The microbiota and microbiome in pancreatic cancer: more influential than expected

P. gingivalis is a pathogenic bacterium that responsible for chronic periodontitis. Through collecting prediagnosis oral wash samples from participants, Fan et al. [43] characterized the composition of the oral microbiota in samples from 361 people who developed PDAC and 371 healthy participants matched by age, sex, race, body mass index, smoking status, alcohol use and diabetes. This study suggested that P. gingivalis was correlated with a 59% greater risk of developing PDAC and played a role in its etiology. Higher levels of ATTC 53978 antibodies against P. gingivalis in blood (> 200 ng/mL) were found in 405 patients with PDAC than in healthy volunteers and were related with a 2-fold higher risk of developing PDAC based on a large European cohort [28], which suggests that ATTC 53978 may serve as the best indicator for aggressive periodontal disease and PDAC susceptibility. The above conclusions are consistent with extensive evidence such as the NHANES I [24] and III data [27] and the Health Professionals Follow-up Study [25]. Thus, P. gingivalis was the periodontal pathogen most strongly associated with an elevated risk of PDAC.

P. gingivalis, Treponema denticola, and Tannerella forsythia, commonly known as the red complex, are the major periodontitis-causing pathogens; they secrete peptidyl-arginine deiminase (PAD) enzymes and have been extensively studied in patients with PDAC. High mutation rates in the tumor suppressor gene p53 and oncogene K-ras, particularly arginine mutations, known as p53 Arg72Pro [44, 45] and K-ras codon 12 arginine mutations [46‐49], respectively, indicates poor prognosis of PDAC patients. These oral bacteria that produce PAD enzymes are capable of degrading arginine, which may result in p53 and K-ras point mutations [50]. Scientists have hypothesized that P. gingivalis plays a critical role in initiating inflammation backed by the evidence that individuals with periodontal disease manifest as elevated markers of systemic inflammation and escaping the immune response related to lipopolysaccharide (LPS) and toll-like receptors (TLRs) [51‐53].

Strains of Fusobacterium is an anaerobic, gram-negative oral bacterium, which has been verified that could cause periodontal diseases and should always be treated as a pathogen [54]. However, some studies have drawn the opposite conclusions. According to a case-control study [43] and a large prospective cohort study [28], Fusobacteria was associated with reduced PDAC risk. Recent studies revealed that Fusobacterium potentiates tumorigenesis [55] and promotes chemoresistance in colorectal cancer [56]. More specifically, Fusobacterium could increase production of reactive oxygen species (ROS) and inflammatory cytokines, and modulate the tumor immune microenvironment and drive myeloid cell infiltration in intestinal tumors. Despite these conflicting results, the paradoxical effects observed in different tumors may provide evidence to unravel the potential mechanisms.

N. elongata and S. mitis were found to be lower in saliva specimens collected after PDAC diagnosis than in controls in both the test cohort (10 pairs) and validation dataset (28 pairs) of a retrospective study [36]. This finding was partly supported by Michaud et al. [28], who showed an inverse association of S. mitis antibodies with pancreatic cancer (but did not measure N. elongata). Farrel et al. [36] found that the combination of N. elongata and S. mitis biomarkers yielded 96.4% sensitivity and 82.1% specificity in distinguishing patients with PDAC from healthy subjects.

The levels of the genera Corynebacterium and Aggregatibacter are present in lower concentrations in PDAC patients than healthy population, while Bacteroides and Granulicatella adiacens are more frequent in PDAC salivary RNA samples [36]. However, another study indicated that Aggregatibacter actinomycetemcomitans was linked with a higher risk of PDAC [43]. In addition, Leptotrichia is considered a protective microbe that dose-dependently decreases the risk of PDAC. Furthermore, a higher ratio of Leptotrichia to Porphyromonas in saliva can be observed in PDAC patients [57].

Researchers are only beginning to explore how H. pylori as an engender of PDAC and its role in manipulating the host immune response. Most studies to date, including case-control studies [62‐64], prospective cohort studies [65, 66], and meta-analyses [67‐69], have confirmed that H. pylori infection is related with increased PDAC risk. However, some studies have found no relationship between the two [70‐73], and several studies have even drawn the opposite conclusion [74, 75]. One of the difficulties for untangling these inconsistent and paradoxical associations is how to exclude confounding factors.

Using the H. pylori IgG antibody level in blood serum from PDAC patients and healthy controls, scientists found that the H. pylori IgG level was higher in PDAC. H. pylori strains that express Cytotoxin-associated gene A (Cag-A) is associated with gastric inflammation and ulceration and promotes malignant transformation in gastric cancer [76, 77]. Case studies of H. pylori antibody-positive in PDAC revealed complicated results related to Cag-A status, and we believe that H. pylori and Cag-A predominance in PDAC microbiota studies. One study [78] proposed that factors including ABO blood type may also participate in this intricate process, while a recent meta-analysis revealed a modestly significant increased risk of Cag-A-negative H. pylori strain [67‐69, 73] with positively correlated factors, including non-O blood type [64, 78] and smoking status [63, 65] in PDAC.

H. pylori causes gastric lesions via directly impairing the gastric mucosa, and its DNA can be detected in infected antrum and corpus stomach tissues. However, the expression of H. pylori DNA cannot be detected in pancreatic juice or tissues by PCR in chronic pancreatitis and PDAC [79], which may suggest that H. pylori cannot trigger pancreatic carcinogenesis directly. Possible indirect mechanisms include inflammation and immune escape. Exposure to carcinogenic nitrosamines is also an underlying mechanism [78]. More specifically, nitrosamine levels are lower in patients with duodenal ulcers than gastric ulcers that characterized by low acidity, which may explain the positive association of PDAC risk with gastric ulcers but not duodenal ulcers [66, 80].

HBV and HCV are hepatotropic viruses that lead to hepatitis and hepatocellular carcinoma (HCC). However, HBV/HCV infection is not restricted to the liver; these viruses can be detected in extrahepatic tissues, including the pancreas [81‐87], which may play a role in the carcinogenesis or development of extrahepatic malignancies [88‐91], including PDAC [85, 86, 89, 91, 92]. Specifically, investigators detected HBsAg and HBcAg in the cytoplasm of pancreatic acinar cells [82], and individuals with chronic HBV infection are accompanied by elevated serum and urinary levels of pancreatic enzymes partly [93, 94]. HBsAg was also observed in pancreatic juice among patients with HBV infection [95] and was associated with the development of chronic pancreatitis [81‐83, 85], which suggests that HBV-related pancreatitis might be a precursor of PDAC [85]. Clinical observations concerning impaired pancreatic exocrine function in patients with chronic HBV infection support this hypothesis. The positive correlation of PDAC risk with HBV infection, especially for long-lasting persistent infection, chronic/inactive HBsAg carriers and occult infection is supported by unanimous conclusions drawn from meta-analyses [96‐104]. Most prominently, HBV is capable of replicating apart from infecting in the tumor and nontumorous pancreatic tissues of PDAC patients [85]. According to the REVEAL-HBV study [105], the association between HBV and PDAC was found in patients with higher viral DNA loads (HBV DNA > 300 copies/mL). The integration of HBV DNA to pancreatic tissue and pancreatic metastases to the liver in patients with HBV infection have been confirmed [83]. Wei et al. [106] discovered that HBV infection in PDAC patients increased the rate of synchronous liver metastasis and this kind of status could be regarded as an independent prognostic factor. However, some studies have drawn conflicting conclusions concerning the association between the presence of HBV, HCV and PDAC [105, 107, 108].

The pancreas and liver share similar features in early embryological and fetal growth and have some common regulatory pathways in PDAC and HCC [109]. Possible mechanisms of HBV or HCV contribution to carcinogenesis include inducing inflammation [85] and modifying tissue viscoelasticity [110], DNA integration in infected cells that delays host immune system clearance of HBV/HCV-containing cells [96], modulating the PI3K/AKT signaling pathway via the HBV X protein (HBX) [86]. Another key finding is that exposure to HBV significantly increased PDAC risk when concomitant with diabetes [85, 96, 97], high frequency of HBV/HCV infection was reported in patients with diabetes [111], and this connection was supported by strong expression of HBsAg and HBcAg in islet cells at histological level [85]. However, pancreatic cancer cells have low levels of HBV replication, so molecular proofs about the potential role of HBV in PDAC remain limited [85, 87]. Nonetheless, available literature supports the potential etiologic and oncogenic role of HBV infection in PDAC. Should such findings be confirmed, they may bring new insights into the etiology and therapeutics of PDAC, and remind clinicians to prevent the reactivation of HBV during chemotherapy in patients with HBV infection.

Bile is a sterile hepatobiliary solution rich in lipid, and microbial colonization in the bile fluid is defined as bactibilia. In a study based on PDAC patients, Maekawa et al. [112] investigated the presence of bacteria in bile samples via genetic sequence analysis, and the result suggested that Enterobacter and Enterococcus spp. were the major microbes. Antibody levels against Enterococcus faecalis capsular polysaccharide (CPS) were increased in serum of PDAC and chronic pancreatitis patients compared with normal subjects, which may indicate that infection with E. faecalis is involved in the progression of pancreatitis-associated PDAC. Escherichia coli (E. coli) is the well-known gut microbe, though outnumbered in the gut approximately one thousand to one by other species. Serra et al. found that pancreas head carcinoma (PHC) is strongly and positively correlated with bactibilia while E. coli and Pseudomonas spp. were the most common microorganisms and were negatively correlated with PHC [113]. Transfusion-transmitted virus (TTV) is considered hepatotropic and a possible cause of acute hepatitis. However, it has also been identified in the pancreas. Clinicians reported a case with both pancreatic cancer and TTV infection [114], which indicates a need for further research.

Some cancer treatments rely on activating the immune system via the gut microbiota [129, 130]. The efficacy of therapies, including alkylating agents, immune checkpoint blockers and adoptive T-cell transfer (ACT), depends on immunity closely related to gut microbiota [131]. Zitvogel [132, 133] found that the chemotherapy drug cyclophosphamide damages the intestinal mucus layer, allowing some gut bacteria enter the lymph nodes and spleen, then specific immune cells being activated. In addition, cyclophosphamide lost its anticancer effects in mice when raised without microbes in guts or given antibiotics. Another study showed similar finding with oxaliplatin and cisplatin and found that the scarce of gut microbes compromises the efficacy of CpG- and anti-IL-10-based antitumor response due to ineffective priming of tumor-infiltrating myeloid cells and a consequent lack of ROS-dependent apoptosis and TNF-dependent necrosis [129].

Scientists are examining how the gut microbiota interplay with immunotherapies response and how these interactions are administrated. Following the observation that only 20~40% of patients respond to immunotherapy, Zitvogel made further efforts to explore whether gut bacteria might influence the response to anti-CTLA-4 and anti-PD-1. Their work found that microbe-free mice failed to respond to one such drug, and mouse response improved when given Bacteroides fragilis [134]. Sivan et al. [135] reported that Bifidobacterium increased cancer immunotherapy response in mice, which suggested that microbes, especially gut bacteria, might be activating the immune response by stimulating enterocytes to generate certain message molecules or provide signals to immune cells that help to supercharge their tumor-fighting efforts. Similar latest studies [136, 137] have linked favorable immunotherapy responses in melanoma patients to specific gut microbes.

Routy and colleagues [138] reported that changes in the gut microbiota composition following antibiotic usage decreased responses to immunotherapy in patients with lung, bladder, and kidney cancer. In contrast, when using antibiotics specific to certain gut bacteria in an HCC mouse model, the proportion of NK T cells in the liver increased, leading to tumor shrinkage [15]. In PDAC, the combination of antibiotics and PD-1 blockade showed that a synergistic antitumoral effect associated with T-cell activation. These confounding results suggest that each cancer type may induce a distinct profile of alterations in gut and tumor microbiota composition that may either attenuate or facilitate the function of immune checkpoint inhibitors.

The protumorigenic effect of the gut microbiota appears linked with its capacity to exert influence on immune response through the tumor microenvironment (TME), making it more tolerant toward cancer. Immune tolerance mechanisms have been implicated as the main barrier to effective antitumor immunotherapy. Pushalkar [60] showed that specific microbes in the gut and intrapancreatic serve as helpers in the establishment of immunosuppressive PDAC tumor microenvironment in spontaneous murine models, enhancing cancer progression and resistance to immunotherapies. Bacterial ablation in models showed antitumor effects and can be reversed or abrogated by the transfer of feces from PDAC-bearing KPC mice, but no difference was found when the transfer came from non-PDAC controls. These experimental data thus provide compelling preclinical evidence for gut microbiota modulation, and the tumor microbiota could sensitize the immune-refractory cancer and convert it into a more responsive one. Because bacterial ablation upregulated PD-1 expression, a clinical trial of antibiotic synergism with checkpoint-based immunotherapy is beginning, using antibiotics and pembrolizumab prior to resection in patients with locally advanced PDAC. These ideas have started to spread and create new possibilities for clear therapeutic applications of microbiota science.

Mechanistically, microorganisms in PDAC differentially activate selective TLRs in monocytes and then produce immune tolerance. TLRs, representing the most acknowledged family of pattern-recognition receptors (PRRs), are a group of pathogen-associated molecular pattern receptors and undertake a certain role in immune response to microbial infection and accelerate tumorigenesis via innate and adaptive immune suppression in PDAC. PRRs reside in most immune cells and can bind a range of microbe-associated molecular patterns (MAMPs, such as LPS), as well as byproducts of dead cells and sterile inflammation called DAMPs (damage-associated molecular patterns) [139]. After binding, the TLRs-DAMPs complex recruit MyD88 or TRIF adaptor molecules as signal transducers to activate signal pathways such as NF-κB and MAPK. The procarcinogenic effects of TLRs can be reversed by inhibiting NF-κB or MAPK pathway [140]. Miller et al. found that TLR4 and TLR7 showed up-regulated expression in the PDAC microenvironment [140, 141], and TLR signaling, such as TLR4/MyD88 [140, 142, 143], plays an important role in pancreatic tumors. Further, animal studies demonstrated that the activation of TLRs could provoke pancreatitis and synergize with K-ras to significantly promotes pancreatic carcinogenesis [140, 141, 144‐146].

Gut microbiota ablation with antibiotics does affect the immune phenotype in the TME, that is, immunogenic reprogramming or reshaping, and then suppress tumor growth by inducing antitumorigenic T-cell activation, boosting immune surveillance and improving sensitivity to immunotherapy in malignances [147]. The immunogenic reprogramming of TME in a murine model manifests a decrease in myeloid-derived suppressor cells (MDSCs), an increase in M1 macrophage polarization, facilitating the T helper 1 (Th1) differentiation of CD4+ T cells and the activation of CD8+ T cells. All changes favor antitumor efficacy. A balance exists between pro- and antitumor T cells in tumor microenvironment. The Th1-type cytokine interferon-γ exhibits antitumorigenic effects, whereas the T helper 2 (Th2)-type cytokines interleukin (IL) 4, IL5, and IL10 and Th17 cells play a protumorigenic role. Specifically, after gut microbiota depletion in a pancreatic cancer murine model, interferon-γ producing T cells Th1 manifest with a significant increase and a corresponding decrease in IL17a and IL10-producing T cells [61]. This result is in accordance with the previous conclusion, and a high Th1/Th2 ratio in TME relates with improved survival in PDAC patients [148].

The immunologic mechanisms that underlie variable responses to systemic therapy upon changes in the microbiota were elucidated in other studies. Certain miRNAs are key regulators of the innate and adaptive immune response [52, 149, 150], and these immune-related miRNAs are secreted by cells and then travel to the pancreatic tissue to alter gene expression. They can modulate host responses to pathogens, and vice versa, pathogens also regulate miRNA expression.