Patients and methods

Since 1998, 18 children have undergone 23 haplo-identical PBSCT, 14 boys and four girls, aged between 1 and 18 years (median 8.5 years). Indication for transplant was malignancy (n=14, ALL: CR1 n=3; CR2 n=5; CR3 n=1; ANLL/2e ALL n=1; MDS n=2; MDS/ANLL n=1; CML n=1) and nonmalignant conditions (n=4). All children with acute leukemia were, at the time of transplantation, in morphological remission. All donors were parents, in 12 instances the father (one father donated twice) and for 10 transplants the mother. In four cases both parents donated.

HLA typing was carried out using standard serological methods for HLA A and B loci and high-resolution typing for HLA DR loci. Peripheral blood stem cells were mobilized using granulocyte colony stimulating factor (G-CSF Neupogen® Amgen) at a dose of 1.0 IU/kg for a minimum of 5 days. Stem cells were harvested using continuous flow centrifugation at day 5. In addition, day 6 harvests were carried out in order to reach a total CD34 cell dose of 10 10⁶/kg until 2001, and thereafter 20 10⁶/kg recipient's body weight. CD34-positive cell selection was carried out using an automated CliniMACS® protocol according to standard published procedures (Figure 1). The total CD34+ cells infused ranged from 6.5 to 31.0 10⁶/kg (median 18.9).

Figure 1.

A picture of a FACS analysis of the stem cell product, before and after CliniMACS® magnetic bead separation. G-CSF mobilized, peripheral blood stem cell product is rich in CD34+ and CD3+ cells as demonstrated in the pre-selection analysis. Fatal GvHD would complicate infusion of such a product in the haplo-identical setting. CliniMACS® treatment effectively enriches the stem cell product by positive selection of CD34+ cells and virtually eliminates the CD3+ T cell population.

Full figure and legend (40K)

Patients were conditioned according to their underlying disease. Children with ALL received etoposide i.v. (350 mg/m²/day, on days -9 and -8), cyclophosphamide i.v. (60 mg/kg/day, on days -5 and -4), and total body irradiation either in one or two fractions depending upon age (7–12 Gy). Conditioning for ANLL was identical other than etoposide was omitted. Busulfan i.v. (3.2 mg/kg/day, from days -9 to -6) together with cyclophosphamide (60 mg/kg/day, on days -4 and -5) and melphalan 140 mg/m² on day –1 was administered to children with MDS. Busulfan i.v. (3.2 mg/kg /day, from days -10 to -7) and cyclophosphamide (50 mg/kg/day, from day -5 to day -2) was standard conditioning in children with nonmalignant conditions.

In patients with rejection or primary nonengraftment further attempt at transplant was undertaken using either OKT3 (from day -1 until day +18) 0.1 mg/kg (maximum dose 5 mg) or fludarabine (30 mg/day, day -5 to day -1) and campath IH (0.2 mg/day, day -5 to day -2).

Results

Sustained engraftment, defined as a persistent absolute neutrophils count >0.5 10⁹/l, was observed in 13 patients at a median of 17.0 days (range 13–26). One of these patients required G-CSF support from day +24 despite a documented 100% donor chimerism, possibly in light of a concurrent adenoviral infection. He later underwent a stem cell boost from the same donor at day +58 and is presently undergoing follow-up to assess his response. Three grafts failed and seven others, although showing initial hematological recovery, rapidly lost their donor cells. Cell doses administered in those with a sustained engraftment (n=13) ranged from 6.5 to 31.0 10⁶/kg (median 19.3), in cases with nonengraftment (n=3) 12.5–16.8 (median 15.2) and in children with early rejection (n=7) 9.3-21.6 10⁶/kg (median 16.7). There was no statistical difference (P=0.4) between the CD34+ cell dose infused and outcome (Figure 2).

Figure 2.

Scattergram of CD34+ cell dose infused in children undergoing haplo-identical PBSCT between 1998 and 2004 related to engraftment. The range of CD34+ cells infused varies considerably, and result from variations in donor mobilization and the weight of the recipient. However, there is no statistical difference between the number of CD34+ cells infused per kg body weight of the recipient and episodes of early rejection and/or engraftment potential (95% CI: -7.703–2.938).

Full figure and legend (14K)

In patients transplanted for malignancy (n=14), seven are presently alive albeit two of these are +2 months and it is too early to say whether they will remain disease free. Of the truly evaluable patients the overall disease-free survival is 42% (Figure 3) at a follow-up of 1.5–6 years. Three children who engrafted relapsed (21%). However, five children (35%) developed graft dysfunction of whom only one survived. Of the remaining four children, all but one had documented engraftment following further re-transplantation, but all died from infection.

Figure 3.

Kaplan Meyer survival of the first 12 children having undergone haplo-identical PBSCT for underlying hematological malignancies. The DFS is 41% with a follow-up of 1.5–6 years. This is comparable to published data, but the high levels of rejection compounded the success of the procedure.

Full figure and legend (12K)

Overall graft dysfunction occurred in 6/18 patients (33%). No acute or chronic GvHD was documented.

In long-term survivors, immune recovery was substantially delayed and no detectable T-cell recovery was evident until 5–6 months post transplant. Functional T and B cell recovery was not evident until considerably later. In contrast, signs of NK activity were detectable as early as 1-month post PBSCT. Reactivation of CMV was evident in two children post transplant and herpes zoster in one other child. A total of five patients developed adenovirus infection. In 4/5 this was following second attempt transplantation either with OKT3 or fludarabine/campath IH conditioning. One patient developed adenoviral reactivation (defined as a plasma DNA load >10³ copies/ml) shortly after engraftment. He remains under antiviral treatment with no evident dissemination. One child developed immune mediated thrombocytopenia, intracranial hemorrhage and PCP following steroids and emergency splenectomy, 3 months after successful engraftment for a secondary leukemia. After 4 years, he remains well with no clinically apparent neurological problems and in complete remission.

Discussion

Our results are comparable to the reported literature in relation to overall and disease-free survival, especially in pediatric patients with underlying hematological malignancies.^{9, 10} The overall survival rates are especially encouraging considering that these children would have certainly died from their initial disease without the transplant procedure. The lack of an otherwise suitable donor was no impediment to HSCT. Although it is generally reported that the problems of engraftment have been overcome in haplo-identical PBSCT, all centers undertaking this procedure have occasional patients with a dysfunctional graft.^{11, 12, 13} However, even though some failures may be explained by differences in immune suppression given pre-transplant and lower doses of CD34+ cells infused,^{14, 15} there is no doubt that not all patients receiving mega doses of CD34+ cells are guaranteed engraftment.¹⁶

Our relatively high levels of graft dysfunction cannot be explained by either low numbers of CD34+ cells infused nor insufficient pre-transplant immune suppression. Despite the introduction of mega doses of stem cells, we are neither able to predict nor overcome graft dysfunction based purely upon the final CD34+ cell count administered, as described by other authors.^{6, 17} We routinely administer antithymocyte globulin (ATG) to all patients pre-haplo-identical transplant as standard therapy. Children with hematological conditions undergo standard myelo-ablative conditioning, inclusive of total body irradiation, albeit as single fraction rather than fractionated radiotherapy.

As graft dysfunction has reduced the success of our haplo-identical transplants, we consider this to be a higher risk procedure than perhaps other centers with less lethal graft failure. Second attempt grafts in our unit have been complicated by lethal infections, mainly adenovirus, despite documented engraftment. It is hoped that with the recent introduction of PCR-monitoring and pre-emptive treatment this, as well as other viral infections, may be controlled. However, in light of the poor immune recovery, especially specific T-cell function, the control of infection may yet prove difficult.^{8, 11}

We observed a substantial number of viral infections in our patient population, but they did not differ substantially from other T-cell depleted HSCTs. Delayed recovery of adaptive immunity results from the degree of HLA disparity, the low numbers of T cells infused in the graft, and the use of ATG, which may inhibit T-cell expansion.¹² Pediatric studies show that the median time to normal T and B cell numbers and proliferative responses was 7–8 months and our patients have similar reconstitution profiles.⁸ Initial T-cell recovery is of oligoclonal memory T cells as a result of expansion of peripheral T cells from the graft. Later, at approximately 6 months, naïve T cells are recognizable presumably resulting from donor precursor cells emigrating from the thymus.¹² In the pediatric setting, immune recovery may be more rapid than in adults because of intact thymic function and because it is feasible to infuse higher doses of CD34+ cells that enhance the recovery of naïve T cells.⁸ The relatively low numbers of serious infections occurring in our patients with a sustained graft reflects this observation, together with the documented rapid neutrophil recovery.

Conclusion

Haplo-identical stem cell transplantation using highly enriched CD34+ cells is a feasible solution for those children requiring HSCT, who otherwise have no suitable donor. Technological advances have made transplantation across major HLA-antigen barriers possible without the induction of fatal GvHD. Delays in immune reconstitution, despite good myeloid engraftment, are significantly greater than in the HLA matched allogeneic transplant setting, but mirror immune-reconstitution in T-cell depleted bone marrow transplantations. Infections have been no worse than in those children undergoing mis-matched unrelated allogeneic transplantation.

In our experience, nonengraftment and rejection cannot always be prevented by mega-doses of CD34+ stem cells. Attempts at re-transplanting children with graft dysfunction have resulted in significant morbidity and mortality.

Apart from relapse, graft dysfunction and delayed immune reconstitution remain significant barriers to the application of this treatment to many more children, which would overcome the considerable time delays and expense of complicated donor searches. New strategies are presently being employed in our unit to combat these problems, namely the role of co-transplantation of expanded donor mesenchymal stem cells. These pluripotent stem cells, known to support hematopoiesis, may have an as yet undetermined role in improving the clinical outcome of children undergoing haplo-identical PBSCT.¹⁸

References

References

Acknowledgements

The medical, nursing and ancillary staff of the IHOBA, Willem-Alexander Kinder en Jeugd Centrum, Leiden University Medical Center, Leiden, the Netherlands. The staff of the Europdonor Foundation, Leiden, the Netherlands. JDJ Bakker-Steeneveld and CM Jol-van der Zijde, Data Managers, IHOBA, Willem-Alexander Kinder en Jeugd Centrum, Leiden University Medical Center, Leiden, the Netherlands. Professor Dr Dietrich Niethammer, Klinik für Kinderheilkunde und Jugendmedizin, Abteilung der Universität Tubingen, Tubingen, Germany. Professor Dr Thomas Klingebiel, Klinik für Kinderheilkunde III, Zentrum für Kinderheilkunde und Jugendmedizin der Universität Frankfurt, Frankfurt, Germany.