Improving bench-to-bedside translation for acute graft-versus-host disease models

Hematopoietic stem cell transplantation (HSCT) is a potentially life-saving treatment for refractory/relapsing hematological malignancies (see Glossary, Box 1), severe blood diseases and primary immunodeficiency diseases (Blazar et al., 2020; Shimoni et al., 2016; Zeiser and Blazar, 2017a; Hall and de Zoeten, 2024; Rao et al., 2007; Nguyen et al., 2015). This procedure involves the transplantation of CD34⁺ progenitor cells (Box 1) derived from three main sources: healthy donor bone marrow, peripheral blood from granulocyte colony-stimulating factor (G-CSF; also known as CSF3)-treated donors or umbilical cord blood (Fig. 1). Before the infusion of these progenitor cells, recipients undergo pretransplant conditioning using whole-body radiation and/or myeloablative chemotherapy (Box 1), which destroys the bone marrow (BM), thereby creating immunological space for engraftment of donor cells (Fig. 1), as well as eliminates cancer cells present within recipients. Patients are then infused with the donor progenitor cells, which engraft, differentiate and proliferate into all of the cellular elements of blood. A recent update from the Center for International Bone and Marrow Transplant Research reported that a total of 23,535 individuals in the USA received HSCT (Phelan et al., 2022). Of these, 14,236 patients received their own (i.e. autologous) progenitor cells, while the other 9299 individuals received allogeneic progenitor cells from related or unrelated donors. For many indications, the preferred hematopoietic stem cell (HSC) donor is a human leukocyte antigen (HLA)-matched sibling (Box 1). However, the chance of finding such a donor varies widely, ranging from 13% to 51% depending on ethnic background and age (Spellman, 2022). The next best option is an HLA-matched unrelated donor, the availability of which in the USA ranges from 29% for African American patients to 78% for white non-Hispanic patients (Spellman, 2022). In the absence of these HLA-matched donors, patients may receive HSCs from either a related or unrelated haploidentical donor who shares ∼50% of the HLA alleles with the recipient. Stem cells from these allogeneic donors can then be used for >95% of all patients in need of HSCT (Baumeister et al., 2020; Spellman, 2022).

Unfortunately, 30-70% of allogeneic HSCT patients will develop a potentially life-threatening, inflammatory disease called acute graft-versus-host disease (aGVHD) (Zeiser and Blazar, 2017a; Hill et al., 2021). This tissue-destructive inflammatory disease usually occurs within 1-3 months following allogeneic HSCT, although it can occur at later timepoints. That said, important new clinical studies demonstrate that post-transplant administration of high-dose cyclophosphamide allows for the use of HLA-disparate donors in HSCT (Shaffer et al., 2024; Shaw et al., 2021; Bolanos-Meade et al., 2023). These clinical studies were based, in part, on the important preclinical findings that cyclophosphamide administration limits aGVHD in mouse models of allogeneic HSCT (O'Donnell and Jones, 2023). In a 2024 study, Shaffer et al. (2024) reported that the 3-year overall survival and aGVHD-free survival of patients who received HLA-mismatched unrelated donor transplants together with post-transplant cyclophosphamide were essentially the same as those of patients who received HLA-matched unrelated donor transplants and cyclophosphamide treatment. This multi-ethnic study [which included Hispanic and non-Hispanic whites, African Americans, Asians, American Indians, Native Hawaiians and Pacific Islanders diagnosed with different leukemias or myelodysplastic syndrome (Box 1)] demonstrated that administration of post-transplant cyclophosphamide could greatly expand the use of HSCT for patients from most ethnic and racial backgrounds.

Both experimental and clinical data demonstrate that allogeneic T cells are the major immune cell type that initiates aGVHD (Hess et al., 2021; Hill et al., 2021; Zeiser and Blazar, 2017a). Although these lymphocytes typically target the skin, gut and liver, they can also damage the lungs and lymphoid tissue (Malard et al., 2023; Zeiser and Blazar, 2017a; Zeiser and Teshima, 2021; Teshima and Hill, 2021). Importantly, allogenic T cell-mediated damage to primary lymphoid tissue (BM, thymus) and secondary lymphoid tissue (lymph nodes, spleen) can result in severe lymphopenia (Box 1), thrombocytopenia (Box 1) and anemia (Muskens et al., 2021; Shono et al., 2014, 2010; Chewning et al., 2013; McDaniel Mims et al., 2019; Cuthbert et al., 1995; Shahin et al., 2017; Zeiser and Teshima, 2021; Enriquez et al., 2023). The immunodeficiency caused by aGVHD markedly increases the risk of infections and bleeding, and of morbidity and mortality, accounting for ∼30% of patient deaths (Muskens et al., 2021; Shono et al., 2014, 2010). Based upon early preclinical studies that demonstrated the ability of corticosteroids to reduce aGVHD in mouse models (You-Ten et al., 1995), high doses of corticosteroids are now used as the first line of defense for patients undergoing HSCT. Nevertheless, 40% of patients with aGVHD do not respond to this therapy, with few second-line therapies available (Malard et al., 2023). Six-month survival for steroid-resistant patients is ∼50%, with survival reduced to 30% in patients surviving >2 years (Zeiser et al., 2023). Although out of the scope of this Review, it is worth mentioning that allogeneic HSCT can also induce a chronic form of the disease called chronic graft-versus-host disease (cGVHD), which affects 30-70% of patients who survive for 5-6 months following this type of transplant (Hill et al., 2021; Zeiser and Blazar, 2017b). cGVHD is a heterogeneous disease that is immunologically different from aGVHD as it can affect any tissue, resulting in fibrotic disease of the face, eyes, musculoskeletal system, genitals and/or joints (Hill et al., 2021; Zeiser and Blazar, 2017b).

Much of our early understanding of aGVHD came from preclinical studies using canine models of allogenic HSCT. Data generated from these large-animal models more than 50 years ago established some of the first HSCT protocols in patients (Teshima and Hill, 2021; Thomas et al., 1975a,b). For example, the prophylactic administration of both cyclosporine A (Box 1) [or tacrolimus (Box 1)] and methotrexate (Box 1) together with pan T cell-depleting antibodies (Box 1) was first proven to be effective in attenuating aGVHD in canine models of HSCT (Deeg et al., 1982; Kolb et al., 1979). These groundbreaking studies culminated in Dr E. Donnall Thomas receiving the Nobel Prize in Physiology or Medicine in 1990. Although canine models of disease continue to be used, the costs of inbred dogs, together with the costs for housing/care in an animal care facility, are far greater than those of rodents. Because the vast majority of all preclinical studies investigating the immunopathogenesis of aGVHD use mouse models, this Review will focus on the different mouse models of disease.

The use of inbred allogeneic HSCT in mice has been instrumental in delineating several of the key aspects of aGVHD pathogenesis as well as providing investigators with potential targets for drug discovery. The major advantages of using these inbred mouse models are their relatively low cost and reproducibility, which have allowed investigators to examine specific pathogenetic mechanisms via the use of a wide variety of readily available pharmacological and immunological reagents (Table 1) (Schroeder and DiPersio, 2011; Stolfi et al., 2016; Zeiser and Blazar, 2016, 2017a; Hess et al., 2021; Markey et al., 2014; Patel et al., 2022; Reddy et al., 2008). Data generated from numerous aGVHD mouse studies have identified and characterized the sequences of events that lead to the development of aGVHD.

Intravenous administration of major histocompatibility complex (MHC; Box 1)-mismatched (Box 1) HSCs induces aGVHD that occur in three major phases: the initiation phase, the activation phase and the effector cell phase (Fig. 2) (Malard et al., 2023; Schroeder and DiPersio, 2011; Blazar et al., 2012). The initiation phase (Fig. 2A) begins following myeloablative pretransplant conditioning of the recipient (Fig. 1). The large majority of mouse models of aGVHD use lethal total-body irradiation (TBI) to ablate the recipient's BM prior to engraftment with allogeneic BM supplemented with a source of allogeneic T cells, such as unfractionated splenocytes (Box 1) or isolated CD4⁺ and/or CD8⁺ T cells (Box 1) (Schroeder and DiPersio, 2011). TBI is a toxic conditioning protocol that damages multiple tissues, which respond with the production of numerous inflammatory mediators, including damage-associated molecular patterns, cytokines, chemokines and chemokine receptors (Khuat et al., 2021; Hill and Koyama, 2020; Blazar et al., 2020; Zeiser et al., 2016; Castor et al., 2012; Mapara et al., 2006; Chewning et al., 2013; Dudakov et al., 2017; Schroeder and DiPersio, 2011; Shono et al., 2014, 2010; Stolfi et al., 2016; Zeng et al., 2017). In addition, TBI injures epithelia of tissues in direct contact with commensal microbiota (Box 1) (e.g. skin and gut), resulting in the translocation of bacteria and their products, which induces microbe-associated molecular patterns and multiple pro-inflammatory mediators, as described above (Blazar et al., 2012; Schwab et al., 2014; Staffas et al., 2017; Zeiser and Blazar, 2017a; Malard et al., 2023). Although the number of commensal bacteria within the healthy liver is relatively low compared to the numbers in the skin and gut, bacterial counts can increase dramatically following TBI owing to the transportation of whole bacteria (and their products) from the damaged gut to the liver via the portal vein (Malard et al., 2023). The resulting inflammatory environment within the different tissues leads to the infiltration and activation of innate immune cells, such as polymorphonuclear leukocytes, monocytes and tissue-residing dendritic cells (DCs). Once activated, DCs migrate from the tissue into the lymphatic vessels, where they are transported into tissue-draining lymph nodes, resulting in initiation of the activation phase (Malard et al., 2023; Schroeder and DiPersio, 2011; Blazar et al., 2012).

This phase is characterized by the upregulation of MHC I and MHC II on activated DCs, which facilitates their binding to, and activation of, donor-derived, alloreactive CD4⁺ and/or CD8⁺ T cells (Fig. 2B). Activation of T cells drives the differentiation and expansion of disease-producing effectors cells that include T helper (Th)1 and Th17 cells (Box 1), as well as cytotoxic CD8⁺ T cells. All three groups of alloreactive T cells then exit the lymph nodes via the efferent lymphatics and enter the systemic circulation, in which they traffic to different target tissues to initiate the effector cell phase (Malard et al., 2023; Schroeder and DiPersio, 2011; Blazar et al., 2012).

During this phase, cytotoxic CD8⁺ T cells damage host cells by releasing tumor necrosis factor-α (TNFα; also known as TNF), interferon-γ (IFNγ), perforin (Box 1) and granzymes (Box 1), and via the interaction of FAS (Box 1) with FAS ligand (Fig. 2C). Alloreactive Th1 effector cells can induce cell and tissue injury by producing IFNγ, TNFα, interleukin (IL)-2 and lymphotoxin-α (Box 1). These inflammatory cytokines can also promote inflammatory tissue damage indirectly via their ability to recruit and/or activate innate immune cells (e.g. macrophages, neutrophils, DCs), as well as activate additional cytotoxic CD8⁺ T cells (Blazar et al., 2012; Hill and Koyama, 2020; Malard et al., 2023; Zeiser and Blazar, 2016, 2017a; Zeiser et al., 2016). Although Th1 responses are considered to be the primary driver of aGVHD, IL-17 (also known as IL17A)-producing T cells also play a significant role by promoting the recruitment of neutrophils and amplifying the inflammatory response (Carlson et al., 2009; Kappel et al., 2009) (Fig. 2C).

There is no question that mouse models of aGVHD have greatly advanced our understanding of disease pathogenesis. However, it has become increasingly apparent that alloimmune responses in mice may differ substantially from those in humans. Indeed, the translation of promising therapeutics that were found to be effective in conventional mouse models of aGVHD has been disappointing in clinical studies (see ‘Therapeutic intervention in mouse models of aGVHD’ section) (Stolfi et al., 2016; Zeiser and Blazar, 2016; Rayasam and Drobyski, 2021; Enriquez et al., 2020; Blazar et al., 2020; Hess et al., 2021; Kennedy et al., 2021). In an attempt to enhance translation of preclinical data to patient treatment, investigators have developed a variety of humanized mouse models of xenogenic graft-versus-host disease (xGVHD) (Table 2) (Chupp et al., 2024; Elhage et al., 2022; Hess et al., 2021; Hess et al., 2020; Ye and Chen, 2022). One of the best-characterized mouse models of xGVHD demonstrates that intravenous administration of human peripheral blood mononuclear cells (PBMCs) into mildly irradiated, nonobese diabetic-severe combined immunodeficient mice lacking the IL-2 receptor common γ chain (Box 1) (NSG mice) induces lethal xGVHD involving the liver, lung and skin (Brehm et al., 2019; King et al., 2009) (Table 2). It should be noted that PBMCs contain T cells, B cells and CD34⁺ progenitor cells. Additional studies by Brehm et al. (2019) using this model reported that when NSG mice were rendered devoid of MHC I and MHC II, engraftment of human T cells failed to induce aGVHD. Furthermore, onset of disease in xGVHD models is associated with enhanced production of several human T cell-derived cytokines and growth factors (Kawasaki et al., 2018). Taken together, these studies demonstrated that human T-cell receptors are capable of recognizing murine MHC I and MHC II, resulting in the activation, polarization and expansion of pathogenic T cells (Brehm et al., 2019). Current evidence shows that the large majority of human T cells derived from xenogeneic mice are Th1 effector cells (Hess et al., 2021, 2020). This is not surprising given the fact that circulating levels of human IL-12 are increased in xGVHD mice (Kawasaki et al., 2018). Another important observation made by investigators using xGVHD models is that tissue inflammation may occur in the absence of pretransplant conditioning, suggesting that nonspecific tissue damage and the consequent cytokine storm are not required for induction of disease (Table 2) (Hess et al., 2021). However, it is highly likely that inflammatory mediators produced during xenogeneic T-cell activation modulate the incidence, progression and severity of disease (Hess et al., 2021).

As with any animal model of human disease, xGVHD has a few noteworthy limitations. For example, although the PBMCs→NSG model is characterized by expansion of xenoreactive human T cells, development and/or function of human B cells, monocytes, macrophages and natural killer (NK) cells is poor (Rongvaux et al., 2013). In addition, these xenogeneic mice maintain their murine innate immune system. It is thought that the paucity of functional human B cells and innate immune cells in this model is due to cytotoxic mechanisms mediated by murine innate immune cells and/or the limited reactivity of several mouse cytokines with corresponding human receptors (Hess et al., 2021; Rongvaux et al., 2014, 2013). Ideally, the ultimate goal for studying human aGVHD would be to create mice with a fully humanized adaptive and innate immune system. Over the past decade, great progress has been made in generating mice with a humanized immune system (HIS) via the engraftment of human CD34⁺ HSCs into immunodeficient NSG recipients (Hess et al., 2021; Rongvaux et al., 2014; Martinov et al., 2021). Although new and important information has been reported using these HIS models, development of fully functional innate and adaptive immunity remains limited (Chupp et al., 2024). Recently, Chupp et al. (2024) reported a major breakthrough in HIS development by engrafting immunodeficient, non-irradiated NBSGW and NSGW41 mice (Box 1) with human neonatal cord blood CD34⁺ cells followed by treatment with 17β-estradiol. They found that 17β-estradiol treatment induced a human lymphoid and myeloid immune system that was characterized by marginal zone B cells, germinal center B cells, follicular Th cells and neutrophils, as well as well as formed lymph nodes and intestinal lymphoid tissue (Chupp et al., 2024). In addition, these mice exhibited diverse human B-cell and T-cell antigen receptor repertoires and were able to mount mature T cell-dependent and T cell-independent antibody responses. It will be interesting to assess the immunopathogenesis of aGVHD using this novel HIS model.

As discussed above, allogeneic mouse models have been invaluable in identifying the role that certain pro-inflammatory cytokines (e.g. IFNγ, IL-1β, IL-2, TNFα and IL-6) and chemokine receptors (e.g. CCR5) play in the pathogenesis of aGVHD. In some cases, some of these inflammatory mediators have been developed for use in clinical studies (Hess et al., 2021; Zeiser and Blazar, 2016, 2017a). Unfortunately, the majority of clinical studies that have evaluated the therapeutic efficacy of immunoneutralizing monoclonal antibodies (mAbs) or selective receptor antagonists for the different inflammatory mediators involved in aGVHD pathogenesis in murine models have shown little or no efficacy in patients (Stolfi et al., 2016; Zeiser and Blazar, 2016; Rayasam and Drobyski, 2021; Enriquez et al., 2020; Blazar et al., 2020; Hess et al., 2021; Kennedy et al., 2021). The reasons for these disappointing results are currently not well understood. It has been suggested that differences in the structure, cell composition, mediators and/or immune responses of the murine versus human immune systems could be responsible for the discouraging therapeutic outcomes. It is known that humans have a more complex lymphatic system that contains larger numbers of lymph nodes than that of mice. The lymph nodes are organized into more complex chains that drain smaller areas of tissue compared to the lymph nodes of mice (Haley, 2003). In addition, the mouse spleen is a major hematopoietic tissue, whereas splenic hematopoietic activity is absent in humans (Haley, 2003; Mestas and Hughes, 2004). In addition to these structural and functional differences, mice and humans exhibit a number of important differences with respect to their innate and adaptive immunity. Mestas and Hughes (2004) provided a detailed overview of these differences, including alterations in the relative abundance of lymphocytes and neutrophils in the systemic circulation, components of B- and T-cell signaling pathways, immunoglobulin subsets, γδ T cells, defensins, cytokines/cytokine receptors, Th1/Th2 differentiation, co-stimulatory molecule expression and function, and chemokine/chemokine receptor expression, to mention just a few. Furthermore, when Shay and coworkers compared the transcriptional profiles of different human and mouse immune cells, they found hundreds of genes with distinctly divergent expression across the different cells (Shay et al., 2013). Using high-throughput sequencing assays on the transcriptome and epigenome of several different tissues obtained from mice and human, Lin et al. (2014) also found extensive RNA expression diversity between the two species. Another feature of aGVHD models that can complicate interpretation of preclinical data generated in allogeneic or xenogeneic mouse models is the fact that human disease is most likely multi-dimensional, involving endless cycles of tissue damage created by alloreactive T-cell activation followed by their infiltration into target tissue, which results in further end-organ damage that promotes the recruitment of a new wave of disease-producing lymphocytes and innate immune cells (Hess et al., 2021). Thus, immunoneutralization or drug blockade of a single mediator may have only a limited effect on overall disease (Gadina et al., 2019; Hess et al., 2021; Schroeder et al., 2018). These concerns have prompted several preclinical studies to assess the efficacy of new therapeutics capable of inhibiting the production of multiple inflammatory mediators.

One class of novel small molecules capable of suppressing inflammatory tissue damage in mouse models of aGVHD is selective Janus kinase (JAK) inhibitors (Kim et al., 2024). There are several different JAK family members that, together with signal transducers and activators of transcription (STATs), regulate the transcription of multiple genes upregulated in inflammatory and autoimmune diseases (Kim et al., 2024). JAK1 and JAK2 inhibitors selectively block the downstream signaling of multiple cytokine receptors (Gadina et al., 2019; Hess et al., 2021; Schroeder et al., 2018; Kim et al., 2024; Zeiser et al., 2023). Based upon the exciting preclinical studies that have demonstrated the efficacy of selective JAK inhibitors in mouse models of aGVHD (Courtois et al., 2021; Gadina et al., 2019; Schroeder et al., 2018), new clinical studies have been initiated to test selective JAK inhibitors in patients undergoing allogeneic HSCT or patients with aGVHD (Kim et al., 2024). Ruxolitinib (a JAK1/2 inhibitor) is the first new US Food and Drug Administration (FDA)-approved drug to successfully treat steroid-resistant graft-versus-host (GVHD) in more than 30 years (Zeiser et al., 2020, 2023; Kim et al., 2024).

In addition to JAK inhibitors, a recent clinical study reported the efficacy of a humanized mAb (vedolizumab) to treat lower-gastrointestinal aGVHD (Chen et al., 2024). This study demonstrated in a randomized, double-anonymized, placebo-controlled phase III clinical study that vedolizumab, an mAb specific for T-cell-associated α4β7 integrin, significantly increased the disease-free survival of gastrointestinal aGVHD patients compared to that of the placebo control group (Chen et al., 2024). This clinical trial was based upon data from preclinical mouse models of aGVHD demonstrating that the expression of integrin α4β7 on activated T cells was required for their infiltration and damage to the intestinal mucosa (Dutt et al., 2005; Fu et al., 2019; Petrovic et al., 2004). In these preclinical findings, the reduced extravasation of alloreactive T cells into the gut was associated with markedly reduced production of multiple inflammatory cytokines and chemokines within the tissue, and diminished recruitment and activation of additional inflammatory cells (e.g. monocytes, neutrophils, macrophages) (Fu et al., 2019). Abatacept, another novel biologic currently approved to treat rheumatoid arthritis, psoriatic arthritis and juvenile idiopathic arthritis, has been shown to be effective in preclinical mouse models of aGVHD (Gao et al., 2021; Hess et al., 2021). This cytotoxic T-lymphocyte antigen-4 (CTLA-4) fusion protein suppresses T-cell activation by binding to CD80 and CD86 on DCs, thereby preventing their interaction with T-cell-associated CD28 and second signal activation of these lymphocytes. Similar to JAK inhibitors and vedolizumab, abatacept prevents activation of alloreactive T cells and subsequent generation of inflammatory mediators (Kean et al., 2024). In a recent clinical study, investigators found that when two young children who were undergoing unrelated donor HSCT were treated prophylactically with abatacept in addition to standard GVHD/immunosuppression therapy, survival outcomes were much improved (Kean et al., 2024).

Although exciting, the translational success of these new treatments is limited, as their efficacy is generally short lived, requiring additional/multiple treatments for controlling disease, as with many conventional therapeutics. Thus, investigators have been evaluating novel cell-based therapies that may provide more long-lasting immunosuppressive effects in mouse models of aGVHD. Taylor et al. (2002) were the first to report that infusion of ex vivo-activated and expanded murine CD4⁺CD25⁺Foxp3⁺ regulatory T cells (Tregs) (Box 1) suppressed murine aGVHD. Following this landmark study, numerous laboratories have demonstrated that infusion of different populations of Tregs can both prevent and reverse pre-existing disease in standard murine and xenogeneic mouse models of aGVHD (Blazar et al., 2018; Riegel et al., 2020; Zeiser et al., 2023; Dittmar et al., 2024; Hippen et al., 2022). These dramatic preclinical results launched a series of early-phase clinical studies to assess the safety and efficacy of Tregs in limited numbers of patients with preexisting aGVHD (or cGVHD) or prophylactically following allogenic HSCT (Hippen et al., 2022). Overall, investigators have found that infusions of Tregs were well tolerated and provided transient/modest improvement of disease (Hippen et al., 2022). Although these studies are quite promising, randomized and placebo-controlled clinical studies will be needed to demonstrate the efficacy and reproducibility of these Treg protocols.

Another population of immunosuppressive cells that are being evaluated for their ability to attenuate aGVHD in mice and humans is the mesenchymal stromal cells (MSCs), which are present in BM, blood, umbilical blood, adipose tissue and placenta (Kadri et al., 2023; Zeiser et al., 2023; Wuttisarnwattana et al., 2023). Upon activation by inflammatory cytokines (such as IFNγ, IL-1, IL-6 and TNFα), MSCs produce immunosuppressive cytokines, such as IL-10, transforming growth factor-β, hepatocyte growth factor, prostaglandin E2, nitric oxide and indoleamine 2,3-dioxygenase (Glenn and Whartenby, 2014; Le Blanc and Mougiakakos, 2012). Although a number of preclinical studies have demonstrated MSC-mediated suppression of aGVHD, others have not (Kadri et al., 2023). Currently, the specific mechanisms responsible for immune suppression in vivo, as well as the location(s) of cell engraftment, have not been definitively identified. Nevertheless, a number of early-phase clinical studies are underway to assess the efficacy of MSCs in suppressing aGVHD (Kadri et al., 2023; Zeiser et al., 2023). As observed in murine models, some, but not all, of the MSC clinical studies have demonstrated significant disease improvement (Kadri et al., 2023). In addition, several other immunosuppressive cell populations have been suggested as candidates for treating aGVHD, including myeloid-derived suppressor cells (Feng et al., 2023), regulatory DCs (Scroggins and Schlueter, 2023), invariant NK cells (Du et al., 2017) and regulatory B cells (Rosser and Mauri, 2015; Rosser and Mauri, 2021).

Despite the wealth of information generated from mouse models and the success of translating a few of the more recent preclinical data into new and effective treatments of human aGVHD, bench-to-bedside translation has remained relatively low (Stolfi et al., 2016; Zeiser and Blazar, 2016; Rayasam and Drobyski, 2021; Enriquez et al., 2020; Blazar et al., 2020; Hess et al., 2021). Accumulating preclinical data suggest that these disappointing outcomes are related to differences in the protocols used in murine models of aGVHD compared to those used in clinical allogeneic HSCT (Table 3) (Blazar et al., 2020; Enriquez et al., 2020; Stolfi et al., 2016; Enriquez et al., 2023; Zeiser and Blazar, 2016). In the next section, we discuss how these differences may affect translation of preclinical data to patient treatment and provide suggestions as to how this information could be used to increase the translatability of mouse models of aGVHD.

As discussed previously, patients that undergo allogeneic HSCT receive a pretransplant conditioning regimen that includes high-dose chemotherapy (busulfan, fludarabine, cyclophosphamide) and/or TBI, which are effective in eradicating malignant cells and preparing the bone marrow for HSC engraftment. In addition, prophylactic administration of one or more immunosuppressive medications (e.g. tacrolimus, methotrexate, anti-thymocyte globulin) is used to suppress alloreactive immune responses, promote engraftment of the transplanted cells and reduce the development of aGVHD (Shaffer et al., 2024). New clinical studies have reported that post-transplant administration of cyclophosphamide reduces the development of aGVHD in recipients who received HLA-disparate HSCs (Shaffer et al., 2024; Shaw et al., 2021; Bolanos-Meade et al., 2023). Of note, these clinical studies were based on the important preclinical findings that post-transplant administration of cyclophosphamide limited the development of aGVHD in mouse models of allogeneic HSCT (O'Donnell and Jones, 2023). Unfortunately, these types of preclinical studies using clinically relevant treatment protocols to model allogenic HSCT in mice are very rare. The paucity of these types of studies could limit the translatability of mouse models of disease.

As noted previously, conventional pretransplant conditioning protocols in mice (e.g. TBI) damage a variety of different tissues, producing numerous inflammatory cytokines and mediators that promote the activation and recruitment of alloantigen-specific T cells to different target tissues (Khuat et al., 2021; Hill and Koyama, 2020; Blazar et al., 2020; Zeiser et al., 2016; Castor et al., 2012; Mapara et al., 2006; Chewning et al., 2013; Dudakov et al., 2017; Schroeder and DiPersio, 2011; Shono et al., 2014, 2010; Stolfi et al., 2016; Zeng et al., 2017; Hill et al., 2021). Several preclinical and clinical studies suggest that the severity of aGVHD is directly related to the magnitude of intestinal injury produced by pretransplant conditioning protocols (Nalle et al., 2014; Nalle and Turner, 2015; Zeiser and Blazar, 2017a; Hill et al., 1997, 2021; Johansson and Ekman, 2007; Hill and Ferrara, 2000). These correlative data suggest that damage to the gut by myeloablative conditioning promotes the translocation of intestinal bacteria and their products into the gut interstitium, mesenteric lymph nodes and systemic circulation, where they initiate alloreactive T cell-mediated aGVHD.

An important question that has been debated over the past few years is whether gut damage is a cause or consequence of multi-organ aGVHD (Nalle and Turner, 2015). In an attempt to differentiate between these two possibilities, Nalle et al. (2014) injected unfractionated, alloreactive splenocytes into NK cell-depleted mice in which recombination activating gene-1 (Rag1) was inactivated (NK⁻/Rag1^−/−) to determine whether they would develop aGVHD (Table 1). Notably, these immunodeficient mice were engrafted with allogeneic T cells without the use of toxic pretransplant conditioning. This study demonstrated that intestinal damage is not required for development of aGVHD when MHC-mismatched splenocytes are engrafted. However, gut injury was required when minor histocompatibility antigen (miHA; Box 1)-mismatched (Box 1) splenocytes were used (Nalle et al., 2014). A major advantage of this model is that it creates immunodeficient mice without the use of tissue-damaging conditioning protocols. One potential limitation with the NK⁻/Rag1^−/− model is whether the pathogenetic immune responses are similar to those described in the more conventional mouse models that use TBI as the conditioning protocol.

Because unfractionated splenocytes were used as a source for disease-producing effector cells in this model, the specific T-cell populations responsible for inducing aGVHD were not determined. To address this question, our group investigated whether allogeneic CD4⁺ T cells alone could induce aGVHD when injected into immunodeficient NK⁻/Rag1^−/− mice or mice devoid of T, B and NK cells by virtue of deletions of their recombination activating gene-2 (Rag2; Box 1) gene and interleukin-2 receptor common γ (Il2rg) gene (McDaniel Mims et al., 2019). We found that adoptive transfer of naïve CD4⁺CD62L⁺CD25⁻ T cells from allogeneic donors into either mouse model induced clinical and histological features of aGVHD, including weight loss, inflammatory cytokine production and tissue inflammation (McDaniel Mims et al., 2019). In addition, engraftment of these alloantigen-specific T cells induced inflammation of the liver and skin, as well as severe anemia due to remarkable BM and spleen damage (McDaniel Mims et al., 2019). Interestingly, these mice did not develop histological evidence of inflammatory tissue damage to the gut or lungs. These findings demonstrated that CD4⁺ T cells are both necessary and sufficient to induce aGVHD in lymphopenic recipients in the absence of toxic pretransplant conditioning and confirm the studies of Nalle et al. (2014) demonstrating that intestinal damage is not required for the development of aGVHD in this model. The findings described above are similar to those reported using mouse models of xGVHD in which injection of human BM mononuclear cells into unconditioned, immunodeficient NBSGW mice induced aGVHD in the absence of tissue damage (Hess et al., 2021, 2020). Based upon these data, Hess et al. (2021) suggested that although cytotoxic inflammatory mediators are not required for initiation of aGVHD, they most likely influence the progression, prevalence and severity of disease.

Historically, the use of myeloablative conditioning has been reserved for relatively young patients because its use in older, medically infirm individuals increases morbidity and mortality (Gyurkocza and Sandmaier, 2014; Shimoni et al., 2016; Stolfi et al., 2016). These clinical observations prompted the development of reduced intensity conditioning (RIC) protocols to increase safety, reduce disease relapse and increase overall survival, thereby expanding the availability of HSCT (Gyurkocza and Sandmaier, 2014; Pingali and Champlin, 2015). We recently reported how RIC prior to engraftment of allogeneic T cells affects the onset, severity and location of aGVHD (Enriquez et al., 2023). Using a modification of a mouse model described by Chen et al. (2004) (Table 1), we found that intravenous administration of small numbers of flow-sorted CD4⁺CD25⁻ conventional T cells (Box 1) from allogeneic donors into recipients subjected to RIC (sublethal irradiation) induced severe wasting disease characterized by anemia and cytopenia (Box 1) that was associated with extensive BM failure and spleen damage (Enriquez et al., 2023). We did not observe damage to those tissues (gut, liver, skin) usually targeted in conventional, lethally irradiated mouse models of aGVHD. Adoptive transfer of allogeneic T cells into RIC-treated recipients also induced striking dysbiosis (Box 1), characterized by marked reductions in commensal/beneficial bacteria that are known to produce lactate and short-chain fatty acids (SCFAs) (Enriquez et al., 2023). It should be noted that lactate and certain SCFAs (e.g. butyrate) are known to promote hematopoiesis, and thus their loss can negatively affect the production of new blood cells (Lee et al., 2021). Variations of this RIC model have also been used to model immune-mediated aplastic anemia via adoptive transfer of unfractionated, allogeneic lymph node or spleen cells into sublethally irradiated recipients (Arieta Kuksin et al., 2015; Chen et al., 2015b). These studies, as well as findings from clinical studies, suggest that hematopoietic tissues, such as the BM and spleen, are particularly sensitive to alloreactive T cell-mediated damage (Chen et al., 2015a; Chewning et al., 2013; McDaniel Mims et al., 2019; Muskens et al., 2021; Shono et al., 2014, 2010; Zeiser and Teshima, 2021). Indeed, alloreactive T cells appear to selectively traffic to and damage BM and spleen in RIC-treated mice (Fig. 3). Taken together, these findings suggest that although RIC dramatically reduces tissue damage in non-lymphoid tissue, hematopoietic tissue injury can persist, rendering mice (and possibly RIC-treated patients) susceptible to recurrent infections and bacteremia (Box 1).

It is well known that, in addition to promoting microbial translocation, myeloablative conditioning protocols can dramatically alter intestinal bacterial populations prior to HSCT. Alterations in gut bacteria (called dysbiosis) produced by these conditioning protocols include marked loss of diversity and abundance of anaerobic bacteria (e.g. Clostridia), whereas the abundance of potential pathogens (e.g. Enterococci, Streptococci and/or Escherichia) is increased (Fredricks, 2019; Peled et al., 2020; Shono et al., 2016; Staffas et al., 2017; van Lier et al., 2023). These data prompted investigators to assess the protective effects of antibiotic administration prior to and/or following pretransplant conditioning. Some studies have reported that antibiotic treatment of mice or humans undergoing HSCT attenuates the onset and/or severity of aGVHD (Fredricks, 2019; Schwab et al., 2014), whereas other studies reported that administration of certain antibiotics can exacerbate aGVHD and increase mortality by damaging bacterial communities that are critical for maintaining intestinal epithelial cell viability and mucosal barrier function (Fredricks, 2019; McDaniel Mims et al., 2021; van Lier et al., 2023). We found that oral administration of a broad-spectrum antibiotic cocktail for 7 days prior to and 4 weeks following engraftment of allogeneic CD4⁺ T cells into non-conditioned, immunodeficient (e.g. NK⁻/Rag1^−/−) mice exacerbated aGVHD-induced BM and spleen damage compared to that in their untreated controls engrafted with allogeneic T cells (McDaniel Mims et al., 2021). Tissue damage was associated with severe anemia and a remarkable reduction in fecal bacterial load and diversity (McDaniel Mims et al., 2021). Our data agree with clinical findings suggesting that prolonged pre- and post-transplant administration of high-dose, broad-spectrum antibiotics in patients undergoing allogeneic HSCT is a major contributor to microbial damage and severe dysbiosis (Fredricks, 2019; Staffas et al., 2017; van Lier et al., 2023). Indeed, these observations have important hematologic implications given that microbe-associated molecular patterns (e.g. lipopolysaccharide, peptidoglycan, flagellin), as well as certain bacterial-derived products (SCFAs, lipopolysaccharide, lactate, etc.), interact with HSCs and hematopoietic progenitor cells (HSPCs) to induce their proliferation and differentiation into all of the cellular elements of blood (Josefsdottir et al., 2017; Lee et al., 2021).

Recent ground-breaking, preclinical studies have reported the efficacy of a novel pretransplant conditioning protocol that uses an anti-mouse CD45 antibody-drug conjugate (ADC) that is coupled to the cytotoxic DNA crosslinker tesirine (Saha et al., 2022, 2024). CD45-ADC binds directly to all BM immune cells, HSCs and HSPCs that express CD45 on their surface, thereby initiating its internalization and cell death, as free tesirine is toxic to these cells. When unconditioned C57BL/6 (H-2^b) (Box 1) mice were treated with a single dose of CD45-ADC, they exhibited a rapid and marked loss of BM HSCs and HSPCs within 3 days post-treatment (Saha et al., 2022). This response could be partially rescued by the engraftment of ADC-treated Bl6 recipients with fully MHC-mismatched BM cells from CByJ.SJL(B6)-Ptprc^a/J CD45.1 (H-2^d) mice (Box 1), resulting in 95% donor chimerism (Box 1) in peripheral blood, with donor myeloid cells, B cells and T cells achieving 95% long-term engraftment (Saha et al., 2022). In a recent follow-up study, these authors compared the development of aGVHD and survival in CD45-ADC-conditioned C57BL/6 mice that received allogeneic BM cells from BALB/c (H-2^d) mice (Box 1) to C57BL/6 mice that were conditioned with lethal TBI and then engrafted with BALB/c BM cells (Saha et al., 2024). They found that 88% of CD45-ADC-conditioned mice (Box 1) survived for more than 18 weeks post-transplant, whereas only 10% of TBI-conditioned recipients survived the same time following transplantation (Saha et al., 2024). Importantly, clinical GVHD scores in CD45-ADC-conditioned mice were significantly reduced compared to those in their TBI-conditioned counterparts (Saha et al., 2024). Similar results were observed when donors and recipients were reversed (C57BL/6→BALB/c). Using these same protocols, Saha et al. (2024) then determined whether CD45-ADC conditioning followed by allogeneic HSCT would prove effective in treating three different mouse models of Fanconi anemia (Box 1). Fanconi anemia is an inherited genetic disease characterized by BM failure, anemia, pancytopenia, myelodysplasia and leukemia, for which HSCT is the only available cure (Saha et al., 2024). Saha et al. (2024) found that CD45-ADC conditioning followed by transplantation of fully allogeneic BM significantly increased survival and reduced aGVHD compared to that in Fanconi anemia mice that received lethal TBI and allogeneic HSCT. In addition, these investigators reported that administration of anti-human CD45-ADC to Rhesus monkeys (Box 1) produced robust myeloablation that was similar to that produced by lethal TBI (Saha et al., 2024). Taken together, these data demonstrate that CD45-ADC conditioning enhances engraftment and hematopoiesis following allogeneic HSCT while significantly reducing aGVHD. Results from these studies are particularly exciting as future preclinical HSCT investigations using this novel conditioning protocol could reveal new strategies to protect recipients from alloreactive T cells.

Another factor that is likely to limit bench-to-bedside translation in mouse models is their housing environment. Specific pathogen-free (SPF) animal care facilities are designed to exclude pathogenic viruses, bacteria and parasites by using multiple ‘barriers’ and restricted access of staff, as well as by employing high-efficiency particulate air (HEPA) filters and sterilized drinking water for all cages. These precautions create a uniform ultra-hygienic environment that provides more reproducible immune responses and phenotypes in the animals housed in this facility. Despite the use of these ultra-hygienic conditions, the phenotypes of different mouse models of inflammatory diseases may be markedly different from one institution to another (Enriquez et al., 2020; Ericsson et al., 2015; Ivanov et al., 2009; Kriegel et al., 2011; Webb et al., 2018). This could be due to differences in animal vendor, diet and/or water treatment, any one of which may markedly alter the gut microbiota of mice (Ericsson et al., 2015; Kriegel et al., 2011; Webb et al., 2018). Major differences in the incidence and severity of chronic gut inflammation in mouse models of inflammatory bowel disease have been shown to occur following relocation of mice to a different animal care facility or change of vendor (Ericsson et al., 2015; Fox et al., 2011; Webb et al., 2018; Stepankova et al., 2007; Yang et al., 2013). An example of how housing conditions can affect the onset and/or severity of aGVHD, Li et al. (2020) recently reported that injection of unfractionated, allogeneic C57BL/6 lymph node cells into CByB6F1 recipients housed under non-SPF conditions [e.g. conventional housing conditions (Box 1)] induced significantly less allogeneic T-cell expansion and BM damage compared to that in mice housed under SPF conditions. These responses were associated with greater microbial diversity within mice housed under convention housing conditions compared to that in SPF-housed recipients (Li et al., 2020). These findings illustrate how differences in the composition of intestinal microbiota in mice housed under different conditions within the same animal facility may alter the onset and/or severity of inflammation, thereby affecting the reproducibility of these models. In contrast to the ‘ultra-clean’ SPF environment, humans are continuously exposed to a spectrum of different pathogenic and nonpathogenic microorganisms that both protect us and shape our immune system to fight future infectious microbes (Enriquez et al., 2020). Investigators have reported that mice housed under SPF conditions have remarkably underdeveloped immune systems compared to those of ‘dirty’ mice housed in pet stores or residing in the wild [i.e. feral mice (Box 1)] (Beura et al., 2016; Liu et al., 2024; Rehermann et al., 2024). Studies have shown that mice need to be exposed to a diverse array of naturally occurring microorganisms to develop a mature immune system that more closely resembles that of humans (Beura et al., 2016; Masopust et al., 2017; Rosshart et al., 2017; Burger et al., 2023; Liu et al., 2024). And, as discussed below, feral mice possess a diverse intestinal microbiota that promotes the development of a fully functional immune system, protecting them from numerous and potentially lethal infectious agents, carcinogens and inflammatory mediators.

There are now numerous examples of the translational importance of microbiota for the modeling of human infectious and inflammatory diseases. For example, Beura et al. (2016) reported that when pet store mice or SPF mice co-housed with pet store mice were infected with Listeria monocytogenes (Box 1), their bacterial load was reduced by more than 10,000-fold compared to that in standard SPF animals. Another study reported that colonic inflammation and numbers of invasive adenomas (Box 1) and adenocarcinomas (Box 1) were significantly reduced in feral microbiota-colonized mice relative to SPF-colonized controls in a mouse model of colorectal cancer (Rosshart et al., 2017). The National Institute of Allergy and Infectious Diseases recently convened a 2-day workshop to discuss how exposure of mice to diverse microbial communities may improve our understanding of human immune system development and disease (Liu et al., 2024). Participants, including several of the investigators referenced above, presented data related to the development and use of microbial-experienced mouse models of infectious and inflammatory diseases. In addition, the consortium discussed the challenges and possible solutions when using ‘dirty’ mice to model human immunity. A recent ground-breaking study by Lo et al. (2024) reported how the use of laboratory mice colonized with feral microbiota could advance our understanding of the immunopathogenesis of immune-related adverse events induced by anti-tumor biologics called immune checkpoint inhibitors. Administration of checkpoint inhibitors such as anti-CTLA-4 (Box 1) mAb induces severe colitis in a subset of cancer patients undergoing treatment with this immunotherapeutic antibody (Waldman et al., 2020). Unfortunately, checkpoint inhibitors fail to induce colitis in mice housed under SPF conditions. Lo et al. (2024) reported that treating feral microbiota-colonized mice with anti-CTLA-4 antibodies induced robust colitis, demonstrating the feasibility of using feral microbiota as models of inflammatory diseases.

Another factor that can affect the translatability of mouse models of disease is housing temperature. All animal care facilities in the USA house mice at a standard temperature (ST) range of 22-24°C. This temperature is below the murine thermoneutral temperature (TNT) range of 30-32°C (Bucsek et al., 2018; Hylander et al., 2019; Hylander and Repasky, 2016; Karp, 2012). It is well known that ST housing creates mild, but chronic, cold stress that can alter a number of different immune responses compared with those in mice housed at their TNT (Enriquez et al., 2020; Bucsek et al., 2018; Hylander et al., 2019, 2023; Hylander and Repasky, 2016; Karp, 2012; James et al., 2023). Both primary and secondary lymphoid tissue are innervated by the sympathetic nervous system, which can be activated by ST-induced cold stress, resulting in the release of norepinephrine, which modulates a number of different immune cell responses in a receptor/ligand-specific manner (Hylander et al., 2023; James et al., 2023). ST-induced, norepinephrine-mediated alterations in disease phenotypes have also been reported in mouse aGVHD, cancer and cardiovascular disease models (Hylander et al., 2023; James et al., 2023). Furthermore, certain gut microbial communities are significantly altered in mice housed at ST versus TNT (Hylander et al., 2023). Indeed, changes in environmental temperature may exert profound effects on intestinal homeostasis and microbial composition in mice (Chevalier et al., 2015; Hylander et al., 2019, 2023; Karl et al., 2018). An interesting observation, reported by numerous investigators using the conventional, lethally irradiated mouse model of aGVHD, is that engraftment of allogeneic BM alone into recipients subjected to lethal irradiation and housed at ST does not induce aGVHD unless the BM is supplemented with a source of allogeneic T cells (e.g. splenocytes and/or isolated T cells) (Hill and Koyama, 2020; Schroeder and DiPersio, 2011; Stolfi et al., 2016; Zeiser and Blazar, 2017a). This contrasts markedly with human HSCT protocols, which attempt to remove all allogeneic T cells prior to HSC administration. Leigh et al. (2015) reported that lethally irradiated mice housed at TNT and engrafted with BM alone did in fact develop aGVHD, suggesting that ST-mediated cold stress was suppressing the development of inflammatory tissue damage. Furthermore, they found that disease suppression at ST was mediated by the interaction between norepinephrine and the β2-adrenergic receptor (Leigh et al., 2015). Because most aGVHD mouse models use lethal TBI as the pretransplant conditioning protocol and ST housing, we investigated whether housing temperature affects the onset and severity of aGVHD in the RIC mouse model of T cell-mediated BM and spleen damage described above. In contrast to the study reported by Leigh et al. (2015), which used a conventional model of aGVHD, we found that the survival of allogeneic T cell-engrafted mice housed at ST to be significantly reduced compared to that of their counterparts housed at TNT (Enriquez et al., 2023). We are currently investigating the mechanisms by which ST decreases survival in this model. Together, these findings suggest that relatively modest differences in housing temperature may produce major changes in disease phenotype.

A report recently estimated that >5 million mice are used every year according to annual use reports for rats and mice from 16 of the top 30 National Institutes of Health (NIH)-funded US research institutions. The author extrapolated this number to all US institutions, estimating that >100 million mice are used in research in the USA annually (Carbone, 2021). Others have argued that this is an overestimation and that annual mouse use lies somewhere between 25 and 40 million/year in the USA (Grimm, 2021). Despite these different estimates, it is fair to say that large numbers of mice are used for biomedical research. Although Carbone (2021) could not provide a breakdown of the inbred versus outbred mice used by the different institutions, it is highly likely that inbred mice are used for the large majority of preclinical studies investigating the immunopathogenesis of inflammatory diseases. Indeed, all mouse models of aGVHD use young, lean and healthy inbred strains of mice as surrogates for genetically diverse humans (Table 3). In contrast, patients undergoing HSCT are likely to be medically infirm, older and/or obese (Blazar et al., 2020; Stolfi et al., 2016). Preclinical studies have shown that altering one or more of these risk factors could have a major impact on disease phenotypes in mouse models (Blazar et al., 2020; Mirsoian et al., 2014). Although the use of inbred mice has been critical to our understanding of immunology and immunopathology, it is not clear whether the immune responses of these genetically constrained animals recapitulate the diverse immune responses found in humans (Enriquez et al., 2020). It is known that human T-cell responses are more heterogeneous than those of inbred mice. In addition, specific immune responses and disease phenotypes can be quite different in inbred versus outbred mice (Graham et al., 2017, 2015; Martin et al., 2017; Nikodemova and Watters, 2011; Reichenbach et al., 2013). Thus, the exclusive use of healthy inbred mice could limit the bench-to-beside translation of data obtained from aGVHD studies (Enriquez et al., 2020; Blazar et al., 2020; Stolfi et al., 2016).

To our knowledge, no studies have used outbred mice to study the immunopathogenesis of aGVHD. Therefore, our group engrafted RIC-treated outbred CD1 mice with inbred (or outbred) allogeneic CD4⁺ T cells and then assessed recipient mice housed at ST or TNT for tissue damage. We found that engraftment of large numbers of allogeneic T cells into RIC-treated CD1 mice did not induce aGVHD in any tissue in mice housed at ST or TNT. These data suggest that outbred mice subjected to RIC are resistant to T cell-mediated aGVHD in this and possibly other models of aGVHD (Enriquez et al., 2023). Because MHC II disparity alone was incapable of driving T cell-mediated tissue damage in these outbred mice, it is reasonable to suggest that uncharacterized, regulatory immune mechanisms contribute to disease resistance. Other investigators have reported similar results when attempting to induce autoimmune encephalomyelitis or diabetes in outbred recipients (Marin et al., 2014; Maron et al., 1999). Although we did not determine the mechanisms by which the outbred CD1 mice remained disease free in our study, we did observe significantly larger numbers of BM- and spleen-residing B cells and myeloid cells relative to those in inbred mice that developed aGVHD (Enriquez et al., 2023). These data, together with those reported by Marin et al. (2014) for their mouse model of autoimmune encephalitis, suggest that within the B-cell and/or myeloid cell populations present in T cell-engrafted CD1 mice reside substantial numbers of immunosuppressive regulatory B cells and/or myeloid derived suppressor cells (Enriquez et al., 2023). It will be interesting to see whether future studies can identify the mechanisms responsible for suppression of aGVHD in outbred mice.

Although outbred mice have been used extensively in toxicological, pharmacological and infectious disease research, their use in modeling inflammatory diseases has been limited (Enriquez et al., 2020; Chia et al., 2005). In addition, the genotypes of commercially available stocks of outbred mice are either unavailable or poorly defined (Chia et al., 2005). These concerns have led investigators to generate genetically diverse reference populations of mice that are referred to as Collaborative Cross mice (Collaborative Cross Consortium, 2012; Threadgill and Churchill, 2012). These highly outbred mice were generated by combinational breeding schemes using eight unique and genetically diverse founder strains of mice that included three inbred strains, two inbred mouse models of diseases (type 1 diabetes and obese/type 2 diabetes mice) and three wild-derived strains (Graham et al., 2017; Leist and Baric, 2018; Phillippi et al., 2014; Threadgill and Churchill, 2012). Using multi-combinational and sibling breeding protocols, more than 60 Collaborative Cross recombinant inbred strains have been generated. Each mouse within a specific Collaborative Cross recombinant inbred strain is genetically identical. It will be interesting to determine how these mice respond to allogeneic HSCT protocols.

The use of mouse models to identify new and more effective therapeutic strategies for the treatment of inflammatory diseases is a mainstay for preclinical studies. Unfortunately, the bench-to-bedside transition of promising therapeutic strategies observed in mouse models of aGVHD has been disappointing. The reasons for the limited translation of preclinical data have not been completely defined; however, several concerns associated with murine models have raised important questions regarding the translational nature of using lethal TBI as a conditioning protocol for inbred mice housed under ultra-hygienic and stress-inducing temperatures. Indeed, these conditioning and housing protocols, as well as the use of inbred strains of mice, markedly alter murine microbiota, their immune responses and the immunopathogenesis of aGVHD. Designing studies that more accurately reflect the human environment and treatment protocols are likely to increase the translatability of data obtained from preclinical studies.