Remodeling of Bone Marrow Hematopoietic Stem Cell Niches Promotes Myeloid Cell Expansion during Premature or Physiological Aging

1Hematopoietic aging is characterized by expansion of hematopoietic stem cells (HSCs) with impaired function, such as reduced engraftment, quiescence, self-renewal, unfolded protein response, and lymphoid differentiation potential, leading to myeloid-biased output both in mice and humans. Myeloid malignancies are more frequent in the elderly, but whether changes in the aged HSCs and/or their microenvironment predispose to these malignancies remains unclear. HSC aging was initially considered to result only from intrinsic changes, such as epigenetic deregulation, replication stress, deficient DNA repair, transition from canonical to non-canonical Wnt signaling, and increased autophagy. However, aging also reduces HSC polarity and capacity to lodge near bone, and microenvironmental alterations have been proposed to contribute to hematopoietic aging. For instance, myeloid bias can result from telomere dysfunction or inflammatory cytokine production in the aged hematopoietic microenvironment, and bone marrow (BM) adrenergic nerve degeneration has been recently proposed to drive HSC aging reversibly. However, the relative contribution of intrinsic and extrinsic mechanisms to HSC aging remains debated. Myeloid-biased HSCs and more prominently platelet-primed HSCs expand in aged mice. However, both platelet-primed and unprimed old HSCs exhibit myeloid bias, possibly suggesting a microenvironmental participation. The standard model of platelet generation (thrombopoiesis) suggests that megakaryocyte precursors migrate from endosteal BM (close to bone) to sinusoids (further from bone) for maturation, which can be promoted by cytokines like interleukin (IL)-1α IL-1β, IL-6, IL-3, and interferon-gamma (IFNγ). Whereas platelet-biased HSCs are located near sinusoids, both megakaryocytes and their precursors have been recently found throughout BM, including the sinusoids where megakaryocyte maturation and thrombopoiesis take place. In this study, we have investigated whether different BM microenvironments regulate myeloid differentiation and megakaryopoiesis during aging.

2To characterize changes in vascular beds during aging, we performed whole-mount immunofluorescence staining of thick femoral sections of young and old wild type (WT) mice using CD31 and endomucin (EMCN) to identify sinusoids, arterioles, capillaries, and transition zone vessels (TZVs), which connect arterioles with sinusoids near bone (endosteum). Consistent with previous findings, endosteal vessels and TZVs were reduced in old mice; however, we also noted that, whereas sinusoidal areas appeared unchanged, arterioles were slightly reduced and capillaries located further from bone increased 4-fold in old mice (Figures 1A–1F). These vascular changes were paralleled by similar alterations of their associated perivascular cells. We used Nes-gfp transgenic mice, which carry HSC niche-forming perivascular BM mesenchymal stem or progenitor cells (BMSCs) labeled with GFP. Nes-GFP+ cells augmented 4-fold specifically in non-endosteal BM, mostly associated with the increased capillaries (Figures 1G–1I and S1A–S1D). These changes correlated with increased inflammatory cytokines that drive myeloid cell expansion. The concentration of IL-1α, IL-1β, and IL-6 increased in the BM during aging, whereas IL-3 and IFNγ showed similar trends (Figures 1J–1N and S1M-S1Q).

3We have previously shown that sympathetic adrenergic signals regulate Nes-GFP+ cell proliferation and are affected during age-related myeloproliferative neoplasms. Additionally, increased sympathetic adrenergic activity has been previously described during aging, chronic stress, and depression, and might increase osteoporosis and fracture risk by restraining bone formation. However, the opposite (decreased BM adrenergic innervation) has been recently suggested as causative of HSC aging. To clarify this, whole-mount preparations of skulls and thick tibial sections of Nes-gfp mice were immunostained for tyrosine hydroxylase (TH), to visualize sympathetic noradrenergic fibers and nestin+ cells in large 3D volumes. This study did not confirm reduced TH+ fibers in the aged BM but found these fibers increased by 50% in the skull of old mice (Figures 2A–2C) and augmented 2.5-fold in the aged tibial BM, compared with the young samples (Figures 2D–2F). In both cases, nestin+ cells were found in proximity of noradrenergic fibers (Figures S1E–S1L). Together, these results suggest contraction of endosteal (bone-associated) HSC niches and expansion of non-endosteal neurovascular HSC niches during aging.

4To study the possible contribution of increased adrenergic innervation to aged hematopoiesis, we analyzed mice lacking β2-ΑR and β3-ΑR (Adrb2−/−Adrb3−/−), the main β-ARs that cooperate in HSC niche regulation. Importantly, myeloid and megakaryocyte progenitors expanded less in the endosteal BM of aged Adrb2−/−Adrb3−/− mice (Figures 3A, 3B, and S2A–S2F). Consequently, megakaryocytes did not increase in Adrb2−/−Adrb3−/− mice during aging (Figures 3C–3G), suggesting a functional role for increased noradrenergic signaling in aged hematopoiesis. Megakaryocytes were less frequently found in apposition to BM sinusoids and could form less protrusions (required for proplatelet formation) in aged Adrb2−/−Adrb3−/− mice (Figures 3H–3M). Accordingly, aged Adrb2−/−Adrb3−/− mice did not show the increased circulating platelets typically observed in aged WT mice (Figure 3N), whereas other circulating blood cell types remained unaffected (Figures S2G–S2J). These results suggest that increased β-adrenergic signaling in the BM promotes myeloid differentiation into platelets during aging.

5We analyzed single Adrb2−/− mice and Adrb3−/− mice to understand the role of each β-AR in myelopoietic regulation. Resembling aged Adrb2−/−Adrb3−/− mice (Figure 3N), aged Adrb2−/− mice did not exhibit increased circulating platelets (Figure 4A), suggesting that increased β2-adrenergic activity promotes thrombopoiesis during aging. Already at adulthood, Adrb2−/− mice showed reduced frequency of myeloid and megakaryocyte progenitors, and this reduction persisted during aging (Figures 4B, 4C, S3A, and S3B). We generated BM chimeras using Adrb2−/− mice or WT mice as donors/recipients to distinguish hematopoietic cell-autonomous from microenvironmental regulation. WT mice carrying β2-AR-deficient hematopoietic cells showed normal frequencies of myeloid and megakaryocyte progenitors (Figures 4B′–4C′ and S3A′–S3B′). In contrast, Adrb2−/− recipients of WT BM cells showed similar reductions as found in primary mice (Figures 4B″–4C″ and S3A″–S3B″). These results suggest that β2-AR signals promote megakaryocyte differentiation through the microenvironment. To obtain mechanistic insight, we treated with selective β2-AR or β3-AR agonists the HSPC-like HPC-7 cell line cultured alone or co-cultured with MS-5 stromal cells, which resemble nestin+ BMSCs (,; Figures S3C and S3D). Whereas the selective β-AR agonists did not affect megakaryocyte differentiation of HPC-7 cells cultured alone (Figure S3E), β2-AR and β3-AR agonists had opposite stage-specific effects on megakaryocytic differentiation from HPC-7 cells co-cultured with MS-5 cells. Whereas β3-AR agonist increased the frequency of undifferentiated c-kithiCD41lo cells and decreased the fraction of c-kitloCD41lo megakaryocyte progenitors, β2-AR agonist did not affect early differentiation, but instead increased the frequency of CD41hi megakaryocytic cells at a later differentiation stage (Figure S3F).

6To test the potential significance of this regulation in the human setting, we treated human umbilical cord blood-derived hCD34+ HSPCs co-cultured with MS-5 cells with a selective β2-AR agonist (Figure 4D). Treatment with β2-AR agonist increased the frequency of human CD61+ megakaryocytic cells (Figures 4E–4G). These results suggest that β2-AR stimulation of stromal cells promotes megakaryocyte differentiation.

7To investigate the underlying mechanism, we measured several cytokines which regulate megakaryopoiesis. Among those, IL-6 concentration increased in WT BM during aging (Figure 1H) and was low in endosteal Adrb2−/− BM (Figures 4H and S3G–S3I). Moreover, β2-AR agonist increased Il6 mRNA 2-fold in MS-5 stromal cells; this effect was abrogated after blocking protein kinase A (Figure 4I) downstream of β2-AR signaling. Adult Adrb2−/− mice and Il6−/− mice showed similarly reduced BM megakaryocytes (Figures 4J–4M), confirming the role of β2-AR and IL6 in megakaryopoiesis. To directly test the role of IL-6 downstream of β2-AR, we established primary BM cultures from Il6−/− or control mice and treated them with β2-AR agonist or vehicle. β2-AR agonist tripled myeloid progenitors in WT samples, but not in Il6−/− samples (Figure 4N). Altogether, these results suggest that β2-AR on stromal cells promotes megakaryocyte differentiation through IL-6.

8Our co-culture experiments suggested that, contrasting β2-AR, β3-AR might inhibit megakaryocytic differentiation (Figures S3C–S3F). Additionally, β3-AR stimulation has been shown to promote HSC lymphoid skewing. Therefore, we measured the frequency of immunophenotypically defined lymphoid-biased HSCs (Figure S4A; ). Lymphoid-biased HSCs were reduced by one-third in adult endosteal Adrb3−/− BM (Figures 5A and S4B). This phenotype was not observed in WT mice carrying β3-AR-deficient hematopoietic cells (Figures 5A′ and S4B′) but was reproduced in chimeric mice lacking β3-AR in the microenvironment (Figures 5A″ and S4B″), implying niche-mediated regulation. Opposite trends were observed for long-term HSCs (LT-HSCs) and myeloid-biased HSCs, with their frequencies tending to increase in endosteal Adrb3−/− BM (Figures S4C–S4F). These results suggest accelerated lymphoid deficiency in the endosteal microenvironment lacking β3-AR.

9We investigated possible mechanisms that might explain the intriguingly opposite effects of β3-AR and β2-AR signaling on lympho-myeloid lineage skewing. Whereas both β-ARs can activate G proteins, their differential cardiovascular effects have been attributed to β3-AR-dependent nitric oxide (NO) generation. Therefore, we measured nitrate concentration in the BM extracellular fluid and found it doubled in endosteal (compared with non-endosteal) WT BM and halved in Adrb3−/− BM (Figure 5B). Among NO synthases, Nos1 (but not Nos2 or Nos3) showed higher mRNA expression in endosteal WT BM and was downregulated in Adrb3−/− BM (Figures S4G–S4I). To further study the role of β3-AR and Nos1 on HSC lineage bias, we established primary BM cultures from transgenic mice expressing GFP under the regulatory elements of Von Willebrand factor (Vwf-eGFP), which discriminate myeloid/platelet-biased (Vwf-eGFP+) HSCs from lymphoid-biased (Vwf-eGFP−) HSCs. Treatment with β3-AR agonist did not affect platelet-biased HSCs (Figures S4J and S4K) but doubled lymphoid-biased HSCs in a Nos1-dependent manner (Figure 5C). Moreover, Adrb3−/− mice and Nos1−/− mice showed similar signs of premature hematopoietic aging; myeloid or megakaryocyte progenitors (Figures 5D–5G) and megakaryocyte apposition to blood vessels (Figures 5H–5K) similarly increased in adult Adrb3−/− mice and Nos1−/− mice. Consistently, transition zone vessels decreased (Figures 5L–5O) whereas capillaries expanded (Figure 5P) in both KO models. Despite the partial reduction of lymphoid-biased HSCs in endosteal Adrb3−/− BM, the frequencies of circulating lymphocytes or myeloid cells appeared unchanged in 5-month-old Adrb3−/− mice (Figures S4L–S4N), contrasting previous findings. Whereas adult Adrb3−/− mice did not yet show in our analysis premature hematopoietic aging in peripheral blood, adult Nos1−/− mice showed reduced lymphocytes and increased neutrophils and platelets in circulation (Figures 6A and 6B), suggesting that other pathways (besides β3-AR) contribute to Nos1-dependent regulation of hematopoiesis. Altogether, these results suggest that microenvironmental β3-AR contributes to balance HSC lineage-bias toward lymphoid production, which is at least partially dependent on Nos1-dependent NO production.

10We next interrogated whether the BM microenvironment might promote pathological (and not only physiological) hematopoietic aging. Premature aging in Hutchinson-Gilford progeria syndrome (HGPS) is caused by the accumulation of a truncated prelamin A protein named “progerin,” which is produced via aberrant splicing resulting from a de novo synonymous c.1824C > T (p.G608G) point mutation in the LMNA gene (encoding lamin A and C). This gene has been recently associated with changes in epigenetic and chromatin architecture in aged HSCs. Progerin also accumulates during normal aging and HGPS displays most aging hallmarks, suggesting that HGPS can inform physiological aging. LmnaG609G knockin mice exhibit key hallmarks of the human disease, including accelerated aging, shortened lifespan, and bone and cardiovascular defects. Certain hallmarks of hematopoietic aging in mice, such as increased circulating platelets, have been observed in HGPS. However, it remains unknown whether premature aging affects the overall hematopoietic system in HGPS and whether this is a consequence of the LMNA mutation in HSCs, other hematopoietic cells and/or their microenvironment.

11We measured peripheral blood counts in progeroid mice and control mice. Remarkably, adult progeroid mice resembled adult Nos1−/− mice (Figures 6A and 6B) in their premature myeloid skewing evidenced by decreased lymphocytes and increased neutrophils, monocytes, and platelets (Figures 6C and 6D), to a similar extent as normally aged mice. Flow cytometry analysis confirmed increased myeloid cells and decreased lymphoid cells in the peripheral blood of progeroid mice (Figures 6E–6G).

12To test HSC function, lethally irradiated CD45.1 recipients were transplanted with CD45.2 BM cells from LmnaG609G/G609G mice or WT mice, together with competitor BM cells from congenic CD45.1 mice (Figure S5A). Long-term engraftment was lower for LmnaG609G/G609G (compared with WT) hematopoietic cells (Figure S5B), as previously observed for normally aged HSCs, albeit to a lower extent. However, myeloid skewing was not observed in blood cells derived from LmnaG609G/G609G HSCs (Figures 6H–6J), suggesting that myeloid skewing is a non-HSC-autonomous aging feature in HGPS. To assess this possibility, we transplanted CD45.2 BM cells from LmnaG609G/G609G mice into lethally irradiated CD45.1 WT recipients. Sixteen weeks later, WT recipients carrying the LmnaG609G/G609G mutation only in hematopoietic cells (Figures 6K and 6L) did not reproduce the myeloid skewing and increased platelet counts observed in constitutive LmnaG609G/G609G mice (Figures 6C and 6D). Together, these results suggest that premature aging in HGPS affects the hematopoietic system but cannot be explained by HSC-autonomous alterations. Consistent with this idea, microenvironmental alterations (such as sinusoidal vasodilation) were found in LmnaG609G/G609G BM (Figures S5D and S5E). Moreover, myelopoietic cytokines augmented in normally aged BM microenvironment (Figures 1J–1N) were also increased in adult progeroid mice (Figures 6M–6Q).

13These results suggested that similar microenvironmental alterations might promote myeloid differentiation during physiological/pathological aging. Therefore, we tested whether targeting the microenvironment could impact hematopoiesis in HGPS. Progeroid mice were chronically treated with a β3-AR agonist, which can rescue nestin+ niches in humans and mice with age-related myeloproliferative disorders and has been recently suggested to rejuvenate normally aged HSCs. Treatment with β3-AR agonist over 2 months normalized circulating granulocytes and lymphocytes and BM neutrophils and partially rescued BM B cells (Figures 7A–7D). This effect correlated with decreased frequency of BM LT-HSCs (Figures 7E and 7F).

14Megakaryocytes have been shown to regulate HSC proliferation and their numbers were found to increase during normal/pathological aging in this study. Therefore, we examined the spatial relationships between megakaryocytes and HSPCs in the different models. Like in aged mice, HSCs expanded but located further from megakaryocytes in adult Adrb3−/−mice, as an additional premature aging feature in these mice (Figures S6A–S6D). Increased LT-HSC frequency in vehicle-treated progeroid mice (compared with WT mice) (Figures 7E and 7F) correlated with reduced co-localization of HSPCs and megakaryocytes, which was rescued by β3-AR agonist (Figures 7G–7J). HSCs located more distantly from megakaryocytes in vehicle-treated progeroid mice (compared with WT mice) but this association was partially restored by β3-AR agonist (Figures 7K–7N). These results suggest the possibility that decreased interactions between megakaryocytes and HSCs might reversibly regulate hematopoiesis during aging.

15This study suggests that alterations of the BM microenvironment during pathological/physiological aging change hematopoiesis. Remodeling of BM HSC niches (namely reduced endosteal niches and increased non-endosteal neurovascular niches) is associated with the overproduction of pro-inflammatory cytokines which drive excessive myelopoiesis in normal/premature aging. However, the latter can be improved by modulating the microenvironment. Together, these results expand previous findings on the microenvironmental contribution to hematopoietic aging to the role of different neurovascular beds in normal and pathological aging, and their potential therapeutic targeting.

16Stem cell aging compromises tissue turnover and regeneration, but the contribution of the aged stem cell niche remains incompletely understood. In the hematopoietic system, HSCs expand during aging but exhibit impaired self-renewal and lymphoid differentiation potential, favoring myeloid output. Moreover, myeloid malignancies are more frequent in the elderly, but whether changes in the aged HSCs and/or their microenvironment predispose to these malignancies has remained unclear. Although HSC aging was initially regarded as solely derived from intrinsic changes in HSCs, several alterations in the BM microenvironment or in its interaction with HSCs have been proposed to promote hematopoietic aging. Our study supports this contention by demonstrating that changes in the microenvironment contribute to physiological hematopoietic aging and pathological premature aging (HGPS). In both settings, the aged microenvironment promotes myeloid differentiation through similar cytokines, such as IL-1β and IL-6.

17Progeroid mice exhibit several hallmarks of hematopoietic aging: expanded HSCs with reduced engraftment, lymphoid deficiency, and myeloid skewing toward platelets. However, chimeric mice carrying LmnaG609G/G609G hematopoietic cells in a WT environment do not reproduce the hematopoietic aging of progeroid mice. Together with the altered vasculature and increased concentration of myelopoietic cytokines in progeroid BM, these results strongly suggest that the BM microenvironment causes premature hematopoietic aging in HGPS. In fact, chronic treatment of progeroid mice with β3-AR agonist (targeting the microenvironment) reduces HSCs and restores their proximity to megakaryocytes and their lympho-myeloid skewing.

18During physiological aging, endosteal BM niches decrease, consistent with previous observations, whereas non-endosteal neurovascular BM niches containing Nes-GFP+ stromal cells show a pronounced expansion. This microenvironmental remodeling might directly favor myeloid expansion during aging, because lymphoid niches have been found near bone, whereas myeloid cell expansion and thrombopoiesis mainly occurs in non-endosteal niches.

19Recently, an age-related reduction of BM adrenergic nerve fibers has been reported to dive most hallmarks of HSC aging. However, studies on BM innervation during aging have provided conflicting results, with BM adrenergic fibers reportedly been dramatically reduced, unchanged, or increased. The whole-mount or thick section imaging and 3D reconstruction of different bones in the present study supports the latter and demonstrates that these fibers are doubled in different bones during physiological aging. Moreover, these fibers are not reduced in adult progeroid mice (Figure S7), which exhibit premature hematopoietic aging. Our findings are consistent with the well-known increase in sympathetic activity in the elderly and suggest a functional change of neurotransmission (β2-AR overriding β3-AR), rather than a general decline of BM innervation during aging. Actually, increased noradrenergic activity in the BM seems to account for augmented thrombopoiesis in aged mice because platelets are not elevated in aged Adrb2−/− mice or Adrb2−/−Adrb3−/− mice. A previous study showed direct effects of α-ARs on mature megakaryocyte adhesion, proplatelet formation, and platelet release, whereas α-AR signaling on more primitive CD34+ progenitor cells did not affect lineage commitment. However, the present study indicates that HSPC lineage commitment is regulated by the microenvironment through β2-AR and β3-AR. Therefore, BM noradrenergic activity appears to regulate lineage commitment and megakaryopoiesis at different stages of maturation through distinct α- and β-ARs. Interestingly, β2-AR and β3-AR exhibit opposite roles on myeloid differentiation: whereas β2-AR signaling promotes megakaryopoiesis through stromal-cell-derived IL6 and becomes predominant during aging, β3-AR inhibits myelopoiesis. Lack of β2-AR in the microenvironment (but not in the hematopoietic system) halves myeloid and megakaryocyte progenitors in adult mice. Decreased frequency of megakaryocyte lineage cells is similarly found in adult Adrb2−/− mice and aged Adrb2−/−Adrb3−/− mice (but not in WT chimeras carrying Adrb2−/− hematopoietic cells), suggesting that predominant β2-AR activation promotes megakaryocyte differentiation during aging. Whereas adult Adrb2−/− mice and Adrb3−/− mice exhibit opposite deregulation of lymphoid/myeloid lineage output, only Adrb2−/− mice retain this phenotype upon aging (Figures 4B, 4C, and 5A), further suggesting that β2-AR prevails over β3-AR in hematopoietic lineage regulation during aging. This regulation requires IL-6, a cytokine known to increase in elderly humans and to be similarly regulated in other cell types. Supporting the role of IL6, megakaryocytes appear decreased in Il6−/− BM. Moreover, β2-AR agonist specifically increased IL-6 expression in stromal cells co-cultured with human/murine HSPCs and consequently stimulates megakaryopoiesis in these co-cultures or in primary BM cultures from WT mice, but not Il6−/− mice.

20In contrast, microenvironmental (but not hematopoietic) lack of β3-AR partially reduces lymphoid-biased HSCs in adult mice, as previously reported. However, the results here differ in that β3-AR loss does not appear overall sufficient to cause premature hematopoietic aging, because the frequencies of circulating lymphocytes and myeloid cells appeared unchanged in 5-month-old Adrb3−/− mice. In vitro, β3-AR agonist-treated stromal cells decrease human and murine HSPC differentiation into megakaryocytes. This effect requires Nos1, because Nos1−/− mice have high circulating platelets and granulocytes, but low circulating lymphocytes. Furthermore, β3-AR agonist specifically increases lymphoid-biased HSCs in primary BM cultures in a Nos1-dependent manner. Moreover, adult Adrb3−/− or Nos1−/− mice share premature microenvironmental aging features in the BM, such as reduced bone-forming transition zone vessels and increased BM capillaries, myeloid progenitors, and megakaryocyte apposition to blood vessels.

21Megakaryocytes inhibit HSC proliferation. Whereas both HSCs and megakaryocytes expand during aging, HSCs locate further from megakaryocytes in adult Adrb3−/− mice and progeroid mice, suggesting that decreased HSC-megakaryocyte interactions might contribute to premature hematopoietic aging features in these mice. Supporting this possibility, chronic treatment of progeroid mice with β3-AR agonist decreased HSCs and corrected lineage skewing, correlated with normalized distribution of HSCs near megakaryocytes. Future studies will be required to investigate changes in the regulation of HSC proliferation by megakaryocytes during aging.

22In summary, this study shows that the aged BM microenvironment promotes myeloid expansion during physiological aging and in premature aging. Normal murine aging concurs with the reduction of endosteal niches and the expansion of non-endosteal niches comprising capillaries and nestin+ cells associated with sympathetic noradrenergic fibers. Interestingly, β2-AR and β3-AR regulate myelopoiesis through opposite and stage-dependent effects on the hematopoietic microenvironment. During normal aging, increased β2-AR activity promotes IL6-dependent myeloid differentiation, whereas decreased β3-AR-Nos1-NO is associated with reduced endosteal niches and increased central niches. Adult Nos1 KO mice and progeroid mice display premature aging in peripheral blood, manifested as reduced lymphocytes and increased myeloid cells. Megakaryocytes move closer to BM sinusoids in Adrb3−/− or progeroid mice, possibly explaining increased thrombopoiesis. Megakaryocytes and HSCs expand but separate from each other during normal aging and in adult Adrb3−/− or progeroid mice. However, chronic treatment of progeroid mice with β3-AR agonist reduces HSCs and restores their proximity to megakaryocytes and their lympho-myeloid skewing. Therefore, normal or premature niche aging promotes normal myeloid expansion, which can be improved by targeting the microenvironment.

23Further information and requests for resources and reagents should be addressed to the Lead Contact, Dr. Simon Mendez-Ferrer, at sm2116@medschl.cam.ac.uk.

24Young mice were analyzed between 8-30 weeks of age, and old mice were 66-120 weeks old. Mice were housed in specific pathogen free facilities. All experiments using mice followed protocols approved by the Animal Welfare Ethical Committees, according to EU and United Kingdom Home Office regulations (PPL 70/8406). Nes-gfp, Adrb2−/−, FVB/N-Adrb3tm1Lowl/J, Lmnatm1.1Otin, Vwf-egfp, B6.129S4-Nos1tm1Plh/J (stock#2986), B6.129S2-Il6tm1Kopf/J (stock#2650) (Jackson Laboratories) and congenic CD45.1 and CD45.2 C57BL/6 mice (Charles River Laboratories) were used in this study.

25In LmnaG609G/G609G mouse model, HSC activity was assessed by long-term competitive repopulation assay using the congenic CD45.1/CD45.2 isotypes. Recipient female CD45.1 C57BL/6 mice (8 week-old) were subjected to lethal irradiation (12 Gy whole body irradiation, split dose 6.0 + 6.0 Gy, 3 h apart). 106 BM nucleated cells from (CD45.2) WT mice or LmnaG619G/G609G mice were mixed with 106 BM nucleated cells from CD45.1 C57BL/6 mice and i.v. transplanted in 200 μL sterile PBS into lethally irradiated CD45.1 C57BL/6 mice. At various time points after transplantation (4, 8, 12 and 16 weeks), peripheral blood nucleated cells were collected from recipients. Cells were stained with fluorescent-conjugated antibodies against B220, Mac1 and CD90 antigens, and analyzed by fluorescence-activated cell sorting. The selective β3-AR agonist BRL37344 (Sigma, St. Louis, MO) was administered at 2mg/kg through intraperitoneal (i.p.) injection once per day. Vehicle (saline solution) daily injections were performed in the same way. Treatment was initiated when LmnaG609G/G609G mice were 7 weeks old and lasted for 8 weeks.

26To generate chimeric mice carrying Adrb2−/− or Adrb3−/− BM cells, two-month-old C57BL/6 mice were lethally irradiated (12 Gy, two split doses) and i.v. transplanted with 2 X 106 nucleated BM cells from two-month-old WT, Adrb2−/− or Adrb3−/− mice. To test microenvironment-dependent effects of β-adrenergic signaling, WT, Adrb2−/− or Adrb3−/− (CD45.2) mice were lethally irradiated (12 Gy, two split doses) and i.v. transplanted with 2 X 106 nucleated BM cells from CD45.1 C57BL/6 mice. Hematopoietic cells in the BM were analyzed 4 months after transplantation in both settings.

27For BM hematopoietic cell isolation, bones were crushed in a mortar, filtered through a 40-μm strainer to obtain single cell suspensions, and depleted of red blood cells by lysis in 0.15 M NH4Cl for 10 min at 4°C. Blood samples were directly lysed. Cells were incubated with the appropriate dilution (2-5 μg/ml) of fluorescent antibody conjugates and 4’,6-diamidino-2-phenylindole (DAPI) for dead cell exclusion, and analyzed on LSRFortessa flow cytometer (BD Biosciences, Franklin Lakes, NJ) equipped with FACSDiva Software (BD Biosciences). The following antibodies were used: fluorescent CD45.1 (A20), CD45.2 (104), B220 (RA3-6B2), CD11b (M1/70), CD3ε (145-2C11), Ly-6G (1A8), CD90 (53-2.1) sca-1 (E13-161.7), CD42d (1C2), biotinylated lineage antibodies (CD11b, Gr-1, Ter119, B220, CD3ε) (BD Biosciences), c-kit (2B8) (eBioscience), CD150 (TC15-12F12.2), CD34 (RAM34) and CD41 (MWReg30) (BioLegend). Biotinylated antibodies were detected with fluorochrome conjugated streptavidin (BD Biosciences).

28To isolate endosteal and non-endosteal HSPCs, long bones were flushed gently to obtain hematopoietic cells less tightly associated with the bone and then flushed-bones were crushed in a mortar to obtain hematopoietic cells enriched in the endosteal compartment. Cells were stained with the above-mentioned antibodies and analyzed by flow cytometry or sorted (FACS Aria cell sorter, BD Bioscience). Long-term HSCs (LT-HSCs) were immunophenotypically defined as lin-sca-1+ckit+CD34-CD150+CD41- cells. Myeloid-biased HSCs were immunophenotypically defined as lin-sca-1+ckit+CD34-CD150+CD41+ cells. Lymphoid-biased HSCs were immunophenotypically defined as lin-sca-1+ ckit+CD34-CD150-CD41- cells.

29For analysis of Nes-GFP+ cell distribution in endosteal and non-endosteal BM fractions, the long bones were flushed with PBS and remaining endosteal part was crushed in a mortar. Both fractions were digested in 2ml of collagenase type I (Stem Cell Technologies, cat. No. 07902) for 30 min at 37°C with agitation. The enzyme was quenched by adding 18ml PBS/2%FCS. Cell suspensions were filtered, pelleted and red blood cell lysis was performed as stated above. Samples were stained with the following antibodies: 1:100 anti-CD45-biotinylated ab (BD Biosciences, cat. No. 553078), 1:100 anti-Ter119 –biotinylated ab (BD Biosciences, cat. No. 553672), 1:100 anti-CD31-APC ab (BD Biosciences, cat. No. 551262). Subsequently, cells were stained with Streptavidin-APCCy7 1:200 (BD Biosciences, cat. No. 554063). DAPI (Sigma Aldrich, cat.no. D9542) was added at 1:10 000 to discriminate dead cells. Samples were acquired on Gallios flow cytometer (Beckman Coulter) and a LSRFortessa cell analyzer (BD Biosciences) and were analyzed with Kaluza analysis software (Beckman Coulter).

30HPC-7 cells were maintained in IMDM medium supplemented with 10% FCS and 100 ng/ml of mSCF (PeProTech 250-03).

31MS-5 cells were maintained in α-MEM medium supplemented with 10% FCS. To induce megakaryocyte differentiation, HPC-7 cells were cultured with or without MS-5 stromal cells in StemSpan H3000 medium, supplemented with 50 ng/ml of TPO (PeProTech 300-18) and 2U/mL of EPO (R&D 287-TC-500) up to 4 days. 10 μM of BRL37344 (β3-AR agonist; Sigma B169), 10 μM of clenbuterol (β2-AR agonist; Sigma C5423), 5 μM of H-89 (Protein kinase A inhibitor; Sigma B1427) and vehicle controls were added to the culture. To examine megakaryocyte differentiation, cells were incubated with anti-ckit-FITC ab (1:200, BD PharMingen 553354) and anti-CD41-PE ab (1:200, BD PharMingen 558040). DAPI was added at 1:10,000 to discriminate dead cells. Samples were analyzed by Gallios flow cytometer (Beckman Coulter).

32Human umbilical cord blood CD34+ HSPCs were isolated using a CD34 Microbead kit (Miltenyi Biotec 130-046-702) following manufacturer’s instructions, and were cultured with MS-5 stromal cells in Cellgro medium (CellGenix cat. no. 20802-0500), supplemented with 50 ng/ml of TPO (PeProTech 300-18) and 5 ng/ml of IL-1β (Cellgenix 1411-050) for 7-10 days, during which vehicle, 10 μM of BRL37344 (β3-AR agonist; Sigma B169) or 10 μM of clenbuterol (β2-AR agonist; Sigma C5423) were added to the culture. Conditioned medium was refreshed at day3 and day7 of the coculture. To examine megakaryocyte differentiation, cells were fixed and stained with anti-human CD61 antibody (Serotec MCA 728).

33For mouse bone marrow long-term culture, femurs and tibias were flushed gently to obtain bone marrow cells. Cells were seeded and cultured in MyeloCult M5300 medium (StemCell Technologies, cat. no. 05350) supplemented with 10−6 M hydrocortisone (StemCell Technologies, cat. no. 07904) at 33°C for 14 days. Half of the medium was refreshed at day 7. At day 14 of culture, half of the medium was refreshed again and added with 50 ng/ml of TPO (PeProTech 300-18) for 4 days, during which vehicle, 10 μM of BRL37344 (β3-AR agonist; Sigma B169), 10 μM of clenbuterol (β2-AR agonist; Sigma C5423), or 100 μM of L-VINO (Insight Biotechnology 728944-69-2) were added to the culture. At day 18 of culture, cells were collected and subjected to flow cytometry analysis.

34Immunofluorescence staining of cryostat sections was performed as previously described, with minor modifications. Briefly, tissues were permeabilized with 0.1% Triton X-100 (Sigma) for 10 min at RT and blocked with TNB buffer (0.1 M Tris–HCl, pH 7.5, 0.15 M NaCl, 0.5% blocking reagent, Perkin Elmer) for 1 h at RT. Primary antibody incubations were conducted for 2 h at RT. Secondary antibody incubations were conducted for 1 h at RT. Repetitive washes were performed with PBS + 0.05% Triton X-100. Stained tissue sections were counterstained for 5 min with 5 μM DAPI and rinsed with PBS. Slides were mounted in Vectashield Hardset mounting medium (Vector Labs) and sealed with nail polish. Skull bones were fixed with 2% PFA for 2h in 4C, washed with PBS and cut through the sagittal suture. Each half was permeabilized with PBS-0.1% Triton with 20% goat serum (Thermo Scientific, cat. No. 9722) overnight at 4°C on the rocker. Endogenous biotin was blocked with the Avidin/biotin blocking kit (Vector Laboratories, cat. No. SP-2001) according to manufacturer’s recommendation. Endogenous peroxidase was blocked with 0.4% peroxide (Sigma-Aldrich, cat. no. H1009) for 2h at RT. Subsequently, the skull bones were stained with rat anti-CD31 (BD Biosciences, cat. no. 551262) and rabbit anti-TH (Millipore, cat. no. AB152) in 0.1% Triton-20%goat serum-PBS (diluted 1:100 and 1:500 respectively) for 3 days, and washed over the next day with 0.05% Triton - PBS- at RT on a rocker. Samples were stained with secondary goat anti-rat-Alexa647 at 1:300 (Life Technologies, cat. no. A21247) and goat anti-rabbit biotinylated at 1:200 (Stratech Scientific, cat. no. 111-066-003) o/n at 4°C. Subsequently, the skull bones were treated with ABC amplification kit (Vector Labs) to detect Cy3-Tyrmide amplified signal from TH staining (Perkin Elmer, cat.no. SAT704A001EA). As a final step, the skull bones were stained with DAPI 1:1000 for 5 min. Images were acquired with Leica TCS SP5 confocal microscope with 10x objective. From each sample, 3 representative images were collected in different areas: frontal bone near bregma, central sinus of parietal bone and interparietal bone near lambda. The TH+ area was analyzed with ImageJ. For whole mount staining of long bones, tibia or femur BM thick sections were obtained with a cryostat and remaining OCT was removed by PBS washes. Samples were blocked and permeabilized in 0.1% Triton X-100 (Sigma) TNB (0.1 M Tris–HCl, pH7.5, 0.15 M NaCl, 0.5% blocking reagent, Perkin Elmer) overnight at 4°C. On the next day, samples were incubated with rabbit anti-TH (Merck AB152; 1:200), goat anti-CD31 (R&D AF3628 1:100), rat anti-EMCN (Insight Biotechnology sc-65495 1:100) or rat anti-CD31 (BD Biosciences Clone MEC13.3; 1:200) diluted in 0.1% Triton X-100 TNB overnight at 4°C. Samples were rinsed with 0.05% Triton X-100 PBS for 8 hours and incubated overnight with secondary antibodies including A546 Donkey anti-rabbit IgG (Invitrogen A10040; 1:200), Dylight650 donkey anti-rat IgG (Thermo Fisher SA5-10029; 1:300), Cy3-donkey anti-goat IgG (Jackson 705-165-147, 1:300) diluted in TNB. Samples were rinsed with 0.05% Triton X-100 PBS for 4 hours, washed with PBS and counterstained with DAPI (Sigma; 1:1000) to label cell nuclei. Images were acquired with Leica SP5 confocal microscope using 10x, 20x and 40x objectives and analyzed with ImageJ. At least 3 independent and randomly selected BM areas were imaged and analyzed per sample. To quantify neural fibers, total TH+area was normalized per total BM area labeled with DAPI. For quantification of sinusoidal vessel area, EMCN+ area > 150 μm distant to the bone surface was measured and normalized to BM area. Arteriolar (> 6 μm diameter) and capillary (< 6 μm diameter) CD31hiEMCN- vessel number was counted and normalized to BM area. To quantify transition zone vessels, CD31hiEMCNhi area < 150 μm close to the bone surface was quantified and normalized to the bone surface length.

35For staining of megakaryocyte and megakaryocyte progenitor (Figures 3C–3F, 4J–4L, 5D–5F, 7G–7I, and S2A–S2D), BM femur thin sections were blocked with TNB for 1h at RT. Samples were then incubated with primary antibodies for 2 h at RT or overnight at 4°C, followed by secondary antibodies incubation for 1h at RT. Finally, samples were counterstained with DAPI to label cell nuclei. Repetitive washes were performed with PBS. At least 3 independent and randomly selected BM areas were imaged and analyzed by ImageJ. The following primary antibodies were used: CD41 (1:200, PE conjugated rat monoclonal antibody, BD PharMingen 558040), CD42d (1:100, APC conjugated armenian hamster monoclonal antibody, eBioscience 17-0421-80) and ckit (1:100, goat polyclonal antibody, R&D AF1356). The following secondary antibodies were used: Alexa Fluor 647 Goat-anti-armenian hamster IgG (1:200, Abcam ab173004) and Alexa Fluor 647 donkey-anti-goat IgG (1:200, Life Technologies A21447).

36For staining of megakaryocyte localization relative to sinusoids (Figures 3I–3L and 5H–5J), BM femur thin sections were blocked with 0.1% Triton X-100 TNB for 1h at RT. Samples were then incubated with rat anti-EMCN (Insight Biotechnology sc-65495 1:100) antibody diluted in 0.1% Triton X-100 TNB overnight at 4°C. On the next day, samples were incubated with Alexa Fluor 647 goat-anti-rat IgG (1:200, Life Technologies, cat. no. A21247) secondary antibody for 1h at RT, followed by rat serum blocking (1:10 rat serum in 0.05% Triton X-100 TNB) for 10 min. Samples were incubated with third antibody, CD41 (1:200, PE conjugated rat monoclonal antibody, BD PharMingen 558040) for 2h at RT. Finally, samples were counterstained with DAPI to label cell nuclei. Repetitive washes were performed with PBS or 0.05% Triton X-100 PBS. At least 3 independent and randomly selected BM areas were imaged and analyzed by ImageJ.

37For HSCs and megakaryocytes staining (Figures 7, S5, and S6), BM femur thin sections were blocked with TNB for 1h at RT and followed by Avidin/biotin blocking. Samples were then incubated with rat anti-CD150 (1:50, Biolegend Cat. No. 115902) and armenian hamster anti-CD42d (1:100, eBioscience 17-0421-80) primary antibodies overnight at 4°C. On the next day, samples were incubated with Alexa Fluor 555 goat-anti-rat IgG (1:200, Life Technologies A21434) and Alexa Fluor 647 Goat-anti-armenian hamster IgG (1:200, Abcam ab173004) secondary antibodies for 1h at RT, followed by rat serum blocking (1:10 rat serum in TNB) for 10 min. Samples were re-stained with biotinylated lineage antibodies (1:100, BD Biosciences) for 2h at RT. Finally, samples were incubated with Alexa Fluor 488 Streptavidin-conjugated antibody (1:200, Invitrogen S32354) for 1h at RT and counterstained with DAPI to label cell nuclei. At least 3 independent and randomly selected BM areas were imaged and analyzed by ImageJ.

38For staining of human cord blood CD34+ HSPC co-culture (Figures 4D and 4E), cells were fixed with 1% PFA for 20 mins at RT. After PBS washes, cells were blocked with 10% donkey serum (Sigma D9663) diluted in PBS for 1h at RT, followed by 2h incubation of mouse-anti-human CD61 primary antibody (1:500, Serotec MCA 728). After PBS washes, cells were incubated with Alexa Fluor 546 donkey-anti-mouse IgG (1:200, Life Technologies, cat. no. A10036) secondary antibody for 1h at RT. Finally, samples were counterstained with DAPI to label cell nuclei. 10 independent and randomly selected fields in culture wells were imaged and analyzed by ImageJ.

39RNA isolation was performed using Trizol Reagent (Sigma T9424). Reverse transcription was performed using the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems 4368814), following the manufacturer’s recommendations. qPCR was performed using the PowerUp SYBR Green Master Mix (Applied Biosystems A25742) and ABI PRISM® 7900HT Sequence Detection System. The expression level of each gene was determined by using the absolute quantification standard curve method. All values were normalized with Gapdh as endogenous housekeeping gene.

40The following primers (mouse) were used:NOS1-Fw: ACTGACACCCTGCACCTGAAGANOS1-Rv: GTGCGGACATCTTCTGACTTCCNOS2-Fw: CAGCTGGGCTGTACAAACCTTNOS2-Rv: CATTGGAAGTGAAGCGGTTCGNOS3-Fw: CCTCGAGTAAAGAACTGGGAAGTGNOS3-Rv: AACTTCCTTGGAAACACCAGGGGapdh-Fw: GCATGGCCTTCCGTGTTCGapdh-Rv: CTGCTTCACCACCTTCTTGAT

41NOS1-Fw: ACTGACACCCTGCACCTGAAGA

42NOS1-Rv: GTGCGGACATCTTCTGACTTCC

43NOS2-Fw: CAGCTGGGCTGTACAAACCTT

44NOS2-Rv: CATTGGAAGTGAAGCGGTTCG

45NOS3-Fw: CCTCGAGTAAAGAACTGGGAAGTG

46NOS3-Rv: AACTTCCTTGGAAACACCAGGG

47Gapdh-Fw: GCATGGCCTTCCGTGTTC

48Gapdh-Rv: CTGCTTCACCACCTTCTTGAT

49MILLIPLEX MAP Mouse Cytokine/Chemokine Magnetic Bead Panel (MCYTOMAG-70K, Merck Millipore) was performed following the manufacture’s protocol. Extracellular fluids of mouse long bones were collected from the supernatants following BM extraction and subjected to the mouse cytokine panel.

50NO(X) content was measured in freshly-thawed samples that had been kept at −80°C for less than 2 months using a nitric oxide analyzer (NOA) 280i (Siever, GE Healthcare) according to the manufacturer’s instructions. Data were collected, processed and analyzed by using liquid software (Siever, GE Healthcare).

51Statistical analyses and graphics were carried out with GraphPad Prism 5 software and Microsoft Excel.

52The published article includes all relevant datasets generated or analyzed during this study.