Brain-Derived Acetylcholine Maintains Peak Bone Mass in Adult Female Mice

Abstract

Preclinical and clinical data support a role of the sympathetic nervous system in the regulation of bone remodeling, but the contribution of parasympathetic arm of the autonomic nervous system to bone homeostasis remains less studied. In this study, we sought to determine whether acetylcholine (ACh) contributes to the regulation of bone remodeling after peak bone mass acquisition. We show that reduced central ACh synthesis in mice heterozygous for the choline transporter (ChT) leads to a decrease in bone mass in young female mice, thus independently confirming the previously reported beneficial effect of ACh signaling on bone mass accrual. Increasing brain ACh levels through the use of the blood brain barrier (BBB)-permeable acetylcholinesterase inhibitor (AChEI) galantamine increased trabecular bone mass in adult female mice, whereas a peripheral increase in ACh levels induced by the BBB-impermeable AChEI pyridostigmine caused trabecular bone loss. AChEIs did not alter skeletal norepinephrine level, and induced an overall increase in osteoblast and osteoclast densities, two findings that do not support a reduction in sympathetic outflow as the mechanism involved in the pro-anabolic effect of galantamine on the skeleton. In addition, we did not detect changes in the commitment of skeletal progenitor cells to the osteoblast lineage in vivo in AChEI-treated mice, nor a direct impact of these drugs in vitro on the survival and differentiation of osteoblast and osteoclast progenitors. Last, ChT heterozygosity and galantamine treatment triggered bone changes in female mice only, thus revealing the existence of a gender-specific skeletal response to brain ACh level. In conclusion, this study supports the stimulatory effect of central ACh on bone mass accrual, shows that it also promotes peak bone mass maintenance in adult mice, and suggests that central ACh regulates bone mass via different mechanisms in growing versus sexually mature mice.

Introduction

The control of organ function is in part driven by the coordinated action of the sympathetic (SNS) and the parasympathetic (PSNS) nervous systems, the two arms of the autonomic nervous system. Sympathetic nerves mediate their actions via norepinephrine (NE), which binds to alpha and beta adrenergic receptors (αARs and βARs, respectively), whereas parasympathetic nerves signal via acetylcholine (ACh), which binds to muscarinic and nicotinic acetylcholine receptors (mAChRs and nAChRs, respectively) (see Robertson and Biaggioni⁽¹⁾ for review).

The skeleton, like most organs of the body, receives cues from the central nervous system (CNS) via hormones and peripheral nerves. Sensory and sympathetic nerves innervate the periosteal and cancellous compartments of bones, as evidenced by classical immunohistochemistry methods and more recently by gene reporter analyses based on nerve-specific markers.^(2–12) Rodent and human osteoblasts express βARs, mAChRs, and nAChRs, as well as enzymes responsible for the uptake and catabolism of NE and ACh, and thus have the potential to respond to and/or to modify the action of neuropeptides released from noradrenergic and cholinergic nerve terminals.^(13–19)

Pharmacological and genetic manipulations of βAR signaling in mice and rats have provided strong evidence to support the model whereby bone remodeling is under the control of SNS signaling in both physiological and pathophysiological conditions (see Elefteriou⁽²⁰⁾ for review). Mice deficient in the NE-synthesizing enzyme dopamine β-hydroxylase (DBH) or for the β2 adrenergic receptor (β2AR) (globally or specifically in osteoblasts) all showed a normal bone mass at 3 months of age but a high trabecular bone mass at 6 months of age.^(21–23) Mice and rats receiving βAR antagonists also exhibited a high bone mass phenotype^(21,24–28) and conversely, the βAR agonist isoproterenol induced bone loss in mice.^(21,29,30)

Experimental evidence to support a functional impact of ACh on the skeleton also exist, although the mechanism of action involved remains less characterized. Genetic loss-of-function studies in mice revealed a low bone mass phenotype upon global ablation of nAChRs.⁽¹⁷⁾ Another study showed that lack of the M₃ muscarinic receptor in the CNS of M₃r-deficient mice reduced bone mass accrual by decreasing bone formation and increasing bone resorption at all ages examined (6, 12, and 24 weeks of age), whereas M₃r-deficiency in osteoblasts (generated by the α1(I)Collagen-Cre transgene) did not affect bone mass.⁽³¹⁾ Conversely, pharmacological gain-of-function approaches to raise ACh levels, based on the use of the acetylcholinesterase inhibitor (AChEI) donepezil in wild-type (WT) mice showed that this drug increased tibial and vertebral bone volume fraction upon administration in growing, sexually immature 5-week-old female mice.⁽³²⁾ Experimental evidence also supported the notion that the central effect of M₃r deficiency on bone mass was mediated by an increase in sympathetic outflow.⁽³¹⁾

Cardiovascular studies have shown that an increased sympathetic tone and a reduction in parasympathetic tone contribute to the elevation of blood pressure associated with age,^(33–35) especially in women.⁽³⁶⁾ Sympathetic activity was found higher in postmenopausal women compared to premenopausal women and was inversely correlated with trabecular bone fraction.⁽³⁷⁾ These observations and the late onset high bone mass phenotype of the Dbh^−/− and β2 adrenoreceptor (Adrb2^−/−) mice suggested a relationship between the impact of the autonomic nervous system on bone remodeling and age or gonadal status. In this study, we asked whether increasing ACh levels in adult mice, after peak bone mass acquisition and sexual maturity, had a beneficial impact on the mouse skeleton, and whether this effect would be mediated by a central or peripheral action of ACh, independently of potential brain plasticity–related compensatory mechanisms induced by embryonic gene deletion approaches.

Materials and Methods

Animals

WT mice used in this study were on the C57BL/6J background and from The Jackson Laboratory (Bar Harbor, ME, USA). Choline transporter heterozygous deficient (ChT^HET) mouse breeders were provided by Dr. Randy Blakely (Florida Atlantic University, Boca Raton, FL, USA). Twelve-month-old female mice/15-month old male mice were treated daily for 6 weeks via intraperitoneal injection (i.p.) with galantamine hydrobromide (Sigma-Aldrich, St. Louis, MO, USA; G1660) at 2.5 mg/kg/day or vehicle, and were euthanized at 13.5 and 16.5 months of age. Fourteen-week-old female/male mice were treated daily for 6 weeks with galantamine hydrobromide or pyridostigmine bromide (Sigma-Aldrich; P9797) at 2.5 mg/kg/day and 1 mg/kg/day, respectively. Pyridostigmine was administered via mini-osmotic pumps implanted subcutaneously (Durect Corporation, Cupertino, CA, USA; 0000298) to reduce biohazard concerns.⁽³⁸⁾ All procedures were approved by the Institutional Animal Care and Use Committee at Baylor College of Medicine.

Micro–computed tomography analysis

Micro–computed tomography (μCT) analyses were performed according to ASBMR guidelines⁽³⁹⁾ using a Scanco μCT 40 system (Scanco Medical, Bassersdorf, Switzerland). Tomographic images were acquired at 55 kVp and 145 μA with an isotropic voxel size of 12 mm and at an integration time of 300 ms. To segment bone from nonmineralized tissue, a Gaussian noise suppression filter (sigma = 0.8, support = 1) was used, and global thresholds were consistent across scans per anatomical site.

Histomorphometry

Bones were dehydrated and embedded undecalcified in methyl methacrylate. Histomorphometric measurements were performed using the Bioquant Image Analysis System (R&M Biometrics, Nashville, TN, USA) according to ASBMR guidelines.⁽⁴⁰⁾

Serum calcium/phosphate and brain/bone ACh and NE assays

Serum calcium was measured with the Calcium Colorimetric Assay kit (BioVision, Milpitas, CA, USA; K380-250). Serum phosphate was measured with the Phosphate Colorimetric Assay Kit (BioVision; K410-500). Brain/bone ACh and NE were measured by HPLC (Vanderbilt University Neurochemistry Core, Nashville, TN, USA) as described in.^(41,42)

Primary cell culture

Bone marrow stromal cells (BMSCs) were isolated from long bones of 20-week old female C57BL6 WT mice, treated with AChEIs for 6 weeks, and plated at a density of 1 × 10⁶ cells/mL. Fifty μg/mL ascorbic acid and 5 mmol/L glycerophosphate were added to culture medium at cell confluence, 7 days after plating. ALP staining and Von Kossa staining were performed 21 days after plating. The number of Cfu-F, Cfu-AP and Cfu-Ob was counted manually.

To assess the effects of AChEIs on osteoblast and osteoclast differentiation in vitro, BMSCs and monocytes were treated with 0.1, 1, 10 μmol/L AChEIs supplemented with 10% fetal bovine serum, 100 IU/mL penicillin, and 100 μg/mL streptomycin. Culture medium and AChEIs were refreshed every other day. BMSCs were isolated from long bones of adult C57 WT mice. For BMSC proliferation, cells were plated at a density of 2 × 10⁶ cells/mL. After BMSCs reached 80% confluence, cells were trypsinized and replated in 24-well plates at a density of 1 × 10⁵ cells/mL. AChEI treatments were started 1 day after plating. The cells were fixed at different time points. Crystal violet staining was performed and the OD of the released dye (OD₅₇₀ nm) was used to quantify relative cell number. For osteoblast differentiation, BMSCs were plated at a density of 1 × 10⁶ cells/mL and differentiated by adding ascorbic acid (50 μg/mL) and glycerophosphate (5 mmol/L) to culture medium 7 days after plating. RNAs were collected 21 days after plating.

Monocytes were prepared from spleens of C57 WT mice by the Ficoll (Lymphocyte Separation Medium; MP Biomedicals, Solon, OH, USA) gradient method and plated at a density of 3 × 10⁶ cells/mL. Osteoclast differentiation was induced with 30 ng/mL of Macrophage Colony-Stimulating Factor (MCSF, Sigma) and 50 ng/mL of Receptor Activator of Nuclear Factor B Ligand (RANKL, R & D Systems, Minneapolis, MN, USA). Osteoclast differentiation induction and AChEIs treatment started the day when cells were plated. Culture medium and AChEIs treatment were refreshed every other day. At day 6, tartrate-resistant acid phosphatase (TRAP) staining was performed. TRAP+ cells containing between 3 and 10 nuclei were counted manually.

Semiquantitative and quantitative real-time RT-PCR

Semiquantitative RT-PCR was performed using the following primers:

• AChE: Fw: 5′-ATGACCCTCGAGACTCCAAA-3′, Rev: 5′-TCCGCCTCGTCCAGAGTAT-3′ (Tm: 55°C);

• BChE: Fw: 5′-ACACAGACCCACTTCCTCCT-3′, Rev: 5′-GTGCATAGGGGATACCGAGA-3′ (Tm: 55°C);

• ChAT: Fw: 5′-CCATTGTGAAGCGGTTTGGG-3′, Rev: 5′-GCCAGGCGGTTGTTTAGATACA-3′ (Tm: 60°C);

• VAChT: Fw: 5′-TTGATCGCATGAGCTACGAC-3′, Rev: 5′-AGGCTCCTCGGGATACTTGT-3′ (Tm: 55°C);

• ChT: Fw 5′-CATCCTCAGCCACCTATGCT-3′, Rev: 5′-TGGATACCCGTAGGCAGTCT-3′ (Tm: 60°C);

• Ocn: Fw: 5′-ACCCTGGCTGCGCTCTGTCTCT-3′, Rev: 5′-GATGCGTTTGTAGGCGGTCTTCA-3′ (Tm: 65°C);

• Ctr: Fw: 5′-TGCTGGCTGAGTGCAGAAACC-3′, Rev: 5′-GGCCTTCACAGCCTTCAGGTAC-3′ (Tm: 64°C);

• Hprt: Fw: 5′-AGCGATGATGAACCAGGTT-3′, Rev: 5′-GTTGAGAGATCATCTCCACC-3′ (Tm: 55°C).

PCR cycling conditions: 5 min at 95°C, 30-s denaturation at 95°C, 30-s annealing at temperature indicated above, 45-s extension at 72°C, repeated 35 times.

Quantitative real-time PCR was performed using TaqMan^® gene expression assays from Applied Biosystems (Hprt1, Mm00446968_m1; Rankl, Mm00441908_m1; Alpl, Mm00475834_m1; Opg, Mm00435451_m1), and iQ SYBR Green Supermix (Cat#1708882, Bio-Rad) with the following primers:

• β2mg: Fw 5′-TTCGGTGCTTGTCTCACTGA-3′, Rev: 5′-CAGTATGTTCGGCTTCCCATTC-3′;

• Il1β: Fw: 5′-GAAATGCCACCTTTTGACAGTG-3′, Rev: 5′-TGGATGCTCTCATCAGGACAG-3′;

• Il6: Fw: 5′-TAGTCCTTCCTACCCCAATTTCC-3′, Rev: 5′-TTGGTCCTTAGCCACTCCTTC-3′;

• TNFα: Fw: 5′-GCTACGACGTGGGCTACAG-3′, Rev 5′-CCCTCACACTCAGATCATCTTCT-3′.

Specificity of amplification was verified by melting curve analysis and presence of a single peak.

Statistics

All data are presented as means SD. One-way ANOVA statistical analysis was performed for multiple-group comparisons and it was followed by post hoc comparison with Dunnett’s t tests (two-sided) when statistical significance was found by one-way ANOVA. Unpaired two-tailed Student’s t tests were performed for two-group comparisons. For all analyses, *p < .05 was considered significant.

Results

Genetic reduction in ACh level via choline transporter (ChT) deficiency inhibits bone mass accrual

The transport of choline into neurons is a rate-limiting step for the synthesis of ACh.⁽⁴³⁾ To determine the effect of reduced ACh level on bone mass, we took advantage of mutant mice heterozygous for a deletion of the high affinity, hemicholinium-3-sensitive choline transporter 1 (Slc5a7, herein called ChT^HET mice).⁽⁴⁴⁾ ChT^KO mice are lethal, but ChT^HET mice are viable and have roughly one-half the levels of brain ACh as compared to WT littermates.⁽⁴⁵⁾ Using three-dimensional micro-computed tomography (μCT) analyses, we detected a 6% reduction in vertebral bone volume fraction in ChT^HET female mice compared to control WT mice at 3 months of age (Fig. 1A), associated with a reduction in trabecular thickness (Table 1) but no significant change in bone cell densities (Supplementary Table 1). No detectable difference in vertebral and femoral bone volume fraction was detected at 6 and 12 months of age in both genders (Fig. 1A,B and data not shown), although ChT ^HET males, like females, showed a reduction in trabecular thickness at 3 months of age (Table 1). These results are in line with the reduced bone mass of M₃r-deficient young female mice,⁽³¹⁾ and although effect size is mild, independently confirm that ACh signaling promotes bone mass accrual in young mice.

Fig. 1.

ChT heterozygosity leads to reduced peak bone mass accrual in young female mice. μCT analyses of vertebral trabecular bone in female (A) and male (B) WT and ChT^HET mice at 3, 6, and 12 months of age (n = 8–17 per group, unpaired two-tailed Student’s t tests). ChT = choline transporter.

Table 1.

μCT Analyses of Vertebral Trabecular Bone Structure in 3-Month-Old WT and ChT^HET Mice

	Female			Male
Parameter	WT	Het	p	WT	Het	p
Tb.Th (mm)	0.051 ± 0.001	0.049 ± 0.002*	.002	0.051 ± 0.004	0.047 ± 0.002*	.004
Tb.N (1/mm)	4.676 ± 0.338	4.698 ± 0.188	.827	5.530 ± 0.405	5.496 ± 0.326	.789
Tb.Sp (mm)	0.205 ± 0.017	0.205 ± 0.009	.983	0.168 ± 0.016	0.171 ± 0.013	.52
BMD (mgHA/cm3)	1012 ± 23	1002 ± 21	.227	1009 ± 20	1005 ± 23	.57

BMD = bone mineral density; HA = hydroxyapatite; Tb.N = trabecular number; Tb.Sp = trabecular separation; Tb.Th = trabecular thickness.

^*Value of p: ChT^HET versus WT, n = 14–18 per group, unpaired two-tailed Student’s t tests.

The AChEI galantamine does not promote bone gain in aging mice

Compensations arising from neuronal plasticity following embryonic genetic loss of ACh transport or signaling can potentially have indirect repercussions on adult physiology and bone homeostasis. To exclude this putative confounding factor of non-inducible genetic models and to determine if a reduction in ACh signaling contributes to age-related bone loss, we sought to increase ACh levels pharmacologically in aging WT mice using AChEIs, which are drugs clinically used to treat the cholinergic deficiency of patients with Alzheimer disease (AD). We chose to administer the AChEI galantamine at a dose of 2.5 mg/kg/day (i.p.) to 12-month-old female mice and 15-month-old male mice, which are time points associated with severe age-related low bone mass in mice. Following a treatment period of 6 weeks and although an effective and selective increase in brain ACh levels was measured for at least 6 hours post-galantamine injection (Fig. 2A), no difference in vertebral and femoral trabecular bone volume (Fig. 2C–F), nor trabecular parameters (Supplementary Tables 2 and 3) was detected between galantamine and vehicle groups, in both males and females. ACh levels did not change in bone upon galantamine treatment (Fig. 2B).

Fig. 2.

Galantamine does not promote bone gain in aging mice. Brain (A) and bone (B) ACh levels over 6 hours following a single injection of Gal in C57BL6 female mice, measured by HPLC and normalized by protein content (n = 4 per group, one-way ANOVA followed by two-sided Dunnett’s t tests). (C–F) μCT analyses of vertebral and femoral trabecular bone in 13.5-month-old WT C57BL6 female mice (C,D) and 16.5-month-old male mice (E,F) treated with Gal for 6 weeks (n = 8–10 per group, unpaired two-tailed Student’s t tests). ACh = acetylcholine; Gal = galantamine.

Peripherally versus centrally-acting AChEIs have differential effects on bone mass in adult mice

Because of the experimental challenge in producing a bone anabolic response in aging mice that have a minimal trabecular bone mass at the start of treatment and within a 6-week time frame compatible with daily injections, we sought to determine whether AChEIs may have a detectable effect on the skeleton of younger mice at the earliest stage of bone loss, which occurs between 2 to 3 months of age in spine and long bones⁽⁴⁶⁾ (and as early as 25 to 30 years of age in humans).⁽⁴⁷⁾ In addition, we sought to determine whether this effect would be mediated via a central or peripheral action of ACh, using AChEIs characterized by their differential ability to cross the blood brain barrier (BBB). We chose the BBB-permeable, centrally-acting drug galantamine (2.5 mg/kg/day, i.p.) to increase ACh signaling in the CNS, and the BBB-impermeable drug pyridostigmine (1 mg/kg/day via a subcutaneous micro-osmotic pump) to increase ACh signaling peripherally.⁽⁴⁸⁾ Dosage was chosen based on previous studies in mice.^(17,49,50)

Over a 6 week-long treatment period, started at peak bone mass in 3.5-month-old female mice, the peripherally-acting drug pyridostigmine induced femoral (−33%) and vertebral (−13.2%) cancellous bone loss compared to the vehicle-treated group, as assessed by 3D μCT analyses (Fig. 3A,B). The effect of pyridostigmine on bone mass was measured in females only because of experimenter safety concerns. This bone loss induced by pyridostigmine was associated with an increase in bone Rankl/Opg ratio and Il1b mRNA expression (but not change in Il6 and Tnfa) (Fig. 3C–F) and an increase in osteoclast surface and number, but no detectable change in osteoblast parameters (Table 2). In contrast, the centrally-acting drug galantamine led to an increase in cancellous BV/TV in both femurs (+24.3%) and vertebrae (+12.3%) compared to vehicle-treated mice (Fig. 3A,B). This bone anabolic effect was associated with an increase in osteoblast surface and number, bone formation rate/bone surface, and mineralized surface/bone surface, but no significant change in osteoclast parameters (Table 2) nor in the expression of osteoclastogenic genes (Fig. 3C–F). AChEIs did not induce significant changes in serum calcium/phosphate level nor body weight (Fig. 3G–I), with the exception of a 6% increase in body weight observed in pyridostigmine-treated mice. Notably, galantamine treatment did not trigger any detectable change in bone volume fraction in male mice (Fig. 3J,K).

Fig. 3.

Centrally-acting Gal increases bone mass whereas peripherally-acting Pyr decreases bone mass. μCT analyses of femoral (A) and vertebral (B) trabecular bone in 20-week-old female mice treated with Vehicle (Con), Gal, or Pyr (n = 9 per group, one-way ANOVA followed by two-sided Dunnett’s t tests). Bone β₂mg-normalized Rankl/Opg ratio (C), Il1β (D), Il6 (E), and Tnfa (F) relative expression measured by RT-qPCR (n = 7–10 per group, one-way ANOVA followed by two-sided Dunnett’s t tests). Serum Ca (G), Ph levels (H), and BW (I) in adult female mice treated with AChEIs for 6 weeks (n = 9–10 per group, one-way ANOVA followed by two-sided Dunnett’s t tests). μCT analyses of the femoral (J) and vertebral (K) trabecular bone in 20-week-old male mice (n = 9 per group, unpaired two-tailed Student’s t tests). BW = body weight; Ca = calcium; Gal = galantamine; Ph = phosphate; Pyr = pyridostigmine.

Table 2.

Histomorphometric Analyses of Vertebral Trabecular Bone in 20-Week-Old Female Mice Treated with Gal or Pyr

Parameter	Con	Gal	p	Pyr	p
Ob.S/BS (%)	15.33 ± 6.84	23.91 ± 4.40*	.008	19.11 ± 3.30	.261
N.Ob/B.Pm (#/mm)	10.34 ± 4.79	16.32 ± 3.12*	.009	13.03 ± 2.53	.262
Oc.S/BS (%)	7.86 ± 2.06	10.97 ± 3.33	.068	11.45 ± 2.65*	.028
N.Oc/B.Pm (#/mm)	4.18 ± 1.14	5.87 ± 1.80	.056	6.15 ± 1.47*	.026
BFR/BS (μm/day)	0.16 ± 0.08	0.34 ± 0.14*	.026	0.17 ± 0.11	.999
MAR (μm/day)	0.82 ± 0.10	0.89 ± 0.09	.284	0.73 ± 0.09	.136
MS/BS (%)	10.08 ± 2.56	13.21 ± 3.30*	.023	9.43 ± 3.11	.905

BFR/BS = bone formation rate per bone surface; Gal = galantamine; MAR = mineral apposition rate; MS/BS = mineralized surface per bone surface; N.Ob/B.Pm = osteoblast number per bone perimeter; N.Oc/B.Pm = osteoclast number per bone perimeter; Ob.S/BS = osteoblast surface per bone surface; Oc.S/BS = osteoclast surface per bone surface; Pyr = pyridostigmine.

^*Value of p: treatment group versus vehicle control, n = 6–8 per group, one-way ANOVA followed by two-sided Dunnett’s t tests.

AChEIs do not affect bone cells directly

Histomorphometric analyses indicated that the bone gain induced by galantamine was associated with an increased number and activity of osteoblasts but no reduction in osteoclasts parameters, an observation inconsistent with low sympathetic outflow, known to alter the density of these two cell lineages in opposite manner. In addition, we did not detect a reduction in bone NE content between treatment groups 24 hours post-AChEI injections (vehicle: 2.72 ± 0.92 versus galantamine: 2.19 ± 0.50). These two sets of outcomes suggested that in adult mice, galantamine treatment increased bone mass via a SNS-independent mechanism, leading us to ask whether AChEIs might exert this effect by a direct action on skeletal cells, because multiple ACh receptor subtypes have been detected in osteoblast and osteoclasts.^{(13–17,19,51–53)} We were able to detect expression of AChE (encoding acetylcholinesterase) and BChE (encoding butyrylcholinesterase) in mouse calvaria osteoblasts and bone marrow stromal cells (BMSCs), but not in monocyte/osteoclast cultures (Fig. 4A). Overall level of expression was much lower than in the brain, although the different nature of the tissues investigated (brain tissue and cells in culture) calls for caution in interpreting this difference. These results are in agreement with previous studies demonstrating immunoreactivity for AChE in osteoblasts^(18,19) and BChE gene expression in human and mouse osteoblast cell lines.⁽¹⁵⁾ In contrast, we did not detect expression of ChAT (encoding choline acetyltransferase) nor VAChT (encoding vesicular acetylcholine transporter) in either osteoblast or osteoclast cultures, regardless of their differentiation stage (expression of both genes was detected, as expected, in the brain), whereas expression of ChT was detected in BMSC cultures but not primary calvarial osteoblasts (Fig. 4A). Osteoblasts thus have the ability to catabolize ACh via actions of AChE and BChE, but cannot synthesize it, and AChEIs have the potential to raise extracellular ACh levels in bone and thus impact bone remodeling via a peripheral action (provided ACh is released by cells of the bone microenvironment other than osteoblasts). It must be emphasized, however, that we detected in the mouse CNS a 50-fold higher level of ACh compared to bone, and galantamine treatment increased this difference to >200-fold, without affecting bone ACh level. (Fig. 2A,B).

Fig. 4.

AChEIs have no direct effects on osteoblast viability and differentiation. (A) RT-PCR analysis of AChE, BChE, ChAT, VAChT, ChT, and Hprt expression. Brain RNA served as positive control. Ctr and Ocn (Bglap) were used to validate the identity of osteoclast and osteoblast cultures, respectively (n = 3). (B) Number of Cfu-F, Cfu-AP, and Cfu-Ob colonies formed by BMSCs prepared from vehicle and AChEIs-treated female mice, following 14 days of in vitro differentiation in osteogenic medium (n = 8–11 per group, one-way ANOVA). (C) Cell number measured by crystal violet staining of BMSCs treated with AChEIs for 1, 3, and 5 days (n = 4, one-way ANOVA). (D) Hprt-normalized Alpl gene expression in BMSCs treated with AChEIs for 14 days (n = 3, one-way ANOVA). (E) Osteoclast differentiation measured by the number of TRAP-positive cells containing three to 10 nuclei per cell, following treatment with AChEIs for 6 days (five areas per dose per group were measured; n = 4, one-way ANOVA). A representative image shows the type of osteoclasts counted. AChE = acetylcholinesterase; Alpl = alkaline phosphatase; BChE = butyrylcholinesterase; ChAT = choline acetyltransferase; ChT = choline transporter; Ctr = calcitonin receptor; Ocn = osteocalcin; VAChT = vesicular acetylcholine transporter; Hprt = hypoxanthine phosphoribosyltransferase 1; TRAP = tartrate-resistant acid phosphatase.

To determine whether AChEIs had an impact on osteoblast progenitors in vivo, BMSCs were isolated from the long bones of 20-week old female WT mice treated with AChEIs for 6 weeks, plated at equal density and differentiated for 2 weeks in osteogenic conditions to enumerate the percentage of fibroblast-positive colony-forming units (Cfu-F), alkaline phosphatase–positive colony-forming units (Cfu-AP), and osteoblast-positive colony-forming units (Cfu-Ob). We did not, however, detect differences in the number of colonies between treatment groups (Fig. 4B), thus indicating that the administration of AChEIs did not affect the in vivo commitment of skeletal stem cells to the osteoblast lineage.

AChEIs are known to modulate AChR signaling in addition to inhibiting AChEs^(54,55) and in vitro evidence suggested that AChE activity promotes osteoclast formation.⁽⁵⁶⁾ Therefore, we asked whether the AChEIs used in this study may have a direct, ACh-dependent or independent effect on osteoblast and osteoclast viability and differentiation. To address this question, BMSCs were prepared and treated with vehicle, galantamine, or pyridostigmine at three different doses (0.1μM, 1μM, and 10μM) reported to enhance ACh signaling in HEK-293 kidney cells in vitro.⁽⁵⁵⁾ This treatment over 5 days did not affect BMSC number, suggesting it does not change the proliferation or survival of these cells (Fig. 4C). Upon growth in osteogenic medium for 14 days, BMSC differentiation assessed by Alpl gene expression did not differ either between groups (Fig. 4D). The same treatment had no major effect on osteoclast differentiation either, measured by the number of tartrate-resistant acid phosphatase (TRAP)-positive spleen monocytes following 6 days of differentiation in the presence of M-CSF and RANKL (Fig. 4E). These results thus do not support a direct, ACh-independent, action of AChEIs on BMSCs or osteoclasts.

Discussion

In this study, a pharmacological gain-of-function strategy comparing the effect of two distinct AChEIs with different BBB permeability provided evidence, in adult mice, that ACh in the CNS promotes the maintenance of peak bone mass in a gender-specific manner, whereas peripheral ACh favors bone loss. The notion that ACh in the CNS promotes the maintenance of peak bone mass is supported by several experimental results, including the selective rise in central ACh levels upon galantamine treatment, the higher expression of AChE and BChE in the CNS versus bone cells and the lack of direct effect of this drug on osteoblast and osteoclast viability and differentiation. These results in sexually mature adult mice and in the context of bone remodeling extend prior findings in young growing mice where another BBB-permeable AChEI (donepezil) was shown to promote bone mass accrual in the period of active bone modeling.^(31,32) Results are also relevant to the use of AChEIs for the treatment of various forms of dementias, including Alzheimer disease, and their protective effect on the skeleton, although they apply to CNS tissues not yet impacted by the disease.^(57–60)

A noticeable difference with prior studies in sexually immature mice^(31,32) is the lack of evidence to support a change in sympathetic outflow as the main mechanism involved in the CNS action of galantamine/ACh on bone mass in adult mice. Shi and colleagues⁽³¹⁾ indeed reported cellular changes consistent with an increase in SNS outflow in young M₃ r^−/− mice (increased osteoclast number and decreased osteoblast numbers associated with low bone mass), whereas Eimar and colleagues⁽³²⁾ showed that donepezil/high brain ACh levels induced bone gain in 5-week-old female mice via a reduction in osteoclast number, but no detectable effect on bone formation parameters. This reduction in osteoclast number is consistent with a reduction in sympathetic outflow; however, reduced NE level in bone is expected to augment osteoblast number and bone formation,⁽⁶¹⁾ which was not observed in their study. In contrast, we observed that the cellular changes associated with galantamine/high ACh brain levels in sexually mature female mice are limited to an increase in osteoblast number and activity but no reduction in osteoclast number. Hence, histomorphometry results, in addition to the normal bone NE levels measured in galantamine-treated mice, do not support a reduction of SNS outflow as the mechanism promoting peak bone mass maintenance in galantamine-treated adult mice, and highlight the importance of age (or sexual hormones, see the last paragraph of the discussion) in the response of bone cells to changes in autonomic function.

The use of pyridostigmine to selectively raise ACh peripherally (ie, outside the CNS) revealed the opposite action of brain and peripheral ACh on bone, because pyridostigmine caused bone loss. One in vitro study showed an inhibitory and direct action of donepezil on osteoclast formation,⁽⁵⁶⁾ which we could not replicate, in line with the undetectable expression of the targeting enzymes (AChEs) in monocytes. The detection of AChEs in osteoblasts and the increase in osteoclast parameters, bone Rankl/Opg and Il1b expression in pyridostigmine-treated mice rather suggested that peripheral ACh promotes bone loss by increasing the osteoclastogenic potential of osteoblasts, which express multiple ACh receptor subtypes, or possibly osteocytes, whose production of RANKL was shown most relevant for osteoclastogenesis and bone/cartilage resorption.^(62,63) Although the ability of BBB-permeable and impermeable AChEIs allowed us to dissociate the central versus peripheral modes of action of ACh on bone remodeling, genetic approaches such as the use of ChT^HET mice provides an integrated sum of these mechanisms, which might explain the mild and transient phenotype of the latter mouse model, where the central bone anabolic effect of ACh might be antagonized by its peripheral catabolic action on the skeleton.

This study showed that an increase in central ACh level following galantamine administration in adult WT mice induced bone gain in females only, in contrast to sexually immature WT mice, in which donepezil-induced bone gain was observed in both genders.^(17,32) Although we cannot exclude intrinsic differences between the AChEIs used in these two studies, these observations support the hypothesis that gonadal hormones modify the CNS or peripheral action of ACh on bone remodeling. Interestingly, estrogen is known to exert trophic effects on the cholinergic system.^(64–68) Thus, not only age but also sex hormones may regulate the response of the skeleton to central ACh signaling in adults, via their action on osteoblasts/osteoclasts or CNS neurons.