1 Dept. of Genetics, Research Institute, Hospital for Sick Children and Dept. of Molecular and Medical Genetics, University of Toronto, 555 University Ave., Toronto, Ontario, Canada M 5G 1 X8.

Introduction

Over the last 40 years the hematopoietic system has provided many of the important paradigms that guide our understanding of stem cell function. Much of our knowledge of the regulation of the hematopoietic system is derived from experiments in the mouse; these studies have involved identification of various classes of progenitor cells, growth factors that stimulate growth and differentiation, and molecular events that underlie the abnormalities that occur in diseases such as leukemia. This information has derived largely from the development of in vivo transplantation assays for normal stem cells and the ability to establish and grow leukemic cells in vitro and in vivo [1]. In contrast, our understanding of the biology of the human hematopoietic system has suffered relative to that in the mouse because of the lack of similar assays for normal stem cells and leukemic cells. Normal and leukemic human cells often appear to have complex growth factor requirements that are not easy to provide in short- or long-term cultures. Furthermore, the difficulties in growing primary human leukemic cells in culture suggest that there are selective processes that may result in alterations of the properties of such cells over time and the resultant cell lines do not accurately re flect the original disease. In an attempt to develop in vivo animal models for human leukemic cells, a large body of literature has accumulated over the past 20 years on the growth of human tumor xenografts in immune-deficient nude mice [2]. However, the growth of human leukemic cells as an ascites or solid subcutaneous tumor in nude mice does not reflect the normal course of the disease in humans. In addition to leukemic cells, normal human hematopoietic cells have also been introduced into nude mice directly or in diffusion chambers. The transplantation of human bone marrow directly into mice generally yielded inconclusive results [3], while the implantation of diffusion chambers demonstrated the development of human progenitors for as long as 28 days in vivo although it was not possible to distinguish between persistence of progenitors and engraftment of stem cells [4]. The recently described approaches to engraft human cells into two novel strains of immune-deficient mice provide the foundation for in vivo assays to characterize the normal developmental program of human hematopoietic stem cells along the myeloid and lymphoid lineages [5, 6, 7] and to develop models of several human hematopoietic diseases. Already, important preliminary experiments have established the feasibility of such models for leukemia [8], cancer [9,10,11], infectious diseases [12, 13, 14], and autoimmunity [15,16]. Here I will briefly review our attempts to develop animal models for normal and leukemic human hema topoiesis .

Engraftment of Human Bone Marrow

Successful engraftment of human cells into mice is dependent on the development of two particular strains ofimmunedeficient mice; scid and bg/nu/xid (abbreviated bnx). The scid mouse has already played an important role in the characterization of the murine lymphoid system because it is an ideal recipient for transplantation experiments designed to detect lymphoid reconstitution without the complications of lethal irradiation [17]. The scidmutation, carried on mouse chromosome 16, prevents the production of nature T and B cells; all other hematopoietic lineages including stem cells, myeloid progenitors, B cell precursors, and natural killer (NK) cells are normal [18]. The action of this gene product is not restricted to normal lymphoid development since scid mice have a generalized radiation repair defect that renders the animals at least two times more sensitive to the effects of y-radiation [19]. In spite of these defects, homozygous scid mice have a normal life span and in all other respects appear normal although they are highly susceptible to infection. The underlying defect and origin of the bg/nu/xid mouse is quite different to that of the scid mouse. I t was constructed by combining three recessive mutations, beige, nude, and xid, to generate a potential recipient mouse for human tumor xenografts that was more immunedeficient then the parental nude mouse [20]. The nude mutation affects the development of the thymic epithelium, thereby preventing T cell differentiation resulting in an athymic mouse; all other lineages appear normal. I t should be noted that some extrathymic processes lead to low levels of mature T cells especially in mice stimulated by exposure to antigens. The bg mice have a deficiency in cytotoxic T cells and NK cells [21]. The mutation is a lysosomal storage defect producing abnormally large cytoplasmic granules, so the effect on the immune system is indirect. The xid defect is an X chromosome linked gene that affects lymphokine activated killer (LAK) cells as well as B cell responses to certain thymus independent antigens [22]. Animals in which the nude and xid mutations are combined are very deficient in B cells that appear to be blocked in development at some point between pro-B cells and before cytoplasmic immunoglobulin expression [23]. In the particular outbred bg/nu/xid mouse used in our experiments, the bg mutation is not as active as if it was alone and the mice have only slightly reduced levels of NK cells, although they are deficient in LAK activity presumably by the action of the xid gene. Although the exact relationship between NK and LAK cells is unknown, both may display antitumor activity and appear to play an important role in host resistance to xenografting [24, 25]. It is important to stress that both scid and bg/nu/xid mice have NK cells and intact non-lymphoid resistance systems such as macrophages, therefore any environmental exposure to antigens stimulate these resistance mechanisms. We have found that animals with even mild subclinical infections are extremely resistant to xenografting [8]. In our initial experiments, human interleukin-3 (IL-3) and granulocyte macrophage-colony-stimulating factor (GM-CSF) were delivered to the animals by implanting an osmotic minipump subcutaneously. The animals were also given sublethal doses of radiation since syngeneic transplantation is faster and more complete if the recipient animals are conditioned with irradiation or chemotherapy prior to transplant. Animals were killed at various lengths of time after transplant and the bone marrow and spleen were analyzed by molecular techniques using human specific probes to determine whether human DNA was present; human cells comprised approximately 0.1%-1.0% of these tissues [6]. Since the goal of these experiments was to determine whether any of the earlier hematopoietic cell types had engrafted the mice, cells from these tissues were plated in in vitro progenitor assays that were selective for the growth of human colony-forming unit -granulocyte macrophage (CFU-GM). Significant numbers of human progenitors were detected in the spleen and bone marrow. N o human DNA or progenitors were detected in wild-type animals prepared the same way and only very low numbers of human progenitors could be detected in the hematopoietic tissues of scid mice transplanted with human bone marrow. There did not appear to be any dif ference in the level or speed of engraftment in the presence or absence of exogenously added human growth factors. This applies only to the two factors tested and on the level of CFU-GM progenitors. It remains to be seen whether other combinations are more effective and whether the presence of these or other factors affect the differentiation of progenitors into mature cell types. There are several explanations for successful engraftment in the absence of exogenous human factors. The human cells could be secreting their own growth factors or the cells are responding to some crossreactive murine factor or the murine hematopoietic microenvironment. We have recently found that some human factor-dependent myeloid leukemic cell lines proliferate in immune-deficient mice (C. Sirard and J. E. Dick, unpublished) and others have reported the stimulation of human lymphoid leukemic cell lines on murine stromal cells [26], lending credence to the idea that the murine environment is capable of stimulating human hematopoietic cells. While these data clearly showed that human progenitors can engraft mice, the more important question was whether human stem cells had also engrafted. This is not an easy question to anwer because stem cells can only be identified by their function; this includes the ability to differentiate into all lineages, high self renewal capacity, slow cell cycling, ability to engraft for long periods of time, etc. [1]. Two lines of evidence suggest that a cell type at least earlier than CFU-GM is responsible for maintaining the engrafted cells. Examination of the kinetics of engraftment indicated there was a rapid increase, of at least 40-fold, in the number of CFU-GM during the first 14 days of engraftment. CFU-GM have a very low self-renewal capacity and would not be able to generate such a large increase, suggesting that an earlier cell type is responsible for this large increase. Furthermore, human progenitors have been detected in animals 7 months after transplantation. Finally we have preliminary evidence that human bone marrow, highly enriched for earlier cell types, can engraft the mice. Taken together this suggests that some earlier cell type can engraft immune-deficient mice, although the exact nature of this cell type is unknown. Gene transfer provides a powerful tool to more conclusively characterize the engrafting cell types. This stem cell marking technology has been powerfully applied to elucidate the stem cell hierarchy in the mouse [27] and a similar strategy in the human system should enable more precise identification of the human cell types responsible for engrafting the mice. Human bone marrow was infected with a retrovirus vector that contained the dominant selectable neo gene using the optimized conditions [31]. Infected and preselected cells were transplanted into immune-deficient mice. A large proportion of the human progenitors detected in the bone marrow for at least 4 months after transplantation contained the retrovirus. In combination with polymerase chain reaction (PCR) technology to clone out the virus integration site in small numbers of cells [29], these gene transfer experiments lay the foundation to determine whether the different progenitors engrafting the mice arise from a common progenitor. The engraftment of mice with both myeloid and lymphoid cells should permit the detection of a pluripotent human stem cell using these approaches.

Models of Human Leukemic Diseases

The growth of normal human hematopoietic cells in scid or bg/nu/xid mice suggested that these mice may also be useful recipients in which to grow human leukemic cells that often are difficult to establish in culture and to provide a system to study the growth of human leukemia in vivo. Non-T acute lymphoblastic leukemia (ALL) is the most prevalent childhood leukemia and is characterized by a pre-B cell phenotype [30]. Our initial studies have focussed on bone marrow taken directly from patients with non- T ALL and cell lines recently established from relapse patients. One such cell line (A-l) is Epstein~Barr virus (EBV) free, has a normal karyotype, and grows autonomously, producing an unidentified factor which augments its growth in semi-solid clonogenic assays and suspension cultures. Not only did A-l grow in scid mice transplanted according to the same procedures as normal bone marrow but they showed a pattern of infiltration reminiscent of that observed in many children with ALL [8]. For the first month post-transplant, the A-l cells could be detected only in the bone marrow, and only after an additional 8 weeks were large numbers ( > 75 % ) of leukemic cells present in the spleen and bone marrow. Small infiltrates were also present in the kidney and liver at this time. High numbers of A-l cells were found in the peripheral circulation several weeks later, coinciding with widely disseminated leukemic infiltrates in many organs, including the brain. The animals began to die at 12 weeks after transplant. The widely disseminated growth particularly in the central nervous system is a feature of the terminal disease in children. Celllines with different growth factor requirements proliferate differently in scid mice. Another cell line (G-2) which responds to different growth factors than A-l and grows slower in culture, proliferates extremely rapidly in mice, killing animals in 6 weeks at comparable cell doses to A-l. Limiting dilution experiments indicate as few as lOO G-2 cells will produce leukemic growth in scid mice. An additional observation has been the infiltration of G-2 into the thymic remnant present in scid mice and elevated levels of the CALLA antigen on the cells that infiltrated the thymus. CALLA is a differentiation antigen expressed on early human B cells and is a marker for non- T ALL. It was expressed on the cells originally obtained from the patient but was completely lost during in vitro cell culture. Independent clones and limiting dilution experiments all suggest that CALLA expression is being modulated in vivo as opposed to selection of rare positive cells in the population. The growth and modulation of differentiation antigens of non- T ALL cells in immunedeficient mice implies that they are responding to some murine growth factor or microenvironmental influence. Interestingly, Gluck et al. found that murine stromal cells can support the growth of factor dependent non- T ALL cell lines, lending further credence to this idea [26]. The ability to engraft bone marrow directly from patients with leukemia, either before or after treatment, into scid mice could be a valuable tool for predicting the clinical course of the disease, detecting residual leukemias, and for developing individualized therapeutic strategies. Toward this objective, bone marrow cells taken directly from patients at diagnosis with non-T ALL were injected into scid mice. Samples taken directly from patients at diagnosis grew little or not at all in the bone marrow and spleen of scid mice even after 8 months of observation. In contrast, all of the bone marrow samples taken from patients with recurrent disease in first or second relapse proliferated extensively into a widely disseminated leukemia in mice. Cells from patients who have relapsed several times and who are not responding well to therapy were the most aggressive in mice with the shortest latency. These data show that some biological parameters associated with poor clinical outcome, such as rapid and widely disseminated proliferation, can be reproduced in SCID mIce. In addition to lymphoid cells, acute myeloblastic leukemia (AM1) and chronic myelocytic leukemia (CM1) cell lines also grow in immune-deficient mice (C. Sirard and J. E. Dick, unpublished). Some of the well-established lines such as K 562 which do grow in nude mice also grew very rapidly in scid mice. As noted earlier, one of the AM1 cell lines (MO7E) which is human growth factor (11-3, GM-CSF) dependent grew extensively in the bone marrow of scid mice without an exogenously added factor . The establishment of an in vivo model for human leukemia presents a unique system in which to address experimentally a number of biological questions governing the clinical outcome and the growth of leukemic cells in vivo. For example, the identification of leukemic cells in bone marrow usually dictates the course of chemotherapy; however, histological methods for detecting low numbers of residual cells are neither sensitive nor precise. Further refinement of the animal model could offer a sensitive method by which to study residual cells from patients undergoing chemotherapy. Furthermore, new chemotherapeutic and immunotherapeutic protocols, combinations of biological response modifiers, or new unconventional therapies that are difficult to develop and evaluate by human experimentation can be tested in an in vivo situation which mimics the progression of human leukemia. Using high-efficiency gene transfer technology, individual leukemic cells can be marked to follow the growth and development of clones during the multistage progression of the disease. Gene transfer in conjunction with this model system should allow the introduction of key growth regulatory genes, such as oncogenes or tumor suppressor genes, into normal human bone marrow, to determine how their aberrant expression affects normal hematopoiesis and leukemic transformation and progression. CM1, in particular, is a good candidate for these gene transfer experiments because some of the genes involved in its etiology (e.g., bcr/abl) and progression (e.g., p 53) have been identified. It should be possible to directly test the role of these genes in the multistage progression of CM1 using gene transfer with the long-term goal of developing an animal model for CM1.

References

1. Till JE, McCulloch EA (1980) Hemopoietic stem cell differentiation. Eiochim Eiophys Acta 605:431-459

2. Fogh J, Fogh JM, Orfeo T (1977) One hundred and twenty-seven cultured human tumor cell lines producing tumors in nude mice. J Natl Cancer Inst 59:221-226

3. Earr RD, Whang PJ, Perry S (1975) Hemopoietic stem cells in human peripheral blood. Science 190:284-285

4. Pojda Z, Szczylik C, Wiktor-Jedrzejczak W (1987) Multiple lineage colony growth from human marrow in plasma clot diffusion chambers. Exp Hemato115:922-927

5. McCune JM, Namikawa R, Kaneshima H, Shultz LD, Lieberman M, Weissman IL (1988) The SCID-hu mouse: murine model for the analysis of human hematolymphoid differentiation and function. Science 241 : 1632-1639

6. Kamel-Reid S, Dick JE (1988) Engraftment of immune-deficient mice with human hematopoietic stem cells. Science 242:1706-1709

7. Mosier DE, Gulizia RJ, Eaird SM, Wilson DE (1988) Transfer of a functional human immune system to mice with severe combined immunodeficiency. Nature 335: 256-259.

8. Kamel-Reid S, Letarte M, Sirard C, Doedens M, Grunberger T, Fulop GM, Freedman MH, Phillips RA, Dick JE (1989) A model of human acute lymphoblastic leukemia in immune-deficient scid mice. Science 246: 1597 -1600

9. Reddy S, Piccione D, Takita H, Eankert RE (1987) Human lung tumor growth established in the lung and subcutaneous tissue of mice with severe combined immunodeficiency. Cancer Res 47:2456-2460

10. Bankert R, Umemoto T, Sugiyama Y, Chen F, Repasky E, Yokota S (1989) Human lung tumors, patient's peripheral blood lymphocytes and tumor infiltrating lymphocytes propagated in scid mice. Curr Top Microbiol Immuno1152:201-210

11. Cannon MJ, Pisa P, Fox RI, Cooper NR (1990) Epstein-Barr virus induces aggressive lymphoproliferative disorders of human B cell origin in SCID /hu chimeric mice. J Clin Invest 85: 1333 -1337

12. Namikawa R, Kaneshima H, Lieberman M, Weissman IL, McCune JM (1988) Infection of the SCID-hu mouse by HIV1. Science 242: 1684-1686

13. McCune JM, Namikawa R, Shih CC, Rabin L, Kaneshima H (1990) Suppression of HIV infection in AZT -treated SCID-hu mice. Science 247:564-566

14. Mosier D, Gulizia R, Baird S, Spector S, Spector D, Kipps T, Fox R, Carson D, Cooper N, Richman D, Wilson D (1989) Studies of HIV infection and the development of Epstein-Barr virus-related B cell lymphomas following transfer of human lymphocytes to mice with severe combined immunodeficiency. Curr Top Microbiol Immuno1152: 195-199

15. Krams SM, Dorshkind K, Gershwin ME (1989) Generation of biliary lesions following transfer of human lymphocytes into scid mice. J Exp Med 170: 1919

16. Duchosal M, McConahey P, Robinson C, Dixon F (1990) Transfer of human systemic lupus erythematosus in severe combined immunodeficient (SCID) mice. J Exp Med 172:985-988

17. Bosma GC, Custer RP, Bosma MJ (1983) A severe combined immunodeficiency mutation in the mouse. Nature 301: 527 530

18. Dorshkind K, Pollack SB, Bosma MJ, Phillips RA (1985) Natural killer (NK) cells are present in mice with severe combined immunodeficiency (scid). J Immuno1134:3798-3801

19. Fulop a, Phillips R (1990). The scid mutation in mice causes a general defect in radiation repair. Nature 347:479-482

20. Andriole a, Mule J, Hansen C, Linehan W, Rosenberg S (1985) Evidence that lymphokine-activated killer cells and natural killer cells are distinct based on an analysis of congenitally immunodeficient mice. J Immuno1135:2911-2913

21. Roder J (1979) The beige mutation in the mouse. I. A stem cell predetermined impairment in natural killer cell function. J Immuno1123:2168-2173

22. Scher I, Steinberg AD, Berning AK, Paul WE (1975) X-linked B-lymphocyte immune defect in CBA/N mice. II. Studies of the mechanisms underlying the immune defect. J Exp Med 142:637-650

23. Karagogeos D, Rosenberg N, Wortis HH (1986) Early arrest of B cell development in nude, X-linked immune-deficient mice. Eur J Immuno116:1125-1130

24. Fodstad 0, Hansen CT, Cannon GB, Statham CN, Lichtenstein oR, Boyd MR (1984) Lack of correlation between natural killer activity and tumor growth control in nude mice with different immune defects. Cancer Res 44:4403-4408

25. Talmadge J, Meyers K, Prieur D, Starkey J (1980) Role of NK cells in tumour growth and metastasis in beige mice. Nature 284: 622- 624

26. Gluck U, Zipori D, Wetzler M, Berrebi A, Shaklai M, Drezen 0, Zaizov R, Luria D, Marcelle C, Stark B, Umiel T (1989) Longterm proliferation ofhuman leukemia cells induced by mouse stroma. Exp Hematol 17:398-404

27. Dick JE, Magli MC, Phillips RA, Bemstein A (1986) Genetic manipulation of hematopoietic stem cells with retrovirus vectors. Trends Genet 2: 165 -170

28. Laneuville P, Chang W, Kamel-Reid S, Fauser AA, Dick JE (1988) Highefficiency gene transfer and expression in normal human hematopoietic cells with retrovirus vectors. Blood 71: 811-814

29. Frohman MA, Dush MK, Martin oR (1988) Rapid production of full-length cDNAs from rare transcripts: Amplification using a single gene-specific oligonucleotide primer. Proc Natl Acad Sci USA 85:8998-9002

30. Greaves MF (1986) Differentiation-linked leukemogenesis in lymphocytes. Science 234:697-704

31. Dick JE, Kamel-Reid S, Murdoch B, Doedens M (1991) Gene transfer into normal human hematopoietic cells using in vitro and in vivo assays. Blood 78: 1-11