Recent Advances in Acute Graft-Versus-Host Disease

1. University of Michigan Medical Center, Ann Arbor, MI

2. University of Regensburg Medical Center, Regensburg, Germany

Allogeneic hematopoietic cell transplantation (HCT) is an important component of the therapeutic armamentarium available for patients with hematologic malignancies, particularly for those with acute leukemia. In the last 10 years, improvements in management of complications of allogeneic HCT have reduced the severity of acute graft-versus-host disease (GVHD); damage to kidneys, lungs, and liver; and deaths due to infections such that transplantrelated mortality has fallen by more than 50 percent.¹  Further, at many centers, long-term survival after matched, unrelated donor transplant is equivalent to that of matched, sibling donor transplant. This mini-review focuses on advances in the understanding and management of acute GVHD with special emphasis on the gastrointestinal (GI) tract, the target organ most resistant to treatment.

Both recipient age and extent of recipient-donor HLA matching have long been used to stratify risk for development of acute GVHD. Recently, a large retrospective study investigated whether other factors (conditioning regimen [myeloablative plus total body irradiation vs. myeloablative without total body irradiation vs. reduced intensity], stem cell source [bone marrow vs. peripheral blood], and donor type [matched sibling vs. unrelated donor]) might have predictive value for acute GVHD.²  Results from that study indicated that regimens that do not use total body irradiation are associated with a lower risk of acute GVHD when bone marrow, but not peripheral blood, was used as the stem cell source, while donor type (matched siblings being more favorable than unrelated donors) and severity of symptoms at the onset were found to have modest predictive value for response to treatment for acute GVHD and non-relapse mortality.

Biomarker combinations that include regenerating islet-derived 3-alpha (Reg3α), tumor necrosis factor receptor 1 (TNFR1), interleukin 2 receptor alpha (IL2Rα), interleukin 8 (IL8), hepatocyte growth factor (HGF), and elafin appear to have predictive value for etiology, response to treatment, and treatment-related mortality in acute GVHD.^3,4  For example, Reg3α may be useful in establishing the etiology of post-transplant diarrhea. Elevated levels of Reg3α identified patients at high risk for GI GVHD, whereas a non-GVHD etiology (e.g., infection with C. difficile) was more likely in patients with post-transplant diarrhea with low Reg3α concentrations. Further, patients with post-transplant diarrhea who have high Reg3α levels are less likely to respond to treatment within a month. A combination of three factors (volume of diarrhea, histologic severity, and Reg3α level) at the onset of symptoms can classify patients into four groups for non-relapse mortality (NRM) (Figure 1). Overall NRM for the study cohort with newly diagnosed GI GVHD was 48 percent, but patients with zero risk factors had an NRM of 25 percent compared with an NRM of 86 percent for patients with all three risk factors. In a separate multicenter study of patients with newly diagnosed acute GVHD, a panel of six biomarkers (TNFR1, IL2Rα, IL8, HGF, reg3α, and elafin) predicted both response to treatment and non-relapse mortality.⁴  If plasma biomarkers can predict GVHD prior to the onset of clinical symptoms, preemptive treatment of acute GVHD on the basis of laboratory testing may be feasible in a manner akin to preemptive treatment of cytomegalovirus on the basis of viral load as determined by quantitative polymerase chain reaction.

Reg3α, a natural antibiotic that targets enterococci, is secreted by Paneth cells (the “guardians” of the intestinal stem cells) and enterocytes. Involvement of the intestinal microbiome in GI GVHD has been postulated since the remarkable observation 40 years ago that mice with a sterile gut are resistant to acute GVHD. We now know that GVHD damages Paneth cells, depleting protective antimicrobial proteins such as Reg3α and defensins and thereby causing loss of diversity of the intestinal microflora. We analyzed by 16s RNA sequencing the architecture of the GI microbiome in 70 stool specimens from patients with acute GI GVHD and found overgrowth of enterococci together with contraction of all other microbial flora (Figure 2). These observations support the concept that the architecture of the microbiome is altered in human GI GVHD and parallel those in experimental models of acute GVHD in which systemic antibiotics amplified GI damage and in which the effects of such antibiotic treatment was ameliorated by oral probiotics.⁶  Together, these studies imply an interplay among gut flora, GI mucosal immunity, and GVHD and suggest therapeutic approaches that target these interactions.

A variety of bacterial ligands and receptors control the responses of regulatory T cells (Tregs) that maintain tolerance of the gut’s immune system for commensal GI microflora.⁷  The key contribution of FoxP3+ Tregs to the homeostasis of the GI mucosal immune system is now well established in multiple mammalian species, including humans. Tregs condition stem cells in a number of microenvironments, and in the bone marrow they co-localize with hematopoietic stem cells to facilitate their protective effects.⁸  Tregs can convert effector T cells into Tregs by modulating the function of antigen-presenting cells, and in in vivo model systems, the adoptive transfer of donor Tregs has been shown both to protect HCT recipients from lethal GVHD and to improve immune reconstitution.⁹  Initial clinical trials of haploidentical HCT appear to confirm these findings in humans. With Treg-enriched donor lymphocytes as the sole prophylaxis for GVHD, only two of 28 patients developed clinically significant acute GVHD, even though large numbers of conventional T cells were co-transplanted with CD34-selected donor stem cells.¹⁰  Non-hematopoietic antigen-presenting cells in the GI tract may also be important participants in this process because they can potently stimulate alloreactive T cells and induce lethal GVHD.¹¹  Prevention of GVHD in the GI tract may also be attributable to the capacity of Tregs to impede the migration of conventional T cells. Functional expression of chemokine (C-C motif) receptor 5 (CCR5) is required for donor effector lymphocytes to home to the GI tract and liver and cause acute GVHD. In a phase II trial involving high-risk patients, blockade of CCR5 with maraviroc for the first 30 days after HCT resulted in a 29 percent incidence of grade II-IV acute GVHD in the skin compared with an 8.8 percent incidence in the GI tract and the liver.¹² 

Our understanding of acute GVHD of the GI tract has rapidly advanced during the last several years. Experimental systems have illuminated unexpected relationships between both the innate and adaptive arms of the GI mucosal immune system and the GI microbiota that influence the development of GVHD. Increased understanding of the complex interactions among different classes of antigen-presenting cells, including Tregs and effector T cells, and their trafficking characteristics has suggested new therapeutic strategies, now in early-phase clinical trials, that do not rely on traditional, non-specific immunosuppressants. The discovery of important predictive and prognostic biomarkers may soon provide an approach to personalization of therapy according to an individual patient’s risk of developing acute GVHD. Such advances are likely to improve outcome broadly by making allogeneic HCT safer and more effective.