Graft-Versus-Host Disease Alters Intestinal Microbial Ecology by Inhibiting Production of Enteric Defensins

Abstract 244

Bacterial infection is a serious complication of bone marrow transplantation (BMT). Intestinal GVHD, in particular, significantly enhances the risk for Gram-negative septicemia. The majority of the intestinal microbiome of the hosts consists of non-cultivable obligate anaerobes, while the Gram-negative bacteria such as Escherichia coli (E. coli) make up a small proportion of the microflora. Thus, it remains unclear why Gram-negative septicemia is dominant in intestinal GVHD, while the role of systemic immunosuppression and use of antibiotics is well-appreciated. We evaluated gut flora changes in the course of GVHD in mouse models of BMT without giving antibiotic or immunosuppresive drugs. Lethally irradiated B6D2F1 (H-2^b/d) or B6C3F1 (H-2^b/k) mice were injected with 5 × 10⁶ T-cell depleted BM alone (non-GVHD controls) or with 2 × 10⁶ T cells (GVHD group) from MHC-mismatched B6 (H-2^b) donors on day 0. Intestinal microflora was identified by using terminal restriction fragment length polymorphism (T-RFLP) analysis of 16S rRNA gene libraries constructed from each sample of gut contents. It consisted of approximately 80% of obligate anaerobe, and 20% of anaerobe such as Lactobacilli and Clostridia with very few E. coli before BMT. After BMT, this diversity of gut flora was preserved in non-GVHD controls (Table). In contrast, GVHD mice showed a marked increase of E. coli with a significant decrease in the members of obligate anaerobe. This was associated with dissemination of E. coli to the mesenteric lymph nodes (mLNs) and liver, with elevated serum levels of lipopolysaccharide (LPS). Such a loss of diversity of gut flora with a flora shift towards E. coli was significantly associated with morbidity and mortality of GVHD. Numbers of T-RFLP peaks that indicate diversity of intestinal flora were inversely correlated with GVHD clinical scores (p<0.001). A degree of E. coli proportion has significant correlation with GVHD clinical scores (p<0.001) and GVHD mortality (80% in high E. coli group vs. 0% in low E. coli group). We then investigated the underlying mechanisms of the disruption of intestinal ecology in GVHD. Paneth cell derived α-defensins are essential regulators of intestinal microbial ecology. We therefore hypothesized that Paneth cell damage in GVHD inhibited production of enteric defensins and disrupted intestinal ecology. Immunohistochemistry for lysozyme that marks Paneth cells showed significant reduction of Paneth cells in GVHD and quantitative real-time PCR analysis showed dramatically reduced expression of enteric defensins including Defa21, Defa1, Defa4, Defa5, and Defcr-rs1 (Table). These results suggest that Paneth cell injury in GVHD could lead to lower expression of enteric defensins and a shift of gut flora from commensal microorganisms towards widespread prevelance of gram-negative bacteria in the intestinal microbiome, and the subsequent high risk for the development of life threatening Gram-negative septicemia. LPS derived from Gram-negative bacteria plays an important role in amplifying systemic GVHD. Thus, such an alteration of intestinal ecology may be related to exaggeration of systemic GVHD. These results thus uncover the previously unrecognized role of the intrinsic antimicrobial peptides and the crosstalk between hosts and the intestinal microbes in the pathogenesis of GVHD and infection after allogeneic BMT.