Introduction

Intestinal bacteria with 7α-dehydroxylating (7αDH) activity are able to produce the secondary bile acids, such as deoxycholic acid (DCA). Higher concentrations of DCA are linked to a higher risk of colorectal cancer (CRC) risk, yet underlying molecular mechanisms demonstrating a causal role of DCA are still unclear. This includes a recently demonstrated immunomodulatory effect of secondary bile acids, which may contribute to CRC pathogenesis. Here, we aim to investigate the tumor-promoting and immunomodulatory role of DCA in a gnotobiotic CRC mouse model.

Aims & Methods

Germ-free wildtype mice were colonized with a simplified bacterial consortium, the bile acid converting microbiota (BACOMI). BACOMI is able to produce DCA due to the 7αDH activity of Clostridium scindens (+7αDH). Control mice were colonized with BACOMI without C. scindens, lacking 7αDH activity (-7αDH) and DCA. A sub-group of mice was treated with AOM/DSS to induce colonic tumors. Bacterial abundances were measured in colonic content via qPCR, cecal bile acid composition was analyzed using LC-MS/MS and colonic tissue was analyzed using qPCR and immunofluorescence-staining. Primary T-cells and macrophages were isolated from the colon as well as mesenteric and iliac lymph nodes and analyzed by flow cytometry.

Results

DCA was exclusively present in mice colonized with the BACOMI consortium and led to higher colonic tumor numbers compared to mice colonized with BACOMI without C. scindens (+7αDH = 20.5 tumors±7.5 SD vs. -7αDH = 7.9 tumors±2.0 SD), in particular for small- and medium-sized tumors. Presence of DCA also resulted in shorter colon lengths and lower levels of butyrate in cecal content. In BACOMI colonized mice, the treatment with AOM/DSS promoted colonic colonization by C. scindens compared to mice without AOM/DSS treatment. Even without AOM/DSS treatment, the presence of C. scindens and DCA resulted in significantly higher numbers of Ki67+ epithelial cells in colonic crypts and more pronounced hypoxia in the colonic epithelium. Colonic expression of genes involved in CRC pathogenesis (e.g., Axin2, Ccnd1, Myc), bile acid metabolism (e.g., Nr1h4, Nr0b2, Slc51b) as well as immune responses (e.g., Nfkb1, Rorc, Itgam) was upregulated in mice with DCA. Sub-types of Foxp3+ regulatory T-cells were more abundant in mesenteric and iliac lymph nodes whereas numbers of Ly6C high macrophages were lower in the colon and mesenteric lymph nodes of mice with DCA compared to mice lacking DCA.

Conclusion

​In a gnotobiotic mouse model for CRC, the presence of microbially-derived DCA promotes colonic tumorigenesis in vivo. Even in the absence of chemically-induced tumorigenesis, DCA stimulates the expression of markers closely associated with CRC and promotes colonic epithelial proliferation. These findings strongly imply a causal role of DCA in the development of CRC. Additionally, DCA shows immunomodulatory effects, which may demonstrate another mechanistic link of how DCA affects intestinal health.

Introduction

A wide range of relationships between the microbial community and various environmental or individual factors have been proposed. Monozygotic and dizygotic twins provide a unique model for assessing host-microbial interactions. The mycobiome has been increasingly implicated in the development of several diseases, but knowledge of the dynamics of the fungal community, particularly in twins, has not been addressed.

Aims & Methods

In this work, we investigated the fungal community in a twin cohort and explored its relationship with associated genetic and environmental factors through in-depth analysis.
The twin cohort consisted of 106 well-characterised and systematically studied monozygotic and dizygotic pairs (212 stool samples). DNA was extracted and PCR reactions were performed to sequence the lower gastrointestinal microbiome and the mycobiome community. The bacterial community was characterised using the V1-V2 region of the 16 rRNA gene and the fungal community was analysed using the ITS1-ITS2 region.

Results

The fungal community, represented by the sum of all fungal sequences, showed a significantly higher heterogeneity than the bacterial community. Several individual and genetic factors, such as the zygotic status of the twins (monozygotic vs. dizygotic), as well as sex, mode of birth, breast-feeding as newborns or living in the same household, were independent of the total number of fungal sequences. The bacterial community was significantly more abundant than the fungal community, the most common representatives of which were Candida and Geotrichum with sometimes very high abundances. Saccharomyces was also common in the cohort, but only in low abundance, other fungi with a wider distribution were Cladosporium or Penicillium. The environmental and individual factors mentioned above were also correlated with the individual fungal genera, but there were no significant correlations found here either. The next step was to correlate the bacterial and fungal communities with each other, and strong negative correlations were found with certain bacteria such as Prevotella.

Conclusion

Genetic factors in twins are not associated with dynamics in the faecal fungal community and were the subject of high inter-individual variability. However, there is a strong association between bacterial and mycobial interaction, which requires further in-depth understanding of microbial interactions.

Disclosure

No conflict of interest to declare

Introduction

Fibroscan is a well-established NIT for the diagnosis of advanced fibrosis (F≥3) in patients with NAFLD, recently defined as compensated advanced chronic liver disease (cACLD). EASL Guidelines¹ proposed 8 and 12 kPa, respectively, as rule-out and rule-in cut-offs for cACLD. Patients with Fibroscan measurement between 8 and 12 fall in a grey zone where further investigations are recommended. We recently proposed the cACLD Risk Score to further stratify this population.²
The main aim of this study was to test the diagnostic performance of the main NITs in a two European cohorts of patients with histological diagnosis of NAFLD. Secondly, we assessed the performance of the cACLD risk score to further stratify patients in the Fibroscan’s grey zone.

Aims & Methods

This is a retrospective observational study. We enrolled consecutive patients with histological diagnosis of NAFLD/NASH from January 2014 to December 2021 at two tertiary liver units in UK (the Royal Free Hospital, London - RFH) and Italy (Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome – FPG). We excluded patients who did not perform at least one of these NITs at time of biopsy (± 6 months): FIB4, NAFLD Fibrosis score (NFS), Fibroscan, APRI, AGILE3+, cACLD Risk Score. We performed a ROC analysis to explore the diagnostic performance of NITs for cACLD (F>2). Secondly, in patients with intermediate Fibroscan results (between 8-12 kPa), we tested the diagnostic performance of cACLD Risk Score, considering the originally proposed cutoff of 0,5.

Results

We included 536 patients;201 were female, median age was 53 years, and 44% had diabetes.
Fibroscan and AGILE3+ score had the best diagnostic performances for cACLD with an AUROC of 0.81 and 0.83 respectively. EASL criteria showed a high sensitivity and high specificity for 8 kPa and 12 kPa Fibroscan’s cutoffs, respectively (Sens 84.3%; Spec 86.6%), a wrong classification rate of 14.2%. 129 patients (24.1%) fell in the grey zone (between 8 and 12 kPa).
Diagnostic performance of cACLD Risk Score in Fibroscan’s grey zone was suboptimal (AUROC: 0.682). (Fig. 2)
The use of Fibroscan+cACLD Risk Score vs Fibroscan alone showed a better overall accuracy (77.8% vs 61.7%) and a slight worsening of overall wrong classification rate (22.7% vs 14.2%) mainly due to a higher number of false negatives than false positives. Among patients with intermediate Fibroscan (8-12 kPa), cACLD risk score correctly classify 64.4% of patients with a PPV of 72.0% and NPV of 62.5%.

Conclusion

Our results suggest that cACLD Risk Score should be used in patients with indeterminate Fibroscan results in order to further stratify their risk of cACLD. The use of cACLD Risk Score improve the performance in identifying patients with cACLD (high PPV) although its false negative rate could lead to missed diagnosis (suboptimal NPV). Patients with intermediate fibroscan and low cACLD risk score should be still considered for further investigation (liver biopsy).

References

¹ Berzigotti A, Tsochatzis E, Boursier J, et al. EASL Clinical Practice Guidelines on non-invasive tests for evaluation of liver disease severity and prognosis – 2021 update. Journal of Hepatology. 2021;75(3):659-689.
² Papatheodoridi M, Hiriart JB, Lupsor-Platon M, et al. Refining the Baveno VI elastography criteria for the definition of compensated advanced chronic liver disease. J Hepatol. 2021 May;74(5):1109-1116.