Introduction

Severe aplastic anemia (SAA) is a rare, life-threatening condition that causes bone marrow failure. Therapeutic options for SAA are primarily immunosuppressive treatment (IST) or hematopoietic stem cell transplantation (HSCT). HSCT from a matched sibling donor (MSD) has been the standard of care, although only 30–40% of patients have a fully matched donor in the family. If an MSD is not available, other options include a second IST cycle or a matched-unrelated donor (MUD) HSCT [1]. Congenital amegakaryocytic thrombocytopenia (CAMT) is a rare autosomal recessive disease causing bone marrow failure that presents with severe thrombocytopenia at birth and related hemorrhagic complications. With time, it can evolve into SAA and even leukemia. Treatment options are limited to transfusion support and HSCT [2, 3].

In absence of a fully human leukocyte antigen (HLA)-matched donor, mismatched unrelated and mismatched related (haploidentical) transplants have occasionally been performed using various approaches for patients with SAA and CAMT. This approach provides the possibility of a donor for virtually all patients, with immediate availability. However, the high rate of graft failure (GF) and graft-versus-host-disease (GvHD), delayed immune recovery (IR), and severe infections can be major drawbacks [4,5,6,7].

One approach using haploidentical donors has been to ex vivo T-cell-depleted grafts using CD34+ cell enrichment, with infusion of high doses of CD34+ cells from mobilized peripheral blood to overcome the HLA barrier. Early studies using mega-doses of CD34+ cells reported rapid engraftment, but with a higher number of associated infectious complications related to the delayed IR [8]. To overcome these problems, selective elimination of αβ+ T cells (which leaves NK cells and γδ+ T cells in the graft) has been evaluated in nonmalignant conditions, including IST-refractory SAA and CAMT, allowing for a faster IR [9].

A novel 2-step partial T-cell-depletion strategy can remove naïve T cells responsible for GvHD and preserve CD34+ progenitor cells and CD45RA− memory T cells with specificity for opportunistic pathogens. When utilized in leukemia patients, rapid engraftment with low rates of corticoid-responsive GvHD and a low incidence of viral complications was observed [10]. Previously, we reported our experience using this strategy with 25 consecutive high-risk patients with leukemia, and CD45RA− depleted haplo-HSCT was well tolerated, with rapid engraftment and a low risk of severe acute and chronic GvHD. Although viral reactivations of cytomegalovirus (CMV) and adenovirus were well controlled, we observed a high rate of human herpesvirus-6 encephalitis [11].

However, this approach has not yet been reported in the context of bone marrow failure syndromes. Therefore, here, we describe our experience with CD45RA+ T-cell-depleted HSCT in patients with SAA and CAMT.

Methods

Patients

From February 2017 to July 2018, we enrolled five pediatric patients diagnosed with nonmalignant diseases who needed allogeneic transplantation but lacked a suitable HLA-matched donor (Table 1). The median age at transplantation was 9 (range 2–13) years. One patient (patient 4) received transplantation from 9/10 mismatch unrelated donors with a mismatch at HLA*A; cell processing was similar to the other four patients who received transplantation from a haploidentical donor. The only exclusion criterion was a poor clinical condition (Lansky score < 60%). The study protocol was approved by local ethics committees; informed consent was obtained from the patients themselves or from their legal guardians.

Table 1 Patient and donor characteristics
Full size table

Conditioning

The conditioning regimen included 800 cGy of total lymphoid irradiation over 4 days, 30 mg/m2/day of fludarabine over 5 days, and cyclophosphamide 50 mg/kg/day for 4 days. On day 0, the patients received their first hematopoietic progenitor cell CD34+-enriched graft. The following day, they received a second infusion of CD45RA-depleted cells.

Supporting treatment

The administration of granulocyte-colony-stimulating factor (G-CSF) was started on day + 7. All patients received infectious prophylaxis with acyclovir from day − 1 to + 180 and if GvHD or immunosuppressive treatment were present. Antifungal prophylaxis with micafungin was prescribed from day-1 until IST was suspended. Trimethoprim-sulfamethoxazole was started with conditioning treatment until day-1 and was reinitiated after the neutrophil engraftment (> 1000/mm3) until IST was stopped. Screening for CMV, Epstein–Barr virus (EBV), and adenovirus was performed weekly until day + 100 using a polymerase chain reaction test. Galactomannan antigen was performed twice a week until day + 100. BK virus study was performed when compatible signs and symptoms were observed.

GvHD prophylaxis and grading

GvHD pharmacological prophylaxis consisted of intravenous cyclosporine A 3 mg/kg/day or mycophenolate 15 mg/kg/bid starting at day − 1, maintained for 3 months and then slowly tapered. Acute GvHD (aGvHD) was graded according to standard criteria [12, 13], whereas chronic GvHD (cGvHD) was defined as mild, moderate, and severe [14] in patients surviving 100 days following transplantation.

Monitoring of immune reconstitution

Immune recovery was evaluated at the time of engraftment and then at 30, 60, 90, 180 and 360 days after transplantation. Immune reconstitution assessment included serum immunoglobulin measurement and total lymphocyte count, T cells, B cells, NK cells, and T-cell subsets (CD4+, CD8+, CD4+ CD45RA+, CD4+ CD45RO+, CD8+ CD45RA+, CD8+ CD45RO+).

Preparation of the CD45RA-depleted graft

Donors were mobilized with G-CSF (10 μg/kg/day for 4 days) and harvested on day + 5 by 1–2 leukapheresis procedures. The first product was T-depleted using a CliniMACS device (Miltenyi Biotec, Bergisch Gladbach, Germany) to enrich the graft for CD34+. The minimum cell dose required for CD34 was 5 × 106/kg. The unselected CD34− fraction was then processed for CD45RA+ cell depletion using the CliniMACS device and Depletion 3.1 software. The maximum CD3+ CD45RA+ dose allowed was 1 × 104/kg and the minimum depletion of CD45RA+ cells was ≥ 2.5 log10 (Fig. 1). The spillover of CD45RA− cells were frozen in separate aliquots. Outcomes were censored as of December 2020.

Fig. 1
figure 1

Flow cytometry charts for leukocytes and T-lymphocytes in pre- and post-CD45RA-depletion cellular product

Full size image

Engraftment definitions

Neutrophil engraftment was defined by the first of 3 consecutive days with an absolute neutrophil blood count > 0.5 × 109/L. Platelet engraftment was defined as the first day of platelet count > 20 × 109/L without transfusional support in the previous 7 days. Primary GF was defined as the absence of neutrophil recovery by day + 28 with platelet count < 20 × 109/L and Hb < 80 g/L. Secondary graft failure was defined as the presence of cytopenias in a patient who was already engrafted.

Telomere length determination

Telomere length was determined by Southern blots of enzymatically digested DNA, as previously reported from patients with inherited bone marrow failure [15, 16].

Results

Patients and cellular doses

Five patients underwent an HSCT from haploidentical (4) or mismatched unrelated donors (1) between February 2017 and July 2018. Donor characteristics are shown in Table 1. Patients received 2 cellular products: the CD34+ fraction contained a mean of 6.58 × 106/kg cells (range 5–7.73) and the CD45RA-depleted fraction contained a mean of 3.84 × 107/kg cells (range 2.8–5). Graft composition is shown in Table 2. Median log10 depletion of CD45RA+ cells was 4.99 (range 3.73–6.61).

Table 2 Graft composition and postransplantation evolution
Full size table

Engraftment and GvHD

All patients engrafted neutrophils (> 0.5 × 109/L) at a median of 10 days (range 9–11). Full donor chimerism was achieved by day + 30 and maintained over time in all patients. Two of 5 (40%) patients (patients 1 and 3) developed aGvHD grade ≥ 2 with gut, liver, and skin involvement (histologically confirmed) rapidly responsive to steroids (details in Table 2). Chronic GvHD was observed in one patient (patient 1), with severe lung involvement confirmed by histopathology.

Infections

CMV reactivation was observed in the blood on days + 21 and + 23 in two patients (patients 1 and 2); and another patient (patient 3) presented with CMV-associated hemorrhagic cystitis on day + 34. Additionally, there was one case of EBV reactivation on day + 135 (patient 3) and two episodes of adenovirus reactivation on days + 53 and + 150 (patients 3 and 5). One patient presented BK reactivation on day + 330 (patient 3), and another showed grade 4 hemorrhagic cystitis on day + 165 (patient 2). Other infectious complications before day + 100 included typhlitis in 1 patient (patient 2), which resolved with antibiotic treatment, and parainfluenza-1 virus pneumonitis and Clostridium difficile infection in another patient (patient 4). Adenovirus, BK virus, and parainfluenza reactivations did not need treatment with antiviral agents.

Follow-up and thrombotic microangiopathy

Three (60%) patients remain alive, with a median follow-up of 881 days post-transplant (range 323–1248 days). Two of the 5 patients (patients 1 and 2) died, on days 321 and 363, respectively. Patient 1 presented with severe lung cGvHD complicated by respiratory syncytial virus (RSV) pneumonia. Patient 2 presented with transplant-related thrombotic microangiopathy (TMA). Both patients needed invasive mechanical ventilation. TMA was observed in 3 (60%) of the five patients (patients 1, 2, and 3); these three patients were previously exposed to cyclosporine A during IST. All the patients with SAA had normal telomere-length values compared with their parents.

Post-transplant adoptive CD45RO + lymphocyte infusion

Three patients received post-HSCT CD45RA− depleted donor lymphocyte infusions (DLIs) due to lymphopenia and viral reactivations (patients 1, 2, and 3) (Table 2). Two patients received 2 infusions (patients 1 and 2) and another patient received only one (patient 3). The mean CD3+ CD45RO+ dose per infusion was 2.38 × 107/kg (range 0.83–4). The median CD4/CD8 ratio infused in the DLI was 3.69. All DLIs were infused during GvHD prophylaxis, except for second DLI in patient 2 (day + 194). After DLI, the viral load decreased in all patients.

A fourth patient (patient 5) received 1 × 106/kg CD3+ CD45RO+ prophylactic DLI on days + 30, + 60 and + 90 (Table 2). After immunotherapy with DLI, only one patient of four who received a DLI presented with de novo grade I skin aGvHD (patient 2). In the other patients, GvHD was not related to DLI.

Immune reconstitution

On day + 30, immune reconstitution showed a mean lymphocyte count of 0.54 × 103/μl (range 0.38–0.67), with a mean of 0.33 × 103/μl (range 0.21–0.47) T cells, a mean of 0.158 × 103/μl (range 0.03–0.35) NK cells, and very low B cells mimicking the graft. T-cell counts increased progressively, achieving a mean of 0.682 × 103/μl (range 0.18–1.49) T cells, a mean of 0.255 × 103/μl (range 0.006–0.59) NK cells, and a mean of 0.15 × 103/μl (range 0.0002–0.58) B cells at 6 months after HSCT (Fig. 2).

Fig. 2
figure 2

Immune reconstitution pre and after transplantation on days + 15, + 30, + 60, + 90, + 180 and + 360 (median values)

Full size image

Discussion

Our results using CD45RA-depleted HSCT in haploidentical and mismatched unrelated donors for children with bone marrow failure syndromes show rapid neutrophil recovery within 10 days, similar to other T-cell-depletion strategies. GF was not observed. Grade > II acute and severe chronic GvHD appeared in 2/5 and 1/5 patients, respectively. Acute GvHD was universally responsive to steroids. Given the frequency of GvHD was higher than expected in the first three patients, we decreased the maximum allowed CD45RA− and CD45RA+ T-cell doses in the graft to 3 × 107/kg and 1 × 104/kg, respectively. In addition, DLI with CD3+ CD45RA− T cells was also decreased to 1 × 106 T cells/kg/dose. No severe infections were observed before day + 100. CMV reactivation occurred in three patients, being the most common infectious complication.

Recently, new strategies for haploidentical HSCT with unmanipulated grafts have been developed for nonmalignant diseases. Post-HSCT cyclophosphamide has been used sporadically for patients with SAA and CAMT, especially in adults [7, 17, 18]. Another strategy includes G-CSF-primed bone marrow and G-CSF peripheral blood stem cells and significant immunosuppression for GvHD prophylaxis. Xu et al. reported 52 pediatric patients with SAA receiving haploidentical HSCT with median myeloid recovery of 12 days (range 10–22 days). No primary GF occurred; however, three patients experienced secondary GF. Acute GvHD, grades II–IV, was observed in 39.2% of patients, and cGvHD occurred in 38.1%. With a median follow-up of 744 days, overall survival (OS) was 84.5% [19].

Haploidentical HSCT strategy for pediatric patients using ex vivo T-cell depletion with purified high doses of CD34+ cells from G-CSF mobilized donors presented high rates of GF and slow IR [9, 20]. Partial T-cell depletion can eliminate alloreactive cells from the graft while retaining cells that help faster IR and sustained engraftment. Several groups reported this approach for patients with nonmalignant disorders [9, 21]. A Korean group recently reported haploidentical HSCT using depletion of αβ+ cells, retaining γδ+ lymphocytes that enhance IR and are not implicated in GvHD. Five patients received αβ-depleted cells and 16 received CD3-depleted grafts [19]. Bertaina et al. reported 23 pediatric patients with nonmalignant disorders, including 4 with SAA and 1 with CAMT, using haploidentical donors and αβ/CD19 depletion. Conditioning differed according to primary disorder; however, all patients received anti-thymocyte globulin for preventing graft rejection and GvHD. There were four cases of GF. Hematological recovery was rapid, with a median of 13 (range 10–20) days to neutrophil engraftment. Three children developed grades I–II acute GvHD, and no cases of chronic GvHD were observed during a median follow-up of 18 months [9]. In this study, the only two patients who died were affected by SAA and died of RSV infection and TMA.

Partial T-cell-depleted transplantation from HLA-identical siblings was recently reported by Sanz et al. in 26 patients with SAA. There were seven pediatric patients and 19 adults. Engraftment was achieved at a median of 11 (range 10–14) days. Acute GvHD grade I was observed in two patients, but there was no grade II–IV GvHD. Only two patients presented mild skin cGvHD, and 1 had moderate skin and ocular involvement. The OS at 10 years was 81% [22].

The high incidence (60%) of TMA in our patients could be related to the prolonged exposure to calcineurin inhibitors prior to the transplant, the use of mismatch donors, and the use of cyclosporine as post-transplant immunosuppression, given the three patients had received two courses of IST with cyclosporine A. Two of the three patients received cyclosporine for GvHD prophylaxis, and the other patient received mycophenolate as prophylaxis and sirolimus for GvHD treatment. SAA has been linked to increased rates of TMA in previous studies, and has been mentioned as an underlying predisposition to complement activations or preexisting endothelial injury [23, 24].

The CD45RA-depletion strategy allows for post-HSCT immunotherapy, and four patients received DLI of CD45RA-depleted cells to boost the IR. In three patients, DLI was prescribed to treat viral reactivations/infections and to increase lymphocyte counts. In these five infusions, the mean dose of CD3+ CD45RO+ cells was 2.38 × 107/kg/dose. In a separate patient, DLI was administered under a prophylactic regime on day + 30, + 60 and + 90 using lower doses (1 × 106/kg) of CD3+ CD45RO+. All DLIs were infused during GvHD prophylaxis, except for 2° DLI in patient 2 (day + 194). Even if immunosuppressive drugs impacted the IR in these patients, DLI showed a decrease in the CMV viral load in all patients. Maschan et al. used low-dose memory T-cell DLI after TCRαβ-depleted HSCT to improve the immune response to common pathogens in 53 patients and 134 infusions. This strategy used escalating doses of memory T cells at monthly intervals, with a dose range of 25–100 × 103/kg and 100–300 × 103/kg for MUD associated with a low risk of de novo GvHD but with a risk of reactivating preexisting GvHD [25]. The largest series of patients reported to date targeting CD45RA+ cells as strategy to deplete naïve T cells but preserving memory T cells in hematological malignancies showed an increase in survival due to a reduction in transplant-related mortality without a significant increase in the rates of GvHD. In that series, the patients also received NK cells, so this strategy could help reduce GvHD [26].

In conclusion, we suggest that CD45RA+ cell depletion in haploidentical and unrelated donors needs further improvement for children with bone marrow failure syndromes such as SAA and CAMT who lack a suitable matched related donor. Despite presenting rapid and sustained engraftment, the incidence of GvHD is still higher than expected. Therefore, reducing CD3+ CD45RA+ cells in the graft and DLI should be investigated. Our results showed lower rates of GvHD and infections using lower doses of CD3+ CD45RO+ cells; however, more experience is needed to confirm these data. In addition, we must mention the high incidence of TMA in this group of patients, which could be related to long-term immunosuppressive treatment. A change in the GvHD prophylaxis protocol avoiding cyclosporine could be beneficial in these patients to lower the risk of TMA, although a greater number of cases need to be evaluated to draw conclusions and to confirm this hypothesis regarding TMA and its association with immunosuppressive treatment in this group. This potential association, however, could underscore the importance of designing an effective GvHD prevention strategy that avoids prolonged use of calcineurin inhibitors in this heavily pretreated population. Although this T-cell-depletion approach also allows selectively depleted DLIs for post-HSCT immunotherapy to boost IR, the ideal dose regimen for minimizing GvHD toxicity and maximizing viral protection remains unclear.