Autocrine transforming growth factor-b regulation of ... - Nature

Oncogene (2005) 24, 5751–5763

& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00 www.nature.com/onc

Autocrine transforming growth factor-b regulation of hematopoiesis: many outcomes that depend on the context Francis W Ruscetti*,1, Salem Akel2 and Stephen H Bartelmez3 1 Laboratory of Experimental Immunology, Center for Cancer Research, National Cancer Institute-Frederick, Frederick, MD 21702-1201, USA; 2Department of Medical Laboratory Sciences, Hashemite University, Zarqa, Jordan; 3Hemogenix, Inc., Colorado Springs, CO, USA

Transforming growth factor-b (TGF-b) is a pleiotropic regulator of all stages of hematopoieis. The three mammalian isoforms (TGF-b1, 2 and 3) have distinct but overlapping effects on hematopoiesis. Depending on the differentiation stage of the target cell, the local environment and the concentration and isoform of TGF-b, in vivo or in vitro, TGF-b can be pro- or antiproliferative, pro- or antiapoptotic, pro- or antidifferentiative and can inhibit or increase terminally differentiated cell function. TGF-b is a major regulator of stem cell quiescence, at least in vitro. TGF-b can act directly or indirectly through effects on the bone marrow microenvironment. In addition, paracrine and autocrine actions of TGF-b have overlapping but distinct regulatory effects on hematopoietic stem/progenitor cells. Since TGF-b can act in numerous steps in the hematopoietic cascade, loss of function mutations in hematopoeitic stem cells (HSC) have different effects on hematopoiesis than transient blockade of autocrine TGF-b1. Transient neutralization of autocrine TGF-b in HSC has therapeutic potential. In myeloid and erythroid leukemic cells, autocrine TGF-b1 and/or its Smad signals controls the ability of these cells to respond to various differentiation inducers, suggesting that this pathway plays a role in determining the cell fate of leukemic cells. Oncogene (2005) 24, 5751–5763. doi:10.1038/sj.onc.1208921 Keywords: transforming growth factor-b1; hematopoiesis; autocrine; stem cells

complex series of events. Hematopoietic stem cells (HSC), which can remain quiescent for months, must retain the abilities to differentiate into all lineages, to self-renew and to maintain long-term repopulation marrow activity (LTMRA). Recent work has shown that a variety of intrinsic transcription factors such as Runx, PU.1, GATA1, c-myb and SCL as well as external cellular and humoral factors present in the bone marrow microenvironment are involved in hematopoietic regulation. In the last 30 years, we and numerous others have shown that the growth and differentiation of hematopoietic cells is positively and negatively regulated by a number of distinct cytokines (Ogawa, 1993). Constitutive hematopoiesis is regulated, in part, by the balance of these opposing signals (Jacobsen et al., 1994). It has been our hypothesis that transforming growth factor-b1 (TGF-b1) is an essential regulator of hematopoiesis. Previous in vitro and in vivo work indicates that TGF-b is a regulator of all stages of hematopoiesis (for reviews, see Fortunel et al., 2000; Hu and Zuckerman, 2001). The function of TGF-b1 in this process is underscored by the observation that TGF-b1-null mice have a significant increase in mature myeloid cells and rapid death due to inflammation (Shull et al., 1992; Kulkarni et al., 1993; Letterio et al., 1994). We proposed that TGF-b1 regulates both HSC quiesence and survival through autocrine and paracrine mechanisms. Here, we will review new data that suggest that TGF-b is not essential for regulation of HSC quiescence, the difficulties involved in defining the outcomes of TGF-b signaling due to crosstalk between signaling pathways and whether the actions of autocrine TGF-b on regulating the heterogeneous HSC compartment will present new therapeutic opportunities.

Introduction In hematopoiesis, the need to continuously generate large numbers of maturing cells of eight distinct lineages from small numbers of stem cells requires a highly

Quiescent HSC are directly and reversibly growth inhibited in vitro by TGF-b

*Correspondence: FW Ruscetti, Leukocyte Biology Section, Laboratory of Experimental Immunology, Bldg 567, RM 251, NCI-Frederick, Frederick, MD 21702-1201, USA; E-mail: [email protected] The content of this publication does not reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products or organizations imply endorsement by the US Government.

In order to determine if TGF-bs directly regulated stem cell proliferation in vitro (Keller et al., 1990), sequential cell sorting of murine lin bone marrow cells was used for low Rhodamine 123 (Rho, a mitochrondial DNA dye) followed by low Hoescht 33342 (Ho, genomic DNA binder) and c-kit þ , to isolate a purified HSC

Autocrine TGF-b1 regulation of hematopoiesis FW Ruscetti et al

5752

population. This method has been recently shown to isolate the most quiescent part of the side population (SP) cells which contains most of the HSC (Bertoncello and Williams, 2004). Ho/Rholow HSC are CD34, SCA1 þ , give long-term donor repopulation with 5–10 cells (Wolf et al., 1993) and 93–100% of these cells can be stimulated by a combination SCF, IL-3 and IL-6 to develop high proliferative potential (HPP) macrocolony (>100 000 cells) formation in vitro (Wolf et al., 1993; Sitnicka et al., 1996). Functionally, Ho/Rholow HSC are in the G0 state of the cell cycle and have no radioprotection activity. In single cell assays, the continuous presence of exogenous TGF-b1 at 10 mg/ml directly inhibited the in vitro cell division of essentially all Ho/ Rholow HSC between their 0–5th division (mean ¼ eight cells) (Sitnicka et al., 1996). The addition of TGF-b1 neutralizing antibodies as late as 3 days after the addition of exogenous TGF-b1 preserved the ability of 80% of the cells to make macroclones, indicating that the effect of exogenous TGF-b1 was reversible and not markedly apoptotic to HSC in the context of other cytokines. Delayed addition of TGF-b1 until the HSC had undergone three divisions (eight cells/well) was still able to cause growth arrest of the cells at that stage. Thus, TGF-b1 inhibits the in vitro growth of HSC in the Ho/Rholow cell population. Direct inhibitory effects of TGF-b1 on purified CD34 þ human stem/progenitor cells has also been reported. Using purified CD34high þ Lin (Lu et al., 1993) and CD34 þ CD38 (Batard et al., 2000) stem/ progenitor cells, it has been demonstrated that TGF-b inhibited 80–90% of the colony formation induced by various cytokine combinations including early acting cytokines (SCF, Flt-3 or TPO), differentiating cytokines (GM-CSF, IL-3 or EPO) plus cell cycle accelerators (IL-6). Recently, a CD34High þ , CD38, c-kitlow, IL-6Rlow and Mpllow cell population was isolated from human bone marrow progenitors (designated HPP-Q). These cells possess many of the characteristics of murine Ho/Rholow HSC. They are quiescent, yet can be stimulated to form HPP macrocolonies (Fortunel et al., 1998). In single cell assays, these cells can be growth arrested by TGF-b. It is not known whether these cells can promote long-term engraftment particularly after in vitro growth arrest with TGF-b1. The observation that proliferation of both human and murine quiescent primitive cells can be rapidly growth arrested by TGF-b could result in the cells maintaining their stem cell nature. Ho/Rholow HSC cultured for 5 days with SCF, IL-3 and IL-6 showed 15% of donor repopulation at 10 months compared to 60% using the same number of uncultured cells. The ability of the cultured cells to compete in a competitive repopulation assay was detectable but less than the uncultured cells. However, the same cells treated for 5 days with cytokines and TGF-b had no activity. Similarly, Ho/ Rholow HSC stimulated to undergo cell division and then growth arrested with TGF-b showed markedly reduced engrafting potential (Wiesmann et al., 2000). In human and murine gene transfer experiments, when TGF-b was used to growth arrest the cells after retroviral infection, Oncogene

severely decreased repopulation was seen (Nolta and Kohn, 1990). TGF-b mediated arrested growth at the G1/S transition and G0 quiescence are not the same. Throughout the G1/S transition, new mRNAs and proteins are being made in contrast to little or no synthesis in G0. Perhaps, this marcomolecular synthesis results in the loss of ‘stemness’ in these cells even in the presence of TGF-b1. Evidence supporting this hypothesis was obtained when 50 Ho/Rhlow HSC were expanded in liquid culture with IL-3, SLF and IL-6 in the presence or absence of TGF-b1. After an 8 day incubation, the number of HPP-colony-forming cells (CFC) in control cultures expanded from 35 to 450, the number of GM-CFC from none to roughly 70 000, and the total cellularity increased to approximately 300 000 (Sitnicka et al., 1996). The addition of optimal doses of TGF-b to these cultures completely inhibited the proliferation and expansion of the initial cell input. In contrast, TGF-b at 0.4 mg/ml stimulated an increase in HPP-CFC expansion and an increase in total cellularity to 450 000 cells, relative to untreated cultures. These results suggest that even when HSC are growth arrested, TGFb treatment results in subtle changes in differentiation state. As discussed previously (Ruscetti and Bartelmez, 2001), TGF-b blocks HSC cell cycle progression at the G1/S boundary not G0 making it likely that TGF-binduced HSC inhibition is not equivalent to in vivo quiescence.

Stem cell regulation in TGF-b-null mice The study of the development of the hematopoietic tissue in vivo in the absence of endogenous TGF-b1 has been greatly aided by the generation of a variety of knockout mice. Homozygous TGF-b1 knockout mice have a 50% intrauterine death rate because of severe developmental retardation. A strain-dependent variation in survival occurs due to trans-placental transport of maternal TGF-b1 (Bottinger et al., 1997). Progressive wasting occurs with death by 5 weeks due a systemic T-cell-dependent massive inflammatory syndrome. MHC class II-dependent autoimmune disease also occurred. Defective hematopoiesis results in plasmacytosis, myeloid hyperplasia, reduced numbers of erythroid cells as well as a lack of Langerhans dendritic cells. These defects correlated with the absence of TGF-b1 (Dickson et al., 1995; Yaswen et al., 1996; Borkowski et al., 1997). These mice still make adequate amounts of TGF-b2 and b3, which shows that the isoforms are not equal in function in vivo. The double TGF-b1/MHC class II knockout mice lose the proinflammatory and autoimmune defects but still retain the massive myeloproliferative syndrome (Letterio et al., 1998). The phenotype of the type II TGF-b receptor (TbRII)-null mice was indistinguishable from that of TGF-b1 knockout mice (Oshima et al., 1996). Attempts at isolating the quiescent Ho/Rholow c-kit þ HSC from the TGF-b knockout mice have resulted in a >90% reduction of these cells


5753 Table 1

Stem cella progenitora phenotypes in TGF-b and Class II KO marrow

% c-kit+ in sample

% Rh low, Ho lowb

% Rh low, Ho low, c-kit+c

Littermate Class II KO TGF-b, Class II KO Normal–3 months

10.271.2 11.9710.6 10.2 9.1

1.170.9 1.270.9 0.5 1.5

0.870.5 1.270.8 0 0.7

Type of mouse Littermate Class II KO TGF-b, Class II KO Normal–3 months

% SCA-1+ in sample 4.372.0 6.271.4 3.1 8.3

% Rh low, Ho low 1.170.9 1.270.9 0.5 1.5

% Rh low, Ho low, SCA-1+ 4.771.8 6.172.6 1.9 0.8

Type of mouse

a (a) Rh low, Ho low are stem cell enriched, (b) Rh low, Ho low, c-kit+ or SCA-1 + are highly enriched for stem cells, (c) c-kit+ contain most progenitors. bThis is percent of all viable cells going to the sorter. cThis is the percent of Rh low Ho low cells that were c-kit+ or SCA-1+

indicating that autocrine TGF-b1 could be a major regulator of stem cell quiescence (Table 1), although other molecules such as interferon g (Yang et al., 2005b), and TNFa (Jacobsen et al., 1992) also contribute to this inhibition. It is not clear whether this lack of quiescent HSC is due solely to the absence of TGF-b1 or subclinical pathologies also contribute to it. It has been proposed that loss of function mutations using conditional deletions is the most accurate to study the physiologic role of a certain gene or a pathway. Given TGF-b’s complex interactions at multiple stages of hematopoiesis, a dynamic cellular process which by itself has many indirect and direct complex interactions with the environment, drawing any clearcut conclusions, may be a ‘leap of faith’. Using such a conditional model, TbRI-null mice have normal hematopoiesis containing normal progenitor numbers and rates of differentiation (Larsson et al., 2001, 2003). Compared to littermate controls, HSC from these mice have normal cell cycle profiles and normal LTMRA following bone marrow transplantation. Furthermore, HSC from these TbRInull mice have normal susceptibility to repeated 5FU treatments. Similarly, hematopoietic failure occurred at a stage identical to normal controls following serial transplants (Larsson et al., 2005). These results clearly show that TGF-b is not an essential regulator of HSC quiescence and self-renewal during constitutive and stress-induced hematopoiesis.

Stem cell quiescence is regulated by a cellular niche How can the starkly opposing effects of TGF-b on HSC in vivo and in vitro be explained? Numerous historical studies have shown that bone marrow has an organized and structured architecture in which close relationships exist between HSC and the microenvironment (Lord, 1990). Recently, following transplantation, stem cells showed selective redistribution and were highly enriched in the endosteal region while the more mature progenitors were enriched in the central marrow region (Nilsson et al., 2001). Using genetic techniques, it was subsequently noted that osteoblasts were a critical component

of the marrow microenvironment (Calvi et al., 2003; Zhang et al., 2003). Elegant work by Arai et al. (2004) showed that the receptor tyrosine kinase Tie2-expressing HSC are quiescent, and adhere to the osteoblasts in the endosteal niche. The interaction of Tie2 with its ligand angiopoietin-1 (Ang-1) maintained the in vivo LTMRA of the HSC. Ang-1 enhanced the ability of the Tie2 HSC to become quiescent resulting in the protection of the HSC compartment from myelosuppressive stress. These cells are resistant to a myelosuppressive dose of 5FU. Furthermore, Arai and his co-workers show that Tie2 signaling in HSC maintains LTMRA in vitro. Tie2 does this by keeping the cells quiescent, clearly demonstrating that HSC self-renewal capacity can be maintained by preventing cell division. Other regulators of the cell cycle such as p21 (Cheng et al. (2000a); p27 Cheng et al. (2000b); Bmi-1(Park et al. (2004)) have been shown to regulate HSC self-renewal (Cheshier et al., 1999). It is essential for an accurate interpretation of the data obtained from disrupting parts of the TGF-b pathway in mice that one determines the effects of any deletion on the HSC-microevironment interactions. It is then reasonable to ask why not a nonessential role for TGF-b in the same HSC self-renewal process? When HSC/progenitor cells from TbRI-null mice (Larsson et al., 2003) and TbRII-null mice (Fan et al., 2002) were stimulated in vitro by a single cytokine, there were significantly increased numbers of actively proliferating cells comparable to the numbers by transient blockage of TGF-b1 (see later section). This suggests that the differences lie in the responses to the microenvironment. Several studies have examined the effects of TGF-b on hematopoietic cell growth in long-term bone marrow cultures (LTBMCs). This in vitro culture system contains a complex mixture of hematopoietic cells and stromal accessory cells that can sustain the growth and proliferation of primitive progenitors over many weeks in culture (Dexter, 1979). Both mRNA for TGF-b and active and latent TGF-b protein have been detected in the stroma of LTBMCs (Cashman et al., 1990; Eaves et al., 1991). Several conclusions have been drawn from these studies. First, the addition of exogenous TGF-b at the initiation of the culture completely arrests hematopoiesis, such that no clonogenic Oncogene


5754

progenitors are found after 2 weeks (Waegell et al., 1994); second, the addition of TGF-b1 inhibits the increased cycling of primitive progenitors initiated in the culture by the addition of IL-1 or a horse serum media; third, the addition of neutralizing TGF-b1 antibodies 3 days after the addition of IL-1, when progenitors were maximally cycling, allowed the cells to remain longer in S phase, whereas progenitor cells in control cultures underwent cell cycle arrest; and fourth, the addition of neutralizing TGF-b antibodies to LTBMCs resulted in an increased output of mature myeloid cells, mature progenitors and day-14 CFUs. Thus, the production of TGF-b by stromal cells results in a negative-feedback loop to maintain stem cell quiescence and to inhibit progenitor cell cycling. Does TGF-b1 regulate or interact with the signaling pathways of Wnt (Staal and Clevers, 2005), Notch (Stier et al., 2002; Mancini et al., 2005) or Tie2 (Arai et al., 2004), thought to be involved in stem cell renewal? Cytokines such as IL-1 and TPO whose stromal production is increased by TGF-b1 (Sakamaki et al., 1999) transiently override this inhibitory effect. These results could be indirect since the TGF-b plays a role in the development and function of osteoblasts in the marrow niche (Lieb et al., 2004). Osteoblasts and adipocytes develop from a common progenitor from the bone marrow stroma/stem cells. TGF-b plays a role in directing the cell fate of these stem cells to become osteoblasts and in bone formation rather than fat tissue formation (Moerman et al., 2004).

Smad crosstalk, differential effects of TGF-b isoforms and other unsolved mysteries Smad 2 and Smad 3 are intracellular signaling molecules for TGF-b1. Smad3-null mice are viable demonstrating defects in immune function (Yang et al., 1999a, b), whereas Smad2-null phenotype is embryonically lethal indicating that Smad 2 and Smad 3 regulate a nonoverlapping set of genes. Recent studies have shown that Smad3 is the major transcriptional activator for TGF-b, while Smad2 is a transmodulator of this transcriptional activity (Yang et al., 2003). Previous studies suggested that Smad5 is involved in the signaling pathway by which TGF-b inhibits proliferation of primitive human hematopoietic progenitor cells, since suppression of Smad5 expression by antisense oligonucleotides reversed the inhibitory effects of TGF-b on hematopoietic colony formation (Bruno et al., 1998). While this study was one of the first to suggest that Smad 5, typically activated by bone morphogenetic proteins (BMPs), might be activated by TGF-b, more recent studies have shown this to occur in other cells as well, including intestinal epithelial cells (Yue and Mulder, 2001) and endothelial cells (Oh et al., 2000; Goumans et al., 2002). It has also been shown that BMPs can regulate the development of human hematopoietic stem cells, presumably through the BMPactivated Smad proteins, Smad1 and Smad5, which they expressed in HSC (Bhatia et al., 1999). In addition, Oncogene

several Smad-independent pathways have been identified (Engel et al., 1999; Yue and Mulder, 2001; Wilkes et al., 2003). The TGF-b-activated receptor complex can signal through AKT, mitogen-activated protein kinases (MAPKs), the phoshoinositol-3-kinase (PI3K) and the PP2A/p70s65K pathways (Wakefield and Roberts, 2002). AKT (Remy et al., 2004) and the tuberous sclerosis complex 2 gene product (Birchenall-Roberts et al., 2005) can both alter TGF-b signaling by binding to Smad3. The numerous pathways involved in crosstalk adds new complexity to the TGF-b regulation of hematopoiesis. The opposing effects of TGF-b1 and 2 on HSC regulation are equally puzzling. Recently, it has been shown that TGF-b2 has a positive regulatory role on HSC/progenitor cells (Langer et al., 2004). Studying the proliferation of these cells, it was found that TGF-b2 had a biphasic dose response (low concentrations were stimulatory while high concentrations were inhibitory). Furthermore, the number and repopulating ability of the HSC in herterozygous null TGF-b2 mice were significantly lower than the littermate controls. There have been several reports that isoforms, both qualitatively and quantitatively differs in their effects on progenitor cells (Ohta et al., 1987; Jacobsen et al., 1991a, b, 1992) The reasons for these opposing properties of TGF-b1 and 2 on HSC regulation are not clear. Even though the nonsignaling binding molecules, betaglycan III (Stenvers et al., 2003) and endoglin (Chen et al., 2002, 2003), can increase the binding affinity of TGF-b2 but not b1 for the receptor on HSC , both isoforms use the same receptor activation and Smad signals, leaving the biochemical mechanism obscure. Another mysterious mechanism involves the autocrine actions of TGF-b by hematopoietic stem/progenitor cells that can negatively regulate the cycling status of progenitors (Hatzfeld et al., 1994; Sitnicka et al., 1996). Specifically, the addition of TGF-b antisense oligonucleotides or TGF-b neutralizing antibodies enhanced the numbers and size of most immature bone marrow colonies from purified human or human stem/progenitor cells in response to various combinations of cytokines. In some experiments, more than half of the detectable CD34 þ CFC were maintained in a growth factor unresponsive state by autocrine TGF-b. The observations that neutralizing antibodies that prevent binding to the cell surface receptor can have differential effects on autocrine vs paracrine TGF-b remain a mystery. The clearest example of this is the HSC from Smad3-null mice which are still sensitive to growth inhibition by exogenous TGF-b1 but are not growth arrested by the autocrine pathway (Ruscetti, Roberts and Letterio, unpublished results). Furthermore, disruption of the TbRII gene in mice leads to a lethal inflammatory disorder that is transplantable (Leveen et al., 2002). In an adoptive transfer, TGF-b1 bone marrow reconstitutes B and T lymphoid tissues in Rag2-null mice. Despite normal levels of TGFb1 expressed in all other tissues, the mice die from a lethal inflammatory, showing an obligate autocrine


5755

function for TGF-b1 in lymphoid homeostasis (Mamura et al., submitted).

Autocrine TGF-b1 is a regulator of the in vitro survival of murine HSC-LTRA Autocrine TGF-b1 has direct effects on the function of murine HSC in vitro. SCF or IL-3 as single agents promoted HSC to survive as single cells or undergo only a few (1–3) cell divisions. Addition of a neutralizing anti-TGF-b1 monoclonal antibody, in conjunction with SCF or IL-3, increased the proportion of HSC that divided as well as increase the average clone size. More striking was the effect of adding anti-TGF-b1 to single HSC cultures in the absence of added growth factors; where a high proportion of cells survived up to 14 days as single cells compared to medium alone in which cell survival was limited to a few days (Ruscetti et al., 1999). Neutralizing of autocrine TGF-b1 was required because a non-neutralizing antibody that binds with the same affinity does not prolong survival and performing the assay in serum-free conditions due to concerns about TGF-b in sera (Dybedal and Jacobsen, 1995) did not affect the results (Table 2). All cells began cell cycle progression (measured as the time to first division) more rapidly when treated with anti-TGF-b1. Cultures of 100 LTR-HSC were grown without cytokine or sera but in the presence of anti-TGF-b1 for 5 days, then assayed in a competitive repopulation assay which indicated that a substantial proportion of the surviving cells retained their LTR ability (Soma et al., 1996; Ruscetti et al., 1999). The competitive repopulating units indicated a relative enrichment of HSC compared to 105 normal bone marrow cells. Moreover, we isolated stem/progenitor cells from TGF-b1-deficient mice. These progenitor cells did not show any increase in survival after neutralizing antibody treatment. Thus, neutralization of autocrine TGF-b1 directly promotes the survival of the HSC and preserves some LTMRA in culture. Further, adoptive transfer experiments using marrow from TGF-b1, MHC class II knockout mice suggest the existence of significant autocrine TGF-b1 functions. Transplantation of TGF-b1-null bone marrow into lethally irradiated wild-type recipients has shown that

bone marrow is the major source of plasma TGF-b1 (recipients of TGF-b1-null donor marrow have fivefold lower levels than recipients of wild-type marrow).

Inhibition of autocrine TGF-b1 in HSC by morpholino antisense oligomers prior to transplant accelerates engraftment while reducing stem cell numbers required for long-term repopulation Even a short (60 min) exposure of highly purified HSC to a neutralizing anti-TGF-b1 antibody dramatically reduced the time required for engraftment of HSC with LTMRA (Bartelmez et al., 2000). Subsequent experiments had shown that morpholino antisense nucleotides (MAS) to TGF-b1 or TGF-bRII induce HSC survival, reduce the number of HSC required for hematopoietic repopulation, and decrease the time to engraftment. Donor-derived peripheral blood T cells, B cells, monocytes and neutrophils total blood count were assayed. As expected, transplantation of 100 untreated LTRHSC or control-treated LTR-HSC (scramble MAS or isotype antibodies) required competitor or support marrow to prevent hematopoietic death due to radiation and engrafted slowly over a period of 1.5–6 months: donor cell chimerism and reached 30–40% of circulating cells at B3 months. In contrast, 60 HSC exposed to TGF-b MAS for 16 h engrafted rapidly to reach donor cell chimerism of 50–60% by 3 weeks and 70–85% by 1.5 months. The early engraftment was predominantly donor neutrophils followed by B cells and then T cells by 1.5 months. The induction of rapid engraftment of LTR-HSC by TGF-b MAS or neutralizing antibody did not require any support cells nor impair their LTMRA ability as measured by sustained high % donor chimeras beyond 1 year post-transplant. Also, a blockade of HSC autocrine TGF-b dramatically reduced the number of HSC required to rescue a lethally irradiated congenic recipient, while more than 1000 control LTR-HSC per mouse were required for hematopoietic rescue, as few as 30 TGF-b MAS-treated cells would rescue a recipient (Bartelmez et al., 2000). Translation of this ex vivo method to human stem cell transplants could reduce the time to engraftment plus lower the number of stem cells required for a durable graft.

Table 2 Effects of Anti TGF-b antribody and serum on cytokine-dependent survival of purified quiescent murine Ho/Rholow stem cells Stem cell factor

No antibody Neutralizing anti TGF-b Binding anti TGF-b Serum-free, no antibody Serum free, neutralizing anti-TGF-b

% Viable cells Day 1

Day 2

Day 4

Day 7

Day10

Day 14

9276I 9871 9871 9772 9870I

8678 9473I 7973S 8173 9372

72715 9373 6570.9 70714 8876

4679 8673 4278 3676 7778

1773 8377 2377 1476 5977

1272 7478 973 674 4378

Murine Lin cells were further purified by FACS separation and seeded as single cells in Terasaki plates. A minimum of 1000 wells was scored per group. Cultures were supplemented with SCF (20 ng/ml) or IL-3 (20 ng/ml), cytokine reported to protect quiescent marrow cells against death (Keller et al., 1995) plus or minus anti-TGF-b (20 mg/ml, ID11 neuralizing or 2G1.12 binding only). Media were IMDM 12.5 % horse serum and 12.5% fetal calf serum; serum free was QBSF alone Oncogene


5756

The mechanism by which a blockade of autocrine TGF-b1 in quiescent HSC occurs remains unclear. Does antibody and antisense treatment block TGF-b1mediated apoptosis in the absence of stromal or cytokine factors? TGF-b1 has been shown to regulate the cell surface expression of many cytokine receptors (Dubois et al., 1990, 1994; Jacobsen et al., 1991a, 1992) as well as Fas (Dybedal et al., 1997). Since antisense treatment of 60 cells rescues a mouse from lethal irradiation, the HSC must undergo differentiation to short-term repopulating cells (Yang et al., 2005a, b). In what progeny and at what time do the cells resume producing autocrine TGF-b? Does this blockade of autocrine TGF-b lead to better homing to the marrow space (Quesenberry and Becker, 1998)? Transient modulation of CD26, a cell surface molecule which negatively regulates engraftment, greatly increases the efficiency of bone marrow transplantation (Christopherson et al., 2004). It is somewhat puzzling that such a brief treatment with antibody or antisense leads to such a profound increase in engraftment. A brief 6 h treatment with anti-TGF-b antibody significantly increased number and size of macrocolonies in a single cell assays, indicating that the transient blockade of autocrine TGF-b in HSC sets in motion a cascade of hematopoietic proliferation and differentiation events that the presence of autocrine TGF-b in the progeny cannot alter.

In vitro neutralization of autocrine TGF-b in umbilical cord blood stem cells: improved routine engraftment of UCB stem cells in allogeneic transplants? Umbilical cord blood cells (UCB) are being used as a source of stem cells in allogeneic bone marrow transplants (Laughlin 2001; Sorrentino, 2004). The good news is that because of the immaturity of the donor immune system, the incidence and severity of graft vs host disease is favorable compared to adult transplants. The bad news is that time to recovery of the graft is delayed leading to more failures. Translation of this murine ex vivo method to human umbilical cord blood transplants could reduce the time to engraftment plus lower the number of stem cells required for a durable graft. The development of methods for in vitro neutralization of TGF-b in hematopoietic stem/progenitor cells is an important goal for clinical research, given the therapeutic potential of these cells. The challenge is to develop defined culture systems in which the cell cycling of primitive quiescent stem/progenitor cells is stimulated in presence of either TGF-b antibodies or TGF-b antisense, while their maturation and senescence are prevented (Hatzfeld et al., 1991, 1994, 1996; Ploemacher et al., 1993; Pierelli et al., 2000). In an attempt to define culture conditions to exploit TGFb1’s autocrine effects on hematopoiesis, the effect of blocking autocrine TGFb1 in serum-free SCF cultures was measured. These conditions allowed increased early erythroid development from primitive Oncogene

human HSC (Akel et al., 2003). UCB CD34 þ CD38 Lin cells were cultured in serum-free conditions using various combinations of SCF and TGFb1 neutralizing antibody. Anti-TGFb1 augmented proliferation of CD34 þ CD38 Lin cells (day 21) in SCF-stimulated cultures. While SCF alone stimulated production of tryptase þ mast cells, while cellsstimulated by SCF/antiTGFb1 were predominantly erythroid (CD36 þ CD14 and glycophorin A þ ). A distinct expansion of erythroid progenitors (CD34 þ CD36 þ CD14) with the potential to form erythroid colonies was seen revealing early EPO-independent erythroid development (Figure 1). In contrast, TGFb1 accelerated the conversion of large BFU-Es into CFU-Es. Finally, TGFb1 accelerated Epoinduced terminal erythroid differentiation and improved enucleation (from 773 to 2276%) in serum-free conditions (Figure 2). Serum addition stimulated enucleation (54718%) that was reduced to 26714% with anti-TGFb1, suggesting that optimal erythroid enucleation is EPO dependent requiring serum factors including TGFb1. Thus, the presence or absence of TGF-b can alter the differentiation outcomes in the same cell lineages.

Efficient retroviral-mediated gene transfer into stem/ progenitor cells released from quiescence by anti-TGF-b Efficient retroviral-mediated gene transfer as therapy for human genetic diseases implies that the host cells are in a cycling state, which is not normally the case for primitive HSC. Since both autocrine and paracrine TGF-b1 produced by hematopoietic progenitors are partly responsible for their maintenance in a quiescent state, the use of anti-TGF-b1 antibodies to neutralize the bioactivity of endogenous TGF-b1 during the retroviral infection procedure was attempted to increase gene transfer into primitive stem/progenitor cells. This strategy has been tested with success on hematopoietic cells of several origins, including human UCB cells (Imbert et al., 1998; Yu et al., 1998). The efficiency of

Figure 1 Effect of TGF-b1 on the ability of EPO to induce red cell enucleation of human red blood cells in serum-free cultures. UCB CD34 þ Lin cells were incubated for 21 days in SCF (50 ng/ml) and anti-TGF-b antibody without EP0. Then EPO were added with out TGF-b (left panel) and with TGF-b (right panel) and incubated for 72 h


5757 HL-60 Bipotential cell

1,25 VitD3 Smad2

TGF-β1 Smad2P

Smad2

t-RA

t-RA

*t-RA CD14+ monocyte

CD15+ granulocyte

Figure 2 A schematic model of the mechanism by which autocrine TGF-b regulates the ability of Vit D3 and ATRA to stimulate terminal differentiation of human promyelocytic HL-60 cell line to either monocytes or granulocytes

this method was improved by the coupled use of antiTGF-b1 blocking antibodies with antisense oligos against the cyclin-dependent kinase inhibitor (CDKI) p27kip1 (Dao et al., 1998). This approach has been extended to other effectors involved in the regulation of hematopoietic cell cycling by TGF-b1, such as the CDKI p21cip1. Stier et al., 2003). Moreover, CD34 þ cells produce TGF-b1 during retroviral transfection protocols, promoting a return to quiescence of the most primitive stem/progenitor cells after gene transfer. As discussed above, this G1/S-induced inhibition is not equivalent to quiescence and may be detrimental to maintaining the ‘stemness’ of the transduced cells.

TGF-b signaling pathway in leukemia: rare mutational inactivation A number of leukemic cell lines blocked in differentiation representing early stages of myelomonocytic development exist (Dexter et al., 1980 and Greenberger et al., 1983). Some absolutely require the presence of growth factors for their growth and survival while others are growth factor independent. Regardless of the factor used to stimulate cell growth of murine leukemic cell lines, their growth is inhibited by TGF-b (Keller et al., 1988), while differentiation-blocked (TGF-bR positive) cell lines were insensitive to TGF-b effects (Hampson et al., 1989). It has been proposed that cells with a more transformed phenotype (differentiation blocked) become refractory to the effects of TGF-b and possibly escape negative regulation. However, growth factorindependent cell lines which can form tumors in nude mice representing a more transformed phenotype can be both TGF-b sensitive and insensitive cell lines (Keller et al., 1988; Sing et al., 1988). Furthermore, infection of IL-3-dependent cell lines with oncogene containing retroviruses (v-abl, v-src and v-fms), to abrogate their requirement for growth factors, produces cell lines

retaining their TGF-b sensitivity in all cases. Consistent with the results above, TGF-b is a potent inhibitor of the growth of freshly aspirated human CML (Sing et al., 1988; Algietta et al., 1989) and AML (Tessier and Hoang 1988; Nara et al., 1989) even in the presence of saturating concentrations of growth simulators. It has been proposed that genomic instability plays a role in the disease progression of CML However, unlike colon cancer (Grady et al., 1999), alterations in the microsatelite regions of the TbRII gene are not seen in CML. However, in all phases of CML and Bcell-CLL, there is a reduction in the number of TbRII transcripts (Rooke et al., 1999). In the case of B-cell CLL, this leads to 30% of the cases being insensitive to TGF-b. In those cases, mutations have been found in the signal sequence of the TbRI gene (Schiemann et al., 1999). Mutations in the Smad pathway are rare in leukemia. However, disruption of Smad transcriptional responsiveness has been associated with leukemic transformation. AML-1, a transcriptional coactivator, has a binding domain with homology to FAST, a protein involved in Smad2 signaling. The most common translocation in AML (t8;21) results in an AML-ETO fusion protein. The physical association of this fusion protein with Smad proteins leads to a repression of constitutive and TGF-b inducible gene transcription (Jakubowiak et al., 2000). The physical association of several oncoproteins with Smad3 blocks the ability of TGF-b to downmodulate c-myc in G1, which is necessary to halt cell cycle progression. The ability of Evi-1(Kurokawa et al. (1998), EIA, c-ski (Shi and Massague, 2003)) and Tax to associate with Smad3 prevents the formation of the transcriptional complex of Smad3/E2F and the corepressor p107(for review, see Kim and Letterio, 2003). p107-null mice get a myeloid hyperplasia. Loss of Smad3 in association with loss of p27 has been associated with murine and human lymphoid leuekmias (Wolfraim et al., 2004, see J Letterio, this volume).

Cell fate stimulated by differentiation inducers in myeloid leukemia cells is determined by endogenous Smad signaling Using the model system of HL-60 cells, a human myeloblastic leukemia with promyelocytic features, the interplay of signals from all trans retinoic acid (ATRA), which specifies differentiation to granulocytes, or TGFb1/Vitamin D3 (Vit D3), which specify commitment to monocytic differentiation, is mediated, in part, through a balance between protein serine/threonine phosphatase activity and levels of phosphorylated Smad2 and Smad3 (Cao et al., 2003). Thus, Vit D3/TGF-b1 induces phosphorylation of Smad2/3 and that addition of ATRA together with TGF-b1 reduces the level of phospho-Smad2/3 and the extent of monocytic differentiation. Conversely, okadaic acid, which inhibits protein serine/threonine phosphatases and which enhances the level of phospho-Smad2/3 in cells treated simultaneously with ATRA and TGF-b, pushes the Oncogene


5758

balance toward monocytic differentiation. Together these data suggest that monocytic differentiation is favored by lower protein phosphatase activity and/or high levels of nuclear Smad2/3 (if the inducing agents are TGF-b or Vit D3), and that granulocytic differentiation is favored by higher protein phosphatase activity and/or reduced nuclear Smad2/3. In the case where ATRA and either TGF-b or Vit D3 are acting on the cell simultaneously, the induction of protein serine/threonine phosphatase activity by ATRA can modulate the levels of phospho-Smad2/3 induced by TGF-b and thereby control the partitioning between the granulocytic and monocytic pathways. Most germane to our findings is the report that ATRA elicits a transient and reversible interconversion of the protein phosphatase 2A (PP2A) holoenzyme at the G1/S boundary during ATRA-induced granulocytic differentiation of HL-60 cells (Zhu et al., 1997). PP2A accounts for the majority of the serine/threonine phosphatases activity in most cells and is specifically inhibited by low concentrations of okadaic acid (Sontag 2001). Although several other studies show downregulation of the catalytic subunit of PP2A beginning about 48 h after treatment with ATRA and continuing for 3–5 days, it is probably the transient changes in the regulatory subunit at 18–24 h, which are predicted to result in a change in substrate specificity, which are most likely to affect levels of phospho-Smads at the times we observed. In HL-60R cells, which are resistant to effects of ATRA on granulocytic differentiation, cytosolic PP2A but not PP1 activity is reduced by almost 50% compared to wild-type HL-60 cells (Giehl et al., 2000). Consistent with our hypothesis that the ability of ATRA to reduce levels of phospho-Smad2/3 induced by TGF-b1 treatment may depend, in part, on alterations in phosphatase activity, ATRA is unable to decrease levels of TGF-b -induced phospho-Smad2/3 in either mutant HL-60R cells or in wild-type HL-60 cells treated with okadiac acid . The ability of okadaic acid to induce both phenotypic and functional attributes of monocytes, even in the absence of ATRA, further suggests that reduction in the levels of phosphatases, independent of Smad phosphorylation, is sufficient to specify differentiation to monocytes. Vit D3 has been shown to induce an autocrine TGF-b pathway in several different cell types. In U937 cells, treatment with Vit D3 induces differentiation to CD14positive cells with phagocytic capacity and this has been shown to result from activation of an autocrine TGF-b pathway (Defacque et al., 1999). In HL-60 cells, the Vit D3 analog, EB1089, has been shown to induce expression of both TGF-b receptors and TGF-b ligand, and its antiproliferative activity is blocked by a TGF-b neutralizing antibody (Omay et al., 1995). Our data now extend these studies and show that in HL-60 cells, the ability of Vit D3 to phosphorylate Smad2/3 and to stimulate monocytic differentiation can both be blocked by neutralizing antibodies to TGF-b, suggesting that Vit D3 acts indirectly by activating signaling from either autocrine or paracrine (exogenous) TGF-b in these cells. Moreover, these data also show that reduction in levels Oncogene

of Smad2/3 phosphorylation is sufficient to reduce the commitment of these cells to differentiate to monocytes, even in the absence of changes in the levels of phosphatases. Whereas treatment with TGF-b alone results in arrest of differentiation at the CD14-negative promonocyte stage, Vit D3 can induce HL-60 cells to express CD14 and differentiate to mature monocytes, an effect that has been shown to be dependent on induction of phosphatidylinositol 3-kinase activity (Danielpour, 1996). These additional effects of Vit D3 are therefore probably independent of TGF-b, or possibly dependent on synergistic interaction of the vitamin D receptor with Smad3 to regulate expression of certain target genes containing both Vit D3 response elements and Smadbinding elements as previously reported (Hmama et al., 1999). Thus, cellular levels of phosphatase activity and of phosphorylated Smad2/3 induced by TGF-b can independently affect the commitment to differentiation. However, in the particular context of treatment of cells simultaneously with ATRA and TGF-b, these two mechanisms are inter-related in that elevation of phosphatase activity by ATRA appears to underlie the decrease in the level of Smad2/3 phosphorylation (Figure 2). It remains to be demonstrated whether this unique mode of integration of signals from ATRA and TGF-b will also be relevant for other myeloid leukemia cells.

Induction of differentiation of erythroleukemia cells occurs by mediating crosstalk between the Smad and MAPK pathways A distinct role for TGF-b and activin in erythropoiesis has been reported in studies of primary and transformed cells. However, these studies were not fully informative about the signaling networks that mediate TGF-b/ activin effects. The importance of crosstalk between receptor-activated Smad signaling and the MAPK pathways in the regulation of erythroid differentiation induced by TGF-b/activin has been studied (Akel et al., submitted). Treatment of cells with TGF-b/activin resulted in inhibition of cell growth and led to erythroid differentiation as evidenced by a significantly increased proportion of Hb-containing cells. Changes in cell fate were preceded by cytokine/receptor-mediated intracellular signaling, which involved activation of the receptor-activated Smads and various cascades of MAPK, ERK, p38 and JNK. Signaling was rapid suggesting that this process might be directly related to receptor signaling and not to transcriptional activation. Direct activation of Smad2/3 by TGF-b type I receptors is well established and recent studies have described links between MKK4/JNK and MKK3/p38 activation and TGF-bR signaling involving XIAP, HPK1 and TAK1 (Engel et al., 1999; Hanafusa et al., 1999). Although no clear link of the Ras-MEK-ERK pathway with TGF-bR signaling has been described yet, active ERK still may negatively or positively modulate


5759

receptor Smad activation and nuclear translocation (de Caestecker et al., 1998; Kretzschmar et al., 1999). Since inhibition of TGF-b type I receptors abrogated TGF-b/ activin-induced activation of ERK, p38 and JNK MAPKs in K562 cells, all these MAPK pathways are probably linked to TGF-b type I receptors. Our results are in accordance with the recent findings of DaCosta Byfield et al. (2004) showing that SB505124, an inhibitor of TGF-bRI signaling interferes with TGF-b-induced activation of MAPKs. In K562 cells that represent human leukemic hematopoietic progenitor cells (HPC), we have explored the role of the TGF-b/activin-induced signaling in mediating cell growth arrest. TGF-b/ activin-induced growth arrest was reversed in cells pretreated with the inhibitor of TGF-b/activin type I receptors and to a lesser extent with a p38 MAPK inhibitor, but not with ERK inhibitors. Recently, it was found that activation of p38 mediates growth inhibition of normal HPC by TGF-b (Verma et al., 2002), thus the suppressive effect of TGF-b in normal and leukemic human HPC seems to involve activation of Smad2/3 and p38, but not ERK. Cooperation between these two pathways was studied in EPO-independent erythroid differentiation. Two categories of Epo-independent erythroid differentiation were evaluated: cytokine induced (TGF-b/activin) and chemically induced (HU and butyrate). Cytokineinduced erythroid differentiation was dependent on the activation of p38. Either selective inhibition of p38 by SB203580 or coinhibition of phosphorylation of Smad2/ 3 and p38 MAPK by SB505124 was sufficient to prevent the formation of glycophorin A þ Hb þ cells. Since it was not possible to achieve selective inhibition of phosphorylation of Smad2/3 by SB505124 without inhibiting p38 MAPK in cells treated with TGF-b/activin, it remains uncertain whether activation of Smad2/3 is a prerequisite for TGF-b/activin-induced erythrodifferentiation. Unfortunately, we found that ectopic stable expression of Smad7, an inhibitor of Smad2/3 phosphorylation, also prevented phosphorylation of p38 MAPK in response to TGF-b/activin treatment. Okadaic acid-induced Smad2/3 and p38 activation in K562 cells was coincident with the induction of erythroid differentiation and promotion of differentiation induced by various other agents. It was found that a selective inhibition by SB505124 of phosphorylation of Smad2/3 but not of p38 MAPK in OA-treated cells was sufficient to prevent OA-induced differentiation. This indicates that activation of p38 MAPK is subthreshold for differentiation and activation of Smad2/3 is essential for erythroid differentiation. HU and butyrates induce cytodifferentiation and growth inhibition of a variety of tumor cells and have been used in cytoreductive and differentiation therapy of malignant disease (Sowa and Sakai, 2000; Tsimberidou et al., 2002). The ability of these chemicals to induce changes in gene expression is mediated in part by changes in signal transduction pathways. Butyrate caused sustained activation of JAK/STAT signaling in murine erythroleukemia cells, moreover, erythodifferentiation of K562 cells by HU and butyrate involved the

phosphorylation of p38 and dephosphorylation of ERK and JNK MAPKs (Witt et al., 2000; Park et al., 2001). Smad activation is also involved in the regulation of HU- and butyrate-induced erythroid differentiation of K562 cells. Although no direct binding was described between HU and butyrate and TGF-b receptors, results of SB505124 indicate that both chemicals activate the serine/threonine TGF-bRI kinase upstream of Smad2/3. It remains unclear whether these chemicals directly or indirectly (through autocrine TGF-b) regulate receptor/ smad Signaling. Attempts to block ligand induction by anti-TGF-b antibody did not prevent HU/butyrateinduced erythroid differentiation. Given that these agents induce phosphorylation of Smads so quickly, it is possible that they, like OA, block phosphatase activity which increases the strength of basal endogenous signaling by autocrine TGF-b/activin in K562 cells leading to erythrodifferentiation. In erythro-megakaryocytic differentiation, several reports demonstrate that ERK negatively regulates erythroid differentiation, and that sustained activation of this MAPK is sufficient to induce a differentiation program along the megakaryocytic lineage (Radtke et al., 2004). Consistent with these reports, we have shown that ERK1/2 is transiently activated by TGF-b/ activin but reduced below basal level during erythroid differentiation. Moreover, inhibitors of ERK1/2 enhanced erythroid differentiation in the absence and the presence of TGF-b/activin stimulation. Similar synergism was reported between ERK inhibitors, and HU and butyrate, showing that ERK negatively regulates this Epo-independent hemoglobin synthesis (Racke et al., 1997). Surprising, ERK inhibitors induced Smad2/3 phosphorylation, which was dependent on TGF-bRI signaling. Similar results were found in HL-60 and TF-1 cells, suggesting that crosstalk between the Smad and the ERK pathways operate in hematopoietic cells and can modulate cell differentiation. Yang et al. (2003) showed that inhibition of ERK in mouse embryo fibroblasts potentiates Smad signaling and activation of Smad-dependent gene targets. In a recent study, ERK inhibition increased both the basal and TGF-b-induced Smad7 promoter activity in rat fibroblasts (Uchida et al., 1993). Based on the findings of our group and others, this unexplained effect might be related to the fact that inhibition of ERK augmented Smad2/3 signaling, which by itself led to increased Smad7 transcripts as part of the negative feedback loop of TGF-b/Smad signaling. Regardless of where in the TGF-b signaling pathway they act, OA, ERK inhibitors and chemical inducers all act to change the intracellular signaling balance in favor of differentiation. The biological effect of HU and butyrate on erythrodifferentiation, growth inhibition, cell cycle arrest at G1 phase and stimulation of fetal hemoglobin synthesis shares some similarity with those of TGF-b supporting the idea that these agents may mediate their effects through common signal transduction pathways. Collectively, our data suggest that simultaneous activation of Smad2/3 and p38 is required during EPO-independent erythroid differentiation. Oncogene


5760

Levels of Smad7 regulate the signaling of Smad2/3 and MAPKs and control erythroid and megakaryocytic differentiation of erythroleukemia cells During the last decade, mechanisms underlying the complexity of the multiple TGF-b responses and their dependence on the type of target cells and environment began to be more apparent. Regulation of TGF-b/Smad signaling by inhibitory Smads (I-Smads) and the crosstalk with other signaling pathways undoubtedly explained in part the pleiotropic effects of the TGF-b family (de Caestecker et al., 1998; Kretzschmar et al., 1999). Recently, it was shown that smad 7 can direct cell fate decisions in hematopoietic cells (Chadwick et al., 2003). I-Smads (Smad7 and Smad6) directly regulate receptor-activated Smads. Increased expression of I-Smads inhibits TGF-b, activin and BMP signaling (Nakao et al., 1997; Afrakhte et al., 1998). Modulation of Smad and MAPK signaling pathways and cell responses by Smad7 were investigated in erythroleukemia cell lines. In one previous study by Kitamura et al. (2000), it was shown that Smad7 is specifically absent in erythroleukemia cells that differentiate in response to activin. Herein, RT–PCR results revealed that Smad7 transcripts are present in all tested cell lines, albeit at low levels. Disruption of endogenous Smad7 expression in these cells by siRNA significantly improved cell differentiation to physiological doses of TGF-b and activin. This indicates that although Smad7 is expressed at low levels in erythroleukemia cells, it still can act as a physiological inhibitor for TGF-b/activin-induced erythrodifferentiation. However, in the presence of high ligand stimulation, activation of R-Smads escaped the inhibition by endogenous Smad7. The cellular response, Smad signaling and expression of physiologically relevant endogenous Smad7 emphasize that TGF-b/ activin signaling is intact in erythroleukemia cells. To examine whether the balance between TbRI/Smad activation and levels of Smad7 regulates cell responses to TGF-b/activin, K562 cells were stably transfected with plasmid containing Smad7. Cells overexpressing Smad7 became resistant to TGF-b/activin-induced Smad2/3 phosphorylation, erythrodifferentiation and growth inhibition. The inhibitory activity of Smad7 might result from a direct association with ligandactivated type I TGF-b receptors that interfere with receptor phosphorylation of Smad2/3. However, it is still possible that Smad7 can form a complex with receptor-activated Smads and compete with Smad4 for oligomer formation and thus prevent movement of R-Smads into the nucleus as has been observed with Smad6 (Massague, 1998, 2000). The regulation of erythroid and megakaryocytic differentiation by possible crosstalk between the Smad and MAPK pathways was studied (Rouyez et al., 1997; Herrera et al., 1998). The overexpression of Smad7 blocked the activation of different MAPKs. It is likely that Smad7 may modulate certain cell responses by its ability to block a specific signaling pathway. One of these pathways is the p38 MAPK, which is known to mediate different actions of TGF-b and activin on cell Oncogene

growth, differentiation and apoptosis. Hemoglobinzation of K562 cells by TGF-b/activin was prevented using either an inhibitor of p38 (SB203580) or an inhibitor of TbRI (SB505124) (see last section). The profound inhibitory effect of Smad7 on the activation of Smad2/ 3 and p38 provides further evidence about the involvement of these two signaling cascades in TGF-b/activininduced erythroid differentiation. The regulatory effect of Smad7 on p38 and erythroid differentiation is not limited to TGF-b/activin stimulation. Induction of p38 activation and erythrodifferentiation using HU and butyrate occurred in control K562 cells but not in Smad7-transfected K562 cells. Collectively, these data suggest that Smad7 may modulate the kinase activity of p38 irrespective of the stimulus that triggers this p38 activation. However, it is not fully understood how Smad7 inhibited p38 activation and other MAPK proteins. Limited recent reports showed a link between Smad7 and MAPK proteins. Ectopically expressed Smad7 enhanced the coimmunoprecipitation of HAMKK3 and flag-tagged p38, suggesting possible direct interaction between Smad7 and p38. Moreover, Smad7 has been shown to modulate apoptosis of certain cells independent of TGF-b signaling but mediated by the JNK pathway (Mazars et al., 2001). Next, the impact of Smad7 expression on megakaryocytic differentiation was studied. K562/7 cells did not have any basal activity of Smad2 and exhibited an increase in cell size and nuclear lobulation compared to control cells. Moreover, megakaryocytes generated from K562/7 cells by TPO and PMA underwent enhanced endomitosis and exhibited more nuclear lobulation than control cells. This may be explained by the ability of Smad7 to counteract the inhibitory signals induced by autocrine TGF-b and other cytokines. The interference of Smad7 with R-Smad activation reversed TGF-binduced growth arrest and enhanced megakaryocyte polyploidy, suggesting that Smad signaling mediates TGF-b-induced inhibition of megakaryocyte proliferation and terminal differentiation. Thus, modulation of Smad7 expression could provide a basis for abrogating the inhibitory effects of TGF-b on megakaryopoiesis. Since activation of ERK MAPK mediates megakaryocytic differentiation of transformed cell lines and enhances endomitosis in cultures of primary cells (Rojnuckarin et al., 1995), we looked at whether ERK mediates effects of Smad7 on terminal megakaryopoiesis. Expression and activation of ERK was not enhanced in relation with high Smad7 levels; moreover, inhibition of ERK signaling did not reduce polyploidization of K562/7 cells. Our results favor the idea that Smad7 regulates megakaryopoieis independent of ERK. Using the K562/7 model, we are currently investigating the molecular basis of the Smad7 effect on megakaryocyte endomitosis (Kuter et al., 1992; Kalina et al., 2001). Acknowledgements This project has been funded in whole or in part with Federal funds from the National Cancer Institute, National Institutes of Health, under Contract No. NO1-CO-56000.


5761 References Afrakhte M, Moren A, Jossan S, Itoh S, Sampath K, Westermark B, Heldin C, Heldin N and ten Dijke P. (1998). Biochem. Biophys. Res. Commun., 249, 505–511. Akel S, Petrow-Sadowski C, Laughlin MJ and Ruscetti F. (2003). Stem Cells, 21, 557–567. Algietta M, Stacchini A, Severino A, Sanavio F, Ferrando M and Piacibello W. (1989). Exp. Hematol., 17, 296–299. Arai F, Hirao H, Ohmura M, Takubo K, Ito K, Koh G and Suda T. (2004). Cell, 118, 149–161. Bartelmez S, Iversen P and Ruscetti F. (2000). Blood, 96, 474A. Batard P, Monier M, Fortunel N, Ducos K, Sansilvestri-Morel P, Phan P, Hatzfeld A and Hatzfeld JA. (2000). J. Cell Sci., 113, 383–390. Bertoncello I and Williams B. (2004). Methods Mol. Biol., 263, 181–200. Bhatia M, Bonnet D, Wu D, Murdoch B, Wrana J, Gallacher L and Dick JE. (1999). J. Exp. Med., 189, 1139–1148. Birchenall-Roberts M, Fu T, Bang O, Dumbach M, Resau J, Sadowski C, Bertolette D, Lee H, Kim S and Ruscetti F. (2005). J. Biol. Chem., 279, 25605–25615. Borkowski TA, Letterio JJ, Mackall CL, Saitoh A, Wang X, Roop D, Gress R and Udley MC. (1997). J. Clin. Invest., 100, 575–581. Bottinger EP, Letterio JJ and Roberts AB. (1997). Kidney Int., 51, 1355–1360. Bruno E, Horrigan SK, Van Den Berg D, Rozler E, Fitting PR, Moss ST, Westbrook C and Hoffman R. (1998). Blood, 91, 1917–1923. Calvi L, Adams G, Weibrecht K, Weber J, Olson D, Knight M, Martin R, Schipani E, Divieti P, Bringhurst F, Milner L, Kronenberg H and Scadden D. (2003). Nature, 425, 841–846. Cao Z, Flanders K, Bertolette D, Lyakh L, Wurthner J, Parks WT, Letterio JJ, Ruscetti F and Roberts AB. (2003). Blood, 101, 498–507. Cashman JD, Eaves AC, Raines EW, Ross R and Eaves CJ. (1990). Blood, 75, 96–101. Chadwick K, Wang L, Li L, Menendez P, Murdoch B, Rouleau A and Bhatia M. (2003). Blood, 102, 906–915. Chen CZ, Li L, Li M and Lodish HF. (2003). Immunity, 19, 525–533. Chen CZ, Li M, de Graaf D, Monti S, Gottgens B, Sanchez MJ, Lander ES, Golub TR, Green AR and Lodish HF. (2002). Proc. Natl. Acad. Sci. USA, 99, 15468–15473. Cheng T, Rodrigues N, Shen H, Yang Y, Dombkowski D, Sykes M and Scadden DT. (2000a). Science, 287, 1804–1808. Cheng T, Shen H, Rodrigues N, Dombkowski D, Stier S and Scadden DT. (2000b). Nat. Med., 6, 1235–1240. Cheshier SH, Morrison SJ, Liao X and Weissman IL. (1999). Proc. Natl. Acad. Sci. USA, 96, 3120–3125. Christopherson KW, Hangoc G, Mantel CR and Broxmeyer H. (2004). Science, 305, 1000–1003. DaCosta Byfield S, Major C, Laping N and Roberts AB. (2004). Mol. Pharmacol., 65, 744–752. Danielpour D. (1996). J. Cell. Physiol., 166, 231–239. Dao MA, Taylor N and Nolta JA. (1998). Proc. Natl. Acad. Sci. USA, 95, 13006–13011. de Caestecker MP, Parks WT, Frank CJ, Castagnino P, Bottaro DP, Roberts AB and Lechleider RJ. (1998). Genes Dev., 12, 1587–1592. Defacque H, Piquemal D, Basset A, Marti J and Commes T. (1999). J. Cell. Physiol., 178, 109–119. Dexter TM. (1979). Acta Haematol., 62, 299–305. Dexter TM, Garland J, Scott D, Scolnick E and Metcalf D. (1980). J. Exp. Med., 152, 1036–1044.

Dickson M, Martin J, Cousins F, Kulkarni A, Karlsson S and Akhurst RJ. (1995). Development, 121, 1845–1854. Dubois CM, Ruscetti FW, Palaszynski EW, Falk LA, Oppenheim JJ and Keller JR. (1990). J. Exp. Med., 172, 737–744. Dubois CM, Ruscetti FW, Stankova J and Keller JR. (1994). Blood, 83, 3138–3145. Dybedal I, Guan F, Borge OJ, Veiby OP, Ramsfjell V, Nagata S and Jacobsen SE. (1997). Blood, 90, 3395–3403. Dybedal I and Jacobsen SE. (1995). Blood, 86, 949–957. Eaves CJ, Cashman JD, Kay RJ, Dougherty GJ, Otsuda T, Gaboury LA, Hogge DE, Landsdorp P, Eaves AC and Humphries K. (1991). Blood, 78, 110–117. Engel M, McDonnell M, Law B and Moses HL. (1999). J. Biol. Chem., 274, 37413–37420. Fan X, Valdimarsdottir G, Larsson J, Brun A, Magnusson M, Jacobsen SE, ten Dijke P and Karlsson S. (2002). J. Immunol., 168, 755–762. Fortunel N, Batard P, Hatzfeld A, Monier MN, Panterne B, Lebkowski J and Hatzfeld J. (1998). J. Cell Sci., 111, 1867–1875. Fortunel NO, Hatzfeld A and Hatzfeld JA. (2000). Blood, 96, 2022–2036. Giehl K, Seidel B, Gierschik P, Adler G and Menke A. (2000). Oncogene, 19, 4531–4541. Goumans M, Valdimarsdottir G, Itoh S, Rosendahl A, Sideras P and ten Dijke P. (2002). EMBO J, 21, 1743–1753. Grady WM, Myeroff LL, Swinler SE, Rajput A, Thiagalingam S, Lutterbaugh JD, Neumann A, Brattain MG, Chang J, Kim SJ, Kinzler KW, Vogelstein B, Willson JK and Markowitz S. (1999). Cancer Res., 59, 320–324. Greenberger JS, Eckner RJ, Sakakeeny M, Marks P, Reid D, Nabel G, Hapel A, Ihle JN and Humphries KC. (1983). Fed. Proc., 42, 2672–2681. Hampson J, Ponting ILO, Cook N, Vodinelich L, Redmond S, Roberts AB and Dexter TM. (1989). Growth factors, 1, 193–202. Hanafusa H, Ninomiya-Tsuji J, Masuyama N, Nishita M, Fujisawa J, Shibuya H, Matsumoto K and Nishida E. (1999). J. Biol. Chem., 274, 27161–27167. Hatzfeld A, Batard P, Panterne B, Taieb F and Hatzfeld J. (1996). Hum. Gene. Ther., 7, 207–213. Hatzfeld J, Batard P, Cardoso A, Li M, Panterne B, Sansilvestri P, Ginsbourg M, Leveesque J and Hatzfeld A. (1994). Blood Cells, 20, 430–434. Hatzfeld J, Li ML, Brown EL, Sookdeo H, Levesque JP, O’Toole T, Gurney C, Clark SC and Hatzfeld A. (1991). J. Exp. Med., 174, 925–929. Herrera R, Hubbell S, Decker S and Pertuzzeli L. (1998). Exp. Cell. Res., 228, 407–414. Hmama Z, Nandsan D, Sly L, Knutson K, Herrera-Velit P and Reiner N. (1999). J. Exp. Med., 190, 1583–1594. Hu X and Zuckerman KS. (2001). J. Hematother. Stem Cell Res., 10, 67–74. Imbert AM, Bagnis C, Galindo R, Chabannon C and Mannoni P. (1998). Exp. Hematol., 26, 374–381. Jacobsen SE, Keller J, Ruscetti F, Kondaiah P, Roberts A and Falk L. (1991a). Blood, 78, 2239–2247. Jacobsen SE, Ruscetti F, Dubois C, Lee J, Boone T and Keller J. (1991b). Blood, 77, 1706–1716. Jacobsen SE, Ruscetti FW, Dubois CM and Keller JR. (1992). J. Exp. Med., 175, 1759–1772. Jacobsen SE, Ruscetti FW, Ortiz M, Gooya J and Keller JR. (1994). Exp. Hematol., 22, 985–989. Oncogene


5762 Jakubowiak A, Pouponnot C, Berguido F, Frank RMS, Massague J and Nimer SD. (2000). J. Biol. Chem., 275, 40282–40287. Kalina U, Koschieder S, Hofmann WK, Wagner S, Kauschat D and Ottmann OG. (2001). Exp. Hematol., 29, 602–608. Keller JR, Mantel C, Sing GK, Ellingsworth LR, Ruscetti SK and Ruscetti FW. (1988). J. Exp. Med., 168, 737–750. Keller JR, McNiece IK, Sill KT, Ellingsworth LR, Quesenberry PJ, Sing GK and Ruscetti FW. (1990). Blood, 75, 596–602. Keller JR, Ortiz M and Ruscetti FW. (1995). Blood, 86, 1757–1764. Kim SJ and Letterio J. (2003). Leukemia, 17, 1731–1737. Kitamura K, Aota S, Sakamoto R, Yoshikawa S and Okazaki K. (2000). Blood, 95, 3371–3379. Kretzschmar M, Doody J, Timokhina I and Massague J. (1999). Genes Dev., 13, 804–816. Kulkarni AB, Huh CG, Becker D, Geiser A, Lyght M, Flanders KC, Roberts AB, Sporn MB, Ward JM and Karlsson S. (1993). Proc. Natl. Acad. Sci. USA, 90, 770–774. Kurokawa M, Mitani K, Irie K, Matsuyama T, Takahashi T, Chiba S, Yazaki Y, Matsumoto K and Hirai H. (1998). Nature, 394, 92–96. Kuter DJ, Gminski DM and Rosenberg RD. (1992). Blood, 79, 619–626. Langer JC, Henckaerts E, Orenstein J and Snoeck HW. (2004). J. Exp. Med., 199, 5–14. Larsson J, Blank U, Helgadottir H, Bjornsson JM, Ehinger M, Goumans MJ, Fan X, Leveen P and Karlsson S. (2003). Blood, 102, 3129–3135. Larsson J, Blank U, Klintman J, Magnusson M and Karlsson S. (2005). Exp. Hematol., 33, 592–596. Larsson J, Goumans MJ, Sjostrand LJ, van Rooijen MA, Ward D, Leveen P, Xu X, ten Dijke P, Mummery CL and Karlsson S. (2001). EMBO J., 20, 1663–1673. Laughlin MJ. (2001). Bone Marrow Transplant, 27, 1–6. Letterio J, Geiser A, Kulkarni A, Dang H, Kong L, Nakabayashi T, Mackall C, Gress R and Roberts AB. (1998). J. Clin. Invest., 98, 2109–2119. Letterio JJ, Geiser AG, Kulkarni AB, Roche NS, Sporn MB and Roberts AB. (1994). Science, 264, 1936–1938. Leveen P, Larsson J, Ehinger M, Cilio CM, Sundler M, Sjostrand LJ, Holmdahl R and Karlsson S. (2002). Blood, 100, 560–568. Lieb E, Vogel T, Milz S, Dauner M and Schulz M. (2004). Tissue Eng., 10, 1414–1425. Lord B. (1990). Int. J. Cell Cloning, 8, 317–331. Lu L, Xiao M, Grigsby S, Wang W, Wu B, Shen R and Broxmeyer H. (1993). Exp. Hematol., 21, 1442–1446. Mancini S, Mantel N, Dumortier A, Suter U, MacDonald HR and Radtke F. (2005). Blood, 105, 2340–2342. Massague J. (1998). Annu. Rev. Biochem., 67, 753–791. Massague J. (2000). Nat. Rev. Mol. Cell Biol., 1, 169–178. Mazars A, Lallemand F, Prunier C, Marais J, Ferrand N, Pessah M, Cherqui G and Atfi A. (2001). J. Biol. Chem., 276, 36797–36803. Moerman E, Teng K, Lipschitz D and Lecka-Czernik B. (2004). Aging Cell, 3, 379–389. Nakao A, Afrakhte M, Moren A, Nakayama T, Christian JL, Heuchel R, Itoh S, Kawabata M, Heldin NE, Heldin CH and ten Dijke P. (1997). Nature, 389, 631–635. Nara N, Tohda S, Nagata K, Suzuki T and Yamashita Y. (1989). Leukemia, 3, 572–577. Nilsson SK, Johnston HM and Coverdale JA. (2001). Blood, 97, 2293–2299. Nolta JA and Kohn DB. (1990). Hum. Gene Ther., 1, 257–268. Oncogene

Ogawa M. (1993). Blood, 81, 2844–2853. Oh SP, Seki T, Goss KA, Imamura T, Yi Y, Donahoe P, Li L, Miyazono K and ten Dijke P. (2000). Proc. Natl. Acad. Sci. USA, 97, 2626–2631. Ohta M, Greenberger JS, Anklesaria P, Bassols A and Massague J. (1987). Nature, 329, 539–541. Omay S, Nishikawa M, Morita K, Toyada H, Nakai K, Shima H, Nagao M, Kitagawa S, Saito M and Shiku H. (1995). Exp. Hematol., 23, 244–251. Oshima M, Oshima H and Taketo MM. (1996). Dev. Biol., 179, 297–302. Ottmann OG and Pelus LM. (1988). J. Immunol., 140, 2661–2665. Park IK, Morrison SJ and Clarke MF. (2004). J. Clin. Invest., 113, 175–179. Park J, Choi H, Jeong J, Han J and Kim I. (2001). Cell Growth Differ., 12, 481–486. Pierelli L, Marone M, Bonanno G, Mozzetti S, Rutella S, Morosetti R, Rumi C, Mancuso S, Leone G and Scambia G. (2000). Blood, 95, 3001–3009. Ploemacher RE, van Soest PL and Boudewijn A. (1993). Stem Cells, 11, 336–347. Quesenberry PJ and Becker PS. (1998). Proc. Natl. Acad. Sci. USA, 95, 15155–15156. Racke F, Lewandowska K, Goueli S and Goldfarb A. (1997). J. Biol. Chem., 272, 23366–23370. Radtke F, Wilson A, Mancini S and MacDonald H. (2004). Nat. Immunol., 5, 247–253. Remy I, Montmarquette A and Michnick S. (2004). Nat. Cell Biol., 6, 358–365. Rojnuckarin P, Drachman JG and Kaushansky K. (1995). Blood, 85, 402–413. Rooke HM, Vitas MR, Crosier PS and Crosier K. (1999). Leukemia, 13, 535–541. Rouyez MC, Boucheron C, Gisselbrecht S, Dusanter-Fourt I and Porteu F. (1997). Mol. Cell. Biol., 17, 4991–5000. Ruscetti F, Sitnicka E, Letterio J, Walz T and Bartelmez SH. (1999). Blood, 94, 466A. Ruscetti FW and Bartelmez SH. (2001). Int. J. Hematol., 74, 18–25. Sakamaki S, Hirayama Y, Matsunaga T, Matsunaga T, Kuroda H, Kusakabe T, Akiyama T, Konuma Y, Sasaki K, Tsuji N, Okamoto T, Kobune M, Kogawa K, Kato J, Takimoto R, Koyama R and Niitsu Y. (1999). Blood, 94, 1961–1970. Schiemann WP, Pfeifer WM, Levi E, Kadin ME and Lodish HF. (1999). Blood, 94, 2854–2861. Shi Y and Massague J. (2003). Cell, 113, 685–700. Shull MM, Ormsby I, Kier AB, Pawlowski S, Diebold RJ, Yin M, Allen R, Sidman C, Proetzel G, Calvin D, Annunziata N and Doetschman T. (1992). Nature, 359, 693–699. Sing GK, Keller JR, Ellingsworth LR and Ruscetti FW. (1988). Blood, 72, 1504–1511. Sitnicka E, Ruscetti FW, Priestley GV, Wolf NS and Bartelmez SH. (1996). Blood, 88, 82–88. Soma T, Yu JM and Dunbar CE. (1996). Blood, 87, 4561–4567. Sontag E. (2001). Cell Signal, 13, 7–16. Sorrentino BP. (2004). Nat. Rev. Immunol., 4, 878–888. Sowa Y and Sakai T. (2000). Biofactors, 12, 283–287. Staal F and Clevers H. (2005). Nat. Rev. Immunol., 5, 21–30. Stenvers K, Tursky M, Harder K, Kountouri N, AmatayakulChantler S, Grail D, Small C, Weinberg R, Sizeland A and Zhu H. (2003). Mol. Cell. Biol., 23, 4371–4385. Stier S, Cheng T, Dombkowski D, Carlesso N and Scadden DT. (2002). Blood, 99, 2369–2378.


5763 Stier S, Cheng T, Forkert R, Lutz C, Dombkowski DM, Zhang JL and Scadden DT. (2003). Blood, 102, 1260–1266. Tessier N and Hoang T. (1988). Blood, 72, 159–164. Tsimberidou AM, Alvarado A and Giles FJ. (2002). Expert Rev. Anticancer Ther., 2, 437–448. Uchida T, Matozaki T, Matsuda K, Suzuki T, Matozaki S, Nakano O, Wada K, Konda Y, Sakamoto C and Kasuga M. (1993). J. Biol. Chem., 268, 11845–11850. Verma A, Deb KD, Sassano A, Uddin S, Varga J, Wickrema A and Platinias L. (2002). J. Biol. Chem., 277, 7726–7735. Waegell WO, Higley HR, Kincade PW and Dasch JR. (1994). Exp. Hematol., 22, 1051–1057. Wakefield L and Roberts A. (2002). Curr. Opin. Genet Dev., 12, 22–29. Wiesmann A, Kim M, Georgelas A, Searles A, Cooper D, Green W and Spangrude GJ. (2000). Exp. Hematol., 28, 128–139. Wilkes M, Murphy S, Garamszegi N and Leof E. (2003). Mol. Cell. Biol., 23, 8878–8889. Witt O, Sand K and Pekrun A. (2000). Blood, 95, 2391–2396. Wolf NS, Kone A, Priestley G and Bartelmez SH. (1993). Exp. Hematol., 21, 614–622. Wolfraim LA, Fernandez TM, Mamura M, Fuller WL, Kumar R, Cole DE, Byfield S, Felici A, Flanders KC, Walz TM, Roberts AB, Aplan PD, Balis FM and Letterio JJ. (2004). N. Engl. J. Med., 351, 552–559.

Yang L, Bryder D, Adolfsson J, Nygren J, Mansson R, Sigvardsson M and Jacobsen SE. (2005a). Blood, 105, 2717–2723. Yang L, Dybedal I, Bryder D, Nilsson L, Sitnicka E, Sasaki Y and Jacobsen SE. (2005b). J. Immunol., 174, 752–757. Yang X, Castilla LH, Xu X, Li C, Gotay J, Weinstein M, Liu PP and Deng CX. (1999a). Development, 126, 1571–1580. Yang X, Letterio JJ, Lechleider RJ, Chen L, Hayman R, Gu H, Roberts AB and Deng C. (1999b). EMBO J., 18, 1280–1291. Yang YC, Pick E, Zavadil J, Liang D, Xie D, Heyer J, Pavidis P, Kucherlapati R, Roberts AB and Bottinger E. (2003). Proc. Natl. Acad. Sci. USA, 100, 10269–10274. Yaswen L, Kulkarni AB, Fredrickson T, Mittleman B, Schiffman R, Payne S, Longenecker G, Mozes E and Karlsson S. (1996). Blood, 87, 1439–1445. Yu J, Soma T, Hanazono Y and Dunbar CE. (1998). Gene Therapy, 5, 1265–1271. Yue J and Mulder K. (2001). Pharmacol. ther., 91, 1–34. Zhang J, Niu C, Ye L, Huang H, He X, Tong WG, Ross J, Haug J, Johnson T, Feng JQ, Harris S, Wiedemann LM, Mishina Y and Li L. (2003). Nature, 425, 836–841. Zhu T, Matsuzawa S, Mizuno Y, Kamibayashi C, Mumby M, Andjelkovic N, Hemmings B, Onoe K and Kikuchi K. (1997). Arch. Biochem. Biophys., 339, 210–217.

Oncogene