Five decades of progress in haematopoietic cell transplantation based on the preclinical canine model


  • M. Lupu,

    1. Clinical Research Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA
    2. Oregon Cancer Center for Animals, Oregon State University, Corvallis, OR, USA
    Search for more papers by this author
  • R. Storb

    Corresponding author
    1. Clinical Research Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA
    2. Department of Medicine, University of Washington School of Medicine, Seattle, WA, USA
    Search for more papers by this author

Dr R. Storb
Fred Hutchinson Cancer Research Center
1100 Fairview Avenue N, D1-100
WA 98109, USA


The preclinical canine model has proved valuable for the development of principles and techniques of haematopoietic cell transplantation (HCT) applicable to human patients. Studies in random-bred dogs concerning the impact of histocompatibility barriers on engraftment and graft-versus-host disease, the kinetics of immunological reconstitution, the efficacy of various pretransplant conditioning regimens, post-transplantation immunosuppression protocols, treatment of malignant diseases, and graft-versus-tumour effects have advanced HCT from an investigational therapy with uncertain clinical benefit half a century ago to an important treatment choice for thousands of patients treated annually in transplantation centres worldwide. More recent preclinical canine studies have resulted in the clinical translation of non-myeloablative, minimally invasive transplantation protocols that have extended allogeneic HCT to include older human patients with malignant and non-malignant, acquired or inherited haematological disorders, and those with comorbid conditions. Here, we review the contributions of the canine model to modern HCT and describe the usefulness of HCT for the treatment of canine haematological disorders.

The beginnings of haematopoietic cell transplantation

The history of haematopoietic cell transplantation (HCT) began with attempts to allay the lethal effects of irradiation that were observed in the wake of the atomic bomb explosions towards the end of World War II. It was soon recognized that bone marrow was the most radiosensitive organ in the body, and that marrow transplantation could rescue radiation victims from the lethal effects of marrow aplasia. This discovery encouraged clinicians and scientists to explore more aggressive protocols in patients with life-threatening haematological malignancies and increase the intensity of cytotoxic therapies beyond levels that were marrow toxic in hopes of eradicating the underlying cancer; rescue from the otherwise lethal marrow aplasia would be accomplished by the intravenous infusion of bone marrow cells.

In 1957, Thomas et al. reported studies of marrow grafting in patients with leukaemia,1 concluding after 2 years2 that marrow transplantation was for the most part unsuccessful, with almost all patients dying from either allograft failure or recurrent leukaemia. The immunological mechanisms involved in graft rejection and graft-versus-host disease (GVHD) were not well understood at that time. Overall, the early clinical results were disappointing and raised the concern that the allogeneic barrier could never be overcome. Therefore, many clinicians abandoned the idea that marrow transplantation could be used for the treatment of haematological malignancies in human patients,3 even though graft-versus-leukaemia effects had already been recognized.4

Preclinical studies of HCT: Murine and canine models

The first studies aimed at elucidating the radiation protection phenomenon were conducted in inbred mice.5–16 Early research showed that the protection of lethally irradiated mice could be achieved not only by shielding the spleen or one femur with lead during the radiation6 but also by the intraperitoneal7 or intravenous16 infusion of syngeneic spleen or bone marrow after irradiation. Jacobson et al.6 hypothesized that the mouse spleen contained a humoral factor capable of stimulating the regeneration of blood-forming tissue, thereby advancing the ‘humoral’ theory of haematopoietic reconstitution after irradiation. As an alternative to the humoral hypothesis, Barnes et al.8 suggested that the active principle in the spleen and bone marrow suspensions might be living cells that would act at least temporarily as a tissue graft, advancing the ‘cellular’ theory, which received little support at that time. The cellular hypothesis received strong support from the observation by Main and Prehn 9 in mice that marrow donor skin grafts survived indefinitely in allogeneic marrow recipients. Subsequently, Trentin et al.5 showed that the skin graft tolerance was specific for the donor strain. In 1956, three independent groups10–12 confirmed the cellular theory by showing marrow repopulation through donor cells using various blood genetic markers. Further studies in inbred mice addressed other fundamental rules of marrow transplantation biology, including the observations that the mechanisms of graft rejection and GVHD13 were controlled by genetic factors14 that were in turn governed by the major histocompatibility complex (MHC).15 However, in 1967, the experimental haematologist Dirk W. van Bekkum drew attention to the fact that the first clinical trials in human patients with haematological malignancies failed mainly because ‘the clinical applications were undertaken too soon, most of them before even the minimum of basic knowledge required to bridge the gap between mouse and patient had been obtained.’17

In the following years, scientists directed their efforts at identifying more suitable preclinical animal models. Dogs appeared especially useful owing to their random-bred nature, large body size, longer life span, wide genetic diversity, and well-mixed gene pool; they are the only mammalian species besides human beings to possess these qualities.18 Further, short gestation time and large litter sizes allowed studies of donor–recipient combinations that were matched or mismatched for MHC antigens, which are known as the dog leukocyte antigen (DLA) system, simulating human leukocyte antigen (HLA)-matched donor–recipient pairs. Early experiments in the canine model using DLA-matched donor–recipient pairs and postgrafting immunosuppression19,20 renewed a sense of confidence towards treating human patients with haematological diseases by allogeneic marrow transplantation. As a consequence, in the late 1960s, the results of the preclinical canine HCT studies began to be translated to patients with aplastic anaemia and haematological malignancies using HLA-matched sibling marrow donors.21,22

Not surprisingly, studies in canines have provided an excellent basis for most of the HCT principles and techniques23 that have been directly translated over the past decades to the clinical setting for a wide spectrum of human diseases.16 Dog models of total body irradiation (TBI),17,24–31 chemical32–35 and radioimmunological36,37 myeloablation, cell dose requirements,38–41in vivo42–47 and in vitro48–51 graft manipulation, engraftment and GVHD,52–54 as well as graft-versus-tumour effects55–57 have been essential to solve the existing clinical problems of human HCT.

DLA typing

Early canine allogeneic marrow transplantation studies showed occasional long-term stable engraftment, even though most animals succumbed to problems with GVHD and graft rejection.58 From these experiments, it was concluded that major efforts would need to be directed towards understanding and defining histocompatibility barriers between donors and recipients.

The use of canine species in the experimental HCT field involved a rigorous study of the DLA system.19,59–68 In 1968, Epstein et al.59 described a serologic canine histocompatibility typing system and showed that compatibility between littermate donors and recipients played a fundamental role in reducing the risks of both graft rejection and GVHD. Further, long-term donor engraftment was shown to occur in both unrelated and related recipients, which were matched either by serologic typing alone or by serology and mixed leukocyte culture non-reactivity.19,61 Thus, the dog was the first random-bred species in which the impact of in vitro donor–recipient matching on HCT outcomes was shown.19

Antigenic canine histocompatibility polymorphisms were first studied by serological59,61 and cellular typing in mixed leukocyte culture systems.60,61 It was not until later that the term dog leukocyte antigen was introduced, and with it, the recognition that the histocompatibility complex could be divided into class I and class II regions. Subsequently, understanding of the molecular organization of the DLA region provided tools for molecular histocompatibility typing, which was facilitated by identification of convenient microsatellite polymorphisms within class I and class II regions that were inherited in a Mendelian fashion.68 As a consequence, molecular assessment of DLA class I and class II microsatellite marker polymorphisms,62,63 combined with DLA class II DRB1 allele sequencing,65,67 enabled high-resolution histocompatibility testing of canine families and rapid selection of DLA-identical donors.

Graft collection

Initial canine HCT studies involved the use of bone marrow as the source of haematopoietic progenitor cells obtained by aspiration from the humeral and femoral bones.23 Marrow cells stored at −80 °C in dimethyl sulfoxide were capable of recovering 80% of the haematopoietic colony forming units in vitro69 and induced haematopoietic reconstitution of lethally irradiated dogs.20,48

Subsequent experiments in the canine model showed that either autologous40,49,70 or allogeneic50,51,71,72 peripheral blood stem cells could be successfully used as substitutes for marrow grafts, leading to recovery of normal haematopoiesis after otherwise lethal marrow ablation. Quantitative studies of peripheral blood mononuclear cell (PBMC) grafting showed that approximately 10-fold more PBMCs than marrow cells were necessary to ensure stable engraftment.39,40 To induce expansion and/or peripheralization of haematopoietic progenitors from the marrow space into the peripheral blood, various mobilizing agents, such as cyclophosphamide (Cy)42; recombinant cytokines, including recombinant canine (rc) granulocyte colony-stimulating factor (G-CSF),43,45 rc stem cell factor (SCF, c-kit ligand),43,45 rc granulocyte–macrophage colony-stimulating factor44; or the recently described CXCR4 antagonist AMD3100,47 were evaluated in the canine model. In addition, to facilitate PBMC collection, numerous apheresis-based techniques have been adapted to dogs.19,23,72–78

Currently, more than 90% of all autologous HCT protocols for adult human patients with haematological malignancies include the use of mobilized PBMC grafts as a source of stem cells because of higher cell yields, decreased procedural risks, faster engraftment and shorter hospitalization compared with the use of marrow grafts. However, human patients with aplastic anaemia79 and those given grafts from unrelated donors80 showed better outcomes after allogeneic marrow rather than PBMC grafts. Although the major disadvantage of allogeneic PBMC grafts was associated with a higher incidence of chronic GVHD,81 an advantage in survival was observed in human patients with advanced haematological cancers.82

Conditioning regimens

Conditioning regimens given before HCT serve to suppress the immune system of the recipient for graft acceptance, while eradicating the underlying disease. Extensive preclinical canine studies investigating various conditioning regimens of chemotherapy, TBI, monoclonal/polyclonal antibody administration, or combinations thereof have set the stage for successful HCT trials in human patients.


In 1957, Thomas et al. established that irradiation itself did not eliminate leukaemia in human patients; it was postulated that eradication of underlying malignancy might be possible by addition of chemotherapy.1

Encouraging results of pretransplant conditioning with Cy in mice and rats83 prompted investigation of high-dose Cy conditioning regimens in dogs and rhesus monkeys, in place of TBI.32,33,83 It has been shown that dogs given a single Cy dose of 100 mg/kg could be rescued from lethal immunosuppression by administration of either autologous33 or allogeneic32 marrow grafts from DLA-matched donors, leading to sustained engraftment and mixed donor–recipient haematopoietic chimerism. High-dose Cy conditioning regimens were subsequently translated to human patients with aplastic anaemia undergoing allogeneic HCT.79

Following studies with busulfan in the rat,83 pre-HCT conditioning with dymethyl myleran (DMM), an intravenously injectable homologue of busulfan known for its marked myelosuppressive activities, was extensively evaluated in both healthy dogs34,35 and in dogs with malignant lymphoma.84 In healthy dogs, DMM given at a dose of 7.5 mg/kg induced profound marrow aplasia, which was reversed by autologous marrow grafting.34 In the allogeneic HCT setting, despite the achievement of a mixed chimeric state, the level of immunosuppression using DMM at a dose of 10 mg/kg as a single agent led to successful engraftment in 50% of the dogs, while the remainder rejected their grafts and died from pancytopenia; however, more consistent engraftment was achieved when either antithymocyte serum (ATS), produced by immunization of rabbits with puppy thymocytes, or a combination of ATS and procarbazine was given in addition to DMM.35 In dogs with malignant lymphoma, DMM was less effective in inducing complete remissions compared with TBI conditioning at similar marrow toxic doses.84,85

Total body irradiation

Conditioning regimens using TBI have been extensively studied in both healthy dogs17,24–31,86 and in dogs with tumours56,57,85,87–89 as part of engraftment, toxicity and malignant disease treatment studies. It was recognized that bone marrow haematopoietic stem cell progenitors were very vulnerable and reacted uniformly to the harmful effects of irradiation. Therefore, for the treatment of lymphoid malignancies, TBI conditioning provided the advantage of eradicating non-cycling malignant cells as effectively as cycling cells.

Thomas et al. were the first to show prompt recovery of bone marrow and lymphoid tissues in lethally irradiated dogs given intravenous infusions of freshly isolated or cryopreserved autologous or allogeneic marrow cells.24,86,90 All dogs that were not given a marrow infusion after a TBI dose of 400 cGy died from complications of marrow failure.24

In the absence of postgrafting immunosuppression, stable allogeneic DLA-identical littermate marrow grafting and full donor chimerism were consistently obtained only after conditioning with TBI doses higher than 900 cGy, delivered at a dose rate of 7 cGy/min.29 However, engraftment of DLA-identical littermate marrow was consistently achieved with a single dose of 450 cGy when the exposure rate was increased to 70 cGy/min.91 Several studies investigating the impacts of dose rate on engraftment and toxicity showed that the exposure rate and total dose were the most important parameters for acute toxicity associated with TBI. It was determined that the overall TBI dose tolerated by healthy dogs given autologous marrow grafts could be increased from 10 to 14 Gy by reducing the exposure rate from 10 to 5 cGy/min. Generally, acute toxicity correlated with the TBI dose rate, while chronic toxicity was related to the total TBI dose.26,27

To decrease TBI-related toxicity rates in human patients, but to retain a strong antitumour effect and promote sustained engraftment, a new approach was investigated in the canine model by using higher TBI doses given in a fractionated fashion.30,31,91 This strategy was based on the principle that haematopoietic cells were less capable of undergoing DNA repair after multiple, fractionated TBI doses compared with other tissues that would thus be affected to a lesser extent. As a consequence, long-term complications were significantly less in dogs given fractionated versus single TBI doses. When total doses of 12–21 Gy were given in multiple fractions of 1.5–2 Gy at intervals of 3–6 h, with dose rates ranging from 2 to 20 cGy/min, an advantage of fractionated doses regarding acute toxicities was noted only at the highest rate of 20 cGy/min.30

Other studies investigated the effect of post-TBI administration of haematopoietic growth factors on engraftment of DLA-identical littermate marrow grafts.92,93 Post-TBI administration of recombinant human G-CSF or rcSCF significantly accelerated recovery of neutrophils and monocytes without altering platelet recovery and without significantly increasing the risks of graft failure or GVHD; however, lymphocyte recovery was more rapid in G-CSF-treated dogs. Although growth factors did not improve engraftment of DLA-identical grafts, they slightly improved survival.

Sustained engraftment of DLA-non-identical grafts was seen only after preparative regimens with higher TBI doses than those used in the DLA-identical HCT setting. TBI doses of 1500–1800 cGy were generally tolerated only when given in fractions.26,28,94,95 When a TBI dose of 1800 cGy was given in three fractions of 600 cGy delivered at 2.1 cGy/min at 48-h intervals, the immunosuppression achieved was sufficient to allow sustained engraftment of unrelated DLA-incompatible marrow grafts.95 Furthermore, attainment of sustained engraftment of histoincompatible marrow grafts following infusion of donor buffy coat cells or lymphocytes19,96 was an important observation that promoted the use of PBMC grafts. Other manipulations, such as the addition of monoclonal antibodies to TBI,97–100 have also been successful in assuring sustained engraftment of DLA-non-identical grafts, as described below.

Monoclonal and polyclonal antibodies

The efforts to minimize TBI-related toxicities prompted the investigation of novel conditioning regimens in the canine model, including the use of monoclonal antibodies, polyclonal antibodies, or fusion peptides targeted towards specific T-cell populations or T-cell costimulatory pathways.

A beneficial effect in overcoming resistance to DLA-mismatched marrow grafts was noted in 50% of the dogs given anti-class II monoclonal antibodies in addition to TBI and postgrafting immunosuppression.97 In the same model, the administra-tion of S5 monoclonal antibody directed against the CD44 antigen led to sustained engraftment in 75% of the transplanted dogs.100 Further canine studies explored the immunosuppressive effects of radiolabelled monoclonal antibodies, such as the 131I-labelled/S536 or bismuth-213-labelled anti-CD45101/anti-TCRαβ37 conjugates, capable of inducing lethal marrow suppression36 or selective T-cell ablation,37 respectively.

A targeted non-myeloablative strategy focused on induction of T-cell anergy was tested in dogs given DLA-identical littermate marrow grafts after conditioning with CTLA4-Ig, a fusion peptide that blocked T-cell costimulation through the B7-CD28 signalling pathway, and an otherwise inadequate TBI dose of 100 cGy; when this approach was combined with postgrafting immunosuppression by mycophenolate mofetil (MMF) and cyclosporine (CSP), stable mixed haematopoietic chimerism was achieved in 70% of the dogs.102 These results have encouraged further efforts in the preclinical canine model towards conditioning strategies that would further reduce the TBI dosing or even eliminate the need for TBI, thereby further reducing or even avoiding both short- and long-term side-effects from radiation.

Immunosuppression after HCT to prevent GVHD

The preclinical canine model has served to investigate various postgrafting immunosuppressive regimens consisting of single drugs or combinations of immunosuppressive agents52–54,102–117; these therapies suppress the immune system, either by affecting all immune reactive tissues in general or by specifically affecting only replicating immune reactive cells. Parameters such as the clinical effectiveness, engraftment, GVHD prevention, immune reconstitution and undesirable side-effects in long-term survivors have all been used to determine whether a given immunosuppressive regimen could be safely translated from the preclinical setting to clinical protocols.

The main factors that influenced the success in crossing the immunological barriers between donor and recipient were the degree of histocompatibility matching between donor and recipient, and the efficacy of pretransplant conditioning and post-transplant immunosuppressive regimens. In both dogs and human beings, even when donors and recipients were rigorously MHC matched, GVHD prevention remained a difficult goal; marrow grafting from MHC-identical donors carried the risk of acute and/or chronic GVHD, despite the administration of postgrafting immunosuppression.52,103 It was first recognized in dogs that GVHD was due to sensitization of donor T lymphocytes not only to disparate major but also minor histocompatibility antigens of the recipients.28 Once activated, lymphocytes attacked host tissues, predominantly the haematopoietic system, skin, gut and liver.52,118 Omission of immunosuppression after grafting has been associated with a high incidence of acute GVHD and adverse effects on survival.59

Studies in unrelated DLA-non-identical dogs have resulted in valuable strategies for GVHD prevention53,54,103,106–109,111,113 and treatment.104,105 In this model, in the absence of postgrafting immunosuppression, acute and rapidly fatal GVHD occurred within 2 weeks of myeloablative HCT.53 Significant delay in the onset of GVHD and prolongation of survival were obtained when postgrafting immunosuppression was given as single-agent therapy with methotrexate (MTX),103 CSP,106 azathioprine,53 succinylacetone,108 FK506 (tacrolimus),109 or as combinations of these agents.107,109,111,119 Based on canine studies, post-HCT immunosuppression regimens consisting of either MTX, CSP, or a brief course of MTX combined with either CSP or FK506, were successfully translated to human patients for GVHD prophylaxis; antithymocyte globulin was used to treat acute GVHD once it was established.52,54 Furthermore, it was found that postgrafting immunosuppression could be discontinued generally after 3–6 months of treatment because of establishment of mutual graft-versus-host tolerance.

Clear evidence of therapeutic graft-versus-tumour effects1,56 mediated by allogeneic effector T-cells that destroy tumour cells of the recipient after recognizing and reacting to disparate tumour-associated antigens, prompted the exploration in the canine model of less aggressive, non-myeloablative HCT regimens or so-called ‘minitransplants’. These regimens relied solely on host immunosuppression to facilitate stable mixed chimerism, prevention of both allograft rejection and GVHD and eradication of haematological malignant disease through a graft-versus-tumour response, following sublethal, minimal-invasive TBI conditioning.102,110,115 MMF,117 a purine synthesis inhibitor, was used in combination with other immunosuppressive drugs,54 leading to reduced frequency and severity of acute allograft rejection and improved survival in dogs given either DLA-identical102,110,112 or unrelated DLA-non-identical53,111 marrow grafts. In the DLA-identical model, pretransplant conditioning with a non-myeloablative dose of 200 cGy TBI and post-HCT immunosuppression with MMF and CSP induced stable mixed chimerism and prevented both allograft rejection and GVHD.110 This protocol has been successfully translated to human patients who were not eligible for conventional myeloablative transplantation because of advanced age or comorbidities.120

The impact of preceding blood transfusions on HCT outcomes

Early canine HCT studies drew attention to the fact that transfusions given before marrow grafts might jeopardize engraftment; it was hypothesized that this phenomenon might be associated with sensitization of recipients to non-DLA-associated polymorphic minor histocompatibility antigens.121 It was furthermore shown that DLA-identical recipients that received either three whole blood transfusions from the marrow donor or nine whole blood unrelated transfusions before TBI conditioning rejected 100% and 40%, respectively, of the marrow grafts.28 These studies showed that several minor polymorphic loci, which were undetected by the usual in vitro histocompatibility typing, were involved in the occurrence of transfusion-induced sensitization. It was almost two decades later when the sensitizing cells responsible for transfusion-associated graft rejection were identified as being dendritic cells contained in the transfusion product.122 These observations prompted the exploration of treatments designed to eliminate or inactivate the cells responsible for the induction of this phenomenon. The incidence of graft rejection was lessened by reducing antigen-presenting mononuclear cells through the use of buffy coat-poor blood transfusion products; transfusion-induced sensitization was successfully overcome by using a combination of an alkylating agent, procarbazine, and ATS as pre-HCT conditioning, or prevented by in vitro treatment of blood transfusions with ultraviolet light or 2000 cGy gamma radiation.28,123,124 These findings were then translated into the clinic, leading to improved management of the multiply-transfused patients with aplastic anaemia or other non-malignant diseases who were candidates for marrow transplantation.125–127

Haematopoietic reconstitution and side-effects after HCT

Haematopoietic reconstitution

Following myeloablative HCT, granulocyte counts achieved normal levels approximately by day 12; during the early postirradiation period, dogs might require whole blood or platelet transfusions. However, following non-myelablative HCT, life-threatening declines of peripheral blood cell counts generally did not occur.110 Although dogs with successful engraftment were profoundly immunodeficient for 200–300 days after myeloablative HCT, long-term survivors recovered their immune function and were not susceptible to increased incidences of infection.128

Conditioning regimen-induced side-effects

The main long-term side-effects after high-dose TBI conditioning in dogs were pancreatic insufficiency and atrophy leading to maldigestion and malnutrition, keratitis, pneumonitis, change in coat colour, cataracts and sterility; in addition, a five-fold increased incidence of spontaneous carcinomas and sarcomas was reported. These findings were not seen in a smaller number of dogs conditioned with either Cy or busulfan.129

Acute side-effects were associated with Cy administration, including anorexia, haematuria, vomiting and diarrhoea. Based on long-term surveillance for more than 10 years after HCT, canine recipients conditioned with Cy regained fertility and sired normal litters.130

Side-effects induced by immunosuppression after HCT

The limiting toxicity of MTX in dogs was gastrointestinal, as evidenced by diarrhoea and vomiting; however, mouth ulceration or so-called mucositis, which is a major side-effect in human patients, was rarely seen in dogs. The side-effects associated with MMF administration in dogs were gastrointestinal, consisting mainly of diarrhoea.131

The administration of calcineurin inhibitors in dogs was also associated with gastrointestinal side-effects, including an increased incidence of intussusception. CSP also caused liver and kidney function changes, although these appeared to be less common in dogs than in human beings at therapeutic levels. Following long-term CSP administration, dogs exhibited problems with papilloma infection; in addition, dogs presented changes in the skin and gums, as well as increased hair growth and blood pressure, all of which were reversible upon discontinuation of CSP administration.132 To reduce toxicities, CSP blood levels should be monitored and the dosing adjusted to keep levels within the therapeutic range.

HCT for canine malignant diseases

Another advantage of the dog model consisted of the availability of animals with spontaneous malignant diseases, resembling closely those found in human beings. In 1974, Weiden et al. established that dogs with malignant lymphoma had significant impairments of humoral and/or cellular immune reactivity, similar to those seen in human patients.133 In addition, it was shown that canine malignant lymphoma was responsive to combination chemotherapy.134 Many other malignant conditions of the dog have well recognized counterparts in humans.135 These observations heightened the value of the canine cancer model in the investigational setting of HCT, including graft-versus-tumour effects, in a randomly bred species.

The attempts to cure spontaneous canine malignancies by marrow transplantation were pioneered by the Seattle marrow transplantation group.55–57,84,85,87–89,136 The dogs involved in those studies were referred by practicing veterinarians with the consent of the dogs’ owners, and the survivors returned to their owners at the completion of the studies. Autologous grafts were given experimentally in dogs with malignant lymphoma,57,84,85,87–89,136 leukaemia85 and solid tumours85; the latter included mammary gland adenocarcinoma, squamous cell carcinoma, undifferentiated carcinoma, mastocytoma, leiomyosarcoma, osteosarcoma and melanoma. These early preclinical trials investigated different conditioning regimens consisting of TBI,85 chemotherapy,84 or combinations thereof.57,87–89,136 Studies showed that 25% of the dogs with malignant lymphoma in chemotherapy-induced remission given high doses of TBI, and either autologous cryopreserved marrow57,85,87 or freshly isolated PBMC136 grafts, became disease-free long-term survivors; however, the solid tumour-bearing dogs showed limited antitumour responses, consistent with the known greater radioresistance of solid non-haematological tumours.85

Three important allogeneic marrow transplantation studies were carried out by the Seattle transplantation group in dogs with malignant lymphoma,55–57 leukaemia56 and miscellaneous solid tumours.55,56 The graft-versus-lymphoma effect of allogeneic grafts was shown by prolonged disease-free survivals (P < 0.005) after allogeneic compared with autologous HCT using similar conditioning regimens.56 However, allogeneic grafts were complicated by high rates of fatal GVHD55 and toxicities related to postgrafting immunosuppression.57 No suggestion of significant graft-versus-tumour effects in dogs with non-haematological solid tumours was obtained, and all the transplanted dogs died with evidence of persistent tumours.56

Overall, it was established that the mortality rate in dogs with haematological malignancies given either autologous or allogeneic HCT was adversely affected by a compromised clinical status at the time of transplantation.56,85 From these studies, it was predicted that HCT could be more effective if performed in patients not only with minimal tumour burden, but before deterioration of their general condition; clinical translation of these observations resulted in improved transplantation outcomes.137

HCT for canine non-malignant diseases

The therapeutic potential of allogeneic DLA-identical HCT has been investigated for a wide range of canine non-malignant haematological and non-haematological disorders, such as congenital immunological and enzymatic deficiencies, and Duchenne muscular dystrophy.

Allogeneic marrow grafts from DLA-identical littermates corrected cyclic neutropenia (Grey Collie disease)138 and X-linked severe combined immunodeficiency139 with full reconstitution of neutrophil, and T and B cell functions, respectively. Allogeneic HCT studies were also performed in Basenji dogs with haemolytic anaemia because of pyruvate kinase deficiency; these studies showed reversal of iron overload after successful transplantation140 and determined the level of donor chimerism needed to prevent haemolysis.141,142 Using DLA-identical marrow grafts for the treatment of lysosomal storage diseases, such as α-L-iduronidase deficiency,143 mucopolysaccharidosis I,144 and fucosidase deficiency,145 the disease-related pathologies were partially corrected. However, the outcomes in dogs with ceroid lipofuscinosis (Batten’s disease),146 GM1 gangliosidosis,147 haemophilia148 and Duchenne muscular dystrophy149 were not improved following allogeneic HCT. For the treatment of Duchenne muscular dystrophy, stable haematopoietic chimerism induced tolerance in the immune system of recipients to both dystrophin and cells from HCT donors, thereby setting the stage for ongoing investigations of additional therapeutic strategies such as gene and muscle stem cell therapies.

HCT as a therapy for pet dogs with haematological malignancies

Cancer is one of the leading causes of death in dogs, showing a one-log higher incidence than in human beings.150 Among canine cancers, haematological malignancies are rarely curable by radio/chemotherapy.151 Although more than 90% response rates were achievable after various combination chemotherapy regimens,152 remissions were short because of the development of multidrug resistance and recurrence of disseminated disease.151 Other therapeutic alternatives for canine malignant lymphoma, such as the use of high-dose half body irradiation given either alone153 or after preceding remission inductions with chemotherapy,154 led to measurable decreases in tumour size; however, besides radiation-induced side-effects, the median remission durations were not significantly prolonged compared with those achieved following conventional combination chemotherapy.154

The preclinical research developed in dogs and summarized in this review has the potential for improving cancer therapies in pet dogs. Haematopoietic stem cell transplantation can cure canine malignant haematological conditions that have been considered fatal.57,87,136,155,156 Molecular DLA typing methods offer a high-resolution technology for donor selection among single- and multigeneration canine families,63,65,156 which is a prerequisite for lower GVHD rates and improved transplantation outcomes. The canine genome map18,64,157 has contributed to identification of new microsatellite repeat sequences that have extended the panel of markers for evaluation of donor chimerism.158 The availability of safe PBMC mobilizing agents and apheresis devices for dogs offers a minimally invasive alternative for graft collection compared with the traditional bone marrow harvest. Less toxic non-myeloablative HCT regimens have been recently developed in the canine model, which permit graft-versus-tumour effects, while controlling GVHD, and which could be applied to veterinary patients.

The continuous advances in veterinary medical oncology, radiation oncology and transfusion medicine, the expansion of canine blood bank networks and rapid progress in canine immunogenetics, all promise to improve HCT-based therapies for canine patients with cancer. In the clinical veterinary field, the treatment of canine malignant lymphoma in chemotherapy-induced remission by autologous marrow transplantation has already been reported.155 We have recently described the successful treatment of a pet dog with malignant T-cell lymphoma in chemotherapy-induced remission by TBI and allogeneic rcG-CSF-mobilized PBMC transplantation.156

From past to present and future

As a consequence of the numerous experimental studies in random-bred dogs, HCT has advanced from an investigational therapy that was thought to be plagued with insurmountable problems in the 1960s, to an important treatment option that produces amazing clinical results. In 1990, Dr E. Donnall Thomas was awarded the Nobel Prize in Medicine for his pioneering work in marrow transplantation. He emphasized that ‘marrow grafting could not have reached clinical application without animal research, first in inbred rodents and then in outbred species, particularly the dog.’159

The efforts continue in the preclinical canine model to improve the HCT regimens for direct translation to human patients. Research has been directed towards designing specific antibodies against costimulatory molecules with key roles in the development of the immunological reactions after allogeneic HCT.160 Potential antileukaemic effects of donor-derived natural killer cells161 are currently being explored in the non-myeloablative HCT setting. The observation that in some human patients graft-versus-tumour effects are associated with GVHD, whereas in others remission can be achieved without GVHD, has led to investigation of strategies to separate graft-versus-tumour effects from GVHD through adoptive transfer of selected T-cell populations.162 The use of adoptive immunotherapy could make allogeneic HCT more effective in recognizing tumour-specific antigens163 and perhaps extend its application to the treatment of metastatic and non-haematopoietic tumours that arise in organs such as the breast, prostate, pancreas, or colon. In this regard, dogs with spontaneous tumours could become ideal candidates for preclinical trials of antitumour vaccination. Furthermore, research involving the recently recognized immune regulatory T-cells that arise after HCT could lead to a better understanding of mechanisms of graft rejection and GVHD and improve strategies to overcome these immunological barriers.162 Promising preclinical results of kidney grafting in a canine mixed haematopoietic chimerism model164 suggest that HCT may serve in the future as an immunologic platform for solid organ grafting without the need for lifelong immunosuppression. Finally, the approach of transferring cloned genes into the haematopoietic stem cells to correct specific genetic defects may be a promising avenue for the therapy of genetic diseases.165

Not only in human history but also in the history of medicine, the dog has been loyal to human beings. It is now well recognized in the HCT field that the dog has contributed to a legacy that saves thousands of patients annually, proving once again the paradigm of ‘man’s best friend’. The remarkable progress in the clinical applications of preclinical HCT protocols established in dogs and the rapid development of the biological sciences predict sustained advances over the coming years in both the human and veterinary oncology fields.


This work was supported by grants CA78902, HL36444, and CA15704 from the National Institutes of Health, Bethesda, MD; M.L. was supported by an award from the Exchange Scientist Program of the Office of International Affairs, National Institutes of Health.