Conditions Controlling the Proliferation of Haemopoietic Stem Cells In Vitro T. M. DEXTER, T. D. ALLEN AND L. G . LAJTHA Paterson Laboratories, Christie Hospital and Holt Radium Institute, Manchester M20 9BZ England
ABSTRACT A liquid culture system is described whereby proliferation of haemopoietic stem cells (CFU-S), production of granulocyte precursor cells (CFU-C), and extensive granulopoiesis can be maintained in vitro for several months. Such cultures consist of adherent and non-adherent populations of cells. The adherent population contains phagocytic mononuclear cells, “epithelial” cells, and “giant fat” cells. The latter appear to be particularly important for stem cell maintenance and furthermore there is a strong tendency for maturing granulocytes to selectively cluster in and around areas of “giant fat” cell aggregations. By “feeding” the cultures a t weekly intervals, between 10 to 15 “population doublings” of functionally normal CFU-S regularly occurs. Increased “population doublings” may be obtained by feeding twice weekly. The cultures show initially extensive granulopoiesis followed, in a majority of cases, by an accumulation of blast cells. Eventually both blast cells and granulocytes decline and the cultures contain predominantly phagocytic mononuclear cells. Culturing a t 33°C leads to the development of a more profuse growth of adherent cells and these cultures show better maintenance of stem cells and increased cell density. When tested for colony stimulating activity (CSA) the cultures were uniformly negative. Addition of exogenous CSA caused a rapid decline in stem cells, reduced granulopoiesis and an accumulation of phagocytic mononuclear cells. A variety of in vitro clonal assay systems are now available which may permit the identification of some of those factors which control proliferation and differentiation of haemopoietic precursor cells. Bradley and Metcalf (‘66) and Pluznik and Sachs (‘66) described the formation of colonies in vitro by granulocyte precursor cells (CFU-C) and showed that the formation of these colonies is dependent upon the presence of a suitable colony stimulating activity (CSA) which can be isolated from a variety of sources (for review see Metcalf and Moore, ’71). More recently, methods have also been described for the clonal growth of erythroid precursor cells (Stephenson e t al., ’71), lymphoid cells (Metcalf e t al., ’75a; Fibach e t al., ’76; Sredni e t al., ’76) and megakaryocytes (Metcalf e t al., ’75b). However, unique chromosome markers have indicated (Wu, ’68; Edward, ’70) that these precursor cells are derived from a comJ. CELL.PHYSIOL., 91: 335-344.
mon pluripotent stem cell-the CFU-S (described by Till and McCulloch, ’61). The factors and intercellular relationships controlling the proliferation of the stem cells and their differentiation into the various committed precursor cells, are of fundamental importance in understanding the mechanisms controlling haemopoiesis. All the in vitro culture methods for haemopoietic tissues hitherto used yielded essentially short term systems which were running down, in which physiological control processes and culture decline phenomena (associated with the decline of the culture) are difficult to separate. Herein we describe some properties of a system, in which in vitro proliferation of pluripotent stem cells and production of granulocytic precursor cells can be maintained for several months, for > 15 population doublings, thus offering a Received June 25,’76. Accepted Oct. 27, ‘76.
335
336
T M DEXTER, T D ALLEN AND I,
C,
I.A.JTHA
Fig. 1 Scanning electron micrograph of t h e attached mononuclear phagocytic cells. T h e cell a r e extended along a single axis, with numerous cytoplasmic projections. Scale bar = 10 urn.
significantly longer period of behavior close to the “in vivo” situation for the study of haemopoietic control mechanisms. MATERIALS AND METHODS
Culture system To establish the culture lo7 pooled femoral bone marrow cells from 6 to 8-week-old virgin BDF, female mice are inoculated into each of five flat bottomed, 4-oz screw-capped glass culture bottles (United Glass, London) in 10 ml of Fischer’s medium supplemented with 20-25%horse serum (Flow Laboratories) 500 unitdm1 penicillin and 50 pg/ml streptomycin sulphate. The cultures are gassed with a mixture of air + 5% CO, and incubated a t 37°C. At 3- to 7-day intervals the cultures are “fed” by removal of half the growth medium (5 ml) and addition of an equal volume of fresh growth medium. Over a 3-week period the numbers of non-adherent cells decline (as do also CFU, and CFUJ and a layer of adherent cells (fibroblastoid, epitheliod and phagocytic mononuclear cells) becomes established. After three weeks the cultures are “recharged”, i.e., “fed” as before, with 5 ml of growth medium containing a further lo7
freshly isolated syngeneic femoral bone marrow cells (i.e,, containing approximately 3,000 CFU-S and 15,000 CFU-C) from comparably aged mice. In such cultures proliferation of suspension (non-adherent) cells and production of stem cells for several months has been observed (Dexter and Lajtha, ’76).
CFU-S and CFU-C assays Generally five culture bottles are used in each experiment. The pooled suspension cells present in the growth medium removed a t weekly or two times weekly intervals are counted directly in an haemacytometer. The growth medium is then centrifuged (800g for 10 minutes) and the cell pellet is resuspended in a small volumn of Fischer’s medium. The cells can be smeared directly onto slides for staining and differential morphological analysis, or assayed for CFU-S and CFU-C, according to the techniques of Till and McCulloch (’61) and Bradley and Metcalf (’66) a s modified by Bradley and Sumner (’68). Scanning electron microscopy Ten milliliters of 6%glutaraldehyde (pH 7.4 in MI15 Sorensens phosphate buffer) was
337
HAEMOPOIETIC STEM CELLS I N VITRO
Fig. 2 Scanning electron micrograph of a single attached “epithelial” cell i E ) showlng flattened morphology. Scale bar = 20 I*.. TABLE 1
Maintenance of CFU-S and CFU-C in cultures Weeks cultured
1 2 3 4 5 6
7 8 9 10 11 12 13 14
Cell count (suspension x 105)
21 13.0 16.0 7.8 8.2 6.4 9.2 5.3 7.0 11.0 58 9.0 8.9 7.6
Morphology
X
CFU-Slculture isuspension)
CFU-Clculture (suspension)
B
Gr
Mono
558 -t 72 500 -+ 40 770 f 96
8,956 -t 766 13,000 % 1,200 ND 3,420 -t 200 2,000 f 180 ND 4,000 f 280 ND ND 2,870 f 190 ND ND < 20 0
10 5 16 31 18 15 23 24 ND 13 3 2 5 3
68 79 70
18 16 14 29 14 47 59 63 ND 76 84 87 93 92
ND 200 f 18 378 -+ 80 300 f 40 ND 371 -t 28 ND 130 -t 45 56 It 12 0
0
40
68 38 18 10 ND 9 12 10 2 5
ND, no data, B, blasts, Gr. granulocytes (all stages), Mono, phagocytic mononuclear cells
gently pipetted into the culture bottles. This was gently mixed with the 10 mls of growth medium, resulting in a final concentration of 3% After ten minutes the mixture of growth medium and fixative was gently decanted and replaced with fresh 3%glutaraldehyde. After a further 30 minutes fixation the monolayer was washed briefly (5 minutes) in three
changes of buffer alone and then post fixed in 1%osmium tetroxide for one hour (in M/15 Sorensens phosphate buffer pH 7.4). After several buffer washes the monolayer was rinsed with distilled H,O; the bottles were broken into small fragments at all times keeping the cells wet to prevent air drying. The bottle fragments (approximately 1 cm x
338
1’. M. DEXTER. T. D. ALLEN AND L. G . LAJTHA
Fig. 3 Scanning electron micrograph o f a “giant fat” cell in early stages of lipid accumulation. The vermiform ridges on t h e surtace iarrowed) a r e t h e mitochondria in the very thin layer of cytoplasm which surrounds t h e lipid containing vacuole. Scale har = 10 p .
1 cm) were then dehydrated through ascending concentrations of acetone and critical point dried from liquid CO,. The cells were coated with a 200 A thickness of gold deposited by sputtering, and examined in a Cambridge ,534-10 Scanning Electron Microscope a t an accelerating voltage of 15 or 20 Kv. RESULTS
The adherent layer The crucial feature of the culture system is the establishment of the adherent layer of cells, without which there is a rapid decline (1-2 weeks) in the numbers of cells and disappearance of CFU-S and CFU-C within two to three weeks. Also, when there is poor development of the adherent layer, or if the bottles are siliconised, (thus preventing the attachment of potentially adherent cells) cell proliferation is poor and short lived. The adherent layer contains three morphological cell types; phagocytic mononuclear cells with numerous cytoplasmic extensions and an overall tendency to spread along a single axis (fig. 11, flat-
tened cells which tend to form a confluent monolayer (fig. 21, and a third cell type, which develops over a 3-week period and forms clusters and is characterised by an enormous accumulation of a lipid type of material which stains intensely with Sudan Black B and is also extremely osmiophilic (fig. 3). The accumulation of lipid within these cells causes an increase in size up to a range of 50-150 p m and these “giant fat” cells themselves form radiating foci (fig. 4) clearly visible in the bottles with the naked eye. There is a marked tendency for maturing granulocytes to cluster in and around the aggregations (fig. 4). In t h e absence of these “giant fat” cells there is no CFU-S maintenance.
CuEture at 37’C-weekly feeding The data from a “typical” experiment are shown in table 1.Cell proliferation is obviously occurring since, after the second week, the numbers of cells in suspension approximately double between each feeding. As the numbers of adherent cells also remain fairly constant throughout the experiment, (as judged by re-
339
HAEMOPOIETlC STEM CELLS IN VITRO
Fig. 4 Scanning electron micrograph of an area of characteristic aggregation of “giant fat” cells (the larger spherical cells) and granulocytes (small spherical cells) clustered around the periphery. The mononuclear phagocytic and “epithelial” cells forming the adherent layer are also visible. Scale bar = 100 pm.
moval of adherent cells with a rubber policeman and counting directly on a haemacytometer) the increase in the numbers of suspended cells represents a true proliferation rather than simply detachment of the adherent population. Similarly the pluripotent stem cells (CFUS ) are proliferating a t least for nine weeks, also showing an approximate weekly doubling. I t is noteworthy that CFU-S and CFU-C are not detectable in the adherent cell population. This was established by removing all the growth medium from cultures which had been maintained for 2 , 6 or 10 weeks. The adherent cells were than washed once in Fischer’s medium removed mechanically with a rubber policeman and assayed for CFU-S and CFU-C content. In three separate experiments they were uniformly negative. Hence to monitor changes occurring in their number it is sufficient to assay the non-adherent cell population which can be suspended by gentle agitation of the bottles. These CFU-S are apparently normal, forming erythroid, gran-
TABLE 2
Experimental uariation in extents of stem cell maintenance in uitro Number of experiments
6 10 8 3
CFU-S maintenance (weeks)
CFU-C maintenance (weeks1
3-4
4-5 7-8
6-7
9-12 > 13
10-13 > 14
Maintenance signifies the time (in weeks of culture1 at which CFUS and CFU-C were no longer detectable.
ulocytic and megakaryocytic colonies in the spleens of potentially lethally irradiated mice, and if sufficient numbers of stem cells are injected (approximately 40 CFU-S), the cultured cells can protect lethally irradiated mice from haemopoietic death. The “committed” granulocytic precursor cells (CFU-C) are also being maintained for several months (table 1).These are presumably being produced by differentiation from the pluripotent stem cell-although a limited
340
T. M. DEXTER, T.
L).
ALLEN AND L. G. LAJTHA
capacity for their self-duplication is evident from their sensitivity to the 3H-thymidine “suicide” (Lajtha, et al., ’69). The colonies formed by these CFU-C i n agar cultures are similar to those produced from freshly isolated bone marrow cells, showing maturation towards granulocytes and macrophages, depending upon t h e addition of appropriate colony stimulating activity (CSA) -in this case, heart conditioned medium (Dexter and Testa ’76). In the absence of CSA, they did not form colonies in agar. In all cultures, extensive granulopoiesis is seen initially with all maturation stages represented. With time in culture, however, granulocytes TABLE 3
Egect of twice weekly feeding on stem cell maintenance (1experiment)
Variations in stern cell maintenance We have previously found that stem cell proliferation in the liquid cultures is dependent to a great extent upon the batch of serum used (Dexter and Testa, ’76) with horse serum (Flow Laboratories) providing most consistent growth. Other sera, e.g., horse (Gibco), foetal calf (Flow and Gibco), baby calf (Gibco) have been found to be unsuitable for maintenance of stem cell proliferation in vitro. How-
~
Days cultured
Number of feedings
7 18 34 49 62 83 109
2 5 10 14 18 24 30
Cell count ( X 10’)
8.0 6.0 1.2 1.2 2.1 2.1 5.1
decline, macrophages increase and in many cultures there is also an accumulation of “blast” cells (mononuclear cells of approximately 8-10y diameter with immature nuclei and one or two nucleoli). The nature of these blast cells is unknown but surface marker studies have shown t h a t they possess neither 0 antigen characteristic of T lymphocytes (Raff and Wortis, ’70) nor surface lg determinants characteristic of B-lymphocytes (Raff et al., ’70). Furthermore they show no myeloperoxidase or S u d a n Black a c t i v i t y (characteristic of maturing granulocytes). With longer time (> 9 weeks) in culture the numbers of such blast cells decrease and phagocytic mononuclear cells become the predominant cell type. Proliferation of such mononuclear cells has been observed in some cultures for periods in excess of 30 weeks.
CFU-Slculture (suspension)
166 2 20 11o-t 12 153 -t 27 28 -t 5 8 56 -t 10 0
(Once weekly feeding “control” culture shown In table 1) I Equivalent to population doublings.
TABLE 4
Effect of temperature on stem cell maintenance Group
37OC 33°C 37oc 33°C 37oc 33°C 37T 33OC 37% 33oc 37°C 33°C 37oc 33°C 37°C 33-2 37°C 33T
Weeks cultured
1 1 3 3 4 4 6 6 7 7 9 9 10 10 11 11 12 12
Cell count ( X 105)
13.0(5-24) 62.0(25-130) 5.0 (2-11) 52.006-98) 5.0 (3-9) 60.0(25-140) 7.5(6-9) 41.5 (40-431 4.5(3-7) 28.0(20-32) 6.5(4-9) 22.0(16-30) 8.5 (5-12) 15.0(11-18) 8.5 (3-16) 22.006-30] 5.0(5) 15.0(9-22)
ND, no data. Results are the mean (and range) of four or five experiments.
CFU-SI culture
290 (97-558) 780 (320-1254) 230 (50-695) 832 (295-1253) 209 (30-384) 811 (170-1944) 270 (60-392) 590 (290-8921 96 (21-71) 197 (154-220) 102 (46-219) 225 (100-356) ND ND 75 (0-250) 125 (35-220) 12 (0-140) 45 (0-180)
CFU-C/culture (1 exwriment)
9,000 21,000 8,000 12,000 ND ND 2,000 3,000 ND ND ND ND 52 1,130 ND ND 0 162
34 1
HAEMOPOIETIC STEM CELLS IN VITRO
ever, even using the same batch of serum there is a variation in the extent of stem cell maintenance. Table 2 shows that approximately 25% of the experiments show only limited stem cell maintenance with a gradual decline in stem cell numbers and rapid conversion to phagocytic mononuclear cells. In about 60%of the experiments, however, stem cell proliferation can been seen for periods between 6 and 12 weeks; an absolute increase in numbers of CFU-S and CFU-C occurs between each feeding. In some experiments there is maintenance of stem cells for up to 16 weeks. In all these long term cultures blast cell accumulation is regularly seen. The observation that CFU-C proliferation persists somewhat longer than that of CFU-S indicates a small but definite self-renewal capability of CFU-C.
Culture at 37°C-twice weekly feeding Feeding a t weekly intervals may not be imposing a great proliferative stress upon the cells, indeed a "plateau" value after one doubling may be reached well before seven days in culture. This "plateau value" if maintained for several days may, in fact, represent suboptimal culture conditions, thus, limiting the overall proliferative capacity of the stem cells. Feeding a t more frequent intervals was therefore used to determine whether the number of "population doublings" of the stem cell population could be increased (all other factors being constant). The data shown in table 3 with two times weekly feeding indicate that higher numbers of "doublings" can be achieved by altering the feeding regime. ?he effect of temperature on stem cell maintenance
In preliminary experiments investigating the effect of temperature on leukaemic transformation in vitro it was noticed that in cultures kept at 33"C, the initial three weeks "setting up period" produced significantly more adherent cells (including giant fat cells) than those kept a t 37°C. When such adherent layers were "recharged" with bone marrow cells and the culture was continued a t 33°C with weekly feeding-there was a striking effect upon the numbers of cells produced in suspension (table 4). For the first five to six weeks almost ten times more cells were present a t 33°C compared with 37"C, still doubling a t weekly intervals. With regard to stem cell maintenance, CFU-S at 33°C are consist-
TABLE 5
Morphology of cells obtained from culture maintained at 37°C or 33°C Morphology Group
37°C 33°C 37T 33°C 37°C 33°C 37T 33°C 37°C 33OC
Weeks cultured
Blasts 9:
Granulocytes %
1 1 3 3 5 5 8 8 10 10
10 7 10 6 31 5 16 10 0 9
78 83 83 80 53 71 6 19 3 30
Mononuclear cells %
12 10
7 14 16 22 78 70 96 61
TABLE 6
Temperature effecton ability ofthe adherent layer to support stem cell proliferation Group
37°C 33°C 330-370c 37°C 33°C 33"-37"C 37% 33oc 33"-37"C 37°C 33oc 33"-37"C
Weeks cultured
Cell count ( x lo5)
CFU-S/ culture
1 1 1 3 3 3
11.6 29.0 59.0
380 700 1,400 240 2,800 1,550 180 700 800
4
4 4 6 6 6
4.0
74.0 40.0 2.0 56.0 2.0 5.3 58.0 4.4
148 986 120
37°C-refers to cultures maintained wholly at 37°C 33'C-refers to cultures maintained wholly at 33OC 33"-37"C-refers to cultures where the adherent population was initiated at 33°C and the cultures transferred to 37OC subsequent with the inoculation of the second bone marrow population
ently maintained a t levels two to four times that seen a t 37°C-at least for 11 weekswhilst the effect on the CFU-C is not so marked. In terms of the morphology of the cells produced-the two cultures show a broadly similar appearance, with granulopoiesis for five to seven weeks being followed by an increase in phagocytic mononuclear cells (table 5). The percentage of blast cells a t 33°C however seldom exceeds 10% for the duration of the cultures. Of particular interest is our preliminary finding that the variation seen in extent of stem cell maintenance a t 37°C (using the same batch of serum) does not occur a t 33°C:
342
T. M. DEXTER, T. D. ALLEN AND L. G. LAJTHA TABLE 7
Effect of feeding cultures with medium containing CSA (colony stimulating activity) Morphology Group
~~~
Weeks cultured
Cell count x 105
CFU-S
1 1 1
7
3 3 3
4
412 44 8 150 5 0
Blasts X
Granulocytes %
Mononuclear cells X
20 5 1 12 2 0
66 48 13 58 20 0
15 40 86 30 I8 100
~
Control MECM MHCM Control MECM MHCM
5 4 3 2
MECM, mouse embryo conditioned medium; MHCM, mouse heart conditioned medium. Data from one experiment are shown. In two further experiments similar results. were obtained.
all cultures so far maintained a t 33°C have shown extensive granulopoiesis and stem cell maintenance for a t least six weeks. Since this may have been due to a more uniformly “effective” adherent layer being produced during the initial three weeks setting-up period we have investigated the capacity of such adherent layers produced a t 33°C but subsequently transferred to 37”C, to maintain the growth of stem cells. Cultures were set up as before a t 33°C (with weekly feeding) but after the initial three weeks, they were transferred to 37”C, fed, re-charged with the second bone marrow population and maintained at 37°C and fed in the usual manner. The results are shown in table 6 and indicate the improved capacity of the adherent layer established a t 33°C to support the subsequent production of granulocytes and stem cells a t 37°C for at least four weeks, however, by six weeks this capacity of the adherent layer appears to be “exhausted” and the situation reverts to the “control” 37°C level.
The role of ‘CSA” in the cultures In view of the extensive granulopoiesis seen for up to two months in some cultures and a view of the proposed role of colony stimulating activity (CSA) in inducing granulocyte development in agar cultures in vitro (Metcalf ’70) and the in vivo evidence supporting the role of CSA as a physiological regulator of granulopoiesis (Moore et al., ’74) the culture media were tested for the presence (?endogenously produced) of CSA. The growth medium (from feeder layers alone or from cultures where the second “target” population of bone marrow cells had been added) removed after various culture periods a t 37”C, was centrifuged and the “cell
free” supernate was tested (in the single layer agar system) for its ability to stimulate the development of CFU-Cfrom either freshly isolated bone marrow cells or from the cultured marrow cells. Various concentrations of such “conditioned” media obtained from our bone marrow cultures were tried 6 7 5 % ) but the results were uniformly negative (i.e., no cluster or colony formation was induced). Even after heat treatment (56°C for 30 minutes) and/or extensive dialysis (thought to remove molecules which can inhibit CSA (Bradley and Sumner, ’68)no colony stimulating activity was detectable. Furthermore, neither the suspended cells nor the adherent layer cells would promote the growth of bone marrow CFU-C when used as a “feeder layer” in the double layer agar system. Since our culture system appeared to deficient in “CSA’, this raised the possibility of further improvement by providing exogenous CSA for the cultures. Consequently, together with the re-charging of the cultures (after the initial three weeks setting up period) the medium was supplemented with mouse embryo or heart conditioned medium (MECM or MHCM) as a source of CSA (Dexter and Testa, ’76). The final concentration of MECM or MHCM in the cultures was 15%-in the single layer agar system this concentration is known to induce optimal, i.e., plateau, numbers of CFU-C using either freshly isolated bone marrow or cultured bone marrow as target cells. The cultures were continued to be fed a t weekly intervals with growth medium supplemented with MECM or MHCM, the cells collected and assayed as before. The results are shown in table 7. Feeding with CSA containing “conditioned” medium caused a marked decrease of the stem cells
HAEMOPOIETIC STEM CELLS IN VITRO
after one week and their virtual disappearance after three weeks in culture. Concomitant with the decline in CFU-S, CFU-C were also no longer detectable after four weeks in culture. Also there was a marked effect upon the morphological appearance of the cells: the CSA supplemented cultures underwent an early transition to mononuclear cells. The addition of MECM or MHCM produced no apparent increase in cell numbers in the cultures, nor did it cause a change in the cell types present to mature forms when added to cultures with accumulated blast cells-even after several days of cultures with high CSA activity in the medium. DISCUSSION
Cultures methods have been described before in which stem cell proliferation could be maintained for a matter of days (Sumner et al., ’72; Cline and Gold, ’76) or even a few weeks (Dexter and Lajtha, ’74).The introduction of an adherent (? nurse) layer from the initial “setting-up’’ bone marrow inoculum (Dexter and Lajtha, ’76; Dexter and Testa, ’76) presents a significant improvement, by maintaining granulopoiesis and stem cell proliferation for two to three months. Whilst this time scale indicates that the culture conditions (and the behavior of cells) may approach physiological states, a t least during the few weeks of culture, the method as used to date is clearly not yet optimal. It has been indicated that two times weekly feeding improves the rate of stem cell production and maintenance, further improvement may be achieved with more frequent medium change. For example, 48 hours may be sufficient to allow stem cell population doubling between medium changes, and to avoid a “plateau” state in the cultures which may result in suboptimal conditions. The accumulation of blast cells, after about five weeks (? maturation block) may be an indication of some suboptimal state, although it should be emphasized that even with such “maturation arrest” the cultures produce normal CFU-S and CFU-C for several weeks, i.e., cells which under appropriate conditions (as splenic or as agar colony formers) do produce normal mature granulocytes. The observation that the cultures do very much better at a lowered incubation temperature incorporating a better maintenance of
343
stem cells, a more effective adherent layer; higher plateau values of cells, and a delayed accumulation of blast cells is relevant in this respect. Under suboptimal conditions the cells may gradually build up increasingly unbalanced states-either in intracellular or in intercellular respects (e.g., numerical imbalance between interacting cell populations). The inevitable slowing down of some metabolic events a t 33”C, compared with 37°C may go some way to prevent or delay the development of such “imbalances.” It is interesting that it can do this without apparently significantly affecting the measurable cell proliferation rates. That the cells can achieve and maintain higher cell concentrations a t the lower temperature is certainly one indication of a generally improved milieu. The absence of detectable amounts of colony stimulating activity (CSA) in the harvested culture media is somewhat surprising. CSA is obligatory in semisolid cultures (agar or methylcellulose) not only for granulocytic colony growth (Metcalf, ’701, but even for temporary maintenance of CFU, (Testa and Lajtha, ’73). It was conceivable that our cultures produced a very low concentration of endogenous CSA and that the blast cell accumulation was, in fact, due to suboptimal CSA in the medium. As has been described in the results, however, addition of exogenous CSA did not help: not only did i t not “revert” the blast cells to mature forms, but it caused a speedy rundown of the cultures. It is possible that this was caused by some impurities in the CSA which would be absorbed by the agar or methylcellulose in the semisolid cultures, but it must be remembered that the semisolid cultures represent a system which is running down within a matter of days while active granulopoiesis persists in our cultures-with undetectable amounts of CSA in the medium-for many weeks. It is also possible that molecules with CSA inhibitory activity are present in these cultures. However if this be the case it would still be difficult to explain why active granulopoiesis is occurring in the presence of inhibitory concentrations sufficient to mask CSA. While our culture system of bone marrow cells settled on a special adherent layer of cells is not a complete representation of bone marrow functions-so far we have been unable to demonstrate, e.g., erythropoiesis or thrombopoiesis-by its long maintenance and extent of stem cell prolifera-
344
T M DEXTER, T. D. ALLEN AND 1, G LAJTHA
tion (over 10-15 population doublings) i t offers a useful system for the study of factors involved in the control of proliferation and differentiation of some of the earliest haemopoietic cells. ACKNOWLEDGMENTS
This work was supported by t h e Medical Research Council and the Cancer Research Campaign. The authors wish to thank Nicola P. J. Higgins and G. R. Bennion for excellent technical assistance. LITERATURE CITED Bradley, T. R., and D. Metcalf 1966 The growth of mouse bone marrow cells in uitro. Austr. J. Exp. Biol. Med. Sci., 44: 287-300. Bradley, T. R., and M. A. Sumner 1968 Stimulation of mouse hone marrow growth in vitro by conditioned medium. Austr. J. Exp. Biol. Med. Sci., 46: 607-618. Cline, M. J.,and D. W. Golde 1976 Improved techniques for liquid culture of human and mouse bone marrow. Blood, 47: 369-379. Dexter, T. M., T. D. Allen, L. G. Lajtha, R. Schofield and B. I. Lord 1973 Stimulation of differentiation and proliferation of haemopoietic cells in uitro. J. Cell Physiol., 82: 461-474. Dexter, T. M., and L. G. Lajtha 1974 Proliferation of haemopoietic stem cells in uitro. Brit. J. Haematol., 28: 525-530. - 1976 Proliferation of haemopoietic stem cells and development of potentially leukaemic cells in uitro. In: Proceedings of VIII International Symposium on Comparative Research on Leukaemia and Related Diseases. Copenhagen, 1975. Publishers: Karger, Basel. Dexter, T. M., and N. G. Testa 1976 Differentiation and proliferation of haemopoietic cells in culture. In: Methods in Cell Biology. Vol. 14. D. M. Prescott, ed. Acadmic Press, New York, pp. 387-405. Edwards, G. E., R. G Miller and R. A. Phillips 1970 Differentiation of rosette-forming cells from myeloid stem cells. J. of Immunology, 105: 719-729. Fibach, E., E. Gerassi and L. Sachs 1976 Induction of colony formation in uitro by human lymphocytes. Nature, 259: 127-128. Lajtha, L. G., L. V. Pozzi, R. Schofield and M. Fox 1969 Kinetic properties of haemopoietic stem cells. Cell & Tissue Kinet., 2: 39-49.
Metcalf, D. 1970 Studies on colony formation in uitro by mouse bone marrow cells. 11. Action of colony stimulating factor. J. Cell Physiol., 76: 89-100. Metcalf, D., and H. R. MacDonald 1975 Heterogeneity of in uitro colony- and cluster-forming cells in the mouse marrow: Segregation by velocity sedimentation. J. Cell Physiol., 85: 643-654. Metcalf, D., and M. A. S. Moore 1971 Haemopoietic Cells. North Holland Publishing Co., Amsterdam. Metcalf, D., H. R. MacDonald, N. Odartchenko and L. B. Sordat 1975a Growth of mouse megakaryocyte colonies in uitro. Proc Nat. Acad. Sci., 72: 1744-1748. Metcalf, D., N. L. Warner, G. J. V. Nossal, J. F. A. P. Miller, K. Shortman and E. Rabellino 19751, Growth of Blymphocyte colonies in uitro from mouse lymphoid organs. Nature, 255: 630-632. Raff, M. C. 1969 Theta isoantigen as a marker of thymus-derived lymphocytes in mice. Nature, 224: 378-379. Raff, M. C., M. Sternberg and R. B. Taylor 1970 Immunoglobulin determinants on the surface of mouse lymphoid cells. Nature, 225: 553-534. Moore, M. A. S., G. Spitzer, D. Metcalf and D. G. Pennington 1974 Monocyte production of colony stimulating factor in familial neutropenia. Brit. J. Haematol., 27: 47-55. Pluznik, P. H., and L. Sachs 1966 The induction of colonies of normal “mast” cells by a substance in conditioned medium. Exp. Cell Res., 43: 553-563. Sredni, B., Y. Kalechman, H. Michlin and L. A. Rozenzajn 1976 Development of colonies in uitro of mitogen stimulated mouse T-lymphocytes. Nature, 259: 130-132. Stephenson, J. R., A. A. Axelrad, D. L. McLeod and M. M. Shreeve 1971 Induction of colonies of haemoglobin synthesizing cells by erythropoietin in uitro. Proc. Nat. Acad. Sci., 68: 1542-1546. Sumner, M. A., T. R. Bradley, G. S. Hodgson, M. J. Cline, A. Fry and L. Sutherland 1972 The growth of bone marrow cells in liquid culture. Brit. J. Haematol., 23: 221-234. Testa, N. G., and L. G. Lajtha 1973 Comparison of the kinetics of colony forming units in spleen (CFU,) and culture (CFU,). Brit. J . Haematol., 24: 367-376. Till, J. E., and E. A. McCulloch 1961 A direct measurement of the radiation sensitivity of normal mouse bone marrow cells. Rad. Res., 14: 213-222. Williams, W., and G. J. van den Engh 1975 Separation of subpopulations of in uitro colony forming cells from mouse marrow by equilibrium density centrifugation. J. Cell Physiol., 86: 237-246. Wu, A. M., J. E. Till, L. Siminovitch and E. A. McCulloch 1968 Cytological evidence for a relationship between normal haematopoietic colony forming cells and cells of the lymphoid system. J. Exptl. Med., 127: 455-463.
ABSTRACT A liquid culture system is described whereby proliferation of haemopoietic stem cells (CFU-S), production of granulocyte precursor cells (CFU-C), and extensive granulopoiesis can be maintained in vitro for several months. Such cultures consist of adherent and non-adherent populations of cells. The adherent population contains phagocytic mononuclear cells, “epithelial” cells, and “giant fat” cells. The latter appear to be particularly important for stem cell maintenance and furthermore there is a strong tendency for maturing granulocytes to selectively cluster in and around areas of “giant fat” cell aggregations. By “feeding” the cultures a t weekly intervals, between 10 to 15 “population doublings” of functionally normal CFU-S regularly occurs. Increased “population doublings” may be obtained by feeding twice weekly. The cultures show initially extensive granulopoiesis followed, in a majority of cases, by an accumulation of blast cells. Eventually both blast cells and granulocytes decline and the cultures contain predominantly phagocytic mononuclear cells. Culturing a t 33°C leads to the development of a more profuse growth of adherent cells and these cultures show better maintenance of stem cells and increased cell density. When tested for colony stimulating activity (CSA) the cultures were uniformly negative. Addition of exogenous CSA caused a rapid decline in stem cells, reduced granulopoiesis and an accumulation of phagocytic mononuclear cells. A variety of in vitro clonal assay systems are now available which may permit the identification of some of those factors which control proliferation and differentiation of haemopoietic precursor cells. Bradley and Metcalf (‘66) and Pluznik and Sachs (‘66) described the formation of colonies in vitro by granulocyte precursor cells (CFU-C) and showed that the formation of these colonies is dependent upon the presence of a suitable colony stimulating activity (CSA) which can be isolated from a variety of sources (for review see Metcalf and Moore, ’71). More recently, methods have also been described for the clonal growth of erythroid precursor cells (Stephenson e t al., ’71), lymphoid cells (Metcalf e t al., ’75a; Fibach e t al., ’76; Sredni e t al., ’76) and megakaryocytes (Metcalf e t al., ’75b). However, unique chromosome markers have indicated (Wu, ’68; Edward, ’70) that these precursor cells are derived from a comJ. CELL.PHYSIOL., 91: 335-344.
mon pluripotent stem cell-the CFU-S (described by Till and McCulloch, ’61). The factors and intercellular relationships controlling the proliferation of the stem cells and their differentiation into the various committed precursor cells, are of fundamental importance in understanding the mechanisms controlling haemopoiesis. All the in vitro culture methods for haemopoietic tissues hitherto used yielded essentially short term systems which were running down, in which physiological control processes and culture decline phenomena (associated with the decline of the culture) are difficult to separate. Herein we describe some properties of a system, in which in vitro proliferation of pluripotent stem cells and production of granulocytic precursor cells can be maintained for several months, for > 15 population doublings, thus offering a Received June 25,’76. Accepted Oct. 27, ‘76.
335
336
T M DEXTER, T D ALLEN AND I,
C,
I.A.JTHA
Fig. 1 Scanning electron micrograph of t h e attached mononuclear phagocytic cells. T h e cell a r e extended along a single axis, with numerous cytoplasmic projections. Scale bar = 10 urn.
significantly longer period of behavior close to the “in vivo” situation for the study of haemopoietic control mechanisms. MATERIALS AND METHODS
Culture system To establish the culture lo7 pooled femoral bone marrow cells from 6 to 8-week-old virgin BDF, female mice are inoculated into each of five flat bottomed, 4-oz screw-capped glass culture bottles (United Glass, London) in 10 ml of Fischer’s medium supplemented with 20-25%horse serum (Flow Laboratories) 500 unitdm1 penicillin and 50 pg/ml streptomycin sulphate. The cultures are gassed with a mixture of air + 5% CO, and incubated a t 37°C. At 3- to 7-day intervals the cultures are “fed” by removal of half the growth medium (5 ml) and addition of an equal volume of fresh growth medium. Over a 3-week period the numbers of non-adherent cells decline (as do also CFU, and CFUJ and a layer of adherent cells (fibroblastoid, epitheliod and phagocytic mononuclear cells) becomes established. After three weeks the cultures are “recharged”, i.e., “fed” as before, with 5 ml of growth medium containing a further lo7
freshly isolated syngeneic femoral bone marrow cells (i.e,, containing approximately 3,000 CFU-S and 15,000 CFU-C) from comparably aged mice. In such cultures proliferation of suspension (non-adherent) cells and production of stem cells for several months has been observed (Dexter and Lajtha, ’76).
CFU-S and CFU-C assays Generally five culture bottles are used in each experiment. The pooled suspension cells present in the growth medium removed a t weekly or two times weekly intervals are counted directly in an haemacytometer. The growth medium is then centrifuged (800g for 10 minutes) and the cell pellet is resuspended in a small volumn of Fischer’s medium. The cells can be smeared directly onto slides for staining and differential morphological analysis, or assayed for CFU-S and CFU-C, according to the techniques of Till and McCulloch (’61) and Bradley and Metcalf (’66) a s modified by Bradley and Sumner (’68). Scanning electron microscopy Ten milliliters of 6%glutaraldehyde (pH 7.4 in MI15 Sorensens phosphate buffer) was
337
HAEMOPOIETIC STEM CELLS I N VITRO
Fig. 2 Scanning electron micrograph of a single attached “epithelial” cell i E ) showlng flattened morphology. Scale bar = 20 I*.. TABLE 1
Maintenance of CFU-S and CFU-C in cultures Weeks cultured
1 2 3 4 5 6
7 8 9 10 11 12 13 14
Cell count (suspension x 105)
21 13.0 16.0 7.8 8.2 6.4 9.2 5.3 7.0 11.0 58 9.0 8.9 7.6
Morphology
X
CFU-Slculture isuspension)
CFU-Clculture (suspension)
B
Gr
Mono
558 -t 72 500 -+ 40 770 f 96
8,956 -t 766 13,000 % 1,200 ND 3,420 -t 200 2,000 f 180 ND 4,000 f 280 ND ND 2,870 f 190 ND ND < 20 0
10 5 16 31 18 15 23 24 ND 13 3 2 5 3
68 79 70
18 16 14 29 14 47 59 63 ND 76 84 87 93 92
ND 200 f 18 378 -+ 80 300 f 40 ND 371 -t 28 ND 130 -t 45 56 It 12 0
0
40
68 38 18 10 ND 9 12 10 2 5
ND, no data, B, blasts, Gr. granulocytes (all stages), Mono, phagocytic mononuclear cells
gently pipetted into the culture bottles. This was gently mixed with the 10 mls of growth medium, resulting in a final concentration of 3% After ten minutes the mixture of growth medium and fixative was gently decanted and replaced with fresh 3%glutaraldehyde. After a further 30 minutes fixation the monolayer was washed briefly (5 minutes) in three
changes of buffer alone and then post fixed in 1%osmium tetroxide for one hour (in M/15 Sorensens phosphate buffer pH 7.4). After several buffer washes the monolayer was rinsed with distilled H,O; the bottles were broken into small fragments at all times keeping the cells wet to prevent air drying. The bottle fragments (approximately 1 cm x
338
1’. M. DEXTER. T. D. ALLEN AND L. G . LAJTHA
Fig. 3 Scanning electron micrograph o f a “giant fat” cell in early stages of lipid accumulation. The vermiform ridges on t h e surtace iarrowed) a r e t h e mitochondria in the very thin layer of cytoplasm which surrounds t h e lipid containing vacuole. Scale har = 10 p .
1 cm) were then dehydrated through ascending concentrations of acetone and critical point dried from liquid CO,. The cells were coated with a 200 A thickness of gold deposited by sputtering, and examined in a Cambridge ,534-10 Scanning Electron Microscope a t an accelerating voltage of 15 or 20 Kv. RESULTS
The adherent layer The crucial feature of the culture system is the establishment of the adherent layer of cells, without which there is a rapid decline (1-2 weeks) in the numbers of cells and disappearance of CFU-S and CFU-C within two to three weeks. Also, when there is poor development of the adherent layer, or if the bottles are siliconised, (thus preventing the attachment of potentially adherent cells) cell proliferation is poor and short lived. The adherent layer contains three morphological cell types; phagocytic mononuclear cells with numerous cytoplasmic extensions and an overall tendency to spread along a single axis (fig. 11, flat-
tened cells which tend to form a confluent monolayer (fig. 21, and a third cell type, which develops over a 3-week period and forms clusters and is characterised by an enormous accumulation of a lipid type of material which stains intensely with Sudan Black B and is also extremely osmiophilic (fig. 3). The accumulation of lipid within these cells causes an increase in size up to a range of 50-150 p m and these “giant fat” cells themselves form radiating foci (fig. 4) clearly visible in the bottles with the naked eye. There is a marked tendency for maturing granulocytes to cluster in and around the aggregations (fig. 4). In t h e absence of these “giant fat” cells there is no CFU-S maintenance.
CuEture at 37’C-weekly feeding The data from a “typical” experiment are shown in table 1.Cell proliferation is obviously occurring since, after the second week, the numbers of cells in suspension approximately double between each feeding. As the numbers of adherent cells also remain fairly constant throughout the experiment, (as judged by re-
339
HAEMOPOIETlC STEM CELLS IN VITRO
Fig. 4 Scanning electron micrograph of an area of characteristic aggregation of “giant fat” cells (the larger spherical cells) and granulocytes (small spherical cells) clustered around the periphery. The mononuclear phagocytic and “epithelial” cells forming the adherent layer are also visible. Scale bar = 100 pm.
moval of adherent cells with a rubber policeman and counting directly on a haemacytometer) the increase in the numbers of suspended cells represents a true proliferation rather than simply detachment of the adherent population. Similarly the pluripotent stem cells (CFUS ) are proliferating a t least for nine weeks, also showing an approximate weekly doubling. I t is noteworthy that CFU-S and CFU-C are not detectable in the adherent cell population. This was established by removing all the growth medium from cultures which had been maintained for 2 , 6 or 10 weeks. The adherent cells were than washed once in Fischer’s medium removed mechanically with a rubber policeman and assayed for CFU-S and CFU-C content. In three separate experiments they were uniformly negative. Hence to monitor changes occurring in their number it is sufficient to assay the non-adherent cell population which can be suspended by gentle agitation of the bottles. These CFU-S are apparently normal, forming erythroid, gran-
TABLE 2
Experimental uariation in extents of stem cell maintenance in uitro Number of experiments
6 10 8 3
CFU-S maintenance (weeks)
CFU-C maintenance (weeks1
3-4
4-5 7-8
6-7
9-12 > 13
10-13 > 14
Maintenance signifies the time (in weeks of culture1 at which CFUS and CFU-C were no longer detectable.
ulocytic and megakaryocytic colonies in the spleens of potentially lethally irradiated mice, and if sufficient numbers of stem cells are injected (approximately 40 CFU-S), the cultured cells can protect lethally irradiated mice from haemopoietic death. The “committed” granulocytic precursor cells (CFU-C) are also being maintained for several months (table 1).These are presumably being produced by differentiation from the pluripotent stem cell-although a limited
340
T. M. DEXTER, T.
L).
ALLEN AND L. G. LAJTHA
capacity for their self-duplication is evident from their sensitivity to the 3H-thymidine “suicide” (Lajtha, et al., ’69). The colonies formed by these CFU-C i n agar cultures are similar to those produced from freshly isolated bone marrow cells, showing maturation towards granulocytes and macrophages, depending upon t h e addition of appropriate colony stimulating activity (CSA) -in this case, heart conditioned medium (Dexter and Testa ’76). In the absence of CSA, they did not form colonies in agar. In all cultures, extensive granulopoiesis is seen initially with all maturation stages represented. With time in culture, however, granulocytes TABLE 3
Egect of twice weekly feeding on stem cell maintenance (1experiment)
Variations in stern cell maintenance We have previously found that stem cell proliferation in the liquid cultures is dependent to a great extent upon the batch of serum used (Dexter and Testa, ’76) with horse serum (Flow Laboratories) providing most consistent growth. Other sera, e.g., horse (Gibco), foetal calf (Flow and Gibco), baby calf (Gibco) have been found to be unsuitable for maintenance of stem cell proliferation in vitro. How-
~
Days cultured
Number of feedings
7 18 34 49 62 83 109
2 5 10 14 18 24 30
Cell count ( X 10’)
8.0 6.0 1.2 1.2 2.1 2.1 5.1
decline, macrophages increase and in many cultures there is also an accumulation of “blast” cells (mononuclear cells of approximately 8-10y diameter with immature nuclei and one or two nucleoli). The nature of these blast cells is unknown but surface marker studies have shown t h a t they possess neither 0 antigen characteristic of T lymphocytes (Raff and Wortis, ’70) nor surface lg determinants characteristic of B-lymphocytes (Raff et al., ’70). Furthermore they show no myeloperoxidase or S u d a n Black a c t i v i t y (characteristic of maturing granulocytes). With longer time (> 9 weeks) in culture the numbers of such blast cells decrease and phagocytic mononuclear cells become the predominant cell type. Proliferation of such mononuclear cells has been observed in some cultures for periods in excess of 30 weeks.
CFU-Slculture (suspension)
166 2 20 11o-t 12 153 -t 27 28 -t 5 8 56 -t 10 0
(Once weekly feeding “control” culture shown In table 1) I Equivalent to population doublings.
TABLE 4
Effect of temperature on stem cell maintenance Group
37OC 33°C 37oc 33°C 37oc 33°C 37T 33OC 37% 33oc 37°C 33°C 37oc 33°C 37°C 33-2 37°C 33T
Weeks cultured
1 1 3 3 4 4 6 6 7 7 9 9 10 10 11 11 12 12
Cell count ( X 105)
13.0(5-24) 62.0(25-130) 5.0 (2-11) 52.006-98) 5.0 (3-9) 60.0(25-140) 7.5(6-9) 41.5 (40-431 4.5(3-7) 28.0(20-32) 6.5(4-9) 22.0(16-30) 8.5 (5-12) 15.0(11-18) 8.5 (3-16) 22.006-30] 5.0(5) 15.0(9-22)
ND, no data. Results are the mean (and range) of four or five experiments.
CFU-SI culture
290 (97-558) 780 (320-1254) 230 (50-695) 832 (295-1253) 209 (30-384) 811 (170-1944) 270 (60-392) 590 (290-8921 96 (21-71) 197 (154-220) 102 (46-219) 225 (100-356) ND ND 75 (0-250) 125 (35-220) 12 (0-140) 45 (0-180)
CFU-C/culture (1 exwriment)
9,000 21,000 8,000 12,000 ND ND 2,000 3,000 ND ND ND ND 52 1,130 ND ND 0 162
34 1
HAEMOPOIETIC STEM CELLS IN VITRO
ever, even using the same batch of serum there is a variation in the extent of stem cell maintenance. Table 2 shows that approximately 25% of the experiments show only limited stem cell maintenance with a gradual decline in stem cell numbers and rapid conversion to phagocytic mononuclear cells. In about 60%of the experiments, however, stem cell proliferation can been seen for periods between 6 and 12 weeks; an absolute increase in numbers of CFU-S and CFU-C occurs between each feeding. In some experiments there is maintenance of stem cells for up to 16 weeks. In all these long term cultures blast cell accumulation is regularly seen. The observation that CFU-C proliferation persists somewhat longer than that of CFU-S indicates a small but definite self-renewal capability of CFU-C.
Culture at 37°C-twice weekly feeding Feeding a t weekly intervals may not be imposing a great proliferative stress upon the cells, indeed a "plateau" value after one doubling may be reached well before seven days in culture. This "plateau value" if maintained for several days may, in fact, represent suboptimal culture conditions, thus, limiting the overall proliferative capacity of the stem cells. Feeding a t more frequent intervals was therefore used to determine whether the number of "population doublings" of the stem cell population could be increased (all other factors being constant). The data shown in table 3 with two times weekly feeding indicate that higher numbers of "doublings" can be achieved by altering the feeding regime. ?he effect of temperature on stem cell maintenance
In preliminary experiments investigating the effect of temperature on leukaemic transformation in vitro it was noticed that in cultures kept at 33"C, the initial three weeks "setting up period" produced significantly more adherent cells (including giant fat cells) than those kept a t 37°C. When such adherent layers were "recharged" with bone marrow cells and the culture was continued a t 33°C with weekly feeding-there was a striking effect upon the numbers of cells produced in suspension (table 4). For the first five to six weeks almost ten times more cells were present a t 33°C compared with 37"C, still doubling a t weekly intervals. With regard to stem cell maintenance, CFU-S at 33°C are consist-
TABLE 5
Morphology of cells obtained from culture maintained at 37°C or 33°C Morphology Group
37°C 33°C 37T 33°C 37°C 33°C 37T 33°C 37°C 33OC
Weeks cultured
Blasts 9:
Granulocytes %
1 1 3 3 5 5 8 8 10 10
10 7 10 6 31 5 16 10 0 9
78 83 83 80 53 71 6 19 3 30
Mononuclear cells %
12 10
7 14 16 22 78 70 96 61
TABLE 6
Temperature effecton ability ofthe adherent layer to support stem cell proliferation Group
37°C 33°C 330-370c 37°C 33°C 33"-37"C 37% 33oc 33"-37"C 37°C 33oc 33"-37"C
Weeks cultured
Cell count ( x lo5)
CFU-S/ culture
1 1 1 3 3 3
11.6 29.0 59.0
380 700 1,400 240 2,800 1,550 180 700 800
4
4 4 6 6 6
4.0
74.0 40.0 2.0 56.0 2.0 5.3 58.0 4.4
148 986 120
37°C-refers to cultures maintained wholly at 37°C 33'C-refers to cultures maintained wholly at 33OC 33"-37"C-refers to cultures where the adherent population was initiated at 33°C and the cultures transferred to 37OC subsequent with the inoculation of the second bone marrow population
ently maintained a t levels two to four times that seen a t 37°C-at least for 11 weekswhilst the effect on the CFU-C is not so marked. In terms of the morphology of the cells produced-the two cultures show a broadly similar appearance, with granulopoiesis for five to seven weeks being followed by an increase in phagocytic mononuclear cells (table 5). The percentage of blast cells a t 33°C however seldom exceeds 10% for the duration of the cultures. Of particular interest is our preliminary finding that the variation seen in extent of stem cell maintenance a t 37°C (using the same batch of serum) does not occur a t 33°C:
342
T. M. DEXTER, T. D. ALLEN AND L. G. LAJTHA TABLE 7
Effect of feeding cultures with medium containing CSA (colony stimulating activity) Morphology Group
~~~
Weeks cultured
Cell count x 105
CFU-S
1 1 1
7
3 3 3
4
412 44 8 150 5 0
Blasts X
Granulocytes %
Mononuclear cells X
20 5 1 12 2 0
66 48 13 58 20 0
15 40 86 30 I8 100
~
Control MECM MHCM Control MECM MHCM
5 4 3 2
MECM, mouse embryo conditioned medium; MHCM, mouse heart conditioned medium. Data from one experiment are shown. In two further experiments similar results. were obtained.
all cultures so far maintained a t 33°C have shown extensive granulopoiesis and stem cell maintenance for a t least six weeks. Since this may have been due to a more uniformly “effective” adherent layer being produced during the initial three weeks setting-up period we have investigated the capacity of such adherent layers produced a t 33°C but subsequently transferred to 37”C, to maintain the growth of stem cells. Cultures were set up as before a t 33°C (with weekly feeding) but after the initial three weeks, they were transferred to 37”C, fed, re-charged with the second bone marrow population and maintained at 37°C and fed in the usual manner. The results are shown in table 6 and indicate the improved capacity of the adherent layer established a t 33°C to support the subsequent production of granulocytes and stem cells a t 37°C for at least four weeks, however, by six weeks this capacity of the adherent layer appears to be “exhausted” and the situation reverts to the “control” 37°C level.
The role of ‘CSA” in the cultures In view of the extensive granulopoiesis seen for up to two months in some cultures and a view of the proposed role of colony stimulating activity (CSA) in inducing granulocyte development in agar cultures in vitro (Metcalf ’70) and the in vivo evidence supporting the role of CSA as a physiological regulator of granulopoiesis (Moore et al., ’74) the culture media were tested for the presence (?endogenously produced) of CSA. The growth medium (from feeder layers alone or from cultures where the second “target” population of bone marrow cells had been added) removed after various culture periods a t 37”C, was centrifuged and the “cell
free” supernate was tested (in the single layer agar system) for its ability to stimulate the development of CFU-Cfrom either freshly isolated bone marrow cells or from the cultured marrow cells. Various concentrations of such “conditioned” media obtained from our bone marrow cultures were tried 6 7 5 % ) but the results were uniformly negative (i.e., no cluster or colony formation was induced). Even after heat treatment (56°C for 30 minutes) and/or extensive dialysis (thought to remove molecules which can inhibit CSA (Bradley and Sumner, ’68)no colony stimulating activity was detectable. Furthermore, neither the suspended cells nor the adherent layer cells would promote the growth of bone marrow CFU-C when used as a “feeder layer” in the double layer agar system. Since our culture system appeared to deficient in “CSA’, this raised the possibility of further improvement by providing exogenous CSA for the cultures. Consequently, together with the re-charging of the cultures (after the initial three weeks setting up period) the medium was supplemented with mouse embryo or heart conditioned medium (MECM or MHCM) as a source of CSA (Dexter and Testa, ’76). The final concentration of MECM or MHCM in the cultures was 15%-in the single layer agar system this concentration is known to induce optimal, i.e., plateau, numbers of CFU-C using either freshly isolated bone marrow or cultured bone marrow as target cells. The cultures were continued to be fed a t weekly intervals with growth medium supplemented with MECM or MHCM, the cells collected and assayed as before. The results are shown in table 7. Feeding with CSA containing “conditioned” medium caused a marked decrease of the stem cells
HAEMOPOIETIC STEM CELLS IN VITRO
after one week and their virtual disappearance after three weeks in culture. Concomitant with the decline in CFU-S, CFU-C were also no longer detectable after four weeks in culture. Also there was a marked effect upon the morphological appearance of the cells: the CSA supplemented cultures underwent an early transition to mononuclear cells. The addition of MECM or MHCM produced no apparent increase in cell numbers in the cultures, nor did it cause a change in the cell types present to mature forms when added to cultures with accumulated blast cells-even after several days of cultures with high CSA activity in the medium. DISCUSSION
Cultures methods have been described before in which stem cell proliferation could be maintained for a matter of days (Sumner et al., ’72; Cline and Gold, ’76) or even a few weeks (Dexter and Lajtha, ’74).The introduction of an adherent (? nurse) layer from the initial “setting-up’’ bone marrow inoculum (Dexter and Lajtha, ’76; Dexter and Testa, ’76) presents a significant improvement, by maintaining granulopoiesis and stem cell proliferation for two to three months. Whilst this time scale indicates that the culture conditions (and the behavior of cells) may approach physiological states, a t least during the few weeks of culture, the method as used to date is clearly not yet optimal. It has been indicated that two times weekly feeding improves the rate of stem cell production and maintenance, further improvement may be achieved with more frequent medium change. For example, 48 hours may be sufficient to allow stem cell population doubling between medium changes, and to avoid a “plateau” state in the cultures which may result in suboptimal conditions. The accumulation of blast cells, after about five weeks (? maturation block) may be an indication of some suboptimal state, although it should be emphasized that even with such “maturation arrest” the cultures produce normal CFU-S and CFU-C for several weeks, i.e., cells which under appropriate conditions (as splenic or as agar colony formers) do produce normal mature granulocytes. The observation that the cultures do very much better at a lowered incubation temperature incorporating a better maintenance of
343
stem cells, a more effective adherent layer; higher plateau values of cells, and a delayed accumulation of blast cells is relevant in this respect. Under suboptimal conditions the cells may gradually build up increasingly unbalanced states-either in intracellular or in intercellular respects (e.g., numerical imbalance between interacting cell populations). The inevitable slowing down of some metabolic events a t 33”C, compared with 37°C may go some way to prevent or delay the development of such “imbalances.” It is interesting that it can do this without apparently significantly affecting the measurable cell proliferation rates. That the cells can achieve and maintain higher cell concentrations a t the lower temperature is certainly one indication of a generally improved milieu. The absence of detectable amounts of colony stimulating activity (CSA) in the harvested culture media is somewhat surprising. CSA is obligatory in semisolid cultures (agar or methylcellulose) not only for granulocytic colony growth (Metcalf, ’701, but even for temporary maintenance of CFU, (Testa and Lajtha, ’73). It was conceivable that our cultures produced a very low concentration of endogenous CSA and that the blast cell accumulation was, in fact, due to suboptimal CSA in the medium. As has been described in the results, however, addition of exogenous CSA did not help: not only did i t not “revert” the blast cells to mature forms, but it caused a speedy rundown of the cultures. It is possible that this was caused by some impurities in the CSA which would be absorbed by the agar or methylcellulose in the semisolid cultures, but it must be remembered that the semisolid cultures represent a system which is running down within a matter of days while active granulopoiesis persists in our cultures-with undetectable amounts of CSA in the medium-for many weeks. It is also possible that molecules with CSA inhibitory activity are present in these cultures. However if this be the case it would still be difficult to explain why active granulopoiesis is occurring in the presence of inhibitory concentrations sufficient to mask CSA. While our culture system of bone marrow cells settled on a special adherent layer of cells is not a complete representation of bone marrow functions-so far we have been unable to demonstrate, e.g., erythropoiesis or thrombopoiesis-by its long maintenance and extent of stem cell prolifera-
344
T M DEXTER, T. D. ALLEN AND 1, G LAJTHA
tion (over 10-15 population doublings) i t offers a useful system for the study of factors involved in the control of proliferation and differentiation of some of the earliest haemopoietic cells. ACKNOWLEDGMENTS
This work was supported by t h e Medical Research Council and the Cancer Research Campaign. The authors wish to thank Nicola P. J. Higgins and G. R. Bennion for excellent technical assistance. LITERATURE CITED Bradley, T. R., and D. Metcalf 1966 The growth of mouse bone marrow cells in uitro. Austr. J. Exp. Biol. Med. Sci., 44: 287-300. Bradley, T. R., and M. A. Sumner 1968 Stimulation of mouse hone marrow growth in vitro by conditioned medium. Austr. J. Exp. Biol. Med. Sci., 46: 607-618. Cline, M. J.,and D. W. Golde 1976 Improved techniques for liquid culture of human and mouse bone marrow. Blood, 47: 369-379. Dexter, T. M., T. D. Allen, L. G. Lajtha, R. Schofield and B. I. Lord 1973 Stimulation of differentiation and proliferation of haemopoietic cells in uitro. J. Cell Physiol., 82: 461-474. Dexter, T. M., and L. G. Lajtha 1974 Proliferation of haemopoietic stem cells in uitro. Brit. J. Haematol., 28: 525-530. - 1976 Proliferation of haemopoietic stem cells and development of potentially leukaemic cells in uitro. In: Proceedings of VIII International Symposium on Comparative Research on Leukaemia and Related Diseases. Copenhagen, 1975. Publishers: Karger, Basel. Dexter, T. M., and N. G. Testa 1976 Differentiation and proliferation of haemopoietic cells in culture. In: Methods in Cell Biology. Vol. 14. D. M. Prescott, ed. Acadmic Press, New York, pp. 387-405. Edwards, G. E., R. G Miller and R. A. Phillips 1970 Differentiation of rosette-forming cells from myeloid stem cells. J. of Immunology, 105: 719-729. Fibach, E., E. Gerassi and L. Sachs 1976 Induction of colony formation in uitro by human lymphocytes. Nature, 259: 127-128. Lajtha, L. G., L. V. Pozzi, R. Schofield and M. Fox 1969 Kinetic properties of haemopoietic stem cells. Cell & Tissue Kinet., 2: 39-49.
Metcalf, D. 1970 Studies on colony formation in uitro by mouse bone marrow cells. 11. Action of colony stimulating factor. J. Cell Physiol., 76: 89-100. Metcalf, D., and H. R. MacDonald 1975 Heterogeneity of in uitro colony- and cluster-forming cells in the mouse marrow: Segregation by velocity sedimentation. J. Cell Physiol., 85: 643-654. Metcalf, D., and M. A. S. Moore 1971 Haemopoietic Cells. North Holland Publishing Co., Amsterdam. Metcalf, D., H. R. MacDonald, N. Odartchenko and L. B. Sordat 1975a Growth of mouse megakaryocyte colonies in uitro. Proc Nat. Acad. Sci., 72: 1744-1748. Metcalf, D., N. L. Warner, G. J. V. Nossal, J. F. A. P. Miller, K. Shortman and E. Rabellino 19751, Growth of Blymphocyte colonies in uitro from mouse lymphoid organs. Nature, 255: 630-632. Raff, M. C. 1969 Theta isoantigen as a marker of thymus-derived lymphocytes in mice. Nature, 224: 378-379. Raff, M. C., M. Sternberg and R. B. Taylor 1970 Immunoglobulin determinants on the surface of mouse lymphoid cells. Nature, 225: 553-534. Moore, M. A. S., G. Spitzer, D. Metcalf and D. G. Pennington 1974 Monocyte production of colony stimulating factor in familial neutropenia. Brit. J. Haematol., 27: 47-55. Pluznik, P. H., and L. Sachs 1966 The induction of colonies of normal “mast” cells by a substance in conditioned medium. Exp. Cell Res., 43: 553-563. Sredni, B., Y. Kalechman, H. Michlin and L. A. Rozenzajn 1976 Development of colonies in uitro of mitogen stimulated mouse T-lymphocytes. Nature, 259: 130-132. Stephenson, J. R., A. A. Axelrad, D. L. McLeod and M. M. Shreeve 1971 Induction of colonies of haemoglobin synthesizing cells by erythropoietin in uitro. Proc. Nat. Acad. Sci., 68: 1542-1546. Sumner, M. A., T. R. Bradley, G. S. Hodgson, M. J. Cline, A. Fry and L. Sutherland 1972 The growth of bone marrow cells in liquid culture. Brit. J. Haematol., 23: 221-234. Testa, N. G., and L. G. Lajtha 1973 Comparison of the kinetics of colony forming units in spleen (CFU,) and culture (CFU,). Brit. J . Haematol., 24: 367-376. Till, J. E., and E. A. McCulloch 1961 A direct measurement of the radiation sensitivity of normal mouse bone marrow cells. Rad. Res., 14: 213-222. Williams, W., and G. J. van den Engh 1975 Separation of subpopulations of in uitro colony forming cells from mouse marrow by equilibrium density centrifugation. J. Cell Physiol., 86: 237-246. Wu, A. M., J. E. Till, L. Siminovitch and E. A. McCulloch 1968 Cytological evidence for a relationship between normal haematopoietic colony forming cells and cells of the lymphoid system. J. Exptl. Med., 127: 455-463.
