Bone and Mineral, 18 ( 1992) 199-2 13 0169-6009/92/$05.00 0 1992 Elsevier Science Publishers B.V. All rights reserved,
199
BAM 00460
one marrow stromal colony formatio stirn~l~t~o~ by ~~ern~~oietic cells
A.J. Friedenstein, N.V. Latzinik, Yu.F. Gorskaya, EA. Luria and I.L. Moskvina
(Received 25 November 1991) (Accepted 14 April 1992)
Summary In mouse bone marrow cultures plated at low cell density, stromal colonies formed from colony-forming unit fibroblastic (CFUf) failed to develop unless the cultures were supplemented with irradiated feeder cells. Colo~ly-stimulating activity was pro&reed by irradiated bone marrow and spleen cells and by platelets, was dose dependent, not species specific and was maximal at high serum con~ntration. The efficiency of CFUf cofony formation was 1.7 x IO+ for mechanically disaggregated and 14.6 x IO* for trypsinised bone marrow cells. The colonies formed in the presence of feeder cells comprised hundreds of tibroblasts. In the absence of feeder cells, small tibrobtast foci and single libroblasts only were present in cultures. PDGF, IL-3 and EGF did not substitute for the colony-stimulating activity of feeder cells. These results suggest that CFUf colony formation requires growth factor(s) released by platelets and megakaryocytes which ~main to be identity.
Key words: Bone marrow; CFUf colonies; Growth factors
Bone marrow stromal fibroblasts (MSF) belang to the osteogenic stromal cell lineage and are histogenetiically distinct from haemopoietic cells [l-3]. They play an important part in establishing the haemopoietic microenvironment both in vivo and in vitro [4-71.The osteogenic potential of MSF was proven the ability of cultured stromal fibroblasts to form bone and cartilage upon tra~lsplantation Correspondence to: Dr. A.J. Friedenstein, Gamalcya Institute for Epidemiology and Microbiology, Academy of Medical Sciences, Gamaleya Str. 18, Moscow D-98, Russian Federation.
both into diffusion chambers and at heterotopic sites in vivo [S-10]. Under steady-state conditions in vivo, MSF have a slow rate of self renewal. Despite this they include the clonogenic precursor cells (the CFUf), which are local, non-repopulating and highly adherent cells [3,11-161. In situ more than 90% of CFUf remain arrested in the Go-period of the cell cycle 117,181while in bone marrow cell cultures they start to proliferate and form adherent cell colonies composed of fibroblasts [8,19].CFUf-derived fibroblasts synthesize collagen type I and III, lack macrophage markers and factor-VIII-associated antigen [20,21]. Approximately 50% of the Iarge CFUf colonies formed have high proliferative paucity and ost~genic potential as shown by transplantation into di~usion chambers of single-colony-derived strains of MSF (221,and a similar proportion of these strains respond to parathyroid hormone (PTH) by an increase in adenylate cyclase activity [23], which is an accepted characteristic of the osteogenic phenotype 1241;approximately 10% of single-colony~erived strains of MSF have both osteogenic and chondro~ni~ potential f22J.These observations infer that the initiating CFUf were highly proliferative cells, some of them having the properties attributed to osteogenic stem cells [16,25,26]. CFUf colony formation requires serum growth factors, PDGF in particular 1271,and when bone marrow cells are explanted in media containing serum, CFUf colonies are formed without additional stimulation [8,20].However, the efficiency of colony formation (ECFf) decreases with decreasing explantation density of marrow cells, but can be increased again by adding irradiated marrow feeder cells (FMC) to the cultures [28,29]. This indicates that the haemopoietic cells which a~ompany CFUf in marrow cell cultures stimulate CFUf colony fo~ation, Here we have studied CFUf colony formation in adherent marrow cell cultures (AMCC) depleted of haemopoietic cells. We conclude that stromal colony formation in fact requires stimulation by haemopoietic marrow cells or by platelets, which release a colony-stimulating factor into the supernatant culture m~ium.
Materials and Methods
Male CBA mice of 20 g weight were used as donors of bone marrow cells (BMC) for the CFUf colony assay [2]. The marrow was flushed from a femur or several femora and single-cell suspensions were prepared in minimum essential medium (aMEM) by repeatedly flushing marrow fragments through a Pasteur pipette (m~hani~al segregation) or alternatively by trypsini~tion for I k in 0.25% trypsin in Hanks sotution on a magnetic mixer [8,13]. BMC suspensions thus prepared were filtered through a nylon filter, 90-pm pore size, and after washing twice with aMEM the cells were explanted into T-25 flasks or 35mm dishes. The culture medium consisted of aMEM with 20% fetal calf serum (FCS) or, in some ex~~ments, with l-S% FCS or 20% heat-inactivated human serum (HS). Testing for CFUf colony formation [28]was carried out in two different ways: by culturing the complete population of marrow cells containing both adherent
201
and non-adherent cells, designated conventional marrow cell cultures (CMCC) or by culturing the adherent BMC only. Adherent marrow cell cultures (AMCC) were established by culturing the explanted BMC for 2 h in medium with 10% FCS and then removing the medium and non-adherent cells. The adherent cells were then washed three times with ~MEM and the flasks filled with fresh medium supplemented with FCS. At this stage, or at intervals between 24 h and 7 days, irradiated feeder cells were added to the A&ICC where indicated. Feeder cells usually remained in AMCC during subsequent colony growth, but in some experiments the culture medium with non-adherent feeder cells was removed and replaced with fresh medium at intervals indicated in the text. Mechanically disaggregated BMC, and spleen, thymus and lymph node cells [31] from CBA mice and guinea pigs, and also leukocytes and platelets isolated from titrated blood of ~~lineapigs and rabbits were used as feeder cells. All feeder cells were irradiated with a dose of 60 Gy f60Co) and some of the feeder cells were also treated with EDTA, 0.4% in liovine serum at 37°C for 15min, to remove platelets adhered to nucleated cells. Purified growth factors, PDGF (human platelets, Sigma), EGF (mouse, Sigma) and IL-3 (mouse, Genzyme) were added to AMCC in 35-mm dishes and in T25 flasks after non-adherent BMC were removed. Colonies were grown at 37°C in a CO2 incubator with 5% CO2 in air. After IO-19 days the cultures were fixed with 10% formalin in phosphate buRer and stained by the Giemsa method. The number of CFUf colonies consisting of more than 50 fibroblasts was counted under a dissecting microscope; fibroblast foci consisting of less than 50 cells and single fibroblasts were counted separately in some of AMCC cultures. ECFf was determined as the number of CFUf colonies containing more than 50 cells, per lo4 explanted BMC. Results
The development and structure of CFUf colonies was similar in mouse AMCC supplemented with either guinea pig or mouse FMC and in CMCC. Developing CFUf colonies could be recognised at 3-6 days of culture as clusters of elongated fibroblasts (Fig. la). Later, these ~brobl~sts increased in number forming looselyknit foci which include dozens of forming loosely ~broblasts; beginning from the 67th days, the size of the fibroblast foci increased rapidly so that by 10 days most of the CFUf colonies were composed of hundreds or thousands of flbroblasts (Fig. lb). Some of the colonies also contained macrophages and myeloid cells in addition to the fibroblasts. In CMCC the ECFf increased with increasing number of BMC explanted (Fig. 2). In AMCC where the non-adherent cells had been removed, the ECFf was less than in CMCC. However, in AMCC supplemented with FMC the ECFf increased in a dose-de~ndent manner up to a maximum value at a concentration of 10’ guinea pig FMC per ml (Fig. 3). Under these conditions the ECFf achieved a stable value over a range of plating densities of BMC which differed by several
Fig, I. CFUf colony formationin adherentbone marrowcell cultures (AMCC)supplementedwith marrow feedercells. (a) Fibroblastfoci in &day living culture ( x 200). (b) CFUf colony in IO-dayliving culture ( x 200).
orders of magnitude (Table 1). Mechanically disaggre~ted or trypsinizcd marrow cells were equally susceptible to stimulation by FMC and the colonystimulating activity of FMC was non species specific for mouse and guinea pig. Data obtained from 20 experiments indicate that ECFf in the cultures supplemented with feeder cells was t .7 + 0.2 (SEM) for mechanically disaggregated and 14.6 + 4.1 for trypsinized BMC. ECFf in FC-supplemented cultures was dependent on the concentration of FCS in the medium and was greatest at high
203
Fig. 2. Mechanically disaggregated bone marrow cells were explanted in T25 culture flasks with culture medjum supplemented with 20% FCS (CMCC). Number of colonies formed is plotted against number of ceils explanted. The cultures were fixed at 10 days, Results shown are for triplicate platings, bars are SEMs. 100
3
‘E
80
s 8
3x10s FC lo’ FC
60
2~10~ FC
“0 $j 40 P z
20 0
CMCCAMCC
AMCC+FC
Fig. 3. 3 x 10’ mechanically disaggregated bone marrow cells were explanted per T25 flask in culture medium supplemented with 20% FCS. The cultures were fixed at 10 days. Number of colonies formed is shown for the following conditions: CWCC, Conventional marrow cell cultures. AMCC, Adherent marrow cell cultures, non-supplemented with feeder cells. AMCC+FC, AMCC supplemented with 3 x IO’, 10’ or 2 x 10’ guinea pig feeder marrow cells.
serum concentration. No colonies were formed in the presence of 1% serum, independent of concentration of FC (Fig, 4). This was also true for cultures with HS. This indicated that serum growth factors do not substitute for colonystimulating activity of feeder cells and vice versa, Single spindle-shaped fibroblasts could be distinguished 24 h after explantation
204 Table 1 ~~6iency of colony formation in AMCC supplemented with 10’ feeder cells is independent of the number of bone marrow cells explanted Number of feeder Number of explanted bone marrow cells a x 103 cells 10’ IO’ IO’ 10’ IO’ -
3 10 30 100 300 300
Number of CFUF colonies in cultures
Efficiency of colony formation (ECFf)’
3,353 10,11,11 21,33,37 101,105,118 298,338
10.0 10.7 10.0 10.8 10.5 0.01
08, 1
” AMCC were established by explantation of try~sinised BMC in T25 flasks in culture medium su~plemcnt~ with 20% FCX Guinea pig feeder marrow cells were added at 2 h and the cultures were fixed at 8 days, h ECFf per 104 bone marrow cells.
of cells in living ~0~9fewer-supplemented AMCC (Fig. 5a). They amounted to 0,02~ of the number of initially explanted m~hanically disaggregated and 0.2% of the number of trypsinized BMC. Isolated fibroblasts remained in the lo-day non-feeder-supplemented AMCC in addition to 2,3-&N-49 cell fibroblast foci (Fig. Sb,c) and occasional small CFUf colonies. Their total number in IO-day 100
0 10"
Sx106
lo7
3xld
Number of feeder cells
Fig. 4 AMCC were established after 5 x 10’ mechanically disaggregated bone marrow cells were explanted per T25 Bask in culture medium supplemented with SK FCS. After 2 h non-adherent ceils were decanted and fresh medium with FCS added together with guinea pig feeder marrow cells (Fe). Number of colonies formed in the AMCC is plotted against the number of FC for different concentration of FCS. Cultures were fixed at 10 days. Means ( f SEM) are shown for triplicate platings.
cultures corresponded with the number of single ~broblasts in the 24-h cultures (compare a with gin Fig. 6) and it is therefore natural to assume that foci of2,3-8 and 16-32 fibroblasts and CFUf colonies were formed following 1,2-3,4-S and over 6 divisions, respectively, of the original single tibroblasts. In fact, marked foci occupied the same position in the flasks as t.hesingle fibroblasts observed in the 24-h vultures. ECFf in non-feeder-supplemented AMCC was 0.02 and 0.4 for mechanically disaggregated and trypsinird BMC respectively, The number of CFUf colonies in AMCC supplemented with FC corresponded to 8@-90%of the number of single fibroblasts in 24-h non-supplemented AFv¶CCfrom mechanically disaggregated and trypsinized marrow cells respectively (compare a with h, Fig. 6). Maximal colony-stimulating effect of marrow feeder cells was observed when
ml* 5. Single fibroblasts and fibroblast foci in
AMCC non-supplemented with feeder cells. (a) Single
fibroblast in 24-h living culture (x 200). (b) Fibroblast foci in IO-day stained cultures (x 200). (c) Fibroblast foci in IO-day living culture ( x 200).
120 100 80 60 40
d8 E=
2% a
20
5
0
abcdefgh
q q
Mechanically disaggregated cells Trypsinised cells
Fig. 6. AMCC were established after explantation of 5 x 10’ mechanically disaggregated or 5 x lo4 trypsinised bone marrow cells in T25 flasks in culture medium supplemented with 20% FCS. The number of single ~broblasts at 24 h and single fibroblasts, fibroblast foci and small colonies at 10 days in AMCC non-supplements with FC, (a) to (g} inclusive, is compared with the number of CFUf colonies at 10 days in AMCC supplemented with FC (h). (a) Number of single fibroblasts in 24-h AMCC. (b) Number of single fibroblasts in IO-day AMCC. (c,d,e) Number of fibroblast foci containing 2, 3-8, and 9- 49 cells respectively in W-day AMCC. (f) Number of small colonies in IO-day AMCC. (g) The total number of single Woblast foci and small colonies (b + c + d + e + f) in IO-day AMCC non-supplemented with FC. (h) Number of CFUf colonies in IO-day AMCC supplemented with 10 ’ guinea pig feeder marrow cells. Means (+ SEM) are shown for triplicate platings.
207 Table Eficiency of colony formation in AMCC supplemented with feeder cells is dependent on the period of association with the feeder cells Period of CFUf-feeder cells association (days) -Exp.I
Number of CFUf colonies in cultures
O-10 O-l O-3 O-4 O-7
Mean no. of CFUf colonies
Percent of CFUf colonies
107,125 25,27 &I,52 s2,54 119,120
2 116 26 49 53 120
100 22 42 46 103
O,l,l 104,113,116 75,98,105 41,43,44 31.3650 2,639 234
ill 93 43 39 6 3
124
1
Exp.11 O-10 2-l 1 3-12 4-13 7-16 IO-19
I
I
100 84 39 35 5 2
AMCC were established by explantation of 8 x 10’ mechanically disaggregated BMC in T25 Basks with culture medium supplemented with 20% FCS. 10’ guinea pig feeder marrow cells per flask were added for the indicated periods of cultivation. In Experiment I, feeder cells were added at 2 h (0 days) and present for 10 days or were decanted and replaced by fresh culture medium after 1 to 7 days; the cultures were fixed at 10 days. In Experiment II, feeder cells were added at 2 h (0 days) and were present for 10 days, or they were added at 2 to 10 days and were present for a further 9 days; the cultures were fixed at 10-19 days.
FMC were added at 2 h after explantation of marrow cells and when they remained in cultures for 7 days, Experiment I (Table 2). The table also demonstrates, Experiment II, that after 3 and 7 days cultivation without feeder cells, from 50% to more than 90% of CFUf respectively lose the ability to be stimulated to colony formation by .FMC. In addition to BMC, spleen cells from mice and quinea pigs and platelets from guinea pigs and rabbits also had a colony-stimulating action: 2 x IO8platelets, moreover, had a stimulating action equal to that of 1.2 x 10’ bone marrow feeder cells (Fig. 7). Cell populations containing megakaryocytes (BMC and mice spleen cells) preserved their colony-stimulating activity after removal of their platelets by EDTA treatment (Fig. 8). Conve~ely, after removal of the platelets, guinea pig spleen cells, among which there were no megakaryocytes, had much reduced activity (Fig. 8). Thymocytes, lymph node cells and blood leukocytes did not stimulate the formation of CFUf colonies (Fig. 7). Addition of either PDGF, K-3, or EGF into AMCC did not lead to the formation of CFUf colonies (Table 3). In one instance where IL-3 and feeder cells were both present there was no added effect of IL-3 (Table 3).
208
0
m
I
S
t
b
P
80
60 40 20 0 0
s
m
I
t
m
2.5 X 1s” FCIflask
Q
lo7 platelets per flask
111
5.0 X 10’ FClflask
[13
3x107 platelets per flask
1.2 X 10’ FC/flask
B
2~10~ platelets per flask
@#
2.5 X 10’ FClflask
Fig+7. The effect of different populations of feeder cells (FC) on ECFF, expressed as percent of maximum value obtained. AMCC were established after explantation of 5 x 10 mechani~aliy disa~regated bone marrow ceils in T25 flasks in culture medium supplemented with 20% FCS. (A) Guinea pig FC, (B) mouse FC, FC were added to the cultures at 2 h. o, AMCC non-supplemented with feeder cells; m, marrow; &spleen: I, lymph node: t, thymus feeder cells; b, blood leucocytes; p, platelets. The cultures were fixed at 10-12 days. Means ( f SEM) are shown for triplicate platings.
209
80 80
B
c
D
Untreated feeder cells Feeder cells treated with EDTA Fig.8. The effect of EDTA treatment of feeder cells on ECFf, expressed as percent of maximum value obtained. AMCC were established after explantation of 6 x IO5 mechanically disaggregated bone marrow cells in T25 flasks with culture medium with 20% FCS. A, AMCC not supplemented with feeder cells; B, AMCC supplemented with IO’ guinea pig feeder marrow cells; C, AMCC supplemented with I.5 x IO’ guinea pig feeder spleen cells; D, AMCC supplemented with 0.8 x IO’ mouse feeder spleen cells. Feeder cells were added into cultures at 2 h. The FC were either treated or untreated with EDTA. Cultures were fixed at I2 days. Means (+ SEM) are shown for triplicate platings.
Discussion
Despite 90% of CFUf sticking to the surface of culture vessels within 1 h [32], CFUf colony formation practically failed in AMCC deprived of non-adherent haemopoietic cells and could be restored if irradiated marrow feeder cells were added into the cultures. Thus, CFUf have to be stimulated by haemopoietic cells to form fibroblast colonies, but so far, the requirement of CFUf for such stimulation has remained unrecognised in conventional marrow cell cultures where CFUf are accompanied by endogenous haemopoietic cells. That CFUf colonies are cell clones was determined previously by chromosome typing [22,28]. The stimulatory effect of marrow feeder cells is dose dependent in AMCC and their concentration can be adjusted independently of the number of explanted CFUf. Therefore in feeder-supplemented AMCC the efficiency of colony formation can exceed that in CMCC many-fold and remains constant independent of the explantation density of the bone marrow cells. There seem to be good reasons to suppose that CFUf are represented by single fibroblasts present at 24 h in AMCC. Indeed, the number of CFUf colonies in feeder-supplemented cultures corresponds to 85% of single fibroblasts in 24-h cultures, many of the colonies being topographically linked to single fibroblasts marked in living cultures 9 days earlier. Unless stimulated by feeder cells these
210 Table 3
The effect of PDGF, IL-3 and EGF on CFUf colony formation in AMCC No. of explanted bone marrow cells
Growth factors
3 x IO’ (35-mm dish)
-
No. of feeder cells 2x106 4x IO6 6x1@
10 ng PDGF/ml 50 ng PDGF/ml
3 x IO’ (T25 flask)
0.12 42,4S 43.48 z&61 0,o 0.0
0.40,1
2 x IO’ (T25 flask)
2 x 10’ (3Smm dish)
No. of CFUf colonies
20 ng 20 ag 3 ng 10 ng 30 ng 70 ng -
IL-3/ml IL-3/ml IL-3/ml IL-3/ml IL-3fml It3/ml
5 ng EGF/ml 20 ng EGF/ml 100 ng EGF/ml
44&O 60 090 o,o,o, 1 OS0 0.0 o,o 0.0 198 22,36 198 757 4*7
AMCC cultures were established by explantation of 5 x IO’ mechanically disaggregated bone marrow cells
in T25 flasks in culture medium supplemented with 20% FCS. Growth factors and guinea pig feeder marrow cells were added in the cultures at 2 h. The cultures were fixed at 10 days.
fibroblasts do not proliferate at all or else they accomplish 1 to 6 cell doublings to produce 2-49 cell foci of fibroblasts instead of forming CFUf colonies composed of hundreds or thousands of fibroblasts in feeder-supplemented cultures. The number of stromal colonies detected by the CFUf colony assay was dependent on the mode of cell di~ggregation in preparing marrow cell suspensions. In trypsindisaggregated marrow cell suspensions the CFUf concentration was much higher than in suspensions prepared by mechanical disaggregation. The differences largely depend on inactivation of a considerable number of CFUf by mechanical t~atment of the marrow [30]. Despite long survival times of single marrow fibroblasts in non-feeder-supplemented cultures their competence to form CFUf colonies gradually diminished and became completely lost after 7 days of cultivation in the absence of feeder cells. Maximal ECFf was reached when feeder cells were present in AMCC during the first 6 or 7 days. It has been also demonstrated that in CMCC the CFUf enter S-phase of their first mitotic period in vitro between 28 and 60 h after explantation [19]. Taken together these results seem to indicate that proliferation of freshly-isolated CFUf which were arrested in vivo of the Go period of the cell cycle, and of their early descendants, both required stimulation by growth factors produced by ~emopoietic cells.
211
The colony-stimulating activity of feeder cells is not species specific, is expressed by cell populations which include megakaryocytes and platelets and can probably be attributed to cells of the megakaryocyte lineage. Indeed, after removal of platelets, the guinea pig spleen cells which lack megakaryocytes lose the colony-stimulating activity, while mice spleen cells which contain both platelets and megakaryocytes retain their stimulating activity. In a previous study [34] it was shown that bone marrow feeder cells stimulated CFUf colony formation when separated from adherent marrow cells by HA millipore filter and by 2 mm of medium, which indicates that they release a diffusible colony-stimulating factor (factors) into the culture medium, Maximal stimulation of colony formation by feeder cells was reached at high serum concentration in the culture medium and feeder cells did not substitute for serum or vice versa. Thus, the stimulatory activity of feeder cells is probably not due to the fibroblast growth factors present in serum of which PDGF is an obvious representative. For the present, the requirements of stromal clonogenic cells for growth factors has been studied only in conventional CMCC, i.e. when CFUf were accompanied by non-adherent haemopoietic cells which may act as feeder cells. It has been demonstrated that CFUf colony formation requires PDGF [32] and in bone marrow cultures with serum-deprived culture medium or with culture medium supplemented with platelet-poor plasma, PDGF had to be added for CFUf colony formation to occur [27], maximal stimulation being achieved at 10 ng/ ml and 30 ng/ml PDGF in mice and human CMCC respectively. Here we have tested in AMCC with serum-rich culture medium whether PDGF, IL-3 or EGF could replace the colony-stimulating activity of feeder cells. In the circumstances, none of these growth factors induced CFUf colonies in non-feeder-supplemented AMCC and the growth factors released by feeder cells which are required for stimulating osteogenic stromal colony formation remain to be identified. Most of the CFUf colonies in mouse feeder-supplemented AMCC comprise hundreds or thousands of fibroblasts. Single-colony-derived strains of rabbit bone marrow fibroblasts undergo 17-22 population doublings and produce more than 10” descendents from the initial CFUf [22]. These observations confirm that CFUf colonies originate from highly proliferative clonogenic stroma1 cells [22]. The colonies are heterogeneous with regard to their potential for osteogenesis and in vitro differentiation [23]. It has been assumed that this diversity depends upon differences in the unequal potential of the cells which initiate CFUf colonies and that these cells represent stromal precursors at different levels in a putative hierarchy of the bone marrow osteogenic cell lineage [22,26]. Recently Owen and collaborators [29] have demonstrated that cell differentiation in CFUf colonies can be reversed and that this depends Upon serum growth factors. It remains to be determined whether colony-stimulating factors produced by haemopoietic cells influence the differentiation pattern of bone marrow fibroblasts and may thus promote diversity of CFUf colonies.
212 References 1 F~edenstejn AJ, Petra~ova KV, Kurafesova AJ, Frolova GP. Heterotopic transplants of bone marrow. Aua\y& of precursor cells for osteogenic and bemopoietic tissues. Transplantation 1968;6:230-247. Friedenstein AJ. Precursor cells of mechanocytes. Int Rev Cytol 1976;47:327-359. Feedenstein AJ, Ivanov-Smolenski AA, Chajlakyan RK. Origin of bone marrow stromal mmhanocytes in radiochimeras and heterotopic transplants. Exp Hematol 1978;6:440-444. F~~~stein AJ, Ch~lakyan RK, Latzinik NV, Panasyk AF, Keilis-Borok JV. Stromai cells responsible for transferor the mi~r~nvironment of hemopoietic tissues. Cloning in vitro and retransplantation in vivo. Transplantation 1974;17:331-340. 5 Friedenstein AI, Latzinik NV, Groshera AG, Gorskaya UF, Marrow microenvironment transfer by heterotopic transplantation of freshly isolated and cultured cells in porous sponges. Exp Hema,tol 1982;10:217-227. 6 Patt HM, Maloney MA, Flannery ML, Hematopoictic microenvironment transfer by stromai ~brobtasts derived from bone marrow va~ng in ~~~u~a~ty.Exp Hematol 1982~1~728-732. 7 Dexter TM, Alien TD, Lajtha LG. Conditions controlling the proliferation of hemopoietic stem ceils in vitro. Cell Physiol 1977;91:335-349, 8 Friedenstein AJ, Chailakyan RK, Lalykina KS. The development of iibroblast colonies in monolayer cultures of guinea-pig bone marrow and spleen cells. Cell Tissue Kinet 1970;3:393-403. 9 Ashton BA, Allen TD, Howlett CR, Eagleson CC, Hattori A, Owen M. Formation of bone and cartilage by marrow stromal cells in di~usion chambers in vivo. Clin Grthop 198O;lS~29~306. 10 Friedenstein AJ. Stromal mechanocytes of bone marrow: cloning in vitro and retransplantation in vivo. In: Thierfelder S, Rodt H, Koilb HJ, eds. Immunoiogy of bone marrow transplantation. Berlin: Springer-Verlag, 1980;t9-29. 11 Friedenstein AJ, Kuralesova Al. Osteogenic precursor cells of bone marrow in radiation chimeras. Transplantation 1971;12:99-108. 12 Golde DW, Hacking WG, Quan SG, Sparkes RS, Gaie RP. Origin of human bone marrow Bb~blasts. Br 3 Haematol ~980;~~83-187. 13 Witson FD, Greenberg GR, Konrad PN, Klein AD, Walling PA. Cytogenetic studies on bone marrow fibrobtasts from male-female haemopoietic chimera: evidence that stromal elements in human transplantation recipients are host type. Transplantation 1978;25:87-92. 14 Lennon JE, Micklem HS. Stromal cells in long-term murine bone marrow culture: FACS studies and origin of stromal cellsin radiation chimeras. Exp Hematol 1986;14:287-292, IS Simmons PJ, Przepilzka 4 Thomas ED, Torok-Storb B. Hostorigin of marrow stromal cells foliowing allo~neie bone marrow t~nsplantation. Nature 1987;328:429~31. t6 F~~enstein AJ, Osteogenic stem ~4s in the bone marrow. Bone Miner Res 1990;7:24.3-272. I7 Epichina SY, Latzinik NV. Proliferative activity of bone marrow stromal clonogenic cells, Bull Exp Biol Med 1976;81:55-58. 18 Kaneko S, Motomura, Ibayashi H. Differentiation of human bone marrow-derived fibroblastoid colony forming ceils (CFUf). Br J Haematol 1982;51:217-305. 19 Keiliss-Borok IV, Latxinik NV, Epichina SY, F~~enstein AJ. Dynamics of the fo~ation of fibroblast Wtonies in monolayer cultures of bone marrow, according to 3H thymidine in~~oratiou experiments. Cytologia 1971;13:1402-141I. 20 Castro-Malaspina H, Gay RE, Resnick J, Kapoor N, Meyers P, Uriariezi D, McKenzie S, Broxmeyer HE, Moore MAS. Characterisacion of human bone marrow fibroblast colony-forming cells (CFU-F) and their progeny. Blood 1980;56:289-301. 21 Simmons PJ, Torok-Storb B. Identification of stromaf cell precursors in human bone marrow by a novel mon~lonal anti~dy, STRO-I, Blood 1991;78:55-62. 22 F&den&n AJ, Chailakhyan RK, Gerasimon HV, Bone marrow osteogenic stem cells: in vitro cukivation and transplantation in diffusion chambers, Cell Tissue Kinet 1987;20:263-272. 23 Barling PM, Bennett JH, TritTitt JT, Gwen ME. The adenylate cyclase response to parathyroid hormone in Cultured rabbit marrow fibroblastic cells. Bone Miner 1989;7:23-30.
213 24 Rodan G, Heath JK, Yoon K, Noda M, Rodan SB. Diversion of osteoblast phenoty~s. C&a Foundation Symposium 1988;136:111-l 70. 25 Owen M. Marrow stromal stem cells. In: Stem Cell Symposium. J Cell Sci. 1988; Suppl.l0:63-76. 26 Owen M, Friedenstein AJ. Stromal stem cells: marrow derived osteogenic precursors. Ciba Foundation Symposium 1988;136~42-60. 27 McIntyre AP, Bjornson BM. Human bone marrow stromal cell colonies: response to hydro~~isone and dependent on platelet-derived growth factor. Exp Hematol 1986;17:833-839. 28 Latzinik NV, Gorskaya UF, Grosheva AG, Friedenstein AJ. The content of stromal colony-forming cells (CFUf) in the mouse bone marrow and the clonal nature of CFUf-derived stromal fibroblast colonies. ~togenesis 1986;17:27-35. 29 Bennett JH, Joyner CJ, Triffitt JT, Owen ME. Adipocytic cells cultured from marrow have osteogenic potential. J Cell Sci 1991;99:131-139. 30 Latzinik NV, Sidorovich SU, Friedenstein AJ. Effect of bone marrow trypsinisation on the efficiency of fibroblast colony formation in monolayer cultures. Bull Exp Biol Med 1981;9:356-360. 31 Fri~enstein AJ, Gorskaja UF, Kulagina NN. Fibroblast precursors in normal and irradiated mouse hemopoietic organs, Exp Hematol 1976;4:267-274. 32 Friedcnstein AJ. Determined and inducible osteogenic precursor cells. Ciba Foundation Symposium 1971;[email protected]. 33 Hirata J, Kaneko S, Nishimura JN, Mot~ura S, Ibayashi H. Effeet of platelet-delve growth factor and bone marrow-conditioned medium on the proliferation of human bone mar;row-derived libroblastoid colony-forming cells. Acta Hematol 1985;74:189-194. 34 Friedenstein AJ, Luria EA, Latzinik NV, Gorskaya YF. Humoral nature of the CSA of bone marrow cells on stromal colony formation in bone marrow cultures. Bull Exp Biol Med 199O;i10509-510.
199
BAM 00460
one marrow stromal colony formatio stirn~l~t~o~ by ~~ern~~oietic cells
A.J. Friedenstein, N.V. Latzinik, Yu.F. Gorskaya, EA. Luria and I.L. Moskvina
(Received 25 November 1991) (Accepted 14 April 1992)
Summary In mouse bone marrow cultures plated at low cell density, stromal colonies formed from colony-forming unit fibroblastic (CFUf) failed to develop unless the cultures were supplemented with irradiated feeder cells. Colo~ly-stimulating activity was pro&reed by irradiated bone marrow and spleen cells and by platelets, was dose dependent, not species specific and was maximal at high serum con~ntration. The efficiency of CFUf cofony formation was 1.7 x IO+ for mechanically disaggregated and 14.6 x IO* for trypsinised bone marrow cells. The colonies formed in the presence of feeder cells comprised hundreds of tibroblasts. In the absence of feeder cells, small tibrobtast foci and single libroblasts only were present in cultures. PDGF, IL-3 and EGF did not substitute for the colony-stimulating activity of feeder cells. These results suggest that CFUf colony formation requires growth factor(s) released by platelets and megakaryocytes which ~main to be identity.
Key words: Bone marrow; CFUf colonies; Growth factors
Bone marrow stromal fibroblasts (MSF) belang to the osteogenic stromal cell lineage and are histogenetiically distinct from haemopoietic cells [l-3]. They play an important part in establishing the haemopoietic microenvironment both in vivo and in vitro [4-71.The osteogenic potential of MSF was proven the ability of cultured stromal fibroblasts to form bone and cartilage upon tra~lsplantation Correspondence to: Dr. A.J. Friedenstein, Gamalcya Institute for Epidemiology and Microbiology, Academy of Medical Sciences, Gamaleya Str. 18, Moscow D-98, Russian Federation.
both into diffusion chambers and at heterotopic sites in vivo [S-10]. Under steady-state conditions in vivo, MSF have a slow rate of self renewal. Despite this they include the clonogenic precursor cells (the CFUf), which are local, non-repopulating and highly adherent cells [3,11-161. In situ more than 90% of CFUf remain arrested in the Go-period of the cell cycle 117,181while in bone marrow cell cultures they start to proliferate and form adherent cell colonies composed of fibroblasts [8,19].CFUf-derived fibroblasts synthesize collagen type I and III, lack macrophage markers and factor-VIII-associated antigen [20,21]. Approximately 50% of the Iarge CFUf colonies formed have high proliferative paucity and ost~genic potential as shown by transplantation into di~usion chambers of single-colony-derived strains of MSF (221,and a similar proportion of these strains respond to parathyroid hormone (PTH) by an increase in adenylate cyclase activity [23], which is an accepted characteristic of the osteogenic phenotype 1241;approximately 10% of single-colony~erived strains of MSF have both osteogenic and chondro~ni~ potential f22J.These observations infer that the initiating CFUf were highly proliferative cells, some of them having the properties attributed to osteogenic stem cells [16,25,26]. CFUf colony formation requires serum growth factors, PDGF in particular 1271,and when bone marrow cells are explanted in media containing serum, CFUf colonies are formed without additional stimulation [8,20].However, the efficiency of colony formation (ECFf) decreases with decreasing explantation density of marrow cells, but can be increased again by adding irradiated marrow feeder cells (FMC) to the cultures [28,29]. This indicates that the haemopoietic cells which a~ompany CFUf in marrow cell cultures stimulate CFUf colony fo~ation, Here we have studied CFUf colony formation in adherent marrow cell cultures (AMCC) depleted of haemopoietic cells. We conclude that stromal colony formation in fact requires stimulation by haemopoietic marrow cells or by platelets, which release a colony-stimulating factor into the supernatant culture m~ium.
Materials and Methods
Male CBA mice of 20 g weight were used as donors of bone marrow cells (BMC) for the CFUf colony assay [2]. The marrow was flushed from a femur or several femora and single-cell suspensions were prepared in minimum essential medium (aMEM) by repeatedly flushing marrow fragments through a Pasteur pipette (m~hani~al segregation) or alternatively by trypsini~tion for I k in 0.25% trypsin in Hanks sotution on a magnetic mixer [8,13]. BMC suspensions thus prepared were filtered through a nylon filter, 90-pm pore size, and after washing twice with aMEM the cells were explanted into T-25 flasks or 35mm dishes. The culture medium consisted of aMEM with 20% fetal calf serum (FCS) or, in some ex~~ments, with l-S% FCS or 20% heat-inactivated human serum (HS). Testing for CFUf colony formation [28]was carried out in two different ways: by culturing the complete population of marrow cells containing both adherent
201
and non-adherent cells, designated conventional marrow cell cultures (CMCC) or by culturing the adherent BMC only. Adherent marrow cell cultures (AMCC) were established by culturing the explanted BMC for 2 h in medium with 10% FCS and then removing the medium and non-adherent cells. The adherent cells were then washed three times with ~MEM and the flasks filled with fresh medium supplemented with FCS. At this stage, or at intervals between 24 h and 7 days, irradiated feeder cells were added to the A&ICC where indicated. Feeder cells usually remained in AMCC during subsequent colony growth, but in some experiments the culture medium with non-adherent feeder cells was removed and replaced with fresh medium at intervals indicated in the text. Mechanically disaggregated BMC, and spleen, thymus and lymph node cells [31] from CBA mice and guinea pigs, and also leukocytes and platelets isolated from titrated blood of ~~lineapigs and rabbits were used as feeder cells. All feeder cells were irradiated with a dose of 60 Gy f60Co) and some of the feeder cells were also treated with EDTA, 0.4% in liovine serum at 37°C for 15min, to remove platelets adhered to nucleated cells. Purified growth factors, PDGF (human platelets, Sigma), EGF (mouse, Sigma) and IL-3 (mouse, Genzyme) were added to AMCC in 35-mm dishes and in T25 flasks after non-adherent BMC were removed. Colonies were grown at 37°C in a CO2 incubator with 5% CO2 in air. After IO-19 days the cultures were fixed with 10% formalin in phosphate buRer and stained by the Giemsa method. The number of CFUf colonies consisting of more than 50 fibroblasts was counted under a dissecting microscope; fibroblast foci consisting of less than 50 cells and single fibroblasts were counted separately in some of AMCC cultures. ECFf was determined as the number of CFUf colonies containing more than 50 cells, per lo4 explanted BMC. Results
The development and structure of CFUf colonies was similar in mouse AMCC supplemented with either guinea pig or mouse FMC and in CMCC. Developing CFUf colonies could be recognised at 3-6 days of culture as clusters of elongated fibroblasts (Fig. la). Later, these ~brobl~sts increased in number forming looselyknit foci which include dozens of forming loosely ~broblasts; beginning from the 67th days, the size of the fibroblast foci increased rapidly so that by 10 days most of the CFUf colonies were composed of hundreds or thousands of flbroblasts (Fig. lb). Some of the colonies also contained macrophages and myeloid cells in addition to the fibroblasts. In CMCC the ECFf increased with increasing number of BMC explanted (Fig. 2). In AMCC where the non-adherent cells had been removed, the ECFf was less than in CMCC. However, in AMCC supplemented with FMC the ECFf increased in a dose-de~ndent manner up to a maximum value at a concentration of 10’ guinea pig FMC per ml (Fig. 3). Under these conditions the ECFf achieved a stable value over a range of plating densities of BMC which differed by several
Fig, I. CFUf colony formationin adherentbone marrowcell cultures (AMCC)supplementedwith marrow feedercells. (a) Fibroblastfoci in &day living culture ( x 200). (b) CFUf colony in IO-dayliving culture ( x 200).
orders of magnitude (Table 1). Mechanically disaggre~ted or trypsinizcd marrow cells were equally susceptible to stimulation by FMC and the colonystimulating activity of FMC was non species specific for mouse and guinea pig. Data obtained from 20 experiments indicate that ECFf in the cultures supplemented with feeder cells was t .7 + 0.2 (SEM) for mechanically disaggregated and 14.6 + 4.1 for trypsinized BMC. ECFf in FC-supplemented cultures was dependent on the concentration of FCS in the medium and was greatest at high
203
Fig. 2. Mechanically disaggregated bone marrow cells were explanted in T25 culture flasks with culture medjum supplemented with 20% FCS (CMCC). Number of colonies formed is plotted against number of ceils explanted. The cultures were fixed at 10 days, Results shown are for triplicate platings, bars are SEMs. 100
3
‘E
80
s 8
3x10s FC lo’ FC
60
2~10~ FC
“0 $j 40 P z
20 0
CMCCAMCC
AMCC+FC
Fig. 3. 3 x 10’ mechanically disaggregated bone marrow cells were explanted per T25 flask in culture medium supplemented with 20% FCS. The cultures were fixed at 10 days. Number of colonies formed is shown for the following conditions: CWCC, Conventional marrow cell cultures. AMCC, Adherent marrow cell cultures, non-supplemented with feeder cells. AMCC+FC, AMCC supplemented with 3 x IO’, 10’ or 2 x 10’ guinea pig feeder marrow cells.
serum concentration. No colonies were formed in the presence of 1% serum, independent of concentration of FC (Fig, 4). This was also true for cultures with HS. This indicated that serum growth factors do not substitute for colonystimulating activity of feeder cells and vice versa, Single spindle-shaped fibroblasts could be distinguished 24 h after explantation
204 Table 1 ~~6iency of colony formation in AMCC supplemented with 10’ feeder cells is independent of the number of bone marrow cells explanted Number of feeder Number of explanted bone marrow cells a x 103 cells 10’ IO’ IO’ 10’ IO’ -
3 10 30 100 300 300
Number of CFUF colonies in cultures
Efficiency of colony formation (ECFf)’
3,353 10,11,11 21,33,37 101,105,118 298,338
10.0 10.7 10.0 10.8 10.5 0.01
08, 1
” AMCC were established by explantation of try~sinised BMC in T25 flasks in culture medium su~plemcnt~ with 20% FCX Guinea pig feeder marrow cells were added at 2 h and the cultures were fixed at 8 days, h ECFf per 104 bone marrow cells.
of cells in living ~0~9fewer-supplemented AMCC (Fig. 5a). They amounted to 0,02~ of the number of initially explanted m~hanically disaggregated and 0.2% of the number of trypsinized BMC. Isolated fibroblasts remained in the lo-day non-feeder-supplemented AMCC in addition to 2,3-&N-49 cell fibroblast foci (Fig. Sb,c) and occasional small CFUf colonies. Their total number in IO-day 100
0 10"
Sx106
lo7
3xld
Number of feeder cells
Fig. 4 AMCC were established after 5 x 10’ mechanically disaggregated bone marrow cells were explanted per T25 Bask in culture medium supplemented with SK FCS. After 2 h non-adherent ceils were decanted and fresh medium with FCS added together with guinea pig feeder marrow cells (Fe). Number of colonies formed in the AMCC is plotted against the number of FC for different concentration of FCS. Cultures were fixed at 10 days. Means ( f SEM) are shown for triplicate platings.
cultures corresponded with the number of single ~broblasts in the 24-h cultures (compare a with gin Fig. 6) and it is therefore natural to assume that foci of2,3-8 and 16-32 fibroblasts and CFUf colonies were formed following 1,2-3,4-S and over 6 divisions, respectively, of the original single tibroblasts. In fact, marked foci occupied the same position in the flasks as t.hesingle fibroblasts observed in the 24-h vultures. ECFf in non-feeder-supplemented AMCC was 0.02 and 0.4 for mechanically disaggregated and trypsinird BMC respectively, The number of CFUf colonies in AMCC supplemented with FC corresponded to 8@-90%of the number of single fibroblasts in 24-h non-supplemented AFv¶CCfrom mechanically disaggregated and trypsinized marrow cells respectively (compare a with h, Fig. 6). Maximal colony-stimulating effect of marrow feeder cells was observed when
ml* 5. Single fibroblasts and fibroblast foci in
AMCC non-supplemented with feeder cells. (a) Single
fibroblast in 24-h living culture (x 200). (b) Fibroblast foci in IO-day stained cultures (x 200). (c) Fibroblast foci in IO-day living culture ( x 200).
120 100 80 60 40
d8 E=
2% a
20
5
0
abcdefgh
q q
Mechanically disaggregated cells Trypsinised cells
Fig. 6. AMCC were established after explantation of 5 x 10’ mechanically disaggregated or 5 x lo4 trypsinised bone marrow cells in T25 flasks in culture medium supplemented with 20% FCS. The number of single ~broblasts at 24 h and single fibroblasts, fibroblast foci and small colonies at 10 days in AMCC non-supplements with FC, (a) to (g} inclusive, is compared with the number of CFUf colonies at 10 days in AMCC supplemented with FC (h). (a) Number of single fibroblasts in 24-h AMCC. (b) Number of single fibroblasts in IO-day AMCC. (c,d,e) Number of fibroblast foci containing 2, 3-8, and 9- 49 cells respectively in W-day AMCC. (f) Number of small colonies in IO-day AMCC. (g) The total number of single Woblast foci and small colonies (b + c + d + e + f) in IO-day AMCC non-supplemented with FC. (h) Number of CFUf colonies in IO-day AMCC supplemented with 10 ’ guinea pig feeder marrow cells. Means (+ SEM) are shown for triplicate platings.
207 Table Eficiency of colony formation in AMCC supplemented with feeder cells is dependent on the period of association with the feeder cells Period of CFUf-feeder cells association (days) -Exp.I
Number of CFUf colonies in cultures
O-10 O-l O-3 O-4 O-7
Mean no. of CFUf colonies
Percent of CFUf colonies
107,125 25,27 &I,52 s2,54 119,120
2 116 26 49 53 120
100 22 42 46 103
O,l,l 104,113,116 75,98,105 41,43,44 31.3650 2,639 234
ill 93 43 39 6 3
124
1
Exp.11 O-10 2-l 1 3-12 4-13 7-16 IO-19
I
I
100 84 39 35 5 2
AMCC were established by explantation of 8 x 10’ mechanically disaggregated BMC in T25 Basks with culture medium supplemented with 20% FCS. 10’ guinea pig feeder marrow cells per flask were added for the indicated periods of cultivation. In Experiment I, feeder cells were added at 2 h (0 days) and present for 10 days or were decanted and replaced by fresh culture medium after 1 to 7 days; the cultures were fixed at 10 days. In Experiment II, feeder cells were added at 2 h (0 days) and were present for 10 days, or they were added at 2 to 10 days and were present for a further 9 days; the cultures were fixed at 10-19 days.
FMC were added at 2 h after explantation of marrow cells and when they remained in cultures for 7 days, Experiment I (Table 2). The table also demonstrates, Experiment II, that after 3 and 7 days cultivation without feeder cells, from 50% to more than 90% of CFUf respectively lose the ability to be stimulated to colony formation by .FMC. In addition to BMC, spleen cells from mice and quinea pigs and platelets from guinea pigs and rabbits also had a colony-stimulating action: 2 x IO8platelets, moreover, had a stimulating action equal to that of 1.2 x 10’ bone marrow feeder cells (Fig. 7). Cell populations containing megakaryocytes (BMC and mice spleen cells) preserved their colony-stimulating activity after removal of their platelets by EDTA treatment (Fig. 8). Conve~ely, after removal of the platelets, guinea pig spleen cells, among which there were no megakaryocytes, had much reduced activity (Fig. 8). Thymocytes, lymph node cells and blood leukocytes did not stimulate the formation of CFUf colonies (Fig. 7). Addition of either PDGF, K-3, or EGF into AMCC did not lead to the formation of CFUf colonies (Table 3). In one instance where IL-3 and feeder cells were both present there was no added effect of IL-3 (Table 3).
208
0
m
I
S
t
b
P
80
60 40 20 0 0
s
m
I
t
m
2.5 X 1s” FCIflask
Q
lo7 platelets per flask
111
5.0 X 10’ FClflask
[13
3x107 platelets per flask
1.2 X 10’ FC/flask
B
2~10~ platelets per flask
@#
2.5 X 10’ FClflask
Fig+7. The effect of different populations of feeder cells (FC) on ECFF, expressed as percent of maximum value obtained. AMCC were established after explantation of 5 x 10 mechani~aliy disa~regated bone marrow ceils in T25 flasks in culture medium supplemented with 20% FCS. (A) Guinea pig FC, (B) mouse FC, FC were added to the cultures at 2 h. o, AMCC non-supplemented with feeder cells; m, marrow; &spleen: I, lymph node: t, thymus feeder cells; b, blood leucocytes; p, platelets. The cultures were fixed at 10-12 days. Means ( f SEM) are shown for triplicate platings.
209
80 80
B
c
D
Untreated feeder cells Feeder cells treated with EDTA Fig.8. The effect of EDTA treatment of feeder cells on ECFf, expressed as percent of maximum value obtained. AMCC were established after explantation of 6 x IO5 mechanically disaggregated bone marrow cells in T25 flasks with culture medium with 20% FCS. A, AMCC not supplemented with feeder cells; B, AMCC supplemented with IO’ guinea pig feeder marrow cells; C, AMCC supplemented with I.5 x IO’ guinea pig feeder spleen cells; D, AMCC supplemented with 0.8 x IO’ mouse feeder spleen cells. Feeder cells were added into cultures at 2 h. The FC were either treated or untreated with EDTA. Cultures were fixed at I2 days. Means (+ SEM) are shown for triplicate platings.
Discussion
Despite 90% of CFUf sticking to the surface of culture vessels within 1 h [32], CFUf colony formation practically failed in AMCC deprived of non-adherent haemopoietic cells and could be restored if irradiated marrow feeder cells were added into the cultures. Thus, CFUf have to be stimulated by haemopoietic cells to form fibroblast colonies, but so far, the requirement of CFUf for such stimulation has remained unrecognised in conventional marrow cell cultures where CFUf are accompanied by endogenous haemopoietic cells. That CFUf colonies are cell clones was determined previously by chromosome typing [22,28]. The stimulatory effect of marrow feeder cells is dose dependent in AMCC and their concentration can be adjusted independently of the number of explanted CFUf. Therefore in feeder-supplemented AMCC the efficiency of colony formation can exceed that in CMCC many-fold and remains constant independent of the explantation density of the bone marrow cells. There seem to be good reasons to suppose that CFUf are represented by single fibroblasts present at 24 h in AMCC. Indeed, the number of CFUf colonies in feeder-supplemented cultures corresponds to 85% of single fibroblasts in 24-h cultures, many of the colonies being topographically linked to single fibroblasts marked in living cultures 9 days earlier. Unless stimulated by feeder cells these
210 Table 3
The effect of PDGF, IL-3 and EGF on CFUf colony formation in AMCC No. of explanted bone marrow cells
Growth factors
3 x IO’ (35-mm dish)
-
No. of feeder cells 2x106 4x IO6 6x1@
10 ng PDGF/ml 50 ng PDGF/ml
3 x IO’ (T25 flask)
0.12 42,4S 43.48 z&61 0,o 0.0
0.40,1
2 x IO’ (T25 flask)
2 x 10’ (3Smm dish)
No. of CFUf colonies
20 ng 20 ag 3 ng 10 ng 30 ng 70 ng -
IL-3/ml IL-3/ml IL-3/ml IL-3/ml IL-3fml It3/ml
5 ng EGF/ml 20 ng EGF/ml 100 ng EGF/ml
44&O 60 090 o,o,o, 1 OS0 0.0 o,o 0.0 198 22,36 198 757 4*7
AMCC cultures were established by explantation of 5 x IO’ mechanically disaggregated bone marrow cells
in T25 flasks in culture medium supplemented with 20% FCS. Growth factors and guinea pig feeder marrow cells were added in the cultures at 2 h. The cultures were fixed at 10 days.
fibroblasts do not proliferate at all or else they accomplish 1 to 6 cell doublings to produce 2-49 cell foci of fibroblasts instead of forming CFUf colonies composed of hundreds or thousands of fibroblasts in feeder-supplemented cultures. The number of stromal colonies detected by the CFUf colony assay was dependent on the mode of cell di~ggregation in preparing marrow cell suspensions. In trypsindisaggregated marrow cell suspensions the CFUf concentration was much higher than in suspensions prepared by mechanical disaggregation. The differences largely depend on inactivation of a considerable number of CFUf by mechanical t~atment of the marrow [30]. Despite long survival times of single marrow fibroblasts in non-feeder-supplemented cultures their competence to form CFUf colonies gradually diminished and became completely lost after 7 days of cultivation in the absence of feeder cells. Maximal ECFf was reached when feeder cells were present in AMCC during the first 6 or 7 days. It has been also demonstrated that in CMCC the CFUf enter S-phase of their first mitotic period in vitro between 28 and 60 h after explantation [19]. Taken together these results seem to indicate that proliferation of freshly-isolated CFUf which were arrested in vivo of the Go period of the cell cycle, and of their early descendants, both required stimulation by growth factors produced by ~emopoietic cells.
211
The colony-stimulating activity of feeder cells is not species specific, is expressed by cell populations which include megakaryocytes and platelets and can probably be attributed to cells of the megakaryocyte lineage. Indeed, after removal of platelets, the guinea pig spleen cells which lack megakaryocytes lose the colony-stimulating activity, while mice spleen cells which contain both platelets and megakaryocytes retain their stimulating activity. In a previous study [34] it was shown that bone marrow feeder cells stimulated CFUf colony formation when separated from adherent marrow cells by HA millipore filter and by 2 mm of medium, which indicates that they release a diffusible colony-stimulating factor (factors) into the culture medium, Maximal stimulation of colony formation by feeder cells was reached at high serum concentration in the culture medium and feeder cells did not substitute for serum or vice versa. Thus, the stimulatory activity of feeder cells is probably not due to the fibroblast growth factors present in serum of which PDGF is an obvious representative. For the present, the requirements of stromal clonogenic cells for growth factors has been studied only in conventional CMCC, i.e. when CFUf were accompanied by non-adherent haemopoietic cells which may act as feeder cells. It has been demonstrated that CFUf colony formation requires PDGF [32] and in bone marrow cultures with serum-deprived culture medium or with culture medium supplemented with platelet-poor plasma, PDGF had to be added for CFUf colony formation to occur [27], maximal stimulation being achieved at 10 ng/ ml and 30 ng/ml PDGF in mice and human CMCC respectively. Here we have tested in AMCC with serum-rich culture medium whether PDGF, IL-3 or EGF could replace the colony-stimulating activity of feeder cells. In the circumstances, none of these growth factors induced CFUf colonies in non-feeder-supplemented AMCC and the growth factors released by feeder cells which are required for stimulating osteogenic stromal colony formation remain to be identified. Most of the CFUf colonies in mouse feeder-supplemented AMCC comprise hundreds or thousands of fibroblasts. Single-colony-derived strains of rabbit bone marrow fibroblasts undergo 17-22 population doublings and produce more than 10” descendents from the initial CFUf [22]. These observations confirm that CFUf colonies originate from highly proliferative clonogenic stroma1 cells [22]. The colonies are heterogeneous with regard to their potential for osteogenesis and in vitro differentiation [23]. It has been assumed that this diversity depends upon differences in the unequal potential of the cells which initiate CFUf colonies and that these cells represent stromal precursors at different levels in a putative hierarchy of the bone marrow osteogenic cell lineage [22,26]. Recently Owen and collaborators [29] have demonstrated that cell differentiation in CFUf colonies can be reversed and that this depends Upon serum growth factors. It remains to be determined whether colony-stimulating factors produced by haemopoietic cells influence the differentiation pattern of bone marrow fibroblasts and may thus promote diversity of CFUf colonies.
212 References 1 F~edenstejn AJ, Petra~ova KV, Kurafesova AJ, Frolova GP. Heterotopic transplants of bone marrow. Aua\y& of precursor cells for osteogenic and bemopoietic tissues. Transplantation 1968;6:230-247. Friedenstein AJ. Precursor cells of mechanocytes. Int Rev Cytol 1976;47:327-359. Feedenstein AJ, Ivanov-Smolenski AA, Chajlakyan RK. Origin of bone marrow stromal mmhanocytes in radiochimeras and heterotopic transplants. Exp Hematol 1978;6:440-444. F~~~stein AJ, Ch~lakyan RK, Latzinik NV, Panasyk AF, Keilis-Borok JV. Stromai cells responsible for transferor the mi~r~nvironment of hemopoietic tissues. Cloning in vitro and retransplantation in vivo. Transplantation 1974;17:331-340. 5 Friedenstein AI, Latzinik NV, Groshera AG, Gorskaya UF, Marrow microenvironment transfer by heterotopic transplantation of freshly isolated and cultured cells in porous sponges. Exp Hema,tol 1982;10:217-227. 6 Patt HM, Maloney MA, Flannery ML, Hematopoictic microenvironment transfer by stromai ~brobtasts derived from bone marrow va~ng in ~~~u~a~ty.Exp Hematol 1982~1~728-732. 7 Dexter TM, Alien TD, Lajtha LG. Conditions controlling the proliferation of hemopoietic stem ceils in vitro. Cell Physiol 1977;91:335-349, 8 Friedenstein AJ, Chailakyan RK, Lalykina KS. The development of iibroblast colonies in monolayer cultures of guinea-pig bone marrow and spleen cells. Cell Tissue Kinet 1970;3:393-403. 9 Ashton BA, Allen TD, Howlett CR, Eagleson CC, Hattori A, Owen M. Formation of bone and cartilage by marrow stromal cells in di~usion chambers in vivo. Clin Grthop 198O;lS~29~306. 10 Friedenstein AJ. Stromal mechanocytes of bone marrow: cloning in vitro and retransplantation in vivo. In: Thierfelder S, Rodt H, Koilb HJ, eds. Immunoiogy of bone marrow transplantation. Berlin: Springer-Verlag, 1980;t9-29. 11 Friedenstein AJ, Kuralesova Al. Osteogenic precursor cells of bone marrow in radiation chimeras. Transplantation 1971;12:99-108. 12 Golde DW, Hacking WG, Quan SG, Sparkes RS, Gaie RP. Origin of human bone marrow Bb~blasts. Br 3 Haematol ~980;~~83-187. 13 Witson FD, Greenberg GR, Konrad PN, Klein AD, Walling PA. Cytogenetic studies on bone marrow fibrobtasts from male-female haemopoietic chimera: evidence that stromal elements in human transplantation recipients are host type. Transplantation 1978;25:87-92. 14 Lennon JE, Micklem HS. Stromal cells in long-term murine bone marrow culture: FACS studies and origin of stromal cellsin radiation chimeras. Exp Hematol 1986;14:287-292, IS Simmons PJ, Przepilzka 4 Thomas ED, Torok-Storb B. Hostorigin of marrow stromal cells foliowing allo~neie bone marrow t~nsplantation. Nature 1987;328:429~31. t6 F~~enstein AJ, Osteogenic stem ~4s in the bone marrow. Bone Miner Res 1990;7:24.3-272. I7 Epichina SY, Latzinik NV. Proliferative activity of bone marrow stromal clonogenic cells, Bull Exp Biol Med 1976;81:55-58. 18 Kaneko S, Motomura, Ibayashi H. Differentiation of human bone marrow-derived fibroblastoid colony forming ceils (CFUf). Br J Haematol 1982;51:217-305. 19 Keiliss-Borok IV, Latxinik NV, Epichina SY, F~~enstein AJ. Dynamics of the fo~ation of fibroblast Wtonies in monolayer cultures of bone marrow, according to 3H thymidine in~~oratiou experiments. Cytologia 1971;13:1402-141I. 20 Castro-Malaspina H, Gay RE, Resnick J, Kapoor N, Meyers P, Uriariezi D, McKenzie S, Broxmeyer HE, Moore MAS. Characterisacion of human bone marrow fibroblast colony-forming cells (CFU-F) and their progeny. Blood 1980;56:289-301. 21 Simmons PJ, Torok-Storb B. Identification of stromaf cell precursors in human bone marrow by a novel mon~lonal anti~dy, STRO-I, Blood 1991;78:55-62. 22 F&den&n AJ, Chailakhyan RK, Gerasimon HV, Bone marrow osteogenic stem cells: in vitro cukivation and transplantation in diffusion chambers, Cell Tissue Kinet 1987;20:263-272. 23 Barling PM, Bennett JH, TritTitt JT, Gwen ME. The adenylate cyclase response to parathyroid hormone in Cultured rabbit marrow fibroblastic cells. Bone Miner 1989;7:23-30.
213 24 Rodan G, Heath JK, Yoon K, Noda M, Rodan SB. Diversion of osteoblast phenoty~s. C&a Foundation Symposium 1988;136:111-l 70. 25 Owen M. Marrow stromal stem cells. In: Stem Cell Symposium. J Cell Sci. 1988; Suppl.l0:63-76. 26 Owen M, Friedenstein AJ. Stromal stem cells: marrow derived osteogenic precursors. Ciba Foundation Symposium 1988;136~42-60. 27 McIntyre AP, Bjornson BM. Human bone marrow stromal cell colonies: response to hydro~~isone and dependent on platelet-derived growth factor. Exp Hematol 1986;17:833-839. 28 Latzinik NV, Gorskaya UF, Grosheva AG, Friedenstein AJ. The content of stromal colony-forming cells (CFUf) in the mouse bone marrow and the clonal nature of CFUf-derived stromal fibroblast colonies. ~togenesis 1986;17:27-35. 29 Bennett JH, Joyner CJ, Triffitt JT, Owen ME. Adipocytic cells cultured from marrow have osteogenic potential. J Cell Sci 1991;99:131-139. 30 Latzinik NV, Sidorovich SU, Friedenstein AJ. Effect of bone marrow trypsinisation on the efficiency of fibroblast colony formation in monolayer cultures. Bull Exp Biol Med 1981;9:356-360. 31 Fri~enstein AJ, Gorskaja UF, Kulagina NN. Fibroblast precursors in normal and irradiated mouse hemopoietic organs, Exp Hematol 1976;4:267-274. 32 Friedcnstein AJ. Determined and inducible osteogenic precursor cells. Ciba Foundation Symposium 1971;[email protected]. 33 Hirata J, Kaneko S, Nishimura JN, Mot~ura S, Ibayashi H. Effeet of platelet-delve growth factor and bone marrow-conditioned medium on the proliferation of human bone mar;row-derived libroblastoid colony-forming cells. Acta Hematol 1985;74:189-194. 34 Friedenstein AJ, Luria EA, Latzinik NV, Gorskaya YF. Humoral nature of the CSA of bone marrow cells on stromal colony formation in bone marrow cultures. Bull Exp Biol Med 199O;i10509-510.
