Alcam Regulates Long-term Hematopoietic Stem Cell Engraftment and Self-Renewal

Abstract

Hematopoietic stem cells (HSCs) reside in a specialized bone marrow (BM) microenvironment that supports the maintenance and functional integrity of long-term (LT)-HSCs throughout postnatal life. The objective of this work is to study the role of activated leukocyte cell adhesion molecule (Alcam) in HSC differentiation and self-renewal using an Alcam-null (Alcam^−/−) mouse model. We show here that Alcam is differentially regulated in adult hematopoiesis and is highly expressed in LT-HSCs where its level progressively increases with age. Young adult Alcam^−/− mice had normal homeostatic hematopoiesis, and normal numbers of phenotypic HSCs. However, Alcam^−/− HSCs had reduced long-term replating capacity in vitro and reduced long-term engraftment potential upon transplantation. We show that Alcam^−/− BM contain a markedly lower frequency of long-term repopulating cells than wild type (WT). Further, the long-term repopulating potential and engraftment efficiency of Alcam^−/− LT-HSCs was greatly compromised despite a progressive increase in phenotypic LT-HSC numbers during long-term serial transplantation. In addition, an age-associated increase in phenotypic LT-HSC cellularity was observed in Alcam^−/− mice. This increase was predominately within the CD150^hi fraction, and was accompanied by significantly reduced leukocyte output. Consistent with an aging-like phenotype, older Alcam^−/− LT-HSCs display myeloid-biased repopulation activity upon transplantation. Finally, Alcam^−/− LT-HSCs display premature elevation of age-associated gene expression, including Selp, Clu, Cdc42, and Foxo3. Together, this study indicates that Alcam regulates functional integrity and self-renewal of LT-HSCs.

Introduction

Hematopoietic stem cells (HSCs) are capable of long-term self-renewal and multipotent differentiation to sustain life-long production of mature blood cells of all lineages. Adult long-term repopulating HSCs reside in specialized BM niches along the endosteum and in perivascular sites adjacent to the endothelium [1–3]. The cell fate decisions of HSCs are controlled by cell intrinsic determinants, as well as cell extrinsic cues provided by the specialized niches. The intricate balance between HSC dormancy and activation, self-renewal and differentiation is critical in maintaining the functional integrity of HSCs throughout postnatal life. As with most tissues, aging of the hematopoietic system is associated with many clinically significant conditions, including diminution of the adaptive immune system, increased incidence of myeloid malignancies, and greater propensity for anemia [4–7]. Consistent with these clinical manifestations, functional studies of HSCs from older mice indicate a decline in stem cell capacity, attenuated lymphoid lineage output and increased myeloid potential with age [8,9]. These age-related alterations of HSCs, however, are associated with an increased frequency of cells that meet the most stringent phenotypic definitions of the HSC population in murine and human BM [10–13]. Furthermore, cell intrinsic alteration, such as skewing in lineage-associated gene expression and epigenetic remodeling, is associated with HSC aging [9,11]. Additional microenvironmental mechanisms that regulate HSC self-renewal and lineage specification may also contribute to aging of the hematopoietic system.

Activated leukocyte cell adhesion molecule (ALCAM or CD166) is a cell surface immunoglobulin superfamily member that mediates homophilic adhesion and heterotypic interactions with CD6 [14,15]. Alcam reportedly functions in biological processes as diverse as axon pathfinding and fasciculation [16,17], vascular-angiogenesis [18], migration of neurons and activated monocytes [19] and trafficking of leukocytes through the blood-brain barrier into the central nervous system [20]. Aberrant ALCAM expression has been associated with many cancer types with variable prognostic values [21–26]. Notably, ALCAM has been reported to mark a tumor-initiating cell population in prostate cancer and in lung cancer [27,28]. In the hematopoietic system, ALCAM is expressed on activated lymphocytes and monocytes, and on primitive human HSCs [29,30]. In addition, Alcam is reportedly required for myeloid colony formation in avian embryonic hematopoietic progenitors [31], and Alcam-expressing endothelial cells can support murine embryonic hematopoiesis [18]. Engagement of ALCAM and CD6 is essential for dendritic cell-mediated T-cell stimulation [32], and ALCAM gene silencing was recently shown to be required for megakaryocytic differentiation [33]. Another recent study showed that an Alcam⁺/Sca1⁻ subset of the endosteal niche cell population robustly supports HSC activity and exhibits differential expression of cell adhesion-related genes, suggesting that these cells regulate HSC function through cell adhesion [34]. Moreover, earlier studies of murine HSC aging suggested that Alcam expression is up-regulated several fold in aged HSCs compared to young HSCs [9,11]. Based on these observations, we hypothesized that Alcam might regulate adult HSC function related to age.

In the study described herein, we comprehensively investigated the role of Alcam in adult hematopoiesis and HSC function using an Alcam-null mouse model. A series of phenotypic and functional studies reveal that Alcam is important in ensuring long-term engraftment and the self-renewal of LT-HSCs during transplantation and aging.

Materials and Methods

Mice

Alcam-null allele [17] backcrossed to C57Bl6 for more than 8 generations were a generous gift from Joshua Weiner (University of Iowa). All mice were genotyped by PCR using primers listed in Supplementary Table 1. Two-month-old (young) or 12–15 month-old mice were used for experiments. All mice were maintained in an AAALAC-accredited animal facility at City of Hope, and all experimental procedures involving mice were performed in accordance with federal and state government guidelines and established institutional guidelines and protocols approved by the Institutional Animal Care and Use Committee at Beckman Research Institute of City of Hope.

Flow cytometry and cell sorting

Femurs and tibias were crushed gently with a mortar and pestle to dissociate the BM fraction. Cells were resuspended in 5 mL PBS/0.5% BSA and filtered through a 70-µm filter (BD Biosciences). The following antibodies were used for fluorescence activated cell-sorting (FACS) analyses: CD117 (C-kit, clone 2B8, eBioscience), Ly-6A/E (Sca-1, clone E13–161.7, BioLegend), CD150 (SLAM, clone TC15-12F12.2, BioLegend), CD48 (clone HM48.1, BioLegend), CD34 (clone RAM34, eBioscience), CD16/CD32 (clone 93, eBioscience), CD127 (IL7R, clone A7R34, eBioscience), CD135 (Flk-2, Flt-3, Ly-72, clone A2F10, eBioscience), CD45.1 (clone A20, BioLegend), CD45.2 (clone 104, BioLegend), CD11b (clone M1/70, eBioscience), Ter119 (clone TER-119, BioLegend), CD45R (clone RA3–6B2, BD Biosciences), CD3 (clone 17A2, eBioscience), Ly-6G/Ly-6C (Gr1, clone RB6–8C5, BioLegend) and goat polyclonal anti-mouse ALCAM (CD166, R&D system). The lineage antibody cocktail included the following biotin-conjugated anti-mouse antibodies: CD19 (clone eBio1D3), NK-1.1 (clone PK136), CD45R (clone RA3–6B2), IgM (clone II/41), CD3 (clone 145-2C11), CD4 (clone GK1.5), CD8 (clone 53-6.7), Gr1 (clone RB6–8C5), CD127 (clone A7R34) at 1 µg/mL, CD11b (clone M1/70) at 2 µg/mL, and Ter119 (clone Ter119, from BioLegend) at 3 µg/mL. Secondary reagents used included streptavidin (SA)-PerCP-Cy5.5 (BioLegend), PE-TexasRed (Invitrogen) or V500 (BD Biosciences). Flow cytometry was performed on a 4-laser, 14-detector FACS-LSRII (BD Biosciences). For cell sorting, lineage negative cells were enriched using EasySep lineage depletion reagents (StemCell Technologies) according to the manufacturer’s instructions. Phenotypic populations were defined as LT-HSCs (Lin⁻/ckit^hi/Sca1⁺/Flt3⁻/CD150⁺/CD48⁻), short-term HSCs (ST-HSCs) (Lin⁻/ckit^hi/Sca1⁺/Flt3⁻/CD150⁻/CD48⁻), multipotent progenitors (MPPs) (Lin⁻/ckit^hi/Sca1⁺/CD150^+/−/CD48⁺), lymphoid-primed multipotent progenitors (LMPPs) (Lin⁻/ckit^hi/Sca1⁺/Flt3⁺/CD150^+/−/CD48⁺), common lymphoid progenitors (CLPs) (Lin⁻/IL7R⁺/ckit^lo/Sca1^lo), myeloid progenitors (MPs) (Lin⁻/ckit⁺/Sca1⁻), common myeloid progenitors (CMPs) (Lin⁻/ckit⁺/Sca1⁻/CD34⁺/FcγR^lo), granulocyte-macrophage progenitors (GMPs) (Lin⁻/ckit⁺/Sca1⁻/CD34⁺/FcγR^hi), and megakaryocyte-erythroid progenitors (MEPs) (Lin⁻/ckit⁺/Sca1⁻/CD34⁻/FcγR^lo). Cell sorting was performed on a 4-laser, 15-detector FACSAria-III or a 6-laser, 18-detector FACSAria II SORP (BD Bioscience).

Transplantation experiments

For competitive repopulation experiments, 2 × 10⁵ unfractionated BM mononuclear cells isolated from Alcam^−/− mice or WT littermates (CD45.2⁺) were transplanted intravenously into lethally irradiated (13 Gy) 6- to 8-week-old congenic C57BL/6 mice (CD45.1⁺/CD45.2⁺) together with 2 × 10⁵ CD45.1⁺ unfractionated BM cells. Secondary transplantation was performed similarly using sorted CD45.2⁺ HSCs isolated from primary recipients 16 weeks after transplantation. Limiting dilution transplantation was similarly performed with three donor cell doses (2 × 10⁵, 4 × 10^4, 8 × 10³). For LT-HSC engraftment, 50 purified LT-HSCs from Alcam^−/− mice or WT littermates (CD45.2⁺) were transplanted into lethally irradiated (13 Gy) 6- to 8-week-old CD45.1⁺ mice together with 2 × 10⁵ CD45.1⁺ supportive cells. Engraftment of CD45.2⁺ cells was analyzed over 6 months and transplantation was repeated with 100 purified CD45.2⁺ LT-HSCs.

Quantitative (q)RT-PCR analysis

RNA was isolated from sorted BM cells by using the RNeasy micro kit (Qiagen) according to the manufacturer’s protocol. First-strand cDNA was generated using 200 U SuperScript III reverse transcriptase (Invitrogen) and 0.5 µg oligo dT primer in a 20 µL reaction. Quantitative (q)RT-PCR was performed using LightCycler 480 SYBR Green I master mix (Roche Applied Science) containing 0.2 µM gene-specific primers and detected with a LightCycler 480 real-time PCR system (Roche Applied Science). Primers used are listed in Supplementary Table 1, and relative expression levels were determined by the standard curve method. Alternative method using the TaqMan assay is described in Supplementary Material and Method.

Statistics

Statistical analyses were performed with Student’s t test or analysis of variance (ANOVA) for normal distribution. Mann-Whitney U tests were performed when normal distribution was not satisfied. p value less than 0.05 was considered statistically significant (*p < 0.05; **p < 0.01; ***p < 0.001). Frequency estimation of limiting-dilution analysis was performed based on Poisson distribution using L-Calc (Stem Cell Technologies).

Results

Alcam is highly expressed in LT-HSCs and is progressively up-regulated with age

As a first step toward understanding the function of Alcam in hematopoiesis, we assessed whether Alcam surface expression is differentially regulated in various phenotypically defined subsets of adult hematopoietic stem and progenitor cells (HSPCs) by immunostaining and flow cytometry (Figure 1A). First, we analyzed young (2 month old) mice and found that Alcam was abundantly expressed in greater than 95% of primitive hematopoietic stem and progenitor cells, including phenotypically-defined LT-HSCs, short-term HSCs (ST-HSCs), multipotent progenitors (MPPs) and lymphoid-primed multipotent progenitors (LMPPs) (Figure 1B and C). Alcam expression was differentially regulated amongst myeloid progenitor subsets and common lymphoid progenitors (CLPs) (Figure 1B and C). Overall, granulocyte-macrophage progenitors (GMPs) expressed high levels of Alcam, while megakaryocyte-erythroid progenitors (MEPs) did not express detectable levels, and common myeloid progenitors (CMPs) expressed intermediate levels. The CMP compartment could be divided into two subsets (Alcam⁺ and Alcam⁻) based on Alcam expression (Figure 1B, top). Similar differential Alcam surface expression was observed in HSPC subsets of 12 month-old mice (Figure 1C). Interestingly, Alcam levels on the cell surface were significantly (p= 0.0159) elevated in 12 month-old LT-HSCs compared to those of 2 month-old (Figure 1C). To determine whether Alcam expression is transcriptionally regulated, we analyzed Alcam mRNA levels in sorted LT-HSCs, ST-HSCs, MPPs, CMPs, MEPs, and GMPs by qRT-PCR, and found a similar differential expression pattern as that observed with cell surface staining (Figure 1D). These results indicate that Alcam is differentially regulated at the transcriptional level, and is most highly expressed in the LT-HSC compartment. We also analyzed Alcam mRNA levels in HSPC subsets from young (2 month old), 12 month old and 16 month old mice by qRT-PCR. Similar preferential expression in LT-HSCs is observed in all age groups, and we find a significant (p< 0.0001) age-associated up-regulation of Alcam expression in LT-HSCs (Figure 1E). An approximately 2-fold and 5-fold increase in Alcam levels was detected at 12 months and 16 months, respectively.

Figure 1. Alcam is highly expressed in primitive HSCs and is progressively up-regulated with age.

(A) Representative FACS profile illustrating gating strategies for HSPC subsets. Gray arrows indicate further separation of the gated population. (B) Representative histograms of Alcam staining in MP (top) (CMP: dotted line; GMP: black shading; MEP: black solid line), LMPP (middle), and LSK subsets (bottom)(LT-HSC: black shading; ST-HSC: dotted line; MPP: black solid line) or isotype control (gray shading). (C) Average Alcam median fluoresence intensity +/− SEM (2-month-old: n=5; 12-month-old: n=4) deteremined by FACS. (D) Average relative Alcam mRNA levels in sorted populations, as detected by qRT-PCR (n=3, each performed in duplicate). (E) Relative Alcam mRNA levels in sorted populations from 2-month- and 12-month-and 16-month old WT, determined by qRT-PCR (n=3, performed in duplicate). *p < 0.05; *** p<0.001

Alcam-deficient HSCs display robust myeloid differentiation and short-term repopulation

Since Alcam is highly expressed in HSCs and subsets of progenitor cells, we used an Alcam-null mouse allele [17] to determine the hematopoietic consequences of Alcam-deletion. First, we assessed whether Alcam-heterozygosity or Alcam-deficiency leads to alteration of the hematopoietic profile in the peripheral blood (PB) of young (2 month old) adult mice. Complete blood count analysis showed no significant changes in hematological parameters for WT, Alcam^+/− (HT) or Alcam^−/− (KO) mice other than slightly reduced leukocyte counts (p=0.0498) in Alcam^−/− (KO) (Supplementary Table 2). The frequencies of B cells (B220⁺), T cells (CD3⁺) and myeloid cells (CD11b⁺ or Gr1⁺), as determined by flow cytometry, were not altered in Alcam^+/− or Alcam^−/− mice (Figure 2A). The cellularity and frequencies of various HSPC subsets were similar in Alcam^+/− or Alcam^−/− bone marrow (Figure 2B). The frequencies of phenotypic progenitor subsets, including LSK (Lin⁻/ckit^hi/Sca1⁺), LMPP, CLP, CMP, GMP, and MEP, were similar in Alcam^+/− or Alcam^−/− mice as compared to WT littermates (Figure 2C and 2D). There was no significant difference in the frequencies of LT-HSCs, ST-HSCs and MPPs within the LSK fraction (Figure 2E). Together, these results indicate that Alcam-deficiency does not alter the steady state hematopoietic profile in young adult mice.

Figure 2. Alcam-deficient HSCs and early progenitors have normal differentiation and proliferation but are compromised in serial-replating capacity.

(A) Frequencies of B-cells (B220⁺), T-cells (CD3⁺) and myeloid cells (Gr1⁺ or CD11b⁺), in the peripheral blood, as determined by immunostaining and flow cytometry. (B) Total BM mononuclear cell number isolated from 1 femur and 2 tibias of mice. Frequencies of LSKs, LMPPs, CLPs (C); myeloid progenitor subsets (D); and LT-HSCs, ST-HSCs, and MPPs (E) in young (2-month-old) adult BM. For A–E, bar graphs represent average +/− SEM (WT, n=5; HT, n=5; KO, n=6). (F) Bars and numbers indicate percent of Meg/E, GM, G and Mix colonies identified by single cell differentiation assay (KO, n=273; WT, n=267). (G) Percent of Meg/E, GM, and Mix colonies derived from Alcam⁺ CMPs and Alcam CMPs in CFU-C assay. (H) Colony-forming progenitor (CFU-C) number per 1,000 cells in each sorted population. Bar graphs show representative average +/− SEM of duplicate assays from three independent experiments. (I) CFU-C number per 10,000 cells replated over 3 rounds of weekly successive replating (P2, P3, P4). Bar graphs show representative average +/− SEM of duplicate assays from three independent experiments. Bar graphs represent average +/− SEM (n=3). *** p<0.001; ns, not significant.

To assess the consequences of Alcam-deficiency on the differentiation potential of HSCs at high resolution, we used a single-cell in vitro differentiation assay [35]. We sorted phenotypic LT-HSCs and plated single cells in erythromyeloid differentiation-inducing media and cytokines (see Supplementary Material and Method). We determined the lineage composition by FACS analysis (c-Kit, CD11b, Gr1, Ter119) and morphological examination of cytospin preparations. Overall, Alcam^−/− (n=273) and WT (n=267) HSCs gave rise to similar phenotypes and numbers of all colony types, including megakaryocyte-erythroid (Meg/E), granulocyte-macrophage (GM), granulocyte (G), and mixed (Mix; contained both Meg/E and GM) (Figure 2F). Although not statistically significant, we found that Alcam^−/− HSCs have enhanced propensity to differentiate into granulocytes (12% in KO compared to 5% in WT), which corresponded to a reduction in immature mixed (Mix) colonies (35% in KO compared to 42% in WT). Since Alcam is expressed in a subset of CMPs, we tested whether they represent functionally distinct progenitors. We performed a colony-forming unit (CFU-C) assay (See Supplementary Material and Method) using sorted Alcam⁺ CMPs and Alcam⁻ CMPs (Supplementary Figure 1A, top). We found that while the total CFU-C numbers are similar (data not shown), Alcam⁺ CMPs are significantly biased towards GM potential (31% for Alcam⁺ v.s. 2% for Alcam⁻) whereas Alcam⁻ CMPs are significantly skewed towards Meg/E lineage (38% for Alcam⁻ v.s. 0% for Alcam⁺) potential (Figure 2G; p<0.0001). In addition, we assessed Alcam surface expression in refined Pre-GM and Pre-MegE subsets within CMP based on additional markers previously described [36] (Supplementary Figure 1A, bottom). Consistent with the functional bias, Alcam expression was high in Pre-GM and low in Pre-MegE (Supplementary Figure 1A, bottom). These results indicate that Alcam expression in myeloid precursors differentially marks cells with a GM lineage potential. We also assessed clonogenic progenitor activity of sorted LT-HSC, MPP and myeloid progenitor (MP) cells. Alcam^−/− LT-HSCs, MPPs, MPs formed similar numbers of colonies as their WT counterparts (Figure 2H). To assess the ability to maintain self-renewing progenitors, every 7 days we serially replated colonies derived from LT-HSCs, and observed a significant reduction in Alcam^−/− CFU-Cs upon replating (Figure 2I; p=0.0002). To examine whether Alcam deficiency leads to altered proliferation of stem and progenitor cells, we performed in vivo EdU incorporation assays (see Supplementary Material and Method). We observed comparable cell cycle distribution for each phenotypic population from Alcam^−/− or WT mice (Supplementary Figure 2A, B, C), suggesting that Alcam deficiency does not alter the proliferation rate. Expression of cell cycle-related genes such as Cdkn2c (p18), Cdkn1b (p27), p53 and Gata2 were similar in Alcam^−/− and WT LT-HSCs (Supplementary Figure 2F, G, H, I).

To further assess the hematopoietic consequences of Alcam deficiency in vivo, we performed a competitive repopulation assay using unfractionated BM cells (Figure 3A). We compared the relative contribution of CD45.2⁺ test donor cells to that of congenic competitor cells (CD45.1⁺) over time. Overall, there was no significant difference in donor chimerism over 16 weeks (Figure 3B). The contribution of peripheral B-cells, T-cells and myeloid cells did not significantly differ between Alcam^−/−, Alcam^+/− and WT donors (Figure 3C), nor did the numbers of donor LT-HSCs (Figure 3D) and progenitor subsets in the BM (16 weeks) significantly differ (Figure 3E). We further assessed the long-term self-renewal potential by secondary-transplantation. Sorted CD45.2⁺ LSK cells (2000 cells) were transplanted together with 2 × 10⁵ CD45.1⁺ radio-protective cells into CD45.1⁺ congenic recipients (n=8), and CD45.2⁺ donor chimerism was analyzed over 16 weeks. Engraftment was considered positive if CD45.2⁺ chimerism in all lineages (B, T, myeloid) was greater than 1%. The number of mice showing multi-lineage engraftment over time was significantly reduced in groups received Alcam^−/− (p=0.0018) compared to WT donor cells (Figure 3F). These results suggest that while Alcam-deficient cells can support short-term repopulation, long-term repopulating activity is compromised.

Figure 3. Alcam-deficient HSCs display robust short-term repopulation.

(A) Schematic representation of competitive repopulation assay design. Unfractionated BM cells (2 × 10⁵) from 2-month-old Alcam^−/− (KO), Alcam^+/− (HT) or WT littermates (CD45.2⁺) were transplanted along with equal numbers of congenic (CD45.1⁺) BM cells as competitors into recipient mice (CD45.1⁺/CD45.2⁺) irradiated at 13 Gy. Donor engraftment in PB was analyzed every 4 weeks, and BM engraftment was assessed at 16 weeks. CD45.2⁺ LSK cells were then sorted and similarly transplanted along with equal numbers of congenic (CD45.1⁺) BM cells into secondary recipients (CD45.1⁺) irradiated at 13 Gy. (B) Time course analysis of CD45.2⁺ to CD45.1⁺ cell ratio in PB of primary competitive repopulation recipients. (C) CD45.2⁺ to CD45.1⁺ ratio in peripheral B-cells (B220⁺), T-cells (CD3⁺) and myeloid cells (Gr1⁺ or CD11b⁺) of primary recipients at 16 weeks. Each dot represents an individual mouse transplanted with WT (n=7), HT (n=8) or KO (n=8) cells. (D) The number of CD45.2⁺ LT-HSCs detected in the BM of primary recipients at 16 weeks. Bar graph represents average +/− SEM (WT, n=4; HT, n=5; KO, n=5). (E) The number of CD45.2⁺ ST-HSC, MPP and LSK cells in the BM at 16 weeks. Bar graph represents average +/− SEM (WT, n=4; HT, n=5; KO, n=5). (F) Percentage of mice engrafted with CD45.2⁺ cells after secondary transplantation (n=8). Engraftment was considered positive if CD45.2⁺ chimerism in all lineages (B, T, myeloid) was greater than 1%. ** p<0.01; ns, not significant.

Alcam deficient BM contains lower frequency of long-term repopulating cells

To determine the frequency of functional long-term repopulating cells in Alcam^−/− BM, we performed a limiting-dilution transplantation experiment using decreasing numbers (2 × 10⁵, 4 × 10⁴, 8 × 10³) of CD45.2⁺ donor cells together with 2 × 10⁵ CD45.1⁺ radio-protective cells (Supplementary Figure 3A). Consistent with the competitive repopulation results, all mice transplanted with 2 × 10⁵ cells showed multi-lineage engraftment up to 16 weeks post-transplant, and there was no significant difference in CD45.2⁺ chimerism between Alcam^−/− and WT donors (Supplementary Figure 3B). Time course analysis showed that overall CD45.2⁺ chimerism in mice transplanted with lower Alcam^−/− donor cell doses was significantly lower (Supplementary Figure 3C and D). We analyzed CD45.2⁺ engraftment, which is considered positive if chimerism is greater than 1% in all lineages (B, T, myeloid) 16 weeks after transplantation. The number of mice showing multi-lineage engraftment in each cell dose is summarized in Table 1. The results show that engraftment of Alcam^−/− cells was significantly (p=0.0036) reduced at lower donor cell doses. Based on the fraction of mice engrafted at each cell number, the estimated frequency of functional long-term repopulating cells in Alcam^−/− mice was significantly lower (1 in 51,885; range 1/35,999-1/74,781) than in WT mice (1 in 15,213; range 1/10,559–1/21,918). These results indicate that Alcam^−/− BM contains approximately three fold lower frequency of long-term repopulating cells.

Table 1.

Estimated long-term repopulating (LTR) cell frequency in Alcam-deficient bone marrow assessed by limiting-dilution transplantation.

Cell number	# Transplanted	# Engrafted	LTR cell frequency
Wild type
200000	7	7
40000	8	8	1 /15.213
8000	8	2	(1/10,559–1/21,918)
Alcam−/−
200000	7	7
40000	8	4	1 /51,885
8000	8	1	(1/35,999–1/74,781)

Alcam-deficient LT-HSCs are compromised in long-term multi-lineage engraftment potential

Based on our phenotypic analysis and estimate of long-term repopulation cell frequency, we predicted that the long-term repopulation potential and self-renewal of Alcam^−/− LT-HSCs could be compromised on a per cell basis. To test this, we transplanted 50 sorted phenotypic LT-HSCs from young Alcam^−/− or WT mice (CD45.2⁺) together with 2 × 10⁵ CD45.1⁺ support cells into lethally-irradiated CD45.1⁺ congenic recipients (n=12) (Figure 4A). Mice that received Alcam^−/− LT-HSCs displayed significantly reduced (p<0.0001) CD45.2⁺ chimerism up to 24 weeks post-transplantation (Figure 4B). Compared to WT transplanted mice, which all showed multi-lineage CD45.2⁺ engraftment, only 8 of 12 Alcam^−/− transplanted mice were CD45.2⁺ engrafted at 24 weeks (Figure 4D, p=0.0079). For the mice that had CD45.2⁺ engraftment at 24 weeks after transplantation, the numbers of Alcam^−/− LT-HSCs in the BM tends to be higher than WT but not statistically different (Figure 4C).

Figure 4. Alcam^−/− LT-HSCs are compromised in long-term multi-lineage engraftment.

(A) Schematic of LT-HSC long-term repopulation assay. LT-HSCs were sorted from WT or KO mice (CD45.2⁺), and 50 cells were transplanted into each CD45.1⁺ congenic recipient (n=12) along with 2 × 10⁵ CD45.1⁺ radio-protective cells. CD45.2⁺ chimerism was assessed over 24 weeks, and BM HSC engraftment was analyzed at 24 weeks. CD45.2⁺ LT-HSCs were then sorted from primary recipients (1°), and 100 LT-HSCs were similarly transplanted into each secondary recipient (2°). Donor chimerism was similarly assessed over time and HSC engraftment in the BM was examined at 24 weeks. (B) Time course of CD45.2⁺ donor chimerism after primary transplantation. (C) The number of CD45.2⁺ LT-HSCs, ST-HSCs, MPPs in the BM after primary transplantation (n=3). (D) Percentage of mice engrafted with CD45.2⁺ donor cells through 1°- and 2°-transplantation. (E) Time course analysis of CD45.2⁺ donor chimerism after secondary transplantation (WT: n=6; KO: n=5). (F) The number of CD45.2⁺ LT-HSCs, ST-HSCs, MPPs in the BM after secondary transplantation (WT: n=6; KO: n=5). All bar graphs represent mean +/− SEM. (G) Representative images of CFSE-labeled WT (left) or KO (right) LSK cells (white arrow) localized close to the endosteum (indicated by red dash line) in the trabecular bone. Scale bars represent 25 µm. *p < 0.05; **p < 0.01; ns, not significant.

To further evaluate the long-term repopulation potential and self-renewal capacity of LT-HSCs, we performed a secondary-transplantation of sorted CD45.2⁺ LT-HSCs (100 cells/mouse) along with 2 × 10⁵ CD45.1⁺ support cells (Figure 4A). Time course analysis of secondary recipients showed that Alcam^−/− LT-HSC-derived chimerism was significantly less (p<0.0001) than that of WT LT-HSCs (Figure 4E). Likewise, the fraction of multi-lineage Alcam^−/− engrafted mice (1 out of 5) was markedly lower than that of WT (4 out of 6) at 24 weeks (Figure 4D; p=0.0079). Although Alcam^−/− LT-HSCs had reduced engraftment potential, we found significantly more (p=0.0107) phenotypic Alcam^−/− LT-HSCs in the BM than WT LT-HSCs (Figure 4F). Since engraftment of transplanted HSCs require homing to the BM cavity, we tested the homing ability of Alcam^−/− HSCs by a short-term homing assay (see Supplementary Material and Method). The number of labeled Alcam^−/− LSK cells in the BM and spleen of transplanted mice 4 hours after transplantation were similar to WT LSK cells (Supplementary Figure 4). We also tested whether Alcam-deficiency alters lodgment of CFSE labeled LSK cells (40,000 cells/mouse) into non-irradiated recipients 16 hours after injection (see Supplementary Material and Method). Overall, Alcam^−/− LSK cells are capable of localizing close to the endosteum both in the midsections of the long bone and the trabecular bone where HSCs have been shown to preferentially localize [37] (Figure 4G). In the trabecular bone region, Alcam^−/− LSK cells appear to distribute somewhat closer to the endosteum compared to WT LSK (Supplementary Figure 4C; p=0.0319). Further, we tested whether Alcam^−/− HSCs show enhanced mobilization induced by G-CSF treatment (300 ug/kg, 5 days). We find that the frequency of phenotypic LT-HSC and ST-HSC in BM, spleen and PB were similar (Supplementary Figure 5A, B, C) and the numbers of CFU-C were not significantly different between Alcam^−/− and WT mice (Supplementary Figure 5D). Collectively, these results provide direct evidence that despite the progressive increase in phenotypic LT-HSC population over time, Alcam^−/− LT-HSCs are defective in long-term multi-lineage engraftment which is not caused by inefficient BM homing, lodgment or enhanced HSC mobilization.

Alcam-deficiency leads to age-associated increase in LT-HSC cellularity and premature up-regulation of Selp

To assess whether Alcam might regulate hematopoiesis in an age-dependent manner, we analyzed the hematopoietic compartment in the BM of 12 month old Alcam^−/− and WT mice. Cellularity in the BM of Alcam^−/− mice did not differ significantly from that of WT mice (Figure 5A), and the frequencies of myeloid and lymphoid progenitors were also comparable (Figure 5B and C). Evaluation of the LSK population revealed that the LT-HSC subset was significantly increased (p=0.0288) in 12-month-old Alcam^−/− mice (Figure 5D and E). Consistent with the reported age-associated increase in phenotypic HSCs, the frequency of LT-HSCs in BM was 2-fold greater in 12-month-old WT mice (0.035 +/− 0.006 %) compared to young (2 month old) WT mice (0.015 +/− 0.002 %)(Figure 5E and 2E). In comparison, the LT-HSC frequency in Alcam^−/− 12-month-old mice (0.065 +/− 0.009%) was 4-fold greater than in young Alcam^−/− mice (0.016 +/− 0.002%). Similar to young LT-HSCs, there was no significant difference in the cell cycle distribution of older Alcam^−/− LT-HSCs compared to WT LT-HSCs of the same age shown by EdU incorporation (Supplementary Figure 2D and E).

Figure 5. Alcam deficiency leads to age-associated expansion of LT-HSCs and reduction in leukocyte output.

(A) BM cellularity in 12-month-old WT (n=6) or KO (n=5) mice. Shown are total numbers of mononuclear cells pooled from 1 femur and 2 tibias. Frequencies of CMP, GMP and MEP (B); and CLP, LMPP and LSK (C) subsets in the BM. (D) Representative FACS profiles and gating strategy for phenotypic HSC subsets (LT-HSC, ST-HSC) and CD150^hi, CD150^lo fractions of LT-HSCs. Numbers shown are the mean frequencies of the cell types in BM of WT (top, n=6) or KO (bottom, n=5) mice. (E) Frequencies of LT-HSC, ST-HSC and MPP subsets in the BM. (F) Frequencies of the CD150^hi and CD150^lo subsets of LT-HSCs in WT and KO mice. All bar graphs represent mean +/− SEM (n=5 or 6). (G) White blood cell (WBC) counts of 2- and 12-month-old WT and KO mice. Differential counts of lymphocytes (LY) (H), neutrophils (NE) (I), monocytes (MO) (J). Each dot represents an individual mouse, lines indicate the median (WT: n=11; KO: n=14). *p < 0.05; **p < 0.01

The increase in LT-HSCs seen in older Alcam^−/− mice was mostly due to an increase in the CD150^hi fraction (Figure 5D and F, KO: 0.048 +/− 0.005% compared to WT: 0.024 +/− 0.004%; p=0.0072), which was reportedly myeloid-biased [10]. Therefore, we assessed whether there was an age-associated alteration of the blood counts in Alcam^−/− mice. We found that leukocyte counts were significantly reduced in 12-month-old Alcam^−/− mice (Figure 5G, p=0.0012). Differential counts revealed that this decrease was mainly due to a prominent reduction in the number of lymphocytes (Figure 5H, p=0.0020). The numbers of neutrophils (Figure 5I, p=0.0209) and monocytes (Figure 5J; p=0.0186) were also significantly lower in Alcam^−/− mice. There was no significant difference in erythrocyte, hematocrit, hemoglobin or platelet counts (Supplementary Figure 6A–D) or the lineage composition of leukocytes in the blood or BM in Alcam^−/− mice compared to WT by FACS analysis (Supplementary Figure 6E, F). We performed a LT-HSC repopulation assay using 100 sorted LT-HSCs from 15 month old Alcam^−/− or WT mice (CD45.2⁺). Although overall Alcam^−/− or WT-derived CD45.2⁺ chimerism in the PB was similar (Figure 6A), we observed a significant (p= 0.0182) biased toward the myeloid lineage within Alcam^−/− derived CD45.2⁺ population in the PB and the BM (Figure 6B and data not shown). Furthermore, there were significantly (p=0.0426) more phenotypic LT-HSCs derived from Alcam^−/− in the transplanted BM 24 weeks after transplantation compared to those derived from WT (Figure 6C). These results are consistent with an age-associated increase of phenotypic LT-HSCs in Alcam^−/− mice.

Figure 6. Alcam-deficiency causes cell autonomous increase of LT-HSC with age and premature up-regulation of age-related gene expression.

(A) Time course analysis of CD45.2⁺ chimerism in PB of mice transplanted with LT-HSCs (100 cells) sorted from 15-month-old WT or KO mice (CD45.2⁺) along with 2 × 10⁵ CD45.1⁺ supportive cells. (B) Percentage of B cells (B220⁺), T cells (CD3⁺) and myeloid cells (Gr1⁺ and/or CD11b⁺) within CD45.2⁺ donor derived BM cells at 24 weeks after LT-HSC transplantation (n=7). (C) Number of CD45.2⁺ LT-HSC in the BM of transplanted mice 24 weeks after LT-HSC transplantation. Each dot represents an individual mouse, lines indicate the median. Relative mRNA levels of Selp (D), Clu (E), Cdc42 (F), and Foxo3 (G) in sorted LT-HSC from 2-month- and 12- or 15-month-old WT or KO determined by qRT-PCR. Bar graph represents mean + SEM (n=2–4, each in duplicates). *p < 0.05; **p < 0.01

To dissect the molecular mechanism underlying the aging-like phenotype, we assessed changes in expression of genes known to associate with age or are functionally important for HSC aging in sorted LT-HSCs from 2-month-old and 12- or 15-month-old Alcam^−/− or WT mice. We analyzed age-related genes including Selp [9,11,38], Clu [39], Cdc42 [40], and Foxo3 [41] by qRT-PCR. Consistent with the reported progressive up-regulation with age, we detected markedly increase expression of these genes in 12–15 months old WT LT-HSCs compared to those of 2 months old (Figure 6D–G). In Alcam^−/− LT-HSCs, expression levels of Selp, Clu, Cdc42, and Foxo3 are substantially elevated at 2 months of age while similar levels are detected at an older age (12- or 15-months) (Figure 6D–G). Taken together, these results indicate that Alcam deficiency leads to an age-associated increase in CD150^hi phenotypic LT-HSCs, and premature up-regulation of age-related gene expression.

Discussion

ALCAM is known to be expressed on primitive human HSCs [29,30], however, its functional contribution to HSC biology has been unclear. The current study provides a comprehensive characterization of the role of Alcam in regulating adult HSC function. We report that Alcam is most highly expressed in the primitive subsets of the hematopoietic hierarchy, and its expression is up-regulated in LT-HSCs with age. In addition to being highly expressed in HSCs, Alcam is differentially expressed during myeloid progenitor specification. We demonstrate that high Alcam expression in CMPs enriches for progenitors with GM potential whereas low Alcam expression enriches for those with MegE potential (Figure 2G), consistent with Alcam levels observed in Pre-GM and Pre-MegE subsets (Supplementary Figure 1A). In agreement with these findings, it was recently reported that down-regulation of Alcam is necessary for megakaryocytic differentiation [33], underscoring the biological significance of differential gene expression.

Our studies indicate that Alcam-deficiency leads to a significantly reduced frequency of long-term repopulating cells, and that Alcam-deficient HSCs are functionally compromised in long-term engraftment on a per cell basis. Analysis of cell proliferation by EdU incorporation showed that cell cycle distribution is not altered in Alcam^−/− HSCs or progenitors, regardless of age (Supplementary Figure 2). Thus, Alcam does not appear to regulate general proliferation rate or cell cycle progression of HSCs on a population level. We tested the possibility that Alcam^−/− HSCs are defective in homing or lodgment in the BM cavity which may contribute to the engraftment deficiency. We found that Alcam^−/− LSK cells had similar homing abilities as WT LSK cells (Supplementary Figure 4A, B) and Alcam^−/− LSK cells was capable of lodging close to the endosteum (Figure 4G). We also show that HSC mobilization induced by G-CSF was not affected in Alcam^−/− mice (Supplementary Figure 5). Therefore, the engraftment deficiency of Alcam^−/− HSCs is not due to inefficient homing, lodgment or altered mobilization. We have, however, observed significantly higher proportions of Alcam^−/− LSK cells located near the endosteum in the trabecular bone region (Supplementary Figure 4C). Alcam is expressed in multiple cell types that constitute the BM microenvironment, including osteoblasts [42], endothelial cells [18], adipose derived stromal cells [43], and perichondrial mesenchymal stem cells (MSCs) [44]. Furthermore, an Alcam⁺ subset of the endosteal population has recently been shown to support HSC activity through cell adhesion [34]. It is therefore possible that Alcam might regulate the composition of BM microenvironment or the localization of specialized HSC niche components. Whether Alcam deficiency leads to alterations in the HSC niche and the microenvironmental consequences of Alcam loss remain outstanding questions. Further understanding of the microenvironmental composition and functional characteristics of these components is needed to address these questions.

In spite of the reduced long-term engraftment capacity measured in Alcam^−/− LT-HSCs, the number of Alcam^−/− LT-HSCs was remarkably higher than their WT counterparts after two rounds of long-term repopulation (6 months each)(Figure 4F). This finding is consistent with increased numbers of phenotypic LT-HSCs detected in 12-month-old Alcam^−/− mice (Figure 5E). Therefore, Alcam-deficiency causes cell autonomous, age-associated increase in phenotypic LT-HSC. A series of recent studies have demonstrated that the HSC compartment consists of several subpopulations of HSCs with intrinsically distinct lineage differentiation and proliferation behaviors, and that lymphoid-biased HSCs predominate in early development through young adulthood, whereas myeloid-biased HSCs become progressively dominant during aging [12,45–48]. The HSC expansion found in 12–15 months old Alcam^−/− mice is primarily within the CD150^hi subset of HSCs (Figure 5F), which is associated with the myeloid-biased HSCs predominating the HSC pool with age [10]. Indeed, we found that these Alcam^−/− LT-HSCs showed myeloid reconstitution bias when compared to age matched WT LT-HSCs in an in vivo repopulation assay (Figure 6B). We did not detect apparent changes in relative lineage composition in the PB or BM of older Alcam^−/− mice (Supplementary Figure 6E, F). This is perhaps not surprising, since it has been shown that despite a clear shift in HSC subtype composition with age in the C57Bl/6 strain of mice, there is no significant change in the blood cell lineage ratio [47]. It has been suggested that a threshold may exist under which homeostatic mechanisms can overcome the differentiation bias of HSCs.

The mechanism underlying age-associated phenotypic HSC expansion is not well understood. Based on our study and the report that myeloid-biased HSCs are able to self-renew more extensively than lymphoid-biased HSCs [48], it is tempting to speculate that the HSC composition in Alcam^−/− mice might shift progressively more towards myeloid-biased with age. We show that Alcam levels on LT-HSC cell surface increases with age (Figure 1C). Conceivably, Alcam might be up-regulated to provide a protective mechanism counteracting HSC aging and maintain a balanced HSC composition. Gene ontology analysis of age-associated gene expression found that NF-κB signaling was the most significantly enriched category activated in HSCs during aging [11]. Meanwhile, NF-κB reportedly regulates ALCAM transcription through direct binding to the NF-κB response elements in the proximal promoter [49]. Whether similar regulatory mechanism by NF-κB is conserved in the mouse remains to be determined. We speculate that age related up-regulation of NF-κB signaling might underlie the increased expression of Alcam with age (Figure 1E). In addition, we find that Alcam deficiency leads to a substantial increase of age-associated gene [9,11,38,50] including Selp, Clu, Cdc42, and Foxo3 in LT-HSCs at a young age (Figure 6D–G). Up-regulation of Selp, which encodes P-selectin, is amongst the most significant age-associated alteration commonly found during HSC aging [9,11,38]. Loss of P-selectin has been shown to result in increased HSC self-renewal and enhanced chronic myeloid leukemia transformation, suggesting that P-selectin might negatively regulate HSC self-renewal [38]. Foxo3, a forkhead transcription factor, is essential for maintenance of HSC quiescence and HSC pool with age [41]. Based on our finding, regulation of HSC pool might be disrupted in Alcam^−/− mice in part through increased expression of Foxo3 and premature up-regulation of P-selectin. Recently, elevated activity of the small RhoGTPase Cdc42 in aged HSCs is linked to a loss of cell polarity and HSC aging [40]. Consistent with an elevated Cdc42 activity in aged HSCs, we found that Cdc42 expression in HSCs increases with age and that Cdc42 level is pre-maturely elevated in Alcam^−/− HSCs. These results suggest that Alcam deficiency might lead to early loss of HSC cell polarity through activation of Cdc42. It has long been suggested that altered interactions with the BM stroma might cause functional changes in aging HSCs [51,52]. The current study provides direct evidence that Alcam-mediated cell adhesion is cell-autonomously required for maintaining a functional HSC pool and a controlled HSC pool size during transplantation and aging.

Conclusions

Alcam is a cell adhesion molecule that is highly expressed in adult LT-HSCs where its expression progressively increases with age. This study provides evidence that Alcam regulates the functional integrity and self-renewal of LT-HSCs upon transplantation and during aging. This work reveals previously unappreciated functions of Alcam in HSC biology, with important implications in hematopoietic transplantation. Further investigation of Alcam signaling effectors, and its role in cell polarity and malignant transformation will be important for the evaluation of potential translational opportunities.