Novel developments in the prophylaxis and treatment of acute GVHD

Allogeneic hematopoietic cell transplant (allo-HCT) is a personalized cellular therapy that induces the highest durable remission rate of all treatment options in patients with high-risk myeloid malignancies.¹ Acute graft-versus-host disease (aGVHD) remains a major barrier for long-term success after allo-HCT. The pathophysiology of aGVHD involves antigen presentation by hematopoietic and nonhematopoietic cells,² inflammation driven by release of damage-associated molecular patterns, effects of chemokines expressed in inflamed tissues, and recruitment of effector cells, including T cells, neutrophils,³ and monocytes, to target tissues.⁴ Conditioning intensity is associated with the risk for aGVHD, which can be explained by the release of sterile triggers of inflammation from stressed or dying cells, which activate the immune system and enhance aGVHD.^5-7 

aGVHD typically presents as a maculopapular rash, hyperbilirubinemia, or large-volume diarrhea.^4,8 The National Institutes of Health consensus criteria are widely used to diagnose aGVHD,⁹ and its severity can be graded using several systems.^10,11 Recently, given the variability in diagnosing and grading aGVHD, the Mount Sinai Acute GVHD International Consortium (MAGIC) established an international expert consensus opinion to standardize the diagnosis and clinical staging of aGVHD.¹² 

The overall incidence of grades 2 to 4 aGVHD after allo-HCT is 40% to 50%, with ≈15% of patients developing grade 3 to 4 (severe) aGVHD, despite prophylaxis.^4,13,14 First-line treatment with systemic, high-dose glucocorticoids results in a 60% response rate in patients with grade 2 disease and a 30% to 40% response rate in those with grade 3 to 4 disease.^15,16 In the Ruxolitinib for Glucocorticoid-Refractory Acute Graft-versus-Host Disease (REACH2) trial, outcomes of patients with corticosteroid-refractory aGVHD (SR-aGVHD) were poor, with a median survival of 11 months, even if treated with ruxolitinib, which was approved for SR-aGVHD in 2019.¹⁷ Real-world evidence of lower gastrointestinal (GI) SR-aGVHD confirms the dismal outcomes for these high-risk patients, with a 1-year nonrelapse mortality (NRM) rate of 39.2% and a median survival of 11.9 months.¹⁸ Several critical advances in the prevention and treatment of aGVHD have been achieved recently (Figure 1), and we discuss these recent developments as well as ongoing and future clinical trials.

Historically, at most centers, standard GVHD prophylaxis has consisted of a calcineurin inhibitor (CNI), such as tacrolimus (TAC) or cyclosporine (CSA), paired with an antimetabolite, such as methotrexate (MTX) or mycophenolate mofetil (MMF), with or without the addition of an antithymocyte globulin (ATG) product.^19,20 Choice of regimen has usually been based on factors such as donor type, degree of human leukocyte antigen (HLA) match, and conditioning intensity.

Most commonly for 8 of 8 HLA-matched related (MRD) and unrelated (MUD) donor allo-HCT, undergoing either myeloablative (MA) or reduced-intensity conditioning (RIC), GVHD prophylaxis consists of MTX with a calcineurin inhibitor (either TAC or CSA). TAC/MTX led to a lower incidence of grade 2 to 4 aGVHD and extensive chronic GVHD (cGVHD), when compared with CSA/MTX in MRD allo-HCT.⁸ In MUD allo-HCT, TAC/MTX led to a significant decrease in grade 2 to 4 aGVHD, compared with CSA/MTX.²¹ Yet, the choice between TAC and CSA is mainly guided by institutional preference and access. A combination of sirolimus and TAC can also be considered in MA MRD allo-HCT based on similar long-term efficacy compared with TAC/MTX, with more rapid engraftment and less mucositis, yet more sinusoidal obstruction syndrome and thrombotic microangiopathy.²² 

The addition of an ATG product in MUD allo-HCT has been commonly used. Both ATLG (anti-T-lymphocyte globulin; Neovii) and thymoglobulin (ATG; Genzyme-Sanofi, Lyon, France) have been investigated in randomized phase 3 trials.^20,23-25 The addition of thymoglobulin led to a decrease in systemic immunosuppression at 12 months without impacting NRM, relapse, or survival.²⁵ Data regarding ATLG are conflicting. Two European studies illustrated a significant reduction in cGVHD with the addition of ATLG in both related²⁰ and unrelated²³ allo-HCT, without compromising survival. However, a predominantly US study found that the addition of ATLG to TAC/MTX, in unrelated allo-HCT, led to lower progression-free and overall survival (OS), despite a reduction in both grade 2 to 4 aGVHD and moderate-severe cGVHD. An unplanned post hoc analysis explored the relationship between ATLG and absolute lymphocyte count (ALC) and found that low ALC count (<0.1 × 10⁹/L) at the time of ATLG administration negatively impacted survival.²⁴ It has been proposed that models incorporating ALC-based dosing could lead to optimum dosing.²⁶ Exposure to ATG before and after HCT has also shown to impact outcomes, with pre-HCT exposure leading to more favorable outcomes, but this needs prospective validation.²⁷ 

Clearly, recipients of mismatched-unrelated donor (mMUD) allo-HCT, receiving conventional GVHD prophylaxis, are at an increased risk for developing severe aGVHD, and prophylaxis with posttransplant cyclophosphamide (PTCy) or abatacept combinations has recently yielded favorable outcomes for these patients.^28-30 For haploidentical (haplo) HCT, PTCy/TAC/MMF was shown to yield a low incidence of GVHD and NRM, establishing it as the standard approach at many centers.³¹ 

Over the past few years, several agents have appeared promising for GVHD prophylaxis, and we discuss the results of recent trials with these agents and how they may impact clinical practice (Table 1).

The Blood and Marrow Transplant Clinical Trials Network (BMT CTN) 1703³² and Dutch-Belgian Cooperative Trial Group for Hemato-Oncology (HOVON) 96³³ phase 3 trials both investigated PTCy in RIC from predominantly 8 of 8 MRD/MUD allo-HCT. In BMT CTN 1703, PTCy/TAC/MMF (n = 214) was compared with TAC/MTX (n = 217); and in HOVON-96, PTCy/CSA (n = 99) was compared with CSA/MMF (n = 52). Both trials demonstrated a significantly higher 1-year GVHD/relapse or progression-free survival (GRFS) with PTCy (BMT CTN 1703, 52.7%; and HOVON-96, 45%) compared with the control arm (BMT CTN 1703, 34.9%; and HOVON-96, 21%). A criticism of both trials is that the control arm did not include ATG, which is the standard at many European centers, and a prospective trial comparing PTCy vs an ATG inclusive regimen is currently accruing in unrelated allo-HCT (NCT05153226). The observed difference in GRFS with PTCy in both studies was mainly attributable to reduction in both aGVHD and cGVHD, with no difference in relapse or survival. Long-term follow-up is needed to determine whether the lower incidence of GVHD translates into a survival benefit or if increased late relapses will occur. Another concern with PTCy is an increased rate of cytomegalovirus reactivation and other infections, which may negate the benefit of decreased cGVHD.^34,35 

Vedolizumab is a humanized monoclonal antibody directed against the α4β7 integrin which is expressed on lymphocytes and is essential to GI trafficking. α4β7 integrin was found to be upregulated on the surface of naive and memory T cells in patients who subsequently developed intestinal aGVHD after allo-HCT.^36,37 Vedolizumab’s mechanism of action, disrupting the homing of T cells to gut lymphoid tissue, may be effective for prevention of intestinal aGVHD.³⁸ The addition of vedolizumab to a CNI with MTX or MMF (vedolizumab = 168; placebo = 165) recently resulted in a higher lower intestinal aGVHD-free survival by day +180 (85.5% vs 70.9%; P < .001), in RIC/MA 7 of 8 or 8 of 8 MUD allo-HCT in an international phase 3 randomized, placebo-controlled, double-blind trial.³⁹ Vedolizumab was also associated with a lower incidence of lower GI aGVHD and lower GI aGVHD-free and relapse-free survival by day +180. Long-term follow-up and further data regarding maximum stage of GI GVHD, impact on skin and liver aGVHD, effect on cGVHD, infectious complications, and relapse incidence are needed to facilitate interpretation.

In a recent phase 3 trial, adding sirolimus to CSA and MMF in RIC 7 of 8 or 8 of 8 MUD allo-HCT resulted in a lower incidence of grade 2 to 4 aGVHD, lower NRM, and higher OS.⁴⁰ No difference was observed for grade 3 to 4 aGVHD and relapse. However, the study was closed early by the Data and Safety Monitoring Board based on a survival advantage with triplet therapy, precluding an analysis for grade 3 to 4 aGVHD. Other phase 3 trials studying the addition of sirolimus in both the MA and RIC settings have not shown significant differences in overall outcomes.^22,41 

Abatacept, a T-cell costimulation blockage agent, was the first agent specifically approved in the United States for GVHD prevention. In a randomized phase 2 trial (backbone of MTX + CSA or TAC) in 8 of 8 HLA-matched unrelated allo-HCTs, abatacept resulted in a numerical (6.8% vs 14.8%; P = .13) reduction in day +100 grade 3 to 4 aGVHD. In a single-arm study of 7 of 8 HLA-matched grafts, the incidence of day +100 grade 3 to 4 was 2.3% with abatacept, comparing favorably with the 30.2% incidence seen in a Center for International Blood and Marrow Transplant Research (CIBMTR) cohort (CNI/MTX).³⁰ One limitation to highlight is the lack of PTCy-based GVHD prophylaxis in the CIBMTR control arm as studies with PTCy, in mMUD allo-HCT, have reported a lower incidence of day +100 grade 3 to 4 aGVHD (≈18%).^28,29,42 Given the promising activity of both abatacept and PTCy in mMUD setting, studies comparing the two, in this population, are warranted.

Using PTCy/TAC/MMF (n = 125) in MUD allo-HCT, investigators from the University of Minnesota reported a cumulative incidence of day +100 grade 3 to 4 aGVHD, 1-year cGVHD requiring treatment, 2-year relapse, OS, and GRFS of 4% (no grade 4), 4%, 25%, 80%, and 57%, respectively. Compared with an institutional cohort using CSA/MTX, PTCy was associated with superior outcomes, including GRFS.⁴³ However, in a phase 3 BMT CTN (1301) trial investigating CNI-free GVHD prevention using bone marrow grafts, single-agent PTCy was not superior to TAC/MTX when evaluating a primary end point of 1-year moderate to severe cGVHD relapse-free survival.⁴⁴ Whether the addition of TAC/MMF to PTCy outperforms CNI/MTX in MA HLA-matched allo-HCT needs to be prospectively investigated.

T-cell–depleting strategies have been long explored to prevent both aGVHD and cGVHD. Ex vivo CD34-positive selection results in up to 5-log reduction of T cells from a peripheral blood stem cell graft. In BMT CTN 1301, ex vivo CD34-selected HLA-matched T-cell–depleted peripheral blood stem cell grafts without additional immunosuppression resulted in lower rates of cGVHD, compared with standard HLA-matched bone marrow grafts with TAC/MTX. Although CD34 selection was not associated with an increased risk of relapse, it did lead to inferior OS because of higher NRM secondary to infection and organ failure when compared with TAC/MTX.⁴⁴ 

The data previously presented have caused many centers to reevaluate long-standing standard platforms of GVHD prevention. Many have adopted PTCy as standard in RIC allo-HCT from both matched and mismatched donors. Both abatacept (in the United States) and PTCy-based regimens are routinely being used in MA allo-HCT. We await full data on vedolizumab to better understand if regulatory approval will be pursued. It also remains to be seen whether these newer agents also mitigate the risk of cGVHD. In the next section, we highlight preliminary data of novel combinations currently under investigation and discuss other potential strategies (Table 2).

One area of exploration is the safety and efficacy of adding novel agents to a PTCy-based platform. A successful combination, however, may need to rely on optimization of PTCy dose. Wang et al added low-dose PTCy (14.5 mg/kg on days 3 and 4) to their standard ATG backbone, in haplo-HCT, and demonstrated lower rates of GVHD and NRM and higher GRFS with this strategy.⁴⁵ Preliminarily, reducing the PTCy dose by 50% on days 3 and 4 in combination with sirolimus/MMF, and employing bone marrow grafts, resulted in earlier engraftment and similar GVHD outcomes in MA haploidentical bone marrow transplant.⁴⁶ Dose optimization of PTCy in HLA-matched allo-HCT is also being investigated. In a phase 2 trial, 1 dose of PTCy (on day 3) with TAC/MMF in MA MUD allo-HCT was effective for aGVHD control but did not reduce the significant risk of cGVHD.⁴⁷ Moving forward, lower doses of PTCy will undoubtedly be employed as well as the addition of novel agents, such as vedolizumab, abatacept, or ruxolitinib, to a reduced-dose PTCy-based regimen. In a phase 1b/2 trial in haplo-HCT (n = 46), PTCy, abatacept, and a short course of TAC resulted in a low incidence of aGVHD (day +100 grade 3-4, 5.1%) and cGVHD (1-year moderate-severe, 17.1%). The 1-year relapse incidence, OS, and GRFS were 8.3%, 90.6%, and 71.8%, respectively.⁴⁸ Outcomes with longer follow-up for this strategy are needed to confirm these promising findings.

The JAK-STAT signaling pathway plays a crucial role in the pathogenesis of GVHD through lymphocyte activation and tissue inflammation. Inhibiting this pathway with ruxolitinib, an oral JAK 1/2 inhibitor, has been successful in both corticosteroid-refractory aGVHD and cGVHD.^17,49 JAK inhibitors are now being investigated as GVHD prophylaxis in multiple ongoing trials. In 20 patients with myelofibrosis, PTCy and ruxolitinib led to a low incidence of GVHD and relapse, but a high rate of myelosuppression, leading to ruxolitinib dose modifications.⁵⁰ Given the cytopenias known to be associated with ruxolitinib use, it is imperative to understand its ideal dose and schedule when used as GVHD prevention. Itacitinib, a selective JAK1 inhibitor, was combined with PTCy/TAC/MMF in haplo-HCT. All patients engrafted with no cases of grade 3 to 4 aGVHD. The 6-month relapse incidence, OS, and GRFS were 5.5%, 90%, and 83%, respectively.⁵¹ 

Adding JAK inhibitors to conventional CNI-based platforms is yet another potential strategy being studied. In a phase 2a study of itacitinib/TAC/sirolimus in RIC-matched HCT, grade 3 to 4 aGVHD was 5% and 1-year GRFS was 51%.⁵² In a pilot study of peritransplant ruxolitinib (D - 3 to D + 30) with TAC/sirolimus in RIC-matched HCT for myelofibrosis, the incidences of day +100 grade 3 to 4 aGVHD and 1-year cGVHD were 11% and 42%, respectively.⁵³ In an interim analysis of 26 patients with myelofibrosis treated with ruxolitinib (D - 1 to 12 months) with TAC/MTX after RIC-matched allo-HCT, very few cases of moderate-severe cGVHD were observed (5%).⁵⁴ Future trials incorporating ruxolitinib with standard GVHD prophylaxis will optimize dosing in the peri-HCT period and continue ruxolitinib for a longer duration after HCT to better prevent cGVHD given the previous compelling results.⁵⁵ 

Orca-T (HLA-matched donors) and Orca-Q (haploidentical donors) are investigational precision-engineered cell therapies manufactured through high-throughput cell sorting technology. In 138 patients undergoing MA HCT receiving Orca-T with either single-agent TAC (n = 127) or sirolimus (n = 7), day +180 grade 3 to 4 aGVHD and 1-year moderate-severe cGVHD were 4% and 5%, respectively, and 1-year OS and GRFS were 90% and 71%, respectively.⁵⁶ In 21 patients undergoing MA haploidentical HCT with Orca-Q and single-agent TAC, 1 patient developed grade 3 aGVHD and none developed moderate-severe cGVHD, leading to 1-year GRFS of 71%.⁵⁷ Orca-T is being investigated in a phase 3 trial (NCT05316701) compared with standard TAC/MTX, and Orca-Q continues to enroll on the current single-arm protocol.

α₁-Antitrypsin (AAT), a serine protease inhibitor, demonstrated a day 28 overall response rate (ORR) of 65% and a complete remission (CR) rate of 35%, in SR-aGVHD, with a tolerable safety profile.⁵⁸ A phase 3 trial investigating AAT for GVHD prevention in MA-matched transplants is ongoing (NCT03805789), and results are awaited.

As several variables, such as disease relapse, aGVHD, cGVHD, and OS, need to be accounted for in allo-HCT trials, the BMT CTN innovated the use of HCT-specific composite end points, such as GRFS and cGVHD relapse-free survival, both of which have served as the primary end point in GVHD prevention trials.⁵⁹ The benefit of composite end points is that they combine several clinically significant end points and reduce the challenge of controlling or censoring competing events, leading to a smaller sample size.⁶⁰ However, the assumption that all components carry equal or even comparable significance is not always satisfied. For example, in GRFS, equal severity is given to aGVHD, cGVHD, relapse, and death. Innovations in therapy of aGVHD and cGVHD or for relapse after HCT alter the gravity of these events. Moreover, as improvements have been achieved in GVHD prevention, continuing to use GRFS as the primary end point might lead to disease relapse differentiating GVHD prevention regimens in a large trial. This is problematic given the heterogeneity of relapse risk and the changing landscape of pre-HCT and post-HCT therapies. Yet, conducting randomized trials using aGVHD or aGVHD-free survival as the primary end point may become exceedingly large with the diminishing number of events.^59,60 

For aGVHD treatment trials, overall response on day 28 is a widely accepted primary end point, given that it is predictive of survival and NRM.⁶¹ However, response assessment on day 28 can be confounded by clinical variables, such as medications or infections, and is subject to interobserver variation.⁶² A composite end point of 6-month freedom from treatment failure, defined as absence of death, relapse/progression, or change in systemic therapy within 6 months of starting initial therapy and before cGVHD diagnosis, can also be considered as the primary end point given that the outcomes being measured are less influenced by the variations of determining clinical response to aGVHD.^62,63 However, limitations of 6-month freedom from treatment failure may include variability in initial corticosteroid dose and duration, treating cGVHD requiring systemic therapy as a competing risk, and, more recently, the early incorporation of ruxolitinib in the treatment for aGVHD.^62,63 

Approximately 30% to 60% of patients develop clinically significant aGVHD despite standard prophylaxis. Systemic glucocorticoids are used initially for grade ≥2 aGVHD with an ORR of 50% to 70%.^64,65 Despite the known metabolic, endocrine, immunologic, neuropsychiatric, cardiovascular, and GI adverse effects, glucocorticoid monotherapy remains the standard for frontline therapy. The addition of MMF in BMT CTN 0802⁶⁵ and itacitinib in GRAVITAS-301⁶⁴ did not improve outcomes, adding to the list of negative studies for upfront therapy for aGVHD.^66-68 Conducting successful trials for aGVHD treatment has been challenging as trial design relies on historical controls, leading to overperformance by the placebo arm. Furthermore, stringent trial eligibility criteria have led to enrolling a higher proportion of low-risk patients, resulting in dilution of any potential effect of an intervention.

BMT CTN 0802 included all grades of aGVHD (66% grade 1-2), whereas GRAVITAS-301 included grade 2 to 4 aGVHD (70% grade 2).^64,65 Although treatment for grade ≥2 aGVHD is glucocorticoids, modern efforts should likely be risk stratified by either clinical or biomarker-based systems. For high-risk patients, this may mean adding novel agents to standard glucocorticoids, whereas for low-risk patients, corticosteroid deescalation or even nonsteroid approaches are being pursued. Clinically, the Minnesota scoring system⁶⁹ is the most widely used, yet <15% of patients will qualify as high risk and no category exists for very low-risk patients. Biomarker-driven stratification using the MAGIC algorithm probability (MAP) has been shown to predict mortality in patients with aGVHD, and dynamic changes in MAP might be an accurate predictor of long-term outcomes of aGVHD, yet the MAP is not aGVHD specific and does require the logistics of waiting for results.⁷⁰ Furthermore, both systems may have to be validated in the context of new GVHD prevention strategies to understand if they retain their prognostic value.

Approximately 50% of cases presenting as grade 3 to 4 aGVHD are refractory to glucocorticoids, and outcomes for these patients are poor. In the randomized REACH2 trial, ruxolitinib led to a higher day 28 ORR, compared with best available therapy (62% vs 39%; P < .001), making it the only US Food and Drug Administration–approved therapy for SR-aGVHD.¹⁷ Despite availability and routine use of ruxolitinib, recent real-world analyses show that patients with severe SR-aGVHD continue to do poorly.^18,71,72 In patients with lower GI SR-aGVHD, the day 28 response to ruxolitinib was only 44.5%, and durable CR at day 56 was only 26.6%.^17,18 In 48 patients with ruxolitinib-resistant SR-aGVHD, the ORR to next treatment was 36%, with a median survival of only 28 days.⁷² 

The key components of aGVHD pathophysiological characteristics are tissue damage, impaired tissue restoration, and alloreactivity. Tissue damage results from conditioning regimens, infections, or other injury, leading to the release of damage-associated molecular patterns, which, along with pathogen-associated molecular patterns, activate the adaptive immune system, causing even further damage and inflammation in the absence of normal tissue repair mechanisms.^73-76 Recent work has also suggested that dysregulation of the commensal microbiome along with depletion of innate lymphoid cells predisposes to GVHD.^74,75 In addition, activation of antigen-presenting cells primes donor T cells to recognize alloantigens, through the combination of cytokines and costimulatory networks, culminating in aGVHD in target tissue.⁷³ Tissue homeostasis and restoration, modulated by the intestinal microbiome and the innate immune system, are therefore integral to GVHD pathogenesis and treatment. These advancements in understanding the pathogenesis of GVHD are now leading to trials focusing on microbiome-centered strategies to overcome dysbiosis, specific kinase inhibition, protease inhibition, cytokine and costimulation targeting, and tissue regeneration.⁷⁶ 

Risk stratification may be able to identify a subset of patients with aGVHD who can achieve better outcomes with other agents in place of corticosteroids. Low-risk patients, based on clinical (Minnesota standard risk) and biomarker (Ann Arbor 1) criteria, treated with itacitinib monotherapy had similar clinical outcomes and less infections compared with matched controls treated with glucocorticoids.⁷⁷ In BMT CTN 1501, patients with standard-risk aGVHD had comparable outcomes with sirolimus alone vs systemic glucocorticoids.⁷⁸ The MAGIC consortium is currently conducting a phase 2 pediatric low-risk aGVHD trial where a rapid corticosteroid taper is guided by both clinical response and serial MAP biomarkers (NCT05090384).

Increasingly, risk-adapted trials for high-risk aGVHD are relying on a strategy of combining novel agents with corticosteroids. Given the past failures of adding additional immunosuppressive agents, these combinations will increasingly focus on anti-inflammation, intestinal stem cell niche preservation, microbiome diversity, and tissue resilience (Table 3).

Itolizumab targets the CD6-activated leukocyte cell adhesion molecule pathway, resulting in hyporesponsive T cells. Itolizumab, combined with corticosteroids, was associated with a high day 29 response (65%) in patients with grade 3 to 4 aGVHD⁷⁹ and is being investigated in a phase 3 randomized trial in patients with grade 3 to 4 aGVHD or grade 2 with lower GI involvement (NCT05263999).

Interleukin-22 (IL-22), a tissue-protective IL-10–family cytokine, improves mucosal healing and intestinal barrier function through nonimmunosuppressive mechanisms. In a study of glucocorticoids plus the recombinant human IL-22 dimer, F-652, in patients with lower GI aGVHD, the day 28 response was 70%. Furthermore, responders demonstrated a distinct fecal microbiota composition featuring expansion of healthy commensal GI flora, indicating improvement of GVHD-associated dysbiosis.⁸⁰ 

Natalizumab, a recombinant humanized IgG4 monoclonal antibody against the α4 chain of α4β7 integrin, with glucocorticoids appeared safe and effective in patients with lower GI aGVHD (day 28 response, 57%).⁸¹ However, the result of a larger phase 2 trial of natalizumab with glucocorticoids for biomarker-defined high-risk aGVHD was negative.⁸² 

AAT has shown durable responses in SR-aGVHD⁵⁸ and is being investigated in combination with glucocorticoids vs glucocorticoids alone in the BMT CTN 1705 phase 3 trial (NCT04167514).

Receptor-interactive protein kinase 1 (RIP1) mediates apoptosis, and its inhibition prevents intestinal stem cell apoptosis in preclinical models of GI aGVHD.⁸³ The combination of a RIP1 inhibitor, GDC-8264, and glucocorticoids is being tested in risk-stratified high-risk aGVHD by the MAGIC consortium (NCT05673876).

Although ruxolitinib was associated with a high day 56 response in patients with SR-aGVHD, there was a subsequent loss of response, highlighting the need for further therapeutics in this high-risk population.¹⁷ The availability, efficacy, and safety of ruxolitinib have led to its early incorporation in the treatment algorithm for aGVHD; however, early use of ruxolitinib (thus regarded as initial treatment failure) results in difficulty observing any benefit from an investigational approach. In SR-aGVHD, there remains debate if future studies need to be compared with ruxolitinib or be used in combination.

Neihulizumab, a monoclonal antibody against P-selecting glycoprotein ligand-1, targets T-cell migration and is being investigated in patients with SR-aGVHD. Preliminary results demonstrate a day 28 response of 69%, supporting further investigation.⁸⁴ 

Glucagon-like peptide 2 (GLP-2), an enteroendocrine hormone produced by intestinal L cells, plays a protective and tissue-regenerative role in aGVHD.⁸⁵ GLP-2 supplementation enhanced the regeneration of Paneth cells and intestinal stem cells in mice with aGVHD. Low numbers of L cells in intestinal biopsies and high GLP-2 blood levels, as an indicator for L-cell damage, were associated with a higher NRM in patients undergoing allo-HCT.⁸⁵ Preclinical studies support the regenerative and protective approach of GLP-2 analogues.^85,86 A proof-of-concept trial with apraglutide (GLP-2 analogue) added to ruxolitinib is underway in patients with lower GI SR-aGVHD (NCT05415410).

Bromodomain and extraterminal domain proteins regulate inflammatory gene transcription, and their inhibition might represent a potential target to mitigate inflammation essential to aGVHD pathogenesis. In murine models of aGVHD, bromodomain and extraterminal domain inhibition improved outcomes without influencing graft-versus-tumor responses,⁸⁷ leading to a phase 1/2 trial of PLX51107 in SR-aGVHD (NCT04910152).

Urinary-derived human chorionic gonadotropin (hCG)/epidermal growth factor (EGF) (commercially available for infertility) has been studied as an adjunct therapy for SR-aGVHD. The theoretical mechanism is via hCG’s regulatory T-cell expansion in support of maternal/fetal tolerance. Furthermore, it contains supplemental EGF, which may promote tissue repair in severe GVHD. In a phase 1 study, hCG/EGF led to a 54% CR rate at day 28 when added to second-line immunosuppression.⁸⁸ 

Fecal microbiome transplant (FMT) allows for fecal transplant from a healthy donor to be given to patients with aGVHD to restore healthy microbiota diversity. Several reports have shown that FMT may be a promising strategy for SR-aGVHD,^89-91 and prospective studies with FMT for SR-aGVHD continue (NCT03359980 and NCT03812705). In another trial, FMT added to ruxolitinib resulted in 71.4% ORR at day 28, including 10 CRs for patients with lower GI SR-aGVHD.⁹² Within the limitations of a small sample size, this combination appears to be promising.

We have made significant progress in recent years leading to the approval of new agents and changes in standards in the prevention and treatment of aGVHD. For GVHD prevention, PTCy makes a strong case as the standard in RIC allo-HCT from matched donors. It is also gaining popularity in the well-matched MA setting, with several trials ongoing. Optimization of cyclophosphamide dose and investigation of the potential of combining novel agents, such as abatacept, vedolizumab, ATG, and JAK inhibitors, with both PTCy and CNI/antimetabolite backbones are the next trials being conducted. How precision-engineered grafts are incorporated is unknown, yet issues with cell-processing logistics and cost will need to be considered. As these new concepts are tested in large, randomized trials, it is vital to realize the limitations of current composite end points as the primary end point.

Treatment of high-risk aGVHD still relies on high-dose systemic glucocorticoids. The lessons learned from negative study results are as follows: (1) not all grade ≥2 aGVHD behaves the same biologically, (2) clinical trials will always enroll more low-risk patients because of eligibility criteria and clinical stability, (3) placebo arms will always outperform historical controls, and (4) risk-stratified approaches may identify high-risk patients who benefit from corticosteroids plus novel agents while omitting corticosteroids and deescalating immunosuppression in low-risk patients. In general, the focus of therapy is shifting from immunosuppression to tissue preservation, organ resilience, and microbiome diversity. The remarkable progress in understanding aGVHD biology is now being matched with therapeutic advancements (Figure 1), which we ultimately hope will translate into improved outcomes for our patients.

This study was supported by the Deutsche Forschungsgemeinschaft, Germany, for the SFB 1479 OncoEscape P01, project ID 441891347 to R.Z., SFB1160 to R.Z., ZE 872/4-2, and TRR167 to R.Z.; the EU: ERC Advanced grant AlloCure-101094168, Deutsche Krebshilfe (grant 70113473) to R.Z.; and the Jose-Carreras Leukemia foundation (grant DJCLS 01R/2019) to R.Z.

Contribution: O.J., R.Z., and Y.-B.C. contributed toward drafting, revising, and approving the manuscript.

Conflict-of-interest disclosure: O.J. reports advisory board participation for Ascentage; R.Z. received honoraria from Incyte, Novartis, Sanofi, and Mallinckrodt; and Y.-B.C. has received consulting fees from Incyte, Takeda, Magenta, Equilium, Actinium, Vor Biopharma, Pharmacosmos, and Celularity.

Correspondence: Yi-Bin Chen, Massachusetts General Hospital, 55 Fruit St, Yawkey 9E, Boston, MA 02114; e-mail: ychen6@partners.org.