A Phase I Trial of SYK Inhibition with Fostamatinib in the Prevention and Treatment of Chronic Graft-Versus-Host Disease

Abstract

Background:

Despite the exciting advancement of novel therapies, chronic graft-versus-host disease (cGVHD) remains the most common cause of non-relapse mortality after allogeneic hematopoietic stem cell transplantation (HCT). Frontline treatment of cGVHD involves systemic steroids, which are associated with significant morbidities. We previously found that inhibition of spleen tyrosine kinase (SYK) with fostamatinib preferentially eradicated aberrantly activated B cells in both ex vivo studies of cGVHD patient B cells as well as in vivo mouse studies. These and other pre-clinical studies implicated hyperreactive B cell receptor signaling and increased SYK expression in the pathogenesis of cGVHD and compelled this first-in-human allogeneic HCT clinical trial. We investigated the safety and efficacy of the oral SYK inhibitor, fostamatinib, for both the prevention and treatment of cGVHD.

Objectives:

The primary objective was to evaluate the safety of fostamatinib and determine its maximum tolerated dose in the post-HCT setting. Secondary objectives included assessing the efficacy of fostamatinib in preventing and treating cGVHD as well as examining alterations in B cell compartments with treatment.

Study Design:

This was a single-institution phase I clinical trial that evaluated the use of fostamatinib in allogeneic HCT patients before the development of cGVHD or at the time of steroid-refractory cGVHD (SR-cGVHD). Patients received fostamatinib at one of three dose levels using a continual reassessment algorithm to determine the maximum tolerated dose. Multiparameter flow cytometry was utilized to evaluate changes in B cell subpopulations over the first year of treatment with fostamatinib.

Results:

Nineteen patients were enrolled in this phase I trial, with 5 in the prophylaxis arm and 14 in the therapeutic arm. One patient (5%) required discontinuation of therapy for a dose-limiting toxicity. At a median follow-up of over three years, no patients had cancer relapse while on fostamatinib treatment, and recurrent malignancy was observed in one patient two years after the end of therapy. In the prophylaxis arm, one of five patients (20%) developed cGVHD while on fostamatinib. In the therapeutic arm, the overall response rate was 77% with a complete response rate of 31%. The median duration of response was 19.3 months and the 12-month failure free survival was 69% (95% CI, 48–100). Patients were able to reduce their steroid dose by a median of 80%, with 73% remaining on a lower dose at 1 year compared to baseline. There was an early reduction in the proportion of IgD⁻CD38^hi plasmablast-like B cells with fostamatinib treatment, particularly in those SR-cGVHD patients who had an eventual response. B cell reconstitution was not significantly impacted by fostamatinib therapy after allogeneic HCT.

Conclusions:

Fostamatinib featured a favorable safety profile in the post-HCT setting. Our data suggests an early efficacy signal that was associated with effects on expected cell targets in both the prophylaxis and treatment of cGVHD, providing rationale for a phase II investigation.

Graphical Abstract

Introduction:

Chronic graft-versus-host disease (cGVHD) is the most common cause of late morbidity and non-relapse mortality among patients who have undergone allogeneic hematopoietic stem cell transplantation (HCT).¹ The presentation of cGVHD is highly heterogeneous and may involve multiple organ systems, often leading to detrimental impacts on quality of life and functional status.^2,3 Standard first-line treatment involves prolonged systemic steroids, which can be complicated by numerous steroid-related toxicities including hyperglycemia, myopathy, and osteoporosis.^4,5 In addition, approximately 50% of patients will not respond to steroids and require additional treatment within two years of diagnosis.⁶

The strategy for treating cGVHD has shifted from long-term immunosuppression with high dose steroids to the targeting of specific mechanistic pathways involved in cGVHD pathogenesis. While alloreactive T cells have traditionally been implicated as the primary driver of cGVHD development, persistent B cell activation and subversion of B cell tolerance have now been identified as important contributors to the disease as well.^7,8 Global B cell targeting utilizing the anti-CD20 monoclonal antibody, rituximab, has demonstrated modest success in treating steroid-refractory cGVHD (SR-cGVHD), though its efficacy may be limited by the persistence of pathogenic long-lived plasma cells and maintenance of an altered B cell homeostatic state.^9–12 Targeting of aberrant signaling downstream of the B cell receptor (BCR) has been achieved through the inhibition of Bruton’s tyrosine kinase (BTK) with ibrutinib, though bleeding and infectious toxicities remain a concern.¹³ In addition, real world experience with this agent that targets IL-2 inducible T cell kinase (ITK) as well as BTK has shown limited steroid sparing effect.¹⁴

IgG-producing B cells play a substantiated role in cGVHD development and perpetuation.^15,16 In particular, plasmablast (PB)-like cells are capable of constitutive IgG production and are significantly increased in humans with active cGVHD compared to controls.¹⁷ Recent preclinical investigations have also revealed that human cGVHD B cells are primed by both B cell-activating factor (BAFF)-associated and hyper-responsive BCR-associated signaling pathways.^12,17,18 Enhanced BCR responsiveness is mediated, in part, through increased expression of the proximal BCR signaling protein, spleen tyrosine kinase (SYK).^18,19 Blockade of SYK kinase activity can abate the aberrant B cell signaling response and induce apoptosis of cGVHD B cells.^18,20 Furthermore, the SYK signaling cascade is known to play a role in long-term serological memory, and its inhibition may circumvent the limitations observed with global B cell depletion.²¹ Finally, SYK is also involved in Fc receptor signaling in monocytes and macrophages, providing a means to target other potentially pathogenic immune subsets beyond B cells.²² In murine models, SYK inhibition has been shown to ameliorate the severity of bronchiolitis obliterans and prevent development of ocular and skin cGVHD.^23,24 Fostamatinib is an oral SYK inhibitor which has demonstrated a favorable safety profile in the treatment of other immune-mediated disorders such as chronic immune thrombocytopenia and rheumatoid arthritis.^25,26 Importantly, treatment with SYK inhibition does not appear to impair the graft-versus-tumor effect and, unlike B cell depletion with rituximab, does not diminish total B cell frequencies.^22,27,28 These findings provide the rationale for an early phase clinical trial of fostamatinib in cGVHD after allogeneic HCT. We hypothesize that SYK inhibition with fostamatinib will be safe and effective in preventing and treating cGVHD.

Materials and Methods:

This was a prospective, open-label, single center, investigator-initiated phase I clinical trial evaluating the use of fostamatinib for both the prophylaxis and treatment of cGVHD. The trial is registered under clinicaltrials.gov as NCT02611063. The primary objectives were to evaluate the safety of fostamatinib and to determine a maximum tolerated dose (MTD) in the post-HCT setting. Secondary objectives included assessing the efficacy of fostamatinib both as prophylaxis against cGVHD after allogeneic HCT and as treatment for SR-cGVHD, estimating the incidence of cancer relapse, and examining alterations in B cell compartments associated with treatment.

Participants

Eligible patients were adults over the age of 18 with an adequate performance status (Karnofsky performance status ≥ 60%) who had undergone allogeneic HCT for the treatment of a hematologic malignancy. Patients without any prior signs or symptoms of cGVHD were enrolled onto the prophylaxis arm. For the therapeutic arm, patients must have a diagnosis of moderate or severe stage cGVHD per the 2014 National Institute of Health (NIH) consensus criteria and steroid refractoriness as defined by stable cGVHD despite the use of steroids equivalent to prednisone ≥ 0.5 mg/kg/day for at least 2 weeks.²⁹ Important exclusion criteria included uncontrolled infections, ongoing gastrointestinal or hepatic acute GVHD, grade 2 or higher cutaneous acute GVHD, B-cell depleting biologic agents within the past 18 months, and the use of Janus Kinase (JAK), Phosphoinositide 3-kinases (PI3K), BTK, or SYK inhibitors within the past 2 weeks. The study protocol was reviewed and approved by the Duke University institutional review board, and all enrolled subjects provided written informed consent.

Treatment and Response Evaluation

Upon enrollment onto the trial, patients received oral fostamatinib at one of three dose levels. As all patients were required to be on a prophylactic extended-spectrum triazole, which are CYP3A4 inhibitors, the doses of fostamatinib had been decreased by half to 100 mg daily, 150 mg daily, and 100 mg twice a day in order to maintain the intended effective doses. Fostamatinib is currently approved by the U.S. Food & Drug Administration for the treatment of relapsed immune thrombocytopenia at a starting dose of 100 mg twice a day.³⁰

For patients in the prophylaxis arm (without cGVHD), treatment with fostamatinib began between 80 and 150 days post-HCT and continued until 1-year post-HCT (i.e., treatment day 215–285). A cGVHD incidence of 25% or lower was the pre-determined threshold considered promising for prophylaxis. For patients in the therapeutic arm (with pre-existing SR-cGVHD), the protocol-defined treatment period continued for one year. Because three patients had developed or progressed in their cGVHD shortly after the end of treatment (EOT), the protocol was amended so that after one year, study participants in the therapeutic arm had the option to continue fostamatinib off-label so long as they continued to derive benefit from the therapy.

Patients were monitored regularly for treatment-emergent adverse events (TEAEs) throughout the course of therapy and at 6- and 12-months after the EOT. TEAEs were graded according to the CTCAE version 4.0.³¹ Dose-limiting toxicities (DLT) are defined in the supplemental appendix. In terms of efficacy, patients in the prophylaxis arm were evaluated during the treatment period for the development of new moderate or severe grade cGVHD, as defined by the NIH consensus criteria for global severity of cGVHD.²⁹ In the therapeutic arm, patients were assessed for response per the NIH consensus response criteria.³² Accrual ended once the study enrolled at least 18 evaluable patients, with evaluability defined as reaching the day 60 safety assessment. Unevaluable patients were included in the reporting of TEAEs but excluded from response assessment. All patients who experienced a DLT while taking the study drug were included in the determination of the MTD.

Statistical Methods

A Bayesian model averaging continual reassessment method with a target DLT probability of 0.20 was employed to determine the MTD in this trial (supplemental appendix).³³ Each cohort size was 2 patients, who may be enrolled in either the prophylaxis or therapeutic arms. Three dose levels were assessed: 100 mg daily, 150 mg daily, and 100 mg twice a day. The dose level with the toxicity rate closest to the target DLT probability was determined to be the MTD. Kaplan-Meier survival analysis was utilized to determine the estimated overall survival and failure-free survival, with failure events defined as progression of cGVHD, relapse of malignancy, or death. Paired t-tests were used to evaluate for significant changes in immune subsets over time. Statistical analyses were performed using R 4.1.2 (R Core Team, Vienna, Austria) and Graphpad Prism (Graphpad Software, San Diego, USA).

Peripheral Blood Flow Cytometry

Whole blood was drawn in K2 EDTA tubes at each study visit, and flow cytometric analysis was performed on these samples on the same day as collection. Upon Fc receptor blocking (TruStain FcX, BioLegend), surface staining was performed for 30 minutes on ice in PBS containing 2% FBS using the following antibodies: CD19 Pacific blue (clone J3-119, Beckman Coulter); IgD FITC (clone IA6-2, BD Biosciences); CD38 PE (clone T16, Beckman Coulter); CD27 PE-Cy7 (Clone O323, eBioscience); CD24 BV510 (clone ML5, BioLegend); CD21 APC (clone HB5, eBioscience); CD14 APC-Cy7 (clone HCD14, BioLegend). Red blood cells were lysed with BD PharmLyse^™ solution for 15 minutes at room temperature. Cells were washed twice, viability stained using 7-AAD (eBioscience), and then acquired on a 3-laser BD FACSCanto flow cytometer (BD Immunocytometry Systems). We used FlowJo v10 software package including the downsample, UMAP, and Phenograph plug-ins for analyses. For all analyses, FSC-A by SSC-A was first used to set a PBMCs gate, followed by gating on single cells and live (7-AAD⁻) cells. The flow cytometry gating strategy is provided in supplemental Figure S1.

For UMAP and phenograph analysis, the total live CD19⁺ B cells from each sample were randomly down-sampled to the same number of events prior to concatenation. UMAP v3.1 was run with the parameters of CD27 PE-Cy7, IgD FITC, CD21 APC, CD24 BV510, and CD38 PE, using default settings as described.^34,35 Phenograph v2.4 was launched holding constant the number of nearest neighbors (K=200) to identify phenotypically distinct B cell populations using an unsupervised approach. Phenograph-determined clusters were then overlaid onto UMAPs for visualization.

Results:

Patient Characteristics

Nineteen patients were enrolled onto the study. One patient failed to meet evaluability criteria due to early discontinuation of therapy, requiring the accrual of an additional patient over the intended enrollment of 18. Five patients without a prior diagnosis of cGVHD were enrolled onto the prophylaxis arm, while 14 patients with SR-cGVHD were enrolled onto the therapeutic arm (Figure 1A).

Figure 1.

Patients enrolled and treated with fostamatinib after allogeneic hematopoietic stem cell transplantation. A. Consolidated Standards of Reporting Trials (CONSORT) diagram depicting enrollment onto trial. The dose level for each patient was determined by the continual reassessment method. B. Distribution of cGVHD severity scores by organ system in the therapeutic arm at the time of enrollment. Percentages indicate the overall proportion of patients with involvement of the respective organ system by cGVHD. cGVHD: chronic graft-versus-host disease, GI: gastrointestinal.

Baseline characteristics of the enrolled patients are described in Table 1. In the prophylaxis arm, all patients had a history of acute GVHD and received stem cells from mobilized peripheral blood, while 40% received myeloablative conditioning. In the therapeutic arm, 36% had severe cGVHD, 50% had involvement of at least 4 organs, and the median number of prior lines of therapy was 3 (range, 1–6). The most commonly involved organs were the skin (71%) and eyes (71%) (Figure 1B). Among the 10 patients with skin involvement, 3 had deep sclerotic features while 2 had superficial non-hidebound sclerotic features. Two patients previously received ruxolitinib and three received ibrutinib.

Table 1. Patient Characteristics.

Baseline characteristics of patients in the prophylaxis and therapeutic arms.

	Prophylaxis Arm (N = 5)	Therapeutic Arm (N = 14)
Median Age, years (range)	66 (39 – 75)	53.5 (29 – 66)
Male Gender, N (%)	4 (80%)	9 (64%)
Median Weight, kg (range)	71.9 (52.1 – 95.5)	73.5 (61.2 – 118.8)
Median KPS (range)	90% (80 – 100)	80% (60 – 100)
Caucasian/White Race, N (%)	5 (100%)	14 (100%)
Hispanic Ethnicity, N (%)	1 (20%)	1 (7%)
Transplant Indication, N (%)
AML	3 (60%)	4 (29%)
MDS*	1 (20%)	3 (21%)
Multiple Myeloma	1 (20%)	2 (14%)
Other‡	0	5 (36%)
Prior Acute GVHD, N (%)	5 (100%)	8 (57%)
Myeloablative Conditioning, N (%)	2 (40%)	11 (79%)
Peripheral Blood Stem Cell Source, N (%)	5 (100%)	13 (93%)▵
Fully HLA Matched Donor, N (%)	3 (60%)	12 (86%)▵
Related Donor, N (%)	2 (40%)	7 (50%)
Female Donor to Male Recipient, N (%)	2 (40%)	4 (29%)
Global Severity of cGVHD
Moderate	N/A	9 (64%)
Severe		5 (36%)
# Organs Involved at Enrollment, N (range)	N/A	3.5 (1 – 6)
≥ 4 Organs Involved, N (%)	N/A	7 (50%)
Median Prednisone Dose at Enrollment, mg/kg (range)	N/A	0.25 (0 – 0.68)
Median Prior Lines of cGVHD Therapy, N (range)		3 (1 – 6)
Ruxolitinib, N (%)	N/A	2 (14%)
Ibrutinib, N (%)		3 (21%)
Rituximab, N (%)		1 (7%)
Concurrent GVHD Medications at Baseline
Calcineurin Inhibitor, N (%)	3 (60%)	4 (29%)
Mycophenolate Mofetil, N (%)	1 (20%)	2 (14%)
Sirolimus, N (%)	0	2 (14%)
Median Time from Transplant to cGVHD Diagnosis, months	N/A	10.4 (4.4 – 19.6)

^*One patient in steroid-refractory group had concurrent diagnoses of both myelodysplastic syndrome and chronic lymphocytic leukemia.

^‡One patient each with angioimmunoblastic T cell lymphoma, hodgkin lymphoma, acute lymphocytic leukemia, chronic lymphocytic leukemia, and chronic myelominocytic leukemia.

^▵Data was unknown for one patient.

KPS: Karnofsky performance status, cGVHD: chronic graft-versus-host disease, AML: acute myeloid leukemia, MDS: myelodysplastic syndrome, HLA: human leukocyte antigen.

The median time from transplant to the start of fostamatinib therapy was 3.9 months (range, 3.2–4.5) in the prophylaxis arm and 37.2 months (range, 11.8–114.9) in the therapeutic arm. In the therapeutic arm, the median time from the diagnosis of cGVHD to start of therapy was 25.0 months (range, 0.7–45.3). The median durations of follow-up for the prophylaxis and therapeutic arms were 43.6 months (range, 31.3–57.6) and 34.9 months (range, 12.8–51.9), respectively.

Safety and Treatment-Emergent Adverse Events (TEAEs)

All 19 enrolled patients received fostamatinib. In the prophylaxis group, one patient discontinued treatment per protocol due to a small decrease in CD3 chimerism, which was present prior to fostamatinib initiation and required donor lymphocyte infusion (DLI). In the therapeutic group, 4 of 14 patients discontinued treatment during the protocol-defined treatment period due to the following occurrences: recurrent non-cardiac chest pain (n = 1, on day 35) and progression of cGVHD (n = 3). Among grade 2 or higher TEAEs, regardless of attributed cause, the most common were hypertension, hyperglycemia, and upper respiratory infections (Table 2). Specifically, grade 2+ infectious complications were reported in 9 (47%) patients, most of which were self-limited viral infections. The three grade 3 infectious events all occurred in the therapeutic arm and were due to pulmonary aspergillosis, bacterial superinfection of influenza, and an infected intraventricular catheter. Twelve patients had abnormal liver function testing, though only two cases were grade 2. At last follow-up, there was one death in the study, which was related to relapse of multiple myeloma approximately two years after EOT. No other deaths or incidences of cancer relapse were reported. Two DLT events were observed for abnormal liver function testing and recurrent non-cardiac chest pain. Only one patient (5%) required discontinuation of therapy due to non-cardiac chest pain. Per the continual reassessment algorithm, the final MTD of fostamatinib was determined to be 100 mg twice a day (Figure S2).

Table 2. Treatment Emergent Adverse Events.

Grade 2 or higher treatment-emergent adverse events which were not present at the start of therapy, found in ≥ 10% of patients, regardless of investigator-attributed cause.

	Grade 2	Grade 3	Grade 4	Total
Hypertension	9 (47%)	2 (11%)	0	11 (58%)
Hyperglycemia	1 (5%)	3 (16%)	0	4 (21%)
Upper Respiratory Infections	4 (21%)	0	0	4 (21%)
Dry Eyes	3 (16%)	0	0	3 (16%)
Fatigue	3 (16%)	0	0	3 (16%)
Pneumonia	2 (11%)	1 (5%)	0	3 (16%)
Cellulitis	2 (11%)	0	0	2 (11%)
Elevated AST or ALT	2 (11%)	0	0	2 (11%)
Lymphopenia	1 (5%)	1 (5%)	0	2 (11%)
Nausea	2 (11%)	0	0	2 (11%)
Other Infections	1 (5%)	1 (5%)	0	2 (11%)
Vomiting	2 (11%)	0	0	2 (11%)
VTE	1 (5%)	1 (5%)	0	2 (11%)

AST: aspartate transaminase, ALT: alanine transaminase, VTE: venous thromboembolism.

Response in Prophylaxis Arm

Five patients without cGVHD at the time of enrollment received fostamatinib prophylaxis for a median duration of 8.1 months (range, 5.2–8.9). The 1-year post-HCT incidence of cGVHD was 20%, with one patient developing elevated liver enzymes attributed to cGVHD. Although no other patients developed cGVHD while on prophylactic fostamatinib, two did eventually develop moderate cGVHD approximately one month after the completion of prophylactic therapy.

Response in Therapeutic Arm

Fourteen patients with moderate or severe SR-cGVHD were treated with fostamatinib for a median duration of 12 months (range, 1–12) in the protocol-defined treatment period. One patient did not reach the pre-defined day 60 threshold for evaluability due to a DLT necessitating early discontinuation of therapy. Ten patients (71%) elected to restart or continue fostamatinib off-label after the planned 12-month treatment period for a median total duration of 23 months (range, 1–41). Five patients still remained on this therapy at last follow-up.

Among the 13 evaluable patients, the overall response rate was 77% (95% CI, 46–95), with six patients (46%) obtaining a partial response (PR) and four patients (31%) obtaining a complete response (CR). Notably, all three of the non-responders were on the lower fostamatinib dose of 150 mg daily, while all patients who received the highest dose level of 100 mg twice a day experienced at least a partial response. The median time to response was 2 months (range, 0.5–4.8) and 60% of responses occurred within the first 3 months of therapy. The median duration of response was 19.3 months (range, 1.2–34.1), with 70% of responses lasting longer than one year (Figure 2A). Responses were comparable among different organ systems (Figure 2B). The 12-month and 24-month failure-free survival were 69% (95% CI, 48–100) and 45% (95% CI, 24–83), respectively.

Figure 2.

Clinical outcomes of patients treated with fostamatinib in both the prophylaxis and therapeutic arms. A. Swimmer’s plot of treatment responses and progression events for patients in the prophylaxis (P) and therapeutic arms (SR). For the therapeutic arm, the protocol-defined one year of treatment is indicated in blue. Due to progression events observed soon after discontinuation of fostamatinib, the study protocol was amended to allow continued off-label treatment with fostamatinib after the end of the 1 year (indicated in green). B. Bar graph depicting response rates by organ system. Responses were generally consistent across organ types. There was no baseline liver involvement by chronic GVHD in this cohort. C. Mean prednisone dose in mg/kg during treatment with fostamatinib, demonstrating an overall downward trend in steroid requirement. Error bars indicate 95% confidence intervals. Of note, Patient SR02 was unevaluable for response and thus excluded from figures 3B and 3C. CR: complete response, PR: partial response, GI: gastrointestinal, cGVHD: chronic graft-versus-host disease.

Among the 11 patients who were taking systemic steroids at the start of treatment, there was a consistent reduction in prednisone usage while on fostamatinib (Figure 2C). When comparing steroid dose nadirs in the first year to baseline, the maximum dose reduction ranged from 50–100% with a median of 80%. Eight of the 11 patients (73%) were still on a lower dose of prednisone at the end of one year. Two of the three non-responders maintained stability of their cGVHD for at least one year while on fostamatinib despite the reduction of steroids. At last follow-up, two responders were successfully weaned from all systemic cGVHD therapy while another maintained a complete response on fostamatinib alone. Three patients who had eventual progression of their cGVHD were treated off-label with a combination of ruxolitinib and fostamatinib. With a median follow-up of 7 months on combination therapy with ruxolitinib (range, 2–24), all 3 patients experienced improvements in their disease and tolerated the combination regimen without any significant unexpected toxicities.

Analysis of Circulating Immune Subsets

We prospectively performed flow cytometric analyses on peripheral blood samples from study participants. In the prophylaxis arm, compared to pre-treatment baseline, there was a significant proportional decrease in IgD⁻CD38^hi (“plasmablast-like”, or PB-like) cells (shown as fold change in Figure 3A) starting on day 3. In the solitary patient who experienced progression while on prophylactic fostamatinib, the PB-like cell frequency increased prior to the time of cGVHD development. Similarly, in the therapeutic arm, PB-like cell frequencies decreased on both day 14 and day 35 of treatment compared to baseline, though the difference was only statistically significant on day 35 (Figure 3B). Interestingly, the only two patients whose PB-like cell frequencies consistently failed to decrease on both day 14 and day 35 were both among the three non-responders in the cohort. Overall, there appeared to be an early reduction in PB-like cells among responders but not among non-responders (Figure 3C).

Figure 3.

Changes in SYK-expressing immune cells with fostamatinib therapy. A. Paired t-test analysis comparing the fold change of plasmablast (PB)-like cells relative to baseline in the prophylaxis arm patients, demonstrating a decrease in PB-like cell frequencies with fostamatinib treatment. The median day post-transplant at which these patients began fostamatinib was post-transplant day 117 (range, 97–134 days). B. Paired t-test analysis comparing the fold change of PB-like cells relative to baseline in the therapeutic arm patients. Patient SR16 was excluded from these analyses due to severe infection and poor drug absorption during this period, while patient SR10 was excluded due to lack of a baseline sample. C. Violin plot of fold change in PB-like cell frequencies among patients in the therapeutic arm, comparing responders to fostamatinib with non-responders. D, E. Paired t-tests demonstrating an early reduction in monocyte counts with fostamatinib treatment in the prophylaxis arm and therapeutic arm, respectively. F. Among patients in the prophylaxis arm, the monocyte counts had an initial decrease but recovered by 6 and 12 months after the end of therapy. Freq: frequency, EOT: end of therapy. P1-P5 refer to the patients enrolled in the prophylaxis arm, while SR1-SR14 refer to patients in the therapeutic arm. * indicates a statistically significant change from baseline, while ns refers to a non-significant change.

As shown in Figures 3D–3E, total monocytes were also significantly decreased in the early treatment period. Since SYK is known to be activated in monocytes through Fc signaling, the lower circulating monocyte counts that associated with treatment initiation also suggest fostamatinib effect. At 6 and 12 months after the end of treatment, monocyte counts were observed to recover back to baseline. Importantly, the frequencies of IgD⁺CD38^lo mature naïve B cell, total B cell, and non-B cell lymphocyte (natural killer and T cell) subsets were not significantly impacted by fostamatinib treatment (Figure 4).

Figure 4.

Non-pathogenic lymphocyte subsets are not significantly impacted by fostamatinib treatment. A, B. Paired t-test of changes in IgD⁺CD38^lo mature naïve B cell subsets in the early treatment period demonstrated no significant change in this subset after starting fostamatinib. C, D. Longitudinal plots of total B cells and total non-B cells (T and NK cells) in the prophylaxis arm. Total B cell counts in the prophylaxis arm were not reduced with fostamatinib therapy, suggesting preserved B cell reconstitution. The median days of study collection for each time point were as follows: 3 (range, 3–3), 8 (range, 8–15), 24 (range, 22–29), 38 (range, 36–43), 59 (range, 57–64), 92 (range, 85–94), 113 (range, 101–121), 141 (range, 141–142), 183 (range, 183–183), 225 (range, 225–232), and 253 for EOT (range, 155–281). Post-HCT: median day after allogeneic stem cell transplantation, Freq: frequency, GC: germinal center, Tx: treatment.

Discussion:

In this phase I trial, SYK inhibition with fostamatinib demonstrated a favorable safety profile and early efficacy signal in the prophylaxis and treatment of cGVHD after allogeneic HCT. The observed TEAEs were consistent with prior investigations of fostamatinib reported in chronic immune thrombocytopenia and rheumatoid arthritis.^25,26 Hypertension was the most common grade 2 or higher TEAE, likely related to fostamatinib’s nonspecific targeting of VEGFR2, but was manageable with antihypertensive agents.³⁶ As elevated hepatic enzymes were common with fostamatinib therapy and after HCT, differentiating drug-related liver toxicity from hepatic cGVHD can be challenging and will require a thoughtful diagnostic approach.

Effective B-cell targeting approaches for the prevention of cGVHD have been elusive. Peri-transplant B cell depletion with rituximab has demonstrated modest efficacy, but its global anti-CD20 effect raises concerns over delayed B cell reconstitution and infectious complications.^1,37 Our study offers a proof of concept for more specific targeting of aberrant B cell signaling in cGVHD via SYK inhibition. In the prophylaxis arm, one of five patients developed cGVHD while receiving fostamatinib, meeting the pre-determined threshold of less than 25% incidence for consideration of additional evaluation. With extended follow-up, malignant relapse was observed in only 1 patient with multiple myeloma that recurred two years after the end of fostamatinib prophylaxis, which may indicate conservation of the graft-versus-tumor effect. Preservation of B cell reconstitution was supported by the lack of significant change in circulating total B cells and mature naïve B cells while on treatment with fostamatinib, comparable to prior reports.³⁸

In the treatment of SR-cGVHD, T cell signaling remains the primary target of current standard therapies such as the Janus kinase (JAK) inhibitor, ruxolitinib, and the rho-associated coiled-coil-containing protein kinase-2 (ROCK2) inhibitor, belumosudil.^39,40 Although a recent trial with a different SYK inhibitor, entospletinib, in combination with steroids as frontline treatment for cGVHD (NCT02701634) was terminated for lack of efficacy, there appears to be a preliminary signal of efficacy with fostamatinib in the steroid-refractory setting, possibly due to differing binding specificities.⁴¹ Fostamatinib produced an ORR of 77% and was able to achieve prolonged disease control in the majority of responders, which allowed for a strong and durable steroid-sparing effect. While patient-reported outcomes were not assessed in this early phase trial, 71% of patients elected to continue the study drug after the protocol-defined 12-month treatment period ended, offering some insight into patient perceptions of the benefits and tolerability of this drug. With its mechanism of action targeting aberrant B cell signaling, fostamatinib may provide a complementary adjunct to current T-cell focused approaches for especially refractory cases of cGVHD. Interestingly, our study reports three patients treated concurrently with both ruxolitinib and fostamatinib (prescribed off-label) for up to 24 months. This limited experience offers some early evidence of tolerability, with all patients continuing on this regimen at last follow-up.

SYK inhibition has previously been shown to preferentially eradicate BCR hyper-responsive B cells taken from patients with active cGVHD.^18,23 For this reason, we expected a selective targeting of BCR-activated subsets with fostamatinib, without disruption of resting B cells. Indeed, our B cell subset analysis revealed that the proportion of post-GC PB-like cells significantly decreased in the early treatment period in both study arms. The proportion of PB-like cells also increased with development of cGVHD in the prophylaxis arm. This is consistent with preclinical findings demonstrating that PB-like cells were associated with an active cGVHD state and elevated BAFF.^17,42 Together, our findings suggest that PB-like cell subsets serve as a pharmacodynamic endpoint as well as a potential mechanistic marker, though future studies are needed to confirm this hypothesis.

Although the sample size was small, there appeared to be a trend indicating a more prominent reduction of PB-like cells among responders compared to non-responders, potentially guiding future investigations on response biomarkers and resistance mechanisms such as reduced SYK expression or alternative activation pathways. Of note, altered B cell homeostasis through an elevated BAFF level and relative naïve B cell lymphopenia provide an alternative pathway in priming these pathogenic B cell subsets for survival.¹⁷ Reassuringly, the preservation of B cell reconstitution with fostamatinib suggests the mediation of a favorable B cell homeostatic state in contrast to the effects of global B cell-targeting strategies. Finally, while this has not been formally evaluated, fostamatinib’s reduction of PB-like cells may conceivably impact response to vaccinations. Encouragingly, a prior animal study did not demonstrate a significant impact of fostamatinib on the humoral immune response to keyhole limpit hemocyanin (KLH) challenge.⁴³ Future studies on pre- and post-treatment antibody testing in humans are needed to assess plasmablast response to vaccination.

Peripheral monocyte counts decreased early in treatment and increased after the development of cGVHD and after discontinuation of therapy. Monocytes are known to express SYK and have been hypothesized to play a role in cGVHD through their functions in antigen presentation and fibrosis.^44,45 While severe monocytopenia was not observed, the known impact of SYK blockade on monocyte chemotaxis and phagocytosis does necessitate ongoing antifungal prophylaxis during treatment. Of note, extended-spectrum azoles may act as CYP3A4 inhibitors that can increase the effective fostamatinib exposure up to two-fold.⁴⁶ Therefore, early discontinuation of azole prophylaxis may impact the effective dose of fostamatinib. The reductions of both PB-like cells and monocytes may suggest simultaneous targeting of multiple pathogenic immune subsets by fostamatinib.

Limitations of this study include the small sample size of 19 patients divided over two study arms. Any inferences on efficacy, while promising, are nevertheless preliminary and will require further validation in later phase trials. In particular, the prophylaxis arm had a limited patient cohort whose primary risk factor for cGVHD was a history of acute GVHD. Future studies will need to evaluate a broader population of high-risk patients. In the therapeutic arm, the inclusion definition of SR-cGVHD was liberalized due to provider and patient reluctance to maintain high doses of glucocorticoids prior to enrollment, which may impact generalizability. However, as 93% of patients in this arm had already trialed two or more lines of therapy, this suggests a high baseline level of steroid refractoriness or dependence.⁴⁷ Due to the exclusion of significant liver lab abnormalities, there were no patients with active chronic GVHD of the liver enrolled. Therefore, fostamatinib’s impact on hepatic cGVHD was not discernible. Of note, only four patients (29%) in the therapeutic arm had previously received a novel targeted therapy, as this trial began enrolling prior to approval of these agents. Patients from racial and ethnic minority groups were underrepresented in this trial, partly due to the limitations of a single institution design. Future trial designs and accrual will require deliberate consideration of diversity in order to ensure equitable representation. Finally, the immune subset analysis, while adequate for this study, could be improved in future trials. The antibody staining panel should be expanded to include other B cell markers of interest (e.g., CD20, TACI, CD95, CD11c, CD10, IgM) as well as markers of both T cell and monocyte subsets.

Rational targeting of biologic pathways implicated in the pathogenesis of cGVHD allows the opportunity to spare patients from broad immunosuppression and steroid-related complications. This phase I study demonstrated a manageable safety profile for fostamatinib in the post-HCT setting. Notably, SYK inhibition led to reductions in plasmablast-like B cells without impacting total B cell and lymphocyte counts, suggesting specific targeting of aberrant B cell subpopulations while preserving immune recovery. These findings provide the rationale for a planned phase II investigation of fostamatinib for the treatment of SR-cGVHD. Nevertheless, given the inherent heterogeneity and pleiomorphism of this disease, not all patients are expected to respond to a single targeted agent. Additional studies are needed to identify predictive biomarkers and investigate novel synergistic combinations that strategically target non-overlapping signaling pathways.

Highlights:

• Fostamatinib had a manageable safety profile in the post-transplant setting

• One of five patients developed cGVHD while on fostamatinib prophylaxis

• Among SR-cGVHD patients, the ORR was 77% with 70% of responses lasting > 1 year

• Fostamatinib allowed for a strong and durable steroid-sparing effect

• Plasmablast-like B cells decreased with therapy in responders

Data Sharing Statement:

Please contact the corresponding author for data requests.