Feasibility of reduced-dose posttransplant cyclophosphamide and cotransplantation of peripheral blood stem cells and umbilical cord-derived mesenchymal stem cells for SAA

Posttransplant cyclophosphamide (PTCy) as graft-versus-host disease (GVHD) prophylaxis is an effective strategie for patients receiving matched sibling donor hematopoietic stem cell transplantation (MSD-HSCT) and haploidentical HSCT (haplo-HSCT). We evaluated the effectiveness and safety of reduced-dose cyclophosphamide, 20 mg/kg for 13 patients in MSD-HSCT cohort and 25 mg/kg for 22 patients in haplo-HSCT cohort, on days + 3, + 4 combined with cotransplantation of peripheral blood stem cells (PBSCs) and human umbilical cord-derived mesenchymal stem cells (UC-MSCs) for severe aplastic anemia (SAA). In MSD-PTCy cohort, the times to neutrophil and platelet engraftment were significantly shorter than those in the MSD-control cohort (P < 0.05). The cumulative incidence of acute GVHD (aGVHD) at day + 100 (15.4%) was lower than that in the MSD-control cohort (P = 0.050). No patient developed chronic GVHD (cGVHD). The 1-year overall survival (OS) and event-free survival (EFS) rates were 100% and 92.3%. In haplo-PTCy cohort, the times to neutrophil and platelet engraftment were significantly shorter than those in the haplo-control cohort (P < 0.05). The cumulative incidences of aGVHD at day + 100 and 1-year cGVHD were 31.8% and 18.2%, and the 1-year OS and EFS rates were 81.8% and 66.9%. Reduced-dose PTCy and cotransplantation of PBSCs and UC-MSCs is an acceptable alternative to patients with SAA.


Results
Patients. 33 Tables 1 and 2. No significant difference was observed between the PTCy cohorts and the control cohorts in terms of baseline demographics except for stem cell source. Eight patients (40%) of MSD-control cohort and all patients of haplo-control cohort received PBSCs and BM, while all patients of PTCy cohorts received PBSCs and UC-MSCs, except for two patients who suffered from primary graft failure received extra BM from a secondary donor in the haplo-PTCy cohort. There were 6 males (46.2%) and 13 males (59.1%) in the MSD-PTCy and haplo-PTCy cohorts and the median ages were 20 (range 9 to 35) and 14 (range 3 to 31) years respectively. One patient of each PTCy cohort was treated with anti-human lymphocyte immunoglobulin pretransplantation. The remaining patients had previously received treatment with cyclosporine (CSA) and/or steroids. In addition, two patients (9.1%) had hospital admissions (range 1.02 to 13.37) × 10 6 /kg). However, we observed that the times to neutrophil and platelet engraftment were shorter in the MSD-PTCy cohort than those in the control cohort [11 days (range 10 to 14) vs 13 days (range 10 to 23) and 12 days (range 9 to 15) vs 14 days (range 10 to 45)]. The difference in graft rejection between the MSD-PTCy and control cohorts was not significant (7.7% vs 15.0%). The rates of chimerism at day + 30 in the MSD-PTCy and control cohorts were 100% and 90% respectively, and there was no significant difference. In the haplo-HSCT cohorts, the median numbers of infused MNCs and CD34+ cells were comparative (MNCs: 18 (1.76-22.53) × 10 6 /kg). Two patients suffered from primary graft failure experienced successful engraftment after receiving transplant from haploidentical donors in the haplo-PTCy cohort. In the haplo-control cohort, secondary graft failure developed in one patient at 1.5 months after transplantation, and this patient sustained engraftment after undergoing MMUD-HSCT. The times to neutrophil and platelet engraftment were shorter in the haplo-PTCy cohort than those in the control cohort [12 days (range 9 to 15) vs 13 days (range 11 to 19) and 11.5 days (range 9 to 17) vs 14 days (range 11 to 62)]. One patient from each cohort suffered graft rejection.
Regimen-related toxicities. The incidence of pulmonary infection in the MSD-PTCy cohort was obviously lower than that in the control cohort (P = 0.022), 7.7% vs 50.0%. Similarly, the incidences of Cytomegalovirus (CMV) and Epstein-Barr virus (EBV) reactivation in the MSD-PTCy cohort were significantly lower compared with those in the control cohort (CMV: 23.1% vs 65.0%; EBV: 0 vs 60.0%), and there were significant differences between the two cohorts (P: 0.032 and 0.001, respectively). Although the incidence of hemorrhagic cystitis was lower in the MSD-PTCy cohort (7.7% vs 35.0%), the difference was not significant (P = 0.108). One patient suffered from pure red cell aplasia (PRCA) after ABO-compatible HSCT, and hematopoietic recovery occurred eighteen months after transplantation. In the control cohort, four of twelve patients with EBV reactivation developed posttransplantation lymphoproliferative disorders (PTLD) and three patients were cured through combination therapy with immunosuppressor withdrawal, rituximab and donor lymphocyte infusion (DLI). Another patient experienced an epileptic seizure because of intracranial EBV infection. Furthermore, two patients developed unexplained seizures that were relieved after expectant treatment. Cardiotoxicity and thrombotic microangiopathy (TMA) each developed in one patient. No patient had veno-occlusive disease (VOD). Table 2. Patient characteristics and transplant-related parameters in the haplo-HSCT groups. Continuous variables were expressed as median values and ranges; Discrete variables were expressed as counts (%). P-value from Fisher exact test and independent samples T test. *P < 0.05, ****P < 0.0001.

Variable
Haplo-PTCy group (n = 22) Haplo-control group (n = 15) P values www.nature.com/scientificreports/  Table 3. Transplantation outcomes in MSD-HSCT groups. 95% CI 95% confdence interval. Continuous variables were expressed as median values and ranges; Discrete variables were expressed as counts (%). Cumulative incidences of aGVHD and cGVHD were estimated using competing risk model. P-value log-rank tests and Fisher exact test. *P < 0.05, ***P < 0.001. www.nature.com/scientificreports/ The incidence of pulmonary infection in the haplo-PTCy cohort was lower than that in the control cohort (40.1% vs 60.0%), while the incidence of CMV reactivation in the haplo-HSCT cohort was higher compared to that in the control cohort (81.8% vs 53.3%), but the differences were not statistically significant (P: 0.325, 0.080). The median times to CMV infection were + 37 days (range 29 to 46) and + 35 days (range 25 to 53) in the haplo-HSCT and the control cohorts respectively. Two of eighteen patients with CMV infection in the haplo-HSCT cohort and one of eight patients with CMV infection in the control cohort developed cytomegalovirus retinitis. The incidence of EBV reactivation in the haplo-HSCT cohort (22.7%) was obviously lower than that in the control cohort (80.0%), and the difference was significant (P = 0.001). There were no PTLD in the haplo-PTCy cohort, but three patients developed PTLD based on their seropositive EBV status in the control cohort. The incidences of hemorrhagic cystitis with BKV seropositivity were comparative between the haplo-HSCT and the control cohorts (31.8% and 26.7%). In the haplo-PTCy cohort, one of seven patients who were seropositive for BKV progressed to BKV encephalitis and suffered from intracranial fungal infection. In addition, one patient had TMA at day + 50. Moreover, posterior reversible encephalopathy syndrome (PRES) developed in four patients in the haplo-PTCy cohort and in two patients in the control cohort. None of all patients developed VOD in the two cohorts.

MSD-PTCy group (n = 13) MSD-control group (n = 20) P values
Survival. The median follow-up time of the MSD-PTCy cohort was similar to that in the control cohort, 22 months (range 10 to 36) vs 21.5 months (range 2 to 36). The median follow-up of the haplo-HSCT and the control cohorts were 11.5 months (range 3 to 27) and 12 months (range 1.5 to 36), respectively.
All thirteen (100%) patients in the MSD-PTCy cohort are now alive with normal hematopoiesis. Interestingly, one female is now three months pregnant. The 1-year overall survival (OS) of the MSD-PTCy cohort was 100% compared to 75.0% (95% CI 50.0-88.7)% in the MSD-control cohort (P = 0.057) (Fig. 3A). The 1-year event-free survival (EFS) of the MSD-PTCy cohort was 92.3% (95% CI 56.6-98.9)% compared to 65.0% (95% CI 40.3-81.5)% in the MSD-control cohort (P = 0.071) (Fig. 3B). In the control cohort, three patients died from serious pulmonary infection at 2, 3 and 5 months, one patient died of intracranial EBV infection at day + 70, and one patient died of TMA at 4 months. In the MSD-PTCy cohort, twelve of thirteen patients had a followup period of more than 12 months at the time of writing. All the patients began to reduce immunosuppressor treatment at 6 months and these agents were withdrawn before 9 months except for the patient with PRCA.
In the haplo-PTCy cohort, two patients died from intractable pulmonary infection at 3 and 9 months, one patient died of cerebral hemorrhage at 3 months, and one patient died of intestinal GVHD at 7 months. In the control cohort, two patients died from pulmonary infection at 4 and 6 months, and one patient died of severe intestinal GVHD at 12 months. The cumulative incidence of 1-year OS was 81.8% (95% CI 58.5-92.8) in the haplo-PTCy cohort compared to 77.1% (95% CI 44.2-92.1) in the haplo-control cohort (Fig. 3C). The 1-year EFS in the haplo-PTCy cohort vs the control cohort was 66.9% (95% CI 39.1-84.2) vs 40.8% (95% CI 20.5-71.2) (Fig. 3D). However, there was no significant difference in 1-year OS or EFS between two cohorts. In the haplo-PTCy cohort, eleven of eighteen living patients had a follow-up period of more than 12 months at the time of writing, those patients began to reduce immunosuppressor treatment at 9 months after HSCT.

Discussion
Effective therapies for SAA patients usually involve immunosuppressive therapy (IST), allo-HSCT and eltrombopag. The application of IST is limited by late sequelae and the high cost of ATG 16 . Although the effective rate of eltrombopag is 20% to 40%, the main problems are relapse and secondary clonal disease 17 . The key challenge of allo-HSCT lies in minimizing complications and maximizing survival. Recently, PTCy has gradually come to be the predominate treatment worldwide due to its favorable prognosis in the allo-HSCT setting 18,19 . However, a high dose of PTCy is accompanied by cardiotoxicity. Therefore, some studies have reported the outcomes of reduced doses of PTCy in malignant and nonmalignant hematological disorders 15,20 . Furthermore, UC-MSCs are popular in allo-HSCT for SAA because of several beneficial effects, including facilitating engraftment and decreasing GVHD 21,22 . At present, there are few evaluations of reduced-dose PTCy for SAA patients receiving cotransplantation of PBSCs and UC-MSCs. Therefore, the aim of this prospective study was to demonstrate the influence of reduced-dose PTCy on early outcomes after cotransplantation of UC-MSCs and PBSCs for patients with SAA receiving transplants from either matched or haploidentical donors. GVHD is a main barrier in patients with SAA undergoing allo-HSCT. PTCy, an attractive option, eliminates early proliferating alloreactive T cells to reduce impaired immune reconstitution 23,24 . However, a high dose of PTCy is accompanied by cardiotoxicity. Then, the doses of PTCy were reduced to 20 mg/kg and 25 mg/kg from the conventional 50 mg/kg dose for MSD-HSCT and haplo-HSCT, respectively. The results showed that the cumulative incidence of aGVHD at day + 100 in the MSD-PTCy cohort was lower compared to that in the control cohort(P = 0.050). In addition, the cumulative incidence of grade II-IV aGVHD was lower in the MSD-PTCy cohort than that in the control cohort, 7.7% (95% CI 0.4-30.3) vs 30.0% (95% CI 11.8-50.7), but we did not observe a significant difference. No patient in the MSD-PTCy cohort developed cGVHD. George et al. reported that 30 patients with SAA received high-dose PTCy as the sole GVHD prophylaxis in PBSC MSD-HSCT. The incidences of grade II-IV aGVHD and cGVHD were 22.0% and 22.7%, respectively 14 . In MSD-HSCT for malignant hematologic disorders, high-dose PTCy has been associated with the occurrence of grade II-IV aGVHD (17% to 45%) and cGVHD (7% to 14%) [25][26][27] . A multicenter study showed that the cumulative incidences of grade II-IV aGVHD and cGVHD at 1 year were 33.7% and 22.4%, respectively, for SAA patients receiving BM and PBSCs in the haplo-HSCT setting 28 . In this study, the cumulative incidence of grade II-IV aGVHD at day + 100 www.nature.com/scientificreports/ and 1-year cumulative incidence of cGVHD in the haplo-PTCy cohort were 27.3% (95% CI 10.8-46.9) and 18.2% (95% CI 5.5-36.8), respectively. Although there were no significant differences in either the cumulative incidence of aGVHD and grade II-IV aGVHD or the 1-year cumulative incidence of cGVHD, reduced-dose PTCy and cotransplantation of PBSCs and UC-MSCs probably overcame the high incidences of aGVHD and cGVHD due to PBSCs. Encouragingly, the 1-year OS in the MSD-PTCy cohort was 100%. We observed no difference in the 1-year OS and EFS values between the haplo-PTCy and haplo-control cohorts, but we believe that the joint use of PBSCs and UC-MSCs leads to decreased donor pain and improves donation rates compared to the use of routine PBSCs and BM. The times to neutrophil and platelet engraftment in the PTCy cohorts were shorter than those in the control cohorts (P < 0.05). The time to neutrophil and platelet engraftment in the haplo-PTCy cohort were 12 days (9-15) and 11.5 days (9-17), respectively. Moreover, the times to engraftment in the haplo-HSCT cohort were heterogeneous, with the time to neutrophil engraftment being 12 to 19 days, and the time to platelet engraftment being 13 to 21 days [28][29][30][31] . It followed that the time to engraftment was not delayed due to reduced-dose PTCy. The application of UC-MSCs was also special in this study. Various abilities of MSCs in the HSCT setting involve facilitating engraftment, remedying graft dysfunction and providing GVHD prophylaxis 32 . A retrospective study demonstrated that UC-MSCs were likely to decrease the occurrence of severe aGVHD, which influenced life quality and prognosis for patients undergoing haplo-HSCT 33 . Another study reached the same conclusion and verified that cotransplantation of UC-MSCs with haplo-HSCT was feasible and safe for patients with SAA 34 . Platelet-derived growth factor (PDGF) secreted from MSCs, a putative anti-apoptotic factor, shows potential capability to stimulate platelet recovery 35,36 . There is controversy regarding the engraftment function of MSCs, but most studies have shown that the role of MSCs in allo-HSCT is worthy of consideration 33,34,37 . However, there are some limitations of adult BM-MSCs, such as quality variation from different donors, limited proliferative potency and stem cell senescence. Same pluripotent stem cells (PSC)-derived MSCs (PSC-MSCs) with batch-tobatch consistency are another alternative for quality control. PSC-MSCs present higher proliferative potential 38 .
A phase I clinical trail demonstrated the potential capability of GMP-grade PSC-MSCs in refractory GVHD 39 . Then, PSC-MSCs may provide another putative cellular source overcome many limitations of adult BM-MSCs.
The incidence of viral infections is another problem for allo-HSCT protocols using PTCy. In this study, the viral infection rates were lower in the MSD-PTCy cohort than those in the control cohort, both for CMV and EBV reactivation (P < 0.05). It is worth noting that no patient underwent EBV reactivation in the MSD-PTCy cohort. Patients with CMV reactivation especially have a risk for developing CMV-related diseases, including CMV pneumonia, retinitis and enteritis. Two patients and one patient with CMV seropositivity in the haplo-PTCy and haplo-control cohorts developed CMV retinitis, respectively. As a result, they developed varying degrees of visual impairment. However, the incidence of CMV reactivation in the haplo-HSCT cohort was higher than that in the haplo-control cohort, although there was no significant difference, which may be due to the additive immunosuppressive effects of ATG and PTCy. How to reduce the incidence of CMV reactivation is an outstanding problem. In the haplo-HSCT cohort, the incidence of EBV reactivation was significantly lower than that in the haplo-control cohort (P < 0.05), which was attributed to the use of rituximab at day + 5. Bacigalupo et al. showed that patients received rituximab at day + 5 after alternative donor HSCT had significantly lower rates of EBV-DNA and lower maximum median EBV copies than the control cohort (P < 0.05). Furthemore, rituximab at day + 5 had no influence on incidences of infections 40 . Pulmonary infection is one of the key life-threatening complications. The incidence of pulmonary infection in the MSD-PTCy cohort was obviously lower than that in the control cohort (P < 0.05). It is worth mentioning that the total dose of ATG was reduced from 10 to 8 mg/ kg in the MSD-PTCy cohort. Though no difference, the incidence of pulmonary infection in the haplo-PTCy cohort was lower than that in the haplo-control cohort, 40.1% vs 60.0%.
There are some limitations in the study. First, the small sample size and short follow-up period were weaknesses. Thus, a prospective study with a large sample size is necessary. In addition, further studies are required to evaluate the clinical utility of reduced-dose PTCy and UC-MSCs in preventing and alleviating GVHD, promoting engraftment and improving outcome. Despite these limitations, the protocol does not increase the high incidence of aGVHD and cGVHD due to PBSCs without compromising efficacy of MSD-HSCT and haplo-HSCT, and also decreases difficulties for donors. In conclusion, it is feasible for patients with SAA to receive reduced-dose PTCy and cotransplantation of PBSCs and UC-MSCs.

Methods
Study design and patients. The study aim was to analyze patients with SAA who underwent MSD-HSCT www.nature.com/scientificreports/ obtained from patients or their guardians. To address potential sources of bias, patients who met the standards were randomly divided into PTCy group and control group in MSD-HSCT and haplo-HSCT cohorts.
Transplantation procedure. Thirteen patients in the MSD-HSCT cohort received a conditioning regimen consisting of fludarabine (FLU) 30 mg/m 2 /day for 5 days given on days − 6 to − 2, ATG 2 mg/kg/day for 4 days given on days − 4 to − 1, and CY 24 mg/kg/day for 5 days given on days − 6 to − 2. Twenty-two patients in the haplo-HSCT cohort received a conditioning regimen consisting of FLU 30 mg/m 2 /day for 5 days given on days − 6 to − 2, ATG 2.5 mg/kg/day for 4 days given on days − 4 to − 1, CY 24 mg/kg/day for 5 days given on days − 6 to − 2, and a single 2.5 ~ 3 Gy of total body irradiation (TBI) on day − 1. Two patients who underwent haplo-HSCT without TBI suffered graft failure and underwent secondary transplantation. The second conditioning regimen included busulfan 0.8 mg/kg, q6h on days − 3 to − 2, and 3 Gy TBI on day − 1. CSA was administered with from day + 5 and PTCy 20 mg/kg was administered on days + 3 and + 4 for patients in the MSD-PTCy cohort. CSA combined with mycophenolate (MMF) was administered from day + 5 and PTCy 25 mg/kg was administered on days + 3 and + 4 for patients in the haplo-PTCy cohort. The consecutive patients in the control cohorts received short-course MTX at a dose of 10 mg/m 2 on day + 1 followed by a dose of 7 mg/m 2 on days + 3, + 6 and + 11, with CSA and/or MMF used as GVHD prophylaxis. The graft sources for all patients were PBSCs and MSCs in the PTCy cohorts. Intravenous infusions of UC-MSCs were given to all patients with PTCy on day 0 at a dose of 1 × 10 7 /kg.

Supportive care.
Patients were hospitalized in a laminar flow ward with positive pressure and high-efficiency particulate air filtration until neutrophil recovery. Standard infection prophylaxis with a quinolone or berberine, trimethoprim-sulfamethoxazole and nystatin was begun on day − 7. Standard Pneumocystis jiroveci prophylaxis was given to patients after engraftment for 1 year. Patients received ganciclovir at 5 mg/kg in the conditioning period. Broad-spectrum antibiotics, antifungals were used for agranulocytosis or fevers. Granulocyte colony-stimulating factor (G-CSF) was given intravenously starting on day + 5 at 5 µg/kg/day until the peripheral blood absolute neutrophil count (ANC) was ≥ 0.5 × 10 9 /L for 3 days. Rituximab was given as prophylaxis for EBV on day + 5 in the haplo-PTCy cohort. Blood products were irradiated with 25 Gy. CMV reactivation was monitored by quantitative PCR twice a week until at least day + 100. Preemptive therapy with ganciclovir or foscarnet was applied. EBV was also monitored weekly by quantitative PCR.

Posttransplantation complications.
Post-HSCT complications mainly involve aGVHD, cGVHD, infections and other toxicities. GVHD was graded in accordance with the consensus criteria for aGVHD 41 and cGVHD 42 . The diagnosis of GVHD depended on clinical symptoms and/or skin, oral mucosa, liver and gut biopsies [42][43][44] . The grade of nonhematologic and noninfectious toxicity was diagnosed according to the Common Terminology Criteria for Adverse Events (CTCAE, v4.0) of the US National Cancer Institute and National Institutes of Health from day 0 to day + 360.
Engraftment and chimerism. Neutrophil engraftment was defined as the first of three consecutive days with an absolute ANC ≥ 0.5 × 10 9 /L. Platelet engraftment was defined as the first day of seven consecutive days with a platelet count ≥ 20 × 10 9 /L. Primary graft failure was defined as failure to achieve an ANC of 0.5 × 10 9 /L by day + 28. Secondary graft failure was defined as an ANC < 0.5 × 10 9 /L after initial engraftment that was independent of infection, medications and insensitive to G-CSF. Full donor chimerism was defined as ≥ 95% leukocytes of donor origin in peripheral blood and/or bone marrow samples 45 . Graft rejection was defined as a progressive decrease in blood counts after initial engraftment with loss of donor chimerism.

Statistical analysis.
Median values and ranges were expressed for continuous variables and percentages for categorical variables. Continuous variables were analyzed by an independent samples T test. Differences in categorical variables were compared by Fisher's exact test. Cumulative incidences of aGVHD and cGVHD were estimated using competing risk model with early death being a competing event. OS and EFS were performed with the Kaplan-Meier method and log-rank tests, while median survival was estimated with a 95% confidence interval. A two-sided P-values less than 0.05 were considered statistically significant. Analyses were performed using GraphPad Prism for Windows, version 8.4.1 and SAS version 9.4 (SAS Institute, Cary, NC).

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request. www.nature.com/scientificreports/