Possible Involvement of the Mucosal Claudin Expression in the Healing of a Refractory Ileocecal Ulcer in a Patient with Crohn's Disease Successfully Treated with Infliximab

Abstract

An impaired intestinal barrier exposes the microbiota to gut immune cells, triggering cytokine responses that cause Crohn's disease (CD). Barrier dysfunction is a prerequisite for CD development. However, the effects of biological therapies on barrier integrity remain poorly understood. We aimed to present a case of CD refractory to adalimumab and risankizumab. Treatment with infliximab improved refractory ulcers, and this effect was associated with alterations in the mucosal mRNA expression of claudins, cell-cell adhesion molecules that contribute to the maintenance of barrier integrity. These findings suggested that infliximab induces CD remission by regulating the expression of claudin.

Introduction

Proinflammatory cytokine responses to intestinal bacteria underlie the immunopathogenesis of Crohn's disease (CD) (1-4). Gut antigen-presenting cells (APCs) are a major cellular source of pathogenic cytokines such as interleukin (IL)-12, IL-23, and tumor necrosis factor alpha (TNF-α) (1-4). In the healthy gut mucosa, the intestinal barrier functions as a boundary that prevents the invasion of bacteria into the intraepithelial and lamina propria immune systems. In CD, disruption of the intestinal barrier exposes gut immune cells to intestinal bacteria, leading to the excessive production of proinflammatory cytokines by APCs (1-4). Thus, intestinal barrier dysfunction is a prerequisite for CD development. Biological therapies targeting IL-12, IL-23, and TNF-α, all of which are produced by APCs, are widely used and have achieved remarkable success in CD treatment (1-4). Although the effects of biologics on cytokine responses are well established, their effects on the intestinal barrier integrity remain poorly understood.

Tight junctions (TJs) form barriers that prevent the passage of ions and molecules, including bacterial cell wall components (2,4,5). TJ proteins (TJPs) are cell-cell adhesion molecules that form a paracellular barrier and are located in epithelial intercellular spaces (2,4,5). Claudins, a major class of TJPs, play a critical role in maintaining paracellular barrier function (5). Claudins are classified into barrier-sealing or pore-forming types (2,4,5). Most claudins, except claudin-2 and claudin-15, support paracellular barrier integrity through their sealing function; therefore, the impaired expression of these claudins disrupts TJ integrity (2,4,5). In contrast, claudin-2 and claudin-15 mediate sodium and water transport by forming pores. The activation of claudin-2 and claudin-15 impairs TJ integrity by promoting sodium and water excretion into the intestinal lumen. Thus, the claudin function is closely linked to the maintenance of TJ integrity. Indeed, patients with CD display downregulation of barrier-sealing claudins and upregulation of pore-forming claudins in comparison to healthy individuals (6,7). These findings led us to hypothesize that biological therapies may induce remission in CD by regulating the expression of claudin. However, it remains unknown whether biologics induce remission of CD through this mechanism. Recently, we encountered a case of CD that was successfully treated with infliximab (IFX) after failure of risankizumab (RZB). Notably, improvement of the refractory ileocecal ulcer in this patient was associated with changes in the mRNA expression of claudins rather than cytokines.

Case Report

A 48-year-old man with CD was referred to our hospital after relocation. He was diagnosed with CD at 46 years of age based on the presence of longitudinal ulcers in the esophagus and ileocecal region. Treatment included oral 5-aminosalicylic acid (5-ASA; 2 g/day) and biweekly subcutaneous injections of adalimumab (ADA; 40 mg). Oral aphthae were present, but genital ulcers and uveitis were not observed. Human leukocyto antigen (HLA)-B51, the strongest HLA allele for Behçet's disease (BD) was negative (8,9). Based on these findings, we diagnosed the present case as CD rather than BD, although granuloma formation was not observed in the pathological examinations. It is worth noting, however, that HLA-A26, the second-strongest HLA allele for BD, was positive (8,9). Thus, we cannot completely exclude the possibility of BD in terms of HLA allele analyses. The clinical course following referral is shown in Fig. 1 (the month of referral is defined as month 0).

Figure 1.

The clinical course of this patient with Crohn’s disease. Month 0 was defined as the time when the patient was referred to Kindai University Hospital. Alterations in peripheral blood white blood cell (WBC) counts and serum levels of albumin (ALB) and C-reactive protein (CRP) are shown. ADA: adalimumab, 5-ASA: 5-aminosalicylic acid, CS: colonoscopy, EGD: esophagogastroduodenoscopy, IFX: infliximab, 6-MP: 6-mercaptopurine, PSL: prednisolone, RZB: risankizumab

Approximately eight months after referral, the patient developed stomatitis, diarrhea, and high-grade fever accompanied by elevated white blood cell (WBC) counts and serum C-reactive protein (CRP) levels. The patient was treated with prednisolone (PSL), starting at 30 mg/day and tapered by 5 mg biweekly. Following this episode, the ADA dose was increased to 40 mg per week. His clinical symptoms improved markedly after dose escalation, and mucosal healing of the esophagus and ileocecal region was observed on endoscopic examinations. However, the patient continued to experience frequent episodes of stomatitis, diarrhea, and fever. Each episode was associated with elevated WBC counts and CRP levels and required PSL treatment (Fig. 1).

Esophagogastroduodenoscopy (EGD) and colonoscopy (CS) were performed after the third episode (EGD#1 and CS#1 in Fig. 1). EGD revealed a white moss-covered ulcer extending from the esophagogastric junction in the esophagus (Fig. 2A) along with multiple ulcer scars in the lower esophagus. CS revealed a deep and longitudinal ulcer extending from the terminal ileum to the ileocecal valve (Fig. 3A). The colorectal mucosa remained intact. Polymerase chain reaction (PCR) testing of the biopsy samples for cytomegalovirus was negative. These clinical courses suggest secondary failure of ADA. The patient, with a body weight of 62 kg, did not take azathioprine due to nausea despite the NUDT15 Arg/Arg allele. Therefore, it is possible that the formation of anti-ADA antibodies may be involved in secondary failure.

Figure 2.

Endoscopic images of esophageal ulcers during treatment with biologics. Longitudinal esophageal ulcers extending from the esophagogastric junction to the lower esophagus during treatment with adalimumab (A) and risankizumab (B). Marked improvement in the esophageal ulcers was observed after treatment with infliximab (C).

Figure 3.

Endoscopic images of the ileocecal ulcer during treatment with biologics. Longitudinal and deep ulcers extending from the terminal ileum to the ileocecal valve during treatment with adalimumab (A) and risankizumab (B). Marked improvement of the ileocecal ulcer was observed after treatment with infliximab (C), and the huge ulcer become a small erosion.

Treatment with RZB, which targets IL-23p19, was initiated upon failure to control esophageal and ileocecal ulcers by ADA targeting TNF-α. The patient received intravenous RZB (600 mg) every 4 weeks for a total of three doses. However, despite RZB treatment, the patient developed stomatitis, diarrhea, and fever. EGD and CS were performed to assess the efficacy of RZB (EGD#2 and CS#2 in Fig. 1). A longitudinal and deep ulcer extended from the esophagogastric junction to the lower esophagus (Fig. 2B), indicating worsening of the ulcer following RZB treatment. Furthermore, no significant alterations were observed in ileocecal ulcers (Fig. 3B). As RZB treatment neither alleviated the patient's symptoms nor decreased the serum CRP levels, it failed to induce clinical remission in this case. In addition, a marked decrease in the serum concentration of albumin was observed during RZB treatment.

IFX has diverse immunoregulatory functions beyond the neutralization of TNF-α-mediated signaling pathways (10-13). IFX administration (5 mg/kg) was initiated to improve the clinical symptoms and refractory ulcers. The dosing administration included infusions at weeks 0, 2, and 6, followed by injections every 4 weeks. After five doses of IFX, improvement in clinical symptoms and normalization of serum CRP and albumin levels were observed (Fig. 1). EGD and CS were performed after six doses of IFX (EGD#3 and CS#3 in Fig. 1). EGD showed no active ulcer in the esophageal mucosa (Fig. 2C), and the large ileocecal ulcer reduced to a small erosion (Fig. 3C). Collectively, these findings demonstrated that CD-associated refractory esophageal and ileocecal ulcers were successfully treated with IFX after ADA and RZB failed. 6-mercaptopurine (6-MP, 30 mg/day) was introduced to avoid the formation of anti-IFX antibodies. No CD flares were observed after 11 doses of IFX, as assessed by clinical symptoms and blood examinations.

Methods

To investigate the molecular mechanisms underlying successful treatment with IFX after failure of ADA and RZB, endoscopic biopsy samples were obtained from the ileocecal mucosa adjacent to the ulcer or erosion. Quantitative reverse transcription PCR (qPCR) was performed as described previously (14,15). The mRNA levels of cytokines, chemokines, and TJPs were measured using SYBR Green-based qPCR on a LightCycler 480 system (Roche, Tokyo, Japan). The primer sequences for the target genes are listed in Table. The mRNA expression levels were normalized to ACTB as a reference gene. Ethical approval for this study was granted by the review board of the Kindai University Faculty of Medicine.

Table.

PCR Primer Sequences.

Target Gene	Primer Sequences
ACTB	F: TCCCTGGAGAAGAGCTACGA
	R: AGGAAGGAAGGCTGGAAGAG
CLDN1	F: GTCTTTGACTCCTTGCTGAATCTG
	R: CACCTCATCGTCTTCCAAGCAC
CLDN2	F: GTGACAGCAGTTGGCTTCTCCA
	R: GGAGATTGCACTGGATGTCACC
CLDN4	F: AGTGCAAGGTGTACGACTCGCT
	R: AGTGCAAGGTGTACGACTCGCT
CXCL10	Qiagen QuantiTect Primer Assays
	Sequence Unavailable
FOXP3	Qiagen QuantiTect Primer Assays
	Sequence Unavailable
IFNB	Qiagen QuantiTect Primer Assays
	Sequence Unavailable
IFNG	F: GTCCAACGCAAAGCAATACA
	R: CTCTTCGACCTCGAAACAGC
IL1B	F: GGACAAGCTGAGGAAGATGC
	R: TCGTTATCCCATGTGTCGAA
IL6	F: AAAGAGGCACTGGCAGAAAA
	R: CACCAGGCAAGTCTCCTCAT
IL10	F: TGCCTTCAGCAGAGTGAAGA
	R: CCCTTAAAGTCCTCCAGCAA
IL12B	Qiagen QuantiTect Primer Assays
	Sequence Unavailable
IL17A	F: CATGAACTCTGTCCCCATCC
	R: CCCACGGACACCAGTATCTT
IL23A	F: TCTCCTTCTCCGCTTCAAAA
	R: TTAGGGACTCAGGGTTGCTG
TJP1	F: GGGGCCTACACTGATCAAGA
	R: AGGACCGTCTAATGGCAGAC
TNFA	F: AACCTCCTCTCTGCCATCAA
	R: GGAAGACCCCTCCCAGATA

Results

No major differences were observed in the mRNA expression of prototypical pro-inflammatory cytokines, including IL1B (IL-1β), IL17A (IL-17), IL23A (IL-23p19), IFNB (IFN-β), IFNG (IFN-γ), and TNFA (TNF-α), regardless of whether the patient was treated with ADA, RZB, or IFX (Fig. 4A). The CXCL10 mRNA expression was lower in the ileocecal mucosa following IFX treatment than in the ADA- or RZB-treated mucosa, whereas the IL12B (encoding IL-12/23p40) expression was higher in the IFX-treated mucosa than in the ADA- or RZB-treated mucosa. IL-10, mainly produced by forkhead box P3 (FOXP3)-expressing regulatory T cells (Tregs), is an immunoregulatory cytokine that potently suppresses proinflammatory cytokine responses against the intestinal microbiota (3,16,17). The ileocecal mucosa treated with IFX showed higher IL10 and FOXP3 mRNA expression levels than the ADA- or RZB-treated mucosa.

Figure 4.

Alterations in the ileocecal mucosal mRNA expression of cytokines and tight junction proteins during treatment with biologics. mRNAs were prepared from the ileocecal mucosa surrounding the ileocecal ulcer or erosion and subjected to quantitative reverse transcription polymerase chain reaction. The relative mRNA expression of cytokines (A) and tight junction proteins (B) during treatment with adalimumab, risankizumab, and infliximab.

TJPs maintain intestinal barrier integrity and immune homeostasis (4,5). The mRNA expression of CLDN1 and CLDN2, which encode claudin-1 and claudin-2, respectively, was markedly lower in the IFX-treated ileocecal mucosa than in the ADA- or RZB-treated mucosa (Fig. 4B). In contrast, the mRNA expression of CLDN4 and TJP1 (encoding zonula occludens-1) was comparable across ADA-, RZB-, and IFX-treated samples. These findings suggest that healing of the ileocecal ulcer following IFX treatment was associated with the upregulation and downregulation of the mRNA expression of immunoregulatory molecules (IL10 and FOXP3) and claudins (CLDN1 and CLDN2), respectively.

Discussion

We report the case of a patient with CD who exhibited refractory ulcers in the esophagus and ileocecal region. Neither ADA nor RZB treatment resulted in mucosal healing at either site. However, IFX treatment led to significant improvement, reducing the size of the large ileocecal ulcer to a small erosion and promoting mucosal healing in the esophagus. These clinical outcomes indicate that IFX is highly effective for refractory CD. Both ADA and IFX target TNF-α, whereas RZB targets IL-23, thereby neutralizing proinflammatory signaling pathways (1,3,18). Therefore, we expected that the IFX-induced downregulation of proinflammatory cytokine responses may contribute to the improvement of esophageal and ileocecal mucosa. However, the mRNA expression of TNFA and IL23A did not differ among ADA-, RZB-, and IFX-treated mucosal samples, suggesting that IFX promoted mucosal healing through a cytokine-independent mechanism. Notably, IFX treatment markedly decreased the mRNA expression of CLDN1 and CLDN2 in the ileocecal mucosa in comparison to ADA or RZB treatment. The clinical course and the mRNA expression profiles support the idea that healing of esophageal and ileocecal ulcers is associated with downregulation of the claudin-1 and claudin-2 expression.

The physical barrier between intestinal epithelial cells is formed by TJs (2,4,5). The claudin family, which consists of 27 proteins, plays a major role in the formation of intercellular TJs located between adjacent epithelial cells (2,4,5,19). Most claudins possess barrier-sealing properties, whereas claudin-2 and claudin-15 exhibit pore-forming functions that facilitate cation and water transport (2,4,5). Intestinal barrier defects leading to increased epithelial permeability are recognized as risk factors for the development and relapse of CD (4). The colonic mucosa of patients with CD exhibits downregulation of barrier-sealing claudins (claudin-3, 5, 7, 8) and upregulation of pore-forming claudin-2 (6). Downregulation of the claudin expression increases macromolecular flux across the epithelium, whereas upregulation of pore-forming claudins allows the backflow of sodium and water into the intestinal lumen (7). In our mRNA study, IFX treatment reduced the mRNA expression of both pore-forming CLDN2 and barrier-sealing CLDN1. Regarding downregulation of the CLDN2 expression, we assume that the reduced expression of claudin-2 contributed to the resolution of the patient's symptoms, particularly diarrhea, by inhibiting sodium and water efflux into the intestinal lumen (19,20). Supporting this idea, Fries et al. reported that the resolution of experimental colitis with IFX was accompanied by a significant decrease in the expression of claudin-2, suggesting a pro-colitogenic role for claudin-2 activation (21). The significance of the IFX-induced downregulation of the CLDN1 mRNA expression in CD remains unclear. Weber et al. observed the increased expression of claudin-1 in the colonic mucosa of patients with CD (22), suggesting that the immune environment associated with active CD may influence the expression of claudin-1.

IFX mediates diverse immunoregulatory effects in the gastrointestinal tract of CD patients (10,11). In addition to neutralizing soluble TNF-α, IFX induces apoptotic cell death in CD4⁺ T cells expressing TNF receptor II by binding to membrane-bound TNF-α expressed on APCs (10,11). APCs expressing membrane-bound TNF-α are also subject to cell death via IFX-mediated complement-dependent and antibody-dependent cellular cytotoxicity (10,11). Furthermore, IFX promotes the expansion of FOXP3-expressing Tregs (12,13). The present case suggests an additional mechanism by which IFX may promote mucosal healing in CD through the downregulation of the mRNA expression of CLDN1 and CLDN2. We assume that IFX-induced the downregulation of the claudin-1 and claudin-2 expression contributes to mucosal healing in CD through maintenance of intestinal barrier integrity. It remains unknown whether downregulation of CLDN1 and CLDN2 mRNA is associated with the normalization of the intestinal barrier function. However, the improvement in the serum albumin concentration supports the restoration of the intestinal barrier by IFX.

It is well established that pro-inflammatory cytokine responses to intestinal microbiota underlie the immunopathogenesis of CD (1,3,18). ADA and IFX are biologics that neutralize TNF-α, whereas RZB neutralizes IL-23p19 (1,3,18). These agents suppress the activation of pro-inflammatory cytokine cascades triggered by TNF-α or IL-23. The successful treatment outcome with IFX led us to hypothesize that IFX effectively downregulates the mRNA expression of proinflammatory cytokines. However, the mRNA expression levels of IL23A and TNFA, which encode IL-23p19 and TNF-α, respectively, were comparable in the ileocecal mucosa treated with RZB or IFX. This suggests that the improvement of the ileocecal ulcer was not dependent on the suppression of IL-23- or TNF-α-mediated signaling pathways. In addition, the RZB-treated mucosa exhibited a marked reduction in the mRNA expression of IL12B, which encodes IL-12/23p40, despite a lack of improvement in the esophageal or ileocecal ulcers. These findings support the notion that the healing of the ileocecal ulcer is not mediated by the downregulation of colitogenic cytokines such as IL-12, IL-23, or TNF-α.

The IFX-treated ileocecal mucosa exhibited the increased mRNA expression of IL10 and FOXP3 in comparison to the ADA- or RZB-treated mucosa. Consistent with our findings, IFX treatment has been shown to increase the expression of FOXP3, thereby inducing clinical remission (12,13). Notably, the increase in the mRNA expression of IL10 and FOXP3 in the IFX-treated mucosa was relatively small. Taken together with the changes in the mRNA expression of proinflammatory cytokines, these results suggest that the regulation of proinflammatory or anti-inflammatory cytokine responses might not have been involved in the improvement of the ileocecal ulcer in this case.

The gastrointestinal tract of patients with active CD is characterized by the accumulation of T helper type 1 (Th1) and Th17 cells, which produce interferon-γ (IFN-γ) and IL-17, respectively (1). Nonetheless, biologics directly targeting IFN-γ (fontolizumab) and IL-17 (secukinumab) have failed to demonstrate clinical efficacy in CD (1,23). The molecular mechanisms underlying the failure of these agents remain largely unclear; however, the direct inhibition of Th1 or Th17 cells has been suggested to impair protective immune responses to the gut microbiota through disruption of intestinal barrier integrity (1,18). Ustekinumab (UST), which targets the p40 subunit shared by IL-12 and IL-23, is widely used in patients with CD (1,18). RZB, which selectively targets the IL-23p19 subunit, inhibits IL-23-mediated signaling while maintaining the IL-12 pathway intact (1,18,24). The differentiation of Th1 and Th17 cells largely depends on IL-12 and IL-23, respectively (1,18). Thus, while UST inhibited both Th1 and Th17 responses, RZB selectively inhibited Th17 responses. This selective suppression of Th17 responses with preserved Th1 activity may contribute to the superior efficacy of RZB over UST observed in the SEQUENCE trial (25). However, preclinical studies suggest that IL-23-independent IL-17 production contributes to maintenance of intestinal barrier integrity (26,27). Therefore, it is possible that RZB's mechanism of action involves not only the neutralization of IL-23-mediated inflammation, but also the preservation of IL-17 production, which is necessary for barrier maintenance (28). Contrary to these experimental findings, in the present case, RZB treatment did not lead to clinical remission. This lack of efficacy was accompanied by the significantly enhanced mRNA expression of CLDN1 and CLDN2, a molecular profile distinctly different from that observed in the IFX-treated mucosa. However, it is worth noting that the therapy sequence (RZB-IFX) might have affected the outcome of IFX treatment. A mild reduction in the IL23A mRNA expression was maintained during treatment with IFX, suggesting that RZB-IFX sequence therapy contributed to the healing of the ileocecal ulcer.

One major question arising from the present case is why IFX, but not ADA, was effective, despite the fact that both biologics target TNF-α. Recently, we reported a case of ulcerative colitis (UC) successfully treated with IFX after ADA treatment failure (29). The previous and present cases share cytokine and chemokine profiles in that the induction of remission by IFX was accompanied by the marked downregulation of the colonic mRNA expression of CXCL10 without significant alterations in TNFA. In our previous study, we showed that the colonic mRNA expression of CXCL10 was significantly higher in patients with active UC and CD than in those with disease remission (14). Thus, excessive CXCL10 production underlies the immunopathogenesis of CD and UC through the recruitment of proinflammatory Th1 cells and macrophages (14,30). Additionally, the elevated expression of claudin-1 and claudin-2 has been reported in patients with active CD and UC, although a mechanistic link between the expression of claudin and CXCL10 has not been established (22). Collectively, these findings support the idea that IFX induces remission of UC and CD through downregulation of the expression of CXCL10, claudin-1, and claudin-2. Verification of this hypothesis awaits future studies addressing the association between the expression of CXCL10 and claudins in patients with CD and UC.

We must be cautious regarding the interpretation of alterations in the mRNA expression of CLDN1 and CLDN2. Alterations in mRNA expression levels do not always indicate changes in the expression of proteins. In this regard, biopsy samples during endoscopy may not be suitable for the assessment of claudin localization by immunohistochemical or immunofluorescence analyses because epithelial architecture is often disrupted in active disease states and the visualization of the claudin distribution by confocal microscopy may be difficult. Alternatively, immunoblotting of biopsy samples may be useful. Unfortunately, we did not prepare samples for immunoblotting. Visualization of the localization and distribution of claudins at the protein and mRNA levels is required to establish a link between the expression of claudin and CD activity. Whether the expression of claudins is differently regulated by biologics awaits future studies comparing the expression of claudins at the mRNA and protein levels before and after the use of each anti-cytokine antibody in a large number of patients with CD and UC.

In conclusion, we report a case of refractory CD that was successfully treated with IFX following treatment failure with RZB. After the administration of IFX, the mRNA expression levels of CLDN1 and CLDN2 in the ileocecal mucosa were markedly reduced. The findings from this case suggest that IFX may induce clinical remission in CD by downregulating the expression of claudin-1 and claudin-2, thereby contributing to the restoration of intestinal barrier integrity. Verification of this hypothesis requires further studies to address the association between the efficacy of IFX and the mucosal expression of claudins.

The authors state that they have no Conflict of Interest (COI).

Funding Statement

This study was supported in part by Grants-in-Aid for Scientific Research, the Japan Society for the Promotion of Science (25K11288), and Kindai University Research Enhancement Grants (KD2405, KD2505).