SECOND ANNUAL INTERNATIONAL UMBILICAL CORD BLOOD SYMPOSIUM

Biology of Blood and Marrow Transplantation 10:728-739 (2004) 2004 American Society for Blood and Marrow Transplantation 1083-8791/04/1010-0009$30.00/0 doi:10.1016/j.bbmt.2004.06.010 SECOND ANNUAL INTERNATIONAL UMBILICAL CORD BLOOD SYMPOSIUM LOS ANGELES, CALIFORNIA, MAY 14-15, 2004 Symposium Summary The Symposium faculty consisted of 23 speakers from the United States, Italy, France, Spain, Germany, and Japan. The program was divided into 7 sessions: (I) unrelated donor cord blood transplantation in adults, (II) practical aspects of cord blood transplantation, (III) optimal supportive care in cord blood transplantation, (IV) cord blood transplantation for childhood/genetic disorders, (V) immune status after cord blood transplantation, (VI) ex vivo expansion and regenerative medicine, and (VII) contemporary issues in cord blood banking. Program details are listed at http://cordbloodsymposium.org/home2004.html. The following comments emphasize significant aspects of selected presentations. Further details are provided in the abstracts that follow. The publication of the abstracts and the summary of the 2004 International Umbilical Cord Blood Transplantation Symposium is supported by an educational grant from StemCyte International Cord Blood Center. Session I reviewed cord blood transplantation (CBT) in adults. A large majority of patients in need of a hematopoietic cell transplant are adults, and a significant body of data is accumulating regarding the effectiveness of cord blood for this population. This is a most critical issue regarding the ability of cord blood units to fulfill society s need for a source of stem cell products for all patients requiring a transplant. Dr. Pablo Rubinstein reviewed data from the National Cord Blood Program of the New York Blood Center on transplantation of 391 (92.9%) of 421 adults who underwent transplantation with cord blood units since 1995. Thus, these data include transplantations performed early in the history of CBT. More than 90% had leukemia (41% advanced), lymphoma, or myelodysplasia. Factors contributing to a favorable outcome were a cell dose of 2.5 10 7 /kg, total body irradiation (TBI) conditioning, and HLA match. Overall survival was 30.2% at 1 year and 23% at 5 years after transplantation. Patients with good HLA matches (6/6 or 5/6 without rejection mismatches) and those conditioned with TBI and fludarabine (FLU) had 68% survival at 1 year. He concluded that remarkable improvement in the overall survival rate of adults who undergo transplantation with cord blood is possible by improvements in HLA matching and conditioning regimens. Dr. John Wagner discussed novel approaches to improving engraftment of cord blood units. This included approaches such as double unit transplantation, ex vivo expansion, and co-infusion of mesenchymal stem cells. He presented data on 23 adult and adolescent patients (median age, 24 years; range, 13-53 years) with high-risk hematologic malignancy who underwent transplantation with 2 partially HLAmatched umbilical cord blood (UCB) units after myeloablative conditioning. The median total infused cell dose was 3.5 10 7 nucleated cells (NCs) per kilogram (range, 1.1-6.3). All evaluable patients (n 21) engrafted at a median of 23 days (range, 15-41 days), with 1 unit predominating. Although there was no association between NC dose, CD34 cell dose, or HLA-A, -B, or -DRB1 match and which unit predominated, the winning unit had a higher CD3 dose. The incidence of severe grade III to IV acute graft-versushost disease (GVHD) was 13% (95% confidence interval [CI], 0%-26%), and disease-free survival (DFS) was 57% (95% CI, 35%-79%) at 1 year. These suggest that (1) double unit UCBT is safe, with 1 unit predominating over time; (2) the unit that disappears over time may facilitate engraftment of the winning unit, possibly by immune-mediated mechanisms; and (3) CD3 dose might predict which unit will ultimately engraft long-term. Most importantly, the use of 2 UCB units extends the application of this treatment to nearly all adults for whom 1 unit would have been insufficient. Dr. Juliet Barker discussed UCB transplantation (UCBT) after nonmyeloablative conditioning in highrisk adults. Fifty-one adults with advanced hematologic malignancies (median age, 50 years; range, 19-60), ineligible for myeloablative conditioning by virtue of advanced age, extensive prior therapy, or serious comorbidities, received UCBT after cyclo- 728

phosphamide 50 mg/kg, FLU 200 mg/m 2, and 200 cgy TBI. Immunosuppression was with cyclosporin A to at least day 100 and mycophenolate mofetil to day 30. Eight patients with no combination chemotherapy in the 6 months before transplantation also received antithymocyte globulin during conditioning. Thirteen patients (25%) received single UCB units, and 38 (75%) received double unit grafts. The median total infused cell dose was 3.4 10 7 NCs per kilogram (range, 1.1-5.7). Units were predominantly 1- or 2-antigen mismatched with the recipient. Fifty patients were evaluable, and neutrophil recovery to 0.5 10 9 /L occurred at a median of 8 days (range, 5-32 days) in 45. Four patients had failure of donor engraftment with autologous recovery, and 1 had late graft rejection. The cumulative incidence of sustained donor engraftment was 90% (95% CI, 82%-98%), with a median chimerism of 100% (range, 72%-100%) by day 100. The cumulative incidences of grade II to IV and grade III to IV acute GVHD were 61% (95% CI, 14%-78%) and 27% (95% CI, 14%-40%), respectively, at day 100, and that of chronic GVHD was 36% (95% CI, 20%-52%) at 1 year. The incidence of transplant-related mortality (TRM) was 18% (95% CI, 8%-28%) at day 180, and that of relapse/progressive disease was 31% (95% CI, 17%-45%) at 1 year. Regression of relapsed or persistent disease has been seen in patients with myelodysplasia and intermediate- and low-grade lymphoid malignancies (n 8). The probability of overall and progression-free survival was 57% (95% CI, 40%-73%) and 48% (95% CI, 32%-63%), respectively, at 1 year. Notably, day 180 TRM in patients aged 45 years was 11% (95% CI, 1%-21%), and in patients with extensive prior therapy it was 25% (95% CI, 6%-44%). However, patients with serious pretransplantation comorbidities had a high TRM of 45% (95% CI, 20%-70%) at day 180 (P.01). These data show that this regimen is both nonmyeloablative and associated with a high incidence of donor engraftment. Further, this approach extends access to transplantation to many adults who would otherwise be ineligible on the basis of a lack of donor or an inability to tolerate high-dose conditioning. Dr. Satoshi Takahashi discussed clinical results in UCB recipients with hematologic malignancies with myeloablative regimens. Sixty-eight adult patients with hematologic malignancies received unrelated CBTs, and results were compared with bone marrow transplantation (BMT; n 45) from unrelated donors. Some critical results were that neutrophil and platelet recoveries were slow in CBT compared with BMT and that the probability of grade III and IV acute GVHD was significantly lower than among BMT recipients. There were no GVHD-related deaths among CBT recipients, compared with 10 deaths out of 24 among BMT recipients, but there was BB&MT no significant difference between CBT and BMT recipients in chronic GVHD incidence. Unrelated CBT showed better TRM, relapse, and DFS results compared with BMT. The conclusion was that the data strongly suggest that CBT could be safely and effectively used for adult patients with hematologic malignancies. Dr. Eliane Gluckman reported a comparison of unrelated cord blood and BMTs in adults with acute leukemia. Ninety-eight adults with acute leukemia given an HLA-incompatible unrelated cord blood transplant (UCBT) were compared with 584 patients who had HLA-matched unrelated BMTs performed between 1998 to 2002 and reported to Eurocord and European Blood and Marrow Transplant Registry. Recipients of UCBT were younger (median, 24.5 versus 32 years; P.001), weighed less (median, 58 versus 68 kg; P.001), and had more advanced disease at transplantation (52% versus 33%; P.001). Multivariate analysis demonstrated lower risks of grade II to IV acute GVHD (P.001) after UCBT; however, neutrophil recovery was significantly delayed (P.001). TRM, relapse, chronic GVHD, and leukemia-free survival were comparable between UCBT and unrelated BMT recipients. Dr. Guillermo F. Sanz described the results of unrelated-donor CBT in 20 adult patients with chronic myeloid leukemia. Fourteen were in chronic phase (CP; 8 in CP1 and 6 in CP2), 2 in accelerated phase, and 4 in blast crisis. With a median follow-up of 49 months (range, 28-82 months), the probability of DFS at 3 years was 40% and was related to age (P.01) and the phase of the disease at transplantation (P.04). The probability of DFS at 3 years was 60% for patients younger than 31 years and 57% for those who underwent transplantation in CP. Six of 8 patients aged 30 years or younger who underwent transplantation in CP remain alive and disease free. The conclusion reached was that unrelated-donor CBT is an acceptable alternative for younger patients with chronic myeloid leukemia in CP who require transplantation and lack an HLA-matched family or unrelated bone marrow donor. Dr. Mary Laughlin reported that UCBs are increasingly used in older subjects. She reported a comparison of the results of 169 1-antigen mismatched UCBTs and 83 1-antigen mismatched and 367 matched BMTs in patients aged 16 to 60 years with acute and chronic leukemia. Rates of treatment-related mortality, treatment failure, and overall mortality were similar after mismatched BMTs and UCBTs. Although outcomes were better after matched transplantations, in a cohort in which 69% of the patients who had UCBTs were mismatched at 2 loci and had advanced leukemia at transplantation, outcomes were comparable to those observed after 1-antigen-mismatched BMT. The con- 729

clusion reached was that use of UCB units mismatched at 2 loci will increase donor availability for older subjects. Session II reviewed the factors important in the selection of the optimal CBU for transplantation. Dr. Effie Petersdorf described a new model for optimal HLA matching of unrelated donors in which quantitative and qualitative measures of donor/recipient disparity are considered and in which stage of disease at the time of transplantation may influence the effect of HLA disparity. Routine testing of prospective patients and their best-matched donors should include high-resolution DNA typing for HLA-A, -B, -C, -DRB1, and -DQB1. If a fully matched donor is not available, then mismatching should be limited to 1 or 2 loci, with preference to allele-level rather than antigen-level mismatches. Dr. Lee Ann Baxter-Lowe pointed out that for the cord blood units currently searched by the National Marrow Donor Program, at least one 4/6 HLAmatched cord blood unit with at least 80 10 7 TNC is available for nearly all patients. Although HLA matching improves outcomes, CBTs can be successful with only 3/6 or 4/6 HLA matches. For BMT, HLA disparity detected by performing high-resolution HLA typing and typing of additional HLA loci has been associated with adverse outcomes, and it is hypothesized that these observations are likely to be applicable to CBT. This hypothesis is being tested by the National Institutes of Health COBLT study, which has reported an interim analysis of high-resolution typing of 188 donor/recipient pairs, the clinical significance of which is under investigation. Until the relationship between HLA matching and transplantation outcomes is well understood, high-resolution typing might be considered for selection of cord blood units for transplantation. Dr. Juliet Barker discussed issues in the selection and availability of UCB versus volunteer donor bone marrow. She highlighted that although major efforts have been made to increase minority representation and speed search times of volunteer donors, many patients in need of urgent transplantation are not able to identify a suitable donor in a timely fashion. In addition, efforts to match donor/recipient pairs at the HLA-A, -B, -C, and -DRB1 alleles may further limit the donor pool and potentially prolong searches. In contrast, UCB has the major advantage of tolerance of limited HLA disparity and rapid availability. At the University of Minnesota, the median time to obtain a UCB graft was 13.5 days, as compared with 50 days for a volunteer bone marrow donor. Clearly, low cell dose is a major limitation of UCBT for adults and larger adolescents. However, strategies such as double unit transplantation and ex vivo expansion may be able to address this problem. For example, of 122 consecutive adult transplant referrals who had a formal UCB search, by using UCB unit-selection criteria of 4/6 to 6/6 HLA-A, -B, and -DRB1 match and combining 2 units for patients with no single unit 3.5 10 7 NCs per kilogram, a satisfactory graft was identified for all patients. A further issue is that UCB searches are not centralized and that banking standards can vary, which can sometimes give rise to concerns about unit quality. Overall, because there have been no randomized trials of volunteer donor BMT versus UCBT, how these 2 stem cell sources compare is not known. Currently, the choice of stem cell source is based on diagnosis, degree of urgency, relative availability of adequately matched bone marrow versus UCB of adequate match and dose, institutional research priority, and physician preference. Session III consisted of a panel discussion regarding the optimal supportive care in CBT. Dr. Richard Champlin was the moderator, and topics discussed included the optimal preparative regimen, requirements for engraftment, the role of nonmyeloablative regimens, the role of antithymocyte globulin, optimal GVHD prophylaxis, growth factors, infection prophylaxis, long-term follow-up, and patient selection. Session IV reviewed CBT for childhood/genetic disorders. Dr. Joanne Kurtzberg indicated that unrelated donor UCB has been used for transplantation in 100 infants and children with inborn errors of metabolism in her medical center since 1995. Cord blood cells exhibit immunologic tolerance, allowing transplantation across partially mismatched HLA barriers and increasing access to allogeneic transplantation therapy for patients in need, and may also be capable of transdifferentiation into nonhematopoietic tissues. All children were prepared for transplantation with myeloablative chemotherapy consisting of busulfan, cyclophosphamide, and antithymocyte globulin. Prophylaxis against GVHD was administered with cyclosporine and methylprednisolone. Thirty-five young children with Hurler syndrome, all with the severe phenotype (MPS I), underwent transplantation with partially HLA-mismatched unrelated donor UCB over the past 8 years. Moderate to severe acute GVHD occurred in 28% of patients. Extensive chronic GVHD was not seen. Twelve percent of patients had serious events: graft rejection (n 1), infectious deaths (n 3), or toxic death (n 1; hyperammonemia). All other patients (87%) are surviving event free for a median 3 years. Additional children (n 60) with lysosomal storage diseases including metachromatic leukodystrophy, adrenoleukodystrophy, and globoid leukodystrophy (Krabbe disease) have undergone transplantation with unrelated donor UCB over the past 9 years. In asymptomatic children, disease was arrested 730

before the onset of neurologic dysfunction. In symptomatic children, disease progression was arrested within 6 to 9 months of the transplantation. In a child with advanced Krabbe disease who died 1 year after transplantation, engraftment of donor cells was noted in the brain. Differentiation to oligodendrocytes was demonstrated in vitro and subsequently in vivo. In a child with MPS III (Sanfilippo syndrome), donor cells differentiated into cardiac myocytes in the heart 6 months after transplantation. Dr. Marco Zecca discussed unrelated donor CBT for children with hematologic malignancies. The experience derived from transplantations in hundreds of children with acute lymphoblastic or myeloid leukemia indicates that the results achieved are substantially comparable to those obtained in children given BMTs. In particular, the risk of leukemia recurrence is not increased after CBT, suggesting a graft-versusleukemia effect as with BMT. Disease status at the time of transplantation is the main factor influencing patients outcome; patients who undergo transplantation in the first or second remission are those with the best results. Because most deaths in children given CBT are due to infectious complications (related to both delayed hematopoietic recovery and lack of adoptive transfer of memory T cells), strategies able to accelerate both hematopoietic and immune reconstitution could widen the use of cord blood cells for transplantation. Dr. Mark Walters discussed the role of CBT in thalassemia and pointed out that 1500 patients with thalassemia have received a hematopoietic cell transplant. The alternative of increasingly effective supportive care also affects the decision to pursue transplantation. Initial results of UCBT for thalassemia (largely with sibling donors) suggest that acceptable outcomes are possible if measures are taken to mitigate the risk of graft rejection. Among 44 patients with sickle cell disease or thalassemia who received augmented conditioning therapy, the event-free survival was 94%, compared to 62% among those who received a standard combination of busulfan and cyclophosphamide, with or without horse antithymocyte globulin. The clinical experience with unrelated UCBT for thalassemia remains very limited. New techniques to prevent GVHD and promote engraftment, coupled with refined donor selection criteria, should expand the availability of transplantation for thalassemia major. Dr. Eliane Gluckman reported results of 44 patients with Fanconi anemia who received an unrelated CBT (UCBT). The median age was 7.7 years, and median follow-up was 20 months (range, 8-83 months). At UCBT, the median number of neutrophils was 620/ ml, and 13 patients had received more than 20 red blood cell transfusions. The cord blood was HLA mismatched in 37 patients. The median number of BB&MT NCs at freezing was 5.5 10 7 /kg and at infusion was 4.7 10 7 /kg. Eighteen patients received cyclophosphamide and irradiation based preparative regimens, 12 patients received FLU-containing regimens, 6 received cyclophosphamide alone, and 8 received other regimens. GVHD prophylaxis consisted of cyclosporin A and methotrexate (43%) or cyclosporin A alone (23%). Neutrophil recovery at day 60 was 56% 8%; cell dose and number of HLA disparities were associated with a higher probability of neutrophil recovery. Acute GVHD (grade II-IV) was 23% 6% (grade II, 7%; grade III, 7%; grade IV, 11%), and chronic GVHD occurred in 4 (19%) of 21 patients at risk. Two-year survival was 36% 7%. In multivariate analysis, 2 factors were associated with better survival: NC at freezing ( 5.5 10 7 /kg; hazard ratio, 0.25; P.004) and FLU-containing preparative regimen (hazard ratio, 0.16; P.018). In patients receiving a higher cell dose, 2-year survival was 55%, versus 16% for those receiving a lower dose (P.005); it was 67% for those receiving a FLU-containing regimen compared with 25% (P.016) for those receiving other regimens. The conclusion reached was that results of UCBT for Fanconi anemia patients are acceptable and can be improved by better selection of the cord blood units and by use of FLU in the preparative regimen. Session V reviewed immune status after CBT. Dr. Bruce Blazar reviewed the role of CD4 CD25 T-regulatory cells in allogeneic BMT. A subpopulation of CD4 T cells that coexpresses CD25, the interleukin-2 receptor chain, without evidence of prior activation, has been shown to be important in self-tolerance by suppressing autoimmune responses. The authors extensive preclinical studies demonstrated that depletion of either host CD25 cells before BMT or donor CD25 cells in the donor graft inoculum markedly accelerated GVHD lethality. In contrast, supplementing the donor graft with fresh CD4 CD25 T-regulatory cells delayed GVHD lethality, and adding ex vivo activated and expanded T-regulatory cells virtually abolished GVHD lethality. Their experiments led them to conclude that the infusion of ex vivo activated and expanded CD4 CD25 cells is a highly potent means of inhibiting GVHD and facilitating alloengraftment, and they suggest that a clinical trial of CD4 25 merits consideration in the context of allogeneic BMT. Dr. Kenneth Weinberg pointed out that immune reconstitution after hematopoietic stem cell (HSC) transplantation is a fundamental problem that affects the clinical outcome of transplantation, regardless of the stem cell source. Some of the features of UCBC that make them attractive as a source of HSCs may also contribute to difficulties in immune reconstitution. Although the number of HSCs in UCBC sources 731

may be limiting, the increased proliferative potential of cord blood HSCs could also be advantageous. A major difference between UCBC and adult marrow or peripheral blood stem cells is the nature of the mature T lymphocytes: the T lymphocytes in UCBC are predominantly naive, whereas adults have mainly memory T lymphocytes. A decreased ability to stimulate naive T cells may account for reduced alloreactivity in UCBC but may also lead to decreased sensitivity to nominal antigens such as viruses. Naive T lymphocytes depend on self-antigen and interleukin-7 as signals for homeostatic proliferation. Interleukin-7 therapy is an attractive strategy for promoting posttransplantation thymopoiesis but may be complicated by exacerbation of GVHD. Some approaches for translation of these experimental concepts into empirical studies in UCBC transplantation are being studied. Dr. Juliet Barker shared the results of a review of serious infectious complications in children who underwent myeloablative unrelated donor transplantation at the University of Minnesota with unmanipulated bone marrow (n 52), bone marrow with T-cell depletion (TCD, n 24), or UCB (n 60) between 1994 and 2004. Analysis of the relative risk (RR) of infection within days 0 to 42, 43 to 100, and 101 to 180 revealed no differences between groups except a greater risk of viral infections from days 0 to 42 in TCD as compared with bone marrow recipients (RR, 3.5; P.02). After day 180, TCD recipients had the greatest infection risk (RR, 3.1; P.03), whereas risk in UCB (RR, 0.5; P.23) was comparable to that in bone marrow recipients and superior to TCD. These data suggest that although serious infection is a major problem for all unrelated donor recipients, regardless of stem cell source, the risk of serious infection after pediatric UCBT is comparable to that of unmanipulated bone marrow and superior to TCD. Session VI reviewed ex vivo expansion and regenerative medicine. Dr. Elizabeth Shpall reported on several trials involving ex vivo expansion of cord blood. In one study, 43 patients with hematologic malignancies (n 40) or breast cancer (n 3) received high-dose therapy followed by unrelated allogeneic CBT. A fraction of each patient s cord blood allograft was CD34-selected and cultured ex vivo for 10 or 14 days before transplantation in defined media with stem cell factor, granulocyte colony-stimulating factor, and megakaryocyte growth and differentiation factor. The remainder of the cord blood graft was infused without further manipulation. This study demonstrated that the CD34 selection and ex vivo expansion of cord blood before transplantation is feasible. The engraftment failure rate was relatively low, and the time to neutrophil in large adults was in the range reported for pediatric recipients of much larger cord blood cell doses. Two additional trials are in progress at the M. D. Anderson Cancer Center. In these trials, patients with hematologic malignancies are being randomized to receive either 2 unmanipulated cord blood units or 1 unmanipulated unit and 1 unit from which all the cells are expanded ex vivo. Future directions include the use of mesenchymal stem cells as a supportive stroma for the expansion of cord blood mononuclear cell progenitors, developed by McNiece et al. Dr. Peter Wernet discussed the isolation of non- HSCs from UCB. He pointed out that an intrinsically pluripotent CD45 population from placental cord blood, termed unrestricted somatic stem cells, grows adherently and can be expanded to large quantities of cells without losing pluripotency. In vitro one can demonstrate unrestricted somatic stem cell homogeneous differentiation into osteoblasts, chondroblasts, adipocytes, and hematopoietic and neural cells, including astrocytes and neurons expressing neurofilament, sodium channel protein, and various neurotransmitter phenotypes. In vivo experiments in animal models demonstrate differentiation into cells of various tissue types after injection into specific cites. No malignant transformation was observed in any of the animals. Morey Kraus discussed potential applications of UCB-derived non-hscs. The potential of such cells has been studied with encouraging results in a rat stroke recovery model, in SOD1 G93A mice in the study of ALS, after myocardial infarction or acute ischemia, and in the Mdx mouse mode of muscular dystrophy. Session VII reviewed contemporary issues regarding cord blood banking. Dr. Michael Creer pointed out that cryopreservation and thawing of cord blood products are the single greatest and most variable sources of HSC loss from collection to infusion. A series of studies examined the effects of cooling rate and type of cryoprotectant (dimethyl sulfoxide or propanediol) on cord blood HSC cell number (TNC and CD34), viability (trypan blue or 7-AAD), and function (colony-forming unit assay) after thawing. Varying the cooling rate from 1 to 20 C/min had little effect on recovery of cord blood HSCs or trypan blue viability. However, cord blood HSC function (colony-forming unit recovery) decreased markedly at cooling rates above 5 C/min. Transfer of cord blood units in Styrofoam containers with liquid nitrogen or segment removal in liquid nitrogen vapor did not cause any detectable loss of HSC number, viability, or function. Similarly, repeated exposure of frozen cord blood units to 10 consecutive 1-minute intervals at room temperature also had no significant effect on HSC number, viability, or function. Dr. Ellen Lazarus discussed the Food and Drug Administration regulatory approach to human cell and 732

tissue products and how this applies to cord blood units. She indicated that peripheral blood stem cells and cord blood are regulated as biologic cell therapies under the tissue approach. Implementation of the regulatory approach is under way, and a period of enforcement discretion for unrelated donor cord blood is ongoing. Cord blood manufacturers currently must register with the Food and Drug Administration. Finalizing regulations is currently a Food and Drug Administration priority. There is a recognized need for international harmonization of cord blood regulations. Symposium Abstracts 1 CORD BLOOD TRANSPLANTATION TO ADULT PATIENTS: A SINGLE- BANK S EXPERIENCE Stevens, C.E., Scaradavou, A., Rubinstein, P. National Cord Blood Program of the New York Blood Center Since 1993, the New York Blood Center has provided cord blood (CB) grafts to more than 1500 recipients, 421 of whom where adults (age 16). These patients were treated in 53 US and 26 non-us Transplant Centers (TCs). Data from 391 (92.9%) are available. The proportion of adult recipients has grown from 9% in 1995 to over 30% in 2001-2003. More than 90% (363/391) had leukemia (41% advanced), lymphoma, or myelodysplasia. No. of patients aged 16-24, 25-39, and 40: 130, 135, and 125, respectively; there were 15% African American recipients, 13.3% Hispanic, and 65.6% Caucasoids, respectively. TNC doses 2.5 10 7 /kg were given to 62%, 25-49 10 7 /kg to 37%, and 5 10 7 /kg to 1.3% of all recipients. HLA was matched at serological levels for A and B and at DNA allele level for HLA-DR. Seven TCs have performed 10 CB transplantations each. Neutrophil engraftment was significantly reduced when TNC doses 2.5 10 7 /kg and stratified significantly with HLA mismatch levels. Engraftment was superior with TBI conditioning compared to no TBI; and regimens including methotrexate associated with reduced engraftment. Overall survival was 30.2% at one year and 23% at five years post-transplant. Centers with 10 CB transplantations had better survival (by Cox Regression analysis) and so did those with better HLA matches, those without advanced stage leukemia, and those whose conditioning included fludarabine. The effect of cell dose was not significant in the Cox regression. Patients with good HLA matches (6/6 and 5/6 without rejection mismatches) and those conditioned with regimens including TBI and fludarabine had excellent survival: 68% at one year, respectively. The data indicate that, despite the low graft cell doses, remarkable improvement in overall survival rate of adults transplanted with cord blood is possible by improvements in HLA matching and conditioning regimens. 2 UMBILICAL CORD BLOOD TRANSPLANTATION: NOVEL APPROACHES TOWARD IMPROVING ENGRAFTMENT Wagner, J.E., Barker, J. Blood and Marrow Transplant Program, University of Minnesota, Minneapolis, MN. Cryopreserved umbilical cord blood (UCB) has been investigated as a potential strategy for augmenting the pool of acceptable donors, reducing the risks of acute and chronic graft-versus-host disease (GVHD) and improving survival. Cell dose, however, has clearly been identified as a major limitation, often preventing the consideration of UCB for adult recipients. Recipients of 1.7 10 5 CD34/kg have slow hematopoietic recovery (median 35 days) and significantly lower incidence of engraftment (68% [95% CI, 46-91]). Therefore, strategies for increasing the cell dose, e.g., ex vivo HSC expansion, transplantation of multiple UCB units, and co-infusion of healthy haploidentical mesenchymal stem cells (MSC), are being explored. Multi-unit UCB transplantation. Twenty-three adult and adolescent patients [median age 24 years (range: 13-53)] with high-risk hematologic malignancy were transplanted with two partially HLAmatched UCB units after myeloablative conditioning. The median total infused dose 3.5 10 7 NC/kg (range 1.1-6.3). All evaluable patients (n 21) engrafted at a median of 23 days (range 15-41) with one unit predominating. While there was no association between nucleated cell dose, CD34 cell dose, or HLA-A,B,DRB1 match and which unit predominated, the winning unit had a higher CD3 dose. Incidence of severe III-IV acute graft-versus-host disease (GVHD) was 13% (0-26%), and disease-free survival was 57% (95% CI: 35-79) at 1 year. These data suggest that 1) double unit UCBT is safe with one unit predominating over time, 2) the unit that disappears over time may facilitate engraftment of the winning unit, possibly by immune mediated mechanisms, and 3) CD3 dose might predict which unit will ultimately engraft long term. Most importantly, the use of two UCB units extends the application of this treatment to nearly all adults for whom one unit would have been insufficient. Co-Infusion of haploidentical MSC and UCB transplantation. Fifteen patients [median age 7.5 years (range: 0.2-16)] with high risk leukemia were transplanted with HLA 0-2 antigen mismatched UCB and haploidentical parental donor MSC after a myeloablative conditioning. No toxicity was observed related to the infusion of MSC. All treated patients (n 8) had hematopoietic recovery and complete chimerism at a median of 19 days (range: 9-28). Platelet recovery occurred at a median of 1.7 months (range: 1.2-3.3). Incidence of grade II-IV acute GVHD was 2 of 8 patients with no patient having chronic disease. With a median followup of 2.9 years, 6 patients are alive and disease free with two patients dying of infections on days 53 and 63, respectively. These data suggest that engraftment, particularly platelet recovery, may be enhanced by the co-infusion of MSC. Clearly larger patient numbers are required to confirm these preliminary observations. 3 UMBILICAL CORD BLOOD TRANSPLANTATION AFTER A NON-MYELOA- BLATIVE THERAPY IN HIGH RISK ADULTS Wagner, J.E., Barker, J. Blood and Marrow Transplant Program, University of Minnesota, Minneapolis, MN. Bone marrow transplantation (BMT) in adults is associated with a high risk of graft-versus-host disease (GVHD), opportunistic infection and regimen related toxicity. For these reasons, umbilical cord blood (UCB) after a non myeloablative preparative regimen has been explored. Fifty-one adults with advanced hematologic malignancies [median age 50 years (range 19-60)], ineligible for myeloablative conditioning by virtue of advanced age, extensive prior therapy or serious co-morbidities received UCB transplantation after cyclophosphamide 50 mg/kg, fludarabine 200 mg/m 2, and 200 cgy TBI. Immunosuppression was with cyclosporine-a to at least day 100 and mycophenolate mofetil to day 30. Eight patients with no combination chemotherapy in the 6 months prior to transplant also received ATG during conditioning. Thirteen patients (25%) received single UCB units and 38 (75%) received double unit grafts. The median total infused cell dose was 3.4 10 7 NC/kg (range 1.1-5.7). Units were predominantly 1-2 antigen HLA mismatched with the recipient. Of 50 evaluable patients, neutrophil recovery occurred at a median of 8 days (range 5-32). Four patients had failure of donor derived engraftment with autologous recovery and 1 had late graft rejection. The cumulative incidence of sustained donor engraftment BB&MT 733