Influence of the hematopoietic stem cell source on early immunohematologic reconstitution after allogeneic transplantation

The use of peripheral blood after granulocyte colony-stimulating factor (G-CSF) mobilization as a source of allogeneic hematopoietic stem cells (HSCs) is being increasingly considered.1-10 The high number of HSCs in such a graft might result in enhanced engraftment and accelerated hematopoietic recovery.6-9 Immune reconstitution and overall alloreactivity after peripheral blood HSC transplantation (PBHSCT) might also be significantly different from what is observed after bone marrow (BM) HSC transplantation (BMHSCT).6-12 Indeed, despite an increase in the number of donor T cells infused with a peripheral blood HSC (PBHSC) graft, the incidence of grade II or greater acute graft-versus-host disease (GvHD) appears to be similar or even less than after BMHSCT. However, a higher risk of chronic GvHD has been reported after allogeneic PBHSCT.7,11 13 

Several cases of severe, potentially lethal cases of acute intravascular hemolysis because of the production of donor-derived antibodies (Abs) directed against ABO antigens (Ags) present on recipient red blood cells (RBCs) and not on donor RBCs (ABO “minor” incompatibility) have been described after PBHSCT14-20(Table 1), thus suggesting an influence of such a type of graft on posttransplantation immunohematologic reconstitution.

A multicenter randomized phase III clinical trial comparing allogeneic BMHSCT to allogeneic PBHSCT has recently been conducted by the Société Française de Greffe de Moelle (SFGM).7 The phenotype and function of immune cells in the HSC grafts as well as the quality of immune reconstitution in patients 30 days after transplantation were evaluated in a cohort of consecutive patients enrolled in the trial.

We have taken advantage of this randomized study to prospectively compare the kinetics of immunohematologic reconstitution after PBHSCT versus BMHSCT. More specifically, and in view of the severe cases of hemolysis recently described, we postulated that the use of PBHSCT might be associated with an accelerated increase in donor-derived anti-RBC Abs.

Between June 1997 and February 1999, 122 patients were enrolled in a clinical multicenter SFGM phase III randomized study that compared allogeneic PBHSCT versus BMHSCT. The protocol was approved by an ethics committee (Comité Consultatif de Protection des Personnes dans la Recherche Biomédicale of Marseille 2) and was conducted with respect to the Helsinki accords for human subject research. All procedures were performed after written, informed consent was given by the donor and the recipient. In June 1998, a prospective immunobiological evaluation was initiated for patients subsequently entered in the study. This evaluation included white blood cell immunophenotyping and immunohematologic assessment of the donor (before G-CSF administration [for PBHSCT recipients] and at the time of HSC graft harvest [for both groups]) as well as of the recipient before and 30 days after transplantation. In addition, cell immunophenotyping was performed on the PBHSC or BM graft. Blood and graft samples were shipped by overnight mail to the Etablissement Français du Sang–Bourgogne/Franche-Comté and were immediately processed or cryopreserved (viable cells and serum). From June 1998 to February 1999, 49 (74%) of 66 consecutive randomized patients (PBHSCT, n = 21; BMHSCT, n = 28) were included in the immunohematologic study. Seventeen patients were not included because of missing (nonharvested) samples.

The clinical study design and results have been previously reported.5 The randomization was stratified by diagnosis and by center to minimize the variations resulting from different practices in terms of supportive care, GvHD prophylaxis, or transfusions policies. Patient and recipient characteristics are described in Table 2. In the PBHSCT arm, donors received 10 μg/kg/d of subcutaneous G-CSF (Filgrastim; Rhône-Poulenc-Rhorer, Montrouge, France) for 5 days. On the 5th day (day −1 of transplantation), the first HSC harvest was performed by apheresis. If the CD34⁺ cell count in the HSC bag was less than 4 × 10⁶/kg of recipient body weight, a second harvest was performed on the 6th day. G-CSF was administered on the 6th day if a third harvest was necessary at day +1. GvHD prophylaxis comprised cyclosporin A (CSA) (initiated at day −1) and methotrexate (MTX) (15mg/m² on day +1; 10 mg/m² on days +3 and +6). The CSA was started intravenously on day −1 at the dosage of 2 to 3 mg/kg/d and was switched to oral formulation as soon as oral intake was satisfactory. The dosage was adapted to whole blood or plasma level and renal function according to each center's practice. No recipient received G-CSF during the immunohematologic study period.

Citrate-anticoagulated peripheral blood and serum samples (Vacutainer, Becton Dickinson, Le Pont de Claix, France) were obtained from the BM donors before or at the time of BM harvest, from the PBHSC donors before and after G-CSF mobilization (ie, before the first apheresis), and from recipients before the conditioning regimen and at day +30 after transplantation. The use of citrate as blood anticoagulant was chosen for the sake of homogeneity between both types of harvests, as PBHSCs were collected in a bag containing citrate.

For anti-A and anti-B titration, 3 suspensions of A1Rh⁺, BRh⁺, and ORh⁻ reagent RBCs (human reagent red blood cells for the reverse determination of ABO grouping, Sanofi Diagnostics Pasteur, Marnes-la Coquette, France) were used. The reagent RBCs were suspended to a concentration of 40% in a buffered preservative solution. The direct antiglobulin test for each reagent RBC sample was negative. Each preparation of reagent RBCs was washed 3 times in 0.15 M saline solution (9% NaCl). The hematocrit of each suspension was then adjusted to 3% in 0.15 M saline solution.

The presence of anti-A or anti-B Abs was determined by a standard indirect antiglobulin test. Titration of the Abs was performed by serial 2-fold dilution. All samples from the same donor/recipient combination were studied in a blinded manner and were evaluated concomitantly by the same person. Three series of 10 tubes were prepared for each tested sample. In the first tube of each series, 50 μL of the tested serum was added. Then, serial dilutions were performed in 0.15 M saline solution to 1:512. Secondarily, 50 μL of 3% A1, B, or O reagent RBC suspension was added in the appropriate sera before a 45-minute incubation at 37°C. After 3 washes in 0.15 M saline solution, the supernatant was carefully removed, and 50 μL anti-immunoglobulin (Ig)G antihuman globulin (Sanofi Diagnostic Pasteur, Marnes-la Coquette, France) was added in each tube. Agglutination was assessed macroscopically and then microscopically after centrifugation (125g for 1 minute) and gentle resuspension. The agglutination of at least 50% ABO reagent RBCs was required to be considered positive. The titers of anti-A and/or B Abs corresponded to the reverse of the last positive dilution. An increase of the anti-A and/or anti-B Ab titers between the day preceding the conditioning regimen and day +30 was defined prior to initiating the study as an increase of 2 or more dilutions.

The low ionic strength solution indirect antiglobulin method and enzymatic technique (DiaMed-ID Micro Typing System, DiaMed AG, Cressier, Switzerland) were used to detect anti-RBC Ag allo-Abs. Detection of Abs was performed on a panel of 3 reagent RBCs (DiaMed-ID Micro Typing System, test cell reagents for antibody screening ID-DiaCell, and ID-DiaCell P). In addition, in case of positivity, identification was performed on a panel of 10 RBCs (reagent red cells; Sanofi Diagnostics Pasteur, Marnes-la Coquette, France). These techniques were performed as described by the manufacturer.

HSC donor– and recipient–ABO compatible RBC concentrates (RBC-Cs) were transfused when the hemoglobin level was below 8 g/dL. Platelet concentrates (PCs) were administered to treat or prevent hemorrhage when the blood platelet count was below 20 × 10⁹/L. If available, PC with the same ABO group as the HSC donor was administered. If unavailable, PC with an ABO Ag compatible with the HSC donor was preferred over PC ABO Ag incompatible with the HSC donor. Platelet products were never washed or plasma-reduced. These rules were applied on initiation of the conditioning regimen. Donor/recipient blood group as well as the number, time of administration, and blood group of transfused PC and RBC-C were recorded for all recipients. Similarly, administration of intravenous polyvalent immunoglobulins (IVIGs) was documented. In addition, the delay between the last PC transfusion containing anti-A and/or anti-B Abs or the last administration of IVIGs and the day +30 immunohematologic testing was calculated.

Continuous variables were compared between the 2 groups by using the Wilcoxon rank-sum test. Qualitative variables were analyzed with a chi-square test, or exact test when expected frequencies were lower than 5. The potential confounding effects of covariables on the relation between source of HSC and increased anti-A and/or anti-B titers in the recipient at day +30 were studied one by one by bivariate analysis (Mantel-Haenszel or exact methods for qualitative variables and exact logistic regression for quantitative variables). Breslow test for homogeneity was used to test whether the type of ABO mismatch was an interaction term in the relation between source of HSC and increased titer of anti-A and/or B Abs.22 

The use of a PBHSC graft resulted in a higher frequency of increased anti-A and/or anti-B Ab titers 30 days after transplantation as compared to a BMHSC graft: 8 (38%) of 21 versus 3 (11%) of 28 (P = .04) (Table 3).

In addition, 3 (14%) of 21 PBHSCT versus 0 of 28 BMHSCT recipients developed an anti-RBC (other than ABO) Ab, detected at day +30 and absent before the conditioning regimen (P = .07) (Table3).

With one exception, anti-A and/or anti-B Abs with increased titers at day +30 after PBHSCT or BMHSCT was always directed against A and/or B Ags absent both on donor and recipient RBCs. One patient, after an ABO minor and major incompatible (A into B) BMHSCT, had increased antidonor (anti-A) Ab titers at day +30. Interestingly, this patient had received anti-A–containing PC 12 hours prior to blood sample harvesting, suggesting passive transmission. In all other patients, the last PC transfusion was performed at least 48 hours before sample harvesting.

To verify that the absence of measurable Ab titers against donor and/or recipient was not due to in vitro absorption on the RBCs present in the tube during sample transportation, we compared Ab titers in the serum of posttransplantation blood samples processed (serum separation) immediately or after a 24-hour delay as in our study. Ab titers were identical in both cases (data not shown), thus eliminating any in vitro artifact.

Comparison of anti-A or anti-B titers in the donor before and after G-CSF treatment was performed in 19 of 21 donors. There were no significant changes in Ab titers for 17 of 19 donors. One donor (blood group O female with no history of pregnancy) had decreased anti-A Ab titers (256 to 64) and one donor (blood group O female with a history of 4 pregnancies) had increased anti-B Ab titers (32 to 128). In all donors, no anti-allo RBC (other than ABO) Abs were found either before or after G-CSF treatment.

Both BMHSCT and PBHSCT donor/recipient groups were compared for a number of parameters that might have influenced the characteristics of immunohematologic reconstitution after transplantation. As detailed in Tables 2 and 4, age, previous pregnancies, ABO mismatch between HSC donor and recipient, diagnosis, conditioning regimen, donor anti-A and/or anti-B Ab titers, episodes of RBC-C transfusion, passive transfusion of anti-A and/or anti-B Ab with PC, ABO Ag compatibility of transfused PC, administration of IVIG, and delay between the last transfusion of an anti-A– and/or anti-B–containing blood product and immunohematologic assessment were parameters that did not differ significantly between both groups. In contrast, the number of episodes of transfused PC was significantly lower in the PBHSCT group than in the BMHSCT group (3 [1-18] versus 6 [3-33]; P = .03) (Table 4). After adjustment for each of these potential confounding variables in bivariate analyses, the use of a PBHSC graft remained significantly associated with increased anti-A and/or anti-B Ab titers 30 days after transplantation.

In PBHSCT recipients, increased titers were observed mostly in the context of a minor ABO mismatch only (O into A or B or AB and A or B into AB) transplant: 5 of 7 versus 3 of 14 in the absence of any minor ABO mismatch (P = .05) (Table 3). This was not the case after BMHSCT: 1 of 8 in the presence of a minor ABO mismatch versus 2 of 20. The interaction test between type of transplant and type of compatibility on the occurrence of increased Ab titers after transplantation approached significance (P = .10).

The presence of high (> 32) anti-A and/or anti-B serum Ab titers in the donor, before as well as after G-CSF administration, tended to be associated with increased anti-A and/or anti-B titers in the recipient at day +30 after transplantation (P = .09).

Donor sex and previous pregnancies did not differ between recipients who had increased or similar Ab titers after PBHSCT (data not shown).

The use of a PBHSC allogeneic graft is being increasingly considered, and several randomized studies comparing a PBHSC graft to a BMHSC graft have been performed.5-10 In addition to the absence of general anesthesia for the donor, accelerated hematopoietic reconstitution as well as reduced transfusion requirements6-9 could be associated with the use of such a source of hematopoietic stem cells. However, several cases of severe, potentially lethal, cases of immune-mediated acute hemolysis have been described after allogeneic PBHSCT14-20 (Table 1). These hemolysis episodes occurred 8 to 14 days after transplantation and are associated with the recognition of residual recipient RBCs by donor-derived Abs. A similar case of severe acute hemolysis has recently been described after nonmyeloablative conditioning regimen and PBHSCT (J. Barrett, personal communication, December 7, 1999).

ABO incompatibility between donor and recipient is not considered as a major obstacle to allogeneic HSC transplantation. ABO incompatibility is said to be major when the recipient has Abs recognizing ABO antigens present on donor RBCs.23,24 In this situation, acute hemolysis, which may occur at the time of HSC graft infusion, may be prevented by using an RBC-depleted graft. In the case of a recipient with high anti-A and/or anti-B Ab titers, a possible later complication is a delayed onset of erythropoiesis after transplantation, associated with an increased RBC-C transfusion requirement after transplantation. ABO incompatibility is termed “minor” when donor-derived anti-ABO Abs can recognize one or several ABO Ags present on recipient erythrocytes. An immune-mediated hemolysis may occur either at the time of allograft transfusion by passive transfer of Abs (this complication is easily prevented by washing the HSC graft), or subsequently, most often 1 to 3 weeks after transplantation, concomitantly to the initiation of donor-derived immune reconstitution. In the latter case, hemolysis is in relation with Ab production by donor-derived lymphocytes. Several risk factors have been associated with the occurrence of delayed hemolysis: T-cell depletion,25 use of FK 506,26 absence of MTX to prevent GvHD,27,28 or use of BMHSCT from an unrelated donor.27 28 

In our study, PBHSCT recipients increased anti-A and/or anti-B Ab titers at day +30 after transplantation more frequently than did BMHSCT recipients (P = .04). Importantly, this result persisted after adjustment for each potential confounding factor, such as age, ABO mismatch, transfusion practices, IVIG administration, and donor anti-A and/or anti-B titers, known to possibly affect immunohematologic reconstitution after transplantation.

In the PBHSCT group, the increase in anti-A and/or anti-B titers after transplantation was associated with minor ABO-incompatible transplantation: 5 (71%) of 7 versus 3 (21%) of 14 in the absence of minor ABO incompatibility (P = .05), which was not the case after BMT: 1 (13%) of 8 versus 2 (10%) of 20. The increased anti-A and/or B serum Ab production after PBHSCT could significantly contribute to enhanced susceptibility to hemolysis, particularly in the case of minor ABO donor/recipient incompatibility. Indeed, as previously mentioned, all reported cases of severe hemolysis after PBHSCT involved a minor ABO mismatch transplantation. In this situation, an accelerated production of antirecipient RBC Abs by donor-derived-B lymphocytes could result in acute hemolysis of the residual recipient RBCs or transfused donor-incompatible RBC-C.

Anti-A and/or anti-B Abs with increased titers at day +30 were almost always directed against A and/or B Ags absent on both donor and recipient RBCs. On the contrary, Abs directed against recipient A and/or B Ags remained undetectable at day +30. Several mechanisms might account for the absence of such Abs. We have eliminated the possibility that in vitro absorption during the 24-hour transport-related conservation of whole blood played a significant role. Low-titer antirecipient RBC Abs produced in vivo might be absorbed in vivo on the RBC Ags present on residual recipient (or transfused) RBCs as well as on the A and/or B Ags present on extra-hematopoietic cells such as endothelial cells.29 In such cases, measurable antirecipient RBC Ab serum titers would be found only when Ab titers increase significantly as in the observed cases of hemolysis (Table 1) or much later on completion of immune reconstitution and disappearance of circulating recipient RBCs.

In addition to minor ABO-incompatible PBHSCT, transfusion of donor-incompatible RBCs as well as the absence of MTX (in the presence of CSA) have been associated with the occurrence of acute hemolysis in most cases. In our study, all patients received MTX, and guidelines, requiring that RBC-C transfusion be both donor and recipient compatible or at least donor compatible for the Rh and Kell Ags, were enforced.30 Indeed, significant hemolysis was observed in none of the recipients of the SFGM trial.

The absence of MTX (in the presence of CSA) has been previously reported as a risk factor for delayed hemolysis.27,28 MTX is cytotoxic for B lymphocytes.31 The immunosuppressive effects of CSA are T-cell specific,32,33 and the use of such a drug, in the absence of MTX, could result in enhanced B-cell proliferation and T-cell–independent Ab response such as anti-RBC IgM isotype Ab production. In addition, because of the mechanisms by which CSA exerts its immunosuppressive action (inhibition of nuclear factor of activated T cells [NFAT] translocation to the nucleus),34 such a drug should be unable to control ongoing T-cell–dependent immune response. This hypothesis implies that transplantation from a donor with high anti-recipient RBC Ab titers might be associated with a higher risk of posttransplantation hemolysis. Interestingly, anti-A and/or anti-B Ab titers after G-CSF administration in the HSC donors of recipients who increased anti-A and/or anti-B Ab titers at day +30 tended to be higher. Such a finding needs be confirmed on a larger number of patients. Nevertheless, the presence of high Ab titers in a putative minor ABO-mismatched HSC donor is a parameter that could contribute to the choice of the HSC graft source as well as the type of posttransplantation immunosuppression.

Several quantitative and qualitative differences in a PBHSC graft versus a marrow graft could contribute to accelerating immunohematologic reconstitution after PBHSCT. The increased number of B cells, T cells, monocytes, and CD34⁺ cells present in the PBHSC grafts as compared to the BM grafts35 could result in accelerated immune-reconstitution with the production of anti-recipient Abs early after transplantation at a time when significant numbers of recipient RBCs are still present. Indeed, we have determined that circulating T cells (CD4⁺, CD8⁺ as well as CD3⁺, CD4⁻, and CD8⁻) and B cells are present in higher numbers 30 days after PBHSCT than after BMHSCT.36 In addition, G-CSF has been reported to directly enhance in vitro immunoglobulin production by activated B cells.37 G-CSF–induced TH2 cytokine profile of the T cells present in the graft could also contribute to enhancing post-PBHSCT Ab responses.38,39 In this respect, analysis of anti-HLA Ab immunization in our cohort of recipients has demonstrated that a PBHSCT is associated with a significantly increased incidence of de novo anti-HLA immunization.40 The frequency of interferon (IFN)-gamma–producing T cells as well as the capacity of producing IFN-gamma at single cell level are indeed reduced in a PBHSC graft versus a BM graft.35 Reduced tumor necrosis factor-alpha and interleukin (IL)-12 production41and increased IL-10 production42 have been attributed to G-CSF exposure and could also contribute to the observed increase in Ab titers after PBHSCT.

ABO compatibility in the context of marrow transplantation has not been conclusively associated with overall outcome.43However, minor ABO-compatible BMT might be associated with an increased risk of GvHD.44 Indeed, the immune activation following the binding of anti-AB Abs to their target Ags (on residual host or transfused RBCs), or extra-hematopoietic cells such as endothelial cells might initiate an alloreactive immune response.45 In view of the accelerated immunohematologic reconstitution we have evidenced after PBHSCT, the issue of ABO incompatibility and posttransplantation alloreactivity clearly needs now to be carefully examined in the setting of PBHSCT. In our study, minor ABO-incompatible PBHSCT was not significantly associated with an increased incidence or severity of acute or chronic GvHD (data not shown).

In conclusion, we observed, in a randomized setting, that the use of G-CSF–mobilized PBSHC was associated with accelerated erythrocyte immunohematologic reconstitution. Larger studies with a longer follow-up will be necessary to confirm this finding. Such an early increase in donor-derived anti-RBC Abs at a time when significant numbers of recipient RBCs are still present can result in severe hemolysis after transplantation. Stringent immunohematologic assessment and transfusion practices should reduce such a risk.30Increased numbers of B and T lymphocytes in the graft, G-CSF-induced functional alterations, and accelerated cellular immune reconstitution could contribute to such findings. Further studies should clarify the mechanisms involved. Overall, different HSC sources are associated with distinct immune reconstitution patterns, including immunohematologic reconstitution.

We wish to acknowledge the contribution of the following BMT centers: Angers: N. Ifrah, S. François, N. Piard; Besançon: J. Y. Cahn; Bordeaux: J. M. Boiron, B. Daze; Grenoble: F. Garban, F. Chenais; Hôtel-Dieu (Paris): B. Rio, M. F. Fruchart; Lille: J. P. Jouet, P. Renom; Lyon: M. Michallet; Nancy: F. Witz; Nantes: N. Milpied, B. David; PitiéSalpétrière (Paris): L. Sutton, A. Verdier; Saint-Etienne: C. Oriol; Toulouse: M. Attal, C. Payen, F. Roubinet; Institut Gustave Roussy (Villejuif): J. H. Bourhis, J. L. Pico; Poitiers: A. Sadoun, C. Giraud; Robert Debré (Paris): M. Duval, F. Sellami and the help of F. Bassompierre (Direction Régionale de la Recherche Clinique, Paris).