Epithelial barrier dysfunction as permissive pathomechanism in human intestinal graft-versus-host disease

Article metrics

Acute graft-versus-host disease (A-GVHD) is characterized by a strong inflammatory reaction of the graft against the recipient, mainly in the skin, the intestine and the liver. The gastrointestinal manifestation of the disease is characterized by nausea, vomiting, abdominal cramps and diarrhea with extensive fluid loss. Histologically, diagnosis of A-GVHD in colonic biopsies is characterized by crypt abscesses and epithelial apoptosis [1]. Pathophysiologically, A-GVHD is described as a three-stage model with activation of antigen presenting cells in the host, donor T-cell activation and target tissue destruction [2]. Within this model barrier dysfunction with translocation of microbial products and luminal antigens is thought to be a major perpetuating factor and a major cause of the “cytokine storm” through immune activation [3]. However, epithelial barrier function in A-GVHD may not only be relevant for the pathophysiology of immune activation but also for the pathomechanism of diarrhea in this disease. Therefore, we investigated the epithelial barrier function in human biopsies ex vivo in Ussing-type chambers with subsequent molecular characterization of potential pathomechanisms.

In this study, 22 patients with diarrhea and clinical suspicion of intestinal GVHD after bone marrow transplantation underwent sigmoidoscopy at the Charité-Universitätsmedizin Berlin. Infectious causes for symptoms were previously ruled out including stool analysis for Clostridium difficile toxin and whole blood CMV-PCR and immunohistochemistry of biopsies. Diagnosis of A-GVHD was made from histological biopsies in 16 patients (A-GVHD; grade I: n = 6, grade II: n = 8, grade III: n = 1, grade IV: n = 1). The remaining six patients without histopathological diagnosis of GVHD served as a further control group (no GVHD). A group of 15 patients which underwent endoscopy within this period of time for tumor exclusion without signs of pathology served also as controls (control). All patients gave their written consent for taking biopsies for scientific purpose and the study was approved by the local ethics committee of the Charité Berlin (229-32). A miniaturized Ussing-type chamber was used for impedance spectroscopy. This technique can differentiate the epithelial and subepithelial portion of the transmural electrical resistance [4]. Paracellular permeability was determined by 3H-Mannitol flux. For macromolecular flux the transcellular passage of horse radish peroxidase (HRP; 44 kDa; Sigma-Aldrich) was analyzed in six consecutive patients with A-GVHD grade I/II. Tight junction protein expression was analyzed by Western blotting. For cryosections, tissues were fixed with 2% paraformaldehyde. Antibodies were rabbit anti-claudin-2, rabbit anti-occludin, mouse anti-claudin-5 (Life Technologies GmbH, Darmstadt, Germany) and mouse anti-ZO-1 (BD Biosciences, San Jose, CA). Secondary antibodies were Alexa Fluor 594 goat anti-mouse or rabbit IgG and 488 goat anti-rabbit or mouse IgG (Molecular Probes, Eugene, OR). Nuclei were counterstained with DAPI. Images were obtained with a confocal laser scanning microscope (Zeiss LSM 510 Meta, Jena, Germany). Results are given as means ± SEM. Differences between groups were tested by one way analysis of variance with Bonferroni-Holm post hoc test.

Representative original impedance locus plots of colonic specimens from a patient suffering from A-GVHD (grade II) and a control patient are shown in Fig. 1a. Electrophysiological data of the colonic biopsies revealed a decrease in transmural wall resistance (Rt) in patients suffering from A-GVHD compared to controls (29 ± 4 vs. 45 ± 3 Ω cm2; p < 0.01). In further analysis using impedance spectroscopy this decrease turned out to be due to a decrease in epithelial resistance (Repi), whereas the subepithelial portion (Rsub) remained unchanged (data not shown). Repi was decreased compared to healthy controls (control), as well as compared to patients with bone marrow transplantation without evidence of A-GVHD (no GVHD; Fig. 1b). Colonic permeability as measured by mannitol flux was found to be markedly increased in A-GVHD (Fig. 1c). To evaluate the permeability to larger molecules passage of HRP was used as a marker of transcytosis. To exclude passive diffusion artifacts, only biopsies without endoscopic or histological gross lesions were used for this analysis. Consequently, there was a non-significant reduction in transmural resistance in this subgroup. However, this group showed a sharp increase in HRP flux compared to control patients (Fig. 1d).

Fig. 1
figure1

a Original impedance spectroscopy plots of human colon in control and patient with acute intestinal graft-versus-host disease. Zreal gives the ohmic component and Zimaginary the reactive component of the complex impedance. Intersections between the semicircle and x-axis at low and high frequencies represent Rt and Rsub, respectively. Rt minus Rsub equals the epithelial resistance (Repi). b Epithelial resistance (Repi) as determined by impedance spectroscopy in colonic biopsies. **p < 0.01 compared with control. *p < 0.05 compared with no GVHD. c Mucosal-to-serosal mannitol tracer flux (JMan) in Ussing-type chambers. *p < 0.05 compared with control. d Horseradish peroxidase (HRP) permeability in early stages (grade I-II) of A-GVHD. Transepithelial resistance was recorded and macromolecule flux (mucosal-to-serosal) was measured in Ussing-type chambers on colonic biopsies (n = 6) over a 30 min time period. *p < 0.05 compared with control

Because of the reduction in Repi in the colon of A-GVHD patients, we went on to analyze the biopsies for changes in tight junction proteins. The expression level of the tight junctional proteins occludin, tricellulin, claudin-1 and claudin-3 to claudin-7 were unchanged (data not shown). Interestingly, claudin-2, a channel-forming protein, was detectable only in severely inflamed tissue but not in controls (Fig. 2a). To analyze the subcellular distribution of claudin-2 we performed immunohistology showing a proper integration into the tight junctional strand complex as indicated by a merge of the claudin-2 immune reactivity and ZO-1 (Fig. 2b). Furthermore, claudin-5, a sealing tight junction protein, showed redistribution from the tight junctional strand complex to a more basolateral staining in patients with A-GVHD (Fig. 2c). However, there was no significant change in the total expression level of claudin-5 (76 ± 12% in A-GVHD vs. 103 ± 8% in controls; p = 0.17; original WB lanes in Fig. 2a).

Fig. 2
figure2

a Western Blot using claudin-2 antibody of patients with severe A-GVHD (grade III/IV) compared to control patients and patients with grade I-II. Original WB lane using anti-claudin-5 antibody (b/c) Immunofluorescence detection of tight junction proteins in confocal LSM in A-GVHD. Merged pictures of ZO-1 or occludin (red) with the respective claudin (green), and nuclei (DAPI, blue). ZO-1 and occludin were used as reference for the tight junction complex. d Apoptotic ratio of controls, patients after bone marrow transplantation without graft-versus-host disease (no GVHD) and patients with acute intestinal graft-versus-host disease (A-GVHD). **p < 0.01 compared with control

Epithelial apoptosis is considered the histological hallmark of A-GVHD. Therefore, we quantified epithelial apoptotic ratio in histological sections by activated caspase-3 staining. We found a fourfold increase in epithelial apoptosis in A-GVHD compared to controls (Fig. 2d).

For A-GVHD, data on epithelial barrier properties are available exclusively from in vivo permeability tests with, e.g., saccharides of different size which have distinct drawbacks in regard to the identification of inherent pathomechanisms and false positive results. In this study, we were able to directly demonstrate colonic barrier dysfunction in A-GVHD patients using an established ex-vivo biopsy model. Furthermore, changes in the assembly of the tight junction molecules and the increase in epithelial apoptosis were the structural basis of this dysfunction.

The first interesting finding was the increased expression of claudin-2 in severe A-GVHD. Claudin-2 is a cation and water channel-forming tight junction protein [5, 6]. The up-regulation of claudin-2 has been shown to be caused by pro-inflammatory cytokines like TNFα and interleukin-13 [7, 8]. Similarly, these proinflammatory cytokines are also up-regulated in A-GVHD and may induce claudin-2 expression in these patients [9, 10]. In contrast, claudin-5 is a sealing tight junction protein [11]. Interestingly, there was no change in immunoblot in A-GVHD but using confocal laser scanning microscopy, we observed redistribution of claudin-5 to the basolateral membrane where it is functionally inactive, a phenomenon that has also been described in patients with active Crohn´s disease [12]. Besides tight junctional changes, up-regulated epithelial apoptosis is another major reason for the barrier dysfunction in A-GVHD. In principle, epithelial apoptosis represents a leak with significant conductivity and therefore contributes to barrier dysfunction for ions [13]. The fact that epithelial apoptoses are functionally relevant for the barrier dysfunction in A-GVHD is further supported by the work of Brown et al. who showed in an animal model that inhibition of TNF-alpha ameliorates not only colonic inflammation with subsequent reduction in epithelial apoptosis but also the elevated intestinal permeability and diarrhea [9].

From our short circuit current data, there was no direct evidence for an activation of active secretory transport processes in A-GVHD (data not shown). Therefore, we conclude that leak-flux diarrhea as the result of a disturbed epithelial barrier is the prominent diarrheal mechanism in A-GVHD.

Besides the increased permeability for ions we also observed an increased macromolecular passage in A-GVHD. Bacteria-derived lipopolysaccharides are thought to play a major role in perpetuating the immune activation in A-GVHD. Evidence for this hypothesis comes e.g., from a mouse model, where LPS antagonization improved histological and clinical A-GVHD score and improved survival in this group [14]. Translocation of lipopolysaccharides is thought to be due to a “leaky barrier”. However, the precise route of LPS permeation is not known. To examine the potential uptake of larger antigens in A-GVHD we used horse radish peroxidase (HRP, 44 kDa). HRP is usually considered not to pass through epithelial apoptotic extrusion sites but to be a marker of gross lesions in the epithelium and to cross the epithelium also transcellularly (via endocytotic uptake) [15]. To exclude gross lesions with consequent passive uptake of HRP we studied a subgroup of patients with early A-GVHD without gross lesions and found a significantly increased HRP permeability. Interestingly, this was highly significant already in early stages of A-GVHD, where the transmural resistance was still not significantly altered, again supporting the concept of a transcellular uptake pathway for HRP.

In conclusion, this small descriptive study in human colonic biopsies shows intestinal barrier dysfunction in acute GVHD. These correlations support the concept of barrier dysfunction to be relevant in the pathogenesis of A-GVHD as well as a role for a leak flux mechanism as important pathomechanism of diarrhea.

References

  1. 1.

    Gallucci BB, Sale GE, McDonald GB, Epstein R, Shulman HM, Thomas ED. The fine structure of human rectal epithelium in acute graft-versus-host disease. Am J Surg Pathol. 1982;6:293–305.

  2. 2.

    Ferrara JL, Levine JE, Reddy P, Holler E. Graft-versus-host disease. Lancet. 2009;373:1550–61.

  3. 3.

    Heidt PJ, Vossen JM. Experimental and clinical gnotobiotics: influence of the microflora on graft-versus-host disease after allogeneic bone marrow transplantation. J Med. 1992;23:161–73.

  4. 4.

    Stockmann M, Gitter AH, Sorgenfrei D, Fromm M, Schulzke JD. Low edge damage container insert that adjusts intestinal forceps biopsies into Ussing chamber systems. Pflug Arch. 1999;438:107–12.

  5. 5.

    Rosenthal R, Milatz S, Krug AM, Oelrich B, Schulzke JD, Amasheh S, et al. Claudin-2, a component of the tight junction, forms a paracellular water channel. J Cell Sci. 2010;123:1913–21.

  6. 6.

    Amasheh S, Meiri N, Gitter AH, Schöneberg T, Mankertz J, Schulzke JD, et al. Claudin-2 expression induces cation- selective channels in tight junctions of epithelial cells. J Cell Sci. 2002;115:4969–76.

  7. 7.

    Heller F, Florian P, Bojarski C, Richter J, Christ M, Hillenbrand B, et al. Interleukin-13 is the key effector Th2 cytokine in ulcerative colitis that affects epithelial tight junctions, apoptosis, and cell restitution. Gastroenterology. 2005;129:550–64.

  8. 8.

    Mankertz J, Amasheh M, Krug SM, Fromm A, Amasheh S, Hillenbrand B, et al. TNFalpha up-regulates claudin-2 expression in epithelial HT-29/B6 cells via phosphatidylinositol-3-kinase signaling. Cell Tissue Res. 2009;336:67–77.

  9. 9.

    Brown GR, Lindberg G, Meddings J, Silva M, Beutler B, Thiele D. Tumor necrosis factor inhibitor ameliorates murine intestinal graft-versus-host disease. Gastroenterology. 1999;116:593–601.

  10. 10.

    Jordan WJ, Brookes PA, Szydlo RM, Goldman JM, Lechler RI, Ritter MA. IL-13 production by donor T cells is prognostic of acute graft-versus-host disease following unrelated donor stem cell transplantation. Blood. 2004;103:717–24.

  11. 11.

    Amasheh S, Schmidt T, Mahn M, Florian P, Mankertz J, Tavalali S, et al. Contribution of claudin-5 to barrier properties in tight junctions of epithelial cells. Cell Tissue Res. 2005;321:89–96.

  12. 12.

    Zeissig S, Bürgel N, Günzel D, Richter J, Mankertz J, Wahnschaffe U, et al. Changes in expression and distribution of claudin 2, 5 and 8 lead to discontinuous tight junctions and barrier dysfunction in active Crohn’s disease. Gut. 2007;56:61–72.

  13. 13.

    Gitter AH, Wullstein F, Fromm M, Schulzke JD. Epithelial barrier defects in ulcerative colitis: characterization and quantification by electrophysiological imaging. Gastroenterology. 2001;121:1320–8.

  14. 14.

    Cooke KR, Gerbitz A, Crawford JM, Teshima T, Hill GR, Tesolin A, et al. LPS antagonism reduces graft-versus-host disease and preserves graft-versus-leukemia activity after experimental bone marrow transplantation. J Clin Invest. 2001;107:1581–9.

  15. 15.

    Wallon C, Braaf Y, Wolving M, Olaison G, Söderholm JD. Endoscopic biopsies in Ussing chambers evaluated for studies of macromolecular permeability in the human colon. Scand J Gastroenterol. 2005;40:586–95.

Download references

Acknowledgements

We thank the team of the endoscopy unit (Campus Benjamin Franklin, Charité .- Universitätsmedizin Berlin) for the excellent cooperation and in memoriam D. Sorgenfrei for his remarkable technical assistance in the lab. This work was supported by grants from the Deutsche Forschungsgemeinschaft (DFG): Schu 559/11-2

Author information

Correspondence to Jörg-Dieter Schulzke.

Ethics declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Additional information

These authors contributed equally: Kathrin Rieger, Jörg-Dieter Schulzke

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Troeger, H., Hering, N.A., Bojarski, C. et al. Epithelial barrier dysfunction as permissive pathomechanism in human intestinal graft-versus-host disease. Bone Marrow Transplant 53, 1083–1086 (2018) doi:10.1038/s41409-018-0144-2

Download citation