Blood and marrow stem cell transplant has improved the outcome for patients with high-risk hematologic malignancies. However, most patients do not have an appropriate HLA (immune type) matched sibling donor available and/or are unable to identify an acceptable unrelated HLA matched donor through the registries in a timely manner. Another option is haploidentical transplant using a partially matched family member donor.

Although haploidentical transplant has proven curative in many patients, this procedure has been hindered by significant complications, primarily regimen-related toxicity including graft versus host disease (GVHD) and infection due to delayed immune reconstitution. These can, in part, be due to certain white blood cells in the graft called T cells. GVHD happens when the donor T cells recognize the body tissues of the patient (the host) are different and attack these cells. Although too many T cells increase the possibility of GVHD, too few may cause the recipient's immune system to reconstitute slowly or the graft to fail to grow, leaving the patient at high-risk for significant infection.

This research project will investigate the use of particular pre-transplant conditioning regimen (chemotherapy, antibodies and total body irradiation) followed by a stem cell infusion from a "mismatched" family member donor. Once these stem cells are obtained they will be highly purified in an effort to remove T cells using the investigational CliniMACS stem cell selection device. The primary goal of this study will be to determine the rate of neutrophil and platelet engraftment, as well as the degree and rate of immune reconstitution in the first 100 days posttransplant for patients who receive this study treatment. Researchers will also study ways to decrease complications that may occur with a transplant from a genetically mismatched family donor.

Secondary outcome evaluations for this clinical study include the following:

• To estimate overall survival, disease free survival and event free survival for these patients
• To estimate the incidence of grade 2 to 4 acute GvHD in these patients
• To estimate the incidence of chronic GvHD and graft failure in these patients
• To estimate the incidence of non-hematologic peri-transplant regimen-related toxicity and regimen-related mortality in the first 100 days after transplantation in these patients
• To estimate the number of patients who require donor lymphocyte infusions and/or stem cell boosts after transplantation in this group of patients and evaluate its impact on immune reconstitution and engraftment
• To estimate the number of patients who develop evidence of EBV reactivation or post transplant lymphoproliferative disease (PTLPD) post transplant in this group of patients
• Describe the pharmacokinetics of rabbit anti-thymocyte globulin (rATG) in patients receiving allogeneic transplantation and determine the frequency of rATG antibody development
• Explore dose and patient characteristics as determinants of active rATG pharmacokinetic parameters

Originally this study began as a randomized comparison of two methods of stem cell selection utilizing the CliniMACS device. Both arms provided a purified product of CD34+ hematopoietic cells that was depleted of T-lymphocytes. One arm utilized a positive selection methodology using an anti-CD34 antibody and the other arm utilized negative selection with the anti-CD3 antibody OKT-3. There were no toxicities noted that would have required the study to be interrupted early, however, due to low accrual, it was decided to redesign the study. The new design focused on the standard arm utilizing negative selection, and all subsequent participants were treated in that manner. The patients who were treated on the positive selection arm are continuing to be monitored as specified in the original protocol, but will be reported in a descriptive manner only. The primary and secondary objectives of the current version of this study will be answered only by those patients receiving a haploidentical stem cell transplant utilizing a negative selection methodology.

Inclusion Criteria:

• Lacking a HLA-identical sibling or unrelated donor matched at 6 HLA loci formally requested within an approximate 90 day period from search initiation and who has a mismatched family member donor available

• At least 2 and less than or equal to 21 years of age

  • Must have one of the following diagnosis:
  • Acute lymphoid leukemia (ALL) in second, third, or subsequent remission.
  • ALL in first remission but high risk for relapse.
  • Acute myeloid leukemia (AML) in relapse or remission.
  • Secondary AML / MDS
  • Chronic myeloid leukemia (CML)
  • Juvenile myelomonocytic leukemia (JMML).
  • Myelodysplastic syndrome (MDS).
  • Paroxysmal nocturnal hemoglobinuria (PNH)
  • Non-Hodgkin lymphoma in second or subsequent CR

• Patients with a shortening fraction ≥ 25%

• Patients with a creatinine clearance ≥ 40cc/min/1.73m^2

• Patients with FVC ≥ 40% of predicted or pulse oximetry ≥ 92% on room air

• Patients with direct bilirubin ≤ 3 mg/dL or SGPT ≤ 500 U/L

• Patients with a Karnofsky or Lansky (age dependent) performance score of ≥ 70

• Mismatched family member donor is available, HIV negative and ≥ 18 years of age

Exclusion Criteria:

• Patients who have received a previous hematopoietic stem cell allograft
• Patients with a known allergy to rabbit or murine products
• Patients with isolated CNS, testicular or other isolated extramedullary site of relapse