COMPARATIVE LEUKEMIA RESEARCH 1973, LEUKEMOGENESIS, BIEL. HAEMAT., NO. 40, ED. Y. ITO ΑND R. M. DUTCHERt, UNIV. OF TOKYO PRESS, TOKYO/KARGER, BASEL, PP. 235-241 (1975)
Growth and Responsiveness of Human Granulocytic Leukemic Cells In Vitro
D. METCALF and Μ. Α.
S.
MOORE
Cancer Research Unit, Walter and Eliza Hall Institute*
The agar culture technique introduced by Bradley and Metcalf (1) and Ichikawa et al. (6) was adapted by Pike and Robinson (17) to permit the growth of granulocytic colonies from human marrow, spleen, and blood cells. Such cultures also develop smaller aggregates of cells termed "clusters," and cluster-forming cells appear to be the progeny of colony-forming cells. Subsequent studies by many groups have documented the following observations; a) The cells in normal marrow populations forming colonies in agar (colony-forming cells, CFC's) are the specific progenitor cells of granulocytes and monocytes (10-50/105 nucleated marrow cells in humans). b) Colony formation requires stimulation by a specific glycoprotein, termed "colony stimulating factor" (CSF). c) Leukemic cells from patients with either chronic myeloid leukemia (CML) or acute myeloid leukemia (AML) will grow in agar, and the proliferating cells are representative of the leukemic * Melbourne, Australia.
Downloaded by: Université de Paris 193.51.85.197 - 1/19/2020 12:41:31 PM
Cells from patients with acute myeloid leukemia (AML) and chronic myeloid leukemia (CIL) can be grown and their properties analyzed in agar gel cultures. Levels of the glycoprotein regulator colony stimulating factor (CSF) were found to be elevated in 19-66% of plasmas tested from patients with various types of granulocytic leukemia, and the growth of AML and CIL cells in vitro was observed to be dependent on, and responsive to, stimulation by CSF-containing material. In both diseases, the leukemic cells appear to be in a responsive state with respect to normal growth regulators, and potentially alterations in regulator levels may therefore be able to achieve sustained arrest of the growth of leukemic populations.
236
METCALF/MOORS
population. d) Proliferation of leukemic cells in vitro is not autonomous and requires stimulation by CSF (8, 14). Growth Pattern of Human Leukemic Cells In Vitro Cells from the blood or marrow of CML patients form large numbers of colonies in agar with a gross appearance identical to normal granulocytic colonies (2, 4, 16). Such colonies contain Philadelphia-positive cells although Philadelphia-negative colonies have been described in cultures of cells from certain patients with CML (4). In an untreated patient with CML, the incidence of CFC's in the marrow is always higher than normal, and whereas in normal blood the incidence of CFC's per 105 nucleated cells is only 1/100 of that in the marrow, in CML the frequency of CFC's is usually higher than in the marrow and can be as high as 20,000 times normal levels. The data from an analysis of 27 pretreatment patients with CML (10) is summarized in Table I. TABLE I. Levels of Colony-forming and Cluster-forming Cells in Normal Humans and Pretreatment Patients with CMLa) Tissue Normal adult marrow Normal adult blood CML marrow CML blood
Colony-forming cells per 2 X 105 cells 56 0.4 460 1,020
Cluster-forming % colony-forming cells cells per 2 X 105 of lighter density cells than 1.062 g/cm3 390 4 1,790 2,100
6 3 61 74
Although differentiation is near normal in CML colonies growing in agar, CML CFC's can be distinguished from normal CFC's because of their lighter buoyant density (14). Thus in the above series, 61-74% of CML CFC's were of lighter density than 1.062 g/cm3 whereas only 3-6% of normal CFC's were of lighter density than 1.062. The in vitro growth characteristics of AML cells are significantly different from normal or CML cells (5, 7, 14, 17). In a study of 100 pretreatment patients with AML or acute myelomonocytic leukemia (AMML) (12), in 13 cases no proliferation occurred in agar. In 18 patients, cells from the marrow or blood formed colonies in agar, but in 7 of the 18 patients these were of smaller size than normal, and colony cells exhibited obvious defects in cellular maturation. The growth pattern at 7 days characteristic of the majority of AML marrow populations was of an absence of colonies with persisting variable numbers of small clusters, the cluster cells again exhibiting abnormal differentiation. The finding of large numbers of clusters in cultures of blood and marrow from a patient with an absence of CFC's is diagnostic of acute myeloid leukemia. Again, as in the case of CML cells, an abnormally high proportion of proliferating AML cells is of lighter buoyant density than 1.062 (14), which distinguishes AML clusterforming cells from normal cluster-forming cells.
Downloaded by: Université de Paris 193.51.85.197 - 1/19/2020 12:41:31 PM
a) Mean data from 27 pretreatment patients with CML. Cultures scored at day 7 of incubation. Colonies >40 cells. Cluster : 3-40 cells. All cultures stimulated by underlayers of human or monkey peripheral blood cells.
HUMAN LEUKEMIC CELLS IN VITRO
237
Plasma CSF Levels in CIL and AML As part of the study on the growth characteristics of AML and CIL cells, CSF levels in the plasma of these patients were determined using dialyzed plasma specimens and the mouse bone marrow assay system (3). Individual assay runs were standardized using 5 standard CSF preparations (4 human, 1 monkey), and data were corrected on the basis of these standards. CSF levels were expressed as the number of colonies stimulated by 0.1 ml of dialyzed plasma from 75,000 C57BL bone marrow cells—the range of CSF levels in normal plasma being 1-20. Repeat assays on plasma samples indicated some variability in assay data even with the use of corrections based on the standard CSF preparations. Thus, concordant data were obtained with only 80% of the plasma samples. Analysis showed that this variability was due in part to loss of CSF during storage at —20°C due to protein precipitation with adsorption of CSF and/or denaturation of the CSF itself. To minimize storage variations, plasma or serum samples should be dialyzed immediately following collection and stored prior to assay at 4°C. In agreement with earlier data on CSF levels in sera from patients with AML (9) the present survey showed (Table II) that CSF levels were elevated above the upper TABLE II.
Plasma CSF Levels in Myeloid Leukemic Patients Disease
AML-pretreatment AML-remission AML-relapse AMML Subacute ML
CIL CIL
acute transformation
Number of patients
% CSF levels elevated above normal
Mean CSF levelsa)
60 58 50 27 16 57 38
33 24 42 33 19 47 66
21±21 13±13 22±22 22±24 14±19 23±18 43±39
limit of normal in 67 of 211 (32%) samples from AML and AMML patients. In remission, plasma CSF levels were less often elevated than in pretreatment or relapse patients, although the differences were of marginal significance only. Plasma CSF levels were elevated in 47% of patients with CIL (including pretreatment cases). It was notable that CSF levels were considerably higher in patients with CIL after the onset of acute transformation. Thus, although patients with acute transformation of CIL resemble in many ways patients with AML, at least as judged by CSF levels the two groups are significantly different. The results from these surveys indicate that for prolonged periods patients with AML and CIL have in their circulation elevated concentrations of CSF, at least as assayed on mouse bone marrow cells. Similar observations have been made on urinary CSF levels, although in individual paired specimens there was often no correspondence between serum and urine CSF levels (9, 18).
Downloaded by: Université de Paris 193.51.85.197 - 1/19/2020 12:41:31 PM
a) Mean number of colonies stimulated by 0.1 ml plasma on 75,000 Ci7BL bone marrow cells±standard deviations. Normal range : 1-20.
238
METCALF/MOORE
Before the significance of elevated CSF levels in leukemic patients can adequately be assessed it is necessary to establish whether leukemic granulocytic cells retain responsiveness to stimulation by CSF. It has been observed that medium conditioned by Rhesus monkey lung tissue is capable of stimulating colony formation by normal human marrow cells. The responsiveness of 150 human marrows has been measured using graded concentrations of monkey lung-conditioned medium. These cultures were scored at day 4 of incubation to determine total cluster counts (aggregates of 3 or 4 cells regardless of size) to avoid a) the difficulty of subjective assessment of colonies where colony size decreases with decreasing CSF concentrations, and b) to avoid progressive loss on continued incubation of aggregates by cellular dispersion, which is a feature of cultures of most human marrow cells. The results (8) indicated that cells from all patients studied with AML or CIL exhibited responsiveness to progressive increases in concentration of monkey lungconditioned medium, although for both diseases the dose-response curves differed in detail from the sigmoid curve obtained for normal marrow cells. CIL cells were marginally less responsive than normal at all concentrations, and AIL cells were more responsive than normal at low-conditioned medium concentrations. Because of the known variability in molecular structure of different CSF's, monkey lung CSF can be presumed to differ somewhat from human CSF, and it is possible that the observed responsiveness curves may have been influenced by the abnormal form of CSF used. This possibility requires that the data be rechecked when a suitable form of human CSF becomes available for study, but it is improbable that the conclusions reached will require substantial revision. These conclusions on the responsiveness of AIL and CIL cells are supported by other observations: a) When CSF-producing cells in bone marrow populations were separated from CFC's or cluster-forming cells by buoyant density centrifugation, no growth of CFC's or cluster-forming cells was observed unless cultures were stimulated by underlayers of white cells or white cell-conditioned medium (15). b) Addition of rabbit antiserum to human urine CSF to cultures of AIL or CIL cells inhibited cellular proliferation stimulated by endogenous CSF-producing cells or by underlayers (15). In current experiments the responsiveness of human leukemic cells to human urine CSF or human plasma of known colony stimulating activity for mouse cells is being investigated. Although human urine CSF appears rarely able to stimulate the formation of large colonies by human marrow cells, as assessed by total cluster counts at day 3, CSFHU does appear to significantly stimulate the proliferation of normal, AIL, or CIL cells (Table III). Using a similar assay system, dialyzed human plasma has been observed to significantly increase the number and size of day 3 clusters in cultures of normal, CIL, and AIL marrow. However, this latter system is highly complex, and it has not yet been clearly established that the stimulating activity is due solely to the CSF in these plasmas. In only some experiments have assays on mouse bone marrow target cells agreed with
Downloaded by: Université de Paris 193.51.85.197 - 1/19/2020 12:41:31 PM
Responsiveness of Human Leukemic Cells to CSF
HUMAN LEUKEMIC CELLS IN VITRO TABLE III.
239
Stimulation of Human Marrow Cultures by Human Urine CSF Type of patient
Miscellaneous hematological diseases AML AML-remission CML Myeloproliferative disorders Normal human fetal liver
Number of patients tested 22 8 3 4 4 4
Stimulation indexa) CSF 4.0±4.1 5.2±5.8 2.4±10 8.9±6.7 2.1 ±0.5 10.0±3.3
Conditioned medium 16.2±12.2 13.9±17.3 7.6±1.0 9.9±7.7 — 34.1±18.7
Mean ratio of total cluster counts at day 3 in cultures containing 0.1 ml of semi-purified human urine CSF or white cell-conditioned medium to clusters in cultures containing 0.1 ml saline ± standard deviations. Cultures contained 25,000 to 100,000 cells in 1 ml. a)
assays on human cells, and using electrophoretically separated fractions of human serum, evidence has been obtained of the existence in human sera of inhibitors that are not detectable using the conventional mouse marrow assay system. Furthermore, the growth of human marrow cells appears to be much more affected by general increases in the protein concentration of the medium than is the growth of mouse bone marrow cells, and it is possible that in the assays for CSF content, the addition of 0.1 ml of plasma to the 1-ml cultures has a nonspecific effect on human cell growth by altering protein concentration. So far as these recent results go, however, they do suggest strongly that normal and leukemic human cells are responsive to human CSF as detected by the mouse assay system, although the effects appear to be complicated by the operation of other factors that are as yet undefined,
With leukemic cells from many patients with AMLor CIL, the agar culture system has an unusually high plating efficiency and more than 60% of unfractionated blood or marrow cells can proliferate (14). Furthermore, karyotypic analysis of cultures from patients whose leukemic cells had mixed karyotypic abnormalities has indicated that these abnormalities are present in similar proportions in the cells proliferating in agar (11). It seems reasonable to conclude therefore that, at least in these patients, the agar culture system is supporting the growth of a genuinely representative selection of the leukemic population. A distinctly different growth pattern is observed between AMLand CIL cells growing in vitro. Typically, CIL cells formed colonies of normal gross appearance, whereas most AMLcells are capable of forming only smaller cell aggregates (clusters) whose cells exhibit abnormal differentiation. In cultures of normal cells, cluster-forming cells are likely to represent the progeny of colony-forming cells, but because of the potential abnormalities present in the leukemic cells it cannot necessarily be concluded from this that in AMLthe abnormal cells are more differentiated than in CIL. It
Downloaded by: Université de Paris 193.51.85.197 - 1/19/2020 12:41:31 PM
Discussion
240 METCALF/MOORE is equally likely that because of intrinsic abnormalities in the metabolism of AML cells the present culture media may not allow AML cells to express their full potential for proliferation. The data from the responsiveness studies on AML and CIL cells have shown clearly that the growth of these cells in vitro is not autonomous and is essentially as dependent on stimulation as is the growth of normal cells. In this context, it should be emphasized that the maximum concentrations of CSF used in the monkey lungconditioned medium experiments were lower than those present in normal human plasma as assessed in both cases on mouse marrow cultures. This, together with the evidence presented here that plasma CSF levels are commonly elevated above normal levels in both diseases, indicates that CSF potentially represents a powerful proliferative stimulus capable of stimulating the progressive growth of the leukemic populations in these patients. However, a number of lines of evidence suggest that the in vivo situation may be quite complex. As mentioned above, individual human plasmas can have quite different levels of stimulating activity for mouse versus human target cells, suggesting the existence of modifying factors in the plasma. Similarly, although brief exposure to CSF in vitro forces CIL cells into cell cycle (13), CML cells in the circulation of patients are mainly not in cell cycle despite the fact that the plasma contains levels of CSF as high as was used in the in vitro experiments. The nature and significance of these modifying or blocking factors is at present under investigation. The in vitro studies on patients with AML and CIL have documented the intriguing situation that the leukemic populations are in a responsive state with respect to normal growth regulators. It is of major importance to characterize the nature of these regulatory factors and to establish whether sustained alterations in regulator balance can control the growth of leukemic populations in vivo and achieve sustained remissions in the progression of these diseases. Acknowledgment This work was supported by the Carden Fellowship Fund of the Anti-Cancer Council of Victoria, the Lady Tata Trust, and the National Cancer Institute, Contract No. NO1-CB33854.
1 BRADLEY, T. R. and METCALF, D. The growth of mouse bone marrow cells in vitro. flust. J. Exp. Biol. Med. Sci., 44: 287 (1966). 2 BRowN, C. H. and CARBONE, P. P. Ιn vitro growth of normal and leukemic human bone marrow. J. Natl. Cancer Inst., 46: 989 (1971). 3 CHAN, S. H., METCALF, D., and STANLEY, E. R. Stimulation and inhibition by normal human serum of colony formation in vitro by bone marrow cells. Brit. J. Haematol., 20: 329 (1971). 4 CHERVENICK, Ρ. A., LAWSON, A. L., ELLIS, L. D., and PAN, S. F. Ιn vitro growth of leukemic cells containing the Philadelphia (Ph) Chromosome. J. Lab. Clin. Med., 78: 838 (1971).
Downloaded by: Université de Paris 193.51.85.197 - 1/19/2020 12:41:31 PM
REFERENCES
HUMAN LEUKEMIC CELLS
IN
VITRO 241
Downloaded by: Université de Paris 193.51.85.197 - 1/19/2020 12:41:31 PM
5 GREENBERG, P. L., NICHOLS, W. C., and SCHRIER, S. L. Granulopoiesis in acute myeloid leukemia and preleukemia. New Engl. J. Med., 284: 1225 (1971). 6 IcHIKAWA, Y., PLUZNIK, D. H., and SAGS, L. Ιn vitro control of the development of macrophage and granulocyte colonies. Proc. Natl. Acad. Sci. U.S., 56: 488 (1966). 7 ISCOVE, N. N., SEIN, J. S., TILL, J. E., and McCuLLOCH, E. A. Colony formation by normal and leukemic marrow cells in culture: Effect of conditioned medium from human leukocytes. Blood, 37: 1 (1971). 8 METCALF, D. Regulation by colony stimulating factor (CSF) of granulocyte and macrophage colony formation in vitro by normal and leukemic cells. In "Control of Proliferation in Animal Cells," ed. by Β. Clarkson and R. Baserga, Cold Spring Harbor Laboratories, New York (1973), in press. 9 METCALF, D., CHAN, S. H., GUwz, F. W., VINCENT, Ρ., and RAVICH, R. Β. M. Colony stimulating factor and inhibitor levels in acute granulocytic leukemia. Blood, 38: 143 (1971). 10 Μοο , M. A. S. In vitro studies in myeloid leukemias. In "Excerpts Medica Reviews in Leukemia and Lymphoma," ed. by F. J. Cleton, D. Crowther, and J. S. Maleas, Excerpts Medica, Amsterdam (1973), in press. 11 MOORE, M. A. S. and METCALF, D. Cytogenetic analysis of human acute and chronic myeloid leukemic cells cloned in agar cultures. Int. J. Cancer, 11: 143 (1973). 12 MOORE, M. A. S., SPITZER, G., WILLIAMS, N., and METCALF, D. Correlation of agar culture analysis and clinical status in patients with acute myeloid leukemia. Ιn "Hemopoiesis in Culture" (Proc. of 2nd Int. Workshop on Hemopoiesis in Culture), ed. by W. A. Robinson, U.S. Government Printing Office, Washington D.C. (1973). 13 MOORE, M. A. S. and WILLIAMS, N. Functional, morphologic and kinetic analysis of the granulocyte progenitor cell. In "Hemopoiesis in Culture" (Proc. of 2nd Int. Workshop on Hemopoiesis in Culture), ed. by W. A. Robinson, U.S. Government Printing Office, Washington D. C. (1973). 14 MooRs, M. A. S., WILLIAMS, N., and METCALF, D. In vitro colony formation by normal and leukemic human hemopoietic cells: Characterisation of the colony forming cells. J. Natl. Cancer Inst., 50: 603 (1973). 15 MOORE, M. A. S., WILLIAMS, N., and METCALF, D. In vitro colony formation by normal and leukemic human hemopoietic cells: Interaction between colony-forming and colonystimulating cells. J. Natl. Cancer Inst., 50: 591 (1973). 16 PARAN, M., SACHS, L., BAιιλκ, Y., and RESNITSKY, P. Ιn vitro induction of granulocyte differentiation in hematopoietic cells from leukemic and non-leukemic patients. Proc. Natl. Acad. Sci. U.S., 67: 1542 (1970). 17 PIKE, B. L. and ROBINSON, W. A. Human bone marrow colony growth in agar-gel. J. Cell. Physiol., 76: 77 (1970). 18 ROBINSON, W. A. and PIKE, B. L. Leucopoietic activity in human urine: The granulocytic leukemias. New Engl. J. Med., 282: 1291 (1970).
Growth and Responsiveness of Human Granulocytic Leukemic Cells In Vitro
D. METCALF and Μ. Α.
S.
MOORE
Cancer Research Unit, Walter and Eliza Hall Institute*
The agar culture technique introduced by Bradley and Metcalf (1) and Ichikawa et al. (6) was adapted by Pike and Robinson (17) to permit the growth of granulocytic colonies from human marrow, spleen, and blood cells. Such cultures also develop smaller aggregates of cells termed "clusters," and cluster-forming cells appear to be the progeny of colony-forming cells. Subsequent studies by many groups have documented the following observations; a) The cells in normal marrow populations forming colonies in agar (colony-forming cells, CFC's) are the specific progenitor cells of granulocytes and monocytes (10-50/105 nucleated marrow cells in humans). b) Colony formation requires stimulation by a specific glycoprotein, termed "colony stimulating factor" (CSF). c) Leukemic cells from patients with either chronic myeloid leukemia (CML) or acute myeloid leukemia (AML) will grow in agar, and the proliferating cells are representative of the leukemic * Melbourne, Australia.
Downloaded by: Université de Paris 193.51.85.197 - 1/19/2020 12:41:31 PM
Cells from patients with acute myeloid leukemia (AML) and chronic myeloid leukemia (CIL) can be grown and their properties analyzed in agar gel cultures. Levels of the glycoprotein regulator colony stimulating factor (CSF) were found to be elevated in 19-66% of plasmas tested from patients with various types of granulocytic leukemia, and the growth of AML and CIL cells in vitro was observed to be dependent on, and responsive to, stimulation by CSF-containing material. In both diseases, the leukemic cells appear to be in a responsive state with respect to normal growth regulators, and potentially alterations in regulator levels may therefore be able to achieve sustained arrest of the growth of leukemic populations.
236
METCALF/MOORS
population. d) Proliferation of leukemic cells in vitro is not autonomous and requires stimulation by CSF (8, 14). Growth Pattern of Human Leukemic Cells In Vitro Cells from the blood or marrow of CML patients form large numbers of colonies in agar with a gross appearance identical to normal granulocytic colonies (2, 4, 16). Such colonies contain Philadelphia-positive cells although Philadelphia-negative colonies have been described in cultures of cells from certain patients with CML (4). In an untreated patient with CML, the incidence of CFC's in the marrow is always higher than normal, and whereas in normal blood the incidence of CFC's per 105 nucleated cells is only 1/100 of that in the marrow, in CML the frequency of CFC's is usually higher than in the marrow and can be as high as 20,000 times normal levels. The data from an analysis of 27 pretreatment patients with CML (10) is summarized in Table I. TABLE I. Levels of Colony-forming and Cluster-forming Cells in Normal Humans and Pretreatment Patients with CMLa) Tissue Normal adult marrow Normal adult blood CML marrow CML blood
Colony-forming cells per 2 X 105 cells 56 0.4 460 1,020
Cluster-forming % colony-forming cells cells per 2 X 105 of lighter density cells than 1.062 g/cm3 390 4 1,790 2,100
6 3 61 74
Although differentiation is near normal in CML colonies growing in agar, CML CFC's can be distinguished from normal CFC's because of their lighter buoyant density (14). Thus in the above series, 61-74% of CML CFC's were of lighter density than 1.062 g/cm3 whereas only 3-6% of normal CFC's were of lighter density than 1.062. The in vitro growth characteristics of AML cells are significantly different from normal or CML cells (5, 7, 14, 17). In a study of 100 pretreatment patients with AML or acute myelomonocytic leukemia (AMML) (12), in 13 cases no proliferation occurred in agar. In 18 patients, cells from the marrow or blood formed colonies in agar, but in 7 of the 18 patients these were of smaller size than normal, and colony cells exhibited obvious defects in cellular maturation. The growth pattern at 7 days characteristic of the majority of AML marrow populations was of an absence of colonies with persisting variable numbers of small clusters, the cluster cells again exhibiting abnormal differentiation. The finding of large numbers of clusters in cultures of blood and marrow from a patient with an absence of CFC's is diagnostic of acute myeloid leukemia. Again, as in the case of CML cells, an abnormally high proportion of proliferating AML cells is of lighter buoyant density than 1.062 (14), which distinguishes AML clusterforming cells from normal cluster-forming cells.
Downloaded by: Université de Paris 193.51.85.197 - 1/19/2020 12:41:31 PM
a) Mean data from 27 pretreatment patients with CML. Cultures scored at day 7 of incubation. Colonies >40 cells. Cluster : 3-40 cells. All cultures stimulated by underlayers of human or monkey peripheral blood cells.
HUMAN LEUKEMIC CELLS IN VITRO
237
Plasma CSF Levels in CIL and AML As part of the study on the growth characteristics of AML and CIL cells, CSF levels in the plasma of these patients were determined using dialyzed plasma specimens and the mouse bone marrow assay system (3). Individual assay runs were standardized using 5 standard CSF preparations (4 human, 1 monkey), and data were corrected on the basis of these standards. CSF levels were expressed as the number of colonies stimulated by 0.1 ml of dialyzed plasma from 75,000 C57BL bone marrow cells—the range of CSF levels in normal plasma being 1-20. Repeat assays on plasma samples indicated some variability in assay data even with the use of corrections based on the standard CSF preparations. Thus, concordant data were obtained with only 80% of the plasma samples. Analysis showed that this variability was due in part to loss of CSF during storage at —20°C due to protein precipitation with adsorption of CSF and/or denaturation of the CSF itself. To minimize storage variations, plasma or serum samples should be dialyzed immediately following collection and stored prior to assay at 4°C. In agreement with earlier data on CSF levels in sera from patients with AML (9) the present survey showed (Table II) that CSF levels were elevated above the upper TABLE II.
Plasma CSF Levels in Myeloid Leukemic Patients Disease
AML-pretreatment AML-remission AML-relapse AMML Subacute ML
CIL CIL
acute transformation
Number of patients
% CSF levels elevated above normal
Mean CSF levelsa)
60 58 50 27 16 57 38
33 24 42 33 19 47 66
21±21 13±13 22±22 22±24 14±19 23±18 43±39
limit of normal in 67 of 211 (32%) samples from AML and AMML patients. In remission, plasma CSF levels were less often elevated than in pretreatment or relapse patients, although the differences were of marginal significance only. Plasma CSF levels were elevated in 47% of patients with CIL (including pretreatment cases). It was notable that CSF levels were considerably higher in patients with CIL after the onset of acute transformation. Thus, although patients with acute transformation of CIL resemble in many ways patients with AML, at least as judged by CSF levels the two groups are significantly different. The results from these surveys indicate that for prolonged periods patients with AML and CIL have in their circulation elevated concentrations of CSF, at least as assayed on mouse bone marrow cells. Similar observations have been made on urinary CSF levels, although in individual paired specimens there was often no correspondence between serum and urine CSF levels (9, 18).
Downloaded by: Université de Paris 193.51.85.197 - 1/19/2020 12:41:31 PM
a) Mean number of colonies stimulated by 0.1 ml plasma on 75,000 Ci7BL bone marrow cells±standard deviations. Normal range : 1-20.
238
METCALF/MOORE
Before the significance of elevated CSF levels in leukemic patients can adequately be assessed it is necessary to establish whether leukemic granulocytic cells retain responsiveness to stimulation by CSF. It has been observed that medium conditioned by Rhesus monkey lung tissue is capable of stimulating colony formation by normal human marrow cells. The responsiveness of 150 human marrows has been measured using graded concentrations of monkey lung-conditioned medium. These cultures were scored at day 4 of incubation to determine total cluster counts (aggregates of 3 or 4 cells regardless of size) to avoid a) the difficulty of subjective assessment of colonies where colony size decreases with decreasing CSF concentrations, and b) to avoid progressive loss on continued incubation of aggregates by cellular dispersion, which is a feature of cultures of most human marrow cells. The results (8) indicated that cells from all patients studied with AML or CIL exhibited responsiveness to progressive increases in concentration of monkey lungconditioned medium, although for both diseases the dose-response curves differed in detail from the sigmoid curve obtained for normal marrow cells. CIL cells were marginally less responsive than normal at all concentrations, and AIL cells were more responsive than normal at low-conditioned medium concentrations. Because of the known variability in molecular structure of different CSF's, monkey lung CSF can be presumed to differ somewhat from human CSF, and it is possible that the observed responsiveness curves may have been influenced by the abnormal form of CSF used. This possibility requires that the data be rechecked when a suitable form of human CSF becomes available for study, but it is improbable that the conclusions reached will require substantial revision. These conclusions on the responsiveness of AIL and CIL cells are supported by other observations: a) When CSF-producing cells in bone marrow populations were separated from CFC's or cluster-forming cells by buoyant density centrifugation, no growth of CFC's or cluster-forming cells was observed unless cultures were stimulated by underlayers of white cells or white cell-conditioned medium (15). b) Addition of rabbit antiserum to human urine CSF to cultures of AIL or CIL cells inhibited cellular proliferation stimulated by endogenous CSF-producing cells or by underlayers (15). In current experiments the responsiveness of human leukemic cells to human urine CSF or human plasma of known colony stimulating activity for mouse cells is being investigated. Although human urine CSF appears rarely able to stimulate the formation of large colonies by human marrow cells, as assessed by total cluster counts at day 3, CSFHU does appear to significantly stimulate the proliferation of normal, AIL, or CIL cells (Table III). Using a similar assay system, dialyzed human plasma has been observed to significantly increase the number and size of day 3 clusters in cultures of normal, CIL, and AIL marrow. However, this latter system is highly complex, and it has not yet been clearly established that the stimulating activity is due solely to the CSF in these plasmas. In only some experiments have assays on mouse bone marrow target cells agreed with
Downloaded by: Université de Paris 193.51.85.197 - 1/19/2020 12:41:31 PM
Responsiveness of Human Leukemic Cells to CSF
HUMAN LEUKEMIC CELLS IN VITRO TABLE III.
239
Stimulation of Human Marrow Cultures by Human Urine CSF Type of patient
Miscellaneous hematological diseases AML AML-remission CML Myeloproliferative disorders Normal human fetal liver
Number of patients tested 22 8 3 4 4 4
Stimulation indexa) CSF 4.0±4.1 5.2±5.8 2.4±10 8.9±6.7 2.1 ±0.5 10.0±3.3
Conditioned medium 16.2±12.2 13.9±17.3 7.6±1.0 9.9±7.7 — 34.1±18.7
Mean ratio of total cluster counts at day 3 in cultures containing 0.1 ml of semi-purified human urine CSF or white cell-conditioned medium to clusters in cultures containing 0.1 ml saline ± standard deviations. Cultures contained 25,000 to 100,000 cells in 1 ml. a)
assays on human cells, and using electrophoretically separated fractions of human serum, evidence has been obtained of the existence in human sera of inhibitors that are not detectable using the conventional mouse marrow assay system. Furthermore, the growth of human marrow cells appears to be much more affected by general increases in the protein concentration of the medium than is the growth of mouse bone marrow cells, and it is possible that in the assays for CSF content, the addition of 0.1 ml of plasma to the 1-ml cultures has a nonspecific effect on human cell growth by altering protein concentration. So far as these recent results go, however, they do suggest strongly that normal and leukemic human cells are responsive to human CSF as detected by the mouse assay system, although the effects appear to be complicated by the operation of other factors that are as yet undefined,
With leukemic cells from many patients with AMLor CIL, the agar culture system has an unusually high plating efficiency and more than 60% of unfractionated blood or marrow cells can proliferate (14). Furthermore, karyotypic analysis of cultures from patients whose leukemic cells had mixed karyotypic abnormalities has indicated that these abnormalities are present in similar proportions in the cells proliferating in agar (11). It seems reasonable to conclude therefore that, at least in these patients, the agar culture system is supporting the growth of a genuinely representative selection of the leukemic population. A distinctly different growth pattern is observed between AMLand CIL cells growing in vitro. Typically, CIL cells formed colonies of normal gross appearance, whereas most AMLcells are capable of forming only smaller cell aggregates (clusters) whose cells exhibit abnormal differentiation. In cultures of normal cells, cluster-forming cells are likely to represent the progeny of colony-forming cells, but because of the potential abnormalities present in the leukemic cells it cannot necessarily be concluded from this that in AMLthe abnormal cells are more differentiated than in CIL. It
Downloaded by: Université de Paris 193.51.85.197 - 1/19/2020 12:41:31 PM
Discussion
240 METCALF/MOORE is equally likely that because of intrinsic abnormalities in the metabolism of AML cells the present culture media may not allow AML cells to express their full potential for proliferation. The data from the responsiveness studies on AML and CIL cells have shown clearly that the growth of these cells in vitro is not autonomous and is essentially as dependent on stimulation as is the growth of normal cells. In this context, it should be emphasized that the maximum concentrations of CSF used in the monkey lungconditioned medium experiments were lower than those present in normal human plasma as assessed in both cases on mouse marrow cultures. This, together with the evidence presented here that plasma CSF levels are commonly elevated above normal levels in both diseases, indicates that CSF potentially represents a powerful proliferative stimulus capable of stimulating the progressive growth of the leukemic populations in these patients. However, a number of lines of evidence suggest that the in vivo situation may be quite complex. As mentioned above, individual human plasmas can have quite different levels of stimulating activity for mouse versus human target cells, suggesting the existence of modifying factors in the plasma. Similarly, although brief exposure to CSF in vitro forces CIL cells into cell cycle (13), CML cells in the circulation of patients are mainly not in cell cycle despite the fact that the plasma contains levels of CSF as high as was used in the in vitro experiments. The nature and significance of these modifying or blocking factors is at present under investigation. The in vitro studies on patients with AML and CIL have documented the intriguing situation that the leukemic populations are in a responsive state with respect to normal growth regulators. It is of major importance to characterize the nature of these regulatory factors and to establish whether sustained alterations in regulator balance can control the growth of leukemic populations in vivo and achieve sustained remissions in the progression of these diseases. Acknowledgment This work was supported by the Carden Fellowship Fund of the Anti-Cancer Council of Victoria, the Lady Tata Trust, and the National Cancer Institute, Contract No. NO1-CB33854.
1 BRADLEY, T. R. and METCALF, D. The growth of mouse bone marrow cells in vitro. flust. J. Exp. Biol. Med. Sci., 44: 287 (1966). 2 BRowN, C. H. and CARBONE, P. P. Ιn vitro growth of normal and leukemic human bone marrow. J. Natl. Cancer Inst., 46: 989 (1971). 3 CHAN, S. H., METCALF, D., and STANLEY, E. R. Stimulation and inhibition by normal human serum of colony formation in vitro by bone marrow cells. Brit. J. Haematol., 20: 329 (1971). 4 CHERVENICK, Ρ. A., LAWSON, A. L., ELLIS, L. D., and PAN, S. F. Ιn vitro growth of leukemic cells containing the Philadelphia (Ph) Chromosome. J. Lab. Clin. Med., 78: 838 (1971).
Downloaded by: Université de Paris 193.51.85.197 - 1/19/2020 12:41:31 PM
REFERENCES
HUMAN LEUKEMIC CELLS
IN
VITRO 241
Downloaded by: Université de Paris 193.51.85.197 - 1/19/2020 12:41:31 PM
5 GREENBERG, P. L., NICHOLS, W. C., and SCHRIER, S. L. Granulopoiesis in acute myeloid leukemia and preleukemia. New Engl. J. Med., 284: 1225 (1971). 6 IcHIKAWA, Y., PLUZNIK, D. H., and SAGS, L. Ιn vitro control of the development of macrophage and granulocyte colonies. Proc. Natl. Acad. Sci. U.S., 56: 488 (1966). 7 ISCOVE, N. N., SEIN, J. S., TILL, J. E., and McCuLLOCH, E. A. Colony formation by normal and leukemic marrow cells in culture: Effect of conditioned medium from human leukocytes. Blood, 37: 1 (1971). 8 METCALF, D. Regulation by colony stimulating factor (CSF) of granulocyte and macrophage colony formation in vitro by normal and leukemic cells. In "Control of Proliferation in Animal Cells," ed. by Β. Clarkson and R. Baserga, Cold Spring Harbor Laboratories, New York (1973), in press. 9 METCALF, D., CHAN, S. H., GUwz, F. W., VINCENT, Ρ., and RAVICH, R. Β. M. Colony stimulating factor and inhibitor levels in acute granulocytic leukemia. Blood, 38: 143 (1971). 10 Μοο , M. A. S. In vitro studies in myeloid leukemias. In "Excerpts Medica Reviews in Leukemia and Lymphoma," ed. by F. J. Cleton, D. Crowther, and J. S. Maleas, Excerpts Medica, Amsterdam (1973), in press. 11 MOORE, M. A. S. and METCALF, D. Cytogenetic analysis of human acute and chronic myeloid leukemic cells cloned in agar cultures. Int. J. Cancer, 11: 143 (1973). 12 MOORE, M. A. S., SPITZER, G., WILLIAMS, N., and METCALF, D. Correlation of agar culture analysis and clinical status in patients with acute myeloid leukemia. Ιn "Hemopoiesis in Culture" (Proc. of 2nd Int. Workshop on Hemopoiesis in Culture), ed. by W. A. Robinson, U.S. Government Printing Office, Washington D.C. (1973). 13 MOORE, M. A. S. and WILLIAMS, N. Functional, morphologic and kinetic analysis of the granulocyte progenitor cell. In "Hemopoiesis in Culture" (Proc. of 2nd Int. Workshop on Hemopoiesis in Culture), ed. by W. A. Robinson, U.S. Government Printing Office, Washington D. C. (1973). 14 MooRs, M. A. S., WILLIAMS, N., and METCALF, D. In vitro colony formation by normal and leukemic human hemopoietic cells: Characterisation of the colony forming cells. J. Natl. Cancer Inst., 50: 603 (1973). 15 MOORE, M. A. S., WILLIAMS, N., and METCALF, D. In vitro colony formation by normal and leukemic human hemopoietic cells: Interaction between colony-forming and colonystimulating cells. J. Natl. Cancer Inst., 50: 591 (1973). 16 PARAN, M., SACHS, L., BAιιλκ, Y., and RESNITSKY, P. Ιn vitro induction of granulocyte differentiation in hematopoietic cells from leukemic and non-leukemic patients. Proc. Natl. Acad. Sci. U.S., 67: 1542 (1970). 17 PIKE, B. L. and ROBINSON, W. A. Human bone marrow colony growth in agar-gel. J. Cell. Physiol., 76: 77 (1970). 18 ROBINSON, W. A. and PIKE, B. L. Leucopoietic activity in human urine: The granulocytic leukemias. New Engl. J. Med., 282: 1291 (1970).
