In vitro culture of bone marrow cells for auto grafting in CML
Eaves, C.J., Barnett, M.J., and Eaves, A.C.     Leukemia Vol 7, Suppl 2

Terry Fox Laboratory and the Leukemia/Bone Marrow Transplantation Program of British Columbia, British Columbia Cancer Agency,
Vancouver General Hospital and the University of British Columbia, Vancouver, B.C.

Long-term maintenance of normal hematopoiesis in vitro is possible when very primitive progenitors are cocultivated with certain non-hematopoietic stromal cells that may co-exist in (or be derived from cells that co-exist in) hematopoietic tissues. Such long-term cultures (L TC) have been used to develop quantitative assays for the most primitive populations of hematopoietic cells currently detectable in adult marrow. In addition they provide a unique model for analysis of the complex molecular mechanisms that may regulate primitive hematopoietic cell population dynamics in vivo. Similar studies with L TC of cells from patients with chronic myeloid leukemia (CML) have made it possible to detect and characterize very primitive neoplastic cell populations in this disease. These latter studies have revealed differences in the properties of primitive CML cells that both reflect and explain their increased turnover and are thus presumably part of the mechanism that enables the neoplastic clone to expand in vivo. In addition, the most primitive neoplastic cells in CML patients are abnormally distributed between the marrow and blood and their ability to maintain their numbers in L TC has also been found to be defective. Assessment of the number and behaviour of primitive cells in L TC of CML marrow has been used to identify those patients most likely to benefit from intensive therapy supported by transplantation of cultured autologous marrow. Twenty-two such CML patients have now been treated with this experimental protocol. The results to date have clearly established the feasibility of this novel treatment strategy and, together with more recent laboratory findings, suggest future avenues for significantly improving the management of CML patients.


Perturbations of the norm frequently provide an opportunity to gain new insights into poorly understood regulatory processes. Where the perturbation has a known genetic basis, this information may serve as a unique clue to the delineation of the underlying molecular mechanism(s) affected. Chronic myeloid leukemia (CML) is a prime example of a disease of perturbed hematopoiesis associated with a consistent and unique genetic abnormality characterized at the molecular level as the creation of a fusion gene involving BCR and ABL.l CML is also of considerable clinical importance as a disease entity since it is inevitably fatal in its acute phase. At the cellular level, CML is recognized to be a clonal, multilineage myeloproliferative disorder arising from the dereglilated proliferation of a pluripotent hematopoictic stem cell and its subsequently amplified progeny' which, by the time of diagnosis, have typically come to dominate the entire hematopoietic sy.stcm.▓ Initially, differentiation processes appear to be relatively' unaffcctcd. Thus early expansion of the CML clone leads to the continued generation of functionally normal mature blood cells altlough the rate of granulocyte and macrophage production may. be increased more than 5O-fold However, on average within 4 years of diagnosis, evolution from this chronic phase to a frank acute leukemia occurs, This change is typically characterized at the genetic level by the acquisition of additional mutations and phenotypically by a breakdown in differentiation leading to a rapid accumulation of non-functional blasts.│ Recent studies of the molecular and cell biology of CML have led to abetter understanding of both norm1al and leukemic hematopoietic stem cell regulation in addition to stimulating the development of new approaches to treatment. This review summarizes some of the progress in these areas that has emerged from a combined laboratory and clinical effort at our centre to improve the management of patients ,with CML using intensive therapy supported by autologous marrow rescue The multilineage nature and likely pluripotent stem cell origin of CML was first suggested by Dameshek long before direct evidence of a normal pluripotent hematopoictic stem cell compartment was obtained.4 This came almost simultaneously about a decade later from two independent lines of ,work. One involved documentation of the multi-lineage differentiation potential of single cell-derived clones generated in the spleens of mice transplanted with small numbers of adult mouse bone marrow cclls5 The second was the recognition of a consistent chromosomal abnormality (the Ph chromosome) in virtually all dividing marrow cells of CML patients, including erythroid cells and megakaryoeytcs6.7 Formal proof that these Ph chromosome-positive cells represented expanded clonal populations was subsequently provided following the development of methodology for analyzing the activity of X-Iinked genes in hematopoietic cells from hetcrozygous females ,with CML▓ These types ofapproaches have now been extended to show that the neoplastic clone may also conmmonly include Band T lymphocytes. as well as members of all of the mycloid lincages.8-10 With the introduction of reproducible in vitro assays for quantitating human progenitors on different hematopoietic pathways and at different stages of differentiation, II and the development of procedures for obtaining cytogenetic data from individual colonies to distinguish their leukemic or norm1al origin,12 considerable information about early leukemic cell compartments has accrued rapidly. For example, it is now well established that, at the level of cells detectable as in vitro colony'-form1ing cells, Ieukemic progenitor numbers on all myeloid lineages appear, on average, to be expanded equally relative to one another and their cycling control is also deregulated in a lineage non-specific fashion. These findings are of interest since, prior to initiation of treatment, platelet counts are usually increased to a much lesser extent that the WBC count and the hematocrit is frequently lower than normal (reviewed in Ref 13). These findings imply that the BCR-ABL gene is active at early. stages of hematopoietic cell differentiation, but in a fashion that this stage does not discriminate between lineages nor affect early' commitment events per se. Ho,w an ovcrproduction of all ty'pcs of early progenitors is translated in vivo into an overproduction of mature cells, to a large extent only on the granulopoietic pathway. is not known. Nevertheless, the large numbers of early types of Ph-positive cells on all lineages explains ,why, it is usually. difficult to detect any' norm1al hematopoietic cells 1ntercstingly. analysis of progenitors from newly diagnosed CML patients with small tumor burdens has shown that normal numbers of normal progenitors are present.14 Thus it could be inferred, as was eventually shown, that even more primitive types of norm1al hematopoietic cells are also likely to be present in substantial numbers in many CML patients. 15,16


Use of the Long-term Culture (L TC) System to Analyze Early Events in Normal and Leukemic Hematopoiesis.

Three features of hematopoiesis in marrow L TC have allowed the unique application of this system to the study of early stages of hematopoiesis. The first is the fact that the system is a dynamic one in which cell proliferation, differentiation and death occur continuously for many weeks.17 Thus, eventually, all of the differentiated hematopoietic cells present should be derived from a very primitive cell type present in the original input suspension. Moreover, if the supportive "stromal " elements required (which are derived from precursors of the fibroblast-adipocyte-endothelial lineages), are provided independently in non-limiting numbers, then after an appropriate interval, the number of differentiated hematopoietic cells present might be expected to be quantitatively related to an input population of very! primitive hematopoietic precursors. In the case of human cells, the number of in vitro colony-forming cells present in L TC initiated under these conditions and assessed after a minimum interval of 5 weeks has been found to be a suitable endpoint for the measurement of an input cell type with properties shared exclusively with long-term in vivo repopulating cells.18,19 Because of the assay. procedure used for their detection and quantitation, these cells are referred to as LTC-initiating cells or L TC-IC. We have shown that the relationship between input L TC-IC and their 5 week clonogenic cell output is linear down to limiting numbers of L TC-IC seeded into the assay cultures. It is therefore possible to use limiting dilution analysis techniques to derive absolute frequencies of L TC-IC and hence to ascertain the output characteristics of individual LTC-IC.18,20 The second feature of the L TC system is that the more primitive types of hematopoietic cells localize within and tend to remain in the adherent layer. This latter fraction of the culture also contains the stromal cells essential for L TC-IC support,21,22 Nevertheless, once the adherent layer is established, the majority of the primitive hematopoietic cells in it are normally maintained in a quiescent state unless the cultures are perturbed by' the addition of specific factors (or fresh horse serum) that activate fibroblasts (and endothelial cells). This leads to the activation of primitive hematopoietic progenitors which is then follow\'ed by' their spontaneous return to a non-cycling state a few days later unless the cultures are again perturbed. The ability to up- and down-regulate primitive progenitor cycling in this way can be repeated for many weeks23 The LTC system has thus provided a convenient model for identifying both stimulating and inhibitory. factors that may be involved in the stromal cell-mediated control of primitive hematopoietic progenitor proliferation.24 A third important feature of the L TC system is its ability' to support not only the production of mature cells from very primitive hematopoietic precursors (L TC-IC) but also the self-renewwal and maintenance of cells with in vivo repopulating potential.25,26 Thus the L TC system should also prove useful for identifying molecular species that regulate this key stem cell function, Assessment of the behaviour of CML cells in L TC has allowed delineation of a number of aspects of early progenitor cell behaviour that are, or are not, perturbed by the neoplastic process, These are summarized in Table I, Of note was the finding early on that the deregulated cycling activity previously documented for primitive Ph-positive colony-forming cells in vivo is reproduced in the L TC system.21 Subsequent studies failed to detect evidence of an underlying autocrine or paracrine mechanism to explain the abnormal cycling behaviour of primitive CML cells both in vivo and in vitro,27 Additional experiments suggested that these cells are normal in their responsiveness to the growth inhibitory' effects of TGF -▀, even under varying conditions of stimulation28 These unexpected findings posed a significant challenge to current models of the molecular mechanisms thought to control early hematopoietic cell turnover and led to the concept of co-operative inhibition, i,e", a mechanism in which the effects of two inhibitors might be additive or even synergistic, Thus one might envisage an additive effect between TGF-▀ and (an)other endogenous inhibitor(s) in the L TC system, neither of which at the levels prevailing in unperturbed L TC, would alone be sufficient to arrest the cycling of primitive normal cells. According to such a model, insensitivity of CML cells to only one component of this co-operating inhibitory' mechanism would then give the observed picture of deregulated cycling

Table 1. Properties of Primitive Ph-positive Populations as Revealed by L TC Studies.

Interestingly, the results of initial studies of the role of MIP-I alfa in the LTC system bear out the predictions of such a model of co-operative inhibition.29 Whether MIP-I alfa ( or other . similarly-acting, factors) normally plays a co-operating inhibitory role in vivo is, of course, still a matter of speculation. However, some support for this possibility has recently been provided by studies in .mice demonstrating predicted effects of appropriately timed administration of MIP-alfa on primitive hematopoietic cell sensitivities to cycle-active drugs.30,31 From the point of view of understanding CML, the hypothesis that MIP-I a is a physiologically relevant negative regulator of primitive normal cells (to which CML cells do not respond) is particularly attractive because it offers an explanation for the deregulation of primitive leukemic cell turnover that is seen in CML patients. Further analysis of how primitive CML cells are able to circumvent MIP-1alfa effects may thus help to pinpoint the molecular mechanisms leading to clonal dominance and hence possibly. to disease progression. Two other points listed on Table I also deserve further comment. One pertains to evidence of a deregulation of proliferation control that is seen in CML at the level of the most primitive neoplastic cells detectable, .i.e., leukemic L TC-IC. It is not difficult to imagine that alterations in mechanism(s) thought to restrict the turnover of primitive normal colony-forming cells might also apply. to leukemic L TC-IC. On the other hand, stage-specific mediators of positive (stimulatory) effects are well documented.22,32 It is thus conceivable that a similar principle might apply to the target cell specificity of different inhibitory. cytokines Also noteworthy is the fact that a substantial proportion of primitive leukemic cells do not show features of activated cells,33 Additional studies will be required to establish unequivocally \whether or not these represent a subpopulation of quiescent primitive leukemic cells and. if so. the mechanisms responsible for such heterogeneity in their c\.cling control. Finally, the self-maintenance of CML L TC-IC in L TC has been found to be highly defective by comparison to normal L TC-IC of either blood or marrow origin, even when these are cultured on normal marrow adherent feeder layers,16 At present there is no direct information as to why this rapid decline of CML L TC-IC in vitro occurs. It is clearly not a technical artifact due simply to the removal of a larger proportion of less adherent leukemic cells when the cultures are fed, as the effect is most dramatic during the first 10 days in L TC prior to any manipulation of the cells It is inviting to speculate that the defective self-maintenance exhibited by leukemic L TC-IC may, at least in part, be secondary to a previous longstanding increased turnover rate of the leukemic L TC-IC population in vivo. However. another possibility is that this represents a more direct and immediate action of the BCR-ABL gene product on intracellular pathways that regulate L TC-IC self-renewal probabilities. Since these alternatives are testable by a number of strategies, it should be possible to resolve this issue in future experiments .

Potential of L TC to Improve the Treatment of CML Patients Using Intensive Therapy with Autologous Bone Marrow Rescue

The biologic selection in vivo against the accumulation of leukemic L TC-IC in the marrow of CML patients \with preferential retention of normal L TC-IC (Table I) provides a rationale for the use of autologous marrow transplants to allow the administration of potentially curative myeloablative treatment regimens. In addition to this naturally occurring benefit can be added a number of in vitro selection strategies. Unfortunately, those that exploit differences likely to be related to an altered proliferative status of primitive leukemic cells, e.g., increased expression of HLA-OR, higher forward light scattering characteristics, increased retention of rhodamine, and increased sensitivity to 4-hydroperoxycclophosphamide ( 4-HC), also have such a small selective potential that their clinical usefulness seems dubious.33 At present, the most significant in vitro purging effect has been obtained by incubating CML marrow under LTC conditions for 10 days.16 Under these conditions leukemic L TC-IC numbers drop 30-fold whereas normal L TC-IC remain at input levels. As a result, for patients whose initial marrows already contain readily detectable frequencies of normal L TC-IC (>2% of normal values) and relatively fewer leukemic L TC-IC, a theoretically attractive autograft. can be obtained by incubating the marrow in LTC for 10 days prior to transplantation In Vancouver, 22 CML patients found to fit into such a group have been subsequently treated \with intensive therapy and transplantation of a cultured marrow autograft. The most dramatic finding in these patients has been the consistent recovery of normal hematopoiesis within the first 2-7 weeks \with a fe\w exceptions (6 patients). usually (5 patients), because the graft appeared to have been inadequate. Of the 15 patients \who were transplanted \with cultured marrow in first chronic phase, 13 are alive and 8 are in hematological remission up to 5 years later In the remaining 7 patients who were transplanted \with more advanced disease survival, although poorer (3 alive), is still encouraging However. late reappearance of some Ph-positive cells has also been a consistent, albeit late, finding and additional strategies appear required to eliminate these cells 11in this regard. the use of interferon post-autografting seems to offer some potential. Clearly. more work remains to refine and evaluate the role of culture purging in the development of a curative treatment. The present findings do, however, underscore the principle of developing new therapies based on careful scientific investigations and highlight the possibility of new avenues on the horizon for significantly improving the management of C M L patients


The work described in this review was supported in part by the National Cancer Institute of. Canada (NCIC) and the British Columbia Health Research Foundation. C.J Eaves is a Terry Fox Research Scientist of the NCIC. We also thank H, Calladine for manuscript preparation and the numerous clinical and support staff of the British Colombia Cancer Agency and Vancouver General Hospital who assisted in patient care and data collection


1. Groffen J, Heisterkamp N In. Goldman JM. Bailliere's Clinical Haematology'. Chronic myeloid leukaemia.
Voi I London; Bailliere Tindall, 1987; p. 983.

2.Raskind WH, Fialkow PJ. Adv Cancer Res 1987:49127

3. Hagemeijer A. In. Goldman JM. Bailliere's Clinical Haematology. Chronic myeloid Ieukaemia Vol I. London: Bailliere Tindall, 1987; p. 963.

4.Dameshek W. Blood 1951 :6372.

5.Till JE, McCulloch EA Radiat Res 1961;14.213

6.Tough IM, Jacobs PA, Brown WMC, Baikie AG, Williamson ERD. Lancet 1963; 1 :844.

7.Whang J, Frei III E, Tjio JH, Carbone PP, Brecher G. Blood 1963;22:664.

8.Bernheim A, Berger R, PreudHonmme JL, Labaume S, Bussel A, Barot-Ciorbaru R. Leuk Res 1981;5:331.

9.Martin PJ, Najfeld V, Hansen JA, Penfold GK, Jacobson RJ, Fialkow PJ. Nature 1980;287:49.

10. Jonas D, Lubbert M, Kawasaki ES et al. 1992;79:1017.

11. Suther1and HJ, Eaves AC, Eaves CJ. In: Gee AP. Bone Marrow Processing and Purging: A Practical Guide.
Boca Raton; CRC Press Inc, 1991; p. 155.

12. Fraser C, Eaves CJ, Kalousek DK. Cancer Genet Cytogenet 1987;24: 1.

13. Eaves CJ, Eaves AC. In: Goldman JM. Bailliere's Clinical Haematology. Chronic Myeloid Leukaemia.
Vol 1. London; Bailliere Tindall, 1987; p. 931.

14. Kalousek DK, Eaves CJ, Eaves AC. 1984;3:439.

15. Coulombel L, Kalousek DK, Eaves CJ, Gupta CM, Eaves AC. N Engl J Med 1983;308: 1493.

16. Udomsakdi C, Eaves CJ, Swolin B, Reid DS, Barnett MJ, Eaves AC. Proc Natl Acad Sci US A 1992;89:6192.

17. Eaves CJ, Cashman JD, Eaves AC. J Tissue Culture Methods 1991; 13:55.

18. Sutherland HJ, Lansdorp PM, Henkelman DH, Eaves AC, Eaves CJ. Proc Natl Acad Sci US A 1990;87.3584.

19. Eaves CJ, Sutherland HJ, Udomsakdi C et al. Blood Cells 1992;18:301.

20. Udomsakdi C, Lansdorp PM, Hogge DE, Reid DS, Eaves AC, Eaves CJ. Blood 1992;80:2513.

21. Eaves AC, Cashm11an JD, Gaboury. LA, Kalousek DK, Eaves CJ. Proc Natl Acad Sci US A 1986;83:5306.

22. Sutherland HJ, Eaves CJ, Lansdorp PM, Thacker JD, Hogge DE. Blood 1991;78:666.

23. Cashman J, Eaves AC, Eaves CJ. Blood 1985;66.1002.

24. Eaves CJ, Cashman JD, Sutherland HJ et al. Ann N Y Acad Sei 1991;628:298.

25. Bamett MJ, Eaves CJ, Phillips GL et al. Transplant 1989;4:345.

26. Fraser CC, Szilvassy SJ, Eaves CJ, Humphries RK Proc Natl Acad Sci US A 1992;89:1968.

27. Otsuka T, Eaves CJ, Humphries RK, Hogge DE, Eaves AC. Leukemia 1991;5:861.

28. Cashman JD, Eaves AC, Eaves CJ. Leukemia 1992~6.886.

29. Eaves CJ, Cashman JD, Wolpe SD, Eaves AC Blood 1992;80 Suppll:155a.

30. Dunlop DJ, Wright EG, Lorimore Set al. 1992;79:2221.

31. Lord BI, Dexter TM, Clements JM, Hunter MA, Gearing AJH. Blood 1992;79:2605.

32. Leary AG, Zeng HQ, Clark SC, Ogawa M. Proc Natl Acad Sci US A 1992;89:4013.

33. Udomsakdi C, Eaves CJ, Lansdorp PM, Eaves AC. Blood 1992;80:2522.