Smurf2 regulates hematopoietic stem cell self-renewal and aging

The age-dependent decline in the self-renewal capacity of stem cells plays a critical role in aging, but the precise mechanisms underlying this decline are not well understood. By limiting proliferative capacity, senescence is thought to play an important role in age-dependent decline of stem cell self-renewal, although direct evidence supporting this hypothesis is largely lacking. We have previously identified the E3 ubiquitin ligase Smurf2 as a critical regulator of senescence. In this study, we found that mice deficient in Smurf2 had an expanded hematopoietic stem cell (HSC) compartment in bone marrow under normal homeostatic conditions, and this expansion was associated with enhanced proliferation and reduced quiescence of HSCs. Surprisingly, increased cycling and reduced quiescence of HSCs in Smurf2-deficient mice did not lead to premature exhaustion of stem cells. Instead, HSCs in aged Smurf2-deficient mice had a significantly better repopulating capacity than aged wild-type HSCs, suggesting that decline in HSC function with age is Smurf2 dependent. Furthermore, Smurf2-deficient HSCs exhibited elevated long-term self-renewal capacity and diminished exhaustion in serial transplantation. As we found that the expression of Smurf2 was increased with age and in response to regenerative stress during serial transplantation, our findings suggest that Smurf2 plays an important role in regulating HSC self-renewal and aging.


Introduction
Aging is a complex process of gradual deterioration of physiological functions that involves virtually all cells and tissues. Tissue-specific stem cells, which are capable of self-renewal to preserve stem cell pools as well as differentiation into a variety of effector cells, play a central role in the normal homeostatic maintenance and regenerative repair of tissues throughout the lifetime of an organism. With advancing age, the selfrenewal capacity of stem cells invariably declines, eventually leading to the accumulation of unrepaired, damaged tissues in old organisms (Rando, 2006;Sharpless & DePinho, 2007;Rossi et al., 2008). Aging of the hematopoietic system is driven by many intrinsic and extrinsic factors that adversely affect the production and functions of blood cells (Van Zant & Liang, 2003;Rossi et al., 2008). Increasing evidence suggests that attenuated hematopoietic stem cell (HSC) function plays an important role in hematopoietic aging, including anemia, decreased immune function, and increased propensity for myeloid malignancies. The ability of HSCs to repopulate and reconstitute lethally irradiated recipients is limited in serial transplantation, reflecting a finite potential of HSC self-renewal under regenerative stress (Ogden & Mickliem, 1976;Harrison & Astle, 1982). In aged mice, HSCs undergo distinct changes: They exhibit decreased repopulating ability, even though the number of phenotypic HSCs based on staining of cell surface markers increases with age in some laboratory mouse strains. Further, aged HSCs show a differentiation preference toward myeloid lineages at the expense of lymphoid differentiation (Morrison et al., 1996;Sudo et al., 2000;Liang et al., 2005;Rossi et al., 2005;. In humans, aged HSCs also show increased frequency and a bias toward myeloid potential (Pang et al., 2011).
The molecular mechanisms responsible for age-dependent decline in HSC self-renewal are complex and many factors have been implicated (Van Zant & Liang, 2003;Rossi et al., 2008). Senescence, a state of irreversible growth arrest that most cells enter at the end of their replicative lifespan, has been proposed to contribute to aging by reducing the renewal capacity of stem cells (Sharpless & DePinho, 2004;Campisi, 2005). Although direct evidence supporting the senescence theory of aging is lacking, several senescence regulators including p16 INK4a (p16) and p53 have been found to play important roles in aging. The expression of p16 increases with age in many human and rodent tissues (Krishnamurthy et al., 2004;Ressler et al., 2006). Age-associated increase in p16 in mice coincides with a decline in the renewal capacity of stem cells in bone marrow, brain, and pancreas (Janzen et al., 2006;Krishnamurthy et al., 2006;Molofsky et al., 2006), although p16 up-regulation in aged HSCs has been challenged . HSCs in old mice lacking p16 have increased regenerative potential, suggesting that p16 plays a critical role in limiting HSC selfrenewal (Janzen et al., 2006). In p53-null mice, the number of HSCs increases and they perform better in competitive repopulation, suggesting an enhanced self-renewal capacity (TeKippe et al., 2003;Liu et al., 2009). Moreover, mice with one wild-type allele and one mutant allele of p53 that lacks the N-terminal transactivation domain maintain cancer protection, but age prematurely including impairment of HSCs (Tyner et al., 2002;Maier et al., 2004;Dumble et al., 2007). These studies support an emerging link between senescence regulation and aging and show the potential importance of senescence regulation in the aging of HSCs.
We have recently shown that the E3 ubiquitin ligase Smurf2 plays a critical role in regulating the senescence response in human and mouse cells (Zhang & Cohen, 2004;Kong et al., 2011;Ramkumar et al., 2012). Adventitious expression of Smurf2 is sufficient to induce senescence in early passage cells (Zhang & Cohen, 2004;Ramkumar et al., 2012), while deficiency in Smurf2 expression impairs the senescence response in culture and in vivo (Kong et al., 2011;Ramkumar et al., 2012). In this study, we found that Smurf2 deficiency led to increased proliferation and an expanded HSC compartment in bone marrow. Surprisingly, increased proliferation did not lead to early HSC exhaustion. Instead, Smurf2-deficient HSCs showed better repopulating ability and multilineage potential than wild-type cells with advancing age or under regenerative stress, suggesting a functional role of Smurf2 in the regulation of HSC self-renewal and aging.
We have generated a Smurf2-deficient mouse model (Smurf2 T/T , T for the trapped allele), in which a gene-trapping cassette is inserted into intron 1 of Smurf2 to disrupt its normal splicing (Ramkumar et al., 2012). The expression of Smurf2 was significantly reduced in total BM and LSK cells of Smurf2-deficient mice compared with wild-type (WT) mice ( Fig. 1). Because of the hypomorphic nature of the trapped Smurf2 allele, there were residual normal splicing and Smurf2 expression in BM, LSK cells (Fig. 1), common lymphoid progenitors, multipotent progenitors, and HSCs ( Fig. S1A) of Smurf2-deficient mice, similar to what we have found previously in other tissues (Ramkumar et al., 2012). Reduced Smurf2 expression in this mouse model allowed us to investigate whether Smurf2 has a functional role in HSC aging.
In aged (24-to 30-month) mice, we also found an increase in BM cellularity in Smurf2-deficient mice (1.59-fold, P = 0.0007) compared with WT mice (Fig. 2B). Similar to what we found in young mice, the frequencies of LT-HSCs, ST-HSCs, MPPs, and LSK population were not significantly different between aged WT and Smurf2-deficient mice (Fig. 2E). Although the average numbers of LT-HSCs, ST-HSCs, MPPs, and LSK populations increased (1.73-2.21-fold) in aged Smurf2-deficient mice, there were great variations among individual aged mice and the increases were not significantly different from those in WT mice ( Fig. 2 and Fig. S2B).
Complete blood count analyses of peripheral blood showed no significant difference between age-matched WT and Smurf2-deficient mice at either young (2 months) or old (18-20 months) age in various lineages, including white blood cells, red blood cells, lymphocytes, granulocytes, monocytes, and platelets (Fig. S3). These results suggest that differentiation and commitment of HSCs to various lineages are not affected by Smurf2 deficiency.

Enhanced cell proliferation of LT-HSCs in Smurf2-deficient mice
The observed increases in BM cells and LT-HSCs in Smurf2-deficient mice suggest an enhanced cell proliferation of these cells. Supporting this notion, we found that the fraction of BM cells in the S phase of cell cycle was increased in 2-month-old Smurf2-deficient mice compared with WT mice ( Fig. S4A; 11.9 AE 0.7% in Smurf2-deficient vs. 8.1 AE 0.4% in WT, P = 0.0012). To further investigate whether cell proliferation is enhanced in Smurf2-deficient LT-HSCs and progenitor populations, we gave young mice 4 injections of bromodeoxyuridine (BrdU) in a span of 24 h to label LT-HSCs (Passegue et al., 2005). BrdU was incorporated in a significant portion of LT-HSCs (Lin À Sca1 + c-kit ++ CD150 + ) using this labeling method (Fig. 3). We found that BrdU incorporation was significantly increased in Smurf2-deficient LT-HSCs (1.8-fold, P = 0.019) compared with WT LT-HSCs (Fig. 3). Increased BrdU incorporation was also observed in Smurf2deficient LSK and ST-HSC/MPP (Lin À Sca1 + c-kit ++ CD150 À ) populations compared with WT cells (Fig. S4B).
Collectively, these data indicate that cell proliferation is enhanced in Smurf2-deficient LT-HSCs.
Enhanced repopulation ability of HSCs in aged Smurf2-deficient mice As quiescence has been postulated to prevent premature stem cell exhaustion during aging (Orford & Scadden, 2008), our observation of decreased frequency of quiescent LT-HSCs in Smurf2-deficient mice suggests that Smurf2-deficient HSCs may experience premature exhaustion. To determine the self-renewal ability and multilineage potential of HSCs, we used competitive repopulation (Harrison, 1980), in which donor BM cells of either WT or Smurf2-deficient mice (CD45.2 + ) were mixed at a 1:1 ratio with competitor BM cells of 2-month-old CD45.1 + / CD45.2 + WT mice and injected into lethally irradiated congenic recipient mice (CD45.1 + ). Contribution of donor, competitor, and recipient in peripheral blood and BM of reconstituted recipients was determined by flow cytometry at different time post-transplantation. As we have found previously that~30% of Smurf2-deficient mice develop tumor spontaneously after 12 month of age, and~70% of tumors are B-cell lymphomas in spleen (Ramkumar et al., 2012), only mice that were tumor free were used in this study. We selected aged mice that were physically fit without any sign of stress or illness and did not have visible tumors upon necropsy. Furthermore, we used flow cytometry to characterize spleen from these aged mice to make sure that there was no lymphoma as described (Ramkumar et al., 2013). Smurf2-deficient HSCs of young (2-month) mice were more efficient in repopulating B and myeloid cells than young WT HSCs and as efficient as WT HSCs in repopulating T cells in the peripheral blood of the reconstituted recipient mice (Fig. 4). Interestingly, Smurf2-deficient HSCs of aged (24-month) animals showed significantly better repopulation ability than the agematched WT HSCs (Fig. 4), despite the decreased quiescence and increased cycling of Smurf2-deficient LT-HSCs. In fact, old Smurf2deficient HSCs had similar repopulating ability as the young HSCs (Fig. 4). Furthermore, analysis of BM and LSK population 6 months after transplantation showed increased contribution from aged Smurf2deficient donor compared with WT donor (Fig. 4). Collectively, these results indicate that HSCs in aged Smurf2-deficient mice have better repopulation ability and suggest that Smurf2 deficiency protects HSCs from early exhaustion during aging.

Smurf2-deficient LT-HSCs exhibit better long-term self-renewal capacity under regenerative stress
To investigate the long-term self-renewal and multilineage potential of HSCs, we used BM serial transplantation, in which LT-HSCs in donor BM are able to reconstitute lethally irradiated recipients in successive but limited transplantations, reflecting the finite potential of HSC selfrenewal (Ogden & Mickliem, 1976;Harrison & Astle, 1982). As the LSK (Lin À Sca1 + c-kit ++ ) population is gated and displayed for staining of CD150 and Flt3 (right panels). LT-HSCs are defined as Lin À Sca1 + c-kit ++ CD150 + Flt3 À , while Lin À Sca1 + c-kit ++ CD150 À Flt3 À and Lin À Sca1 + c-kit ++ CD150 À Flt3 + are used to enumerate short-term HSCs (ST-HSCs) and multipotent progenitors ( frequency of LT-HSCs in Smurf2-deficient mice was not significantly different from WT mice, we used the same number of BM cells from 2-month-old WT and Smurf2-deficient mice in transplantation. We used either female or male donors in two independent experiments. During the 1st, 2nd, or 3rd transplantation, 100% of recipient mice receiving donor BM from WT or Smurf2-deficient mice survived. Recipient mice from the 3rd transplantation were followed for a year and none of them died during observation. Staining of BM cells in the recipient mice during these successive transplantations revealed that >75% of the LT-HSC population had been reconstituted by the donor (Fig. S5A). Three months after transplantation, complete blood count analysis of peripheral blood of recipient mice showed that various blood cell lineages were successfully reconstituted by both the WT and Smurf2-deficient donors (Fig. S5B). During the 4th transplantation, all recipients receiving BM from WT mice died within 3 weeks post-transplantation. In contrast, recipients receiving BM from Smurf2-deficient mice showed significantly better survival ( Fig. 5; 63-67% survival, P < 0.001), suggesting that Smurf2deficient HSCs have an enhanced long-term self-renewal capacity. None of these recipient mice died during 1 year of observation. We carried out a 5th transplantation with Smurf2-deficient donors and found that 60% of the recipient mice survived for 1 year post-transplantation without signs of illness or stress (Fig. S6).
Similarly, we carried out BM serial transplantation using aged mice as donors. All recipients receiving donor BM from either aged WT or Smurf2-deficient mice survived during the first 3 cycles of transplantation. Complete blood count analysis of recipient peripheral blood 3 months after transplantation showed that various blood cell lineages were successfully reconstituted by both the WT and Smurf2-deficient donors (Fig. S5C). During the 4th transplantation, all recipients receiving BM from WT mice died within 3 weeks post-transplantation, while 100% recipients receiving BM from Smurf2-deficient mice survived more than 6 months post-transplantation ( Fig. 5; P < 0.0001). Collectively, these results indicate that Smurf2-deficient LT-HSCs have a better capacity in the long-term self-renewal and multilineage potential than the WT LT-HSCs. Furthermore, this stem cell-intrinsic functional enhancement is maintained in aged Smurf2-deficient mice.

Smurf2 regulates p16 expression during aging and under regenerative stress
Enhanced long-term self-renewal of Smurf2-deficient HSCs despite increased cycling and decreased quiescence in these cells prompted us to examine the expression of cyclin-dependent kinase inhibitors, as several of them have been found to regulate HSC self-renewal. HSCs lacking p18 INK4c (p18) have increased self-renewal ability accompanied by increased cycling (Yuan et al., 2004), and p16 has been shown to limit HSC self-renewal during aging (Janzen et al., 2006). We found that p18 expression was not changed, while the expression of p16 and p19 Arf (p19), which are located in the same INK4a locus, was increased in BM of aged (24-month) WT C57BL/6 mice compared with young (2-month) mice (Fig. 6), consistent with a previous report (Krishnamurthy et al., 2004). Interestingly, the expression of p16 and p18 was increased in LSK cells with age, whereas p19 expression was undetectable in LSK cells (Fig. 6). As our recent studies show that Smurf2 regulates the expression of p16 during senescence (Kong et al., 2011), we found that p16 expression was significantly suppressed in BM and LSK cells in Smurf2deficient mice (Fig. 6). In contrast, the expression of p19 showed a slight decrease and p18 was largely unchanged in Smurf2-deficient cells (Fig. 6).
We then analyzed the expression of these genes as well as Smurf2 in BM of reconstituted recipients during BM serial transplantation. We found that Smurf2 expression was increased during successive transplantation in mice receiving WT donor BM, concomitant with an increase in the expression of p16, p18, or p19 (Fig. 6). While the increase in p18 and p19 expression was not consistently affected by Smurf2 deficiency, the increase in p16 expression was greatly attenuated in mice receiving the Smurf2-deficient donor BM (Fig. 6), suggesting a possible link between p16 and Smurf2-mediated regulation of HSC self-renewal and aging.

Discussion
An important characteristic of HSCs in adult animals is their relative quiescence (Cheshier et al., 1999;Mahmud et al., 2001). It is thought  that quiescence prevents premature stem cell exhaustion during aging (Orford & Scadden, 2008). Many quiescence regulators have been found to control stem cell aging. For example, cyclin-dependent kinase inhibitor 1A (CDKN1A or p21) is required to maintain quiescence of HSCs. In the absence of p21, increased cell cycling leads to premature exhaustion and impaired self-renewal of HSCs (Cheng et al., 2000). Conversely, deletion of transcriptional factor ELF4, a negative regulator of quiescence, leads to increased quiescence and reduced exhaustion of HSCs (Lacorazza et al., 2006). In this study, we found that Smurf2 deficiency led to enhanced cell proliferation in BM. HSCs in Smurf2deficient mice exhibited increased cell cycling and decreased quiescence. Surprisingly, increased proliferation of HSCs in Smurf2-deficient mice did not lead to premature stem cell exhaustion. Instead, Smurf2-deficient HSCs showed increased long-term self-renewal and multilineage repopulating capacity under regenerative stress in serial transplantation compared with WT HSCs. Furthermore, HSCs in aged Smurf2-deficient mice displayed better repopulating capability than HSCs in aged WT mice. Taken together, our results indicate that Smurf2 deficiency mitigates age-dependent decline in HSC self-renewal and function despite increased HSC cycling. We found that Smurf2 expression was increased in BM and LSK population enriched for HSCs with advancing age or in response to regenerative stress in serial transplantation, concomitant with increases in the expression of p16, p19, and p18. HSCs lacking p18 have increased long-term self-renewal capacity, even though these cells show increased proliferation (Yuan et al., 2004). Our finding that p18 expression was unchanged by Smurf2 deficiency suggests that p18 is unlikely responsible for Smurf2-regulated decline in HSC self-renewal with age. It is documented that p16 expression increases with age in many human and rodent tissues, including BM and HSCs (Krishnamurthy et al., 2004;Janzen et al., 2006;Ressler et al., 2006), although age-dependent increase in p16 in HSCs has been challenged . In old mice, p16 À/À HSCs perform much better than their WT counterparts in serial transplantation, suggesting that age-dependent increase in p16 expression limits HSC self-renewal (Janzen et al., 2006). We have recently found that Smurf2 regulates p16 expression during senescence (Kong et al., 2011). In Smurf2-deficient mice, age or regenerative stressinduced p16 expression was significantly attenuated, suggesting a possible link between HSC self-renewal and p16 expression regulated by Smurf2. Supporting this notion, we found that Smurf2-deficient HSCs of aged mice substantially outperformed aged WT HSCs in competitive repopulation and serial transplantation. In fact, old Smurf2-deficient HSCs performed as well as young HSCs, a phenotype similarly observed in old p16 À/À HSCs (Janzen et al., 2006). However, Smurf2-deficient HSCs in young mice behaved differently from young p16 À/À HSCs, which are more readily depleted in serial transplantation than WT HSCs (Janzen et al., 2006). In contrast, HSCs in young Smurf2-deficient mice also outperformed WT HSCs in serial transplantation and competitive repopulation. It is possible that Smurf2 limits HSC self-renewal in a p16dependent manner only in aged mice, or Smurf2 deficiency protects agedependent decline in HSC self-renewal independently of p16. Additional Hematopoietic stem cells (HSCs) in aged Smurf2-deficient mice exhibit enhanced repopulating ability. Donor bone marrow (BM) from wild-type (+/+) or Smurf2 T/T (T/T) mice (CD45.2 + ) mixed at 1:1 ratio with competitor BM of 2-monthold wild-type (CD45.1 + /CD45.2 + ) mice are injected into lethally irradiated recipient mice (CD45.1 + ). Contribution of donors is determined by CD45 staining. Both young (2-month) and old (24-month) donors were used. (A) Donor (CD45.2 + ) contributions to B (B220 + ), T (CD3 + ), and myeloid (CD11b + ) lineages in peripheral blood of recipient mice (N = 4) at indicated time post-BM transplantation. Twoway ANOVA is used for statistical analysis. *P < 0.05, **P < 0.01, ****P < 0.0001, ns: not significant. (B) Donor contribution to BM and LSK (Lin À Sca1 + c-kit ++ ) cells in recipient mice (N = 3) at 6 months post-BM transplantation. Error bars are SD. Student's t-test is used in statistical analysis.
studies are needed to address the precise role of p16 in Smurf2mediated regulation of HSC self-renewal and aging.
In complex organisms such as mammals, many somatic tissues such as bone marrow are capable of renewal, repair, and even regeneration. Renewable tissues allow the replacement of damaged cells, offering a clear advantage over postmitotic tissues. It is important for stem cells to maintain lifetime self-renewal and proliferation. In this study, we found that Smurf2 expression was increased with advancing age in BM and LSK cells. More importantly, HSCs in old Smurf2-deficient mice maintain their self-renewal capacity at the level similar to young HSCs, suggesting that functional decline in HSC self-renewal during aging is Smurf2 dependent. We have previously shown that Smurf2 is an important regulator of senescence (Zhang & Cohen, 2004;Kong et al., 2011;Ramkumar et al., 2012). Increased expression of Smurf2 is sufficient to induce senescence (Zhang & Cohen, 2004;Ramkumar et al., 2012), while loss of Smurf2 expression impairs the senescence response in culture and in vivo (Kong et al., 2011;Ramkumar et al., 2012). Senescence is proposed to contribute to aging by depleting the renewal capacity of tissues and/or by interfering with tissue homeostasis and functions (Sharpless & DePinho, 2004;Campisi, 2005). Although cells with characteristics of senescence have been detected to accumulate with age in vivo, their functional contribution to aging is still not completely clear (Sharpless & DePinho, 2004;Campisi, 2005). Our study provides evidence for Smurf2 in regulating HSC self-renewal and aging, further strengthening the emerging link between senescence and stem cell aging.

Experimental procedures
Smurf2-deficient mice Smurf2-deficient mice as described previously (Ramkumar et al., 2012) had been backcrossed to C57BL/6 (CD45.2 + ) for more than 10 generations. All mouse studies were carried out according to guidelines approved by the Institutional Animal Care and Use Committee of University of Massachusetts Medical School.

Complete blood count
Peripheral blood was bled from mouse tail veins into EDTA-coated tubes, and complete blood count analysis was performed using the Heska CBC-Diff Veterinary Hematology System (Heska, Loveland, CO, USA).

Cell cycle and cell proliferation analyses
Mice were injected with 1 mg BrdU intraperitoneally every 6 h for 24 h as described previously (Passegue et al., 2005). BM cells were stained for HSCs as described above, fixed and permeabilized using the BrdU flow kit (BD Biosciences). These cells were then stained with anti-BrdU-FITC or anti-Ki-67-PerCp-Cy5.5 (Sola15) antibody for 20 min. Cells stained for Ki-67 were further incubated with DAPI for 30 min at room temperature as described (Yang et al., 2008). Stained cells were analyzed by flow cytometry as described above.

Statistical analysis
Data were presented as mean AE SD. Two-tailed and unpaired Student's t-test was used for pairwise comparisons. Two-way ANOVA was used for multiple comparisons. Kaplan-Meier survival curves were plotted using GraphPad Prism 5.0, and statistical significance was analyzed using the log-rank test. P < 0.05 was considered as statistically significant.

Supporting Information
Additional Supporting Information may be found in the online version of this article at the publisher's web-site.