Dietary Oxalate and Intestinal Inflammation: Evidence From Experimental Colitis and Inflammatory Bowel Disease Patient Cohorts

Abstract

Background & Aims

The role of diet in the pathogenesis of inflammatory bowel diseases remains unclear. Most dietary interventions for inflammatory bowel disease improve symptoms without consistent mucosal healing; exceptions include exclusive enteral nutrition and exclusion diets in patients with Crohn’s disease. Oxalate, a naturally occurring compound found in all plant foods, is absorbed through the gut and has been shown to activate systemic and renal proinflammatory immune responses. However, the impact of intestinal oxalate on innate immune response in the context of inflammatory bowel disease is unknown.

Methods

We measured gene expression, stool oxalate content, and dietary intake in people with inflammatory bowel diseases and controls. Complementary studies were conducted in mouse models of chemically induced colitis, spontaneous colitis, and ex vivo cell culture systems to evaluate the relationships observed between oxalate transporter expression, stool oxalate content, and mucosal inflammation.

Results

Intestinal oxalate transporters, SLC26A2 and SLC26A3 are consistently downregulated across inflammatory bowel disease subtypes (Crohn’s disease, ulcerative colitis), inflammatory bowel disease tissues (colon, ileum), and in mouse models of experimental colitis irrespective to experimental diet assignment. In patients with Crohn’s disease, we observed higher stool oxalate content despite variable dietary oxalate intake. Altered expression of SLC26A6 corresponded to stricturing behavior of disease in patients with Crohn’s disease. Altered expression corresponded more directly to higher stool oxalate content and increased disease activity in mice fed an oxalate supplemented diet.

Conclusions

Dietary oxalate may exacerbate innate immune responses in susceptible individuals. Expression of oxalate transporters may be a disease-associated biological signal of both sensitivity to dietary oxalate and possibly, pathogenesis and clinical trajectory of patients with inflammatory bowel disease. This warrants future exploration of the role of dietary oxalate in mucosal inflammation in inflammatory bowel diseases.

Summary.

This study identifies altered intestinal oxalate handling as a potential contributor to inflammatory bowel disease pathogenesis. We demonstrate that impaired oxalate transport is associated with increased stool oxalate and may represent a novel therapeutic target for precision nutrition strategies through integrated human cohort analyses and experimental models.

What You Need to Know.

Background

Diet influences gut health, yet few dietary interventions have demonstrated consistent benefit in inflammatory bowel diseases, with the exception of elemental diets that are difficult to sustain.

Impact

These findings identify dietary oxalate as a potentially modifiable factor in intestinal inflammation and underscore the need to consider variation in dietary exposures and oxalate metabolism in inflammatory bowel diseases.

Future Directions

Future studies will utilize molecular measurements of dietary intake from stool to determine how dietary oxalate influences microbial populations and metabolic functions, revealing new therapeutic targets to reduce intestinal inflammation.

Inflammatory bowel disease (IBD) encompasses 2 disorders, ulcerative colitis (UC) and Crohn’s disease (CD), marked by chronic inflammation of the digestive tract. Susceptibility is thought to include genetics and diet.1, 2, 3, 4 Despite hundreds of previous dietary studies, no oral dietary intervention can be recommended to resolve inflammation across the broad IBD population, likely because response to diet varies greatly among genetically diverse individuals as well as within molecular subtypes of IBD.^1,2,5 To date, exclusive enteral nutrition is the only dietary intervention capable of resolving inflammation and promoting mucosal healing in CD.^1,2 In adults, exclusive enteral nutrition can reduce inflammation and induce remission in some patients, but adherence is difficult, and mucosal healing rates are sometimes lower and less consistent than in pediatric populations.⁶ Recently, the CD Exclusion Diet has been used in combination with partial enteral nutrition and improved clinical outcomes in a subset of patients with CD, but evidence for maintenance of remission is limited.²

Precision nutrition is a component of precision medicine that aims to improve the efficacy of dietary interventions for the prevention and treatment of nutrition-related diseases by leveraging genetics, sex, dietary intake, lifestyle, and disease status.7, 8, 9 Developing precision nutrition strategies for IBD will help to resolve the role of diet in molecular subtypes of IBD. A cohort study recently demonstrated that higher amounts of oxalate in the feces of patients with IBD corresponded to higher intestinal inflammation.¹⁰ Oxalate is a naturally occurring plant compound, present in variable amounts in all plant foods.¹¹ Several members of the solute carrier 26 (SLC26) family of genes are associated with regulating oxalate homeostasis by both absorption and secretion of oxalate in the gastrointestinal (GI) tract.^12,13 We hypothesized that oxalate induces mucosal inflammation in IBD. We used human observational data followed by complementary studies in mice and cell culture systems to examine how dietary oxalate and oxalate transporter expression relate to mucosal inflammation in the context of IBD. The study design and experimental procedures are represented in Figure 1.

Figure 1.

Study design.

Results

Reduced Expression Pattern of Intestinal Oxalate Transporters SLC26A2 and SLC26A3 is Observed Across Inflammatory Bowel Disease Subtypes and Disease Activity

The SLC26 gene family regulates intestinal oxalate homeostasis.12, 13, 14, 15, 16, 17, 18 Impaired expression of SLC26A2 and SLC26A3 in patients with IBD has been observed by others.18, 19, 20 We performed RNA sequencing (RNA-seq) on colon tissue from patients with UC and non-IBD (NIBD) control patients (NIBD, n = 18; UC uninflamed, n = 40; UC inflamed, n = 40) and observed that SLC26A2, SLC26A3, and SLC26A6 were consistently downregulated in the colon of patients with UC relative to patients with NIBD, with the lowest expression observed in inflamed colonic tissue (Figure 2A). The difference between the paired inflamed and uninflamed samples in individuals with UC was statistically significant (Figure 2B). This suggests that expression of the transporters is further reduced during active disease.

Figure 2.

Expression of oxalate transporters is reduced in IBDs. (A) Expression of SLC26A2, SLC26A3, and SLC26A6 from bulk RNA-seq in colonic mucosa from patients with NIBD and UC (NIBD, n = 18; UC uninflamed, n = 40; UC inflamed, n = 40). (B) Paired analysis of expression of SLC26A2, SLC26A3, and SLC26A6 in uninflamed/inflamed colonic mucosa from patients with UC (UC uninflamed, n = 40; UC inflamed, n = 40). (C) Expression of SLC26A2, SLC26A3, and SLC26A6 in ileal mucosa from patients with NIBD and CD (NIBD, n = 28; CD uninflamed, n = 26; CD inflamed, n = 78). (D) Expression of SLC26A2, SLC26A3, and SLC26A6 in uninflamed colonic mucosa from patients with NIBD and CD (NIBD, n = 80; CD, n = 131). Asterisks denote ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, and ∗∗∗∗P < .0001.

To determine whether these findings were consistent across other types of IBD, we performed RNA-seq on the ileum and colon from patients with CD and NIBD (ileum: NIBD, n = 28; CD uninflamed, n = 26; CD inflamed, n = 78; Colon: NIBD, n = 80; CD uninflamed, n =131) as well. We observed that SLC26A2 and SLC26A3 were consistently downregulated in the ileum of patients with CD relative to those with NIBD (Figure 2C), and SLC26A2, SLC26A3, and SLC26A6 were downregulated in the colon of patients with CD relative to patients with NIBD (Figure 2D). The expression of SLC26A2 and SLC26A3 was lower in inflamed ileal tissue (Figure 2C). These results show that reduced expression of SLC26A2 and SLC26A3 is consistent across tissues and IBD subtypes and further support the observation that expression of the transporters is further reduced during active disease.

Higher Stool Oxalate Content in Crohn’s Disease Independent of Habitual Dietary Intake

Downregulation of SLC26A2 and SLC26A3 suggests that when oxalate is encountered in the diet, less dietary oxalate would be absorbed by patients with IBD, leading to more oxalate retained within the intestines for excretion with stool. We measured the stool oxalate content via untargeted metabolomics in patients with NIBD and CD who responded to a dietary survey measuring habitual dietary oxalate intake (Diet History Questionnaire III [DHQ3]; NIBD, n = 4; CD, n = 11). Higher stool oxalate content is observed in patients with CD despite no difference in habitual dietary oxalate intake (Figure 3A). Given the small number of DHQ3 respondents, we used DNA metabarcoding to measure the diversity of plant intake between patients with NIBD and CD in another cohort. The Shannon diversity of dietary plants was similar, suggesting that dietary intake of plants (oxalate-containing foods) is similar between patients with NIBD and those with CD (NIBD, n = 14; CD, n = 17) (Figure 3B).

Figure 3.

Clinical outcomes. (A) Dietary oxalate intake measured by DHQ3 vs stool oxalate content in patients with CD and NIBD (NIBD, n = 4; CD, n = 11). (B) Dietary plant intake measured by metabarcoding in patients with CD and NIBD (NIBD, n = 14; CD, n = 17). (C) Expression of SLC26A6 from bulk RNA-seq in uninflamed colon by Montreal class for behavior of disease (nonstricturing B1, n = 32; B3, n = 18; stricturing B2, n = 49; B2B3, n = 32). (D) Percentages of patients with CD who have stricturing disease or nonstricturing disease in low-SLC26A6 and high-SLC26A6 CD subsets. Percentages are calculated within each SLC26A6 expression group (low or high) and represent the distribution of disease phenotypes within each group (summing to 100% per group).

Low SLC26A6 Expression Associates With Stricturing Disease Behavior in Crohn’s Disease

We hypothesized that impaired expression of oxalate transporters would be associated with poor clinical outcomes in IBD. Lower colonic expression of SLC26A6 coincided with stricturing behavior of disease in patients with CD (Figure 3C). CD samples were stratified into low and high expressers based on the mean expression within the CD group and were evaluated for associations with Montreal classification. The majority of patients with CD (74.5%) categorized as low SLC26A6 have stricturing disease (Figure 3D). These results suggest that the expression of SLC26A6 may associate with disease behavior and severity.

Oxalate Supplementation Increases the Severity of Dextran Sodium Sulfate–Induced Colitis

To test whether dietary oxalate influences intestinal inflammation more directly, we fed male C57BL/6J (B6) mice an oxalate-supplemented or standard diet, along with drinking water containing ±2.5% dextran sodium sulfate (DSS) for 7 days (injury window), followed by normal drinking water for 5 days (recovery window). The mice whose diet was supplemented with oxalate and 2.5% DSS were 60% less likely to survive than all other groups (χ² P = .0104) (Figure 4A). Oxalate-supplemented diets and 2.5% DSS resulted in greater weight loss by day 3, which continued through the recovery period (Figure 4B), and shorter colon length (Figure 4C). Increased expression of the proinflammatory cytokine, Il1b, was observed with DSS drinking water during the injury phase relative to the standard diet, standard drinking water group (Figure 4D). The expression of Il1b remained higher during the recovery phase in mice that received oxalate with DSS, suggesting that inflammation persisted in the oxalate with DSS drinking water group relative to the standard diet with DSS drinking water group throughout the recovery phase (Figure 4D). A similar trend was observed for Tnfa expression (Figure 4D). Even oxalate with regular drinking water caused some barrier damage during the injury window (Figure 4E). We measured serum lipopolysaccharide (LPS) binding protein during injury to further validate changes in barrier injury (Figure 4F). These results suggest that early epithelial damage conferred by the oxalate supplemented diet may increase the severity of chemically induced intestinal epithelial barrier damage, leading to increased disease severity marked by changes in weight and survival.

Figure 4.

Reduced survival to DSS colitis observed with dietary oxalate. (A) Survival curve in mice ± oxalate ± DSS drinking water. (B) Percent weight change during the injury and recovery windows. (C) Colon lengths. (D) Expression of Il1b and Tnfa in mouse colon measured by quantitative polymerase chain reaction. (E) Histopathology scores reviewed by 2 blinded clinical pathologists. The total score represents the sum of epithelial barrier damage and inflammation observed in each tissue (0–12). (F) Serum LPS binding protein. Asterisks (∗) denote differences compared with standard + water, whereas pound signs (#) denote differences compared with standard + DSS.

Dietary Oxalate Accelerates Colitis Onset in Genetically Susceptible Models of Spontaneous Colitis

Dietary oxalate accelerated the onset of colitis in 2 independent spontaneous models of colitis, namely CC011/Unc/J (CC011) and B6.129P2-Il10tm1Cgn/J (Il10^−/−) mice, with increases in epithelial barrier damage and immune infiltration (Figure 5A). Mild inflammation was observed in B6 mice in response to the oxalate-supplemented diet (Figure 5A). These results suggest that baseline differences in epithelial barrier integrity are contributory to increased inflammation in genetically susceptible hosts exposed to dietary oxalate.

Figure 5.

Earlier onset of spontaneous colitis observed with dietary oxalate. (A) Histopathology scores reviewed by 2 blinded clinical pathologists (n = 5 mice/sex/strain/diet). The total score represents the sum of epithelial barrier damage and inflammation observed in each tissue (0–12). (B) Expression of Slc26a2, Slc26a3, and Slc26a6. Data are described as fold change relative to B6 for each diet (n = 5 mice/sex/strain/diet). (C) Fecal water oxalate content measured by cage (n = 2 cages/strain/diet).

Reduced Intestinal Oxalate Transporter Expression in Genetically Susceptible Models of Spontaneous Colitis

Slc26a2 and Slc26a3 were inherently downregulated in Il10^−/− and CC011 mice relative to B6 mice, irrespective of diet assignment (Figure 5B). Downregulation of these transporters suggests that less oxalate is absorbed from the oxalate-supplemented diet in Il10^−/− and CC011 mice than in B6 mice. Slc26a6 was upregulated in CC011 mice relative to Il10^−/− mice, irrespective of diet (Figure 5B). This transporter has been associated with both absorption and secretion of oxalate back into the lumen.^12,21,22

Increased Fecal Oxalate Is Associated With Disease Severity in Genetically Susceptible Models of Spontaneous Colitis

More oxalate was detected in the feces from cages of CC011 and Il10^−/− mice than from B6 mice (Figure 5C), with CC011 mice manifesting the worst colitis symptoms and highest oxalate in their stool (Figure 5A and C). These findings implicate decreased expression of oxalate transporters Slc26a2 and Slc26a3 in genetically susceptible models of spontaneous experimental colitis with concomitant increases in stool oxalate content and an accelerated onset of intestinal inflammation.

Oxalate Enhances Innate Immune Responses in Macrophages and Dendritic Cells

Oxalate has previously been shown to stimulate inflammation via NLRP3-mediated interleukin (IL)1b secretion in dendritic cells.²³ We isolated bone marrow–derived macrophages (BMMs) and dendritic cells (BMDCs) and intestinal epithelial cells (IECs) from B6 mice and exposed the cells to either untreated media or media supplemented with and without LPS and oxalate for 24 hours. Oxalate exaggerated inflammatory responses in both macrophages and dendritic cells (Figure 6A). Representative Western blot images were chosen to demonstrate that oxalate, LPS, and oxalate with LPS stimulate pro-IL1b and IL1b secretion in dendritic cells (Figure 6B). We measured lactate dehydrogenase (LDH) as a marker of cell death or tissue injury²⁴ and showed that there was no direct cytotoxic effect of oxalate on the intestinal epithelial cells (Figure 6C). These data suggest that oxalate contributes innate immune responses.

Figure 6.

Oxalate is a factor contributing to innate immune responses in cells relevant to IBD. (A) IL1B measured by enzyme-linked immunosorbent assay in supernatants macrophage and dendritic cell cultures (n = 6 cultures of primary cells pooled from 2–3 mice/strain). (B) Representative Western blots from dendritic cells depicting pro-IL1β and mature IL1β production from the exposure to oxalate with and without LPS. (C) Lactate dehydrogenase (LDH) released in intestinal epithelial cells measured at optical density (OD) 490 nm. LDH-positive control is included for reference.

Discussion

To our knowledge, this is the first study to implicate a specific dietary component, oxalate, as a potential exacerbating factor in the pathogenesis and clinical trajectory of IBD. Intestinal oxalate transporters SLC26A2, SLC26A3, and SLC26A6 have been implicated to varying degrees in barrier integrity, GI inflammation, and diarrhea.^12,13,18,20,25, 26, 27, 28, 29, 30, 31, 32 The expression and function of SLC26 family members is altered by inflammation.^13,17,18,33

SLC26A2 is an anion exchanger capable of chloride, sulfate, and oxalate transport, although its role in oxalate handling is less well-defined than SLC26A3 and SLC26A6. Expression has been proposed as a therapeutic target for patients with UC by others, who suggested that SLC26A2 negatively regulates inflammation- and immune-related pathways.¹⁹ Downregulation of SLC26A2 has also been associated with reduced oxalate absorption in patients with UC, highlighting the potential contribution to oxalate absorption.^12,26 Mechanistic studies of this transporter and oxalate absorption capacity have been conducted in oocytes.¹²

SLC26A3 is a major transporter for chloride and bicarbonate exchange and plays an important role in sodium chloride absorption in the colon and distal ileum. It has been implicated in chloride diarrhea and is suspected to be functionally linked to diarrhea phenotypes in IBD.^28,30,34 The loss of SLC26A3 expression and protein has been observed in colonic tissue from patients with UC with active disease.^28,34,35 This transporter has also been implicated in the maintenance of epithelial barrier function, mucus layer integrity, and epithelial differentiation.^17,31,32 Mechanistic studies have demonstrated the SLC26A3 mediates a substantial portion (∼40%–60%) of oxalate absorption in the ileum and colon, with corresponding decreases to serum and urinary oxalate content.^12,15,36 The expression has also been highlighted as a potential therapeutic target for UC by others.¹⁸

SLC26A6 mediates chloride and oxalate exchange and contributes to oxalate secretion. SLC26A6 was recently directly linked to gut microbial dysbiosis, compromised barrier integrity, and inflammation.²⁹ Mechanistic studies have implicated SLC26A6 in oxalate secretion by the duodenum and distal ileum.^12,21,37 However, the mouse ortholog may be more efficient at secreting oxalate in than the human ortholog, which may explain the inconsistencies we observed between mouse and human SLC26A6 expression.^12,38

In our studies, we provide evidence that oxalate transporters are consistently downregulated across all subtypes of IBD, particularly during active disease. We also showed that there is a functional difference in the metabolic fate of dietary oxalate in patients with CD despite variable, habitual intake of dietary oxalate and no differences in diversity of dietary plant taxa. The observations made in patients with CD were well-supported by the experimental colitis studies in mice. Oxalate-supplemented diets resulted in reduced survival to DSS-induced colitis, and reduced transporter expression corresponded to increased stool oxalate content in CC011 and Il10^−/− mice. CC011 mice fed an oxalate-supplemented diet had the highest stool oxalate content and the highest disease activity measured by histopathology. A further increase in human samples, for which there is availability of dietary intake measurements, mucosal gene expression, and stool metabolite measurements, will enhance our understanding of the true effect of reduced expression of SLC26A2, SLC26A3, and SLC26A6 on the metabolic fate of dietary oxalate in humans.

These studies also suggest that the expression of oxalate transporters might serve as biological signals predicting the response to dietary oxalate and the susceptibility to oxalate-induced intestinal inflammation. In paired samples from patients with UC, we demonstrated that, within an individual, expression of each transporter is even lower in inflamed tissue. This finding increases the likelihood that people with IBD may be more susceptible to dietary oxalate during active disease. These observations were well-supported by the experimental colitis study in CC011 and Il10^−/− mice, where these 2 models of spontaneous colitis also showed baseline differences in expression of oxalate transporters relative to B6 mice and irrespective of diet assignment. When we examined the expression of oxalate transporters in patients with CD, lower colonic expression of SLC26A6 coincided with developing fibrostenotic disease, suggesting that expression of the transporters might be predictive of clinical trajectory. Patients with ileal CD and particularly those requiring a bowel resection are at risk of enteric hyperoxaluria; however, the RNA-seq datasets analyzed here were obtained preoperatively, limiting confounding from postoperative effects. However, these findings are considered exploratory and will require validation in larger cohorts enriched for fibrostenotic disease.

Oxalate has been linked to systemic and localized inflammation and oxidative stress.^23,25,39, 40, 41, 42 Using an ex vivo system, we confirmed previous observations that oxalate enhances LPS-induced IL1β secretion by monocytes.²³ This interaction between oxalate and LPS is biologically relevant, given that the intestinal environment is continually exposed to microbially derived LPS and intermittently exposed to dietary oxalate.⁴¹ We did not observe a direct effect of oxalate on IECs in this culture system, despite the pronounced epithelial barrier disruption observed in mouse models. This suggests that oxalate-induced immune activation may exacerbate pre-existing epithelial barrier damage or contribute to disease progression in vivo indirectly. Further studies will be needed to explore the mechanisms impacting oxalate on specific IEC subtypes and to understand its relationship to nascent vs ongoing epithelial injury with subsequent repair. Additionally, alterations in the gut microbiota can independently compromise epithelial integrity.^10,43 Such changes may result directly from oxalate’s presence in the gut lumen or indirectly via the increased production of microbial oxalate degradation byproducts.^10,44

Although humans lack the enzymes necessary to metabolize oxalate, microbes can degrade oxalate to protect hosts against renal oxalate toxicity.^10,16,44, 45, 46, 47 However, the direct impact of intestinal oxalate on the growth of microbes is not fully characterized.^10,44 The excess epithelial barrier damage we observed in vivo may be, in part, the result of a microbially mediated epithelial barrier injury due to changes in the overall composition of the microbiome or a byproduct of microbial degradation of oxalate. Oxalate metabolizing microbes include pathogens like Eshericia coli as well as commensal bacteria like Oxalobacter formigenes, Lactobacillus, and Bifidobacterium.^10,44, 45, 46 O formigenes is reduced among patients with IBD, and this has been shown to coincide with enrichment for E coli, increased stool oxalate, and higher fecal calprotectin.^10,46, 47, 48, 49 One of the possible byproducts of oxalate metabolism by microbes is formate, which has been shown to increase growth of E coli in the setting of inflammation.¹⁰ E coli colonization is associated with barrier dysfunction.⁴³ This is one example of how increased intestinal oxalate content may lead to a microbially mediated barrier defect and warrants future exploration of diet–host–microbiome interactions as they relate to dietary oxalate.

The gut microbiome represents a promising target for novel therapeutic strategies aimed at enhancing the degradation of dietary oxalate safely, rather than increasing the absorption by targeting the transporters or relying solely on dietary restriction. Increased absorption would be detrimental to the renal system, and dietary restrictions are often difficult to sustain, highlighting the potential advantages of future, microbiome-based approaches. Other opportunities for future mechanistic studies and novel therapeutics exist in understanding the relationship of oxalate to other micronutrients in this disease context. Oxalate is known to bind to other minerals like calcium and magnesium.¹¹ Whether the effects of dietary oxalate are mediated through another micronutrient-associated mechanism remains to be determined.

The typical oxalate intake is considered to be 100–150 mg/day.^48,50, 51, 52 Individual food choices and dietary preferences can lead to high or excessive intake of dietary oxalate, even when specific oxalate-containing foods are not consumed in excess.⁵⁰ Consequently, any otherwise healthful dietary pattern may be confounded by elevated oxalate intake or poor physiological response to dietary oxalate even at normal or moderate levels of consumption. It is possible to achieve a healthful, plant-based dietary pattern while simultaneously reducing or restricting total oxalate intake.^53,54 Adherence to a low oxalate diet might still be challenging, and other therapeutic avenues like microbiome-based approaches warrant future studies into disturbed oxalate transport, dietary oxalate intake, and disease outcomes.

Limitations

Several limitations should be acknowledged. First, our human data are primarily hypothesis-generating and derived from a relatively small sample size, which limits statistical power and generalizability. However, the observed associations are supported by complementary evidence from controlled mouse experiments. This strengthens the biological relevance and highlights promising future directions for investigations in humans.

Second, habitual dietary intake was not temporally linked to the stool samples analyzed. To address this limitation, we included molecular detection of dietary components through metabarcoding and measurement of the diversity of plant taxa in a larger cohort. Although not a perfect alternative for temporally matched dietary records and being unable to directly evaluate oxalate intake, this represents one of the most direct and objective alternatives available for assessing dietary intake at the time of sample collection currently available.

Third, mouse models require diets with relatively high oxalate content to illicit measurable physiological responses. This is consistent with prior literature citing organismal differences in microbiota and clearance thresholds, and our approach reflects a common experimental strategy to model pathophysiological events, rather than dietary oxalate load. Nevertheless, these experimental diets may not fully recapitulate the typical human exposure levels, underscoring the need for future studies directly measuring oxalate intake and outcomes in humans.

Conclusions

Experimental colitis models support a functional role for dietary oxalate in amplifying intestinal inflammation and are consistent with observations in human cohorts, demonstrating altered oxalate handling and the reduced expression of key intestinal oxalate transporters. This study identifies modifying dietary oxalate and/or the metabolic fate of dietary oxalate as potential, novel therapeutic areas for IBDs and the expression of SLC26A2 and SLC26A3 as potential disease-associated, biological signals predicting response to dietary oxalate. More broadly, this work highlights diet as a measurable and biologically actionable component of IBD pathogenesis and supports the incorporation of dietary exposures into translational research aimed at developing diagnostic and precision nutrition approaches. Future research should focus on validation in larger, longitudinal patient cohorts integrating stool oxalate measurements, rigorously captured dietary intake, molecular profiling, and microbiome analyses. Such efforts will be essential to define the clinical relevance of oxalate metabolism in IBD and to determine whether targeted dietary or metabolic interventions can be rationally implemented prior to formal clinical recommendations regarding oxalate intake.

Materials and Methods

Study Approval

This study was approved by the University of North Carolina Institutional Review Board (19-2519, 15-0024, 10-0355). All animals were maintained in accordance with the University of North Carolina University Institution Animal Care and Use Committee guidelines (IACUC protocol number: 23-211) and Texas A&M University IACUC guidelines (IACUC protocol number: 2022-0273). At the end of each feeding trial, mice were euthanized by carbon dioxide asphyxiation with a secondary procedure, blood was collected, and harvested tissues were immediately flash-frozen in liquid nitrogen.

The study design and experimental procedures are described below and are represented in Figure 1.

Patient Samples

Study participants were recruited through the University of North Carolina, Chapel Hill Health System and consented to stool and biopsy collection prior to bowel resection. Colonic and ileal mucosa were obtained from patients with an established diagnosis of IBD (CD or UC) or NIBD controls receiving screening colonoscopies between April 2012 and May 2024. This study includes 3 primary RNA-seq cohorts (UC, CD ileum, CD colon) and contemporaneously processed NIBD controls. All CD colon samples were collected from disease-unaffected regions without macroscopic inflammation. The CD ileum and UC colon samples were collected from both disease-affected (inflamed) and disease-unaffected (uninflamed) regions as we were able. The UC colon uninflamed and inflamed samples are paired for individuals.

Clinical phenotyping captured demographic and clinical variables including age, sex, disease duration, age at diagnosis, age at sample acquisition, disease location, and disease behavior. Montreal scores were determined by extracting information from endoscopy reports. We grouped the Montreal B1 (inflammatory) and B3 (penetrating) groups together to form the “nonstricturing” classification and the Montreal B2 (stricturing) and B2B3 (stricturing and penetrating) groups together to form the “stricturing” classification. The demographic information for patient samples is summarized in Table 1. In addition to transcriptomic profiling, subsets of patients contributed stool samples and dietary intake data. However, these datasets were not uniformly available across all individuals, and this precluded integration of all data types. Accordingly, each dataset was analyzed and interpreted as orthogonal to the transcriptomic analyses.

Table 1.

Clinical Characteristics at the Time of Collection

Characteristics	NIBD	UC	P value
Cohort of patients with UC (colon)
Age at sampling, y	59.5 (47.5–72)	38 (28.5–54)	.0002
Female sex	9 (50)	16 (40)	.571
Treatment history
Steroids	3 (16.7)	40 (100)	<.0001
5-ASA	1 (5.6)	38 (95)	<.0001
Immunomodulators	1 (5.6)	25 (62.5)	<.0001
Anti-TNF	0 (0)	36 (90)	<.0001
Other biologics	0 (0)	18 (45)	.0004
Small molecules	0 (0)	4 (10)	.3
Other diagnoses
Hypertension	11 (61.1)	10 (25)	.0166
Diabetes	7 (38.9)	3 (7.5)	.0067
Hyperlipidemia	6 (33.3)	4 (10)	.0555
Smoking status
Never	9 (50)	26 (65)	.385
Former	5 (27.8)	12 (30)	1
Current	4 (22.2)	2 (5)	.0679
Disease extent
Montreal extent, E1	NA	1 (1.7)	–
Montreal extent, E2	NA	10 (17.2)	–
Montreal extent, E3	NA	29 (50)	–

Cohort of patients with CD (ileum)	NIBD	CD
Age at sampling, y	57 (49.5-65.25)	42 (29-51)	.0002
Female sex	9 (32.1)	58 (55.8)	.0334
Treatment history
Steroids	4 (14.3)	88 (84.6)	.0001
5-ASA	0 (0)	53 (51)	.0001
Immunomodulators	0 (0)	67 (64.4)	.0001
Anti-TNF	0 (0)	79 (76)	.0001
Other biologics	0 (0)	26 (25)	.0012
Small molecules	0 (0)	0 (0)	1
Other diagnoses
Hypertension	9 (32.1)	21 (20.2)	.207
Diabetes	1 (3.6)	2 (1.9)	.514
Hyperlipidemia	7 (25)	10 (9.6)	.0514
Smoking status
Never	18 (64.3)	53 (51)	.286
Former	7 (25)	31 (29.8)	.814
Current	3 (10.7)	20 (19.2)	.404
Disease extent
Behavior, B1	NA	3 (2.9)	–
Behavior, B2	NA	48 (46.6)	–
Behavior, B3	NA	16 (15.5)	–
Behavior, B2B3	NA	36 (35)	–
Perianal disease	NA	15 (14.4)	–

Cohort of patients with CD (colon)	NIBD	CD
Age at sampling, y	54 (45-67)	40 (28.5-52)	.0001
Female sex	50 (61.7)	77 (58.8)	.773
Treatment history
Steroids	8 (9.9)	120 (91.6)	.0001
5-ASA	1 (1.2)	98 (74.8)	.0001
Immunomodulators	3 (3.7)	95 (72.5)	.0001
Anti-TNF	0 (0)	102 (77.9)	.0001
Other biologics	0 (0)	26 (19.8)	.0001
Small molecules	0 (0)	0 (0)	1
Other diagnoses
Hypertension	41 (50.6)	25 (19.1)	.0001
Diabetes	13 (16)	7 (5.3)	.0143
Hyperlipidemia	31 (38.3)	13 (9.9)	.0001
Smoking status
Never	40 (49.4)	73 (55.7)	.397
Former	25 (30.9)	38 (29)	.877
Current	16 (19.8)	20 (15.3)	.453
Disease extent
Behavior, B1	NA	32 (24.4)	–
Behavior, B2	NA	49 (37.4)	–
Behavior, B3	NA	18 (13.7)	–
Behavior, B2B3	NA	32 (24.4)	–
Perianal disease	NA	37 (28.2)	–

NOTE. Continuous variables are presented as median (interquartile range) and compared with Wilcoxon rank-sum test. All other variables are presented as number (%) and compared with Fisher’s exact test.

Rows with – indicate descriptive-only variables without statistical testing.

5-ASA, 5-aminosalicylic acid; CD, Crohn’s disease; NIBD, noninflammatory bowel disease; TNF, tumor necrosis factor; UC, ulcerative colitis.

Dietary Intake and Stool Oxalate Content

We measured habitual dietary intake with the validated DHQ3.55, 56, 57, 58 As a secondary measure of dietary intake, we used metabarcoding to molecularly determine the plant composition of diets from stool samples.59, 60, 61 For stool oxalate content in humans, metabolomic profiling was performed by Metabolon, Inc using their global untargeted metabolomics platform. Samples were processed using standard protocols developed by Metabolon. The normalized peak area ratio for oxalate was used for stool oxalate analyses (Chemical ID: 100,000,841).

RNA Isolation, Sequencing, and Processing

For quantitative polymerase chain reaction of mouse and human tissues, RNA from flash-frozen colon tissue was isolated using the Norgen Total RNA Purification kit (Norgen; 48300). Complementary DNA (cDNA) was generated using the High Capacity cDNA Reverse Transcription kit (Thermo Fisher; 4368814). Quantitative polymerase chain reaction analyses were performed on a QuantStudio 3 Real Time PCR (Thermo Fisher) with Powerup SYBR Green Master Mix (Applied Biosystems; A25742). Actb was used as a housekeeping gene to correct for starting amounts of cDNA. Primers were purchased from Integrated DNA technologies as custom DNA oligos, and sequences are provided in Table 2.

Table 2.

Primer Sequences

Gene symbol	Forward	Reverse
Actb	AGCCATGTACGTAGCCATCCAG	TGGCGTGAGGGAGAGCATAG
Il1b	ACGGACCCCAAAAGATGAAG	TTCTCCACAGCCACAATGAG
Slc26a2	AACCATACATTAGACGGACTCTG	AGCCCATTGCTACCTGATAAA
Slc26a3	GAGGAGTTTAAGAAGACGCACAG	TCTGTGAGGAGCAGCTACAA
Slc26a6	TTGGTACTGGTGAAGCTACTTAAT	CCCAATGAGCGTGAGTAGTT
Tnfa	GCCTCTTCTCATTCCTGCTTG	CTGATGAGAGGGAGGCCATT

RNA-seq was performed on the Illumina NovaSeq, NovaSeq6000, HiSeq2500, HiSeq4000, or MGI Tech DNBSEQ-T7. Bulk RNA-seq data were analyzed with modifications. Covariates, including patient sex, year of sample collection, batch, and transcript integrity number scores, were accounted for using the removeBatchEffect() function from the limma package. Differential gene expression analysis was then performed using DESeq2, correcting for the same covariates. Genes with a false discovery rate–adjusted P value (Padj) < .05 were considered differentially expressed. Normalized counts were obtained using DESeq2 and transformed using the variance-stabilizing transformation (vs) with a design matrix including patient sex, year of sample collection, batch, and transcript integrity number.

Patient Population Stratification

In a secondary, exploratory analysis, samples from patients with CD were classified as “low expressers”’ or “high expressers” based on the mean expression of oxalate transporters. This mean expression value within the CD cohort was used as a threshold, with samples below the mean classified as low expressers and those at or above the mean classified as high expressers. This classification was derived exclusively within the CD group; no expression-based subgrouping was performed in NIBD samples. Expression-based groups were subsequently linked to clinical disease characteristics using the Montreal classification system. This approach was used to enable exploratory stratification within the CD cohort and was not intended to define a clinical cutoff.

Metabarcoding

Fecal samples were transferred to a 2-mL tube containing 200 mg of 106- to 500-μm glass beads (Sigma) and 0.5 mL of Qiagen PM1 buffer. Mechanical lysis was performed for 20 minutes on a digital vortex mixer. After 5 minutes of centrifugation, 0.45 mL of the supernatants was aspirated and transferred to a new tube containing 0.15 mL of Qiagen IRS solution. The suspension was incubated at 4 °C overnight. After 5 minutes of centrifugation, the supernatant was aspirated and transferred to deep-well plates containing 0.45 mL of Qiagen binding buffer supplemented with Qiagen ClearMag Beads. The DNA was purified using the automated KingFisher Flex Purification System and eluted in DNase-free water. Metabarcoding was performed by established methods.59, 60, 61 Age, sex, and comorbidities identified in the demographics tables were considered during analyses.

Animals

Dextran sodium sulfate–induced colitis model

Five-week-old male B6 mice were obtained from The Jackson Laboratory and acclimated to experimental diets for 2 weeks prior to the introduction of DSS. B6 mice are resistant to developing experimental colitis in the absence of a chemical stimulus. Colitis-grade DSS salt hydrate was purchased from Sigma-Aldrich (A6419). Mice were assigned to either a standard diet (n = 20) or an oxalate-supplemented diet (n = 20). Following dietary acclimation, mice received either 2.5% DSS in drinking water or control drinking water for 7 days (B6 standard + DSS, n = 10; B6 oxalate + DSS, n = 10; B6 standard + water, n = 10; B6 oxalate + water, n = 10). After 7 days, one-half of the mice were sacrificed during the acute injury phase, whereas the remaining mice were allowed to recover on normal drinking water for an additional 5 days.

Spontaneous colitis models

Three- to 5-week-old CC011 mice were obtained from the Systems Genetics Core at the University of North Carolina. Il10^−/− and B6 mice were obtained from The Jackson Laboratory. CC011 and IL10^−/− mice are established models of spontaneous colitis; however, the immunopathology underlying disease in CC011 mice remains incompletely characterized. Mice were assigned to either a standard diet or an oxalate-supplemented diet (CC011 standard, n = 10; CC011 oxalate, n = 10; IL10^−/− standard, n = 10; IL10^−/− oxalate, n = 10; B6 standard, n = 10; B6 oxalate, n = 10) and maintained on experimental diets for 10 to 16 weeks. One-half of the mice in each group were male (n = 5 mice per sex, strain, and diet).

Phenotyping and outcome measures

Body weights were measured prior to dietary intervention and monitored daily during DSS experiments or at 2, 5, and 10 weeks during spontaneous colitis experiments. Histopathologic evaluation was performed by 2 blinded clinical pathologists using established scoring criteria. Total histology scores (0–12) reflected epithelial barrier damage (0–6) and inflammation (0–6). Serum LPS binding protein was quantified using a commercial enzyme-linked immunosorbent assay kit (Abcam; ab269542). Fecal samples were collected at the final timepoint in spontaneous colitis to quantify fecal water oxalate content using methods described by Jiang et al. Fecal water oxalate concentrations were measured using a colorimetric assay (Trinity Biotech; 591-D) according to the manufacturer’s instructions.

Experimental Diets

Pelleted diets were designed in collaboration with Research Diets Incorporated. The 1% oxalate supplemented diet (D21092702) differs from the standard diet (D17031601) only by the incorporation of oxalate. All other dietary components are controlled between the 2 diets. In line with previous literature, oxalate was provided as sodium oxalate, which allows dietary oxalate to interact with other minerals in the feed like calcium, magnesium, and iron during digestion and metabolism.¹¹ The detailed diet compositions are provided in Table 3. The oxalate-supplemented diet developed for our mouse experiments was adapted from the 1% oxalate diet used by Liu et al,¹⁰ which demonstrated less severe renal outcomes compared with those reported by Hatch et al in 2006⁴⁶ and 2011⁴⁷ using 1.5% oxalate supplementation. These high concentrations have historically been employed to overcome species-specific differences in oxalate absorption and metabolism for mechanistic investigation.^{10,12,15,21,38,46,47,}62, 63, 64

Table 3.

Diet Compositions

Product	Standard	Oxalate
	D17031601	D21092702
Ingredients	gm	gm
Casein	188	187
L-Cystine	3	3
Corn starch	463	458
Maltodextrin	132	131
Cellulose, BW200	71	70
Inulin	24	23
Soybean oil	66	65
t-BHQ	0	0
Mineral mix S10026	9	9
Dicalcium phosphate	12	12
Calcium carbonate	5	5
Potassium citrate, 1 H2O	16	15
Vitamin mix V10001	9	9
Choline bitartrate	2	2
Sodium oxalate	0	10
FD&C yellow dye #40	0	0
FD&C red dye #5	0	0
FD&C blue dye #1	0	0
Total	1000	1000
Nutrients, g
Protein	167	165
Carbohydrate	538	532
Fat	68	68
Cholesterol	0	0
Fiber	88	88
Nutrients, g%
Protein	16	16
Carbohydrate	51	50
Fat	6	6
Cholesterol	0	0
Fiber	8	8
Sodium oxalate (added)	0	1
Calcium (total)	1	1
Phosphorous (‘available’/nonphytate)	0	0
Nutrients, kcal
Protein	667	661
Carbohydrate	2150	2129
Fat	614	608
Total	3431	3398
Nutrients, kcal%
Protein	19	19
Carbohydrate	63	63
Fat	18	18

Cell Culture Studies

Six- to 12-week-old male and female B6 mice were provided by Drs June Brickey and Jenny Ting for all cell culture studies. Bone marrow–derived monocytes were isolated, counted, and plated in 6 wells at 9.9 × 10⁵ cells/well. Monocytes were cultured for 6 days in complete media supplemented with macrophage colony-stimulating factor (M-CSF) or granulocyte-macrophage colony-stimulating factor (GM-CSF) to promote differentiation to BMMs and BMDCs respectively. M-CSF was provided in conditioned media from a previous cultured NCTC clone of strain L (L929). L929 secretes M-CSF among other growth factors.⁶⁵ Conditioned media was provided with growth medium at a ratio of 1:10. GM-CSF was provided at 0.5 ng/mL from recombinant mouse GM-CSF (Thermo Fisher Scientific; RMGMCSF20). The exposure was conducted on day 6 when monocytes were morphologically distinct. BMMs and BMDCs were exposed to either LPS (from Escherichia coli O111:B4; Sigma; L4391), calcium oxalate (CaOx), LPS + CaOx, LPS + adenosine triphosphate, or untreated growth media to serve as a control. The monocytes were stimulated with 1 μg/mL LPS for 3 hours followed by 100 μg/mL CaOx or 1 mM adenosine triphosphate for an additional 21 hours. At 24 hours after stimulation, the supernatants and cell lysates were collected for downstream analyses. The intestinal crypts were counted and plated in Matrigel at 400 crypts/well. The crypts were proliferated for 6 days prior to passage and plated on collagen in 24 wells at 25,000 crypts/well in a 2-dimensional monolayer. The crypts were proliferated for an additional 6 days prior to exposure. The exposure was conducted using the methods as described above for BMMs.

Western Blotting

Post treatment, BMDCs were washed twice with ice-cold phosphate-buffered saline and lysed in cell lysis buffer (Thermo Fisher Scientific; 78510) containing 1:100 protease inhibitor cocktail (Sigma-Aldrich; P8340) and 1:100 phosphatase inhibitor cocktail (Sigma-Aldrich; P5726). The protein lysate concentration was determined by BCA Assay (Thermo Fisher Scientific; 23227). The lysate concentrations were equalized with additional lysis buffer containing phosphatase and protease inhibitors. The lysates were then mixed with 4× Laemmli buffer (Bio-Rad; 1610747) and 2-mercaptoethanol (Sigma-Aldrich; M3148) and boiled for 8 minutes. The samples (5 mg/lane) were resolved via sodium dodecyl sulfate-polyacrylamide gel electrophoresis using 10% polyacrylamide gels (Bio-Rad; 4561034) and transferred to polyvinylidene fluoride membranes (Bio-Rad; 1704156). The membranes were blocked for 1 hour at room temperature with 5% w/v bovine serum albumin (BSA; Boston BioProducts; P-753) reconstituted in Tris-buffered saline containing 0.1% Tween-20 (TBS-T) (Sigma-Aldrich; P9416). The membranes were incubated in primary antibodies diluted in 3% w/v BSA in TBS-T for 18 hours at 4 °C followed by 5 5-minute washes in TBS-T. The membranes were then incubated in secondary antibodies diluted in 3% w/v BSA in TBS-T for 1 hour at room temperature followed by an additional 5 5-minute washes in TBS-T. Chemiluminescent imaging was performed using luminol (Thermo Fisher Scientific; 34577) and a blot imager (Invitrogen; A44115). Densitometry was performed using the software ImageJ. The primary antibodies included IL1β, at 1:000 dilution (Proteintech; 26048-1-AP) and β-actin at 1:10,000 dilution (Proteintech; 60008-1-Ig). The secondary antibodies used were goat anti-rabbit IgG (H+L) horseradish peroxidase conjugate (Genesee Scientific; 20-303) and goat anti-mouse IgG (H+L) horseradish peroxidase conjugate (Genesee Scientific; 20-304).

Statistical Analyses

Continuous variables were assessed for normality using the Shapiro–Wilk test. Data that followed a normal distribution was analyzed using parametric tests (Welch 2-sample t test), whereas non-normally distributed and ordinal data were analyzed using nonparametric tests (Wilcoxon rank-sum with Holm–Bonferroni or Benjamini–Hochberg for multiple comparisons, Kruskal-Wallis rank sum with Dunn’s test for multiple comparisons). Histology scores and Montreal scores were treated as ordinal variables (age, behavior, extent) or categorical variables (location, upper GI, perianal disease). Outliers in human paired diet and stool data were identified using an interquartile range (IQR)-based approach (Q1 − 1.5 × IQR or Q3 + 1.5 × IQR) within each disease group and were excluded from analysis due to limited survey response. The fecal water oxalate content from the spontaneous colitis experiments is measured by cage and was limited to 2 samples per diet/strain, and a 1-way analysis of variance was performed using 10,000 permutations to assess group differences, given the limited sample size and to avoid assumptions of normality. A P value of less than .05 was considered statistically significant for all tests. A single asterisk indicates a P value less than .05, double asterisks indicate a P value less than .01, and triple asterisks indicate a P value less than .001, unless otherwise noted. Statistical analyses were conducted using R version 4.4.1 and JMP Pro version 17.2.0.

CRediT Authorship Contributions

Anna C. Salvador, PhD (Conceptualization: Lead; Data curation: Lead; Formal analysis: Lead; Investigation: Lead; Project administration: Lead; Writing – original draft: Lead; Writing – review & editing: Lead)

Zena Khaled (Investigation: Supporting; Writing – review & editing: Supporting)

Ayesh Awad (Data curation: Supporting; Formal analysis: Supporting; Writing – review & editing: Supporting)

Benjamin Huan (Data curation: Supporting; Writing – review & editing: Supporting)

Gwen Lau (Data curation: Supporting; Writing – review & editing: Supporting)

Sophie Silverstein (Data curation: Supporting; Writing – review & editing: Supporting)

David Weaver (Data curation: Supporting; Writing – review & editing: Supporting)

Lee-Ching Zhu (Formal analysis: Supporting; Writing – review & editing: Supporting)

Surekha Bantumilli (Formal analysis: Supporting; Writing – review & editing: Supporting)

Viguna Thomas (Investigation: Supporting; Writing – review & editing: Supporting)

Emanuele Baldassarri (Investigation: Supporting; Writing – review & editing: Supporting)

Jace McGovern (Investigation: Supporting; Writing – review & editing: Supporting)

Ezan Chaudhry (Investigation: Supporting; Writing – review & editing: Supporting)

Nathan VanLandingham (Investigation: Supporting; Writing – review & editing: Supporting)

Dorothy K. Superdock (Data curation: Supporting; Formal analysis: Supporting; Investigation: Supporting; Writing – review & editing: Supporting)

Brady Furey (Investigation: Supporting; Writing – review & editing: Supporting)

Gloria Hayun Lee (Data curation: Supporting; Formal analysis: Supporting; Writing – review & editing: Supporting)

Benjamin McMichael (Data curation: Supporting; Investigation: Supporting; Writing – review & editing: Supporting)

Samantha Hicks (Investigation: Supporting; Writing – review & editing: Supporting)

Erin C. Steinbach (Funding acquisition: Supporting; Methodology: Supporting; Resources: Supporting; Writing – review & editing: Supporting)

Matthew R. Schaner (Methodology: Supporting; Writing – review & editing: Supporting)

Jeremy Herzog (Methodology: Supporting; Writing – review & editing: Supporting)

Leslie Garry Adams (Investigation: Supporting; Writing – review & editing: Supporting)

W. June Brickey (Resources: Supporting; Writing – review & editing: Supporting)

Jenny P. Y. Ting (Funding acquisition: Supporting; Resources: Supporting; Writing – review & editing: Supporting)

Lawrence A. David (Data curation: Supporting; Funding acquisition: Supporting; Writing – review & editing: Supporting)

David W. Threadgill (Funding acquisition: Supporting; Resources: Supporting; Writing – review & editing: Supporting)

Terrence S. Furey (Data curation: Supporting; Resources: Supporting; Writing – review & editing: Supporting)

Shehzad Z. Sheikh (Funding acquisition: Lead; Investigation: Supporting; Methodology: Supporting; Resources: Lead; Supervision: Lead; Writing – original draft: Supporting; Writing – review & editing: Supporting)