Association of the gut microbiota with colorectal cancer in a South Asian cohort of patients

Background As the gut microbiome is thought to play a role in the pathogenesis of colorectal carcinoma (CRC) and affected by the diet and the genetic composition, we sought to investigate the patterns of gut microbiota that associate with CRC in a South Asian cohort of patients with CRC. Methodology The relative abundance of 45 types of gut microbial species were determined in faecal samples of CRC patients (n=24), DM (n=20) and healthy age matched controls (n=44), using a PCR array. Data was analyzed using the specific software for analysis of bacterial DNA quantification. Results The species Bacteroides fragilis (23.9-fold), Bacteroides thetaiotaomicron (8-fold) and Akkermansia muciniphila (5.9 fold) were several-fold over expressed in patients with CRC compared to healthy individuals, whereas bacterial species of the Phylum Proteobactria were under expressed. There was no difference in the abundance of these 3 species of bacteria with tumour stage or gender and age of patients. Aeromonas species, Enterococcus faecium and Shigella dysenteriae (Proteobacteria) were over 100-fold over abundant in those with DM compared to healthy individuals. Although 70.83% of those with CRC also had diabetes, the relative abundance of microbiota in CRC patients were different to those who had diabetes and no CRC. Conclusions Patients with CRC and DM harbor a markedly different gut microbiota patterns compared to their healthy counterparts. Similar patterns of gut microbial dysbiosis that associate with CRC and DM appear be seen in South Asian populations, compared to Western countries, despite differences in the diet and ethnicity.

demographics [2]. As this increase in the incidence of CRC in developing countries is likely to 48 result in a huge burden to their economies, there is an urgent need to implement programs that 49 reduce its occurrence and adopt novel diagnostic and treatment methods of CRC.  [3][4][5][6]. Metabolic diseases and CRC are associated with microbial dysbiosis, which is characterized 54 reduced diversity of the gut microbiome with an overabundance of the genera Proteobacteria and 55 Firmicutes [7]. Several types of microbiota have been shown to associate with CRC such as pks-56 positive E.coli, enterotoxigenic Bacteriodes fragilis, Fusobacterium nucleatum and Streptococcus 57 gallolyticus [8][9][10][11][12]. While some of these microbes were overabundant in the gut microbiome of 58 patients with CRC, some have been detected specifically in tumor tissue and also in distance 59 metastasis, suggesting that they may play a role in the pathogenesis of this cancer [12]. They are 60 thought predispose to the development of CRC by inducing epigenetic changes and thereby 61 affecting gene transcription, inducing DNA damage and reactive oxygen species and by inducing 62 procarcinogenic cytokines [12]. 4

63
Of the factors that affect the diversity of the gut microbiome, the diet plays a central role. Although 64 the relative abundance of gut microbiota depends on an individual's genetic composition (12%), 65 the influence of the diet is much greater (57%) [13]. Individuals who predominantly consume a 66 Mediterranean diet, rich in grains, legumes, nuts, vegetable and fruits were found to have a gut 67 microbiome which reduced the risk of metabolic diseases, inflammatory bowel disease and colonic 68 cancer compared to those who consume a typical Western diet [14]. South Asian individuals have 69 a very different diet than those of Western and the South East Asian populations, due to differences 70 in religious and cultural practices. Their diets are typically rich in grains, pulses, vegetables and 71 fruits with a low intake of red meat. These vast differences in the diet are likely to influence the 72 microbial composition and thus protect or predispose to the development of CRC. In fact, is has 73 been shown that the gut microbiome is significantly different in individuals of different ethnicity, 74 living in the same geographical area [15]. Since dietary factors are likely to directly contribute to 75 the microbial composition and thus to the risk of developing CRC, change in the dietary patterns 76 can be an important strategy in the prevention and treatment of CRC [16]. Therefore, in order to 77 implement such preventive and therapeutic strategies, it would be important to initially 78 characterize the gut microbial patterns in South Asian individuals with CRC living in those 79 countries.

81
In this study, we have determined the relative abundance of 45 species of gut microbiota on 82 patients with CRC, aged matched healthy individuals and also in patients with metabolic diseases 83 such as diabetes. We found that that pattern of the gut microbiota was significantly different in 84 those with CRC and diabetes compared to healthy individuals.   96 In order to compare the changes that associate with CRC, we recruited healthy individuals (n=44) 97 who underwent colonoscopy and were found not to have any bowel pathology and who were non-98 obese, (BMI <23.9), waist circumference <80cm for females and <90cm for males and who did 99 not have diabetes or hyperlipidemia. Again, the stool samples were obtained from these patients 100 two weeks after colonoscopy.

102
In addition to the above controls, as most of the patients with CRC also had diabetes mellitus 103 (DM), in order to differentiate the changes in the patterns of gut microbiota with those that are 104 specific to CRC, we also recruited patients with DM (n= 20) who underwent colonoscopy and 105 were found not to have any CRC or gut pathology, between January 2017 to April 2018. As in 106 patients with CRC, stool samples were obtained from these individuals, two weeks following 107 colonoscopy. Patients who had adenomas, who received antibiotic therapy for more than 1 week 6 108 prior to stool sample collection, who received chemotherapy and/or radiation or with a history of 109 CRC and inflammatory bowel disease were excluded.        We observed marked differences in the gut microbiota patterns in patients with CRC when 197 compared to healthy individual groups (Fig 1).

Changes in the patterns of gut microbiota in patients with diabetes mellitus when compared to healthy individuals
The relative abundance of the gut microbiota has shown to be different in those with metabolic diseases such as DM [7,18]. As 17/24 (70.83%) patients with CRC also had DM, we proceed to determine if these changes in the gut microbiota observed in those with CRC were associated with CRC or with DM. In order to determine the changes associated with DM, we compared the patterns of gut microbiota in patients with DM (n= 20) with the group of healthy volunteers (n=44). Again.
we observed marked differences in the gut microbiota pattern in patients with DM compared to healthy individuals (Fig 2).

Comparison of the gut microbial patterns of CRC patients compared to those with DM
As we observed marked differences the relative abundance of the gut microbiota between those with CRC and healthy individuals and those with DM and healthy individuals, and since 17/24 (70.83%) of the individuals with CRC had DM, we proceeded to analyze the gut microbial patterns specific to CRC by comparing the microbial patterns of those with CRC (n=24) with those with DM (n=20).
Although all 45 bacterial species were overabundant in patients with DM compared to those with CRC, of whom 70.8% had DM, Bacteroides fragilis, thetaiotaomicron and Akkermansia muciniphila were least expressed ( Figure 3). As, these 3 species of bacteria appear to be associated with CRC, we evaluated the expression of these 3 species related to tumour grade and site of the tumor. There was no difference in the abundance of these 3 species of bacteria with the stage 0 to ii (n=9), iii (n=9) and iv (n=6), when analysed using the Kruskal-Wallis test. There was also no difference in their abundance based on gender or age. 14

Discussion
In this study we investigated the differences in the gut microbial patterns in a South Asian cohort of individuals with CRC and found that the gut microbial patterns were vastly different between those with CRC and healthy individuals. As a large proportion (70.83%) of those with CRC also had diabetes, in order to determine if the changes in the relative abundance of different bacteria were due to the presence of diabetes, we compared the gut microbial patterns in patients with DM with age matched healthy individuals. Again, we found that the gut microbial patterns were indeed markedly different in those with DM compared to healthy individuals. Further comparison of the microbial patients with DM with those with CRC showed that three bacterial species were more likely to be associated with CRC. These are Bacteroides fragilis, Bacteroides thetaiotaomicron and Akkermansia muciniphila. However, the relative abundance of these three bacterial species did not differ based on the tumour stage, age or the gender of patients.
Many previous studies have shown the association between CRC and Bacteroides fragilis [19].
The toxin produced by Enterotoxigenic Bacteroides fragilis has shown to change gene transcription in the colons of mice models by inducing epigenetic changes, which subsequently result in development of tumors [20]. The toxin producing strains of Bacteroides fragilis has shown to be more prevalent in patients with CRC [10]. Although we did not specifically assess toxin production by Bacteroides fragilis in this study, this bacterium was found to be 23.9 times overabundant in those with CRC compared to healthy individuals. Although we found that Bacteroides vulgatus was also two-fold over abundant in those with CRC compared to healthy 15 individuals, other studies have shown that Bacteroides vulgatus was more abundant in healthy individuals [21]. In addition, we found that Bacteroides thetaiotaomicron was 8-fold more over abundant in those with CRC when compared to those with CRC, which has not been reported before.
Ethnic differences have shown to play a significant role in the outcome of CRC, which has been attributed to possible differences in the gut microbiome [22]. For instance, while Bacteroides species, Fusobacterium nucleatum and Enterobacter species were more abundant in African-American patients with CRC, Akkermansia muciniphila and Bifidobacterium species were more abundant in Caucasians [22]. As our PCR array did not have Fusobacterium nucleatum and Bifidobacterium species included, we could not assess their relative abundance. However, Akkermansia muciniphila was 5.9-fold over abundant in those with CRC, compared to healthy individuals. Therefore, it would be important to carry out 16S sequencing of the whole gut microbiome to derive better data to find out the microbiota that associate with CRC in the Sri Lankan population.
As 70.8% of the CRC patients also had DM, in order to identify the gut microbiota patterns related to DM, we assessed the patterns between healthy individuals and those with DM. We found that all the 45 types of gut microbiota assessed in this study, were several folds over expressed in those with DM compared to healthy individuals. Specifically, the Aeromonas species, Enterococcus faecium and Shigella dysenteriae were over 100-fold over abundant in those with DM compared to healthy individuals. Therefore, bacteria of the Phylum Proteobacteria do appear to be 100-fold more abundant in South Asian patients with DM, similar to the observations in Western countries, despite differences in their diets [23]. This gut microbial dysbiosis that occurs due to the overgrowth of bacteria of the Phylum Proteobacteria has shown to associate with low grade endotoxaemia [24]. The presence of low levels of bacterial lipopolysaccharide (LPS) has shown to associate with DM and other metabolic diseases such as non-alcoholic steatohepatitis, which are rapidly increasing in all South Asian countries [25][26][27]. Therefore, similar patterns of gut microbial dysbiosis that associate with CRC and DM appear be seen in South Asian populations as in other countries, despite differences in the diet and ethnicity.
In summary, we studied the gut microbial patterns in a South Asian cohort of patients with CRC and found that Bacteroides fragilis, Bacteroides thetaiotaomicron and Akkermansia muciniphila were several folds over abundant in those with CRC when compared to healthy individuals.
However, as this study was limited to studying only 45 genera and species of microbiota, it would be important to study the whole gut microbiome by carrying out 16S sequencing to identify other possible microbes that associate with CRC. Early identification of such gut microbial dysbiosis could lead to prevention and treatment strategies in populations by possible nutrition interventions.