Osteoimmunology and aging — a frontier to explore

Abstract

Osteoimmunology is an interdisciplinary branch of immunology which studies the interplay of skeletal and immune systems. Both spatial and functional connections exist between the two systems, as most immune cells are generated in the bone marrow microenvironment, which facilitates the communication between the two systems. Moreover, immune cytokines such as RANKL (receptor activating Nf-kB ligand) and non-immune soluble mediators such as osteoprotegrin (OPG), made by immune and bone cells, respectively, interact to influence differentiation and activation of each other. The above interactions become of particular importance in the old age, when dysregulation of both systems yields changes affecting both length and quality of life. This perspective paper will outline both our current understanding as well as general gaps in knowledge, on geriatric osteoimmunology. We will also specifically address two highly prevalent diseases of aging, osteoarthritis and osteoporosis, as major sources of disability, loss of independence and increased morbidity and mortality in older adults, because cellular senescence appears to play a substantial pathogenetic role in both conditions, potentially opening new avenues for diagnosis and treatment.

Introduction

Age-related decrease in mobility is one of the most important contributors to decreased quality and length of life in older adults [1, 2]. Musculoskeletal changes occurring with aging are many, and affect muscles, bones and joints alike [3]. Aging of the muscles, including sarcopenia and frailty, were extensively covered by many reviews [2, 4–9], and will not be further explored here. In this perspective article, we will cover aging of bones with a specific focus on the interplay of the aging immune system with the aging skeletal system. We believe that the discipline of osteoimmunology [10] and the emerging intersect between osteoimmunology and gerontology both have the potential to improve quality of life of older adults by promoting mobility and activity levels.

Skeletal and immune systems share several anatomical and functional features. The bones are the home of all blood lineages, by virtue of housing bone marrow, the primary hematopoietic organ active throughout the life of mammals [11]. As such, differentiation of the bone-forming osteoblasts and bone-resorbing osteoclasts and differentiation of various blood lineages from the hematopoietic stem cells (HSC) can be impacted and modulated by the same soluble mediators, including IL-7 [12], stem cell factor, CXCL12 [13], and others. Moreover, stromal compartments of primary and secondary lymphoid organs as well as the bone and joint cells are producing and/or are responsive to the members of tumor necrosis factor (TNF) family molecules, including but not limited to RANKL (receptor activating Nf-kB ligand), and express competing receptors for this molecule, RANK and osteoprotegerin (OPG), a RANKL decoy receptor [14–16]. These common regulators provide the molecular basis for cross-talk and cross-influence between the two systems. Moreover, age-related changes affect both these molecules and the cells that secrete or receive signals from them [3, 13, 17, 18], providing at least one mechanistic input into age-related alterations of bone and immune tissues. At present, however, molecular details of the above crosstalk and of its changes with aging require further studies.

Osteoporosis and osteoarthritis (OA) are two debilitating chronic diseases affecting the musculoskeletal system in older adults [2, 19]. The involvement of the immune system and its changes with aging in these two diseases remain incompletely understood. While osteoporosis in its fundamental mechanisms does not begin as an inflammatory process, the end result of imbalance between the bone-forming osteoblasts and bone-resorbing osteoclasts intersects with inflammation, as synovial and synovial fluid macrophages constitute the osteoclast precursor pool, that respond to many systemic and inflammatory local cues in a similar manner to other tissue-residing macrophages [14].

New research is showing that the pathobiology of osteoporosis substantially involves cellular senescence of bone and joint tissues. OA, meanwhile, was for many decades and even centuries believed to be a degenerative disease fundamentally different from rheumatoid arthritis and other autoimmune arthritic conditions [1, 20, 21]. While fulminant autoimmunity does not underlie the pathogenesis of OA, it is now very clear that osteoarthritis has an identifiable secondary inflammatory component and may also be mediated or accompanied by cellular senescence, and these factors, alone or together, may propagate or even cause OA. Below, we will discuss these issues as related to aging and older adults and seek to outline the most important specific questions of aging biology that the field of osteoimmunology is positioned to answer.

Osteoarthritis (OA)

Epidemiology of OA

OA is the most prevalent type of arthritis, impacting 14.8% of individuals over the age of 30 globally in 2020 [22, 23]. The incidence of OA is rising due to population aging, increasing obesity rates, and longer lifespans. By 2050, nearly one billion people are expected to suffer from OA. OA prevalence rises with age, with rates of 3.5% for ages 30–60, 23.2% for ages 50–69, and 38.4% for those 70 + [22, 23].. Of the affected joints, the knee is most common, followed by the hand and hip, with knee OA peaking in the 80–84 year old group. Among individuals aged 70 and older, the condition ranked seventh among causes for years lived with disability (YLDs) [22, 23].

Risk factor impact analysis for OA in 2020 showed that obesity accounted for approximately 20% of the OA-associated disability [24]. The most common locations for osteoarthritis are the knees and hands. By 2050, the osteoarthritis of the knee is expected to increase by 75%, and osteoarthritis of the hand by 50%. Osteoarthritis affects women more than men, so that in 2020, women accounted for 61% of osteoarthritis cases, compared to 39% in men, and projections indicate that a higher proportion of women will continue to experience this condition[22, 24]. Studies are underway to investigate the reasons for gender differences in the prevalence of OA, including hormonal factors and anatomical differences. Other contributing factors to OA pathogenesis and prevalence include joint injuries and genetics [22–25]. At present, there is no effective cure for osteoarthritis; therefore, it is crucial to emphasize patient-centered prevention strategies, early intervention, and reassessment and modification of treatment plans to slow its progression. Additionally, efforts should be made to improve access to costly treatments, such as joint replacements, especially in low- and middle-income countries.

Diagnosis and pathogenesis of OA

Osteoarthritis (OA) involves the degeneration of articular cartilage, subchondral bone remodeling, leading to secondary synovial inflammation, which results in joint pain, stiffness, and reduced mobility. The initial insult that begins the cascade of OA pathogenesis is not well understood, although factors mentioned above such as injury, joint pressure due anatomical stress or excessive weight, as well as genetic, hormonal, and other influences, all have the potential to impact microtrauma and joint healing process. More recently, cellular senescence changes have been described in OA-affected bone and joint tissues, raising the possibility that accumulation of senescent cells above some critical level may impair the joint function significantly to become the initiating process in OA. Cellular senescence is a cell fate affecting many mesenchymal tissues, including tendons and synovium, characterized by irreversible cell cycle arrest mediated by p16 or p21 [Kirkland; Campisi], and is often but not always accompanied by altered secretome, also called senescence-associated secretory phenotype or SASP [Campisi]. This altered secretome is almost always characterized by production of inflammatory cytokines and chemokines, including but not limited to IL-6, IL-8, IL-1β, and TNFα, as well as by extracellular matrix-remodeling enzymes such as metalloproteinases, all of which have the potential to propagate inflammation and delay or alter healing [Campisi; Francheschi].

While much remains to be learned in preclinical and clinical studies on the interactions of these cells and molecules with OA as a function of age, the initial studies of senescent cells have ushered a new era in OA research, diagnostics, and treatment. The importance of identifying senescent cells, including their exact phenotype, lies in the increased availability of senolytic and senomorphic treatments [REF Kirkland] that could consequently be beneficial to OA treatment. Indeed, research into chondrocyte plasticity has demonstrated that modulating connexin43 can mitigate cellular senescence and facilitate cartilage regeneration, presenting possible therapeutic opportunities, that are further discussed below under AO therapies [25].

Osteoarthritis (OA) diagnosis has also evolved significantly between 2023 and 2025, integrating traditional clinical assessments with advanced biomarker and imaging technologies. Clinically, OA is diagnosed in individuals aged 45 and over who experience activity-related joint pain with either no morning stiffness or stiffness lasting no longer than 30 min, without the routine need for imaging unless atypical features are present [26]. Diagnosis is made through a combination of medical history, physical examination, and imaging of the affected joints. X-rays provide the images necessary to identify osteoarthritis [27]. Occasionally, blood tests are performed to exclude other conditions. Imaging is essential for diagnosing OA. X-rays can detect joint space narrowing and osteophytes but may miss early cartilage changes. MRI provides detailed views of cartilage, bone marrow lesions, and synovitis, helping in early detection [27].

Recent studies have found new biomarkers—proteins and metabolites—that improve early diagnosis and help categorize patients for personalized treatments [28]. Specifically, genomic studies have identified 26 new genes associated with osteoarthritis (OA), including MGP and ALDH1A2, which are presently undergoing confirmatory and clinical validation tests. Other biomarker research has developed a dual-biomarker algorithm that accurately distinguishes OA from inflammatory arthritis by measuring cartilage oligomeric matrix protein and interleukin-8 levels in synovial fluid [29].

Biophysical emerging technologies, such as bioimpedance combined with deep learning algorithms, have also shown considerable potential in the non-invasive detection of osteoarthritis (OA), achieving accuracy rates of up to 98% in identifying knee OA. These integrated diagnostic methodologies enhance early detection and personalized management of OA, thereby improving patient outcomes [30].

Advances in OA treatment—repurposed drugs, biologicals, and senolytics

Therapeutic strategies for OA are also advancing. As mentioned, no curative treatment for OA exists at the present. Beyond symptomatic treatments to alleviate pain and stiffness, which also fail in many patients, repurposing of drugs has found its use in OA. Methotrexate, commonly used for rheumatoid arthritis, has been found to alleviate hand OA symptoms. Semaglutide, a GLP-1 receptor agonist, aids in weight loss and reduces knee OA pain, indicating a connection between metabolic health and joint function. Moreover, preliminary research points to the possibility that semaglutide may offer neuroprotective effects, possibly safeguarding against Alzheimer’s disease [31, 32]. Related to that, off-label use of Ozempic in OA has also shown efficacy. Besides the clinical trials, semaglutide usage has been correlated with enhanced exercise capability in individuals diagnosed with diabetes and heart failure. Semaglutide (wegovy) has received approval for reducing the risk of stroke, myocardial infarction, and other critical cardiovascular events in patients who are overweight or obese [32].

Innovative cell and biological treatments are also being developed. One promising method involves the transplantation of lab-grown cartilage originating from nasal cells into damaged knees, which has demonstrated significant improvements in mobility and reduction of pain. Additionally, liposomic lubricants have been found to suppress shear-stress-induced inflammatory gene expression within joints, presenting a potential strategy for slowing the progression of osteoarthritis [33].

Several other strategies targeting senescent cells that cause inflammation and cartilage damage in joints are being tested, using monoclonal antibodies and senolytic agents [34]. For instance, anti-DPP4 monoclonal antibodies can induce antibody-dependent cellular cytotoxicity, leading to the clearance of senescent cells by natural killer cells. Similarly, chimeric antigen receptor (CAR) T cells targeting urokinase-type plasminogen activator receptor (uPAR), a marker upregulated in senescent chondrocytes, have shown potential in preclinical studies [34]. While these therapies are still under investigation, they represent a shift towards addressing the underlying cellular mechanisms of OA, with the goal of not only alleviating symptoms but also modifying disease progression.

Another promising approach uses of UBX0101, a small molecule that disrupts the interaction between MDM2 and p53, to induce apoptosis of senescent cells. In preclinical models, UBX0101 reduced OA-related pain and cartilage erosion, and promoted cartilage regeneration [35, 36]. These effects were also observed in human knee tissue cultures [35], suggesting translational potential.

Finally, ABT-263 (navitoclax), a senolytic of the BCL-2 family inhibitor class, has demonstrated efficacy in selectively eliminating senescent synovial mesenchymal stem cells (MSCs) from OA patients. Treatment with ABT-263 improved the chondrogenic and adipogenic capacities of these MSCs and reduced the expression of senescence markers and inflammatory cytokines [37].

Overall, there is a substantial and renewed promise in the new generation treatments, and their clinical testing is now pending to validate the safety and efficacy of these promising approaches to ameliorate and cure OA.

Osteoporosis

Epidemiology of osteoporosis

Osteoporosis is a systemic disease characterized by reduced bone mass and deterioration in bone tissue microarchitecture, which increases bone fragility and, subsequently, the risk of fracture. With the demographic shift towards the older population worldwide, osteoporosis is becoming a devastating threat to public health, as it has enormous clinical and economic consequences. The health burden of osteoporosis is estimated to be greater than that of cancer, with the exception of lung cancer [38].

Prevalence of osteoporosis is estimated at 21.7% in the older adult population worldwide, with 35.3% in women and 12.5% in men [39]. Prevalence rates vary across the world due to different genetics, nutrition, lifestyle, environmental factors, and demographics. Some studies found that the prevalence rate in women in China to be as high as 23.4%, with age-specific prevalences going up and reaching 63.1% in women over 80 years of age. In men, prevalence also increases with age, but it remains lower throughout and is estimated at 38.3% in men over 80 [39]. However, the real prevalence is very difficult to establish with accuracy, as osteoporosis is a silent disease with clinical symptoms occurring only after an incident fracture [40]. On physical examination, consequences of prior (or so-called prevalent) vertebral fractures can be noted, such as back pain, loss of height, or kyphosis [21]. The clinical importance of osteoporosis therefore lies in greatly increased fracture risk, especially of the hip, spine, and wrist. These fractures often occur with minimal stress or strength force impact [40].

Globally, women have twice the risk of hip fractures compared to men, with Caucasian women being at highest risk [41]. Women over 50 years of age have a chance of fracture as high as 40% or even more [39, 42]. This severely impacts both the quality and length of life in older adults, as hip fractures are associated with hospitalization, autonomy loss, and high mortality rates [40]. Around 50% of deaths due to fractures in women were due to hip fractures, 28% due to spine and 22% to other fractures, due to immobility and consequent infections, dominantly pneumonia [38]. With the overall prevalence being higher in women, it is important to note that mortality rates are still higher in men. This can be explained by present comorbidities and higher risk of infection [43]. Early internal fixation can allow patients to be mobilized faster and therefore reduce the risk of postoperative complications, which can lessen mortality [44].

Diagnosis and pathogenesis of osteoporosis

Bone strength is defined by the qualitative and quantitative properties of bone tissue. Quality of bone is determined by its structural and material characteristics. Structural characteristics refer to macro- (hip axis length, cross-sectional moment of inertia) and microarchitecture (trabecular and cortical thickness, trabecular connectivity), while material characteristics involve mineralization, collagen type and configuration, and existing damages to bone tissue (micro fractures) [21, 38]. The diagnosis of osteoporosis is made by measurement of bone mineral density (BMD), which represents bone mineral content per unit of volume (g/cm³) or area (g/cm²) and can be measured by several densitometric techniques, of which dual-energy X-ray absorptiometry (DEXA or DXA) is the gold standard. Other methods include quantitative ultrasound, quantitative computed tomography radiographic absorptiometry, and others [38]. Serum levels of bone metabolic markers can also be used in predicting fracture risk, especially when combined [44].

Osteoporosis can be divided into primary and secondary. Primary osteoporosis is more common [45] and it is the result of age-dependent changes. It can be further divided into postmenopausal, older-age (also termed senile, a term we strongly believe should be abandoned due to its pejorative and ageistic connotation) and idiopathic, which is usually seen in children and has no known etiology besides genetics [39, 42].

Etiology of osteoporosis involves genetic predisposition, endocrine factors (levels of estrogen, testosterone, calcitonin, and parathyroid hormone — PTH), nutritional factors (protein, calcium, magnesium, phosphorus, and vitamin D intake), age-dependent changes, and lifestyle factors (smoking, alcohol use, low physical activity level) [42].

In a normal skeleton, bone formation and bone resorption are balanced with each other and together constitute the process of bone remodeling, which is necessary to maintain adequate bone quality. If bone resorption outpaces bone formation, the result is bone loss. Postmenopausal osteoporosis is characterized by high bone turnover rate, as bone resorption vastly outpaces bone formation. On the other hand, older-age osteoporosis is not characterized by high bone turnover, but rather by reduced bone formation and resorption, but at different rates so that bone formation is outpaced by resorption [15]. Older-age osteoporosis also exhibits significant changes in number and function of bone marrow stromal cells (BMSCs), which normally differentiate into osteoblasts, chondrocytes, and adipocytes. Specifically, with aging, adipogenic differentiation becomes dominant over osteoblastic [46]. It is here that additional research into the interplay between an aging stromal bone marrow niche and the BMSC is needed to better understand the age-related changes in osteoporosis.

Similar to OA, cellular senescence of BMSCs is gaining attention as a major contributor to older-age osteoporosis. As mentioned, cellular senescence is a cell fate response to accumulation of macromolecular damage due to different factors (oxidative stress, radiation, macromolecular damage) [47–49]. In mesenchymal cells, this often, but not always, leads to activation of the p53/p21 and/or p16/Rb signaling pathway [50–55]. Following activation of these cell cycle inhibitors, senescent cells enter a state of irreversible cell cycle arrest. A substantial, but variable, number of these cells also assume an altered secretory phenotype, secreting of pro-inflammatory cytokines, chemokines, and extracellular matrix (ECM)-modifying enzymes, which is referred to as senescence-associated secretory phenotype (SASP) [48, 49, 55]. Via its soluble products, SASP cells can also induce senescence and/or alter the function of neighboring cells [49], including BMSCs [45, 56].

Pathogenesis of osteoporosis can be summarized by its most important pathogenetic mechanisms, inseparable from one another, including estrogen deficiency, inflammation, oxidative stress, cellular senescence, and epigenetic factors [57]. Bone loss due to estrogen deficiency is the main initiating pathogenetic mechanism in postmenopausal osteoporosis. It is more prominent in women as the decrease in estrogen level is greater, but it affects men as well, as estrogen is a major bone regulator regardless of sex [42]. Age-related changes in bone microarchitecture differ between men and women. Women lose trabecular connectivity, while men lose trabecular thickness and cortical bone mineral density (BMD), with conversion of cortical bone to trabecular bone [43].Decrease in estrogen leads to bone loss in cancellous and cortical bone, as it increases apoptosis of osteoblasts and inhibits their differentiation and proliferation, acting through estrogen receptor α. Estrogen deficiency also increases production of pro-inflammatory cytokines, including tumor necrosis factor α (TNF α), interleukins (IL) 1, 4, 6, and 17, and interferon γ (IFN γ), leading to increased osteoclast maturation and bone resorption [57]. Estrogen deficiency both directly and indirectly (through IL 17) increases production of RANKL, further promoting bone resorption. There is also a notable increase in reactive oxygen species (ROS) during the estrogen-deficient state. Aging itself decreases mitochondrial efficiency due to mitochondrial DNA damage and elevated mitochondrial ROS production during ATP synthesis. ROS decrease osteoblastic and increase adipogenic differentiation of bone marrow stromal cells, cause apoptosis of osteoblasts and osteocytes, and increase osteoclast differentiation and bone resorption. Epigenetic factors, especially dysregulation in DNA and histone methylation, can lead to osteoporosis, as osteogenic and adipogenic differentiation of BMSCs depend on opposing patterns of methylation [57].

With age, there is also an increased production of trench-forming osteoclasts, which may be more aggressive than regular osteoclasts [56], and there are changes to bone microvasculature including reduction of type H endothelium and the number of pericytes, which reduces proliferation of BMSCs [58]. Moreover, aging by itself leads to increased and dysregulated production of proinflammatory (IL-1b, IL-6, IL-8, TNFα), as well as wound healing mediators (TGFβ, type 2 cytokines), that are conducive to both propagation of inflammatory processes as well as to collagen deposition, fibrosis and scarring [59, 60]. It will be of critical importance to evaluate by deep molecular profiling what is the quantitative and qualitative impact of these changes on the progression and trajectories of older-age osteoporosis.

Secondary osteoporosis is caused by certain medications (glucocorticoids, antiepileptics) or diseases (hyperthyroidism, hypogonadism, diabetes mellitus, connective tissue disorders). Glucocorticoid-induced osteoporosis is the most common and is attributed to abnormal bone metabolism, with inhibition of osteoblast differentiation and stimulation of osteoclast maturation. This leads to rapid bone loss and very high risk of fractures, with rates as high as 30–50% among patients on long-term glucocorticoid treatment [61]. While of considerable clinical interest, these conditions will not be discussed further.

Treatment of osteoporosis — a role for senolytics?

Treatment of osteoporosis must be personalized and individually tailored to each patient to include nutrition and lifestyle modification, exercise, and pharmacological treatment. The main goal of osteoporosis management is fracture prevention. Primary prevention involves screening for patients without symptoms. Older adults and especially postmenopausal women need to be assessed for risk factors and family history, and a thorough physical examination must be performed. BMD testing should be done for women over 65, younger postmenopausal women with additional risk factors, and anyone with an osteoporotic fracture [21]. Secondary prevention aims to eliminate refractures, especially in the first 2 years after an initial fracture when the risk is highest. Thus, early identification, assessment, and treatment of these patients is crucial, with the main goal of rapidly reducing that risk [62]. Fracture prevention also requires advocating for a balanced diet, with adequate calcium and protein intake, and vitamin D supplementation if necessary [43]. Additionally, falls prevention programs are essential. This requires medication review for side effects that may promote falling, use of canes or other walking aids, home safety adaptations, exercises of strength and balance, vision correction, and fall prevention education [38, 63, 64].

On the pharmacological side, by mechanism of action, commonly used drugs belong to antiresorptive agents (bisphosphonates and denosumab), anabolic agents, and selective estrogen receptor modulators (SERMs). Bisphosphonates (alendronate, risedronate, ibandronate, zoledronic acid) induce apoptosis of osteoclasts by inhibiting enzyme farnesyl pyrophosphate synthase. They are ordinarily used as first-line therapy, especially in postmenopausal osteoporosis [57], and can be used orally or intravenously, albeit with significant side effects, particularly on prolonged [42, 65]. Of interest, zoledronic acid appears to confer prolonged, residual antiresorptive effects for several years, which can enable periodic yearly use to maintain BMD [62]. Denosumab is a monoclonal antibody that binds to RANKL, thereby inhibiting osteoclastogenesis [57]. It can be sequentially combined with bisphosphonates [62] to prolong effects and reduce toxicity.

Anabolic agents (teriparatide, abaloparatide, romosozumab) prevent fractures better than antiresorptive agents in very high-risk patients. In addition, there is a confirmed benefit of using anabolic agents first, before antiresorptive medicine, as that will maximize BMD increase and limit the bone loss after discontinuation of anabolic drugs [57, 62]. Teriparatide is a parathyroid hormone analog, while abaloparatide is an analog of PTH-related peptide. They bind to PTH receptor, inducing anabolic effects and osteoblast differentiation. Of the two, abaloparatide causes larger ostegenic effects, can increase bone formation without increasing bone resorption, and is tolerated better [42]. Romosozumab is an antisclerostin antibody that inhibits the osteocyte protein sclerostin, a negative regulator of bone formation; it carries an increased risk of adverse events [42, 43, 57].

Selective estrogen receptor modulators (SERM) act as partial estrogen agonists in bone tissue (to prevent bone loss), while simultaneously beneficially affecting both lipid metabolism (decreasing LDL and lipoprotein A, and increasing HDL cholesterol), and the uterus and breast tissue, eliminating the risk for higher breast and endometrial cancer incidence. This group includes raloxifene and bazedoxifene. Raloxifene was found to decrease vertebral fractures by 30–50% [66], but there were no notable reductions in other fractures [38, 42]. The main side effect is a higher risk for venous thromboembolism [38, 57].

The holy grail of osteoporosis therapy lies in understanding its pathogenetic origins at the molecular and cellular level. Here again, the exciting and promising new treatments include senolytics and senomorphics, which can eliminate or modify senescent cells and help with bone mass preservation [57]. There have been numerous attempts at BMSC transplantation, as well as newer cell-free stem-cell-based therapies that use extracellular vesicles with paracrine factors, which are isolated from mesenchymal stem-cells generated from human embryonic stem-cells. Further studies will be necessary to both standardize these therapies and to evaluate their preclinical and clinical efficacy [42].

Future directions in gerontological osteoimmunology

Overall, gerontological osteoimmunology is emerging as an exciting and promising area of science, poised to positively impact billions of lives. This subdiscipline has much potential to grow. Below we outline what we see as the five most immediate and most critical and exciting basic and translational goals that should be addressed in the near future. We specifically focus on molecular interactions, cellular senescence, and future treatments, with the hope that federal and other funding agencies will support these goals.

• Undertake systematic fundamental studies of immune-skeletal molecular interactions during aging

• Explore the cross-talk of the two systems in OA and osteoporosis

• Establish the role of cellular senescence and of SASP in OA and osteoporosis

• Test the safety and efficacy of senolytic and senomorphic treatments in randomized clinical trials

• Optimize the combinations of senolytics, senomorphics, and/or targeted anti-inflammatory biologicals to efficaciously treat OA and osteoporosis.

Abbreviations

HSC

Hematopoietic stem cells

OA

Osteoarthritis

OPG

Osteoprotegrin

RANKL

Receptor activator of nuclear factor kappa-B ligand

SERM

Selective estrogen receptor modulators

Author contribution

The authors roles were as follows: MI and JZN conceptualized and wrote and edited the manuscript. IDA and MJF wrote and edited the manuscript.

Funding

Supported by USPS awards AG020719, OT2HL161847, AG052359, and the Bowman Foundation for Medical Research (JZN) AR062173 (IEA) and R25AG076387 and Ann and Alden Hart Endowment (MJF).

Declarations

Competing interests

The authors declare no competing interests.