Society: AGA

Background: Diets high in animal fat and low in plant fiber increase colorectal cancer (CRC) risk, while diets high in fiber, but low in animal fat, are protective. Bacteria present in the colon produce beneficial and harmful metabolites from our diets. Fiber is fermented by these bacteria, producing short-chain fatty acids (SCFA). One of them, butyric acid, is the major energy source for colonic cells and has strong anti-inflammatory/anti-carcinogenic properties. Excess consumption of animal fats is known to increase levels of carcinogenic secondary bile acids in the colon, namely lithocholic acid (LCA) and deoxycholic acid (DCA). Alaska Native (AN) people, at highest risk globally for CRC, were interested in whether increasing dietary fiber intake could reduce CRC risk, as their diets are high in meat and fat, but deficient in fiber.
Purpose: Evaluate whether a fiber supplement added to the usual diet of AN volunteers will increase the microbial production of butyrate and suppress fecal bile acids, thus reducing CRC risk.
Methods: Forty-eight healthy AN adult volunteers were enrolled a 4-week randomized, double-blind, placebo-controlled trial in which they received either a resistant starch supplement, RS, a non-chemically modified food starch extracted from high amylose corn, or a digestible starch supplement, DS, an isocaloric amylopectin supplement, occurring naturally as a branched glucose polymer. Quantitation of fecal SCFA and bile acids was performed by on stool samples collected pre- and post-supplementation by gas chromatography-flame-ionization detection (GC-FID) and gas chromatography-mass spectrometry (GC-MS) (Figure 1). The primary endpoint was to evaluate whether RS could suppress epithelial proliferation measured by immunohistological staining of proliferative cells with Ki67 in well orientated crypts. The magnitude of the Ki67 suppression was also used to define clinically significant responses of <20%, termed “responders” (Figure 2). Statistical analyses were performed with paired t-tests on all SCFA and bile acids.
Results: Butyrate levels showed increases in both groups, significant in the RS group. Conversely, there were reductions in LCA and DCA in both groups, significant for DCA in the RS group.
Conclusions: RS significantly increased fecal anticarcinogenic butyrate and suppressed inflammatory and potentially carcinogenic secondary bile acids which might explain the suppression of cancer biomarkers in the colonic mucosa. Interestingly, DS had similar but less pronounced effects. It is noteworthy that previous dietary analysis indicated the AN diet was also low in carbohydrate (Ocvirk et al. AJCN 2016) and so it is possible that their diet could be improved by the increased consumption of all forms of complex carbohydrate.
Acknowledgements: NIH grant R01 CA204403

Background: Obesity, particularly abdominal obesity, is a major risk factor of cancer. However, it remains unknown how visceral fat may influence cancer risk independently of traditional measures of overall and abdominal obesity, due to the lack of large-scale data on visceral fat. Methods: Leveraging the body (abdominal) magnetic resonance imaging (MRI) measurements among 23,148 participants who attended the first imaging test since 2014 in the UK Biobank (UKB) cohort, we prospectively examined visceral adipose tissue volume (VAT) in relation to the risk of overall, obesity-related, and gastrointestinal (GI) cancer and major individual cancers in men and women, using multivariable Cox proportional hazards regression (see the table 1 footnote for the covariate list). We also trained and validated a prediction equation for VAT based on demographic factors and traditional anthropometric measures (see the predictor list and brief summary of process in the table footnote). We compared the association of the measured and predicted VAT with cancer risk among individuals with the MRI data, and then assessed the predicted VAT in relation to cancer risk among 471,089 participants in the overall UKB cohort who attended the initial assessment visit (2006-2010). To examine the added value of VAT beyond traditional measures of overall and abdominal obesity, we conducted a joint association analysis for predicted VAT with body mass index (BMI) and waist circumference (WC) in relation to total cancer risk in the overall UKB cohort. Results: In the MRI subcohort and overall UKB cohort, respectively, we documented 719 and 47,940 cancer cases during a median follow-up of 4 and 12 years. In the MRI subcohort, each unit (90^th to 10^th difference) increase in the measured VAT was associated with 33% and 74% higher risk of total cancer in men (multivariable HR=1.33, 95% CI: 1.05-1.67) and women (1.74, 1.11-2.72), respectively. The final equation for predicted VAT showed a similarly high R² in the training (0.76) and validation (0.74) sets; no difference was found between the measured and predicted VAT in the validation set (p-value>0.05), with Spearman correlation (r=0.87). In the MRI subcohort, the predicted VAT showed similar HRs to those of the measured VAT. After applying the prediction equation to the overall UKB cohort, we found that the predicted VAT was associated with higher risk of total cancer and many individual cancers, particularly GI, kidney, liver, and endometrial cancers. Furthermore, within each of the BMI and WC groups, predicted VAT was associated with a higher risk of total cancer. Conclusions: VAT was associated with an increased risk of total cancer and several individual cancers. VAT also provided additional information regarding total cancer risks within each of the BMI or WC groups.

Background and aims: Colorectal cancer (CRC) cells experience fluctuating metabolic vulnerability to facilitate tumor progression and metastasis. However, the molecular mechanism underlying CRC metabolism reprogramming remains unclear. Kruppel-like factor (KLF) family proteins play crucial roles in carcinogenesis and progression. Here we report a novel function of KLF4 in promoting CRC aggressiveness by regulating glucose metabolism.
Methods: KLF4 expression was examined by Real time-PCR, Western blot and Immunohistochemical staining in CRC tissue and cells. The clinical significance of KLF4 in two independent cohorts of CRC patients was assessed by Kaplan-Meier analysis and the Multivariate Cox proportional hazards model. The biological function of KLF4 in tumor growth and metastasis was detected both in vivo and in vitro. Glucose consumption, lactate production and extracellular acidification rate (ECAR) were analyzed by the specific fluorescent probe. Serial deletion, site-directed mutagenesis luciferase report assays and chromatin immunoprecipitation (ChIP) were used to determine transcriptional regulation of TKT, PGM1 and ACO2 promoters by KLF4.
Results: The key enzymes of glucose metabolism were differentially expressed in CRC and adjacent non-tumorous tissues, in which the ectopic expression of KLF4 is closely related to multiple glucose metabolic pathways. Downregulation of KLF4 was significantly correlated with more aggressive CRC and indicated a poorer prognosis in CRC patients. KLF4 downregulation was an independent and significant risk factor for recurrence and poor survival. Knockdown of KLF4 promoted CRC proliferation and metastasis both in vitro and in vivo, whereas upregulation of KLF4 hampered CRC aggressiveness. Mechanically, dysregulation of KLF4 facilitated CRC glucose metabolism by transcriptionally inhibiting TKT, whereas transactivating PGM1 and ACO2. Deregulation of TKT, PGM1 and ACO2 blocked KLF4 downregulation-mediated CRC progression, while ectopic expression of TKT, PGM1 and ACO2 rescued KLF4-suppressed proliferation and metastasis. Furthermore, the cAMP-response element binding protein (CREB) that is inhibited by lactate receptor GPR81, binds directly to the KLF4 promoter and activates its expression. Inhibitor of GPR81 attenuated KLF4 deregulation-mediated tumor growth and metastasis.
Conclusions: Lactate-induced downregulation of KLF4 promotes CRC proliferation and metastasis by regulating the expression of three key enzymes for glucose metabolism TKT, PGM1 and ACO2. The lactate receptor GPR81 inhibitor deregulates KLF4 deficiency-promoted CRC proliferation and metastasis, implicating KLF4 as a potential tumor-suppressive prognostic biomarker and a novel therapeutic target for CRC.

Figure 1. KLF4 downregulation correlates with glucose metabolism and promotes CRC progression. (A-C)TCGA dataset shows that the key enzymes of glucose metabolism were differentially expressed in CRC and adjacent non-tumorous tissues. (D)The expression profile of KLF gene family in 473 CRC tissues and 41 normal colorectal epithelial tissues. (E)Gene set enrichment analysis shows that the dysregulation of KLF4 is closely related to multiple glucose metabolic pathways. (F-G)The downregulation of KLF4 indicates a poor prognosis for CRC in TCGA datasets. (I-J)Immunohistochemistry staining analysis of KLF4 expression in adjacent non-tumorous tissues and CRC tissues. (K)Correlation between KLF4 expression and prognosis in CRC patients. (L-O)Cell counting kit-8, clone formation, transwell analysis of CRC cell proliferation, migration and invasion. (M) In vivo tumorigenesis assay. Representative bioluminescent images, tumor growth curves, and the overall survival in different groups are shown.

Figure 2. KLF4, transcriptionally inhibiting TKT whereas transactivating PGM1 and ACO2 promotes glucose metabolism in CRC cells. (A)Transcriptome screening in KLF4 knockdown CRC cells vs control. (B)The sequence of KLF4 DNA-binding sites predicted by JASPAR, ENCODE, MotifMap, TRANSFAC, TRUST and CHEA. (C-D)Venn diagram and volcano plot for KLF4 target genes. (E-G)Glucose consumption, lactate production, and ECAR analysis for glucose metabolism. (H-I)KLF4 represses TKT while activating PGM1 and ACO2 promoter activities. (J-O)Deletion and selective mutation analysis and ChIP assays identify the KLF4-responsive regions and their direct binding sites to the TKT, PGM1 and ACO2 promoters. (P) KLF4 expression was inhibited by lactate stimulation. (Q)GPR81 and PKA inhibitors inhibited lactate-induced downregulation of KLF4 expression. (R-S)Methylation analysis indicates hypermethylation of the KLF4 promoter. Protein interaction analysis implies KLF4 may interact with HDAC2, TP53 and LIN28A.

Figure 2. KLF4, transcriptionally inhibiting TKT whereas transactivating PGM1 and ACO2 promotes glucose metabolism in CRC cells. (A)Transcriptome screening in KLF4 knockdown CRC cells vs control. (B)The sequence of KLF4 DNA-binding sites predicted by JASPAR, ENCODE, MotifMap, TRANSFAC, TRUST and CHEA. (C-D)Venn diagram and volcano plot for KLF4 target genes. (E-G)Glucose consumption, lactate production, and ECAR analysis for glucose metabolism. (H-I)KLF4 represses TKT while activating PGM1 and ACO2 promoter activities. (J-O)Deletion and selective mutation analysis and ChIP assays identify the KLF4-responsive regions and their direct binding sites to the TKT, PGM1 and ACO2 promoters. (P) KLF4 expression was inhibited by lactate stimulation. (Q)GPR81 and PKA inhibitors inhibited lactate-induced downregulation of KLF4 expression. (R-S)Methylation analysis indicates hypermethylation of the KLF4 promoter. Protein interaction analysis implies KLF4 may interact with HDAC2, TP53 and LIN28A.

Background: The secondary bile acid deoxycholic acid (DCA) is produced by intestinal bacteria with 7α-dehydroxylating (7αDH) activity. Higher concentrations of DCA are positively associated with colorectal cancer (CRC) risk, but underlying molecular mechanisms demonstrating a causal role of DCA are still not fully revealed. In this study, we investigate the tumor-promoting role of DCA in vivo in a gnotobiotic CRC mouse model.
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. Control mice were colonized with BACOMI without C. scindens, thus lacking 7αDH activity and DCA. Some mice were treated with AOM/DSS to induce colonic tumors. Cecal bile acid composition was analyzed using LC-MS/MS, bacterial abundances were measured in colonic content via qPCR and colonic tissue was analyzed using immunofluorescence-staining and qPCR. Primary T -cells were isolated from mesenteric and iliac lymph nodes and further analyzed by flow cytometry.
Results: Successful colonization of BACOMI revealed distinct bile acid profiles of BACOMI vs. control mice as well as higher abundances of C. scindens in mice treated with AOM/DSS compared to mice without AOM/DSS treatment. In the AOM/DSS experiment, presence of DCA led to significantly higher colonic tumor numbers (+7αDH = 20,5 tumors ±7,5 SD vs. -7αDH = 7,9 tumors ±2,0 SD), and shorter colon lengths compared to mice devoid of DCA. The difference was most pronounced for small- and medium-sized tumors. In mice colonized with BACOMI without AOM/DSS treatment, the presence of DCA lead to significantly higher numbers of Ki67+ cells in colonic crypts and more pronounced hypoxia in the colonic epithelium. Colonic expression of genes associated with CRC (e.g., Axin2, Ccnd1, Myc) and bile acid metabolism (e.g., Nr1h4, Nr0b2, Slc51b) were significantly upregulated in mice having DCA compared to control mice lacking DCA. Different subtypes of Foxp3+ regulatory T -cells were significantly more abundant in mesenteric and iliac lymph nodes of mice where DCA was present compared to mice lacking DCA.
Conclusions: Microbially-derived DCA promotes colonic tumorigenesis in vivo in an experimental gnotobiotic model for CRC. Even in the absence of chemically-induced tumorigenesis, DCA promotes markers associated with CRC, suggesting a causal role of DCA in colonic tumorigenesis.

Background: Gut microbiota imbalances associated with inflammation and altered metabolism are prominent risk factors for GI cancers, such as colorectal cancer (CRC) and hepatocellular carcinoma (HCC). CEACAM proteins are known immune and epithelial cell regulators, and CEACAM1, 3, 5, and 6 bind to pathogenic microbes. Mutations of the CEACAM-encoding genes produce significant alterations in the gut microbiome through altered regulation of host-immune response, and changes correlate with scores of TGF-β pathway activity in GI cancers. Nearly 40% of human GI cancers have genetic alterations in the TGF-β pathway members, and cancers occur in mouse models with disruption of TGF-β signaling only in the presence of an altered microbiome. Yet, in-depth analyses of how cancer pathways such as TGF-β signaling are involved remain to be elucidated. We have recently found that in NASH-HCC murine model, hepatocyte-specific βII-spectrin knockout (LSKO) blocked diet-induced NASH and HCC (Sci Transl Med. 2021;13(624): eabk2267). We, therefore, investigated crosstalk between pathogen-binding CEACAM1, the gut microbiome, and the multifunctional intrinsically-disordered protein βII-spectrin (SPTBN1), a Smad3 adaptor, in driving GI cancers.
Methods: DNA from fecal samples collected from wild-type mice with disruption of TGF-β signaling (Smad4^+/-Sptbn1^+/-) were analyzed through shotgun metagenomics sequencing to identify differences in gut microbiota composition. We examined CEACAM expression, changes in the T cell populations, liver inflammation, and cancer in tissues from wild-type mice and liver-specific βII-spectrin knockout (LSKO) mice fed either a normal chow diet or DEN and Western diet (WD) induced NASH-HCC model.
Results: Doubly heterozygous Smad4^+/-Sptbn1^+/- mice (C57BL/6J) spontaneously develop GI cancers. These mice exhibit increased bacterial species associated with CRC in humans (C. septicum) and decreased commensal gut microbes associated with a healthy microbiome (B. vulgatus, P. distasonis). Single-cell RNA sequencing analyses revealed that Ceacam1 expression decreases in early CRC and in hepatocytes of WD fed mice (NASH). This result was corroborated using immunohistochemical analyses showing a decreased CEACAM1 expression in NASH livers of WD fed wild-type mice. Strikingly, CEACAM1 levels are increased and restored to normal in tissues from liver-specific βII-spectrin knockout (LSKO) mice. In addition, diet-induced changes in the T cell populations, liver inflammation, and cancer are blocked in LSKO mice.
Conclusion: Our results indicate that two critical pathways, TGF-β signaling, and immune mediator CEACAM1 can suppress inflammation and early GI cancers that arise with microbiome alterations.

Introduction: NASH is primary risk factor for hepatocellular carcinoma (HCC). In tumors, a rate-limiting enzyme in glycolysis, pyruvate kinase isozyme M2 (PKM2) mediates inflammation and aerobic glycolysis in Warburg effect that can induce ferroptosis. PKM2 increased in hepatic macrophages is associated with poor prognosis in HCC. We previously reported that liver-specific TGF-β/SMAD4 knockout mice develop iron deposition and hemochromatosis by a dramatic loss of a key regulator of the iron circulation, hepcidin. Moreover, Smad4^+/-Sptbn1^+/- mice develop gastrointestinal cancers with an altered microbiome. Recently, we found that hepatocytes-specific knockout (LSKO) of a TGF-β/SMAD3/4 adaptor protein, βII-spectrin blocks diet-induced NASH and HCC in mice, and in siRNA-treated human NASH 3D cultures (Sci Transl Med. 2021;13(624): eabk2267). Here, we hypothesized that βII-spectrin modulates PKM2 expression in NASH fatty liver, promoting proinflammatory cytokine release, inflammation, and tumorigenesis, and that microbiome change could alter hepcidin function and modulate ferroptosis.
Methods: SPTBN1^Flox (control) and liver-specific βII-spectrin knockout (SPTBN1^LSKO) mice were fed Western diet (WD) to develop NASH mouse model, and they were fed WD and treated with Diethylnitrosamine (DEN) to develop a NASH-associated HCC mouse model. Body weight, total cholesterol, and triglyceride concentrations in serum were measured, PKM2 expression level in liver tissue was analyzed, and gut microbiome profiling was performed in these mice. To evaluate PKM2 expression in the NASH-associated HCC mouse model, immunohistochemical labeling was performed on liver sections of these mice. Inflammation, lipidosis, and fibrosis were determined with H&E, Oil Red O and Sirius Red staining.
Results: Upregulation of PKM2 expression was shown in NASH and HCC Kupffer cells in SPTBN1^Flox mice fed on WD. In contrast, PKM2 expression in the liver was markedly reduced in SPTBN1^LSKO mice fed on WD in which NASH is blocked, In addition, PKM2 expression increased in SPTBN1^Flox mice and markedly decreased in SPTBN1^LSKO mice fed on a normal diet and treated with DEN. Microbiome profiles and TGF-β/SMAD3–regulated fibrosis were altered, and expression of inflammatory genes was significantly reduced in the LSKO mice compared to the NASH mice.
Conclusions: Our results suggest that the SMAD3/4 adaptor, βII-spectrin knockdown decreases PKM2 expression in Kupffer cells, thereby suppressing pro-inflammatory cytokine production, blocking NASH-associated HCC. Hepatic macrophages such as Kupffer cells could therefore play an important role in the ferroptosis-mediated protumor immune microenvironment. In addition, our study will provide new insights into the molecular mechanisms in the development of NASH and HCC.

Adrian Krainer

Appearances