Introduction, results, and discussion
Although HSCs can function throughout and even beyond the lifetime of an individual (Harrison & Astle, 1982), these cells are not spared from the aging process. Within cells, LNK (SH2B3) acts to dampen multiple extracellular signaling pathways. In its absence, multilineage hematopoiesis is exaggerated (Takaki et al., 2002) and includes strikingly elevated HSC numbers in young mice (Ema et al., 2005). Murine aging is also associated with an enhancement in the numbers of candidate HSCs, although their functionality is significantly reduced (Beerman et al., 2010).
We began to investigate hematopoiesis in aged (21–24 months) LNK−/− mice. Analysis of hematopoietic progenitors by fluorescence-activated cell sorting (FACS) (Pronk et al., 2007) (Fig. 1A) revealed a clear reduction in the bone marrow (BM) frequency of CFU-Es (2.1-fold), a 1.5-fold increase in MkPs, and an even greater (2.5-fold) increase in HSC frequency in aged LNK−/− mice (Fig. 1B).
Competitive in vivo reconstitution experiments were next conducted to assess the intrinsic impact of LNK deficiency on aging HSC function. Sixteen weeks post-transplantation, a 3.4-fold increase in overall chimerism was observed in recipients of aged LNK−/− HSCs (Fig. 2A). In contrast to aged WT HSCs, which reconstituted the lymphoid lineages poorly with a skewed ratio toward myelopoiesis (Fig. 2A), aged LNK−/− HSCs displayed high chimerism in all lineages.
There might be an intrinsic limitation in the potential of aged LNK−/− HSCs to self-renew that can only be provoked by further stress such as serial transplantation (Harrison et al., 1978). Serially transplanted recipients of aged WT HSCs displayed drastically lower total reconstitution at 16 weeks, and lymphoid levels were never significant in these hosts (Fig. 2B). By contrast, recipients of serially transplanted LNK−/− HSCs displayed high overall multilineage chimerism (Fig. 2B). Therefore, while serial transplantation compromised aged WT HSCs function, aged LNK−/− HSCs appear insensitive to such effect.
We next examined the transcriptome of aged WT or aged LNK−/− HSCs by gene expression microarray analysis. This identified 95 genes that were differentially expressed more than 1.5-fold (Table S1). Several highly significant and overlapping biological processes were found to associate with genes down-regulated in aged LNK−/− HSCs and indicated that aged LNK−/− HSCs are characterized by the repression of genes involved in regulating cell cycle progression (Fig. 2C).
Because telomere length has been proposed as a critical regulator of longevity (Monaghan, 2010), we hypothesized that altered cell cycling of LNK−/− HSCs might cause alterations in telomere length. We established an assay that allowed telomere length estimations to be performed at a clonal level (Fig. 2D, left panel). However, when single aged WT and aged LNK−/− HSCs isolated from primary recipient hosts were analyzed, we could not detect differences in telomere length (Fig. 2D, right panel).
To investigate whether LNK−/− HSCs displayed altered proliferation in vivo, we applied a noninvasive biotin-based method (Nygren & Bryder, 2008). We here utilized mice previously transplanted with aged WT or LNK−/− HSCs (Fig. S1). As expected, these analyses revealed that HSCs displayed the highest biotin intensities, regardless of genotype, when compared to GMLPs, CD150-positive myeloid progenitors destined for megakaryocytic or erythroid fates (MPCD150+), and CD150-negative myeloid progenitors (MPCD150−) (Fig. S1A). However, no difference was observed in biotin dilution when comparing aged WT or LNK−/− HSCs, with the majority of cells (WT: 57 ± 18%, LNK−/−: 59 ± 9%) having divided only once during the chase period (Fig. S1B). Similarly, we could not detect significant differences between the two genotypes in the proliferative activity of other progenitor compartments, although there was a trend toward more cycling of MPCD150− cells (Fig. S1B). Therefore, in the setting of transplantation, LNK deficiency does not appear to lead to an increased proliferation of HSCs.
In contrast to young LNK−/− mice where HSC numbers are elevated 10- to 20-fold (Ema et al., 2005), candidate, HSC numbers are increased only about 2.5-fold in aged LNK−/− mice (Fig. 1B). In the context of the expansion of HSCs pool size with normal aging (Beerman et al., 2010), such a ‘roof’ for HSC numbers appears to be driven by environmental cues (Krosl et al., 2003). LNK has previously been shown to negatively regulate several cytokine pathways such as c-kit (Takaki et al., 2002), thrombopoietin (TPO) (Tong & Lodish, 2004), and erythropoietin (EPO) (Tong et al., 2005), with at least the former two being well-established positive regulators of HSC function. Furthermore, Suzuki et al., (2012) recently proposed that LNK deficiency results in increased survival of HSCs without an effect on the lineage distribution of their cellular output, akin to what has previously been described for BCL2 transgenic mice (Domen & Weissman, 2000). It appears reasonable that a higher proliferative competence of downstream progenitor cells in LNK deficient mice can provide an explanation for the reconstitution phenotype. For instance, increased proliferation of B cell–restricted progenitors has been noted previously in the absence of LNK (Takaki et al., 2000).
Aged LNK−/− HSCs are far superior to aged WT HSCs in their abilities to repopulate both primary and secondary hosts. Additionally, aged WT HSCs displayed a lineage output profoundly skewed toward myelopoiesis, a phenomenon that has been attributed to the expansion of a specific subset of HSCs clones (Beerman et al., 2010), contrasting the balanced lineage program exhibited by aged LNK−/− HSCs (Fig. 2A, B). It remains to be seen whether knockdown of LNK in aged WT HSCs would lead to a similar repopulating potential upon transplantation as that of aged LNK−/− HSCs that have never ‘seen’ LNK throughout their ontogeny. This could reveal whether LNK deficiency protects HSCs clones with a balanced lineage potential from exhaustion or maintains the multipotentiality of individual HSCs clones.
Although we find no evidence for differential telomeric maintenance, our gene expression profiling indicates that aged LNK−/− HSCs as a cell pool are less proliferative in steady state compared to aged WT HSCs. This was a rather surprising finding not only because it has previously been demonstrated that LNK−/− HSCs proliferate more rapidly in response to cytokine signaling in vitro (Seita et al., 2007), but also because aged HSCs are less proliferative compared to their young counterparts (Attema et al., 2009). Thus, the aged LNK−/− HSC pool can be regarded as ‘hyper-quiescent’, although it is important to consider that any HSC compartment is heterogeneous, at least in terms of their proliferative status (Nygren & Bryder, 2008). Thus, when translating the number of cycling HSCs into actual numbers and not merely as fraction of the total HSC pool, at least as many HSCs should be cycling in aging LNK−/− animals as in WT mice in steady state.
Although it is possible that aged LNK−/− HSCs proliferate less than aged WT HSCs in steady state, perhaps because of an altered microenvironment, the question remains how to reconcile the gene transcription profile obtained from aged LNK−/− HSCs and the proliferation kinetics observed upon transplantation. While our analyses established many cell cycle–associated genes to be enriched in the comparison between aged WT and aged LNK−/− HSCs, a different scenario arises when the expression levels of these genes are put in context to other hematopoietic progenitors (Fig. S2). The expression levels observed in aged HSCs of both genotypes are on average lower when compared to young HSCs and other more proliferative hematopoietic progenitors (Passegue et al., 2005). Therefore, aged WT and aged LNK−/− HSCs resemble each other more than any of the other progenitor cell subsets analyzed (Fig. S2). Importantly, however, our data demonstrate that despite the extensive repopulating capacity displayed by aged LNK−/− HSC, LNK deficiency does not associate with increased proliferation of these cells.
In summary, the results presented here unequivocally demonstrate that LNK deficiency can functionally counteract most if not all of the hematopoietic stem cell-repopulating defects that associate with normal chronological aging.