this holistic view of cancer and microbiology, several comprehensive cancer centers have recently disbanded their research programs that focused on infection and cancer. At Yale, this decision was spurred in part because reviewers questioned the cancer relevance of studies on tumor virus and HIV replication! Joint workshops and funding programs between the National Cancer Institute and the National Institute for Allergy and Infectious Diseases will help galvanize studies at this crucial interface. Because the great majority of infection-associated cancers and cancer deaths occur in LMICs, cancer microbiology is closely tied to global oncology, another emerging field. Given the epidemiologic overlap between the unmet needs and research priorities of cancer microbiology and global oncology, synergistic approaches in these fields should be emphasized, and programs to teach and support cancer microbiology should include a substantial global oncology component and perspective.

Ultimately, much infectious disease control is cancer control. The most important laboratory-based advances in cancer prevention in recent decades have been approaches that reduce the incidence of infection-associated cancers or precancerous changes: vaccines that prevent HBV or HPV infection and their resulting cancers, drugs that cure HCV, antibiotics that cure H. pylori infection, and drugs that suppress HIV and herpesvirus replication. Widespread use of these and future vaccines and treatments will substantially reduce the worldwide cancer burden and may lead to the eradication of cervical cancer. Public health scientists and practitioners should be educated in microbiology and cancer so that they can determine how best to implement these measures.

Academia and the pharmaceutical and biotechnology industries should vigorously develop new molecules and strategies to prevent or treat infectious disease with the awareness that some may have utility against cancer. There are practical issues that facilitate the discovery of antimicrobial agents that prevent and treat cancer. In contrast to most cancer treatments, the effectiveness of antimicrobial agents against their primary targets is relatively easy to assess in short-term experiments. Furthermore, because infectious agents are distinct from their host cells, medicines including immune therapeutics that target microbial proteins may have fewer serious side effects than those that target cellular components. Finally, we note that the ability of an antimicrobial intervention to reduce the incidence of specific cancers might identify instances of hit-and-run carcinogenesis.

Recognition of cancer microbiology as a field will also spur new outreach and public education efforts. Many people are still surprised to learn that infections can cause cancer. Clearer understanding and explanation of the links between microbes and cancer will help individuals accept vaccination and other efforts to limit pathogen exposure to prevent cancer.

Cancer microbiology has emerged from the historically productive areas of tumor virology and bacteriology. Microbiologists, infectious disease specialists, oncologists, epidemiologists, and other scientists and clinicians should embrace this evolution and exploit this new expansive perspective to generate novel insights into carcinogenesis and translate this new understanding to the benefit of humanity.

Funding

Cancer Microbiology research in the authors’ laboratories is supported by the National Institutes of Health to DD (R35 CA242462 and R01 AI150897); the National Institutes of Health (P30CA016359, R01CA206483 and P50CA196530) and a Team Challenge Award from Yale Cancer Center to BE; a Team Challenge Award from Yale Cancer Center (to ALG); and the National Institutes of Health to AJ (R01 CA206465, U24 AA020794, U01 AA020790, and U10 AA013566).

Notes

Role of the funders: The funders had no role in data collection and analysis, decision to publish, or preparation of the manuscript.

Disclosures: The authors declare no conflicts of interest.

Author contributions: Conceptualization: DD, BE, ALG, WM, AJ; Investigation: DD, BE, ALG, WM, AJ; Writing—original draft: DD; Writing—review & editing: DD, BE, ALG, WM, AJ.

Acknowledgments: We thank Caroline Hendry for valuable suggestions, Jan Zulkeski for assistance in preparing this manuscript, and Aaron Kantor for original artwork.