Journal Logo

Clinical and Translational Research

High Risk of Graft Failure in Patients With Anti-HLA Antibodies Undergoing Haploidentical Stem-Cell Transplantation

Ciurea, Stefan O.1; de Lima, Marcos1; Cano, Pedro2; Korbling, Martin1; Giralt, Sergio1; Shpall, Elizabeth J.1; Wang, Xuemei3; Thall, Peter F.3; Champlin, Richard E.1; Fernandez-Vina, Marcelo2,4

Author Information
doi: 10.1097/TP.0b013e3181b9d710
  • Free


Hematopoietic stem-cell transplantation (HSCT) from mismatched relatives represents an alternative treatment for patients with hematologic malignancies when no matched sibling or unrelated donor exists (1–3). Historically, its use has been limited because of the high rates of graft rejection and acute graft-versus-host disease (GVHD), the latter primarily because of the presence donor T cells in the graft (4, 5). Decreased rates of GVHD have been accomplished with T-cell depletion, but these transplants have been complicated by a high rate of graft rejection. Increased rates of engraftment have occurred with the use of “megadoses” of hematopoietic progenitor cells (2, 6). However, approximately 10 to 20% of patients still develop graft failure (7–9).

The degree of human leukocyte antigen (HLA) matching and the presence of anti-HLA antibodies have been associated with solid organ rejection and survival (10–16). Anasetti et al. (17) reported almost two decades ago that primary graft failure (PGF) in patients with unmodified marrow from a mismatched donor was higher (12.3%) compared with patients with a fully HLA-matched related donor (2%), and that the incidence of graft failure correlated with the degree of donor HLA incompatibility. A positive crossmatch for antidonor lymphocytotoxic antibody associated strongly with graft failure, and the risk of graft failure in alloimmunized patients with a positive donor crossmatch was significantly greater among nonalloimmunized patients (39 vs. 10%; 17). The association between a positive crossmatch and PGF in HSCT has been confirmed by other investigators (18). The use of novel humanized monoclonal antibodies in transplantation may interfere with the classical cross-match method by reacting with donor lymphocytes and may affect the sensitivity and specificity of the test for the presence of anti-HLA antibodies (19). Preformed antibodies present at the time of marrow infusion in multiply transfused mice, rather than primed T-cells, have been shown to be a major barrier against marrow engraftment (20). The antibody-mediated rejection was extremely rapid in primed mice and a high bone marrow cell dose could partly overcome the predisposition to graft failure in sensitized mice (20).

In recent years, novel solid-phase immunoassays that use purified preparations of molecules corresponding to a single HLA antigen allow accurate identification of HLA-antibody specificities have been developed (21). The information gained from the application of these sensitive technologies can be used to predict cross-match results in solid organ transplantation (16); however, the accuracy of these predictions may be suboptimal when HLA typing of the donor is incomplete (low resolution, loci not tested).

We proposed that with the advent of accurate HLA typing, the use of solid-phase immunoassays with fluorescent beads coated with single HLA antigens can be applied with high precision for the detection and characterization of donor-specific anti-HLA antibodies (DSA) in HSCT and hypothesized that prospective testing for the presence of DSA detected in this manner would be associated with a higher risk of PGF after T-cell depleted haploidentical HSCT.



We evaluated the incidence of graft failure in 24 consecutive patients with hematologic malignancies treated at the University of Texas M. D. Anderson Cancer Center with a T-cell depleted HSCT from a haploidentical donor tested for the presence of DSA from September 2005 to August 2008. All patients provided informed consent approved by the University of Texas M. D. Anderson Cancer Center Institutional Review Board according the Declaration of Helsinki. Six patients developed graft failure and five were identified to have DSA. Characteristics of these patients are presented in Table 1.

Characteristics and factors known to influence engraftment in 24 consecutive HLA haploidentical stem-cell recipients receiving their first transplant

Transplant Conditioning Regimen

All patients received a haploidentical graft from a mismatched-related donor (at least 2-antigen mismatch at HLA-A, -B, -C, and DR loci), and “megadoses” of CD34+ stem cells. Peripheral blood progenitor cells were mobilized from donors using filgrastim (Neupogen, Amgen, Thousand Oaks, CA) and collected by apheresis. CD34+ progenitor cells were positively selected using a CliniMACS system as a means of T-cell depletion of the donor graft (22). A minimum acceptable postselection CD34+ cell dose was 5.0×106 CD34+ cells/kg. The conditioning regimen included 140 mg/m2 of melphalan administered intravenously on day −8, 10 mg/m2 of thiotepa administered intravenously on day −7, 160 mg/m2 of fludarabine in divided doses on days −6, −5, −4, and −3, and 1.5 mg/kg per day of rabbit antithymocyte globulin (ATG) on days −6, −5, −4, and −3 (23). No posttransplant immunosuppression therapy and no growth factors were administered. All patients received antimicrobial prophylaxis with trimethoprim-sulfamethoxazole, gancyclovir or valacyclovir, and voriconazole or caspofungin (23).

Four patients who experienced graft failure received second T-cell replete haploidentical transplants, three from the same donor, and one with a different donor. Conditioning for the second transplants included alemtuzumab (Campath, Genzyme, Cambridge, MA) for three patients and fludarabine and ATG in the remaining patient. Tacrolimus was used for GVHD prophylaxis.

Measurement of Anti-HLA Antibody Levels

The presence of DSA was determined by testing the patients’ sera with a panel of fluorescent beads coated with single HLA antigen preparations (LABScreen Single-Antigen, One Lambda, Canoga Park, CA). The anti-HLA antibody reactivity was detected in a Luminex platform (Luminex, Austin, TX) (24). The reactivity of HLA antibodies was defined by testing two panels of HLA molecules, which included a set of single-antigen preparations for assignment of specificity and confirmed by comparison with the reactivity against a panel of HLA class I or II antigen preparations extracted from lymphocytes of single subjects. The final DSA was assessed by comparing the high-resolution types of the donor and antigen panels. Antibody levels were interpreted as normalized fluorescence intensity (FI) as defined by the kit’s manufacturer against DSA mismatch. FI less than 500 was considered negative, whereas positive FI ranging from 500 to 1500, 1500 to 3000, 3000 to 7000, and more than 7000 were classified as weak, intermediate, strong, and very strong, respectively. HLA-A, -B, -C, DRB1, DRB3/4/5, DQB1, and DPB1 high-resolution typing was accomplished by polymerase chain reaction amplification combined with nucleotide sequencing. Retrospective major histocompatibility complex class I-related chain A (MICA) typing of one patient and the corresponding donor was kindly performed by oligonucleotide hybridization of polymerase chain reaction products by Dr. Jar How Lee (One Lambda, Canoga Park, CA).

Treatment With Rituximab and Plasma Exchange

After the first patient was identified to have PGF in the presence of DSA, the subsequent four patients were treated with a combination of rituximab (Rituxan, Genentech, San Francisco, CA) and plasma exchange in an attempt to decrease DSA levels and prevent rejection. Two weekly doses of 375 mg/m2 of rituximab and two sessions of plasma exchange (1× total plasma volume each) starting 2 weeks before transplant were administered in such patients.

Statistical Considerations

Proportions were compared between groups using Fisher’s Exact test (25). The ability of patient covariates (DSA before transplant, CD34+ cell count, number of HLA mismatch sites, and disease status at transplant) to predict the probability of engraftment was evaluated by exact logistic regression (26). Covariates also were compared between the 20 patients without and four patients with graft failure using a Wilcoxon-Mann-Whitney test for quantitative variables or exact test for discrete variables (27). Exact computations were carried out using StatXact and LogXact (Cytel Corp.). All the reported P values are two sided.


Twenty-four patients were treated with this regimen. Five patients (21%), all females, were identified to have intermediate or strong DSA levels (Table 2). We identified DSA against antigens of HLA class I, class II, or both. The most common reactivity was detected against anti-HLA DR (Table 3).

Characteristics of five patients who received a haploidentical stem-cell transplant at UTMDACC and had donor-specific anti-HLA antibodies
Donor-specific HLA antibody levels and engraftment in five patients treated with haploidentical transplantation

Four (17%) of the 24 patients experienced PGF. Overall, three of four (75%) patients with DSA treated with haploidentical transplants developed graft failure, when compared with 1 of 20 (5%) patients who did not have DSA (P=0.008, Fisher’s Exact test). Similarly, initial DSA (before the treatment with rituximab and plasma exchange) and first graft failure were also associated (P=0.02, Fisher’s Exact test). After the first patient was identified to have graft failure in the presence of DSA, a combination of plasma exchange and rituximab was used to decrease the DSA levels and prevent graft failure. Four patients were treated in this fashion, two patients (nos. 3 and 5) had a significant decrease in the antibody levels and achieved engraftment; one of these patients became negative for DSA, whereas the other had persistence of only low levels of DSA (Table 3). The other two patients (nos. 2 and 4) maintained high levels of DSA and experienced PGF (Table 3).

All four patients with graft failure underwent second non-T-cell depleted haploidentical transplants, three of them from the same donor; three engrafted, two without DSA, and one in the presence of DSA. The remaining patient failed to engraft in the presence of persistent DSA. In one of the patients with graft failure, DSA FI decreased significantly within a few days after transplantation; this patient successfully engrafted after the second transplant (Table 3).

To exclude other potential causes for PGF, we analyzed the degree of HLA matching, the number of killer immunoglobulin-like receptors (KIR)-ligand mismatches in the rejection direction (host-versus-graft), number of progenitor cells infused, degree of ABO mismatching, and the disease status at transplant between the study and the control group (Table 1). No significant differences were identified for any of these parameters (Table 1). Because the number of CD34+ cells were actually higher in patients with PGF (mean=12.47, SD=4.64) than in patients without graft failure (mean=9.62, SD=4.07), the significantly higher graft failure rate in patients with DSA versus without DSA could not be explained by number of CD34+ cells. To assess the possibility that other characteristics may have been associated with graft failure, the probability of graft failure was modeled, in turn, as a function of degree of HLA mismatch (<3 vs. ≥3 antigen mismatch) and the disease status at the time of transplant using exact logistic regression. Neither the number of HLA mismatches (P=1.00) nor the remission status at transplant (P=1.00) were associated with the development of PGF.

The fact that all patients with DSA were females prompted a review of the pregnancy history in the 11 female patients in the analysis. The median number of pregnancies was 3 in the PGF group (n=5) when compared with 0 in the control group (n=6; P=0.10). A review of the transfusion history to assess possible alloimmunization, identified a median of 37 units of packed red blood cells transfused in the DSA group (n=5) when compared with 17 units the control group (n=17; P=0.10).


We have analyzed prospectively the relationship between the development of PGF and the presence of DSA in 24 consecutive patients receiving T-cell depleted haploidentical HSCT at a single institution. A high rate of graft failure was identified in patients with DSA present at the time of transplant. Three of four patients with DSA experienced PGF. Interestingly, DSA levels became undetectable in two patients shortly after transplant, indicating that they may have been absorbed during this procedure.

Serum/cell crossmatches were not consistently performed during this time. Notably, in the HSCT setting, some patients with hematologic malignancies undergo treatment with various agents such as ATG or humanized monoclonal antibodies (like rituximab or alemtuzumab), that may affect the viability of lymphocytes, and interfere with the cross-match testing. Therefore, the use of flowcytometric or complement dependent cytotoxicity assays in this setting may produce results that are difficult to interpret, as recently reviewed by Zachary and Leffell (19).

The most common DSA identified in this study involved reactivity against the DR molecules carrying a subunit encoded by the DRB1 in the patients who rejected their grafts. The deleterious effect of anti-DR (DRB1) antibodies on engraftment seemed to be as important as the effect of the anti HLA-class I antibodies. The DQ, DRB4, and DP loci encode for proteins that are found at lower density on the cell membrane (low expression). It is possible that the anti-HLA antibodies directed against antigens encoded by DQ, DRB4, and DP are less deleterious, as seen in case no. 5, in which persistence of high levels against DP was not associated with graft failure (Table 3), probably as the result of a lower target density.

The mechanism by which DSA may cause PGF is debatable. Both humoral and cellular immunity have been invoked as playing a role in graft rejection; however, humoral rejection, through preformed antibodies, seemed to be the primary mechanism responsible for graft failure in models of xenotrasplantation (28). Taylor et al. (20) showed that preformed antibodies in mice can mediate rapid rejection in allogeneic bone marrow transplantation. They also demonstrated that antibody neutralization by immunoglobulin administration may facilitate engraftment in allosensitized hosts (20). Removal of preformed antibodies through plasmapheresis and intravenous gammaglobulin has been used to prevent rejection in solid organ transplantation (29), whereas rituximab, in addition to the elimination of B-cells, can decrease de novo antibody production and facilitate graft tolerance in newly transplanted organs (30). Four of our patients were treated with a combination of rituximab and plasma exchange when high DSA levels were identified. Overall, this intervention decreased the DSA levels substantially in two patients, who both achieved engraftment. Conversely, DSA levels remained at higher levels in two patients who both failed to engraft.

No differences were identified between the number of mismatches in the high expression or low expression HLA loci, and the number of KIR-ligand mismatches in the host-versus-graft direction, suggesting that graft failure was not mediated by mismatches in the rejection direction. Moreover, patients in the graft failure group received a greater median numbers of CD34+ cells when compared with patients who engrafted, and rejection could not be explained by any other parameters that could have negatively influence engraftment in these patients. Antibodies against MICA have been recently identified as a cause for graft failure in kidney transplantation (31). Our patients were not tested for MICA antibodies as this study was initiated before such testing was available. In the single case, where there was a graft loss with no detectable DSA the patient and the donor had identical MICA types (008,012); therefore, MICA allorecognition cannot be invoked as causing PGF in this case. The cause for graft failure in this case is unknown. The role for alloreactive natural killer cells can be ruled out because the donor carried all known inhibitory KIR-ligand mismatches (data not shown), whereas the effect of alloreactive T-cells could not be excluded in this case. Future studies that investigate the role of antibodies against MICA are warranted not only in solid organ but also in HSCT.

All five patients with DSA were young women, four with at least two prior pregnancies and all heavily transfused before transplant, suggesting that alloimmunization occurred because of at least one of these two events.

We found that the assay using fluorescent beads coated with the single-antigen mismatch may be a more practical approach for evaluating risk of humoral causes of graft rejection in allogeneic HSCT. Ideally, the mismatched HLA alleles of the donor defined by high-resolution typing methods should be represented in the panel of fluorescent beads coated with single HLA antigen molecules based on the characterization of the specificity of the antibodies in epitopes born by different HLA alleles.

The DSA could also be important in other categories of HLA-mismatched HSCT including umbilical cord blood transplantation where the use of units presenting several HLA mismatches is a common practice. Because the levels of HLA expression may differ substantially in hematopoietic stem cells from adults and umbilical cord blood, studies of HLA antibody effects should be performed separately according to the cell source.

In conclusion, a high rate of graft rejection was observed among patients receiving a T-cell depleted haploidentical hematopoietic stem-cell transplant performed in the presence of DSA detected by modern techniques. This suggests that donors should be selected, if possible, based on a negative DSA screen. For cases with donor-specific antibodies, an attempt to decrease the antibody levels before transplant may be required for successful engraftment. Therapeutic strategies for such cases need to be assessed in prospective clinical trials.


We thank Dr. Jar How Lee (One Lambda Inc., Canoga, CA) for his kind help in performing typing of MICA alleles for some subjects included in this study.


1. Kernan NA, Bartsch G, Ash RC, et al. Analysis of 462 transplantations from unrelated donors facilitated by the National Marrow Donor Program. N Engl J Med 1993; 328: 593.
2. Aversa F, Tabilio A, Velardi A, et al. Treatment of high-risk acute leukemia with T-cell depleted stem cells from related donors with one fully mismatched HLA haplotype. N Eng J Med 1998; 339: 1186.
3. Copelan EA. Hematopoietic stem-cell transplantation. N Engl J Med 2006; 354: 1813.
4. Powles R. Long-term remissions in acute myelogenous leukaemia. Lancet 1984; 1: 800.
5. Beatty PG, Clift RA, Mikelson EM, et al. Marrow transplantation from related donors other then HLA-identical siblings. N Engl J Med 1995; 313: 765.
6. Bachar-Lustig E, Rachamim N, Li HW, et al. Megadose of T-cell depleted bone marrow overcomes MHC barriers in sublethally irradiated mice. Nat Med 1995; 1: 1268.
7. Lang P, Greil J, Bader P, et al. Long-term outcome after haploidentical stem cell transplantation in children. Blood Cells Mol Dis 2004; 33: 281.
8. Ciceri F, Labopin M, Aversa F, et al. A survey of fully-haploidentical hematopoietic stem cell transplantation in adults with high-risk acute leukemia: A risk factor analysis of outcomes for patients transplanted in remission. Blood 2008; 112: 3574.
9. Koh LP, Rizzieri DA, Chao NJ. Allogeneic hematopoietic stem cell transplant using mismatched/haploidentical donors. Biol Blood Marrow Transplant 2007; 13: 1249.
10. Patel R, Terasaki PI. Significance of positive crossmatch in kidney transplantation. N Engl J Med 1969; 280: 735.
11. Held PJ, Kahan BD, Hunsicker LG, et al. The impact of HLA mismatches on the survival of first cadaveric kidney transplants. N Engl J Med 1994; 331: 765.
12. Suciu-Foca N, Reed E, Marboe C, et al. The role of anti-HLA antibodies in heart transplantation. Transplantation 1991; 51: 716.
13. Terasaki PI, Ozawa MA. Predicting kidney graft failure by HLA antibodies a prospective trial. Am J Transplant 2004; 4: 438.
14. Mao Q, Terasaki PI, Cai J, et al. Extremely high association between appearance of HLA antibodies and failure of kidney grafts in a five-year longitudinal study. Am J Transplant 2007; 7: 864.
15. McKenna RM, Takemoto SK, Terasaki PI. Anti-HLA antibodies after solid organ transplantation. Transplantation 2000; 69: 319.
16. Bray RA, Nolen JD, Larsen C, et al. Transplanting the highly sensitized patient: The Emory algorithm. Am J Transplant 2006; 6: 2307.
17. Anasetti C, Amos D, Beatty PG, et al. Effect of HLA compatibility on engraftment of bone marrow transplants in patients with leukemia and lymphoma. N Engl J Med 1989; 320: 197.
18. Ottinger HD, Rebmann V, Pfeiffer KA, et al. Positive serum crossmatch as predictor for graft failure in HLA-mismatched allogeneic blood stem cell transplantation. Transplantation 2002; 73: 1280.
19. Zachary AA, Leffell MS. Detecting and monitoring human leukocyte antigen-specific antibodies. Hum Immunol 2008; 69: 591.
20. Taylor PA, Ehrhart MJ, Roforth MM, et al. Preformed antibody, not primed T cells, is the initial and major barrier to bone marrow engraftment in allosensitized recipients. Blood 2007; 109: 1307.
21. Pei R, Lee JH, Shih NJ, et al. Single human leukocyte antigen flow cytometry beads for accurate identification of HLA antibody specificities. Transplantation 2003; 75: 43.
22. Miltenyi S, Muller W, Weichel W, et al. High gradient magnetic cell separation with MACS. Cytometry 1990; 11: 231.
23. Ciurea SO, Qureshi S, Rondon G, et al. Sustained engraftment using fludarabine, melphalan and thiotepa conditioning for haploidentical stem cell transplantation. Blood 2007; 351B: 5081a.
24. Smith JD, Hamour IM, Banner NR, et al. C4d fixing, luminex binding antibodies a new tool for prediction of graft failure after heart transplantation. Am J Transplant 2007; 7: 2809.
25. Fisher RA. On the interpretation of χ2 from contingency tables, and the calculation of P. J R Stat Soc 1922; 85: 87.
26. Karim HF, Mehta CR, Patel NR. Computing distributions for exact logistic regression. J Am Stat Assoc 1987; 82: 1110.
27. Randles RH, Wolfe DA. Introduction to the theory of nonparametric statistics. New York, John Wiley 1979.
28. Xu H, Chilton PM, Tanner MK, et al. Humoral immunity is the dominant barrier for allogeneic bone marrow engraftment in sensitized recipients. Blood 2006; 108: 3611.
29. Montgomery RA, Zachary AA, Racusen LC, et al. Plasmapheresis and intravenous immune globulin provides effective rescue therapy for refractory humoral rejection and allows kidneys to be successfully transplanted into cross-match-positive recipients. Transplantation 2000; 70: 887.
30. Bearden CM, Agarwal A, Book BK, et al. Rituximab (Rituxan) inhibits the in vivo primary and secondary antibody response to a neoantigen, bacteriophage phiX174. Am J Transplant 2005; 5: 50.
31. Mizutani K, Terasaki P, Bignon JD, et al. Association of kidney transplant failure and antibodies against MICA. Hum Immunol 2006; 67: 683.

Donor-specific anti-HLA antibodies; Primary graft failure; Haploidentical stem-cell transplantation

© 2009 Lippincott Williams & Wilkins, Inc.