Different Recovery Patterns of Mouse Haemopoietic Stem Cells in Response to Cytotoxic Agents N . M. BLACKETT AND R. E . MILLARD Biophysics Division, In stitute of ' Cancer Research, Sutton, Surrey, England SM2 5 P X
The response and subsequent recovery of mouse haemopoietic ABSTRACT progenitor cells (spleen colony forming cells and agar colony forming cells) has been studied following two cytotoxic agents. Busulphan was administered to normal mice and vinblastine to mice where the progenitor cell proliferation rate had been increased by a period of continuous y-irradiation. With both these agents there is a difference between the response of the spleen colony forming cells and the agar colony forming cells during the first five days. They then recover together, but much more slowly after busulphan than after vinblastine even though their proliferation rate is increased. The rate of progenitor cell recovery after busulphan is increased if the progenitor cells are depleted further by vinblastine. However, methotrexate, which severely depletes the peripheral blood count and bone marrow cellularity but not the progenitor cells, has no effect on the recovery following busulphan. These results suggest that following cytotoxic agents the agar colony forming cells ("committed" stem cells) are not self-maintaining but are dependent on a supply of cells from the pluripotential spleen colony forming cells. In addition it appears that the depletion of the progenitor cells of the bone marrow and not the depletion of the maturing cells, provides a stimulus for stem cell recovery.
Several groups of workers have compared the effect of different cytotoxic agents upon mouse haemopoietic progenitor cells, for example on spleen colony forming cells (CFU), agar colony forming (ACU) and erythroid repopulating cells (Bruce et al., '66; Eaves and Bruce, '72; Blackett and Adams, '72; Millard et al., '73). In these studies cell survival was measured 4-24 hours after administration of the agent, and the chief object of the experiments was to compare the dose response curves of the different cell populations. However, there have been comparatively few studies of the recovery of the progenitor cells following cytotoxic agents, and how this is related to the number of mature cells in the bone marrow and peripheral blood, and also to the recovery of mature blood cell production. Observations on the recovery of haemopoietic tissue may be able to provide information on the interrelationship between different progenitor cell populations, and may clarify the role they play in the recovery of mature blood cell production. Furthermore, the changes occurring during J CELL. PHYSIOL.,89: 4 7 3 4 8 0
recovery are likely to influence the effect of subsequent administration of the same or some other cytotoxic agent. Previous workers (Chen and Schooley, '72) have observed a rapid recovery of both CFU and ACU following vinblastine, compared with a slow recovery of CFU, but a rapid recovery of ACU following whole body y-irradiation. The recovery of CFU following busulphan appears to be even slower and more prolonged (Udapa et al., '72; Josvasen and Boyum, '73); although Reissmann and Udapa ('72) have shown that after busulphan, red cell production can be maintained by injection of erythropoietin while the stem cell level is very low. In the present studies we have investigated the recovery of haemopoietic tissue (progenitor cells and mature blood cells) following busulphan and also after vinblastine. These agents were chosen because each of them has a differential effect upon CFU and ACU (Millard et al., '73; Blackett and Millard, '73), thereby providing a means of investigating the reReceived Nov. 1 1 , ' 7 5 . Accepted Feb. 27, '76
4 73
474
N. M. BLACKETT A N D R. E. MILLARD
ied in mice which had been irradiated continuously for a period of two to three weeks and then given 10 mg/kg vinblastine. The continuous irradiation depletes the stem cells to about 10% of normal and causes a n increase in their rate of proliferation, so that vinblastine has a much greater effect than in normal mice (Millard MATERIALS A N D METHODS et al., ’73). The assays were carried out at intervals from 12 hours to 27 days €01C5,BL mice (10-12 weeks old) were used throughout. The effect of cytotoxic agents lowing vinblastine, and the results for was studied in male mice and female mice bone marrow are presented in figure 1. The were used as recipients for the spleen col- survivals for ACU and CFU are compared ony assay. Continuous ?-irradiation was with the control values for normal, unirfrom a Cs’37 source, at a dose-rate of 38 radiated mice. The results show that CFU are depressed rads per day for a period of two to three weeks. Vinblastine dissolved in saline was to very low values; the nadir of 0.0001 of administered intravenously. Busulphan normal is reached two days after vinblaswas dissolved i n dimethylsulphoxide and tine. Nevertheless, recovery is very rapid, injected intraperitoneally as an emulsion reaching 0.01 survival by day 6, 0.1 by day in arachis oil. Cytosine arabinoside was 9 and the normal value by about day 17. dissolved in saline and injected intraperi- The CFU doubling time between days 2 toneally. Methotrexate was dissolved in 2 % and 9 is about 17 hours. The results for ACU show rather more sodium bicarbonate and injected intraperitoneally. scatter than the CFU results, particularly Red cell production was measured by during the first three days. Initially, ACU incorporation of FeSYinto red cells in pe- survival falls to considerably lower values ripheral blood, following the method of than CFU survival. At 12 hours and at one Blackett and Adams (‘72). The activity was day there was less than one colony per measured two days after the injection of plate for the maximum number of cells that Fe”. could be plated (-lo6 cells), so an upper Agar colony forming cells in bone mar- limit for the survival is indicated assuming row and in spleen were assayed by a meth- a n average of one colony per plate. From od previously described (Millard et al., ’73), day 1 to day 2 there is a rapid rise in ACU. the source of colony stimulating factor be- The ACU level appears to be higher than ing pooled serum from endotoxin treated the CFU level between day 2 and day 3 as mice (Metcalf, ’71). in only one of the eight groups of animals All the results have been expressed as a assayed during this interval was the ACU proportion of the values obtained for un- value lower than the CFU value. From day treated control animals. For the spleen 4 the ACU survival follows the CFU recovcolony assay values for normal animals in ery curve. the various experiments were in the range The expected rapid proliferation rate of 5,0004,500 colonies per femur while for ACU and CFU during recovery was conthe “in vitro” agar colony assay they were firmed at day 5 by experiments with cytobetween 21,000 and 29,000 colonies per sine arabinoside given repeatedly over a femur. The radioactive iron incorporation period of six hours (3 injections at 2-hourinto red blood cells for normal animals was ly intervals). Cytosine arabinoside adminbetween 30% and 40% of the injected istered in this way to normal mice gave a dose. survival of 0.78 t 0.02 for ACU and 0.47 t 0.06 for CFU. Five days after vinblastine, RESULTS the survivals were very much lower, namely 0.02 -+ 0.01 for ACU and 0.013 ? 0.08 Recovery following vinblastine for CFU, which indicates that the proliferStem cells ation rate is greatly increased. The recovery of ACU and CFU was studAssays for ACU and CFU were also done
lationship between these cell populations when one is depressed more than the other. Vinblastine was administered to mice which had received prior y-irradiation; this allowed recovery to be studied when the progenitor cells had been depressed to extremely low levels.
CYTOTOXICITY AND HAEMOPOIETIC STEM CELL RECOVERY
4 75
* 0
2
4
6
8
1
0
2' 0
15
-/
D A Y S
Fig. 1 Recovery of ACU ( o - - - O ) and CFU (0-0) in bone marrow following administration of vinblastine (10 mg/kg) to mice that had received continuous y-irradiation (38 radslday) during the previous two weeks. In addition recovery of radioactive iron incorporation into red blood cells ( A -- -A ). All results expressed as a fraction of normal untreated control animals. The vertical bars represent -1- one standard error.
0 2
4
6
10
8
14
18
D A Y S
Change i n bone marrow cellularity (0-0) and peripheral blood white cell count (0- - - 0)following administration of vinblastine (10 mglkg) to mice that had received continuous y-irradiation. Vertical bars represent k one standard error. Fig. 2
on the spleen following vinblastine (data not presented). The initial depression (0.01 survival) was less than for bone marrow. A pronounced overshoot to several times normal was observed at nine days for both ACU and CFU, and a return to normal by 27 days. The levels of ACU and CFU in the spleen during recovery were about a decade higher than those in bone marrow.
Red cell production Red cell production as a fraction of normal control value is shown i n figure 1. This is depressed to 0.01 of normal, thereafter recovery is rapid, returning to normal by about day 10. Bone marrow and blood cell counts Figure 2 shows the change in bone marrow cellularity and total white blood cells
4 76
N. M. BLACKETT AND R. E. MILLARD ,ANd
1"
10
30
40
D A Y S
in bone marrow following adminFig. 3 Recovery of ACU (0- - - C )and CFU (0-0) istration of busulphan (40 mglkg) Also recovery of radioactive iron incorporation into red blood cells ( A - - -A ) . The vertical bars represent t one standard error.
as a fraction of normal control values. Differential counts on blood smears showed that there was no significant change in the proportion of granulocytes. The circulating leukocytes and bone marrow cellularity are depressed to about 20% of normal within two days and there is no recovery until after six days, normal values being regained by about 14 days. Recovery following busulphan Stem cells The recovery of ACU and CFU in the bone marrow following a single injection of 40 mglkg busulphan is presented in figure 3. It can be seen that CFU are depressed to about 0.02 of normal, and individual measurements fluctuate between 0.01 and 0.04 until about day 16. Thereafter, there is a very slow recovery, and even at the longest time interval studied (42 days), CFU have not returned to normal. The ACU are initially less depressed than CFU (to about 0.25of normal) but fall rapidly until they reach the same level as CFU at day 5. The ACU then show the same slow recovery as the CFU. In general the ACU survivals tend to be higher than the CFU survivals. The considerable scatter of ACU and CFU survivals could be due to difficulty in dissolving the busulphan and a variable re-
lease of the cytotoxic agent from the arachis oil-dimethylsulphoxide emulsion. The effect of busulphan on the spleen was found to be greater, the ACU being depressed to about 0.02 and CFU to about 0.005of normal at two days, and recovery was more rapid, normal values being reached by 20 days, the ACU (but not CFU) showing an overshoot to about four times normal at 25 days. Both assays gave normal values at 35 and 42 days. Repeated cytosine arabinoside was given at different time intervals after busulphan, in order to determine whether the stem cells were proliferating more rapidly than normal. The results for bone marrow and spleen are presented in figure 4. The height of the columns represents the percentage of cells killed per hour by repeated cytosine arabinoside, which was administered every two hours over a period of between six and ten hours. The results show that for both CFU and ACU there is increased cell la11 in the marrow at 5 and 14 days and in the spleen at 14 days, compared to normal control animals. The proliferation rate is therefore increased in the busulphan treated animals. By day 26 and day 42, there is no difference between the treated and the control group, showing that the proliferation rate of ACU and CFU has returned to norm a1.
CYTOTOXICITY AND HAEMOPOIETIC STEM CELL RECOVERY
477
4
14
'6
42
D A Y S
Fig. 4 Effect of repeated administration of cytosine arabinoside (200 mglkg per injection) every two hours for six to ten hours expressed as the percentage of cells killed per hour, at various intervals after busulphan (40 mg/kg). Shaded column busulphan-treated mice, unshaded column controls. Vertical bars represent k one standard error.
Red cell production This is presented i n figure 3 and shows a rapid return to normal, compared to the delayed recovery of the stem cells. Bone marrow cellularity and peripheral blood counts These are not shown graphically as there was little depression from normal levels; the bone marrow nucleated cell count, the total white cell count and granulocyte count i n peripheral blood all fall to about 0.7-0.8 of normal and remain at this level for about 26 days.
TABLE 1
Incorporation of 5QFe( % of injected dose) i n t o circulating red blood cells after b u s u l p h a n
Controls Busulphan 4 days 6 days 8 days 10 days 12 days
+
Intact mice
Splenectomised mice
34.2 f 2 . 2
26.9k 2.6
23.1 f 5.2 28.0 2 2 . 3 48.6 k 2.9 47.5 f 1.8 52.2 f 1.7
37.5 f 12.3 2 9 . 9 2 6.5 6 3 . 4 k 5.1 4 9 . 7 f 2.9 51.9 k 8.0
Recovery of ACU and CFU after busulphan Busulphan
+ 14 days ACU 0.08 k O . 0 1 0.10 k 0 . 0 1 3 Effect of splenectomy CFU 0.04 f 0.006 0.05 2 0.007 To investigate the role of the spleen + 24 days ACU 0.48 f0.006 0.69 2 0.09 0.25 f 0.03 CFU 0.23 f 0 03 during recovery after busulphan the incorporation of 59Fe into circulating red blood cells was measured in intact mice splenic erythropoiesis i n normal animals. and mice splenectomised one day before In the busulphan treated animals, splenecthe injection of iron. In addition, ACU and tomy did not reduce the 59Fe uptake but at CFU were measured 14 and 24 days after eight days there was a significant increase. busulphan in intact mice and in mice The ACU and CFU measurements show splenectomised seven days before the bu- that splenectomy has no effect on stem sulphan administration. cell recovery. The results presented i n table 1 show These results indicate that in busulphan that splenectomy slightly reduced the in- treated mice it is the bone marrow rather corporation of "Fe in the untreated con- than the spleen which is responsible for trols, as expected due to the low level of the recovery of red cell production.
478
N. M. BLACKETT AND R. E. MILLARD
5
15
10
20
D A Y S
Fig. 5 Recovery of ACU and CFU in busulphan (40 mg/kg) treated mice given either vinblastine (10 mg/kg) or methotrexate (200 mg/kg) five days later. Busulphan only, CFU (0). Busulphan methotrexate, CFU (0- - - 0). Busulphan vinblastine, CFU (0- - -0) and ACU (A-A). Vertical bars represent ? one standard error.
+
+
5.
10
15
20
D A Y S
Fig. 6 Change in bone marrow cellularity ( o - - - O ) and peripheral blood white cell count (0- - -0) following administration of methotrexate. Vertical bars represent k one standard error.
The effect of vinblastine or methotrexate after busulphan The state of haemopoietic tissue in busulphan treated mice was further studied by administration of either vinblastine or methotrexate five days after busulphan. Vinblastine was employed to deplete the rapidly proliferating stem cells still further, while methotrexate was employed to depress the maturing cells but not the stem cells (Millard et al., '73). Figure 5 shows that vinblastine reduces
survival of ACU and CFU to levels between 0.0001 and 0.001 survival with a subsequent rapid recovery to normal levels at about 18 days, so that the curve crosses the curve for busulphan alone at about 12 days. This response to busulphan plus vinblastine is very similar to that observed in continuously irradiated mice given vinblastine. Methotrexate given five days after busulphan did not depress the stem cells further and had no influence on the slow recovery
CYTOTOXICITY AND HAEMOPOIETIC STEM CELL RECOVERY
(fig, 5). Methotrexate, however, has a marked effect on the maturing haernopoietic cells since the bone marrow cellularity and the peripheral nucleated cells (fig. 6) are depressed. DISCUSSION
Two quite different responses of haemopoietic tissue to cytotoxic agents have been demonstrated. During the first four days after administering vinblastine to previously irradiated mice, the ACU were initially depleted to a much greater extent than CFU, but then increased rapidly (by about 50 times between day 1 and day 2) until they reached the CFU level. This increase is too rapid to be accounted for by cell proliferation, but could be due to enhanced differentiation of CFU into ACU because during this period the CFU were decreasing but ceased to do so once the ACU had reached the CFU level. In fact the ACU appear to overshoot the CFU level before the two cell populations recover more or less together. The rapid rise in ACU between day 1 and day 2 might also be explained by the repair of potentially lethal damage but then some other explanation would be required for the loss of CFU during this period. Furthermore one would also expect CFU to exhibit potentially lethal damage but this would have to be masked by the loss of CFU. In contrast, to these results the initial effect of busulphan on ACU is less than on CFU, and the ACU level then decreases until it reaches the CFU level at about day 5 . These two responses suggest that the ACU population is not self-maintaining after drug treatment and relies on cells being fed i n from the CFU population. The results also suggest that the rate of differentiation of CFU into ACU is not constant, being temporarily increased after administration of vinblastine to continuously irradiated mice but not after administration of busulphan. After the first five days CFU and ACU maintain similar levels during recovery but the rate of stem cell recovery is quite different for the two drug treatments. In the vinblastine-treated irradiated mice recovery is rapid from a survival of about 0.0001, with a doubling time of about 17 hours between two and nine days. Recovery following busulphan is very much slower, even though the initial depletion
4 79
is very much less, and normal levels are not reached until after 40 days, which corresponds to a doubling time of about eight days. Although after busulphan the stem cells i n the spleen recover sooner than those in the bone marrow, the spleen does not make a major contribution to haemopoiesis since splenectomy does not alter the rate of recovery of red cell production as measured by 59Fe incorporation, nor does it affect the recovery of stem cells in the bone marrow. The slow recovery after busulphan is not due to a failure of the stem cells to increase their rate of proliferation because the results with repeated cytosine arabinoside show that these cells are proliferating rapidly. Neither is it solely due to the small depletion of the stem cells since, with other cytotoxic agents at doses that give a comparable initial stem cell depletion, recovery is much more rapid, for example following BCNU, CCNU and MeCCNU (Blackett et al., '76), cyclophosphamide, and also vinblastine without prior irradiation (unpublished). The stem cells are not incapable of rapid recovery after busulphan because further depletion to very low levels, by administration of vinblastine five days after busulphan, produces a very rapid recovery which crosses the busulphan-only curve at about 12 days. This rapid recovery is likely to be due to the greater stem cell depletion rather than a depletion of the maturing cells causing a feedback stimulation of the stem cells. This conclusion is supported by the measurements with methotrexate which when administered after busulphan, produces no further depression of the stem cells and no change in the recovery rate, although the maturing and peripheral blood cells are severely depleted. Recovery from lower stem cell levels could not be determined by using higher doses of busulphan because ofits short term toxicity, and so it is not possible to say whether greater depletion of stem cells by busulphan alone would have produced a more rapid recovery. This has been shown to occur with radiation (Guzman and Lajtha, '70). It is evident from our results that the rate of recovery of stem cells depends to some extent on the drug employed as well as on the degree of stem cell depletion, but
480
N . M. BLACKETT AND R. E. MILLARD
not apparently on the depletion of the maturing cells. The recovery of mature cell production, as shown by peripheral blood counts and red cell Fes9incorporation, does not depend on the stem cells reaching some specified level as has been suggested by Chervenick and Boggs (‘71), and there must be marked changes in the proliferation kinetics of the maturing cells to compensate for the reduction in stem cell numbers. Although no explanation has been found for the indolent recovery of stem cells following busulphan, it is not due to the inability of the stem cells to proliferate rapidly, nor is it due to a n irreversible derangement of the population size control, since further depletion of the stem cells with vinblastine brings about a rapid recovery. ACKNOWLEDGMENTS
The authors wish to thank Professor L. F. Lamerton for encouragement in carrying out the work described and for many useful discussions of the results. Mrs. S. F. OKell and Mrs. H. M. Belcher provided excellent technical assistance. LITERATURE CITED Blackett, N. M., a n d K. Adams 1972 Cell proliferation and the action of cytotoxic agents on haemopoietic tissue. Brit. J. Haematol., 23: 751758. Blackett, N . M., V . D. Courtenay and S . M. Mayer 1976 Differential sensitivitv to three nitrosoureas of colony forming cells for haemopoietic tissue, Lewis Lung and B16 melanoma tumours in C57BL mice. Cancer Chemotherapy Reports, 5 . Q . 929-933. _ ~ .. .
Blackett, N. M., and R. E. Millard
1973
Differ-
ential effect of Myleran on two normal haemopoietic progenitor cell populations. Nature (London), 2 4 4 : 300-301. Bruce, W . R., B. E. Meeker and F. A. Valeriote 1966 Comparison of the sensitivity of normal haemopoietic and transplanted lymphoma colony forming cells to chemotherapeutic agents administered ‘in vivo’. J. Nat. Cancer Inst., 37: 233-245. Chen, M. G., and J . C. Schooley 1970 Recovery of proliferative capacity of agar colony forming cells and spleen colony forming cells following ionising radiation and vinblastine. J. Cell. Physiol., 75: 89-95. Chervenick, P. A., and D. R. Boggs 1971 Patterns of proliferation and differentiation of hematopoietic stem cells after compartment depletion. Blood, 37: 5 6 8 5 8 0 . Eaves, A. C., and W. R. Bruce 1974 Altered sensitivity of haematopoietic stem cells to 5fluorouracil following endotoxin, cyclophosphamide and irradiation. Cancer Chemotherapy Reports, 5 8 : 813-820. Guzman, E., and L. G. Lajtha 1970 Some comparisons of the kinetic properties of femoral and splenic haemopoietic stem cells. Cell and Tissue Kinet., 3: 91-98. Josvasen, N . , and A. Boyum 1973 Haematopoiesis i n busulphan-treated mice. Scand. J. Haemat., 1 1 : 78-83. Metcalf, D. 1971 Acute antigen-induced elevation of serum colony stimulating factor (CSF) levels. Immunology, 21 : 427-436. Millard, R. E., N. M . Blackett and S. F. Okell 1973 A comparison of the effect of cytotoxic agents on agar colony forming cells, spleen colony forming cells and the erythrocytic repopulating ability of mouse bone marrow. J. Cell Physiol., 8 2 : 309-317. Reissmann, K. R., and K. B. Udupa 1972 Effect of erythropoietin on proliferation of erythropoietin-responsive cells. Cell and Tissue Kinet., 5 : 481-489. Udapa, K. B., H. Okamura and K . R. Reissmann 1972 Granulopoiesis during Myleran-induced summession of transdantable hematoDoietic stem cells. Blood, 39: 3i7-325.
The response and subsequent recovery of mouse haemopoietic ABSTRACT progenitor cells (spleen colony forming cells and agar colony forming cells) has been studied following two cytotoxic agents. Busulphan was administered to normal mice and vinblastine to mice where the progenitor cell proliferation rate had been increased by a period of continuous y-irradiation. With both these agents there is a difference between the response of the spleen colony forming cells and the agar colony forming cells during the first five days. They then recover together, but much more slowly after busulphan than after vinblastine even though their proliferation rate is increased. The rate of progenitor cell recovery after busulphan is increased if the progenitor cells are depleted further by vinblastine. However, methotrexate, which severely depletes the peripheral blood count and bone marrow cellularity but not the progenitor cells, has no effect on the recovery following busulphan. These results suggest that following cytotoxic agents the agar colony forming cells ("committed" stem cells) are not self-maintaining but are dependent on a supply of cells from the pluripotential spleen colony forming cells. In addition it appears that the depletion of the progenitor cells of the bone marrow and not the depletion of the maturing cells, provides a stimulus for stem cell recovery.
Several groups of workers have compared the effect of different cytotoxic agents upon mouse haemopoietic progenitor cells, for example on spleen colony forming cells (CFU), agar colony forming (ACU) and erythroid repopulating cells (Bruce et al., '66; Eaves and Bruce, '72; Blackett and Adams, '72; Millard et al., '73). In these studies cell survival was measured 4-24 hours after administration of the agent, and the chief object of the experiments was to compare the dose response curves of the different cell populations. However, there have been comparatively few studies of the recovery of the progenitor cells following cytotoxic agents, and how this is related to the number of mature cells in the bone marrow and peripheral blood, and also to the recovery of mature blood cell production. Observations on the recovery of haemopoietic tissue may be able to provide information on the interrelationship between different progenitor cell populations, and may clarify the role they play in the recovery of mature blood cell production. Furthermore, the changes occurring during J CELL. PHYSIOL.,89: 4 7 3 4 8 0
recovery are likely to influence the effect of subsequent administration of the same or some other cytotoxic agent. Previous workers (Chen and Schooley, '72) have observed a rapid recovery of both CFU and ACU following vinblastine, compared with a slow recovery of CFU, but a rapid recovery of ACU following whole body y-irradiation. The recovery of CFU following busulphan appears to be even slower and more prolonged (Udapa et al., '72; Josvasen and Boyum, '73); although Reissmann and Udapa ('72) have shown that after busulphan, red cell production can be maintained by injection of erythropoietin while the stem cell level is very low. In the present studies we have investigated the recovery of haemopoietic tissue (progenitor cells and mature blood cells) following busulphan and also after vinblastine. These agents were chosen because each of them has a differential effect upon CFU and ACU (Millard et al., '73; Blackett and Millard, '73), thereby providing a means of investigating the reReceived Nov. 1 1 , ' 7 5 . Accepted Feb. 27, '76
4 73
474
N. M. BLACKETT A N D R. E. MILLARD
ied in mice which had been irradiated continuously for a period of two to three weeks and then given 10 mg/kg vinblastine. The continuous irradiation depletes the stem cells to about 10% of normal and causes a n increase in their rate of proliferation, so that vinblastine has a much greater effect than in normal mice (Millard MATERIALS A N D METHODS et al., ’73). The assays were carried out at intervals from 12 hours to 27 days €01C5,BL mice (10-12 weeks old) were used throughout. The effect of cytotoxic agents lowing vinblastine, and the results for was studied in male mice and female mice bone marrow are presented in figure 1. The were used as recipients for the spleen col- survivals for ACU and CFU are compared ony assay. Continuous ?-irradiation was with the control values for normal, unirfrom a Cs’37 source, at a dose-rate of 38 radiated mice. The results show that CFU are depressed rads per day for a period of two to three weeks. Vinblastine dissolved in saline was to very low values; the nadir of 0.0001 of administered intravenously. Busulphan normal is reached two days after vinblaswas dissolved i n dimethylsulphoxide and tine. Nevertheless, recovery is very rapid, injected intraperitoneally as an emulsion reaching 0.01 survival by day 6, 0.1 by day in arachis oil. Cytosine arabinoside was 9 and the normal value by about day 17. dissolved in saline and injected intraperi- The CFU doubling time between days 2 toneally. Methotrexate was dissolved in 2 % and 9 is about 17 hours. The results for ACU show rather more sodium bicarbonate and injected intraperitoneally. scatter than the CFU results, particularly Red cell production was measured by during the first three days. Initially, ACU incorporation of FeSYinto red cells in pe- survival falls to considerably lower values ripheral blood, following the method of than CFU survival. At 12 hours and at one Blackett and Adams (‘72). The activity was day there was less than one colony per measured two days after the injection of plate for the maximum number of cells that Fe”. could be plated (-lo6 cells), so an upper Agar colony forming cells in bone mar- limit for the survival is indicated assuming row and in spleen were assayed by a meth- a n average of one colony per plate. From od previously described (Millard et al., ’73), day 1 to day 2 there is a rapid rise in ACU. the source of colony stimulating factor be- The ACU level appears to be higher than ing pooled serum from endotoxin treated the CFU level between day 2 and day 3 as mice (Metcalf, ’71). in only one of the eight groups of animals All the results have been expressed as a assayed during this interval was the ACU proportion of the values obtained for un- value lower than the CFU value. From day treated control animals. For the spleen 4 the ACU survival follows the CFU recovcolony assay values for normal animals in ery curve. the various experiments were in the range The expected rapid proliferation rate of 5,0004,500 colonies per femur while for ACU and CFU during recovery was conthe “in vitro” agar colony assay they were firmed at day 5 by experiments with cytobetween 21,000 and 29,000 colonies per sine arabinoside given repeatedly over a femur. The radioactive iron incorporation period of six hours (3 injections at 2-hourinto red blood cells for normal animals was ly intervals). Cytosine arabinoside adminbetween 30% and 40% of the injected istered in this way to normal mice gave a dose. survival of 0.78 t 0.02 for ACU and 0.47 t 0.06 for CFU. Five days after vinblastine, RESULTS the survivals were very much lower, namely 0.02 -+ 0.01 for ACU and 0.013 ? 0.08 Recovery following vinblastine for CFU, which indicates that the proliferStem cells ation rate is greatly increased. The recovery of ACU and CFU was studAssays for ACU and CFU were also done
lationship between these cell populations when one is depressed more than the other. Vinblastine was administered to mice which had received prior y-irradiation; this allowed recovery to be studied when the progenitor cells had been depressed to extremely low levels.
CYTOTOXICITY AND HAEMOPOIETIC STEM CELL RECOVERY
4 75
* 0
2
4
6
8
1
0
2' 0
15
-/
D A Y S
Fig. 1 Recovery of ACU ( o - - - O ) and CFU (0-0) in bone marrow following administration of vinblastine (10 mg/kg) to mice that had received continuous y-irradiation (38 radslday) during the previous two weeks. In addition recovery of radioactive iron incorporation into red blood cells ( A -- -A ). All results expressed as a fraction of normal untreated control animals. The vertical bars represent -1- one standard error.
0 2
4
6
10
8
14
18
D A Y S
Change i n bone marrow cellularity (0-0) and peripheral blood white cell count (0- - - 0)following administration of vinblastine (10 mglkg) to mice that had received continuous y-irradiation. Vertical bars represent k one standard error. Fig. 2
on the spleen following vinblastine (data not presented). The initial depression (0.01 survival) was less than for bone marrow. A pronounced overshoot to several times normal was observed at nine days for both ACU and CFU, and a return to normal by 27 days. The levels of ACU and CFU in the spleen during recovery were about a decade higher than those in bone marrow.
Red cell production Red cell production as a fraction of normal control value is shown i n figure 1. This is depressed to 0.01 of normal, thereafter recovery is rapid, returning to normal by about day 10. Bone marrow and blood cell counts Figure 2 shows the change in bone marrow cellularity and total white blood cells
4 76
N. M. BLACKETT AND R. E. MILLARD ,ANd
1"
10
30
40
D A Y S
in bone marrow following adminFig. 3 Recovery of ACU (0- - - C )and CFU (0-0) istration of busulphan (40 mglkg) Also recovery of radioactive iron incorporation into red blood cells ( A - - -A ) . The vertical bars represent t one standard error.
as a fraction of normal control values. Differential counts on blood smears showed that there was no significant change in the proportion of granulocytes. The circulating leukocytes and bone marrow cellularity are depressed to about 20% of normal within two days and there is no recovery until after six days, normal values being regained by about 14 days. Recovery following busulphan Stem cells The recovery of ACU and CFU in the bone marrow following a single injection of 40 mglkg busulphan is presented in figure 3. It can be seen that CFU are depressed to about 0.02 of normal, and individual measurements fluctuate between 0.01 and 0.04 until about day 16. Thereafter, there is a very slow recovery, and even at the longest time interval studied (42 days), CFU have not returned to normal. The ACU are initially less depressed than CFU (to about 0.25of normal) but fall rapidly until they reach the same level as CFU at day 5. The ACU then show the same slow recovery as the CFU. In general the ACU survivals tend to be higher than the CFU survivals. The considerable scatter of ACU and CFU survivals could be due to difficulty in dissolving the busulphan and a variable re-
lease of the cytotoxic agent from the arachis oil-dimethylsulphoxide emulsion. The effect of busulphan on the spleen was found to be greater, the ACU being depressed to about 0.02 and CFU to about 0.005of normal at two days, and recovery was more rapid, normal values being reached by 20 days, the ACU (but not CFU) showing an overshoot to about four times normal at 25 days. Both assays gave normal values at 35 and 42 days. Repeated cytosine arabinoside was given at different time intervals after busulphan, in order to determine whether the stem cells were proliferating more rapidly than normal. The results for bone marrow and spleen are presented in figure 4. The height of the columns represents the percentage of cells killed per hour by repeated cytosine arabinoside, which was administered every two hours over a period of between six and ten hours. The results show that for both CFU and ACU there is increased cell la11 in the marrow at 5 and 14 days and in the spleen at 14 days, compared to normal control animals. The proliferation rate is therefore increased in the busulphan treated animals. By day 26 and day 42, there is no difference between the treated and the control group, showing that the proliferation rate of ACU and CFU has returned to norm a1.
CYTOTOXICITY AND HAEMOPOIETIC STEM CELL RECOVERY
477
4
14
'6
42
D A Y S
Fig. 4 Effect of repeated administration of cytosine arabinoside (200 mglkg per injection) every two hours for six to ten hours expressed as the percentage of cells killed per hour, at various intervals after busulphan (40 mg/kg). Shaded column busulphan-treated mice, unshaded column controls. Vertical bars represent k one standard error.
Red cell production This is presented i n figure 3 and shows a rapid return to normal, compared to the delayed recovery of the stem cells. Bone marrow cellularity and peripheral blood counts These are not shown graphically as there was little depression from normal levels; the bone marrow nucleated cell count, the total white cell count and granulocyte count i n peripheral blood all fall to about 0.7-0.8 of normal and remain at this level for about 26 days.
TABLE 1
Incorporation of 5QFe( % of injected dose) i n t o circulating red blood cells after b u s u l p h a n
Controls Busulphan 4 days 6 days 8 days 10 days 12 days
+
Intact mice
Splenectomised mice
34.2 f 2 . 2
26.9k 2.6
23.1 f 5.2 28.0 2 2 . 3 48.6 k 2.9 47.5 f 1.8 52.2 f 1.7
37.5 f 12.3 2 9 . 9 2 6.5 6 3 . 4 k 5.1 4 9 . 7 f 2.9 51.9 k 8.0
Recovery of ACU and CFU after busulphan Busulphan
+ 14 days ACU 0.08 k O . 0 1 0.10 k 0 . 0 1 3 Effect of splenectomy CFU 0.04 f 0.006 0.05 2 0.007 To investigate the role of the spleen + 24 days ACU 0.48 f0.006 0.69 2 0.09 0.25 f 0.03 CFU 0.23 f 0 03 during recovery after busulphan the incorporation of 59Fe into circulating red blood cells was measured in intact mice splenic erythropoiesis i n normal animals. and mice splenectomised one day before In the busulphan treated animals, splenecthe injection of iron. In addition, ACU and tomy did not reduce the 59Fe uptake but at CFU were measured 14 and 24 days after eight days there was a significant increase. busulphan in intact mice and in mice The ACU and CFU measurements show splenectomised seven days before the bu- that splenectomy has no effect on stem sulphan administration. cell recovery. The results presented i n table 1 show These results indicate that in busulphan that splenectomy slightly reduced the in- treated mice it is the bone marrow rather corporation of "Fe in the untreated con- than the spleen which is responsible for trols, as expected due to the low level of the recovery of red cell production.
478
N. M. BLACKETT AND R. E. MILLARD
5
15
10
20
D A Y S
Fig. 5 Recovery of ACU and CFU in busulphan (40 mg/kg) treated mice given either vinblastine (10 mg/kg) or methotrexate (200 mg/kg) five days later. Busulphan only, CFU (0). Busulphan methotrexate, CFU (0- - - 0). Busulphan vinblastine, CFU (0- - -0) and ACU (A-A). Vertical bars represent ? one standard error.
+
+
5.
10
15
20
D A Y S
Fig. 6 Change in bone marrow cellularity ( o - - - O ) and peripheral blood white cell count (0- - -0) following administration of methotrexate. Vertical bars represent k one standard error.
The effect of vinblastine or methotrexate after busulphan The state of haemopoietic tissue in busulphan treated mice was further studied by administration of either vinblastine or methotrexate five days after busulphan. Vinblastine was employed to deplete the rapidly proliferating stem cells still further, while methotrexate was employed to depress the maturing cells but not the stem cells (Millard et al., '73). Figure 5 shows that vinblastine reduces
survival of ACU and CFU to levels between 0.0001 and 0.001 survival with a subsequent rapid recovery to normal levels at about 18 days, so that the curve crosses the curve for busulphan alone at about 12 days. This response to busulphan plus vinblastine is very similar to that observed in continuously irradiated mice given vinblastine. Methotrexate given five days after busulphan did not depress the stem cells further and had no influence on the slow recovery
CYTOTOXICITY AND HAEMOPOIETIC STEM CELL RECOVERY
(fig, 5). Methotrexate, however, has a marked effect on the maturing haernopoietic cells since the bone marrow cellularity and the peripheral nucleated cells (fig. 6) are depressed. DISCUSSION
Two quite different responses of haemopoietic tissue to cytotoxic agents have been demonstrated. During the first four days after administering vinblastine to previously irradiated mice, the ACU were initially depleted to a much greater extent than CFU, but then increased rapidly (by about 50 times between day 1 and day 2) until they reached the CFU level. This increase is too rapid to be accounted for by cell proliferation, but could be due to enhanced differentiation of CFU into ACU because during this period the CFU were decreasing but ceased to do so once the ACU had reached the CFU level. In fact the ACU appear to overshoot the CFU level before the two cell populations recover more or less together. The rapid rise in ACU between day 1 and day 2 might also be explained by the repair of potentially lethal damage but then some other explanation would be required for the loss of CFU during this period. Furthermore one would also expect CFU to exhibit potentially lethal damage but this would have to be masked by the loss of CFU. In contrast, to these results the initial effect of busulphan on ACU is less than on CFU, and the ACU level then decreases until it reaches the CFU level at about day 5 . These two responses suggest that the ACU population is not self-maintaining after drug treatment and relies on cells being fed i n from the CFU population. The results also suggest that the rate of differentiation of CFU into ACU is not constant, being temporarily increased after administration of vinblastine to continuously irradiated mice but not after administration of busulphan. After the first five days CFU and ACU maintain similar levels during recovery but the rate of stem cell recovery is quite different for the two drug treatments. In the vinblastine-treated irradiated mice recovery is rapid from a survival of about 0.0001, with a doubling time of about 17 hours between two and nine days. Recovery following busulphan is very much slower, even though the initial depletion
4 79
is very much less, and normal levels are not reached until after 40 days, which corresponds to a doubling time of about eight days. Although after busulphan the stem cells i n the spleen recover sooner than those in the bone marrow, the spleen does not make a major contribution to haemopoiesis since splenectomy does not alter the rate of recovery of red cell production as measured by 59Fe incorporation, nor does it affect the recovery of stem cells in the bone marrow. The slow recovery after busulphan is not due to a failure of the stem cells to increase their rate of proliferation because the results with repeated cytosine arabinoside show that these cells are proliferating rapidly. Neither is it solely due to the small depletion of the stem cells since, with other cytotoxic agents at doses that give a comparable initial stem cell depletion, recovery is much more rapid, for example following BCNU, CCNU and MeCCNU (Blackett et al., '76), cyclophosphamide, and also vinblastine without prior irradiation (unpublished). The stem cells are not incapable of rapid recovery after busulphan because further depletion to very low levels, by administration of vinblastine five days after busulphan, produces a very rapid recovery which crosses the busulphan-only curve at about 12 days. This rapid recovery is likely to be due to the greater stem cell depletion rather than a depletion of the maturing cells causing a feedback stimulation of the stem cells. This conclusion is supported by the measurements with methotrexate which when administered after busulphan, produces no further depression of the stem cells and no change in the recovery rate, although the maturing and peripheral blood cells are severely depleted. Recovery from lower stem cell levels could not be determined by using higher doses of busulphan because ofits short term toxicity, and so it is not possible to say whether greater depletion of stem cells by busulphan alone would have produced a more rapid recovery. This has been shown to occur with radiation (Guzman and Lajtha, '70). It is evident from our results that the rate of recovery of stem cells depends to some extent on the drug employed as well as on the degree of stem cell depletion, but
480
N . M. BLACKETT AND R. E. MILLARD
not apparently on the depletion of the maturing cells. The recovery of mature cell production, as shown by peripheral blood counts and red cell Fes9incorporation, does not depend on the stem cells reaching some specified level as has been suggested by Chervenick and Boggs (‘71), and there must be marked changes in the proliferation kinetics of the maturing cells to compensate for the reduction in stem cell numbers. Although no explanation has been found for the indolent recovery of stem cells following busulphan, it is not due to the inability of the stem cells to proliferate rapidly, nor is it due to a n irreversible derangement of the population size control, since further depletion of the stem cells with vinblastine brings about a rapid recovery. ACKNOWLEDGMENTS
The authors wish to thank Professor L. F. Lamerton for encouragement in carrying out the work described and for many useful discussions of the results. Mrs. S. F. OKell and Mrs. H. M. Belcher provided excellent technical assistance. LITERATURE CITED Blackett, N. M., a n d K. Adams 1972 Cell proliferation and the action of cytotoxic agents on haemopoietic tissue. Brit. J. Haematol., 23: 751758. Blackett, N . M., V . D. Courtenay and S . M. Mayer 1976 Differential sensitivitv to three nitrosoureas of colony forming cells for haemopoietic tissue, Lewis Lung and B16 melanoma tumours in C57BL mice. Cancer Chemotherapy Reports, 5 . Q . 929-933. _ ~ .. .
Blackett, N. M., and R. E. Millard
1973
Differ-
ential effect of Myleran on two normal haemopoietic progenitor cell populations. Nature (London), 2 4 4 : 300-301. Bruce, W . R., B. E. Meeker and F. A. Valeriote 1966 Comparison of the sensitivity of normal haemopoietic and transplanted lymphoma colony forming cells to chemotherapeutic agents administered ‘in vivo’. J. Nat. Cancer Inst., 37: 233-245. Chen, M. G., and J . C. Schooley 1970 Recovery of proliferative capacity of agar colony forming cells and spleen colony forming cells following ionising radiation and vinblastine. J. Cell. Physiol., 75: 89-95. Chervenick, P. A., and D. R. Boggs 1971 Patterns of proliferation and differentiation of hematopoietic stem cells after compartment depletion. Blood, 37: 5 6 8 5 8 0 . Eaves, A. C., and W. R. Bruce 1974 Altered sensitivity of haematopoietic stem cells to 5fluorouracil following endotoxin, cyclophosphamide and irradiation. Cancer Chemotherapy Reports, 5 8 : 813-820. Guzman, E., and L. G. Lajtha 1970 Some comparisons of the kinetic properties of femoral and splenic haemopoietic stem cells. Cell and Tissue Kinet., 3: 91-98. Josvasen, N . , and A. Boyum 1973 Haematopoiesis i n busulphan-treated mice. Scand. J. Haemat., 1 1 : 78-83. Metcalf, D. 1971 Acute antigen-induced elevation of serum colony stimulating factor (CSF) levels. Immunology, 21 : 427-436. Millard, R. E., N. M . Blackett and S. F. Okell 1973 A comparison of the effect of cytotoxic agents on agar colony forming cells, spleen colony forming cells and the erythrocytic repopulating ability of mouse bone marrow. J. Cell Physiol., 8 2 : 309-317. Reissmann, K. R., and K. B. Udupa 1972 Effect of erythropoietin on proliferation of erythropoietin-responsive cells. Cell and Tissue Kinet., 5 : 481-489. Udapa, K. B., H. Okamura and K . R. Reissmann 1972 Granulopoiesis during Myleran-induced summession of transdantable hematoDoietic stem cells. Blood, 39: 3i7-325.
