Stem Cells, Vol. 17, No. 1, 9-18,
January 1999
© 1999 AlphaMed Press
Relapse in Chronic Myeloid Leukemia after Bone Marrow Transplantation: Biomathematical Modeling as a New Approach to Understanding Pathogenesis
P.C. Vincenta,
L. Rutzen-Loesevitzb,
B. Tibkenc,
B. Heinzeb,
E.P. Hoferc,
T.M. Fliednerb
a The Kanematsu Laboratories, Royal Prince Alfred Hospital, Camperdown, Australia;
b Department of Clinical Physiology, Occupational and Social Medicine;
c Department of Measurement, Control and Microtechnology, University of Ulm, Ulm, Germany
Key Words. Biomathematical modeling • Chronic myeloid leukemia • Relapse • Ph-chromosome • Bone marrow transplantation
Prof. Dr. med. Dr. h.c. mult. Theodor M. Fliedner, Radiation Medicine Research Group, Medical Center of the University of Ulm, Helmholtzstrasse 20, D-89081 Ulm, Germany.
 |
Abstract
|
|---|
A biomathematical model was developed to simulate relapse development in patients with chronic myeloid leukemia (CML) following bone marrow transplantation (BMT). The purpose of this study was to better understand the pathophysiology of the time evolution of CML relapse and to provide means whereby the outcomes of patients with CML relapse can be projected and treatment modified accordingly. The model consists of three parallel series of catenated compartments representing granulopoiesis in normal (donor) cells from the marrow, in CML cells from the marrow, and in CML cells from extramedullary sites. It was assumed that CML stem cells were resistant to feedback control and that CML-derived neutrophils, as well as normal neutrophils, exercised feedback regulation of normal stem cells. The known longer generation times for CML neutrophil precursors compared with normal neutrophil precursors were used, and it was assumed that 107 pluripotential stem cells were infused with BMT.
The model was evaluated for its ability to simulate the reappearance of CML (Philadelphia chromosome positive) metaphases in the marrow and the recovery pattern in the blood neutrophil count in six patients who had relapsed following BMT (allogeneic in three patients, allogeneic with T-cell depletion in two patients, and syngeneic in one patient). The variables tested included the site of origin of the CML stem cells responsible for relapse (marrow alone versus marrow and extramedullary sites), the minimum number of CML stem cells responsible for relapse, and the time delay between BMT and the onset of relapse.
Model profiles based on the observed values were obtained in each case. The simulations pointed to the fact that relapse began from a small number of CML cells in medullary and extramedullary sites. The time delay between BMT and the onset of relapse varied from 15 to 240 days.
We suggest that this biomathematical model should be further investigated as a possible means of predicting outcome and guiding the treatment for patients with CML relapsing after BMT.
|
Introduction
|
|---|
Chronic myeloid leukemia (CML) is a clonal disorder characterized in more than 90% of cases by the presence of the Philadelphia (Ph) chromosome in all hematopoietic cells. The reciprocal translocation between chromosomes 9 and 22 that generates the bcr-abl fusion gene occurs in a single cell which, over a period of time, proliferates and ultimately comes to replace normal hematopoietic cells. Little is known, however, of the proliferative phase in CML that predates clinical presentation of the disease. In particular, it is not clear why Ph-positive cells come to replace almost completely the normal hematopoietic cells, since CML cells do not proliferate more rapidly than normal cells [1-4]. It is well known that the Ph transformation occurs in a pluripotential stem cell and that the cytogenetic abnormality is found in all hematopoietic cells at clinical presentation [5]. All of these cells belong to the malignant clone, and their reappearance signals recurrence of the disease.
Allogeneic bone marrow transplantation (BMT) is considered by many to be the treatment of choice for young (<45 years) patients suffering from CML [6]. A reconstitution of normal (donor) hematopoiesis is possible, and a cure can be achieved in 50% to 60% of cases. Studies using the reverse-transcriptase polymerase chain reaction (RT-PCR) have shown a sporadic and transient appearance of bcr-abl expression in the first 12 months after BMT in some patients who remain in complete remission and fail to show any further expansion of the bcr-abl positive clone [7, 8]. Unfortunately, however, approximately 20% of patients transplanted in chronic phase still relapse after BMT [6], with the reappearance of Ph-positive metaphases in all bone marrow cells.
For all of the patients in our study, systematic cytogenetic follow-up studies were carried out. Subsequent reappearances of Ph-positive metaphases were detected and documented in detail. Because the reappearance and steady increase of Ph-positive cells in CML patients relapsing after BMT mimics, at least in part, the initial development of the disease, it was decided to simulate the sequence of events using a biomathematical model derived from a recognized model of granulopoiesis and to compare the results with data of patients who had been studied sufficiently closely following post-BMT relapse. We elected to analyze the proportion of Ph-positive metaphases rather than the level of expression of bcr-abl m-RNA in order to estimate the proportion of CML cells present; some of these would have been of the erythroid, megakaryocytic, or monocytic lineages, but the great majority would be myeloid. In any case, the ratio of Ph-positive to Ph-negative cells would be similar in all hematopoietic cell lines. The results based on the new biomathematical model yielded predicted outcomes which approximated closely the observed increases in Ph-positive cells after relapse and pointed to the significance of extramedullary hematopoiesis in the initiation of relapse. The results also closely simulated the observed changes in blood neutrophil numbers.
|
Methods
|
|---|
Patients Studied
From 1984 to 1994, 55 patients in the chronic phase of CML underwent treatment with HLA-identical allogeneic BMT (54 patients) or syngeneic BMT (one patient) at the medical hospital of the University of Ulm using standard protocols [6] with myeloablative conditioning, including total body irradiation. These protocols have been described elsewhere [9, 10]. In follow-up studies done for 16 to 139 months (median 70 months), 23 patients had relapsed 3 to 61 months after BMT.
Cytogenetic analyses before and after BMT had been carried out sufficiently frequently in six of the relapsing patients to allow comparisons with the predictions derived from biomathematical modeling. As part of patient management following BMT, bone marrow aspirations were performed a minimum of six times and a maximum of 13 times during the course of the study. After the first two years, bone marrow aspirations were performed one to four times per year until the end of the follow-up. In these six patients, the cytogenetic relapse preceded the hematological relapse by 12 to 38 months, and in each case the proportion of Ph-positive metaphases increased steadily during that time period. bcr-abl studies were also performed since 1988 using either Southern blotting or PCR analysis of mRNA as described elsewhere [9, 10]. The present analysis, however, was particularly concerned with the numbers of cells which originated from the primary CML clone. Therefore, the calculations focused only on the proportions of Ph-positive metaphases that were observed.
Details of the six patients who were studied are shown in Tables 1 and 2. All were transplanted in the chronic phase after conventional conditioning regimens including total body irradiation, 12 Gy in six fractions over three days. Five of the six patients received HLA-identical marrow, and one patient (Patient 3) received marrow from an identical twin. Four patients received donor lymphocytes [11, 12] and chemotherapy as treatment for relapse. At the time this study was done, two (Patients 1 and 5) were in Ph-negative remission, one (Patient 3) was alive in Ph-positive chronic phase, and one (Patient 6) had died. The remaining two patients were treated with chemotherapy alone; one was alive in chronic phase and the other had died (Table 2).
Cytogenetic analyses were performed using conventional techniques [13]. A minimum of 50 bone marrow cells were analyzed in each sample. The number of Ph-positive cells increased steadily over time in all six patients, and the time-dependent increase was analyzed. Additional chromosomal abnormalities were observed in all patients, some of which were clonal as reported elsewhere [14]. It should be noted, however, that none of these additional chromosomal abnormalities was suggestive of an acute transformation during cytogenetic follow-up of relapse.
Biomathematical Model of CML Relapse
The biomathematical model used to simulate CML relapse following BMT is shown schematically in Figure 1. A detailed description of the model will be reported elsewhere (Tibken et al., manuscript in preparation). The basis of this model is the Fliedner-Steinbach model of granulopoiesis [15, 16], which was further developed by Tibken and Hofer [17]. This model allows for the compartments of primitive stem cells, committed stem cells, proliferative cells, maturing cells, reserve cells, and functional cells to be subjected to feedback regulation. In the present work, the model was modified to simulate CML relapse after BMT by combining a series of catenated compartments which represent normal granulopoiesis in the marrow with a second series representing Ph-positive granulopoiesis in the marrow and a third series representing Ph-positive granulopoiesis in extramedullary sites. It is assumed that Ph-positive cells move nearly instantaneously from marrow to extramedullary sites via the blood and vice versa and that this traffic is particularly relevant in the case of Ph-positive stem cells. In the case of more mature proliferating cells (myeloblasts through myelocytes), however, it is likely that most of these cells seen in the blood have originated in extramedullary sites [4]. Normal granulopoiesis, by contrast, is limited to the bone marrow, and cells are not released into the blood until they are mature.
View larger version (32K):
[in this window]
[in a new window]
|
Figure 1. The biomathematical model of CML relapse after BMT. Cells move from the stem cell (S), through the committed stem cell (C), proliferative (P), maturing (M), and reserve (R) compartments before release into the blood as functional (F) cells. Information flows through two regulatory systems (Reg I and Reg II) responding to the numbers of precursor cells (Reg I) and to the number of functional cells (Reg II). Normal and CML cells coexist in the marrow and blood but only CML cells proliferate in extramedullary sites.
In the model, it was assumed that 107 normal stem cells were infused at the time of BMT, that small numbers—in a range between 1 and 500—of CML (Ph+) cells survived the pretransplant conditioning and that, in patients who relapse, these Ph+ cells began to proliferate after a delay of up to 240 days. It was assumed: A) that CML (Ph+) stem cells were not susceptible to feedback regulation; B) that CML granulopoiesis contributed to feedback regulation of normal granulopoiesis, and C) that the compartment transit time for Ph+ cells was longer than for normal cells.
Different values for the numbers of surviving Ph+ cells and for the time delay were substituted iteratively until the best representation of the observed values for the reappearance of Ph+ cells, and for the rise in numbers of neutrophils, were obtained.
|
|
The biomathematical model has four assumptions built into it, namely: A) that CML (Ph-positive) stem cells are not susceptible to feedback regulation; B) that CML granulopoiesis, however, contributes to feedback regulation of normal granulopoiesis; C) that the compartment transit time (CTT) for Ph-positive cells is longer than normal, and D) that 107 normal pluripotential stem cells are transplanted (in the adult).
The first two assumptions—concerning feedback regulation in CML—are important features in the pathophysiology of CML. They are, therefore, of crucial importance in simulating the way in which Ph-positive cells can eventually come to dominate normal hematopoiesis, despite their longer than normal generation times. Functionally mature neutrophils appear in the blood; some of these originate from normal marrow granulopoiesis, while others originate from Ph-positive granulopoiesis in the marrow and extramedullary sites. It is assumed that the total neutrophil pool consisting of normal and CML-derived neutrophils exerts negative feedback regulation of normal committed stem cells. Ph-positive stem cells are, however, assumed to be resistant to this feedback regulation. This assumption is consistent with studies of the growth of CML progenitor cells in cultures and is also consistent with the results of studies of leukemia in animal models [18].
It has long been known that the generation times, and hence the CTT for the proliferative compartment in CML, are longer than for normal granulopoiesis. The values used in the present model were derived by averaging the measured mean grain count halving times for myeloblast, promyelocyte, and myelocyte development in the bone marrow as reported by others [2, 19, 20] and then summing the values. (Since myelocytes divide twice, their average generation time was counted twice). This approach yielded a total CTT for the proliferative compartment of 336.5 h.
The fourth assumption—namely, that 107 pluripotential stem cells are transplanted—is based on the typical bone marrow transplant harvest of 2.6 x 108 nucleated marrow cells per kg (recipient) [21], equivalent to 1.8 x 1010 cells in a 70 kg adult. The precise proportion of marrow cells that are pluripotential stem cells capable of reconstituting hematopoiesis in man is not known, but a reasonable estimate can be based on data from mice, which indicate there are 500 marrow-repopulating (Thy-1.1lo Lin– Sca-1+) cells per 106 nucleated marrow cells [22]. This is equivalent to approximately 107 cells in a 70 kg adult recipient.
The questions to be answered were: A) whether relapse began from Ph-positive cells in the bone marrow or in extramedullary sites; B) the effect of the number of Ph-positive cells surviving conditioning therapy, and C) the effect of time delay during which leukemic cells were apparently quiescent.
Values for the numbers of Ph-positive cells that might have survived conditioning therapy and for the length of time delay were substituted iteratively until a best representation of the actual data was obtained. The computer simulation of the biomathematical model was used to compute the percentage of leukemic cells capable of division in the bone marrow, and of the total number of segmented neutrophils (leukemic + normal) circulating in the blood.
|
Results
|
|---|
Simulation of CML Relapse
Simulation curves were generated for each patient, modeling the rise in the proportion of Ph-positive metaphases in the bone marrow and the rise in total blood neutrophils that followed post-engraftment relapse.
The first question to be answered was whether relapse began from Ph-positive cells in the bone marrow or in extramedullary sites. The analysis made ( Fig. 2 and Table 3) indicates that there is a rapid migration between the stem cell pools. These findings are in agreement with clinical experience, supporting the assumption of nearly instantaneous mixing of stem cells. In contrast, it was impossible to simulate closely either the rising proportions of Ph-positive cells or the blood neutrophil levels if it was assumed that relapse began exclusively from cells in the bone marrow (data not shown).
View larger version (14K):
[in this window]
[in a new window]
|
Figure 2. Simulation of residual leukemia, originating in extramedullary sites (top curve) or in the bone marrow (bottom curve), simulated with a time delay of 240 days for five residual leukemic cells. This analysis indicates that there is a rapid migration between the two stem cell pools in bone marrow and in extramedullary sites.
|
|
The second question related to the residual number of Ph-positive cells remaining after completion of BMT—i.e., the number of cells surviving the conditioning regimen. Here, the biomathematical model was run iteratively with the numbers of residual Ph-positive stem cells being progressively reduced from 500, and the results for Patient 1 are shown in Figure 3. This patient's data are best represented by the simulation using five residual cells. Similarly small values were obtained for the other patients ( Table 4).
View larger version (16K):
[in this window]
[in a new window]
|
Figure 3. The effects of assuming different numbers of residual Ph+ stem cells on the rate of reappearance of Ph+ cells in the bone marrow of Patient 1. The patterns simulating for 5, 10, or 500 cells (broken lines) are compared with the cytogenetically evaluated proportions of Ph+ cells seen in serial marrows (unbroken line). Simulations were calculated for a time delay of 240 days. Best simulations were achieved with five cells.
|
|
The third question dealt with the time delay between transplantation and the process of re-activation of the Ph-positive cells that go on to cause the relapse. The results of iterative substitution of time delays between 10 and 240 days for Patient 1 are shown in Figure 4. This patient's data are best represented by the simulation with a time delay of 240 days. Similar iterative substitutions were carried out for all six patients and compared with the rise in Ph-positive metaphases observed in each case. The results can be seen in Figure 5.
View larger version (17K):
[in this window]
[in a new window]
|
Figure 4. The effects of assuming different time delays on the rate of reappearance of Ph+ cells in the bone marrow of Patient 1, calculated for five residual Ph+ cells surviving conditioning ( Fig. 3). Calculated curves for delays of 10, 170, or 240 days (broken lines) were compared with cytogenetically evaluated data (unbroken line). This patient's data were best represented by the simulation with a time delay of 240 days.
|
|
View larger version (34K):
[in this window]
[in a new window]
|
Figure 5. Predicted (broken lines) compared with cytogenetically evaluated (unbroken lines) data in Patients 1 through 6 for the reappearance of Ph+ cells in BM. (Therapy is indicated at times shown by triangles, and the deaths of Patients 2 and 6 are marked with a cross. For therapy see Table 2. For simulation results and cytogenetic data, see Figs. 3 and 4.)
For Patient 1, five residual Ph+ cells and a time delay of 240 days were assumed (see text). The patient was treated with donor lymphocyte infusion at 74 months (not shown here). She is alive and now Ph–.
The observed data for Patient 2 were best represented by assuming five residual Ph+ cells and a time delay of 240 days.
The observed data for Patient 3 were best represented by assuming a single residual Ph+ cell and a time delay of 180 days.
The observed data for Patient 4 were best represented by assuming a single residual Ph+ cell and a time delay of 140 days.
The observed data for Patient 5 were best represented by assuming five residual Ph+ cells and a time delay of 80 days. This patient was treated with donor lymphocyte infusion at 32 months, followed by complete cytogenetic response. He is alive and Ph–.
The observed data for Patient 6 were best represented by assuming five residual Ph+ cells and a time delay of 15 days.
|
|
The time delay factor varied substantially among the patients ( Table 4). Patients 1 through 4 suffered relapses which were predicted to involve long delays (240, 180, or 140 days) between transplantation and the earliest evidence of a Ph-positive reactivation. By contrast, Patient 5 had a much shorter delay (80 days), while Patient 6 had the most rapid relapse of all, with a delay of only 15 days. On the other hand, in all six patients, the simulation results indicated that a very small number of Ph-positive stem cells was responsible for relapse.
Simulation of Neutrophil Regeneration
In the case of a patient with CML relapsing after BMT, the blood neutrophils consist of those derived from normal (donor) stem cells and those derived from the re-emerging (host) Ph-positive clone. The observed numbers of neutrophils are likely to be subject to greater variation than the numbers of Ph-positive cells in the marrow. The predicted numbers of all blood neutrophils were simulated using 107 normal (donor) stem cells and the individually determined number of Ph-positive (host) stem cells as well as time delay ( Table 4). The results of neutrophil simulation are shown in Figure 6. In Patients 1, 2, 4, 5, and 6, there is a good correlation between the predicted and the observed values. In Patient 3, the observed numbers of neutrophils were lower than the predicted values from 24 months onward.
View larger version (35K):
[in this window]
[in a new window]
|
Figure 6. Predicted (broken lines) compared with cytogenetically evaluated (unbroken lines) data in Patients 1 through 6 for the numbers of blood neutrophils. As in Figure 5, therapy is indicated at times shown by triangles, and the deaths of Patients 2 and 6 are marked with a cross. The number of residual Ph+ cells and the time delay for each patient are indicated in the legend for Figure 5. (For therapy, see Table 2. For simulation results and cytogenetic data, see Figs. 3 and 4.)
|
|
It is known that leukemic hematopoiesis takes place in the spleen and liver as well as in the marrow. In the analyses presented here, satisfactory representation of the data could only be achieved if Ph-positive stem cells were assumed to be able to move freely between extramedullary and medullary sites. This concept is not inconsistent with the notion that relapse is largely the result of the activation of extramedullary cells.
In biomathematical modeling, the parameters, "residual leukemic stem cells" and "time delay", were estimated individually for each patient. The simulation of relapse development based on these estimates was very close to cytogenetic data for all six patients. In addition, the neutrophil blood cell counts were represented very well in five of six patients using the estimated values of the parameters. Neutrophil blood cell data thus serve as an independent test for the quality of this modeling.
|
Discussion
|
|---|
We have constructed a biomathematical model for relapse development of CML after BMT which includes a number of known pathophysiological components. An important aspect of our modeling was that we were able to test our results with the cytogenetic and hematologic data of six CML patients who relapsed after BMT. Despite the significant advances that have been made in detecting and analyzing the bcr-abl rearrangement [7, 8], identification of the Ph chromosome remains the best way of proving that an individual cell belongs to the CML clone. Successful allogeneic BMT in CML results in a complete cytogenetic response, defined as the absence of detectable Ph-positive metaphases, while relapse after transplantation is accompanied by the reappearance of Ph-positive cells.
The number of available cytogenetic data proved to be particularly valuable for testing our model. We were able to analyze the cytogenetic data obtained from six patients who underwent bone marrow evaluations several times per year over a period of years, beginning shortly after BMT and continuing through and beyond the stage of full hematological relapse. In these patients, cytogenetic relapse preceded hematological relapse by 12 to 38 months, and in each case the proportion of Ph-positive metaphases increased slowly and steadily. We were able to gain insight into the relationships between leukemic and normal (donor) cells over a period of years after BMT.
A reasonable simulation could be achieved using the assumptions mentioned above: A) that CML relapse originates from both extramedullary sites and from the bone marrow; B) that there is a rapid exchange of leukemic cells between extramedullary organs, the blood and the bone marrow; C) that the spleen can release immature leukemic cells into the bloodstream, which can travel into the marrow or back into the spleen and continue to proliferate; D) that leukemic and normal neutrophils exert negative feedback on normal granulopoiesis; E) that leukemic granulopoiesis is resistant to the feedback control of neutrophils, and F) that leukemic progenitor cells have longer than normal generation times.
Our model fitted the actual results very well when it was assumed that the CML relapse originated from extramedullary sites (principally the spleen) and from the bone marrow; the fit was far less satisfactory if relapse was assumed to begin in the marrow alone. A number of studies have shown that leukemic cells can proliferate in the spleen [2, 23, 24]. It used to be thought that CML cells had an intrinsic abnormality with fetal characteristics that enabled them to multiply in fetal hematopoietic organs [25]. While this might be possible, it is more likely that CML cells can express adhesion molecules that allow them to find niches in the splenic vascular bed in which they can proliferate. Unlike the bone marrow, the spleen does allow the release of immature granulopoietic cells (myeloblasts through metamyelocytes) into the blood [2, 4], and these can be taken up by the marrow even at high marrow cell densities [26].
The parameters that might influence the timing with which Philadelphia-positive mitoses appeared in the bone marrow of each individual patient were analyzed. Two parameters that seemed to play an appreciable role in determining the timing and behavior of re-emerging Ph-positive cells were A) the residual number of Ph-positive cells that survived the original conditioning regimen, and B) the time delay following BMT, during which Ph-positive cells were present but not dividing at any significant rate. The results suggested that CML relapse can originate from very small numbers of Ph-positive cells which survive pretransplant conditioning.
The time delay in our model is comparable to Killmann's "Sleeper-to-Feeder" hypothesis [27]. According to this theory, stem cells can either be in a dormant state, or they can be actively dividing and thus "feeding" subsequent cell pools. This conversion of stem cells from a dormant to a replicating state appears to be random. In the case of CML relapse after BMT, the time when the first leukemic stem cell passes from a dormant to an active state could vary considerably among patients, as seen in the wide range of time delays between the bone marrow transplant and the reactivation of Ph-positive cells in the patients analyzed here.
Another factor contributing to the time delay in leukemic patients could be the graft-versus-leukemia effect, due largely (but not exclusively) to differences in minor histocompatibility antigens, which has been shown to play a significant role in suppressing relapse after allogeneic bone marrow transplantation [11, 12]. It should be noted, however, that one patient in the present series received syngeneic marrow, and two patients received T-cell-depleted marrow; a graft-versus-leukemia effect is unlikely to have been significant in these three patients.
An important step forward achieved by this model has been the ability to link two independently observed sets of patient data (the proportion of Ph-positive metaphases in the marrow and the numbers of neutrophils in the blood) in simulating the observed results. The success achieved in replicating both data sets is encouraging and suggests that the biomathematical model is a reasonable description of a pathophysiological process, namely, the relapse of Ph-positive CML after BMT.
|
Acknowledgments
|
|---|
Paul C. Vincent was Visiting Professor at the University of Ulm while this work was carried out. The authors are indebted to Prof. Herrmann Heimpel, former Director of the Department of Internal Medicine III, University of Ulm, and to Prof. Renate Arnold as well as to Prof. Donald Bunjes, who have been and are responsible for the Bone Marrow Transplantation Unit of this department, for providing us with the data, and for helpful discussions and constructive criticisms that allowed us to perform this study effectively.
The research work of this study has been supported by special university funds and the German Federal Government as well as the European Commission.
|
References
|
|---|
-
Vincent PC, Cronkite EP, Greenberg ML et al. Leukocyte kinetics in chronic myeloid leukemia: I. DNA synthesis time in blood and marrow myelocytes. Blood 1969;33:843-850.[Abstract/Free Full Text]
-
Ogawa M, Fried J, Sakai Y et al. Studies of cell proliferation in human leukemia. VI. the proliferative activity, generation time, and emergence time of neutrophilic granulocytes in chronic granulocytic leukemia. Cancer 1970;25:1031-1049.[Medline]
-
Vincent PC. Granulocyte kinetics in health and disease. Clin Haematol 1977;6:695-717.[Medline]
-
Vincent PC. Kinetics of leukemia and control of cell division and replication. In: Henderson ES, Lister TA, eds. Leukemia. Philadelphia, London, Toronto, Montreal, Sydney, Tokyo: W.B. Saunders Company, 1990:55-104.
-
Heim S, Mitelman F. Cancer Cytogenetics. New York, Chichester, Brisbane, Toronto, Singapore: Wiley-Liss, 1995:33-68.
-
Champlin R, McGlave P. Allogeneic bone marrow transplantation for chronic myeloid leukemia. In: Forman SJ, Blume KG, Thomas ED, eds. Bone Marrow Transplantation. Boston: Blackwell Scientific Publications, 1994;595-606.
-
Hughes TP, Morgan GJ, Martiat P et al. Detection of residual leukemia after bone marrow transplant for chronic myeloid leukemia: role of polymerase chain reaction in predicting relapse. Blood 1991;77:874-878.[Abstract/Free Full Text]
-
Delage R, Solffer RJ, Dear K et al. Clinical significance of bcr-abl gene rearrangement detected by polymerase chain reaction after allogeneic bone marrow transplantation in chronic myelogenous leukemia. Blood 1991;78:2759-2767.[Abstract/Free Full Text]
-
Arnold R, Bartram CR, Heinze B et al. Evaluation of remission state in chronic myeloid leukemia patients after bone marrow transplantation using cytogenetic and molecular genetic approaches. Bone Marrow Transplant 1989;4:389-392.[Medline]
-
Arnold R, Janssen JWG, Heinze B et al. Influence of graft-versus-host disease on the eradication of minimal residual leukemia detected by polymerase chain reaction in chronic myeloid leukemia patients after bone marrow transplantation. Leukemia 1993;7:747-751.[Medline]
-
Kolb HJ, Mittermuller J, Clemm C et al. Donor leukocyte transfusions for treatment of recurrent myelogenous leukemia in marrow transplanted patients. Blood 1990;76:2462-2465.[Abstract/Free Full Text]
-
Kolb H-J, Schattenberg A, Goldman JM et al. Graft-versus-leukemia effect of donor lymphocyte transfusions in marrow grafted patients. Blood 1995;86:2041-2050.[Abstract/Free Full Text]
-
Heinze B, Arnold R, Kratt E et al. Clonal evolution of chromosomal anomalies in leukemic patients after bone marrow transplantation. In: Chadwick KH, Seymour C, Barnhart B et al., eds. Cell Transformation and Radiation-Induced Cancer. Bristol: Adam Hilger Ltd., 1989:177-184.
-
Heinze B, Bink K, Bunjes D et al. Relapse in CML after BMT: Chromosomal abnormalities in addition to the ph do not necessarily implicate the transformation of the disease. Ann Hematol 1996;73(suppl 2):450.
-
Fliedner TM, Steinbach K-H. Simulationsmodelle von Perturbationen des granulozytären Zellerneuerungssystems. In: Doerr W, Schipperges H, eds. Modelle der Pathologischen Physiologie. Berlin: Springer, 1987:89-106.
-
Fliedner TM, Steinbach K-H. Repopulating potential of hematopoietic precursor cells. Blood Cells 1988;14:393-410.[Medline]
-
Tibken B, Hofer EP. A biomathematical model of granulocytopoiesis for estimation of stem cell numbers. STEM CELLS 1995;13(suppl 1):283-289.
-
Hoelzer D, Harriss EB. The failure of normal haematopoiesis in rats during the development of acute leukemia. Acta Haematol 1973;49:36-47.[Medline]
-
Fliedner TM, Cronkite EP, Bond VP. Das Studium der Proliferationsdynamik der Myelopoese unter Verwendung der Einzelzellautoradiographie. Folia Haematologica 1961;6:210-228.
-
Killmann SA, Cronkite EP, Robertson JS et al. Estimation of the life cycle of leukemic cells from labeling in human beings in vivo with tritiated thymidine. Lab Invest 1963;12:671-684.[Medline]
-
Buckner CD, Petersen FB, Bolonosi BA. Bone marrow donors. In: Forma SJ, Blume KG, Thomas ED, eds. Bone Marrow Tranplantation. Boston: Blackwell Scientific Publications, 1994; 259-269.
-
Baum CM, Uchida N, Peault B et al. Isolation and characterization of hematopoietic progenitor and stem cells. In: Forman SJ, Blume KG, Thomas ED, eds. Bone Marrow Transplantation. Boston: Blackwell Scientific Publications, 1994;53-71.
-
Muller-Bérat CN, Wantzin GL, Philip P et al. Agar culture studies of bone marrow, blood, spleen and liver in chronic myeloid leukemia. Leuk Res 1977;1:123-131.
-
Morris TCM, Vincent PC, Gunz FW et al. Evidence following splenic radiotherapy for a highly dynamic traffic of CFU-GM between the spleen and other organs in chronic granulocytic leukemia. Leuk Res 1987;11:109-117.[Medline]
-
Bagby GC. Stem cell (CFU-C) proliferation and emergence in a case of chronic granulocytic leukemia: the role of the spleen. Scand J Haematol 1978;20:193-199.[Medline]
-
Adler SS. The pathogenesis of spleen mediated phenomena in chronic myeloid leukemia and agnogenic myeloid metaplasia: a non abscopal mechanism. Scand J Haematol 1976;17:153-159.[Medline]
-
Killmann SA. Acute leukemia: development, remission/relapse pattern, relationship between normal and leukemic hemopoiesis, and the ‘sleeper-to-feeder’ stem cell hypothesis. Series Haematol 1968;1:103-128.
accepted for publication November 16, 1998.
Top
Abstract
Introduction
