Peripheral Stem Cell Apheresis is Feasible Post ¹³¹Iodine-Metaiodobenzylguanidine-Therapy in High-Risk Neuroblastoma, but Results in Delayed Platelet Reconstitution

Neuroblastoma (NBL) is the most common extracranial solid tumor in children, accounting for 7% to 10% of all childhood malignancies (1). The majority of children presenting with NBL have “high-risk (HR) disease” with amplification of the MYCN oncogene and/or distant metastases at diagnosis, mainly involving bone marrow (BM; ref. 2). Despite the implementation of a multimodal therapy, including induction chemotherapy, surgery, autologous stem cell transplantation (ASCT), and immunotherapy, the prognosis of patients with HR-NBL is still poor. More than half of the patients with HR-NBL experience disease recurrence and long-term survival remains less than 40% (2). This poor outcome necessitates the search for new therapies.

An alternative treatment modality involves metaiodobenzylguanidine (MIBG), a norepinephrine analogue. Approximately 90% of patients with NBL have “MIBG-avid” disease, that is, MIBG will accumulate in the NBL cells (3). MIBG is therefore used as an imaging agent for diagnostic purposes, when radiolabeled with iodine-123, but is also used as a form of targeted radiotherapy when labeled with iodine-131 (¹³¹I). In recurrent or refractory NBL, response rates of ¹³¹I-MIBG treatment range from 20% to 40% (4–9). Dose-limiting toxicity is myelosuppression and support with ASCT at ¹³¹I-MIBG doses of ≥12 mCi/kg is advised (6). When used as upfront therapy in newly diagnosed patients with HR-NBL, thus prior to chemotherapy, objective response rates of up to 70% have been reported (10–12). Recently, administration of ¹³¹I-MIBG during induction chemotherapy and prior to myeloablative therapy (MAT) was shown to be feasible (www.clinicaltrials.gov: NCT01175356; ref. 13), and will be further studied in a prospective randomized trial (NCT03126916, ref. 1). Moreover, a combination of ¹³¹I-MIBG and Topotecan will be studied as an intensification treatment strategy for patients with inadequate response after induction to proceed to MAT and ASCT (NCT03165292). Thus, optimal scheduling of ¹³¹I-MIBG in the high-risk treatment plan is currently being investigated.

As ¹³¹I-MIBG-therapy-related hematologic side effects have been reported, we questioned if ¹³¹I-MIBG, when given upfront, would affect hematopoietic stem and progenitor cells (e.g., CD34⁺ cells) and/or the BM microenvironment, hence impairing the ability to harvest mobilized CD34⁺ cells. In a pilot study, that mainly focused on upfront ¹³¹I-MIBG-therapy toxicity and efficacy, we observed a CD34⁺ cell harvest failure in only 2 of 21 patients (14). The primary aim of this study was to evaluate feasibility of stem cell apheresis after upfront ¹³¹I-MIBG-therapy in a larger cohort of patients with HR-NBL, and determine the effect on hematologic reconstitution after ASCT. This was combined with an in-depth analysis of the quality of the harvested CD34⁺ cells by studying post-thaw viability, clonogenic capacity, and phenotype.

All patients included in this study were patients with HR-NBL (≥1–19 years, stage 4 or MYCN-amplification) treated according to the prospective Dutch Childhood Oncology Group (DCOG), multicenter cohort protocols: pilot phase (2005–2011) and NBL-2009 (2011–October 2015). In these protocols, patients with MIBG-avid disease were treated with two courses of upfront ¹³¹I-MIBG-therapy, followed by standard HR-therapy, called: “MIBG-therapy” group (Fig. 1). In the pilot phase, ¹³¹I-MIBG was administered as a fixed dose: first course ¹³¹I-MIBG dose was 7.4 GBq (200 mCi) and second course 5.5 GBq (150 mCi; ref. 14). In the NBL-2009 study, we aimed to limit the whole-body dose to 4 Gy for the two consecutive ¹³¹I-MIBG administrations. After the first administration (444 MBq/kg), the second dose was based on the total-body radiation dose calculated from the first therapeutic administration. Patients with MIBG-non-avid disease and patients that were too ill for protective nuclear isolation (e.g., superior vena cava syndrome, risk of optic nerve compression) or with uncontrollable hypertension, were excluded to receive ¹³¹I-MIBG-therapy and directly treated with standard HR chemotherapy: the “chemotherapy-only” group (Fig. 1). Thus, patients were not randomly assigned to a patient group. Standard HR-therapy consisted of induction chemotherapy, surgery, MAT with ASCT and radiotherapy to the primary tumor site [identical to the Gesellschaft fur Pädiatrische Onkologie und Hämatologie (GPOH) NB2004 NBL-HR protocol, as previously described; ref. 14)]. Number of patients included in a previous cohort: Gooskens and colleagues (15): 24 patients, Kraal and colleagues (14): 32 patients. The study was conducted in accordance with the ethical guidelines of the Declaration of Helsinki. Written informed consent was obtained from patients, parents, or legal representatives.

As is common practice in DCOG HR-NBL treatment protocols, CD34⁺ cells were harvested after the BM was cleared from tumor. In case BM was not cleared after standard induction chemotherapy (N5/N6), patients received an additional N8 chemotherapy course. Post-chemotherapy, 10 μg/kg G-CSF was administered subcutaneously. Plerixafor was not used. When CD34⁺ cell blood counts reached >20/μL, apheresis was performed, aimed at collecting ≥2 × 10⁶ CD34⁺ cells/kg. The sequential number of days needed for collecting sufficient CD34⁺ cells was registered. In case of insufficient yield, a second apheresis session was attempted after the subsequent chemotherapy course.

Patients with good response to induction therapy [complete response (CR), very good partial response (VGPR) or partial response (PR)] were allowed to proceed to ASCT, with reinfusion of ≥2 × 10⁶/kg CD34⁺ cells (as measured prior to cryopreservation). Hematologic reconstitution post-ASCT was defined as a platelet count >20 × 10⁹/L (without transfusions) and neutrophil count >0.5 × 10⁶/L. In case of thrombocytopenia, platelet transfusions were not standard of care, only in case of severe hemorrhage platelet transfusions were given.

Viability and vitality testing of harvested CD34⁺ cells and nucleated blood cells (NBC), respectively, was routinely performed prior to cryopreservation. NBC vitality testing was performed using trypan blue exclusion. Cell recovery after cryopreservation was calculated as the number of nucleated cells post-thawing divided by the number of cells prior to cryopreservation. CD34⁺ cell viability was determined as previously described (ISHAGE guidelines; ref. 16), combined with 7-amino actinomycin D (7-AAD) staining (BD biosciences), and measured using a CANTO ll flow cytometer (BD Biosciences). Minimal two-hundred thousand CD45⁺ events were collected. Viable CD34⁺ cells were defined as 7-AAD negative. On a selection of 19 patients (9 ¹³¹I-MIBG and 10 chemotherapy-only–treated patients), CD34⁺ cell viability was tested post-thawing. This “subgroup” was selected based on availability of separate cryopreserved reference aliquots from the apheresis, harvest yield and dose of re-infused CD34⁺ cells (evenly distributed between the two groups). Clinical patient characteristics of the subgroup were comparable to the other patients.

Progenitor capacity of collected CD34⁺ cells was assessed using the CFU-GM assay, which was performed standard prior to cryopreservation (n = 81 samples of 44 patients). Additionally, one of the centers performed CFU-GM assays using post-thaw CD34⁺ cells of the above described “subgroup” of 19 patients. Nucleated cells were plated in duplicate in 35 mm tissue culture plates (concentrations: 1.0, 0.5, and 0.25 × 10⁵ cells/mL), in MethoCult GF 4534 (StemCell Technologies). Cultures were incubated for 12 to 14 days at 37°C (5% CO₂). CFU-GM colonies, containing at least 40 translucent cells, were scored in triplicate by microscopy (Leica). CFU-GM recovery was calculated as the number of colonies formed post-thawing divided by the number of colonies prior to cryopreservation.

Of the “subgroup” of 19 patients, post-thaw CD34⁺ cells were characterized for surface marker expression by flow cytometry. Cells were washed, re-suspended in PBS containing 0.2% BSA and incubated (20 minutes, room temperature) with the following monoclonal-antibodies: Antibodies purchased from BD biosciences: CD45-PerCP (clone 2D1), CD34-APC (clone 8G12), CD62L-FITC (clone SK11), CD33-PE-Cy7 (clone p67.6), IgG2a-FITC, IgG1-PE, IgG1-PeCy7. Purchased from Dako: CD45-PB (clone T29/33). Purchased from Beckman Coulter: CD41-PE (clone P2). Isotype controls were used to set gating thresholds.

Groups were compared using the Chi square test for categorical variables and the independent Student t test for continuous variables. A multivariate linear regression model was used to study the association between patient characteristics, treatment and CD34⁺ cell harvest. To account for repeated measures, a generalized linear mixed model (GLMM) was used to estimate marginal mean harvest quality (CFU-GM per CD34⁺ cell) of the first apheresis day for each group. GLMM is a well-known statistical methodology used to study data that are correlated within subjects (17). The adjusted mean with corresponding standard error and confidence intervals were computed for each group. Percentage of CD33-, CD41-, and CD62L-expressing CD34⁺ subsets and cell vitality/viability were compared between the two groups using the Mann–Whitney U test and t test. Survival analysis techniques were used to compare time to platelet and neutrophil reconstitution for patients treated with ¹³¹I-MIBG or chemotherapy-only. The log-rank test has been used to assess the statistical significant difference between the two groups. Time to event was defined as time from infusion of CD34⁺ cells (ASCT) until time of platelet or neutrophil reconstitution. Patients who did not engraft after the first ASCT were censored at time of second infusion. A multivariate Cox proportional hazards regression model was used to estimate the effect of risk factors on platelet and neutrophil reconstitution. Results are presented as hazard ratios (HR) with the corresponding 95% confidence interval (CI).

Eighty-one children were included: 38/81 (47%) treated with upfront ¹³¹I-MIBG-therapy and 43/81 (53%) received chemotherapy-only. The median age (range) at diagnosis was 3.3 (0.1–16.4) years (Table 1). Nearly all patients had BM metastases at diagnosis (72/81; 89%). MYCN-amplification was detected in 9/36 (25%) of ¹³¹I-MIBG-treated patients, compared with 19/38 (50%) of chemotherapy-only patients (P = 0.034). The enclosed CONSORT figure (Fig. 2) shows the flow of the patients from enrollment to collection and reinfusion of CD34⁺ cells.

The first median ¹³¹I-MIBG dose was 0.42 GBq/kg (range 0.13–0.56)/11.2 mCi/kg (3.5–15.2). For patients treated with two courses, the second median dose was 0.37 GBq/kg (range 0.12–0.69)/9.9 mCi/kg (3.2–18.7) and the total cumulative median dose was 0.81 GBq/kg (range 0.26–1.10)/22.1 mCi/kg (7–29.8; Table 1). Eight patients received only one course of ¹³¹I-MIBG, with a median cumulative dose 0.41 GBq/kg (0.17–0.56).

Seventy-one patients underwent apheresis: 34 (89%) of the ¹³¹I-MIBG-therapy group and 37 (86%) of the chemotherapy-only group. There were no significant differences in timing of apheresis between the chemotherapy-only and ¹³¹I-MIBG-therapy group (P = 0.890, Fisher exact test). In both groups, median timing of apheresis was after the fourth chemotherapy course (Table 2). Apheresis in ¹³¹I-MIBG and chemotherapy-only patient groups yielded a comparable total number of CD34⁺ cells/kg: median of 5.4 × 10⁶ (range 0.9–32.3) in ¹³¹I-MIBG- compared to 5.6 × 10⁶ (range 0.5–44.5) in chemotherapy-only–treated patients (Table 2). The number of apheresis days and sessions required to collect sufficient CD34⁺ cells were also comparable between both groups: one apheresis day was sufficient to collect ≥2 × 10⁶/kg CD34⁺ cells in 59% ¹³¹I-MIBG-therapy and in 65% chemotherapy-only patients, 2 days in respectively 74% and 76% (Table 3). For 4% of the patients, additional BM harvesting was performed because the number of collected CD34⁺ cells by apheresis was not sufficient: one patient of the ¹³¹I-MIBG -therapy group and two patients of the chemotherapy-only group. A multivariate regression analysis of CD34⁺ cell harvest yield was performed, showing no association with the cumulative ¹³¹I-MIBG dose (Supplementary Table S1). Instead, CD34⁺ cell harvest yield did significantly associate with BM infiltration at diagnosis, when adjusted for age, gender, MYCN-amplification, LOH of chromosome region 1p, and cumulative dose of both ¹³¹I-MIBG and Cisplatin prior to apheresis (P = 0.004). Taken together, total harvest yield and collection time (number of days and sessions) of apheresis were comparable between both patient groups, indicating that apheresis is feasible after upfront ¹³¹I-MIBG-therapy.

Fifty-nine patients underwent ASCT: 28 (74%) of the ¹³¹I-MIBG-therapy group and 31 (72%) of the chemotherapy-only group. Patients that did underwent stem cell harvest, but did not proceed to ASCT, had progressive disease (PD; ¹³¹I-MIBG group: n = 8, chemotherapy-only group: n = 6) or died (chemotherapy-only group: n = 1; Fig. 2). Median dose (range) of infused CD34⁺ cells was 3.4 × 10⁶/kg (1.2–10.5) in ¹³¹I-MIBG patients and 3.5 × 10⁶/kg (1.2–11.6) in chemotherapy-only patients (Table 2). After ASCT, the median time (95% CI) to platelet reconstitution was 29 days (11–47) and 15 days (12–18) for ¹³¹I-MIBG and chemotherapy-only group, respectively (log-rank overall 0.037; Table 2; Fig. 3). The delayed time to platelet reconstitution in ¹³¹I-MIBG-treated patients was statistically but not clinically significant, as it did not result in hemorrhages or an extended length of hospital stay. Time to neutrophil reconstitution was respectively 11 days (10–12) and 10 days (refs. 9–11; log-rank overall 0.734; Table 2; Supplementary Fig. S1). A multivariate Cox's regression model was performed to estimate the effect of cumulative ¹³¹I-MIBG dose, number of infused CD34⁺ cells at ASCT and BM infiltration at diagnosis, on both platelet and neutrophil reconstitution. A significant statistical association was found between both cumulative dose of ¹³¹I-MIBG (HR 0.395; 95% CI, 0.19–0.85; P = 0.017) and number of infused CD34⁺ cells at ASCT (HR 1.242; 95% CI, 1.1–1.4; P = 0.001) with platelet reconstitution (Table 4). Concerning neutrophil reconstitution, there was a significant association with both BM infiltration at diagnosis (HR 0.377; 95% CI, 0.16–0.89; P = 0.026) and the number of infused CD34⁺ cells (HR 1.282; 95% CI, 1.13–1.46; P < 0.0001), but not with the cumulative dose of ¹³¹I-MIBG (Table 4).

In two patients (¹³¹I-MIBG group) successful hematologic reconstitution was only achieved after a second stem cell infusion. A third patient (chemotherapy-only group) suffered from failure to engraft after two autologous stem cells infusions. Although additional allogeneic cord blood transplantation resulted in neutrophil reconstitution within 12 days, the patient died after 1 month due to septic disease and multiorgan failure, before platelet recovery was achieved.

In conclusion, treatment of patients with HR-NBL with upfront ¹³¹I-MIBG results in timely myeloid but delayed platelet reconstitution after ASCT.

In search of a possible explanation for the delayed platelet recovery after ASCT in ¹³¹I-MIBG-treated patients, we compared the quality of the harvested cells of the two patients groups by analyzing viability. In addition, functional activity of the harvested CD34⁺ cells was assessed using a colony-forming unit assay that determines clonogenic capacity, that is the capacity to differentiate into granulocyte/macrophage progenitors (CFU-GM). Quality assessment was routinely performed prior to cryopreservation (pre-cryo). Analysis of 81 pre-cryo apheresis samples obtained from 44 (54%) patients showed no significant difference in NBC vitality and clonogenic output (CFU-GM/CD34⁺ cell) between the¹³¹I-MIBG and chemotherapy-only group (Supplementary Fig. S2; Supplementary Table S2). Moreover, CD34⁺ cells that were collected during the first apheresis did not differ in their clonogenic capacity compared with cells collected after multiple apheresis days (Supplementary Table S2). For a selection of 19 patients (9 ¹³¹I-MIBG and 10 chemotherapy-only), CD34⁺ cell viability and functioning was additionally tested post-thawing, on a separate apheresis aliquot. Although NBC vitality (Fig. 4A) and recovery (Fig. 4B) were comparable, we found a significant lower percentage of viable CD34⁺ cells in post-thaw apheresis samples of ¹³¹I-MIBG- compared with chemotherapy-only–treated patients, 63% and 83% respectively (Fig. 4C). Clonogenic output of CD34⁺ cells of these 19 patients was highly variable (as commonly observed for CFU-GM), in both pre-cryo and post-thaw samples, and did not significantly differ between the two groups. Median CFU-GM potential (range) prior to cryopreservation was 30.2 × 10⁴/kg (9.0–173.8) in ¹³¹I-MIBG- treated patients versus 71.1 × 10⁴/kg (33.0–378.1) in chemotherapy-only patients (P = 0.203) and CFU-GM recovery after cryopreservation was comparable (Fig. 4D).

To assess whether the delay in platelet recovery may additionally be due to exhaustion of specific progenitor cell subsets in the transplant, we next tested CD34⁺ cells phenotypically. Viable CD34⁺ cells in post-thaw apheresis samples were characterized by markers that indicate early myeloid (CD33) or megakaryocytic (CD41) differentiation using flow cytometry. Cell surface expression of CD33 and CD41 was not significantly different on CD34⁺ cells of the two patient groups (Fig. 4E and F). We also compared the percentage of viable CD34⁺ cells expressing the adhesion molecule CD62L (L-selectin), which appeared to be lower in the ¹³¹I-MIBG compared with the chemotherapy-only group: 37% and 54%, respectively (P = 0.0481; Fig. 4G). Interestingly, CD62L is proposed to be a predictive marker for platelet recovery after ASCT (18). In line, our analysis showed a moderate negative correlation (r = −0.627, P = 0.009) between the percentage of re-infused CD62L-expressing CD34⁺ cells and the time to platelet recovery (Fig. 4H). Thus, the post-thaw viable CD34⁺ cell count was lower in apheresis samples of ¹³¹I-MIBG-treated patients and expression of CD62L, a predictive marker for platelet recovery, was reduced.

¹³¹I-MIBG is an important established treatment for relapsed or refractory NBL and its efficacy is currently investigated in front-line setting. The optimal timing of ¹³¹I-MIBG-therapy during front-line treatment is not yet established. Pilot studies have demonstrated feasibility when given at the time of diagnosis (10, 14, 19) and cooperative groups in both Europe and North America currently investigate its use as part of induction or consolidation therapy (www.clinicaltrials.gov: NCT03126916; NCT01175356, NCT03165292; ref. 13). When given as front-line treatment, ¹³¹I-MIBG-therapy is mostly followed by ASCT. Therefore, there is an urgent need to get insight in the impact of ¹³¹I-MIBG on stem cell apheresis and on engraftment after reinfusion.

By studying our unique upfront ¹³¹I-MIBG-therapy cohort, we found that stem cell apheresis is feasible post-MIBG. Treating patients with ¹³¹I-MIBG early in induction did not affect the total CD34⁺ cell harvest yield and did not extend the apheresis episode. Failure to harvest sufficient CD34⁺ cells by apheresis occurred in only one ¹³¹I-MIBG- and two chemotherapy-only–treated patients. Of interest, our findings indicate that BM tumor infiltration at diagnosis did impair the mobilization of CD34⁺ cells, as described for other tumors (20), even though apheresis only started after clearing of initial BM disease. Concerning the timing of apheresis, there are different approaches: harvesting is performed after two induction chemotherapy courses in North America, as the Children's Oncology group previously showed that this was safe and feasible (21), whereas the consensus in Europe is still to harvest stem cells after the BM is cleared from tumor cells or post induction therapy. The cumulative median ¹³¹I-MIBG dose administered to the newly diagnosed patients in our study was relatively high compared to the reported maximum tolerated dose of 12 mCi/kg for intensively pretreated patients (6), but no stem cell rescue was required. Toxicity and efficacy of upfront ¹³¹I-MIBG-therapy, also for part of this cohort, has been previously described (10, 14, 19). Of note, comparisons between the two patients groups should be interpreted with caution as this study was nonrandomized and patients of the chemotherapy-only group were excluded to receive ¹³¹I-MIBG-therapy because of poor clinical condition or non-MIBG avid disease.

After reinfusion of the collected CD34⁺ cells, time to neutrophil reconstitution was similar in ¹³¹I-MIBG- compared with chemotherapy-only –treated patients, but time to platelet reconstitution was prolonged. More prominent thrombocytopenia than neutropenia has been previously described for ¹³¹I-MIBG in intensively pretreated NBL patients (6, 22–24). This differential toxicity towards platelets and neutrophils might, in part, be related to selective uptake of ¹³¹I-MIBG by platelets or their precursors (25). The prolonged time to platelet reconstitution that we observed was not major, that is, it did not result in hemorrhages or an extended length of hospital stay. Nevertheless, the duration of thrombocytopenia after treatment with ¹³¹I-MIBG and ASCT could delay additional treatment of an aggressive tumor. Hence, in light of shortening of platelet engraftment periods, we further searched for potential explanations for the ¹³¹I-MIBG-related delay in recovery.

Our in-depth analysis of the quality of harvested cells from a subgroup of 19 patients revealed that post-thaw aliquots of ¹³¹I-MIBG-treated patients contained lower viable CD34⁺ cell counts. As no significant differences in harvest quality were observed in pre-cryo samples, this suggests that CD34⁺ cells of ¹³¹I-MIBG-treated patients are more sensitive to cryopreservation, which might result in reinfusion of a lower actual number of viable CD34⁺ cells than estimated. A dose–response relationship between re-infused CD34⁺ cells and hematologic recovery was found by us and others (26, 27). Below a threshold of 1 × 10⁶ CD34⁺ cells/kg, the likelihood of delayed recovery of platelets was demonstrated to increase significantly (28). We therefore attempted to achieve a minimum number of 2.0 × 10⁶. However, these thresholds are set based on the amount at the time of collection. Based on our findings, it would be valuable to include quantification of post-thaw viable CD34⁺ cells, which is also proposed by others as a more accurate predictor of hematologic reconstitution (29).

Delay in platelet recovery may additionally be explained by exhaustion of specific progenitor cell subsets (18, 30, 31). We showed that the percentage of CD62L-expressing viable CD34⁺ cells was reduced in apheresis aliquots of ¹³¹I-MIBG-treated patients. A correlation between the number of re-infused CD34⁺/CD62L⁺ cells and platelet recovery was previously described, and suggests a role for CD62L in engraftment (18, 31, 32). CD62L-mediated rolling of CD34⁺ cells on the endothelium is suggested to be a critical step in the homing process to the BM. Although involvement of CD62L in megakaryopoiesis has also been proposed, this requires further investigation as blocking of the CD62L-ligand interaction in CFU-megakaryocyte (CFU-MK) assays did not impair clonogenic outgrowth of CD34⁺ cells into megakaryocyte progenitors (33).

Considering that therapy for HR-NBL is intense with high doses of different chemotherapeutics, and the need to harvest 6 × 10⁶/kg CD34⁺ cells in current high-risk protocols for tandem transplants, we advise that determination of viable CD34⁺ cell counts in post-thaw samples should be part of the routine quality assessment. Nevertheless, solely post-thaw CD34⁺ cell viability does not necessarily correlate with engraftment for some patients. For example, one patient had to undergo allogeneic cord blood stem cell transplantation after engraftment failure of both autologous harvests, despite adequate post-thaw viable CD34⁺ cell counts. Unfortunately, no functional testing could be performed because no aliquots remained after the two reinfusions. Combining post-thaw values of CD34⁺ cell counts with functional (CFU-GM) testing is expected to further improve routine quality assurance (34). Larger prospective cohort studies should be performed to explore whether determination of CD62L status is a useful addition to CD34⁺ cell testing.

In conclusion, we provide evidence that CD34⁺ cell harvesting is feasible after upfront ¹³¹I-MIBG-therapy in newly diagnosed patients with HR-NBL. After reinfusion, timely neutrophil but delayed platelet reconstitution occurred in ¹³¹I-MIBG- compared with chemotherapy-only–treated patients. Our findings suggest that ¹³¹I-MIBG-treated patients with prior BM tumor infiltration should be monitored more closely and the minimum acceptable number of CD34⁺ cells/kg for reinfusion should be based on post-thaw viability counts, but the impact does not seem to be so great as to preclude the upfront use of ¹³¹I-MIBG in these patients. Nevertheless, in light of our findings, ¹³¹I-MIBG administration post CD34⁺ cell collection is preferred, as will be further studied in upcoming prospective trials (www.clinicaltrials.gov: NCT03126916; NCT03165292).

No potential conflicts of interest were disclosed.

Conception and design: K.C.J.M. Kraal, H.M. Kansen, C. van den Bos, J. Zsiros, H. van den Berg, S. Somers, M.M. van Noesel, M. Fiocco, H.N. Caron, C. Voermans, G.A.M. Tytgat

Development of methodology: K.C.J.M. Kraal, H.M. Kansen, J. Zsiros, H. van den Berg, M.M. van Noesel, C.E. van der Schoot, G.A.M. Tytgat

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K.C.J.M. Kraal, C. van den Bos, J. Zsiros, H. van den Berg, E. Braakman, A.M.L. Peek, M.M. van Noesel, C.E. van der Schoot, G.A.M. Tytgat

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): K.C.J.M. Kraal, I. Timmerman, H.M. Kansen, J. Zsiros, M.M. van Noesel, C.E. van der Schoot, M. Fiocco, C. Voermans, G.A.M. Tytgat

Writing, review, and/or revision of the manuscript: K.C.J.M. Kraal, I. Timmerman, H.M. Kansen, C. van den Bos, J. Zsiros, H. van den Berg, S. Somers, E. Braakman, M.M. van Noesel, C.E. van der Schoot, M. Fiocco, H.N. Caron, C. Voermans, G.A.M. Tytgat

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K.C.J.M. Kraal, C. van den Bos, C. Voermans, G.A.M. Tytgat

Study supervision: H. van den Berg, H.N. Caron, C. Voermans, G.A.M. Tytgat

This work was supported by the Landsteiner Foundation for Blood Transfusion Research, LSBR grant F1101 (to I. Timmerman and C. Voermans) and KiKa grant 96 (to K.C.J.M. Kraal and G.A.M. Tytgat). The authors thank Lieke M.J. van Zogchel, Denise Stalder, and Naomi Weterings from Sanquin for technical support.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.