Crosstalk between bone and brain in Alzheimer's disease: Mechanisms, applications, and perspectives

Abstract

Alzheimer's disease (AD) is a neurodegenerative disease that involves multiple systems in the body. Numerous recent studies have revealed bidirectional crosstalk between the brain and bone, but the interaction between bone and brain in AD remains unclear. In this review, we summarize human studies of the association between bone and brain and provide an overview of their interactions and the underlying mechanisms in AD. We review the effects of AD on bone from the aspects of AD pathogenic proteins, AD risk genes, neurohormones, neuropeptides, neurotransmitters, brain‐derived extracellular vesicles (EVs), and the autonomic nervous system. Correspondingly, we elucidate the underlying mechanisms of the involvement of bone in the pathogenesis of AD, including bone‐derived hormones, bone marrow‐derived cells, bone‐derived EVs, and inflammation. On the basis of the crosstalk between bone and the brain, we propose potential strategies for the management of AD with the hope of offering novel perspectives on its prevention and treatment.

Highlights

• The pathogenesis of AD, along with its consequent changes in the brain, may involve disturbing bone homeostasis.

• Degenerative bone disorders may influence the progression of AD through a series of pathophysiological mechanisms.

• Therefore, relevant bone intervention strategies may be beneficial for the comprehensive management of AD.

1. INTRODUCTION

Alzheimer's disease (AD) is the most common type of dementia. It is clinically characterized by progressive memory loss and cognitive decline in older adults. With an increasing number of patients being diagnosed with AD, it constitutes a serious burden to caregivers and society. ¹ , ² Senile plaques composed of amyloid beta (Aβ) peptide and neurofibrillary tangles consisting of hyperphosphorylated tau are the main neuropathological features of AD. ³ AD is conventionally regarded as a neurodegenerative disease of the brain. In recent years, numerous studies have shown that abnormalities in systemic organs and tissues contribute to brain AD‐type pathologies and cognitive impairment. ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ Large epidemiological studies have shown the associations of systemic comorbidities with AD and dementia risk. ¹¹ , ¹² , ¹³ Systemic factors from the peripheral system can enter the brain and mediate brain aging, neurodegeneration, and AD pathogenesis. ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ Thus, systemic organs and tissues are involved in the pathogenesis and progression of AD. Revealing the mechanisms by which the peripheral system affects AD occurrence and development could provide new insights for developing novel intervention strategies. ¹⁸

The skeletal system is one of the largest organs in the body and is traditionally viewed as a static organ that primarily provides support, protection, and movement. Bone functions are accomplished by osteoblasts (derived from mesenchymal cells with osteogenic functions), osteocytes (also thought to be bone‐resident cells), and osteoclasts (also thought to be bone resorption cells). In addition, bone marrow mesenchymal stem cells (BMSCs) and the bone microenvironment composed of inflammatory cells, endothelial cells, and Schwann cells may also be involved in maintaining bone homeostasis and bone repair. ¹⁹ , ²⁰ Bone can be directly innervated by efferent nerves originating from the central nervous system (CNS), while various central neurohormones (eg, follicle‐stimulating hormone, thyroid‐stimulating hormone, adrenocorticotrophic hormone), neuropeptides (eg, neuropeptide Y, kisspeptin, vasoactive intestinal peptide), neurotransmitters (eg, glutamate, acetylcholine, dopamine), and their intracellular signaling pathways are also implicated in the regulation of bone metabolism. ²¹ , ²² , ²³ , ²⁴ Moreover, increasing evidence has indicated the importance of bone as an endocrine organ that connects systemic organs of the whole body including the brain. ²⁵ , ²⁶ The relationships between AD and bone diseases in mice and humans have been reported, emphasizing the crosstalk between bone and the brain in AD. ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² Therefore, we reviewed the clinical evidence for the associations between bone diseases and AD and investigated the impact of AD on bone metabolism and the underlying mechanisms involved. Here, we also discuss the underlying mechanisms through which bone influences AD pathogenesis, including mechanisms involving bone‐derived proteins, bone marrow‐derived cells, bone‐derived extracellular vesicles (EVs), and inflammation. In light of the crosstalk between bone and brain, we propose some bone‐based intervention strategies for the prevention and treatment of AD.

RESEARCH IN CONTEXT

1. Systematic review: Recently, increasing evidence has indicated a relationship between Alzheimer's disease (AD) and bone diseases in mice and humans. This literature review was performed by searching PubMed, Embase, and Web of Science, aiming to discuss the crosstalk between AD and bone and propose potential bone‐based intervention strategies for the prevention and treatment of AD.

2. Interpretation: In this review, we summarize the effects of AD from the perspectives of AD pathogenic proteins, AD risk genes, neurohormones, neuropeptides, neurotransmitters, brain‐derived extracellular vesicles (EVs), and the autonomic nervous system. Moreover, bone‐derived hormones, bone marrow‐derived cells, bone‐derived EVs, and inflammation can influence brain AD pathology and cognitive function.

3. Future directions: More animal and clinical studies are needed to investigate the contribution of bone diseases and bone‐derived substances to AD pathogenesis. The identification of key molecules that promote AD in different bone diseases is also important. The active prevention and timely treatment of bone diseases in elderly people may be beneficial for the prevention and treatment of AD.

2. CLINICAL EVIDENCE FOR ASSOCIATIONS BETWEEN BONE DISEASES AND AD

Osteoporosis is a prevalent bone disorder characterized by a decrease in bone density and bone mass due to a variety of etiologies. A cohort study in 2005 demonstrated that low femoral neck bone mineral density (BMD) was associated with approximately twice the risk of AD and dementia in women, but no such association was observed in men. ³³ Subsequent studies and meta‐analyses further confirmed an increased incidence of cognitive decline in osteoporosis patients. ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ Therefore, it can be inferred that certain pathophysiological changes occurring in bones may impact brain function. Conversely, AD is also a risk factor for osteoporosis. Clinical research has indicated that individuals with dementia have a greater likelihood of poor bone health and are more prone to developing fractures. ³⁹ BMD was lower in an early AD group and was related to brain volume and memory ability, and higher BMD was associated with better memory performance in AD patients. ⁴⁰ , ⁴¹ These findings suggest that changes in the brain of AD patients might influence bone functions. Recently, a Mendelian randomization analysis of large‐scale BMD and AD gene datasets revealed that low BMD was a risk factor of AD. ⁴² Systematic reviews and meta‐analyses have shown associations among osteoarthritis, bone loss, and AD. ⁴³ , ⁴⁴ , ⁴⁵ A longitudinal study revealed that osteoarthritis was associated with earlier Aβ deposition and greater Aβ‐dependent tau deposition during follow‐up. ²⁸ Alterations in hippocampal functional connectivity and levels of brain‐derived neurotrophic factor (BDNF) are associated with future cognitive decline in patients with osteoarthritis. ⁴⁶ Therefore, a large amount of clinical evidence has associated bone diseases with AD and dementia (Table 1).

TABLE 1.

Clinical studies of an association between bone degenerative changes, cognitive impairment, and Alzheimer's disease pathologies.

Study	Type	Participants	Variables	Correlations
[4]	Cross‐sectional study	650 Koreans	BMD, MMSE	Positive
[28]	Observational study	374 participants from ADNI database	Osteoarthritis, Aβ and tau accumulation in primary motor and somatosensory regions	Positive
[32]	Prospective cohort study	3651 participants from Rotterdam Study	BMD, risk of developing AD	Negative
[33]	Prospective cohort study	987 participants from the Framingham study	BMD, risk of developing all‐cause dementia and AD	Negative
[34]	Retrospective cohort study	59966 Germans	Osteoporosis diagnosis, risk of developing dementia	Positive
[35]	Retrospective cohort study	157988 Koreans	Osteoporosis diagnosis, risk of developing AD	Positive
[36]	Prospective cohort study	2019 Chinese	BMD, risk of developing AD	Negative
[37]	Cross‐sectional study	535 Chinese	BMD, MMSE	Positive
[39]	Retrospective cohort study	8448 Chinese	Dementia diagnosis, risk of developing bone fracture	Positive
[40]	Cohort study	140 Americans	BMD, cerebral atrophy and cognitive decline	Negative
[41]	Cohort study	13113 participants from UK Biobank	BMD, gray matter volume	Positive
[42]	Mendelian randomization study	394929 participants from UK Biobank	Heel BMD, risk of developing AD	Positive

Abbreviations: AD, Alzheimer's disease; Aβ, amyloid beta; BMD, bone mineral density; MMSE, Mini‐Mental State Examination.

3. THE EFFECTS OF AD ON BONE FUNCTIONS

The brain is the central control center of systemic tissues and organs. The brain has been shown to control the growth and differentiation of bone cells and impact bone functions through chemical and electrical signals. ⁴⁷ AD patients exhibit brain atrophy, neural circuit damage, and neurotransmitter imbalance. ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ Functional magnetic resonance imaging (MRI) revealed that AD patients exhibited no or less activation of the hippocampus and other medial temporal lobe structures during the completion of memory tasks. ⁵² Additionally, increased bone fracture rates and reduced BMD are commonly observed in patients with AD. ³⁹ , ⁴² Therefore, neuropathological changes in the AD brain may lead to bone dysfunction.

3.1. Effect of AD pathogenic proteins on bone

Aβ peptides are proteolytic fragments of amyloid precursor protein (APP). Aβ42 is located mainly in the membrane and cytoplasm of osteocytes and the extracellular matrix. ⁵³ In one study, the levels of APP and Aβ42 were markedly elevated in the osteoporotic bone tissues of humans and animals compared with those of controls. ⁵³ Moreover, the Aβ42 levels in bone tissues were negatively correlated with BMD in patients. ⁵³ Aβ42 was reported to enhance osteoclast differentiation and activation, suggesting that Aβ may play an important role in the pathogenesis of osteoporosis. ⁵³ The Swedish mutation of APP (APPswe) could cause early‐onset AD. ⁵⁴ APPswe can suppress osteoblast differentiation and cause osteoporosis, indicating that AD and osteoporosis may share conserved pathogenic mechanisms. ⁵⁵ The osteoporotic deficit in young APPswe mice could be prevented by treatment with cognition‐improving antioxidants, suggesting that reactive oxygen species may be involved in the underlying mechanism of APPswe‐induced osteoporosis. ⁵⁵ , ⁵⁶ , ⁵⁷ A subsequent study revealed that APP plays a physiological role in osteoblast survival and bone formation by regulating mitochondrial function. ⁵⁸ Osteoclast activity was increased in younger APPswe mice (<4 months old) and decreased in older Tg2576 mice (>4 months old) compared to that in age‐ and sex‐matched wild‐type (WT) mice. ⁵⁸ However, bone loss can still be observed in Tg2576 mice after 4 months of age or even at older ages. The decrease in osteoclast formation and activity in aged Tg2576 mice may be due to an increase in soluble advanced glycation end products. This dysregulation of bone remodeling in aged mice may lead to an increased incidence of fractures without altering overall bone density. ⁵⁸ However, another study revealed that Aβ42 did not influence receptor activators of NF‐κB ligand (RANKL)‐induced osteoclastogenesis. ⁵⁹ Instead, it promoted osteoclast differentiation by enhancing RANKL‐mediated activation of NF‐κB and TRAF6‐MAPK signaling pathways and increasing RANKL‐induced calcium oscillation. ⁵⁹ In addition to amyloid plaques, intracellular neurofibrillary tangles consisting of hyperphosphorylated tau are also pathological features of AD. ⁶⁰ Related studies indicated that tau may affect AD through Aβ‐dependent and Aβ‐independent pathways (eg, apolipoprotein E [APOE], the endocytic system, and cholesterol metabolism. ⁶¹ , ⁶² , ⁶³ Early BMD reduction was observed in tauopathy‐specific AD model mice. ⁶⁴ Another study revealed that arthritis model mice with tau gene knockout exhibited less bone loss, cartilage damage, and osteoclast activity due to reduced macrophage polarization. ⁶⁵ Therefore, it can be inferred that two common pathological hallmarks of AD are involved in regulating bone homeostasis. So far, there are no reports about skeletal manifestations in familial AD, to which more attention should be paid in the future.

3.2. Effect of AD risk genes on bone

Several genetic factors have also been implicated in the association between AD and bone‐related diseases. The APOE gene is involved in more than half of AD cases, and its gene polymorphisms are considered the primary genetic risk factors for late‐onset AD. ⁶⁶ , ⁶⁷ APOE ε4 increases the risk of AD in a dose‐dependent manner by enhancing toxicity and inhibiting protective function. ⁶⁷ A previous meta‐analysis indicated insufficient evidence linking the APOE genotype to BMD and fracture incidence. ⁶⁸ In another cohort study, APOE ε4 was indicated to be associated with osteoporosis and an increased risk of fractures, whereas APOE ε3 had the opposite effect. ⁶⁹ While APOE‐deficient mice and WT mice exhibited similar morphological changes in bone tissue, APOE‐deficient mice showed greater bone resorption than bone formation. ⁷⁰ Another study demonstrated that the uptake of triglyceride‐rich lipoproteins by osteoblasts was reduced in APOE‐deficient mice. ⁷¹ This reduction resulted in elevated serum osteocalcin levels and ultimately led to increased bone formation. ⁷¹ A single‐cell atlas of the human infrapatellar fat pad and synovium implicated APOE signaling in osteoarthritis pathology, and inhibiting APOE signaling decreased the progression of osteoarthritis. ⁷² The complex role played by genetic polymorphisms of APOE underscores its plausible connection between AD and bones from various perspectives.

Triggering receptor expressed on myeloid cells 2 (TREM2) is expressed in microglia in the brain and in macrophages in the periphery, and it has been indicated to be involved in the pathogenesis of AD. ⁷³ In bone tissue, TREM2 is expressed in osteoclasts and is a key regulator of osteoclast production. ⁷⁴ , ⁷⁵ In the periodontitis model, the TREM2‐mediated reactive oxygen species signal amplification cascade is important for osteoclast generation, and the accumulation of soluble Aβ42 oligomers in the periodontitis microenvironment further enhances this signal by direct interaction with TREM2. ⁷⁴ Additionally, there is evidence that the TREM2 R47H variant can cause significant age‐ and sex‐dependent musculoskeletal changes in mice. ⁷⁶ Nasu–Hakola disease, also called polycystic lipomembranous osteodysplasia with sclerosing leukoencephalopathy (PLOSL), is caused by mutations in either the TREM2 or TYROBP gene and manifests primarily as progressive presenile dementia along with bone cysts and fractures. ⁷⁷ , ⁷⁸ Given that TREM2 expression predominantly occurs within microglia in the brain, some researchers have questioned whether Nasu–Hakola disease could be considered a “microglial disease.” ⁷⁷ In conclusion, AD risk genes including APOE and TREM2 may be involved in AD, but the underlying mechanisms need further study.

3.3. Neurohormones connect AD and bone

Several studies have suggested that osteoporosis and AD often occur during perimenopause and menopause in women. ⁷⁹ , ⁸⁰ , ⁸¹ , ⁸² , ⁸³ , ⁸⁴ Perimenopause is characterized by increased follicle‐stimulating hormone (FSH) levels and decreased estrogen levels. With aging, a decrease in estrogen receptor α may contribute to cognitive impairment. ⁸⁵ Additionally, estrogen was found to be involved in inducing APP degradation and regulating oxidative damage and energy metabolism. ⁸⁵ , ⁸⁶ In addition, estrogen can also exert neuroprotective effects by attenuating Aβ‐induced lipid peroxidation and glutamatergic excitotoxicity. ⁸⁶ Overall, estrogen is important for maintaining normal brain function, and estrogen deficiency is associated with an increased risk of AD. ⁸³ , ⁸⁷ , ⁸⁸ , ⁸⁹ Similarly, estrogen can exert an anti‐osteoporotic effect by decelerating the rate of bone remodeling, inducing osteoclast apoptosis, and inhibiting osteoblast and osteocyte apoptosis. ⁹⁰ , ⁹¹ , ⁹² Compared with senile osteoporosis, postmenopausal osteoporosis is characterized by an increased number of osteoclasts, leading to enhanced bone resorption. ⁹³ Furthermore, in addition to directly impacting bone remodeling, a decrease in estrogen levels can result in elevated levels of certain inflammatory cytokines (such as interleukin 1 [IL‐1], IL‐6, and tumor necrosis factor α [TNF‐α]) and alterations in the immune cell profile (including increased T cells), ultimately affecting bone homeostasis. ⁹³ , ⁹⁴ In addition, studies have also demonstrated that mice with estrogen deficiency exhibit cartilage degeneration, subchondral bone damage, and other changes, which ultimately lead to osteoarthritis. ⁹⁵ As a superior regulator of estrogen, FSH levels were elevated in the blood of AD patients. ⁹⁶ FSH can directly act on brain neurons to accelerate APP and tau cleavage through the C/EBPβ‐δ‐secretase pathway, leading to AD pathologies and cognitive deficits. ⁹⁷ FSH can promote bone loss through the stimulation of osteoclast production, but it is difficult to separate the action of FSH from that of estrogen. ⁹⁸ Thus, estrogen and FSH are involved in the pathogenesis of both AD and bone diseases.

Thyroid‐stimulating hormone (TSH) is a pituitary gland‐secreted hormone that not only regulates thyroid function but also plays a role in bone remodeling. Clinical studies have shown that increased levels of free T3 and free T4 are associated with decreased BMD and that increased levels of TSH are protective for bone. ⁹⁹ , ¹⁰⁰ , ¹⁰¹ The anti‐osteoclastic effect of TSH reduces bone resorption, although its impact on osteoblasts varies depending on the stage of cell differentiation. ¹⁰¹ , ¹⁰² Similarly, low TSH levels are associated with a high incidence of AD. ¹⁰³ A Mendelian randomization study suggested that genetically predicted elevated levels of TSH were associated with a reduced risk of AD, even when these levels were within normal ranges. ¹⁰⁴ However, it has also been proposed that thyroid dysfunction may be a consequence rather than a cause of AD since the decrease in TSH could be related to pituitary degeneration due to AD pathology. ¹⁰⁵ , ¹⁰⁶ Similarly, hyperactivation of the hypothalamic–pituitary–adrenal (HPA) axis and elevated glucocorticoids can exacerbate the pathological changes in AD and contribute to bone loss. ¹⁰⁷ , ¹⁰⁸ , ¹⁰⁹ , ¹¹⁰ , ¹¹¹ , ¹¹² , ¹¹³ Based on previous studies, the glucocorticoid cascade is hypothesized to have a causal relationship with hippocampal injury, and overactivation of the HPA axis in AD may contribute to tau phosphorylation and enhance Aβ toxicity. ¹¹⁴ , ¹¹⁵ In bone tissue, adrenocorticotropin is thought to stimulate osteoblast proliferation, and glucocorticoid can inhibit osteoblast differentiation, increase osteoclast production, and promote osteocyte apoptosis. ¹⁰¹ , ¹¹⁶ The hypothalamic–pituitary–target organ axis plays an important role in neuroendocrine regulation, and it may also play a ubiquitous and important role in the brain–bone axis, which requires further investigation.

3.4. Neuropeptides connect AD and bone

As endogenous active substances, neuropeptides such as neuropeptide Y (NPY) and kisspeptin are involved in several pathophysiological processes. Within the CNS, NPY is predominantly expressed in the hypothalamus and exerts neuroprotective effects against AD. ¹¹⁷ Clinical studies have consistently demonstrated a general decline in NPY levels among AD patients. ¹¹⁸ , ¹¹⁹ , ¹²⁰ , ¹²¹ Animal models of AD have shown conflicting findings regarding cortical and hippocampal NPY levels. Some studies reported an increase, but others indicated a reduction in NPY‐responsive neurons during the early stages. ¹²² , ¹²³ , ¹²⁴ , ¹²⁵ However, NPY is generally believed to play a neuroprotective role in AD by increasing the expression of nerve growth factor, alleviating the neurotoxic effects of Aβ, modulating neurogenesis, regulating calcium homeostasis, and alleviating neuroinflammation. ¹²⁶ , ¹²⁷ , ¹²⁸ , ¹²⁹ , ¹³⁰ , ¹³¹ , ¹³² In bone tissue, receptors for NPY are expressed, for example, in osteoblasts and osteocytes. ¹¹⁷ NPY appears to play a paradoxical role in maintaining skeletal homeostasis. Researchers discovered that NPY mediated glucocorticoid‐induced trabecular bone degeneration through post‐translational modification of peroxisome proliferator‐activated receptor γ. ¹³³ Senescent osteocytes secrete NPY, which can inhibit the expression of pro‐osteogenic and anti‐adipogenic transcription factors that enable the differentiation of BMSCs from osteoblasts to adipocytes. ¹³⁴ A recent study reported that exogenous overexpression of NPY aggravated postmenopausal osteoporosis by regulating gut microbiota. ¹³⁵ Downregulation of hypothalamic NPY expression can promote bone formation. ¹³⁶ In addition, NPY receptor antagonists promoted bone formation and suppressed bone resorption in a rat model of osteoporosis. ¹³⁵ , ¹³⁷ Thus, NPY could inhibit osteogenesis and promote bone resorption. Moreover, NPY can enhance BMSC differentiation into osteoblasts by activating the Wnt pathway. ¹³⁸ A study using rat models of femoral fractures revealed that antagonists targeting NPY receptors impeded fracture healing. ¹³⁹ Furthermore, NPY can ameliorate osteoporosis by reducing osteoclast numbers through mobilizing hematopoietic stem/progenitor cells. ¹⁴⁰ Therefore, NPY could also promote bone formation to some extent.

Kisspeptin, a hypothalamic neuropeptide, has traditionally been thought to be important for the regulation of reproduction. ¹⁴¹ An in vitro study showed that Aβ could stimulate the release of kisspeptin, which can bind to Aβ and antagonize its neurotoxicity. ¹⁴² In bone, loss of function of the kisspeptin signaling pathway was associated with delayed skeletal maturation. ¹⁴³ In addition, kisspeptin increased osteoblast production and inhibited bone resorption by osteoclasts in vitro, and the effect of kisspeptin on osteoblasts was verified in healthy men. ¹⁴⁴ Another study demonstrated that the deletion of estrogen receptor α with the Kiss1‐Cre knock‐in allele increased bone density in female mice. ¹⁴⁵ Therefore, it can be inferred that kisspeptin is involved in the regulation of the brain–bone axis.

3.5. Regulation of peripheral nerves and neurotransmitters on bone in AD

Bone also receives innervation from the autonomic nervous system, including the sympathetic nervous system and parasympathetic nervous system. ³⁰ , ⁴⁷ , ¹⁴⁶ Clinical evidence has shown that AD patients may have varying degrees of autonomic dysfunction, including orthostatic hypotension, decreased heart‐rate variability, constipation, and urinary incontinence. ¹⁴⁷ With aging, bone density decreases, and degenerative changes in autonomic distribution occur. ¹⁴⁸ In peripheral nerves, acetylcholine serves as the main neurotransmitter of parasympathetic nerves, and cholinergic receptors include muscarinic acetylcholine receptors (mAChRs) and nicotinic acetylcholine receptors (nAChRs). The cholinergic hypothesis is one of the pathogenic hypotheses for AD, postulating that the degeneration of cholinergic neurons in the brain, along with the depletion of presynaptic cholinergic transmitters and reduced activity of cholinesterase, leads to AD development. ¹⁴⁹ , ¹⁵⁰ It has been demonstrated that osteocytes express both mAChR and nAChR and that acetylcholine regulates osteocyte function through interactions with the AChR. ¹⁵¹ The AChR is composed of multiple subunits, and studies have suggested that M3 subunit of the mAChR has a positive influence on cancellous bone microstructure, flexural stiffness, and bone matrix synthesis. ¹⁵² , ¹⁵³ Additionally, the α9 subunit of the nAChR has been implicated in maintaining osteocyte homeostasis and regulating bone mass. ¹⁵⁴ Studies have shown that nAChR can increase bone mass by reducing osteoclast production. ¹⁵⁵ , ¹⁵⁶

The bone region with the highest metabolic activity has the richest sympathetic distribution. ¹⁴⁶ The main neurotransmitter of sympathetic nerves is norepinephrine, which activates α and β adrenergic receptors (ARs). ¹⁵⁷ The βAR subtype may be the primary ARs that mediate sympathetic remodeling of bone. ¹⁵⁷ Stimulation of βARs in osteoblasts can enhance osteoclast production while inhibiting osteoblast proliferation through circadian genes, ultimately leading to the development of low bone mass. ¹⁵⁷ , ¹⁵⁸ , ¹⁵⁹ In addition, the activation of paracrine proinflammatory factors in sympathetic nerves impaired the osteogenic ability of mesenchymal stem cells. ¹⁶⁰ Another study revealed that sympathetic nerves induced osteoporosis by upregulating microRNA 21 expression in osteoblasts via βAR activation and stimulating osteoclasts through exosome‐mediated transport. ¹⁶¹ These findings indicate that dysregulated sympathetic activity in AD may lead to bone loss.

Neurotransmitters are also thought to be important in the brain–bone axis. Glutamate is an excitatory neurotransmitter responsible for excitatory synapse activity that is involved in the maintenance of neuronal function through the crucial glutamate–glutamine cycle. Promoting the glutamate–glutamine cycle in neuron‐astrocytes has been shown to be beneficial in alleviating AD. ¹⁶² Neurons exhibiting heightened baseline activity levels exhibit hyperactivity in response to Aβ, which disrupts glutamate reuptake and leads to neuronal overstimulation. ¹⁶³ , ¹⁶⁴ Numerous studies have shown the dysregulation of glutamate in AD patients and animal models. ¹⁶⁵ In bone tissue, glutamate receptors are expressed by osteoblasts and osteoclasts. ¹⁶⁶ The differentiation of osteoblasts is upregulated by glutamate through the alpha‐amino‐3‐hydroxy‐5‐methylisoxazole‐4‐propionic acid receptor and N‐methyl‐D‐aspartate (NMDA) receptor, and the NMDA receptor is also involved in the regulation of osteoclast production. ¹⁶⁷ , ¹⁶⁸ Furthermore, serotonin and dopamine are also implicated in the regulation of both brain and bone physiology. ³⁰

3.6. Effect of brain‐derived EVs on bone

EVs are a heterogeneous group of membrane‐bound vesicles released by cells and include various subtypes such as exosomes, microvesicles, oncosomes, and apoptotic bodies. ¹⁶⁹ Among these subtypes, exosomes have been extensively investigated due to their wide distribution and heterogeneous functions. Exosomes can be secreted by many tissues and contain membrane proteins, cytosolic and nuclear proteins, extracellular matrix proteins, metabolites, and nucleic acids. Injury to hippocampal neurons can release small EVs containing microRNAs, which subsequently target bone progenitor cells to promote osteogenesis and accelerate bone healing. ¹⁷⁰ Therefore, it can be hypothesized that brain‐derived EVs have the ability to traverse the blood–brain barrier (BBB) and reach bone tissue to regulate its function. Recently, a study confirmed the BBB‐crossing potential of brain‐derived EVs in both AD and WT mice. ¹⁷¹ This study revealed that both brain‐derived EVs and plasma‐derived EVs in AD exert a significant inhibitory effect on osteogenic differentiation while promoting adipogenic differentiation of BMSCs, which ultimately leads to bone loss and marrow adiposity. ¹⁷¹

4. REGULATORY ROLES OF BONE IN AD

We have already discussed the possible effects of AD on bone. As patients with bone disorders including osteoporosis and osteoarthritis have a greater incidence of AD, we further investigated the possible effects of bone‐derived hormones, bone marrow‐derived cells, bone‐derived EVs, and inflammation on AD pathogenesis and cognitive function.

4.1. Bone‐derived hormones

4.1.1. Osteocalcin

Osteocalcin (OCN) is a 5‐kDa protein composed of 49 amino acids that is encoded by the Bglap gene. ¹⁷² OCN is synthesized by osteoblasts and serves as a marker for bone formation. ¹⁷² It exists in both carboxylated and undercarboxylated forms of OCN. ¹⁷³ , ¹⁷⁴ cOCN is primarily deposited in the bone matrix, and ucOCN is released into circulation to perform regulatory functions on multiple organs. ¹⁷³ , ¹⁷⁴ Clinical studies have shown that plasma OCN levels are correlated with cognitive function and brain microstructural changes. ²⁹ , ¹⁷⁵ A recent Mendelian randomization study also suggested that OCN may exert a protective effect against AD. ¹⁷⁶ However, whether OCN levels are altered in the AD brain remains unknown. Several studies have been conducted to investigate the role and underlying mechanism of OCN in cognition. OCN can traverse the BBB and ultimately modulate spatial learning and memory in mice by influencing neurotransmitter synthesis. ¹⁷⁷ G protein‐coupled receptor 158 (GPR158) was demonstrated to be the receptor of OCN in neurons of the hippocampus and mediated the regulation of hippocampal‐dependent memory by OCN through BDNF. ¹⁷⁸ Moreover, the OCN/GPR158 signaling pathway has been shown to mediate a regulatory loop controlling the histone‐binding protein RbAp48, which in turn modulates the expression of GPR158. ¹⁷⁹ Activation of the OCN/GPR158 pathway elicits elevated RbAp48 levels in middle‐aged animals and rescues age‐related cognitive decline. ¹⁷⁹ A recent study showed that OCN could enhance the energy metabolism of astrocytes and microglia through the GPR158 protein, thereby promoting glycolysis, reducing Aβ deposition in the brain, and alleviating cognitive deficits in AD mice. ¹⁸⁰ However, a previous study demonstrated that GPR158 is mainly expressed in neurons and is largely absent in glia. ¹⁸¹ In addition, OCN can regulate oligodendrocyte differentiation and myelin homeostasis via GPR37 in the mouse brain. ¹⁸² As myelin dysfunction is involved in AD pathogenesis, improving oligodendrocyte health and myelin integrity could alleviate cognitive impairment of AD. ¹⁸³ , ¹⁸⁴ Thus, OCN may protect against AD by enhancing myelination. Considering the function of OCN in bone‐related diseases and the brain, some researchers have proposed that OCN could be identified as an antigeronic hormone. ²⁶ This also implies the role of aging in bone and AD. However, the underlying mechanisms need further study.

4.1.2. Lipocalin‐2

Lipocalin‐2 (LCN‐2) is a glycoprotein consisting of 198 amino acids and has been identified as a molecule secreted by osteoblasts. ⁴⁷ The expression level of LCN‐2 is at least 10‐fold greater than that in other tissues. ⁴⁷ The extensive literature on LCN‐2 has shown that it is involved in various pathophysiological processes, including innate immunity, inflammation, metabolism, and apoptosis. ¹⁸⁵ The expression of LCN‐2 is upregulated during osteoblast differentiation and related to bone development and turnover. ¹⁸⁶ , ¹⁸⁷ In addition, LCN‐2 is also highly expressed in the osteoblasts of AD. ¹⁸⁸ Osteoblast‐derived LCN‐2 can cross the BBB and suppress appetite by binding to the melanocortin‐4 receptor in the hypothalamus. ¹⁸⁷ In the preclinical stage of AD, blood LCN‐2 levels are elevated and correlated with Aβ42 levels in cerebrospinal fluid (CSF). ¹⁸⁹ Another study also revealed increased blood LCN‐2 levels in AD patients compared with healthy control individuals and patients with other types of dementia; however, no correlation was found between plasma LCN‐2 and CSF AD biomarkers or cognitive function. ¹⁹⁰ LCN‐2 levels in osteoblasts can promote AD development by upregulating hippocampus collapsin response mediator protein 2. ¹⁸⁸ Moreover, decreasing the expression of LCN‐2 in osteoblasts via the overexpression of Fork‐head box O1 could prevent AD progression. ¹⁸⁸ These findings highlight that LCN‐2 may be a potential therapeutic target for AD. Due to ongoing controversies surrounding the study of LCN‐2 in the brain, including questions about which cells produce it and its role during inflammation, ¹⁹¹ more studies are warranted to explore the relationship between bone‐derived LCN‐2 and AD.

4.1.3. Fibroblast growth factor 23

Fibroblast growth factor 23 (FGF23), which is mainly secreted by osteocytes and osteoblasts in bone, ¹⁹² is thought to be involved in bone mineralization. ¹⁹³ It can interact with the klotho protein, which has been implicated in aging and cognition. ¹⁹⁴ However, limited research has been conducted on the role of FGF‐23 in AD. A clinical study revealed an association between elevated circulating FGF23 levels and increased dementia risk. ¹⁹⁵ AD patients exhibit increased concentrations of FGF23, which are correlated with inflammatory cytokine levels. ¹⁹⁶ An animal study showed that overexpression of FGF23 led to impairment of long‐term potentiation in the hippocampal CA1 region. ¹⁹⁷ These findings suggest that FGF23 may be involved in the development of AD, but the underlying mechanism is not fully understood.

4.1.4. Osteoprotegerin

Osteoprotegerin (OPG), a soluble glycoprotein of the TNF receptor superfamily, is synthesized by osteoblasts, B‐lymphocytes, and articular chondrocytes and plays a role in bone homeostasis. ¹⁹⁸ , ¹⁹⁹ , ²⁰⁰ The role of OPG in AD is unclear. While one study reported significantly elevated blood levels of OPG in AD patients, another study reported no statistically significant difference in blood OPG levels between AD patients and healthy controls individuals. ²⁰¹ , ²⁰² Elevated circulating OPG levels were reported to be associated with decreased total brain volume. ²⁰³ Increasing evidence suggests that OPG plays an important role in vascular injury and inflammatory processes within the CNS. ²⁰⁴ However, more clinical and experimental studies are needed to investigate the role of OPG in AD.

4.1.5. Osteopontin

Osteopontin (OPN) was the first extracellular matrix component identified in bone tissue and is widely distributed across various tissues. ²⁰⁵ OPN plays an important role in bone remodeling. ²⁰⁶ High levels of blood OPN in postmenopausal women were associated with low BMD and osteoporotic vertebral fractures. ²⁰⁷ , ²⁰⁸ Clinical investigations demonstrated elevated levels of OPN in both the plasma and CSF of AD patients, ²⁰⁹ , ²¹⁰ , ²¹¹ suggesting that OPN might play some role in the pathophysiology of AD. An experimental study revealed that microglia‐derived OPN levels were increased in both AD mouse models and AD patients and contributed to AD pathology. ²¹² Moreover, blocking OPN inhibited the microglial proinflammatory response and decreased Aβ plaque pathology in AD mice. ²¹² Whether bone‐derived OPN contributes to AD progression remains largely unknown and warrants further investigation.

4.1.6. Osteocrin

Osteocrin (OSTN, also known as musclin) is a bone‐derived protein consisting of 103 amino acids with a molecular weight of 11.4 kDa. While mice express OSTN primarily in bones and muscles rather than in the brain, primate brains express OSTN within the neocortex. ²¹³ In bone, OSTN is predominantly expressed in osteoblasts and osteocytes, particularly during the early stages of bone formation, and its expression gradually declines with age. ²¹⁴ Due to its homology to natriuretic peptide, it has been proposed that periosteal osteoblast‐produced OSTN may modulate growth plate development by augmenting C‐type natriuretic peptide (CNP)‐dependent proliferation and maturation of chondrocytes. ²¹⁵ By binding to the natriuretic peptide receptor 3 (NPR3) in periosteal bone progenitor cells, OSTN impedes NPR3‐mediated CNP clearance, thereby promoting bone growth. ²¹⁵ , ²¹⁶ Additionally, overexpression of OSTN partially mitigated glucocorticoid‐induced osteocyte degeneration in mice. ²¹⁷ Furthermore, overexpression of OSTN can rescue osteocyte dendritic defects caused by Sp7 gene deficiency, and Sp7‐dependent genes that mark osteocytes are highly enriched in neurons, suggesting potential associations between Sp7‐dependent traits and neuronal development. ²¹⁸ The evolution of the OSTN gene led to the emergence of a novel primate‐specific enhancer sequence capable of binding to myocyte enhancer factor 2 (MEF2). ²¹³ MEF2 and OSTN act as regulators of dendritic growth in the developing cerebral cortex in response to sensory experience. ²¹³ Recent literature has shown that OSTN can prevent depressive‐like behavior in male mice by activating urinary corticin‐2 signaling in the hypothalamus. ²¹⁹ In addition, OSTN ameliorated the expression of inflammatory markers in monocytes, ²²⁰ suggesting a protective role against neuroinflammation. However, whether bone‐derived OSTN is involved in AD pathogenesis remains unclear and needs to be investigated in the future.

4.1.7. Sclerostin

Sclerostin (SOST), a protein expressed by mature bone cells, is an inhibitor of the Wnt signaling pathway. ²²¹ SOST can bind to low‐density‐lipoprotein receptor‐related proteins on the surface of osteoblasts, block the Wnt signaling pathway, and inhibit osteoblast proliferation and differentiation. ²²² Elevated blood SOST levels were correlated with low BMD in patients with prediabetes ²²³ and with worse fracture healing in patients after surgical stabilization of long bone fractures. ²²⁴ SOST levels in blood and CSF are positively correlated with age. ²²⁵ , ²²⁶ Blood SOST was positively associated with brain Aβ burden and cognitive impairment in both aged individuals and patients with AD. ²²⁵ , ²²⁶ A recent study demonstrated that bone‐derived SOST can cross the BBB, increase Aβ production through β‐catenin–β‐secretase 1 signaling, and impair synaptic plasticity and memory function in mice. ²²⁶ More clinical and experimental studies are needed to investigate the role of SOST in AD.

4.1.8. Bone morphogenetic proteins

Bone morphogenetic proteins (BMPs) are the largest subgroup of the transforming growth factor‐beta superfamily, with more than 20 identified BMPs. ²²⁷ BMPs are widely expressed in various tissues throughout the body including bone and brain and are important for bone development and regeneration and maintenance of bone homeostasis. ²²⁸ Specifically, BMP4 is implicated in bone repair and regeneration; however, it can inhibit neurogenesis and oligodendrogenesis in induced pluripotent stem cells (iPSCs) of AD patients. ²²⁸ , ²²⁹ Decreased hippocampal cell proliferation was associated with increased expression of BMP4 in AD. ²³⁰ In addition, BMP4 transgenic mice exhibit increased levels of APP and tau proteins along with cognitive impairment, and in vitro experiments have shown that the downregulation of BMP4 reduces the levels of AD pathogenic proteins. ²³¹ Furthermore, BMP2 and BMP4 can upregulate the expression of aquaporin‐4 in astrocytes, which is also involved in AD pathogenesis. ²³² A clinical study revealed that AD patients had lower blood BMP6 levels, which are positively associated with cognitive function. ²³³ However, the role of bone‐derived BMPs in the pathogenesis and progression of AD needs further study.

4.1.9. Bone‐derived platelet‐derived growth factor‐BB

Platelet‐derived growth factor (PDGF)‐BB, a member of the PDGF family, has been widely studied in the field of orthopedics. ²³⁴ , ²³⁵ , ²³⁶ A study demonstrated that osteoclasts are the primary source of elevated circulating levels of PDGF‐BB during aging. ²³⁶ , ²³⁷ Bone‐derived PDGF‐BB is implicated in the regulation of angiogenesis, arteriosclerosis, bone remodeling, and joint structure damage. ²³⁸ , ²³⁹ A study revealed that aged male mice had increased blood PDGF‐BB levels, which could induce brain vascular calcification. ²⁴⁰ Furthermore, elevated concentrations of bone‐derived PDGF‐BB could induce hippocampal BBB impairment and memory deficits by promoting the shedding of PDGF receptor β from the pericytes' surface. ²⁴¹ Therefore, future investigations should explore the role of bone‐derived PDGF‐BB in other age‐related diseases such as AD.

4.2. Bone marrow‐derived cells

4.2.1. Bone marrow‐derived immune cells

Dysfunction of microglia in the AD brain is closely related to the occurrence and development of AD. There is evidence that bone marrow‐derived macrophages can migrate into the brain to become microglia under physiological conditions, but these microglia account for only approximately 1% of the total number of microglia in the brain. ²⁴² The effect of bone marrow‐derived immune cells on AD is controversial. Early studies revealed that infiltrating monocytes in the blood of AD patients did not perform long‐term functions in clearing Aβ in the brain. ²⁴³ A recent study found that increased infiltration of bone marrow‐derived grancalcin (GCA)‐positive immune cell subpopulation in the brain parenchyma of AD model mice promoted the progression of AD through the secretion of GCA, which could inhibit the phagocytosis and clearance of Aβ in microglia by competitively binding to lipoprotein receptor‐related protein 1. ²⁴⁴ Therefore, the complicated roles of bone marrow immune cells in the pathogenesis of AD remain unclear and need further study.

4.2.2. Bone marrow mesenchymal stem cells

As pluripotent stem cells, BMSCs are also involved in the bone microenvironment and have extensive applications in skeletal system diseases and AD. ²⁴⁵ , ²⁴⁶ , ²⁴⁷ One study demonstrated that intracerebral injection of BMSCs resulted in a significant reduction in Aβ accumulation and an increase in proteins facilitating synaptic transmission. ²⁴⁸ Additionally, intracerebral administration of BMSCs enhanced microglial phagocytosis of Aβ, possibly through microglial activation, upregulation of Aβ‐degrading enzymes, and increased expression of Aβ binding receptors. ²⁴⁹ , ²⁵⁰ Furthermore, transplantation of BMSCs was found to alleviate tau hyperphosphorylation in the brains of AD model mice. ²⁴⁹ , ²⁵¹ BMSC transplantation prevents microglia in the classical M1 state from secreting proinflammatory factors and stimulates M2 microglia to produce anti‐inflammatory cytokines. ²⁵² Nevertheless, both injection methods resulted in decreased expression of inflammation‐related genes such as TNF‐α and IL‐1β and alleviated pathological changes associated with AD. ²⁴⁹ , ²⁵³ , ²⁵⁴ Co‐culture of BMSCs with Aβ‐treated primary cortical and hippocampal neurons in vitro also demonstrated elevated levels of anti‐inflammatory factors, increased telomere length and telomerase activity, and activated signaling pathways that have been linked to aging. ²⁵⁵ In conclusion, BMSC transplantation can alleviate cognitive deficits by regulating Aβ clearance, tau hyperphosphorylation, neurogenesis, neuroinflammation, immunomodulation, apoptosis, autophagy, and angiogenesis in AD mouse models. ²⁴⁸ , ²⁴⁹ , ²⁵⁰ , ²⁵³ , ²⁵⁴ , ²⁵⁶ . Moreover, the effect of BMSCs on AD has been verified in different transplantation approaches and different animal models. ²⁴⁷ Microscopic vascular channels were found between the bone marrow of the skull and the brain. ²⁵⁷ , ²⁵⁸ These channels allow CSF to enter the skull bone marrow and can also affect the migration of cells in the bone marrow to the brain. ²⁵⁷ , ²⁵⁹ Therefore, whether the changes in CSF of AD patients affect the skull bone marrow cells to enter the brain and participate in the neuroimmune process of AD needs to be further studied. ²⁶⁰

4.3. Bone‐derived EVs

Osteoblasts, osteoclasts, and osteocytes can release EVs, which are widely involved in the regulation of bone metabolism. ²⁶¹ , ²⁶² It was observed that plasma levels of osteocyte‐derived EVs were greater in AD patients than in cognitively normal individuals. ²⁶³ Compared with aged osteocyte‐derived EVs, young osteocyte‐derived EVs exhibited an enhanced ability to cross the BBB. ²⁶³ In vivo and in vitro experiments demonstrated that young osteocyte‐derived EVs from WT mice can mitigate brain Aβ deposition and cognitive impairment in both the early and late stages of AD. ²⁶³ Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis revealed that multiple functional factors related to Aβ degradation and mitochondrial energy metabolism were highly enriched in young osteocyte‐derived EVs compared with aged osteocyte‐derived EVs. ²⁶³ This study suggested that young osteocyte‐derived EVs played a protective role in AD by preventing Aβ deposition and neuronal damage. However, whether aged osteocyte‐derived EVs from AD transgenic mice exacerbate AD progression remains unknown and deserves further investigation. In addition, EVs derived from aged bone matrix have been found to promote adipogenesis and exacerbate vitamin D3‐induced vascular calcification, ²⁶⁴ which could contribute to AD dementia. ²⁶⁵ , ²⁶⁶ BMSC‐derived EVs are also involved in the pathogenesis of AD. Studies have indicated that BMSC‐derived EVs effectively reduce Aβ accumulation in the brains of AD mice, with a more pronounced effect observed in young mice, suggesting the potential of BMSC‐derived EVs to mitigate or prevent Aβ formation during the early stages of AD. ²⁶⁷ , ²⁶⁸ In addition, EVs from BMSCs can attenuate the expression of inducible nitric oxide synthase induced by Aβ oligomers, potentially ameliorating hippocampal synaptic plasticity deficits and cognitive impairment. ²⁶⁹ Additionally, BMSC‐derived exosomes alleviate neuroinflammation in AD animal models by reducing neurotoxicity resulting from the activation of microglial and astrocytes in the hippocampus and significantly increasing BDNF expression levels. ²⁷⁰ Therefore, we speculate that bone‐derived EVs may play complicated roles in the crosstalk between bone and brain in AD.

4.4. Inflammation

Chronic inflammation is a hallmark feature of both bone diseases and AD. During the inflammatory response, the NF‐κB signaling pathway is activated, promoting osteoclasts and simultaneously inhibiting osteoblasts. ²⁷¹ , ²⁷² The complement system associated with NF‐κB signaling activation is also involved in regulating bone loss. ²⁷³ In addition, proinflammatory factors such as IL‐1, IL‐6, and TNF have also been found to be associated with bone‐related diseases such as osteoporosis or osteoarthritis. ²⁷⁴ , ²⁷⁵ , ²⁷⁶ , ²⁷⁷ , ²⁷⁸ Similarly, in AD, the activation of inflammation‐related signaling pathways including the NF‐κB pathway, the complement system, and proinflammatory factors are also activated. ²⁷⁹ , ²⁸⁰ , ²⁸¹ A genome‐wide association study also revealed that some risk genes for AD are also associated with inflammation. ²⁸² In general, inflammation is involved in the pathogenesis of bone‐related diseases and AD. Investigations into whether bone‐related diseases are related to systemic inflammation have been conducted. ²⁷⁷ , ²⁸³ , ²⁸⁴ , ²⁸⁵ Peripheral inflammation has been suggested to exacerbate AD‐type pathologies and induce cognitive impairment. ²⁸⁰ , ²⁸⁶ , ²⁸⁷ , ²⁸⁸ Therefore, we speculate that systemic inflammation could be regarded as a common risk factor for bone disease and AD. Peripheral inflammation in osteoarthritis exacerbates brain inflammation and subsequent AD pathology in AD mice. ²⁸⁹ , ²⁹⁰ Therefore, the inflammatory mediators produced during bone diseases may contribute to AD by promoting neuroinflammation and neurodegeneration through the activation of glial cells. However, the recently reported relevant research is limited and needs to be further explored.

Microglia and osteoclasts are tissue‐specific macrophages and functionally similar cell types in the brain and bone. The inflammatory pathway induced by microglia and osteoclasts may mediate shared risk factors for AD and osteoporosis. ²⁹¹ A recent transcriptional sequencing study of microglia in AD mice revealed that seven of the differentially regulated genes in microglia were related to osteoclast generation, recruitment, and inflammation, ²⁹² suggesting that osteoclasts and microglia may represent a common pathological cell type linking osteoporosis and AD.

5. PERSPECTIVE ON NOVEL AD INTERVENTION STRATEGIES TARGETING THE BONE–BRAIN AXIS

5.1. Prevention and management of bone diseases

AD frequently coexists with bone disorders, which could promote brain AD pathology and increase the risk of AD, as discussed earlier. Thus, preventing and managing bone diseases such as osteoporosis and osteoarthritis has become increasingly crucial. There is evidence indicating a negative association between sedentary behavior and BMD. ²⁹³ Another study suggested that increased time spent in sedentary behavior was significantly associated with a greater incidence of dementia among older individuals. ²⁹⁴ In fact, one review concluded that sedentary behavior could impact protein homeostasis, mitochondrial function, cell‐to‐cell communication, and other aspects to accelerate biological aging. ²⁹⁵ This link between bone diseases and AD underscores the importance for AD patients to avoid sedentariness and increase physical activity in AD patients. Studies have indicated that physical exercise can enhance BMD in patients with osteoporosis ²⁹⁶ , ²⁹⁷ and improve fracture resistance and bone strength in AD model mice. ²⁹⁸ Furthermore, a meta‐analysis also corroborated the notion that exercise could enhance cognitive function, physical capability, and functional independence in individuals with AD. ²⁹⁹ Consequently, exercise is generally considered the most cost‐effective and accessible strategy for mitigating osteoporosis and AD.

In addition, some anti‐osteoporotic therapies show great potential for the prevention and treatment of AD. Bisphosphonates are available therapeutic drugs for the management of osteoporosis that inhibit bone resorption. At present, bisphosphonates are believed to prevent AChR inactivation and inhibit acetylcholinesterase activity, thereby exerting a protective effect against AD. ³⁰⁰ , ³⁰¹ , ³⁰² , ³⁰³ However, the other underlying mechanism remains unclear and needs further research. Studies have shown that vitamin D plays a role in the absorption of calcium and phosphate and that vitamin D supplementation can reduce bone turnover and increase BMD. ³⁰⁴ In addition, vitamin D deficiency has also been associated with several psychiatric disorders (eg, schizophrenia, autism spectrum disorder) and neurodegenerative diseases (eg, AD, Parkinson's disease). ³⁰⁵ , ³⁰⁶ , ³⁰⁷ , ³⁰⁸ Research has indicated that vitamin D is associated with processes such as neurological development, oxidative stress, calcium homeostasis, and inflammation. ³⁰⁵ , ³⁰⁹ One study showed that vitamin D supplementation exacerbated Aβ deposition in animal models and increased the risk of death in dementia patients. ³¹⁰ However, some scholars noted that this study equated vitamin D and calcitriol and drew inappropriate clinical conclusions. ³¹¹ The potential benefits of vitamin D supplementation in young and middle‐aged individuals should not be disregarded, as previous studies also indicated its potential efficacy in the early stages prior to Aβ formation. ³¹² , ³¹³ , ³¹⁴ , ³¹⁵ Recently, a prospective cohort study that included 269,229 participants in the UK Biobank showed that vitamin D supplementation could decrease by 17% the risk of AD. ³⁰⁸ In general, although vitamin D connects bones and the brain, the effects of vitamin D in the brain are diverse. Thus, its role in different pathophysiological processes of AD remains unclear. It may be important to investigate whether vitamin D supplementation could decrease AD risk in patients with osteoporosis in the future. A previous study revealed that aged patients who used heavy amounts of non‐steroidal anti‐inflammatory drugs (NSAIDs) had a greater incidence of dementia and AD. ³¹⁶ A meta‐analysis suggested a correlation between exposure to NSAIDs and a reduced risk of AD. ³¹⁷ In addition, two NSAIDs have been found to improve cognitive dysfunction and attenuate Aβ deposition in AD mice. ³¹⁸ The contradictory results of clinical studies may be related to the study population, ³¹⁶ , ³¹⁹ but the relationship between NSAID use and AD risk in bone‐related diseases still needs further investigation. Consequently, the prevention and timely treatment of bone diseases may be essential for AD management.

5.2. Targeting bone‐derived proteins

Based on the dual role of bone‐derived proteins, we propose that modulating the levels of bone‐derived proteins may be a potential therapeutic approach for AD. As discussed earlier, OCN has been verified as a neuroprotective and cognition‐enhancing peptide. Thus, the potential of OCN‐based therapy for AD should be investigated in in‐depth basic and clinical studies. The antagonism of bone‐derived proteins that exacerbate AD may also be promising for AD intervention. Animal experiments have shown that reducing bone‐derived SOST could reduce synaptic damage. ²⁷ Romosozumab, a monoclonal antibody that inhibits sclerostin, has gained approval for osteoporosis treatment, but there remains a lack of research on its treatment of AD or dementia. ³²⁰ , ³²¹ , ³²² Similarly, other bone‐derived proteins have similar limitations. Nevertheless, it can be speculated that augmenting protective bone‐derived proteins and inhibiting destructive bone‐derived proteins may have an impact on AD from the perspective of the bone‐brain axis.

5.3. BMSCs‐based therapy

The strong differentiation potential of BMSCs and the close communication between bone and brain led researchers to investigate whether BMSCs could be used to treat AD. A recent study revealed that transplantation of normal bone marrow  could replace more than 90% of the genetically mutated microglia in the mouse brain, restore the function of microglia, and reduce the pathological and cognitive function of AD in the mouse brain. ³²³ Several promising clinical trials of BMSC intervention in AD are currently under way (NCT02795052; NCT03724136). Compared with BMSC transplantation, BMSC‐derived EVs have greater stability, lack of immunity, lack of carcinogenicity, lack of ethical controversy, and easier access and preservation. ³²⁴ , ³²⁵ , ³²⁶ However, perhaps due to the complexity of the process of extracting EVs from BMSCs and the lack of standardized isolation, storage, and purification protocols, there are no relevant clinical trials on BMSC‐derived EV intervention in AD. Overall, to some extent, BMSC‐based therapy is a promising treatment for AD but still needs further development.

FIGURE 1.

Crosstalk between bone and brain in Alzheimer's disease (AD). AD can affect bone functions through various pathways, including AD pathogenic proteins, AD risk genes, neurohormones, neuropeptides, neurotransmitters, brain‐derived extracellular vesicles (EVs), and the autonomic nervous system. On the other hand, bone‐derived proteins, bone marrow‐derived cells, EVs, and inflammation may also influence brain AD pathology and cognitive function. Some intervention strategies targeting the bone–brain axis may be beneficial for the comprehensive management of AD. Aβ, amyloid beta; APOE, apolipoprotein E; APP, amyloid precursor protein; AD, Alzheimer's disease; BBB, blood–brain barrier; BMP, bone morphogenetic protein; BMSC, bone marrow mesenchymal stem cell; EV, extracellular vesicle; FGF23, fibroblast growth factor 23; FSH, follicle‐stimulating hormone; LCN‐2, lipocalin‐2; NPY, neuropeptide Y; NSAID; nonsteroidal anti‐inflammatory drug; OCN; osteocalcin; OPG, osteoprotegerin; OPN, osteopontin; OSTN, Osteocrin; PDGF, platelet‐derived growth factor; SOST, sclerostin; TREM2, triggering receptor expressed on myeloid cells 2.

6. CONCLUSION AND PERSPECTIVES

In this review, we provided an overview of clinical evidence supporting a strong link between AD and bone health. We discussed the bidirectional effects between AD and bone. AD pathogenic proteins, AD risk genes, neurohormones, neuropeptides, the autonomic nervous system, neurotransmitters, and brain‐derived EVs can impact bone health, and bone‐derived proteins, bone marrow‐derived cells, EVs, and inflammation may also influence brain AD pathology and cognitive function. In addition, the common risk factors for both bone diseases and AD, including aging, female sex, and sedentariness, may also be the connections and promoters of these two diseases. Therefore, the crosstalk between bone and brain provides new insight into dealing with AD and bone disease, and the two may interact to promote each other. The active prevention and timely treatment of bone diseases in elderly people may be beneficial to the prevention and treatment of AD. We summarized the potential mutual mechanisms of bone and brain in AD as well as related intervention strategies in Figure 1.

The crosstalk between bone and brain in AD holds great importance in the prevention and management of AD, so we propose possible perspectives for future research in this area. Additional animal and clinical studies are needed to investigate the contribution of bone homeostasis and bone diseases to AD pathogenesis. First, because the recently reported relevant clinical studies are heterogeneous, larger prospective cohort studies are warranted to elucidate the causal relationship between bone diseases and AD pathology as well as AD progression. Second, considering the significant disparities between animal models and humans, macaque and organoid models may be better able to explain the roles of bone‐derived substances in AD. Third, single‐cell transcriptome and protein sequencing could be used to clarify the key brain regions, cell types, and molecular pathways that link bone and AD. Fourth, the potential utilization of bone‐derived proteins or their antibodies as therapeutic targets for AD also needs to be validated in AD patients to translate the findings in animal experiments to clinical studies. In addition, considering the complex crosstalk between bone and brain, the identification of key molecules that promote AD in different bone diseases will be important for future treatments.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest. Author disclosures are available in the Supporting information.

Supporting information

Supporting Information

Liu Z‐T, Liu M‐H, Xiong Y, Wang Y‐J, Bu X‐L. Crosstalk between bone and brain in Alzheimer's disease: Mechanisms, applications, and perspectives. Alzheimer's Dement. 2024;20:5720–5739. 10.1002/alz.13864