Glucocorticoids are steroid hormones synthesized and secreted by the adrenal glands under the direct control of the hypothalamus and pituitary within a negative feed-back loop known as the hypothalamic-pituitary-adrenal (HPA) axis (Figure 1). In addition, the body’s master clock, which is located in the suprachiasmatic nucleus (SCN) of the brain, also modulates the activity of the HPA axis, so that under normal conditions adrenal glucocorticoid secretion follows a diurnal rhythm of approximately 24 hours [6, 7]. In humans, circulating glucocorticoid levels peak in the early morning, with a second smaller peak in the late afternoon, followed by a nadir during the night. In contrast, nocturnal rodents exhibit only one daily peak in the evening upon awakening. The pivotal role of the central master clock in maintaining diurnal glucocorticoid rhythmicity is evidenced by the observation that disruption of normal SCN signaling in rats abolishes the diurnal variation in circulating glucocorticoid levels [6, 8, 9].

In addition to circadian time cues (i.e., light exposure), various other stimuli including stress, inflammation, illness, or exercise can stimulate the hypothalamic paraventricular nucleus to release corticotropin-releasing hormone (CRH) and arginine vasopressin (AVP) into the hypophysial portal vessels that connect to the anterior pituitary gland [10, 11]. In the anterior pituitary, CRH and AVP bind to their respective receptors and trigger the release of adrenocorticotropic hormone (ACTH), which stimulates the synthesis of glucocorticoids in the adrenal zona fasciculata, and their release into the circulation [10–14]. However, glucocorticoid release can also occur through direct SCN-autonomic innervation of the adrenals via the splanchnic nerve [15].

As well as the classical 24-hour circadian rhythm, glucocorticoids are also secreted in a pulsatile fashion, displaying an ultradian rhythm of short-lived bursts of glucocorticoid release from the adrenals [16–18]. The pulsatile frequency of glucocorticoids is 90 min in humans and 60 min in rodents, likely due to differences in the biological half-lives of cortisol and corticosterone [19]. It has been proposed that the ultradian pulsatility in glucocorticoid secretion occurs due to delays between the activation of the HPA axis and the secretion of glucocorticoids, as well as delays in the negative feedback of glucocorticoid inhibition on the hypothalamus and pituitary [20].

At the tissue level, glucocorticoids freely diffuse across the cell membrane and into the cytoplasm where their concentrations are further finetuned by the actions of 11β-hydroxysteroid dehydrogenase (11β-HSD) isoenzymes 1 and 2 (Figure 2). 11β‐hydroxysteroid dehydrogenase type 1 (11β-HSD1), also known as a cortisone reductase, interconverts inactive cortisone into its active form, cortisol. Conversely, 11β-HSD2 catalyzes the conversion of cortisol into inactive cortisone [21]. While 11β-HSD1 is widely expressed in most tissues including the liver, adipose and bone, expression of 11β-HSD2 is confined to mineralocorticoid target tissues such as the kidneys, colon, and salivary glands [22]. In such tissues, 11β-HSD2 primarily acts to prevent cortisol preferentially binding to the mineralocorticoid receptor (MR), enabling aldosterone to bind to the MR instead [23].

Active glucocorticoids exert most of their functions by binding to the cytoplasmic glucocorticoid receptor (GR), which then undergoes a conformational change where it dissociates from accessory proteins, enabling the GR-ligand complex to translocate into the nucleus with the help of chaperone proteins [24, 25]. Within the nucleus the glucocorticoid-activated GR regulates gene expression through different mechanisms such as direct heterodimeric binding to glucocorticoid response elements (GREs) in the DNA or indirect monomeric tethering to other transcription factors bound to the DNA [25]. Growing evidence suggests that the GR can also act via non-genomic mechanisms through interactions with membrane lipids and cytoplasmic proteins to elicit rapid cellular responses that occur within a few minutes and do not require changes in gene transcription [26, 27].

In young adults, circulating glucocorticoid levels peak early in the morning upon awakening with a smaller second peak in the afternoon. Thereafter, glucocorticoid concentrations gradually decline with lowest levels observed around midnight [40]. In contrast, glucocorticoid rhythmicity in the elderly is characterized by a disturbed pattern of secretion whereby the morning peak in circulating glucocorticoids occurs earlier and is reduced, with higher cortisol secretion at night [4, 5, 41, 42] (Figure 1). Clinical studies suggest that compared to young adults, the nocturnal nadir of circulating glucocorticoid levels can be up to four-fold higher in people over the age of 70 years, translating into an overall increase in mean 24-hour plasma cortisol levels in older individuals [4, 5, 41]. Of note, some studies have found no differences in nadir or mean 24-hour serum cortisol levels between younger and older adults, possibly due to differences in age stratification or fewer intervals between blood sampling [43, 44]. Since cortisol levels are also highly variable within individuals and fluctuate due to changing metabolic demands or stress, it is recommended that these factors be controlled and samples be collected over several days to improve the validity of studies examining diurnal cortisol profiles [45].

There is good clinical evidence that the increase in circulating glucocorticoid levels with age occurs from a blunted and delayed inhibition of ACTH secretion within the HPA feedback loop [12–14]. This age-related impairment in the sensitivity of the HPA axis to circulating glucocorticoids concentrations could be due to factors such as reduced expression of the GR in various regions of the brain or neuronal degeneration. Studies in animals have shown that GR levels are decreased with ageing in the hippocampus, which is a key brain region known to exert a negative feedback response on the HPA axis in conditions of stress [46–49]. Supporting this, Murphy et al. [50] found that chaperone proteins responsible for GR nuclear translocation are also decreased in the hippocampus of aged rats, manifesting in deficient binding of the GR to GRE regions in the DNA. In addition to playing a key role in the stress response, the hippocampus is also particularly vulnerable to the effects of increased stress. In 1986, Sapolsky et al. [51] proposed the “Glucocorticoid Cascade Hypothesis” in which elevated glucocorticoid levels from impaired negative feedback during ageing may cause damage and atrophy to the hippocampus and in turn worsen the inhibitory regulation of glucocorticoid release. Moreover, this damage to the hippocampus would reduce the expression of GRs, resulting in decreased HPA sensitivity to glucocorticoids that could potentiate damage to the hippocampus in a feed-forward cycle [51]. While elegant, the Glucocorticoid Cascade Hypothesis was found to be inconsistent with other observations and was later adapted as the “Glucocorticoid Vulnerability Hypothesis” which recognizes that elevated glucocorticoid levels can make the hippocampus vulnerable to damage but are not the only determinant for neuronal loss in the aged hippocampus [52]. In addition to the hippocampus, decreased GR expression has also been observed in the prefrontal cortex, amygdala, and hypothalamus of aged compared to young animals [46, 48], which could contribute to reduced sensitivity of the HPA axis with ageing, although some studies do not support these findings [53, 54]. Although the SCN (master clock) does not express the GR, its degeneration and decreased sensitivity to light cues with ageing have also been linked to reduced amplitudes in daily circadian rhythms [55, 56] and could contribute to the dampening of glucocorticoid rhythms with age, considering light is the major cortisol zeitgeber (time cue) [6]. In summary, there are multiple factors that may contribute to ageing-related changes in the sensitivity of the HPA axis that drives increases in circulating glucocorticoid levels. These include reduced expression of the GR or neuronal degeneration in various regions of the brain such as the hippocampus. Additionally, increased inflammation during ageing could also be involved.

The term ‘Inflammaging’ first described by Franceschi et al. [57] refers to a state of low-grade chronic inflammation that develops with age. Inflammaging involves a steady increase in circulating cytokines including interleukin-6 (IL-6), IL-1 and tumour necrosis factor α (TNF-α) that indicate changes in innate and adaptive immune function and can predict morbidity, frailty and mortality in elderly people [58, 59]. In response to inflammation and increased cytokine levels, the HPA axis is activated to release glucocorticoids into the circulation [60]. At the tissue level, inflammatory cytokines can also stimulate 11β-HSD1 to increase local glucocorticoid concentrations [61, 62]. This mechanism has been described as ‘anti-inflammaging’, whereby the increase in glucocorticoid levels, attempts to counteract the inflammaging process in an adaptive defence response to inhibit immune activity [63]. However, chronically elevated endogenous glucocorticoid levels may also have negative implications and can lead to the development of prolonged immunosuppression that increases the risk of infection [64–66]. Additionally, glucocorticoid resistance may occur in which glucocorticoids are no longer able to exert their anti-inflammatory effects on target tissues due to alterations in the expression and function of the glucocorticoid receptor [63, 67].

On the other hand, there is some evidence that suggests increased glucocorticoid levels in older adults can induce significant changes in innate and adaptive immune responses and drive inflammaging. For instance, higher levels of perceived stress characterised by increased serum cortisol levels is associated with upregulation of inflammatory biomarkers, oxidative stress and immunosenescence [68]. Notably, elderly caregivers exhibiting greater distress with increased salivary cortisol levels, display blunted lymphocyte proliferation, lower IL-2 levels and reduced lymphocyte sensitivity to glucocorticoids compared to age-matched non-stressed caregivers [69]. In a different study, elderly caregivers were found to have increased IL-6 levels as high as four-fold compared to non-caregivers [70]. Other studies have also shown that distress can increase an individual’s risk of acquiring acute respiratory infections [65, 66] as well as inhibit the immune antibody response to vaccine administration [71]. Together these studies provide evidence that increased glucocorticoid levels from chronic distress can alter cell-mediated immunity and may accelerate premature ageing of the immune response [69]. As such, it is unclear whether the age-related rise in glucocorticoid levels is simply a consequence of inflammaging, or invertedly plays an active role in driving chronic inflammation with age.

While it is evident the interactions between the immune and endocrine systems are bimodal, several questions remain. 1) What is the timing of onset of inflammaging, neurodegeneration and increased glucocorticoid levels with age? 2) How does inflammaging affect diurnal rhythms in glucocorticoid secretion? 3) How do disturbances in glucocorticoid diurnal rhythmicity influence immune activity with ageing? Further studies that measure 24-hour diurnal rhythms in cortisol and inflammatory biomarkers over time will help unravel the complex interactions between immune and endocrine systems during ageing and may provide some important insights into the development of ageing-related pathologies.

Glucocorticoid levels in peripheral tissues such as the skeleton have been shown to increase with ageing, via the action of the glucocorticoid activating enzyme, 11β‐hydroxysteroid dehydrogenase type 1 (11β-HSD1) (Figure 2). Cooper et al. [72] showed in primary human osteoblast cultures that 11β-HSD1 reductase activity was positively correlated with the age of the bone tissue donor. Studies in aged mice have confirmed that 11β-HSD1 gene expression is increased with ageing at skeletal sites of the vertebrae and tibia compared to younger animals [3, 73]. Muscle, skin, adipose and hippocampal tissue also show elevated 11β-HSD1 expression with ageing [74–77]. Increases in cellular 11β-HSD1 levels and activity with age could be driven by higher circulating glucocorticoid concentrations since dexamethasone and cortisol treatment have been shown to enhance 11β-HSD1 actions in primary human osteoblast cultures [72]. In this way, 11β-HSD1 acts to increase glucocorticoid activation in the presence of elevated glucocorticoid levels [72]. Alternatively, local ageing-related changes in cells such as increased expression of inflammatory pathways and cytokines can also increase 11β-HSD1 levels in tissues independently of disturbances in circulating glucocorticoids [62, 78].

Osteoporosis is the most common bone disease globally, affecting approximately 1 in 3 postmenopausal women and 1 in 5 men over the age of 60 years [88]. With rapid population ageing, the incidence of osteoporosis is expected to markedly increase and place a significant burden on public healthcare systems. Osteoporosis is characterized by a reduction in bone mass and a deterioration in bone microarchitecture that compromises bone strength and increases the risk of fractures, particularly as people age [89]. Similar to humans, mice also exhibit trabecular and cortical bone loss and reduced bone density with age, making them suitable models to investigate the pathophysiology of osteoporosis. In the majority of murine strains, bone mass peaks at 4–6 months of age and then decreases thereafter [90–92].

Sarcopenia is a progressive skeletal muscle disorder involving the accelerated loss of muscle mass, strength and function [100]. As with bone mass, muscle mass declines substantially after 50 years of age [101]. Sarcopenia is primarily driven by ageing-related changes in muscle metabolism including a reduction in the number of myofibers and satellite cells, the infiltration of fat and fibrous tissue, and reduced muscle innervation that results in decreased protein turnover [101]. These changes can also be driven by multiple factors such as lack of physical activity, malnutrition or inadequate protein intake, chronic diseases, systemic inflammation, glucocorticoid exposure and hormonal changes [101]. The health implications of sarcopenia are often severe in older adults resulting in increased falls, disability, frailty and mortality [100]. As a result of population ageing, sarcopenia is expected to place a significant economic burden on health care systems. Given there is no approved therapeutic drug to treat Sarcopenia, research focusing on understanding how sarcopenia progresses with age will assist advances in diagnosis and therapies [100].

Osteoarthritis (OA) is a chronic degenerative joint disease involving the irreversible loss of cartilage, formation of abnormal bone and secondary synovial inflammation. Osteoarthritis is the most common form of joint disorders, affecting over 600 million people worldwide in 2020. The risk of developing OA increases with age, with approximately 40% of the population aged 70 years or older being affected [115, 116]. Karlson et al. [117] found women aged 70 years or older, were nine times more likely to have a hip replacement than those aged less than 55 years. Studies in mice show that ageing also affects the severity of OA induced by surgical destabilization of the medial meniscus (DMM) [3, 118]. The DMM model, developed by Glasson et al. [119] mimics a meniscal tear injury and is a commonly used technique to induce posttraumatic OA in mice. It has been shown that cartilage loss and abnormal bone formation are more severe when DMM is performed in 6 to 12-month-old mice, compared to younger animals [3, 118]. Intriguingly, microarray data from Loeser et al. [118] revealed that 493 genes displayed differential expression between young and old DMM mice, indicating that ageing affects the way in which cells respond to OA-inducing injuries. In fact, there are several ageing-related changes in the joint that contribute to the development of ‘spontaneous’ OA including cellular senescence and the formation of advance glycation end products that alter the cartilage and bone matrix, making it more susceptible to damage and injury. These findings demonstrate the importance of investigating OA in older mice as critical molecular pathways could be missed if only younger mice are studied in OA pathogenesis. Studies utilizing older mice could improve translatability to humans.

The severity of OA is also closely associated with several other risk factors including obesity, female sex, genetic predisposition, injury, and certain occupations that involve repetitive use of joints. As such, the multifactorial nature of OA and complexities surrounding what drives the progression of the disease have hindered the development of effective disease modifying treatments. As such, current treatment relies on the use of non-steroidal anti-inflammatory drugs and intra-articular glucocorticoid injection for primary pain relief, with end stage patients requiring joint replacement surgery [120, 121].

The use of therapeutic intra-articular glucocorticoid injections in the treatment of OA has been studied extensively but is not without controversy [122]. Recent studies have reported that corticosteroid injections, especially given continuously or over the long-term can increase the risk of disease progression [123–126]. Furthermore, 20–30% of patients do not respond to glucocorticoid treatment, possibly due to differences in the degree of inflammation and pain, stage of the disease or demographic characteristics such as BMI, sex, and age [127]. Given the widespread use of intra-articular glucocorticoid injections and the projected disease burden of OA in the ageing population, there is a need to establish best practice for intra-articular glucocorticoid injections including identifying which types of patients are most and least likely to benefit, to avoid adverse outcomes.