Developmentally regulated responsiveness to transforming growth ...

Leukemia (1999) 13, 1266–1272  1999 Stockton Press All rights reserved 0887-6924/99 $12.00 http://www.stockton-press.co.uk/leu

Developmentally regulated responsiveness to transforming growth factor-␤ is correlated with functional differences between human adult and fetal primitive hematopoietic progenitor cells SFA Weekx1, J Plum2, P Van Cauwelaert3, M Lenjou1, G Nijs1, M De Smedt2, M Vanhove3, P Muylaert3, DR Van Bockstaele1, ZN Berneman1 and H-W Snoeck1,4 1

Laboratory of Experimental Hematology, University of Antwerp, Belgium; 2Dept of Clinical Chemistry, Microbiology and Immunology, University of Ghent, Belgium; 3Dept of Cardiac Surgery, Middelheim General Hospital, Antwerp, Belgium; and 4Institute for Gene Therapy and Molecular Medicine, Mount Sinai School of Medicine, New York, NY, USA

Important functional differences exist between primitive CD34++CD38− hematopoietic progenitor cells derived from human fetal liver (FL) and adult bone marrow (ABM). FL progenitors are known to have higher proliferative capacities and lower cytokine requirements than their ABM counterparts. In this study, we isolated FL and ABM CD34++CD38− cells and used a two-stage culture system to investigate the effects of transforming growth factor-␤ (TGF-␤) and blocking anti-TGF-␤ antibodies (anti-TGF-␤) on these cells. First, we demonstrate that FL progenitors are significantly less sensitive to the inhibitory effects of TGF-␤ than ABM cells. Second, whereas ABM cells are significantly stimulated by anti-TGF-␤, only very limited effects are seen on FL cells. Third, we show that the effect of anti-TGF-␤ is mainly situated at the level of the initial cell cycles of very primitive progenitor cells with a high proliferation potential. Fourth, we demonstrate that blocking the effects of endogenous TGF-␤ reduces the growth factor requirements of ABM cells in order to proliferate and differentiate. Based on these data, we hypothesize that at least part of the functional differences that exist between adult and fetal stem cells can be accounted for by a developmental different responsiveness to TGF-␤. Keywords: human fetal liver; human adult bone marrow; hematopoiesis; CD34++CD38− cells; TGF-␤

Introduction Human hematopoiesis occurs at three major sites during embryonic and fetal development. It starts in the third week of gestation in the blood islands of the yolk sac, and then migrates during the sixth week to the FL and spleen. Subsequently, at week 20, hematopoiesis homes to the bone marrow which finally takes over the whole blood cell production.1–4 Recently, Tavian et al5 identified a CD34expressing hematopoietic cell compartment associated with the ventral endothelium of the aorta in 5-week-old human embryos. Moreover, additional sites of hematopoiesis (eg dorsal aorta, cardinal veins and proximal umbilical and vitelline arteries) were found in day 8 to day 10 mouse embryos,6 while primitive hematopoietic stem cells also were found in adult mouse liver.7,8 At least part of human hematopoietic stem cells have been shown by in vivo experiments (myelopoietic and lymphopoietic capacity) and in vitro experiments (engraftment of long-term sustainable human hematopoiesis in fetal sheep and in immunodeficient mice) to be contained within the CD34++CD38− fraction.9–12 However, when compared to ABM CD34++CD38− cells, phenotypically identical cells Correspondence: ZN Berneman, Laboratory of Experimental Hematology, Antwerp University Hospital (UZA), Wilrijkstraat 10, B-2650 Edegem, Belgium; Fax: 32 3 825 11 48 Received 28 January 1999; accepted 15 April 1999

derived from ontogenically earlier sources such as cord blood (CB) and FL show important intrinsic functional differences. Indeed, the maximal proliferative capacity of CD34++CD38− cells is inversely correlated with their ontogenic age.13,14 Moreover, we have recently shown that the growth factor requirements of CD34++CD38− cells derived from FL and CB are much less stringent than those for their ABM counterparts.15 In order to better understand these differences between FLand ABM-derived CD34++CD38− cells, we investigated whether factors, known to modulate the growth of hematopoietic progenitors, have different effects on ABM as opposed to FL cells. It has been shown previously by Hatzfeld et al16 that primitive ABM multipotent progenitor cells produce TGF␤ in an endogenous fashion, keeping these cells in a quiescent state. Therefore, we wanted to compare the effects of exogenously added TGF-␤ on the cell proliferation and CFU generation of primitive FL and ABM cells. We furthermore used anti-TGF-␤ antibodies to study the effects of blocking endogenously produced TGF-␤ on these cells populations. Third, we investigated whether these blocking anti-TGF-␤ antibodies might reduce the growth factor requirements of the ABM cell population in order to proliferate and generate secondary colonies. The results from these experiments support the model of the group of Hatzfeld et al16–18 that TGF-␤ plays a central role in controlling the quiescence of hematopoietic primitive cells, and led us to extend their results with the fact that the TGF␤ system, and more specifically the developmental different responsiveness to TGF-␤, is an important determinant of the intrinsic functional characteristics of human hematopoietic stem cells derived from ontogenically different sources. Materials and methods

Fetal liver and adult bone marrow cell purification and culture Human FL (18–22 weeks gestation) and ABM cells (sternal puncture from hematologically normal patients undergoing cardiac surgery) were obtained, stored, purified and cultured as described before.15 In brief, cells were labeled with 43A1 supernatant as a source of anti-CD34 antibodies (IgG3, kindly donated by Dr HJ Bu¨hring, University of Tu¨bingen, Germany),19 stained with fluorescein isothiocyanate-conjugated rabbit anti-mouse immunoglobulins (RAM-FITC) and with phycoerythrin (PE)-conjugated anti-CD38. Cells were sorted on a FACStarPlus Cells Sorter (Becton Dickinson, Erembodegem, Belgium). Cells with a low to medium forward scatter and a low side scatter, a highly positive green (CD34) flu-

Human ABM and FL stem cell responsiveness to TGF-␤ SFA Weekx et al

orescence, and an orange (CD38) fluorescence signal lower than the mean fluorescence of cells labeled with an irrelevant isotype-matched control antibody +2 standard deviations were retained as CD34++CD38− cells. Purities were always ⬎95%. Sorting regions are depicted in Figure 1. Cells (200 per well) were cultured in pre-colony-forming unit (pre-CFU) assays. In this assay, liquid cultures were performed in duplicate in 96-well flat-bottomed plates in IMDM (GIBCO, Paisley, UK) containing 10% fetal calf serum (FCS), 1% bovine serum albumin (BSA) and different combinations of recombinant human (rhu) interleukin (IL)-1, IL-6, IL-3, rhu stem cell factor (SCF) and varying concentrations of TGF-␤ and antiTGF-␤. ABM and FL cells were cultured at 37°C in 7.5% O2 and 5% CO2 in a fully humidified incubator for 7 and 14 days, after which the cells were counted, harvested and plated in duplicate in secondary methylcellulose cultures (0.9%) supplemented with 20% FCS, 1% BSA, 10% conditioned medium of the 5637 bladder carcinoma cell line (containing G-CSF and GM-CSF), IL-6, IL-3, erythropoietin (epo) and 2-mercaptoethanol. These cultures were microscopically scored for colony formation after 14 days of culture at 37°C in 7.5% O2 and 5% CO2 in a fully humidified incubator. All samples were obtained after informed consent according to the guidelines of the Medical Ethics Committees of the University Hospitals of Ghent and Antwerp.

Cytokines and monoclonal antibodies RAM-FITC [F(ab⬘)2 fragments] was purchased from Dako (Glostrup, Denmark). CD38-PE as well as isotype-matched control antibodies were purchased from Becton Dickinson (Erembodegem, Belgium), mouse gammaglobulins from Jackson Immuno Research Laboratories (West Baltimore Pike, PA, USA). IL-1 (specific activity (sa) ⬎5 × 107 U/mg), rhuIL-6 (sa ⬎1 × 108 U/mg), and SCF (sa ⬎1 × 105 U/mg) were obtained from Boehringer Mannheim (Penzberg, Germany), epo (sa 苲1 × 105 U/mg) from Cilag (Brussels, Belgium). rhuIL-3 (biological activity 14 × 103 U/ml) was a kind gift of Dr SC Clark (Genetics Institute, Cambridge, MA, USA). Ultrapure natural human TGF-␤ (sa 苲1 × 106 U/mg), monoclonal mouse anti-TGF-␤1, ␤2, ␤3 and monoclonal mouse anti-human IL-2 (irrelevant control antibody) were purchased from Genzyme (Cambridge, MA, USA).

TGF-␤ measurements TGF-␤ levels in (1) culture medium containing FCS and BSA and in (2) the supernatant of 3000–5000 FL and ABM CD34++ CD38− cells, cultured during 36–48 h in serum-free medium (X-VIVO 15; Biowhittaker, Walkersville, USA), containing 100 ng/ml IL-1, 200 U/ml IL-6, 30 U/ml IL-3 and 100 ng/ml SCF were measured using a sensitive (⬎25 pg/ml) immunoenzymatic method (TGF-␤1 Emax; Promega, Madison, WI, USA), performed according to the instructions of the manufacturer.

Statistics Intra-FL or intra-ABM comparisons were validated using the Student’s t-test for paired samples. All other comparisons were based on the unpaired t-test. Results are expressed as mean ± standard error of the mean (s.e.m.).

Results We investigated the effects of TGF-␤ and blocking anti-TGF␤ antibodies in a pre-CFU assay. In this assay, CD34++ CD38− cells from FL and ABM are cultured in a primary liquid culture, containing an optimal stimulatory combination of early acting cytokines (IL-1 + IL-3 + IL-6 + SCF), after which the cells are counted and plated in semisolid cultures. In this way, the generation of progenitor cells from their precursors, undetectable in direct semisolid assays, can be quantitated so that the effects of TGF-␤ and anti-TGF-␤ on the initial phase of the proliferation and differentiation of very primitive progenitor cells are assessed.10,15 In a first set of experiments, we evaluated the maximal cell proliferation and CFU generation at two time points: day 7 and day 14. As shown in Figure 2a and b we found that both at day 7 and day 14 CD34++CD38− FL cells had a significantly higher proliferative capacity and secondary CFU output than their ABM counterparts. Both ABM and FL progenitors showed a multilineage content as was evidenced by the output of myeloid (macrophage, granulocyte and granulocyte– macrophage), erythroid and mixed erythroid-myeloid colonies, although no significant differences in colony type or col-

Figure 1 Sorting criteria for CD34++CD38− cells. (a) Side (SSC) vs forward (FSC) scatter for ABM mononuclear cells. The rectangular region (R1) defines the gated population used for sorting. (B) Fluorescent intensities of CD34-FITC and CD38-PE for all cells within the gated region R1. R2 defines the gated population used for sorting the CD34++CD38− cells. Dot plots represent the sorting criteria for a single ABM experiment, but similar plots were obtained for the other ABM and FL experiments.

1267


1268

at day 14 (Figure 3a and b). This was in strong contrast with the impact of anti-TGF-␤ on FL CD34++ CD38− cells, where a far less pronounced effect was noted at both time-points: 1.9 ± 0.3-fold (P ⬍ 0.005) cell number expansion and 2.5 ± 0.7-fold (P = 0.03) CFU expansion at day 7, and 1.4 ± 0.1fold (P = 0.047) cell number expansion and 0.7 ± 0.2-fold (P = NS) CFU expansion, respectively, at day 14 (Figure 3a and b). This means that when FL and ABM progenitors are compared, the difference in maximal proliferative capacity between these cell fractions is significantly reduced from 174.0-fold (no anti-TGF-␤) to 74.5-fold (+ anti-TGF-␤) at day 7 (reduction factor: 2.3), and from 182.9-fold to 15.5-fold at day 14 (reduction factor: 11.9). This is paralleled by an equally strong reduction in the difference in secondary CFU generation between FL and ABM cells: from 268.8-fold to 109.7fold at day 7 (reduction factor: 2.5), and from 47.8-fold to 4.3fold at day 14 (reduction factor: 11.1). Thus, at the end of the 14-day culture period, the cell proliferation and especially the production of secondary CFU of FL and ABM cells approached each other when anti-TGF-␤ was present in the cultures. In ABM, anti-TGF-␤ induced the highest increase in cell number and secondary CFU generation during the second week of culture (Figure 3a and b). This may indicate either that anti-TGF-␤ especially induces cycling of very primitive progenitors with a higher proliferation potential, and therefore a large progeny, or that it supports especially the proliferation at the level of more mature progenitors. We therefore added anti-TGF-␤ during the first 36 h of the primary liquid culture,

Figure 2 Proliferation capacity (a) and secondary CFU generation (b) of CD34++ CD38− cells derived from ABM (——) and FL (– – –) in the absence (왖) and presence (쎲) of anti-TGF-␤. Results (mean ± s.e.m.) are expressed as (a) cell number/100 input cells and (b) as CFU output/100 input cells. (n = x,y): number of independent individual experiments at day 7 and at day 14.

ony size were noted between both cell fractions, which was in agreement with our previous results.15 These experiments also demonstrated that the highest rate of proliferation and generation of secondary CFU of FL cells was situated in the first 7 days of the culture period, whereas ABM cells needed another 7 days to reach their highest proliferation rate. In a next set of experiments we investigated the effects of neutralizing anti-TGF-␤ antibodies on FL and ABM CD34++CD38− cells. It has been shown previously by Hatzfeld et al16 that primitive ABM multipotent progenitor cells produce TGF-␤ in an endogenous fashion, keeping these cells in a quiescent state. Addition of anti-TGF-␤, but not of an irrelevant control antibody (anti-IL-2, not shown), to IL-1 + IL3 + IL-6 + SCF-stimulated cells resulted in a very potent stimulation of the proliferation and of the secondary CFU output of ABM cells, especially at day 14: 3.3 ± 0.2-fold (P ⬍ 0.001) cell number expansion and 5.2 ± 1.8-fold (P = 0.03) CFU expansion at day 7, and 22.5 ± 4.5-fold (P ⬍ 0.001) cell number expansion and 21.6 ± 4.5-fold (P = 0.002) CFU expansion

Figure 3 Effects of anti-TGF-␤ on the primary liquid culture total cell number (a) and subsequent secondary CFU output (b) of CD34++ − CD38 cells derived from ABM and FL after 7 days (쐽) and 14 days (쏔) liquid culture supported by IL-1 + IL-3 + IL-6 + SCF. Results (mean ± s.e.m.) are expressed as % vs control without anti-TGF-␤. (n = number of independent individual experiments).


1269

after which the cells were washed and cultured for another 12 days without anti-TGF-␤, and compared this with a 14 day culture period in the presence of anti-TGF-␤. As shown in Figure 4, no significant difference in cell number was seen between both conditions, indicating that anti-TGF-␤ is only essential during the first 36 h of the primary culture, confirming the results of Sansilvestri et al20 and of Hatzfeld et al.21 These data clearly show that the effect of anti-TGF-␤ is mainly situated at the level of the initial cell cycles of very primitive progenitor cells with a high proliferation potential. Addition of exogenous TGF-␤ to the pre-CFU cultures resulted in a strong inhibition of cell proliferation and of secondary CFU generation from both ABM and FL cells (Figure 5a and b). However, at the highest TGF-␤ concentration used, FL cells were significantly less sensitive to the overall inhibitory effects of TGF-␤ than ABM cells (day 7: −53.8 ± 16.7% vs −77.4 ± 4.5% (P = not significant (NS) for FL and ABM cell proliferation, respectively, and −52.4 ± 10.5% vs −82.7 ± 6.3% (P = 0.026) for FL and ABM CFU generation, respectively; day 14: −70.9 ± 8.0% vs −94.0 ± 0.9% (P ⬍ 0.001) for FL and ABM cell proliferation, respectively, and −53.3 ± 17.1% vs −82.3 ± 4.9% (P = 0.04) for FL and ABM CFU generation, respectively). The fact that FCS and BSA used in our cultures contained on average 0.13 ng/ml TGF-␤ (final concentration in the culture medium), a concentration at which exogenously added TGF␤ has little or no effect (Figure 5), indicates that these effects of anti-TGF-␤ are mainly mediated through the blockade of autocrinely and/or paracrinely produced TGF-␤, but not of TGF-␤ present in the culture medium. We furthermore measured TGF-␤ production by 3000–5000 ABM or FL CD34++ CD38− cells after 24–48 h culture in serum-free medium. Using a highly sensitive ELISA technique, the levels of TGF-␤ were near the detection limit (not shown), but no difference in TGF-␤ production level between FL and ABM could be demonstrated. As we have shown previously, an important functional difference between FL and ABM cells are the less stringent growth factor requirements of FL CD34++ CD38− cells when compared to their ABM counterparts.15 Indeed, whereas optimal cell proliferation of ABM cells requires the presence of a

Figure 5 Dose-response curve of the effects of TGF-␤ on the primary liquid culture total cell number (a) and subsequent secondary CFU output (b) of CD34++CD38− cells derived from ABM (——) and FL (– – –) after 7 days (왖) and 14 days (쎲) liquid culture, supported by IL-1 + IL-3 + IL-6 + SCF. Results (mean ± s.e.m.) are expressed as % vs control without TGF-␤. n = number of independent individual experiments.

Figure 4 Primary liquid culture total cell number of CD34++CD38− cells, derived from ABM, after 14 days liquid culture supported by IL1 + IL-3 + IL-6 + SCF in the presence of anti-TGF-␤ during 36 h and 14 days. Results (mean ± s.e.m.) are expressed as % vs control without anti-TGF-␤. n = number of independent individual experiments.

4-cytokine cocktail (IL-1 + IL-3 + IL-6 + SCF), the combination of IL-3 + SCF is already sufficient to obtain maximal cell proliferation of FL cells.10,15 We hypothesized that blocking the effects of endogenous TGF-␤ by adding anti-TGF-␤ antibodies to the stimulatory cytokine cocktail might reduce the cytokine requirements of the ABM cell population. To test this hypothesis, we limited the combination of early acting cytokines in the pre-CFU assay from IL-1 + IL-3 + IL-6 + SCF to IL3 + SCF ± anti-TGF-␤ (Table 1). Interestingly, already at day 7 the use of the 2-cytokine cocktail reduced the ABM cell proliferation with 50.9 ± 6.0% (P ⬍ 0.001) vs control (ie the 4-cytokine cocktail). This reduction was completely abrogated after the addition of anti-TGF-␤ to the two cytokines, which even resulted in a significant stimulation of cell proliferation of 160.9 ± 11.9% (P ⬍ 0.001) vs the 4-cytokine control, although it did not reach the cell proliferation seen with the combination of the 4-cytokine cocktail + anti-TGF-␤. At the level of CFU generation, anti-TGF-␤ again abrogated the reduction noted after incubation with two cytokines (from −86.8 ± 6.1% (P ⬍ 0.001) to 102.5 ± 30.0% vs control (P = NS)). However, as could be anticipated for FL CD34++ CD38− cells, reducing the 4-cytokine cocktail to the combination of


1270

Table 1 Cell proliferation after 7 days of liquid culture containing IL-3/SCF or IL-3/SCF/anti-TGF-␤ and subsequent secondary CFU output of CD34++ CD38− cells from ABM and FL

ABM

FL

IL-3/SCF

Cell No. CFU

49.1 ± 6.0a 13.2 ± 6.1a

103.6 ± 33.9 124.4 ± 43.5

IL-3/SCF/ anti-TGF-␤

Cell No. CFU

160.9 ± 11.9a 102.5 ± 30.0

79.6 ± 53.9 84.7 ± 59.9

Results are expressed as mean percentage ± s.e.m. of control (ie the 4-cytokine cocktail containing IL-1/IL-3/IL-6/SCF). a Significantly different vs control (P ⬍ 0.05). CFU, secondary CFU output. ABM: n = 9; FL: n = 2.

IL-3 + SCF did not result in a significant alteration of cell proliferation (103.6 ± 33.9% (P = NS) vs control), thus confirming our previous results.15 Addition of anti-TGF-␤ to IL-3 + SCF resulted in a minor, albeit not significant, increase in FL cell proliferation (124.4 ± 43.5% (P = NS) vs control). Again for the secondary CFU output, no significant changes were seen when reducing the cytokine cocktail, with or without the addition of anti-TGF-␤. Discussion In this report we show for the first time that the TGF-␤ system is an important determinant of the intrinsic functional characteristics of primitive hematopoietic progenitor cells derived from ontogenically different sources. It is important to note that the developmentally regulated responsiveness to TGF-␤ is more dramatic than it appears from experiments where the sensitivity of FL and ABM cells to the effects of exogenously added TGF-␤ are compared. Since exogenous TGF-␤ reduces the proliferation of ABM 17-fold (−94.0%) over baseline (Figure 5), and since neutralization of endogenously produced TGF-␤ increases proliferation 22.5fold over baseline (Figure 3), TGF-␤ actually reduces the maximal proliferation of ABM CD34++CD38− cells by a near 400fold (17 × 22.5). In FL on the other hand, anti-TGF-␤ results in a 1.4-fold increase (Figure 3), and exogenously added TGF␤ in a 3.4-fold (−70.9%) decrease in proliferation (Figure 5), so that the maximal effect of TGF-␤ is at best a five-fold (1.4 × 3.4) reduction in proliferative capacity. This implies that the sensitivity of CD34++CD38− cells from ABM to TGF-␤ is at least 80- to 100-fold higher compared to their FL counterparts, indicating a developmentally regulated responsiveness to TGF-␤. It should be kept in mind that the differences in TGF␤ responsiveness between FL and ABM can only be defined as being a global effect, since clonal heterogeneity in the purified CD34++CD38− cells was not excluded. Therefore clonal analysis of the TGF-␤ effect is currently under investigation. This increase in the response to TGF-␤ with ontogenic age may possibly explain two major functional differences that exist between FL and ABM multipotent progenitor cells, ie the lower proliferative capacity and the higher growth factor requirements of adult cells when compared to fetal liver cells. First, we show in our study that blocking the TGF-␤ effects by using monoclonal anti-TGF-␤ antibodies results in a spectacular reduction of the difference in maximal proliferative capacity and secondary CFU generation between ABM and FL cells, indicating that the 80- to 100-fold difference in TGF-

␤ sensitivity between ABM and FL CD34++CD38− cells might account for the major part of the more than 100-fold difference in maximal proliferative capacity between FL and ABM CD34++CD38− cells reported here and previously.14,15 Second, our study demonstrates that the elimination of TGF-␤ activity on primitive ABM cells results in a reduced growth factor requirement, comparable to the requirements of FL cells. It has previously been shown that TGF-␤ is an important regulator of the responsiveness of human hematopoietic stem/progenitor cells to several cytokines such as SCF, IL-3, IL-1, and Flt3 ligand. The inhibitory action of TGF-␤ on cell growth is probably associated with the fact that TGF-␤ has direct effects on cell surface expression of some cytokine receptors, which is already been demonstrated by the down-modulation of the receptors for IL-1, IL-3, M-CSF and SCF on hematopoietic cells, depriving them of the ability to respond to these cytokines.22–28 Sansilvestri et al20 showed that two subpopulations of CD34high cells can be distinguished: one subpopulation that expresses c-kit that can be downmodulated by exogenous TGF-␤, and another subpopulation that expresses a low level of c-kit that can be upregulated by anti-TGF-␤. Interestingly, these CD34+Kitlow cells have been reported to contain significant levels of long-term marrow-engrafting hematopoietic stem cells.29 Recently, Fortunel et al18 demonstrated that antiTGF-␤ also allows the rapid expression of both the Flt3 and the Il-6 receptors in quiescent cells which can then respond to these cytokines present in the culture medium. TGF-␤ is a ubiquitously present cytokine produced by progenitor cells, stromal cells, and many other cells and tissues, making it reasonable to accept that the functional differences between FL and ABM progenitors are regulated at the level of response to TGF-␤ rather than at the level of production of TGF-␤ be it acting autocrinely, paracrinely or even endocrinely. However, since we could not demonstrate a different TGF-␤ production level between both cell fractions, we are currently investigating whether this difference in responsiveness to TGF-␤ is regulated at the level of receptor modulation, or rather at the level of the downstream signalling pathways. What might be the biological significance of our findings? As shown by Vaziri et al,30 and recently confirmed by Engelhardt et al,31 adult stem cells have significantly shortened telomeres when compared with FL cells. The shortening of telomeres with age in human stem cells indicates that stem cell populations probably cycle throughout life time, and that hematopoietic stem cells have a limited and defined replicative life span.32 It can be anticipated that the highest level of stem cell expansion occurs during fetal and postnatal growth of the individual since it has been demonstrated that the cycling activity of stem cells in fetal and young mice is higher than in old mice, whereas the size of the stem cell pool increases with age.33–36 Based on our data in the human system, we speculate that the induction of an increased sensitivity to TGF-␤ could possibly be one of the molecular changes associated with replicative aging of stem cells. This implies that TGF-␤ responsiveness might be related with the cycling status and the pool size of stem cells. According to this hypothesis, the relative resistance to the inhibitory effects of TGF-␤ would allow the stem cell compartment to expand in young individuals, whereas in older individuals, the expanded and now quiescent stem cell pool would decrease its rate of cycling due to an increased responsiveness to TGF␤. This would be in accordance with the working model of Fortunel et al18 who introduced the ‘high proliferative potential-quiescent cells’, being primitive progenitor cells that are highly sensitive to the growth inhibitory effect of TGF-␤. It has


furthermore been shown that FL stem cells in mice have a higher reconstituting potential than ABM cells, and that this reconstituting capacity of stem cells decreases upon serial transplantation.37,38 Therefore the responsiveness to TGF-␤ may determine the ‘quality’ of stem cells in terms of reconstituting capacity and developmental potential, indicating that an increased responsiveness to TGF-␤ after extensive renewal, and thus replicative senescence of stem cells, may explain the decreasing reconstituting capacity of stem cells upon serial transplantation.38,39 Telomeres are thought to function as a mitotic clock, and telomere shortening is thought to induce the phenotypic changes associated with cellular senescence.32 Therefore, our data suggest that the induction of a functional inhibitory TGF␤ loop, associated with increasing replicative age of stem cells, is paralleled by telomere shortening. The increased sensitivity of ABM CD34++CD38− cells to endogenously produced TGF-␤ limits unnecessary further expansion of the stem cell compartment. This may serve as a protective mechanism against recruitment and subsequent critical telomere shortening, and therefore against genomic instability and transformation. Indeed, the TGF-␤ system is clearly emerging as an important tumor suppressor mechanism in several organ systems, and very recently, TGF-␤1 has been shown to be a new form of tumor suppressor in the liver and lung.40–42 Taken together, our data clearly show for the first time that at least part of the physiological differences between human adult and fetal primitive hematopoietic progenitor cells can be explained by a developmentally regulated increase in the responsiveness to TGF-␤ from FL to ABM stem cells.

Acknowledgements This work was supported by grant No. G.0096.95 of the Fund for Scientific Research, Flanders (Belgium) (FWO – Vlaanderen), and by a gift from the ‘Nationale Vereniging tot Steun aan Gehandicapte Personen’ and the Rotary Clubs of Belgium to the HEBA Foundation. SFA Weekx is a research assistant with the Fund for Scientific Research, Flanders (Belgium) (FWO – Vlaanderen).

References 1 Tavassoli M. Embryonic and fetal hematopoiesis: an overview. Blood Cells 1991; 1: 269–281. 2 Huang H, Auerbach R. Identification and characterization of hematopoietic stem cells from the yolk sac of the early mouse embryo. Proc Natl Acad Sci USA 1993; 90: 10110–10114. 3 Migliaccio G, Migliaccio AR, Petti S, Mavilio F, Russo G, Lazzaro D, Testa U, Marinucci M, Peschle C. Human embryonic hemopoiesis. Kinetics of progenitors and precursors underlaying the yolk sac → liver transition. J Clin Invest 1986; 78: 51–60. 4 Godin I, Dieterlen-Lie`vre F, Cumano A. Emergence of multipotent hemopoietic cells in the yolk sac and paraaortic splanchnopleura in mouse embryos, beginning at 8.5 days postcoitus. Proc Natl Acad Sci USA 1995; 92: 773–777. 5 Tavian M, Coulombel L, Luton D, San Clemente H, DieterlenLie`vre F, Pe´ault B. Aorta-associated CD34+ hematopoietic cells in the early human embryo. Blood 1996; 87: 67–72. 6 Wood HB, May G, Healy L, Enver T, Morris-Kay GM. CD34 expression patterns during early mouse development are related to modes of blood vessel formation and reveal additional sites of hematopoiesis. Blood 1997; 90: 2300–2311. 7 Taniguchi H, Toyoshima T, Fukao K, Nakauchi H. Presence of hematopoietic stem cells in the adult liver. Nature Med 1996; 2: 198–203.

8 Watanabe H, Miyaji C, Seki S, Abo T. c-kit+ stem cells and thymocyte precursors in the livers of adult mice. J Exp Med 1996; 184: 687–693. 9 Terstappen LWMM, Huang S, Safford M, Lansdorp P, Loken MR. Sequential generations of hematopoietic colonies derived from single nonlineage-committed CD34+CD38− progenitor cells. Blood 1991; 77: 1218–1227. 10 Snoeck H-W, Van Bockstaele DR, Nijs G, Lenjou M, Lardon F, Haenen L, Rodrigus I, Peetermans ME, Berneman ZN. Interferon␥ selectively inhibits very primitive CD34++CD38− and not more mature CD34+ CD38+ human hematopoietic progenitor cells. J Exp Med 1994; 180: 1177–1182. 11 Civin CI, Almeida-Porada G, Lee MJ, Olweus J, Terstappen LWMM, Zanjani ED. Sustained, retransplantable, multilineage engraftment of highly purified adult human bone marrow stem cells in vivo. Blood 1996; 88: 4102–4109. 12 Bhatia M, Wang JCY, Kapp U, Bonnet D, Dick JE. Purification of primitive human hematopoietic cells capable of repopulating immune-deficient mice. Proc Natl Acad Sci USA 1997; 94: 5320–5325. 13 Hao QL, Shah HJ, Thiemann FT, Smogorzewska E, Crooks GM. A functional comparison of CD34+CD38− cells in cord blood and bone marrow. Blood 1995; 86: 3745–3753. 14 Lansdorp PM, Dragowska W, Mayani H. Ontogeny-related changes in proliferative potential of human hematopoietic cells. J Exp Med 1993; 178: 787–791. 15 Weekx SFA, Van Bockstaele DR, Plum J, Moulijn A, Rodrigus I, Lardon F, De Smedt M, Nijs G, Lenjou M, Loquet P, Berneman ZN, Snoeck HW. CD34++CD38− and CD34+CD38+ human hematopoietic progenitors from fetal liver, cord blood and adult bone marrow respond differently to hematopoietic cytokines depending on the ontogenic source. Exp Hematol 1998; 26: 1034–1042. 16 Hatzfeld J, Li ML, Brown EL, Sookdeo H, Le´vesque JP, O’Tolle T, Gurney C, Clark SC, Hatzfeld A. Release of early human hematopoietic progenitors from quiescence by antisense transforming growth factor ␤1 or Rb oligonucleotide. J Exp Med 1991; 174: 925–929. 17 Cardoso AA, Li M-L, Batard P, Hatzfeld A, Brown EL, Le´vesque JP, Sookdeo H, Panterne B, Sansilvestri P, Clarck SC, Hatzfeld J. Release from quiescece of CD34+CD38− human umbilical cord blood cells reveals their potentiality to engraft adults. Proc Natl Acad Sci USA 1993; 90: 8707–8711. 18 Fortunel N, Batard P, Hatzfeld A, Monier M-N, Panterne B, Lebkowski J, Hatzfeld J. High proliferative potential-quiescent cells: a working model to study primitive quiescent hematopoietic cells. J Cell Sci 1998; 111: 1867–1875. 19 Bu¨hring HJ, Ullrich A, Schaudt K, Mu¨ller CA, Busch W. The product of the proto-oncogene c-kit (P145c-kit) is a human bone marrow surface antigen of hematopoietic precursor cells which is expressed on a subset of acute non-lymphoblastic leukemic cells. Leukemia 1991; 5: 854–860. 20 Sansilvestri P, Cardoso AA, Batard P, Panterne B, Hatzfeld A, Lim B, Le´vesque J-P, Monier M-N, Hatzfeld J. Early CD34high cells can be separated into KIThigh cells in which transforming growth factor␤ (TGF-␤) downmodulates c-kit and KITlow cells in which antiTGF-␤ upmodulates c-kit. Blood 1995; 86: 1729–1735. 21 Hatzfeld A, Batard P, Panterne B, Taieb F, Hatzfeld J. Increased stable retroviral gene transfer in early hematopoietic progenitors released from quiescence. Hum Gene Ther 1996; 7: 207–213. 22 Dubois CM, Ruscetti FW, Palaszynski W, Falk LA, Oppenheim JJ, Keller JR. Transforming growth factor-␤ is a potent inhibitor of interleukin-1 (IL-1) receptor expression: proposed mechanism of inhibition of IL-1 action. J Exp Med 1990; 172: 737–744. 23 Jacobsen SEW, Ruscetti FW, Roberts AB, Keller JR. TGF-␤ is a bidirectional modulator of cytokine receptor expression on murine bone marrow cells. J Immunol 1993; 151: 4534–4544. 24 Jacobsen SE, Ruscetti FW, Dubois CM, Lee J, Boone TC, Keller JR. Transforming growth factor-beta transmodulates the expresion of colony-stimulating factor receptors on murine hematopoietic progenitor cell lines. Blood 1991; 77: 1706–1716. 25 Jacobsen SE, Veiby OP, Myklebust J, Okkenhaug C, Lyman SD. Ability of flt3 ligand to stimulate the in vitro growth of primitive murine hematopoietic progenitors is potently and directly inhibited by transforming growth factor-␤ and tumor necrosis factor-␣. Blood 1996; 87: 5016–5026.

1271


1272

26 Dubois CM, Ruscetti FW, Stankova J, Keller JR. Transforming growth factor-␤ regulates c-kit message stability and cell-surface protein expression in hematopoietic progenitors. Blood 1994; 83: 3138–3145. 27 Ploemacher RE, van Soest PL, Boudewijn A. Autocrine transforming growth factor-beta 1 blocks colony formation and progenitor cells generation by hematopoietic stem cells stimulated with steel factor. Stem Cell (Dayt) 1993; 11: 336–347. 28 Jacobsen SE, Ruscetti FW, Ortiz M, Gooya JM, Keller JR. The growth response of Lin-Thy1+ hematopoietic progenitors to cytokines is determined by the balance between synergy of multiple stimulators and negative cooperation of multiple inhibitors. Exp Hematol 1994; 22: 985–989. 29 Kawashima I, Zanjani ED, Almaida-Porada G, Flake AW, Zeng HQ, Ogawa M. CD34+ human marrow cells that express low levels of kit protein are enriched for long-term marrow-engrafting cells. Blood 1996; 87: 4136–4142. 30 Vaziri H, Dragowska W, Allsopp RC, Thomas TE, Harley CB, Lansdorp PM. Evidence for a mitotic clock in human hematopoietic stem cells: loss of telomeric DNA with age. Proc Natl Acad Sci USA 1994; 91: 9857–9860. 31 Engelhardt M, Kumar R, Albanell J, Pettengell R, Han W, Moore MAS. Telomerase regulation, cell cycle, and telomerase stability in primitive hematopoietic cells. Blood 1997; 90: 182–193. 32 Lansdorp PM. Lessons from mice without telomerase. J Cell Biol 1997; 139: 309–312. 33 Morrison SJ, Prowse KR, Ho P, Weissman IL. Telomerase activity in hematopoietic cells is associated with self-renewal potential. Immunity 1996; 5: 207–216. 34 de Haan G, Nijhof W, Van Zant G. Mouse strain-dependent changes in frequency and proliferation of hematopoietic stem cells

35

36 37

38

39 40

41

42

during aging: correlation between lifespan and cycling activity. Blood 1997; 89: 1543–1550. Flemming WH, Alpern EJ, Uchida N, Ikuta K, Spangrude GJ, Weissman IL. Functional heterogeneity is associated with the cell cycle status of murine hematopoietic stem cells. J Cell Biol 1993; 122: 897–902. Morrison SJ, Wandycz AM, Akashi K, Weissman IL. The aging of hematopoietic stem cells. Nature Med 1996; 2: 1011–1016. Rebel VI, Miller CL, Thornsbury GR, Dragowska WH, Eaves CJ, Lansdorp PM. A comparison of long-term repopulating hematopoietic stem cells in fetal liver and adult bone marrow from the mouse. Exp Hematol 1996; 24: 638–648. Rebel VI, Miller CL, Eaves CJ, Lansdorp PM. The repopulation potential of fetal liver hematopoietic stem cells in mice exceeds that of their adult bone marrow counterparts. Blood 1996; 87: 3500–3507. Van Zant G, de Haan G, Rich IN. Alternatives to stem cell renewal from a developmental viewpoint. Exp Hematol 1997; 25: 187– 192. Tang B, Bo¨ttinger EP, Jakowlew SB, Bagnall KM, Mariano J, Anver MR, Letterio JJ, Wakefield M. Transforming growth factor-␤1 is a new form of tumor suppressor with true haploid insufficiency. Nature Med 1998; 4: 802–807. Hahn SA, Schutte M, Hoque ATMS, Moskaluk CA, da Costa LT, Rozenblum E, Weinstein CL, Fischer A, Yeo CJ, Hruban RH, Kern SE. DPC4, a candidate tumor supressor gene at human chromosome 18q21.1. Science 1996; 271: 350–353. de Winter JP, Roelen BAJ, ten Dijke P, van der Burg B, van den Eijnden-van Raaij AJM. DPC4 (SMAD4) mediates transforming growth factor-␤1 (TGF-␤1) induced growth inhibition and transcriptional response in breast tumour cells. Oncogene 1997; 14: 1891–1898.