Multi-omics approaches to studying gastrointestinal microbiome in the context of precision medicine and machine learning

The human gastrointestinal (gut) microbiome plays a critical role in maintaining host health and has been increasingly recognized as an important factor in precision medicine. High-throughput sequencing technologies have revolutionized -omics data generation, facilitating the characterization of the human gut microbiome with exceptional resolution. The analysis of various -omics data, including metatranscriptomics, metagenomics, glycomics, and metabolomics, holds potential for personalized therapies by revealing information about functional genes, microbial composition, glycans, and metabolites. This multi-omics approach has not only provided insights into the role of the gut microbiome in various diseases but has also facilitated the identification of microbial biomarkers for diagnosis, prognosis, and treatment. Machine learning algorithms have emerged as powerful tools for extracting meaningful insights from complex datasets, and more recently have been applied to metagenomics data via efficiently identifying microbial signatures, predicting disease states, and determining potential therapeutic targets. Despite these rapid advancements, several challenges remain, such as key knowledge gaps, algorithm selection, and bioinformatics software parametrization. In this mini-review, our primary focus is metagenomics, while recognizing that other -omics can enhance our understanding of the functional diversity of organisms and how they interact with the host. We aim to explore the current intersection of multi-omics, precision medicine, and machine learning in advancing our understanding of the gut microbiome. A multidisciplinary approach holds promise for improving patient outcomes in the era of precision medicine, as we unravel the intricate interactions between the microbiome and human health.

Abstract 1405: Colorectal carcinogenesis in Smad4/SPTBN1 mutants alter the intestinal microbiome and resistance to 5FU

Background/Aims: Emerging data shows a rise in colorectal cancer (CRC) incidence in young men and women that is often chemoresistant, with potential new risk factors including alterations in the microbiome that remain understudied. We have recently observed altered microbiome with modulation of the gut immune response through crosstalk between sensors, microbes, and the TGF-β signaling pathway. Interestingly, we observed that human CRC cell lines with disruption of TGF-β had increased sensitivity to cisplatin and other DNA cross-linking agents. Yet at the same time, human CRC with loss of Smad4 portends poor prognosis and resistance to chemotherapy. Therefore, it was unclear whether the epithelial loss of Smad4 or stromal loss of Smad4 expression could be responsible for chemoresistance. Here, we investigated the role of chemotherapy with disruption of TGF-β signaling and an altered intestinal microbiome in colorectal carcinogenesis.

Methods: CRCs induced by azoxymethane (AOM)/dextran sodium sulfate (DSS) in wild type (WT) and TGF-β signaling deficient mice (SKO: Smad4+/- and Smad4+/-/Sptbn1+/-) were treated with 5-Fluoro-Uracil (5FU). Shotgun metagenomic sequencing was performed in fecal samples from WT and SKO mice before and after treatment.

Results: Our analyses revealed that SKO mice are more susceptible to AOM/DSS induced CRC as demonstrated by increased multiplicity (SKO vs WT: 10.25±1.5 vs 5.5±1.4, p=0.01) and tumor size (SKO vs WT: 3.63±0.7mm vs 2.17±0.3mm, p=0.02). CRC that develops in mice with disrupted TGF-β signaling is chemoresistance to 5FU and progress to liver metastasis confirmed by histological and immunohistochemical analysis. Interestingly, SKO mice display a unique gut microbiome signature compared to the WT mice. SKO mice had significantly reduced abundances of beneficial species of B. vulgutus (0.056±0.0078 vs 0.018±0.0052) and P. distasonis (0.007±0.001 vs 0.003±0.001). In addition to E. Boltae, and Mordavella sp., the relative abundance of Bacteroides was significantly reduced in AOM/DSS induced tumors and recovered to basal levels after 5FU treatment in WT mice but not in SKO mice with deficient TGF-β signaling (e.g. Bacteroides dorei: WT basal vs WT-CRC vs WT-CRC-5FU: 0.28±0.08 vs 0.05 ±0.01 vs 0.45±0.17; SKO basal vs SKO CRC vs SKO-CRC-5FU: 0.13±0.02 vs 0.06 ±0.007 vs 0.07±0.01).

Conclusions: Our study identified unique gut microbiome species corresponding to 5FU resistance through interactions with key immune-related pathways such as TGF-β members that are implicated in CRC development and drug resistance. The in vivo studies suggest a cell non-autonomous role of the TGF-β pathway in CRC chemoresistance. Collectively, the altered microbiome composition from impaired TGF-β signaling could promote colorectal carcinogenesis and drug resistance. Manipulating these specific species associated with 5FU could potentially increase drug response.

Citation Format: Shuyun Rao, Zhuanhuai Wang, Xiaochun Yang, Lindsay M. Hopson, Stephanie S. Singleton, Wilma Jogunoori, Raja Mazumder, Vincent Obias, Paul Lin, Bao-Ngoc Nguyen, Michael Yao, Larry Miller, Jon White, Lopa Mishra. Colorectal carcinogenesis in Smad4/SPTBN1 mutants alter the intestinal microbiome and resistance to 5FU [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 1405.

Mice with dysfunctional TGF-β signaling develop altered intestinal microbiome and colorectal cancer resistant to 5FU

Emerging data show a rise in colorectal cancer (CRC) incidence in young men and women that is often chemoresistant. One potential risk factor is an alteration in the microbiome. Here, we investigated the role of TGF-β signaling on the intestinal microbiome and the efficacy of chemotherapy for CRC induced by azoxymethane and dextran sodium sulfate in mice. We used two genotypes of TGF-β-signaling-deficient mice (Smad4^+/- and Smad4^+/-Sptbn1^+/-), which developed CRC with similar phenotypes and had similar alterations in the intestinal microbiome. Using these mice, we evaluated the intestinal microbiome and determined the effect of dysfunctional TGF-β signaling on the response to the chemotherapeutic agent 5-Fluoro-uracil (5FU) after induction of CRC. Using shotgun metagenomic sequencing, we determined gut microbiota composition in mice with CRC and found reduced amounts of beneficial species of Bacteroides and Parabacteroides in the mutants compared to the wild-type (WT) mice. Furthermore, the mutant mice with CRC were resistant to 5FU. Whereas the abundances of E. boltae, B.dorei, Lachnoclostridium sp., and Mordavella sp. were significantly reduced in mice with CRC, these species only recovered to basal amounts after 5FU treatment in WT mice, suggesting that the alterations in the intestinal microbiome resulting from compromised TGF-β signaling impaired the response to 5FU. These findings could have implications for inhibiting the TGF-β pathway in the treatment of CRC or other cancers.

Bioinformatics and machine learning in gastrointestinal microbiome research and clinical application

The scientific community currently defines the human microbiome as all the bacteria, viruses, fungi, archaea, and eukaryotes that occupy the human body. When considering the variable locations, composition, diversity, and abundance of our microbial symbionts, the sheer volume of microorganisms reaches hundreds of trillions. With the onset of next generation sequencing (NGS), also known as high-throughput sequencing (HTS) technologies, the barriers to studying the human microbiome lowered significantly, making in-depth microbiome research accessible. Certain locations on the human body, such as the gastrointestinal, oral, nasal, and skin microbiomes have been heavily studied through community-focused projects like the Human Microbiome Project (HMP). In particular, the gastrointestinal microbiome (GM) has received significant attention due to links to neurological, immunological, and metabolic diseases, as well as cancer. Though HTS technologies allow deeper exploration of the GM, data informing the functional characteristics of microbiota and resulting effects on human function or disease are still sparse. This void is compounded by microbiome variability observed among humans through factors like genetics, environment, diet, metabolic activity, and even exercise; making GM research inherently difficult to study. This chapter describes an interdisciplinary approach to GM research with the goal of mitigating the hindrances of translating findings into a clinical setting. By applying tools and knowledge from microbiology, metagenomics, bioinformatics, machine learning, predictive modeling, and clinical study data from children with treatment-resistant epilepsy, we describe a proof-of-concept approach to clinical translation and precision application of GM research.