Engraftment Effects of Intra Bone Marrow Compared to Intravenous Transplantation of Allogeneic Hematopoietic Stem Cells Following a Reduced Intensity Conditioning in a DLA-Identical Canine Littermate Model

Introduction: Successful engraftment following hematopoietic stem cell transplantation (HSCT) depends on factors like immunosuppression, graft composition and number of infused HSC. Whereas the immunosuppression as well as the type and composition of the graft are influenceable low numbers of available HSCs i.e. “weak grafts” remain a clinical challenge. Weak grafts are accompanied by increased graft failure rates and longer cytopenias associated with increased morbidity. Intra bone marrow (IBM) infusion of HSC might be an approach to overcome these problems. Studies in rodents demonstrated faster engraftment with an IBM HSCT approach compared to intravenous (IV) HSCT following myeloablative conditioning. Studies of IBM HSCT following non-myeloablative or reduced intensity conditioning (RIC) are missing.

Aims: Exploring the feasibility and efficiency of IBM allogeneic HSCT in comparison to IV HSCT in dog leukocyte antigen (DLA) identical canine littermates using a RIC regimen.

Methods: DLA-identical siblings were used as donor/recipient pairs for HSCT. Recipient dogs were conditioned with 4.5 Gy total body irradiation before HSCT (d0) and received 15 mg/kg Cyclosporin A BID as pre- and postgrafting immunosuppression (d-1 to d+35). BM grafts were harvested at d0. In the control group (CON, n=7) unmodified BM was transplanted IV. In the IBM group (n=7) BM harvests were centrifuged and buffy coat of the BM was then transfused simultaneously into the recipient humeri and femura (50 ml, 10 min). 10 dogs are currently evaluable. Chimerism of the peripheral blood mononuclear cells (PBMC) and granulocytes (G) were tested weekly until week 8 and afterwards in larger intervals. Blood cell counts and clinical toxicities such as weight loss were monitored.

Results: Infusion of BM directly into the bone was feasible. All animals engrafted. Median number of infused total nucleated cells was 4.0*10⁸/kg (range 2.3-6.0*10⁸/kg, IBM) and 3.3*10⁸/kg (range 1.9-5.0*10⁸/kg, CON, IBM vs CON: p=0.4). Median CD34+ numbers infused were 3.1*10⁶/kg (range:1.2-10.0*10⁶/kg, IBM) and 3.9*10⁶/kg (range: 1.0-7.2*10⁶/kg, CON; IBM vs CON: p= 0.8).

Hematopoietic recovery in the IBM and CON groups were similar. Leukocytes recovery (>1.0*10⁹/l) occurred at median d+11 (range: d+10 - d+16, IBM) and d+10 (range: d+9-d+12, CON; IBM vs CON: p=0.3). Median leukocytes nadirs amounted to 0.23*10⁹/l (IBM) and 0.28*10⁹/l (CON; IBM vs CON: p=0.3) and median duration of leukopenia (<1.0*10⁹/l) were 6 days (range: 5.0–11.0, IBM) and 4 days (range: 3.0–6.0, CON; IBM vs CON: p=0.1). Median platelet nadir after IBMT was 10.0*10⁹/l (range: 0.0 - 25.0*10⁹/l) and 6.0*10⁹/l (range: 3.0-15.0*10⁹/l, CON; IBM vs CON: p=0.8). Period of thrombocytopenia (≤50.0*10⁹/l) lasted for 12 days in both groups (p=0.7).

Chimerism analyses showed an early and fast increase in donor chimerism in both groups. The PBMC donor chimerism at d+14, d+28 and d+56 were 46% (range: 30-53%), 57% (range: 40-73%), 64% (range: 60-83%) for IBM. Results in CON were 37% (range: 17-93%), 60% (range: 49-100%), 57% (range: 40-100%) (IBM vs CON, p=n.s. (all time points)). The G chimerism values at that specific points were 95% (range: 53-100%), 100% (range: 53-100%), 96% (range: 88-100%) for IBM and 100% (range: 93-100%), 99% (range: 92-100%), 98% (range: 93-100%) for CON (IBM vs CON, p=n.s. (all time points)).

Primary goal of the study was the feasibility of the IBM approach. Ethics regulations did not allow to use weak grafts (≤2.0*10⁶/kg) intentionally. However, 4 animals received weak grafts (CON n=2, 1.0 and 2.0*10⁶/kg; IBM n=2, 1.2 and 1.3 *10⁶/kg). Of interest, comparing data of these dogs showed that durations of leukopenia were similar (median 10 days, both groups), but duration of thrombocytopenia were different (median 8 days, IBM vs 22 days, CON). Additionally, long term donor chimerism was higher in the IBM (median 80% PBMC, 100% G) vs CON (median 61% PBMC, 42% G).

Conclusion:

First, IBM HSCT is a feasible and effective method to deliver HSC directly into the bone marrow following RIC in a canine HSCT model. Second, our preliminary data suggest that IBM HSCT reveals advantageous engraftment differences in regards to platelet recovery and donor chimerism kinetics compared to the IV HSCT when grafts with low HSC numbers were infused. Follow up data of this study and future studies will have to clarify these observations further.