The Aging Thyroid: A Reappraisal Within the Geroscience Integrated Perspective

Abstract

The thyroid plays a crucial and pervasive role in physiology (metabolism, thermogenesis, and immunity, among others) and its aging and related changes in thyroid hormone production contribute to the common occurrence of thyroid diseases in elderly and to age-associated changes in other organs and systems. We address the complexity of thyroid aging following the basic suggestions of geroscience. This integrative new perspective identifies a few basic molecular mechanisms or “pillars” (inflammation, adaptation to stress, loss of proteostasis, stem cell exhaustion, metabolism derangement, macromolecular damage, and epigenetic modifications) as a unifying conceptual framework to understand the aging process and age-associated diseases. Within this scenario, we review available data on the presence and role in the thyroid of alterations of such mechanistic pillars, paying particular attention to (i) inflammation, focusing on cellular senescence and age-associated dysbiosis (alteration of gut microbiota); (ii) telomere shortening as an example of macromolecular damage; (iii) proteasomal function, including mitophagy and autophagy; (iv) stem cells and cell renewal; (v) energy metabolism and mitochondrial dysfunction; and (vi) age-related epigenetic changes, focusing on DNA methylation. Overall, the study of these topics in the thyroid is in its infancy and deserves much more attention. Finally, thyroid function in centenarians as a model of healthy aging is reviewed within the framework of possible adaptive mechanisms involving the thyroid to attain longevity. Accordingly, the concept of “thyroid biography” is proposed to grasp the complex combination of factors (including endocrine disruptors and lifestyle habits) impinging lifelong on thyroid function at the individual level.

Essential Points

• Thyroid aging is fully understood when investigated within the integrated perspective of geroscience, which suggests that aging and chronic age-associated diseases share a common set of basic molecular mechanisms (“pillars”)

• Available data on the thyroid and these mechanisms (inflammation, adaptation to stress, loss of proteostasis, stem cell exhaustion, metabolism derangement, macromolecular damage, epigenetic modifications) are thoroughly scrutinized

• The emerging scenario is that the knowledge on these pillars in the thyroid is still scarce, suggesting that the fundamental mechanisms of thyroid aging are still largely unexplored

• The thyroid status of centenarians, assumed as the best model of healthy aging in humans, is reviewed within the framework of possible lifelong adaptive mechanisms involving the thyroid to attain longevity

• The comprehensive concept of “thyroid biography” is suggested to grasp the combination of genetics, lifestyle habits, and environmental factors impinging lifelong on thyroid function at the individual level, thus explaining the large heterogeneity of thyroid function in the elderly

The global aging of humans is considered one of the major challenges owing to its pervasive economic, medical, and cultural implications. People >85 years of age, namely the “oldest old,” represent the segment of the population that is growing faster worldwide, and one or more age-associated diseases (AADs) affect most of them. Thus, it is urgent to grasp the complexity of the aging process and its underpinning molecular mechanisms. According to the integrated view of geroscience, (i) aging is the predominant risk factor for AADs, and aging and AADs share a common set of basic mechanisms (1); (ii) aging and AADs are not separate events but rather parts of a continuum where precise boundaries do not exist, and where the two extremes are the centenarians, who largely avoided or postponed most AADs and are characterized by decelerated aging, and persons who suffered one or more severe AADs in their 60s to 80s and show signs of accelerated aging (2, 3); and (iii) AADs can be conceptualized as a manifestation of accelerated aging (2).

Thyroid aging within the new perspective of geroscience

With age, changes occur in all body systems, including the endocrine system, and thyroid function is particularly important owing to its central role in metabolism, thermogenesis, and immunity, among others, as well as its contribution to most common chronic AADs. The network of basic mechanisms involved in the aging process identified by a group of international experts, and collectively indicated as the “pillars” of geroscience, includes inflammation, adaptation to stress, loss of proteostasis, stem cell exhaustion, metabolism derangement, macromolecular damage, and epigenetic modifications (1). These pillars do not operate separately but are interconnected, influencing and modulating each other, thus constituting an integrated network (1).

In this review, we propose a reappraisal of thyroid aging and thyroid age-related dysfunction from the new geroscience perspective, focusing on the above-mentioned mechanistic molecular pillars that drive the aging process, as illustrated in Fig. 1, and on centenarians and their offspring as the best model of healthy aging in humans (4). Moreover, taking into account that a major characteristic of old subjects is the large heterogeneity of their phenotype, including thyroid function, and accepting the challenge of personalized medicine, we propose the concept of “thyroid biography” to grasp the complex combination of lifestyle habits and environmental factors impinging lifelong on thyroid function at the individual level.

Figure 1.

Molecular mechanisms of aging (pillars) involved in thyroid aging, according to the new geroscience perspective.

Thyroid Aging Within the Context of the Basic Mechanisms of Biological Aging

In the following paragraphs, we present a detailed analysis of the literature focused on the above-mentioned pillars of geroscience to check whether and how much the thyroid aging fits this unifying conceptual framework of the aging process.

Inflammation

Inflammation is one of the geroscience pillars, and accumulating evidence indicates that aging is associated with a chronic, low-level inflammation termed “inflammaging” that represents a major contributor to the pathogenesis of AADs (2). The peculiarity of inflammation is that alterations in any one pillar converge and fuel inflammation, which in turn affects all other pillars (2). Inflammaging involves basically the innate immune system, but acquired immunity also contributes to this phenomenon, which is deeply related to all of the changes occurring with age in the immune system, collectively indicated as immunosenescence (5). Recently we have proposed that inflammaging can be considered as a complex mechanism of adaptation that depends on the context in which it develops and that can be interpreted as a negative (favoring AADs) or positive (promoting health) phenomenon. Therefore, inflammaging is an overall adaptation of the entire body within an integrated view of organs and systems and is the result of the continuous activation of mechanisms to establish progressively new homeostatic equilibria (6, 7).

Immune effects of thyroid hormones

A plethora of data indicates that thyroid hormones (THs) influence innate and acquired immune functions [chemotaxis, phagocytosis, generation of reactive oxygen species (ROS), and cytokine synthesis/release], as particularly evidenced in hypothyroid and hyperthyroid conditions. Although these data show that THs modulate both innate and acquired immune responses, it has not been clearly demonstrated whether changes in thyroid function with age are correlated with an inflammatory state of the gland and/or with the systemic inflammation or inflammaging.

The available data regarding the effects of THs on immune responses in elderly subjects are summarized as follows:

• In 93 healthy late-middle-aged euthyroid subjects, TH concentration was positively correlated with immune functions, including the level of complement proteins C3 and C4, C-reactive protein (CRP), phagocyte activity, percentage and number of natural killer cells and T-cells, and IL-6 expression by activated monocytes (8).

• THs, particularly T3, increase metabolic activity and oxygen consumption and contribute to oxidative stress both in the short-term and in the long-term range (9). ROS production may facilitate various immune functions, such as the bactericidal activity of macrophages through the reduced form of NADPH oxidase activation. Additionally, in immune cells, free T4 (FT4) is able to increase ROS production, leading to cell migration in tissues in response to chemoattractant molecules (10).

• THs can bind to integrin α_vβ₃ on the macrophages and activate phosphoinositide 3-kinase and extracellular signal-regulated protein kinase 1/2 pathways followed by the upregulation of inducible nitric oxide synthase favoring the intracellular killing of bacteria (11). Alternatively, FT4 enters the macrophage through monocarboxylate transporters MCT8 or MCT10, where the prohormone T4 is converted to active hormone T3 by deiodinase (D2), resulting in an increase in phagocytosis and cytokine response (12). These effects are mediated, at least in part, by thyroid hormone receptor (TR)α, which is the predominant TR isoform in macrophages, and knockout mice for TRα have an aberrant macrophage function (12).

• Macrophages contribute to immune system surveillance by sensing and adapting to local stimuli and microenvironmental signals (13). It has been recently shown that free T3 (FT3) negatively contributes to the differentiation of bone marrow–derived monocytes into nonpolarized macrophages (14). FT3 promotes the generation of M1 macrophages (proinflammatory phenotype), even after the differentiation and activation of monocytes into M2 macrophages (14). In vivo, FT3 increases the number of resident macrophages in the peritoneal cavity, whereas it reduces the content of the recruited monocyte-derived cells in the inflamed locus (potentially damaging). In an in vivo model of lipopolysaccharide-induced endotoxemia, FT3 protects mice from developing endotoxic shock. Whereas low FT3 levels increase inflammatory cell recruitment into tissues, an opposite phenomenon occurs when FT3 levels are restored (14).

• In the Leiden 85-Plus Study, Rozing et al. (15) demonstrated that higher levels of circulating CRP and IL-6 were significantly related to lower serum levels of FT3. On the contrary, after lipopolysaccharide stimulation of whole blood in vitro, higher levels of serum FT3 were associated with a higher production of proinflammatory cytokines (IL-1β, IL-6, TNF-α). The authors postulated that serum FT3 stimulates the production of the proinflammatory cytokines, while proinflammatory cytokines, in turn, blunt the stimulatory effect of FT3 by lowering peripheral TH levels. This influence of cytokines probably occurs through regulation of peripheral deiodinase activity, although this putative mechanism is not yet demonstrated. The stimulatory effect of T3 on cytokine production is likely mediated via nuclear receptors regulating genes involved in the cell-mediated immune response (15).

• TSH is able to increase IL-6 production from adipocytes derived from abdominal subcutaneous fat but not from the omental deposit. The basal IL-6 release is higher for preadipocytes than for differentiated adipocytes, independently from their origin, indicating an effect of TSH on adipocyte differentiation (16).

Overall, circulating THs have profound effects on neutrophil, macrophage, and dendritic cell function and, generally, a rise in TH levels results in an amplification of the proinflammatory response of these cells. In this framework, a reduction of THs during aging could be considered a form of adaptation to reduce inflammation/inflammaging and could play a role in immunosenescence. Thus, thyroid disorders might be involved in immunosenescence (17), and the maintenance of normal thyroid function could, therefore, contribute to preserving immune responses in the elderly. THs can modulate inflammation, even stimulating adipocytes to produce adipokines acting on several homeostatic aspects of metabolism and energy that influence body weight, thermogenesis, and lipolysis (described in detail in “THs, thermogenesis, and mitochondria” below). From the available data, the link between thyroid function, inflammaging, and immunosenescence is yet unclear, and it is still uncertain whether the hypothyroidism of the elderly represents an anti-inflammatory adaptation.

Cellular senescence, telomere shortening, DNA damage, and the thyroid

The accumulation of senescent cells with age is a hallmark of aging and a pillar of geroscience. Such accumulation occurs in a variety of organs and tissues and represents another major stimulus that fuels inflammaging. Senescent cells are characterized by cell cycle arrest and telomere shortening as a consequence of DNA damage (18), and they develop a distinct secretome profile characterized by a persistent proinflammatory phenotype (19, 20) known as the senescence-associated secretory phenotype, which includes a variety of proinflammatory cytokines as well as growth factors and extracellular matrix–degrading proteins. The “chronic” senescence-associated secretory phenotype induces senescence in adjacent young cells, contributing to the propagation of inflammation and tissue dysfunction to neighboring cells (21, 22). Because of their low but chronic inflammatory phenotype, persistent senescent cells are thought to accelerate aging and the onset of age-related diseases (2). Cellular senescence not only plays a role in aging and contributes to the appearance of AADs (23, 24) but is also an important antiproliferative process that acts as a strong barrier against cellular transformation and cancer progression (25).

As far as we know, most of these main topics, such as the possible age-related accumulation of senescent cells in the thyroid, have not been directly addressed in humans despite their importance for the physiopathology of the gland and taking into account that the age-associated TH changes can play a role in the accumulation of senescent cells in other organs and tissues.

The available data on this topic are summarized as follows:

• Progressive telomere shortening, a well-known cellular senescence biomarker, has been evaluated in different tissues, including the thyroid from individuals of different age (0 to 98 years) (26). Telomere erosion in the thyroid was evident after 50 years of age, likely owing to slow cell turnover rate, at variance with other human organs where an earlier reduction of telomere occurs, suggesting that the rate of telomere shortening is tissue specific (26).

• THs appear to be able to induce senescence in vitro and in vivo. THs can activate metabolism by binding to two receptors, that is, TRα and TRβ, but depending on which receptor is engaged, opposite effects can ensue. T3 induces DNA damage by oxidative stress and drives mouse embryonic fibroblasts to premature senescence, by binding to TRβ and not to TRα and involving the DNA repair ataxia telangiectasia mutated protein (27). Ataxia telangiectasia mutated protein detects genomic damage, activates mitochondria to produce dangerous ROS, and consequently augments the numbers of DNA double-strand breaks, favoring cellular senescence. Similar results were obtained in vivo using genetically altered mice lacking TRβ, which displayed a reduced number of senescent cells in the liver in comparison with wild-type animals (27).

Overall, emerging hits suggest that the thyroid could play a role in the systemic accumulation of senescent cells with age, but the crucial topic of the presence of senescent cells in normal thyroid gland during aging has not been addressed in animal models or in humans. The occurrence of senescent cells in the thyroid can be predicted taking into account that thyroid epithelial cells are constantly exposed to ROS through dual oxidases for the synthesis of THs, thus producing large amounts of H₂O₂, which can induce genomic damage and telomere erosion. To our knowledge, this topic, which could help in clarifying the adaptive or maladaptive role of hypothyroidism in old and very old subjects (see “Thyroid aging in centenarians and their offspring” below) and could pave the way for senolytic trials targeted either to the thyroid or to AADs (28), has not been investigated in humans and represents an unmet need.

THs and stem cell renewal

One of the aging hallmarks and geroscience pillar is the decline in the regenerative potential of organs and tissues thoroughly described in many organs and compartments, including bone marrow, intestine, brain, muscle, and bone, among others (29). There is a vast literature showing that THs have a major role in cell division and differentiation and in the development of the nervous system, intestine, bone, and muscle during organogenesis and development as well as in the regulation of adult stem cell function in the intestine, muscle (30, 31), and brain (32–35). However, the presence of thyroid stem/progenitor cells in the adult organ is controversial, even though a population of cells with stem properties, which are activated upon tissue regeneration after partial thyroidectomy (36, 37), has been described (38). This topic is difficult to address owing to the low turnover of adult thyrocytes, which can be estimated in several years (38).

The available data on this topic are summarized as follows:

• Hypothyroidism reduced proliferation and apoptosis of stem cells in the subventricular zone as well as migration of transgene-tagged neuroblasts out of the stem cell niche, inhibiting the generation of new cells. These effects were mediated by TRα, but not by TRβ (33).

• The regeneration of cardiotoxin-injured skeletal muscle of mice without D2 was markedly delayed in comparison with wild-type mice (35). D2 generates intracellular active THs in muscle and is essential for normal mouse myogenesis and muscle regeneration. Indeed, a D2-mediated increase in FT3 levels is essential for the enhanced transcription of myogenic differentiation 1 (MyoD1) and for the execution of the myogenic program. Therefore, the retardation of regeneration of cardiotoxin-injured skeletal muscle of mice without D2 was associated with a failure of main markers of terminal differentiation (35). Moreover, in the same model, it has been demonstrated that the satellite cells augmented their proliferative capacity in response to attenuation of TH signaling, suggesting that low T3 levels are required in the early phase of muscle regeneration (39). Recently, Ambrosio et al. (30) published a very interesting and clarifying systematic review on the functional role of deiodinases in muscle stem cells and on the ability of THs to affect the composition, contraction force, glucose metabolism, and energy metabolism.

• A variety of experimental data show that the stem cell population requires low levels of THs to maintain its stemness and renewal capacity (40). This consideration is particularly important taking into account the age-related modifications of THs that in turn can affect the adult stem cell renewal in a variety of organs and tissues.

Overall, the presence and function of thyroid stem/progenitor cells in aged thyroid is a rather neglected issue, despite its potential interest for thyroid function in the elderly as well as for thyroid pathologies such as cancer. Within this complex scenario, we can also hypothesize that the complex remodeling of thyroid function with age is adaptive and contributes to the new systemic homeostatic equilibrium involving several organs and tissues in advanced age, as suggested by studies on centenarians discussed below.

Thyroid and the proteostasis network, including the ubiquitin–proteasome system, autophagy, and mitophagy

An accumulation of damaged proteins is a hallmark of the aging process, and proteostasis is one of the geroscience pillars (1). All cells exploit a series of quality control mechanisms to preserve the stability and functionality of their proteomes. Proteostasis involves a set of molecular components and mechanisms devoted to protein clearance (proteostasis network) that prevents the toxicity associated with protein misfolding and accumulation of toxic aggregates in different subcellular compartments and tissues (41). Proteolytic systems such as the ubiquitin–proteasome system, autophagy, and mitophagy (the selective clearance of damaged mitochondria) are the main molecular components of the proteostasis network, and their decreased activity is a central characteristic of aging contributing to the onset of AADs (41). In particular, mitophagy eliminates dysfunctional or damaged mitochondria, thus counteracting degeneration, dampening inflammation, and preventing unwarranted cell loss. Overall, a combination of insufficient autophagy and mitophagy contributes to multiple AADs (41). In this scenario, T3 regulates lipid homeostasis by stimulating the shuttling of free fatty acids into mitochondria (β-oxidation) (42), and this process is coupled with induction of hepatic autophagy (43) and an increase in oxidative phosphorylation–generating ROS that damage mitochondria.

The topic of thyroid proteostasis is critical for thyroid physiology, including age-related changes, as well as for thyroid pathology, but few data focusing on the thyroid are available, and they are summarized as follows:

• In a study of transcriptomics on 322 normal thyroid glands from subjects of different age, the most significant age-related change was the downregulation of genes related to the mitochondrial and proteasomal functions, loss of differentiation, and activation of autoimmune processes (44). Cho et al. (44) demonstrated that the thyroid age–associated gene expression profile was associated with the upregulation of immune activity and overlapped with gene expression patterns in tissues affected by autoimmune thyroiditis (AITD). These data indicate a possible “link” between aging and AITD.

• In another study on age-dependent transcriptomic changes, Yang et al. (45) reported similar results for multiple organs, but not for thyroid, a discrepancy likely due to a different analytic method and the size of the data set.

• A study exploiting in vitro and in vivo hepatic cell models showed that T3 induces ROS production leading to initiation of mitophagy, a major mechanism to remove severely damaged mitochondria during cell stress or excess mitochondria during development and for sustaining efficient oxidative phosphorylation (46).

• In skeletal muscle, THs control cellular growth, regeneration, and differentiation and induce autophagy by producing ROS, activating AMP kinase, and stimulating unc-51–like autophagy activating kinase 1 (ULK1) and autophagosome formation by inhibiting mammalian target of rapamycin (mTOR) (47). The TH-induced autophagy in skeletal muscle is essential for stimulation of mitochondrial biogenesis and activity (48).

Overall, the scarce above-mentioned data suggest that thyroid proteostasis is likely profoundly affected by aging, but such knowledge is still in its infancy, particularly in humans.

THs, metabolism, and aging

THs strongly control key metabolic pathways responsible for energy balance by regulating energy storage and expenditure. THs act on liver, white and brown adipose tissue, skeletal muscle, and pancreas, modulating plasma glucose levels, insulin sensitivity, and carbohydrate metabolism. Thus, understanding how age-associated TH changes affect central and peripheral mechanisms of metabolism in homeotherms is essential.

“…epigenetic modifications may represent another mechanism of the pervasive influence of THs….”

THs, thermogenesis, and mitochondria

THs have a role in the adaptation of the organism to changing environmental conditions, including cold acclimation by modulating metabolic rate, muscle force production, and cardiac performance (40). THs are able to increase metabolic rate and thermogenesis, including maintenance of body temperature, directly modulating the transcription of nuclear and mitochondrial genes. During cold exposure, thyroid function is activated through the stimulation of TRH synthesis, mediated by the catecholaminergic neurons, and THs synergistically interact with the sympathoadrenal system to induce thermogenesis (49–51), whereas the activity of deiodinase is upregulated in brown adipose tissue (52). These TH functions are critical on aging as cold adaptation change in old animals, in association with alteration of thyroid function (53). It is possible to hypothesize that the increased sensitivity of old people to cold could be the price to pay for the possible beneficial effect of mild hypothyroidism on longevity by reducing metabolic rate, ROS generation, and oxidative damage. Mitochondrial metabolism is the best recognized link between THs and longevity (40). THs are able to increase metabolic rate and thermogenesis through multiple mechanisms involving mitochondrial function (the “uncoupling hypothesis”) (54). A major geriatric syndrome is sarcopenia, and importantly note that in addition to its metabolic activity, T3 is considered an important regulator of muscle development as described above.

Many are the unanswered questions related to the thyroid, thermogenesis, and mitochondria, as this topic concerns domains critical for survival and healthy aging, such as the physiological metabolic and thermoregulatory conditions of the elderly, and particularly of the oldest old.

The available data on this topic are summarized as follows:

• Old animals are more susceptible to cold stress than are young ones, and cold-induced TH release occurs independently of TSH (55).

• The expression of uncoupling protein 1 (UCP1) in mitochondria is activated in response to cold by T3-dependent mechanisms driving the heat production by brown adipose tissue independently from shivering or other muscular processes (56, 57). In humans, two variants located in the upstream enhancer region of the UCP1 gene affect gene expression and are correlated with human longevity (58). Brown adipose tissue and mitochondrial uncoupling can be targeted for interventions to prevent and treat obesity and AADs (59).

• T3 increased fatty acid oxidation and mitochondrial respiration as well as autophagic flux, mitophagy, and mitochondrial biogenesis, with no significant induction of intracellular ROS despite high mitochondrial respiration and UCP1 induction by T3 (53). However, when cells were treated with Autophagy-related 5 (Atg5) small interfering RNA to block autophagy, induction of mitochondrial respiration by T3 decreased and was accompanied by ROS overproduction, demonstrating a critical role for autophagic mitochondrial turnover (60) (see “Thyroid and the proteostasis network, including the ubiquitin–proteasome system, autophagy, and mitophagy” above).

• In male Wistar rats receiving T4 in drinking water, it was reported that the induced hyperthyroidism increased the content of mitochondria in liver, changed the structure of mitochondrial membranes, and uncoupled oxidative phosphorylation system with an increase of 50% in the generation of superoxide radicals, resulting in accelerated aging and decrease of lifespan (61). On the contrary, on the same animal model, calorie restriction was accompanied by an increase in lifespan and a reduction of body temperature significantly correlated with a decrease in T3 and T4 levels (61).

• p43, the ligand-binding form of TH receptor (TR)α1, is located in the mitochondria and plays a major role in the crosstalk between nucleus and mitochondria. p43 responds to T3, acting as a transcription factor for nuclear and mitochondrial genome, and in vivo studies on murine models have shown that p43 overexpression in skeletal muscle increases mitochondrial transcription and biogenesis, inducing a stimulation of mitochondrial respiration and a shift in metabolic and contractile features of muscle fibers toward an oxidative phenotype (62). On the same model, p43 overexpression, after an early rise in mitochondrial DNA and mass, induces oxidative stress characterized by a strong increase of lipid peroxidation and protein oxidation in quadriceps muscle, eventually resulting in muscle atrophy, probably through stimulation of the ubiquitin–proteasome pathway. Therefore, prolonged stimulation of mitochondrial activity by p43 contributes to the insurgence of muscle atrophy, stressing the importance of tight control of p43 expression by the mitochondrial pathway regulated by T3 as one of the processes involved in sarcopenia (62).

• Peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α) is actively involved in the modulation of the mitochondrial biogenesis by THs. The expression of PGC-1α is activated endogenously by T3 and coactivates several nuclear transcription factors, including TR. Therefore, also PGC-1α plays a key role in the crosstalk between nuclear and mitochondrial aging pathways. During aging, telomere shortening causes p53-dependent repression of PGC-1α, resulting in a reduction of mitochondrial biogenesis and an impairment of mitochondrial functions (63). These findings strongly support the importance of the strict interactions among PGC-1α, p53, and THs, and their role in the mechanisms by which the hypothalamic-pituitary-thyroid axis affects and regulates metabolic homeostasis during aging (40).

THs affect a complex circuitry critical for adaptation to cold, metabolic rate, and mitochondrial function/biogenesis, but the impact of aging on such molecular and cellular crosstalk is still unclear and deserves further studies in both animal models and in humans. In this regard, interestingly note that lower basal body temperature appears to be associated with healthy aging (64), and long-term calorie restriction lowers core body temperature in humans (65). However, women appear to have a slightly higher body temperature and yet live longer than men (66). Thus, aged thyroids and age-related TH remodeling can have both detrimental and adaptive effects, but additional studies in humans, particularly in the oldest, are warranted. In support of an adaptive hypothesis, it has been reported that mean temperature decreased with age, with a difference of 0.3°F between the oldest and youngest groups after controlling for many confounders. These results are consistent with low body temperature as a biomarker for longevity, but prospective studies are needed to confirm whether a lower body temperature lifelong is a survival advantage and has a potential relationship with age-related changes in thyroid function (67).

THs and the gut microbiota

The gut microbiota (GM) is a complex, highly dynamic, and evolutionarily shaped ecosystem, recognized as an integral and active “organ” contributing to physiological, metabolic, and immune functions. GM composition is affected lifelong by individual lifestyle (e.g., nutrition, physical activity) and environmental variables (68, 69). GM and the immune system establish a constant lifelong interplay, and conditions that alter GM homeostasis and increase the permeability of intestinal epithelial barrier have been associated with the onset of local and systemic inflammatory and autoimmune disorders (70). Another important gut function is the gastrointestinal absorption of iodine and selenium, essential nutrients for the maintenance of thyroid functions. The role of GM in iodine uptake is still poorly studied, whereas the link between GM composition and host availability of selenium, an essential constituent of deiodinase isoforms, is stronger (71).

The available data on this topic are summarized as follows:

• The study of GM of young adults, elderly, and centenarians highlighted that the mutualistic changes in composition and diversity of GM occurring with age are nonlinear, remaining highly similar between young adults and older adults until 70 years of age and markedly changing in centenarians, suggesting that GM undergoes a profound and possibly adaptive remodeling in the last decades of life (72). In subjects of extreme age, the loss of important core components is accompanied by the gain of new microbial subdominant components, including potentially beneficial but also pathobionts and allochthonous bacteria, resulting in an overall increase in GM diversity that has been observed in centenarians from different ethnicities (68, 73–77). Besides such commonality regarding the increased GM diversity, bacterial signatures that are common among centenarians of different ethnicities, such as Italian, Chinese, and Japanese (68), as well Indians (78), have been reported, despite their consistent diversity regarding genetics, nutrition, and many other context-dependent variables.

• Patients with chronic AITD showed alteration of the intestinal mucosal morphology (increased space between adjacent microvilli and augmented thickness of microvilli) and an impaired intestinal permeability (79). Autoimmune overt hypothyroidism is a risk factor for abnormal bacterial overgrowth in the small intestine (80). Hypothyroidism was associated with decreased frequency of rhythmic colonic activity and slower orocecal transit time (81) predisposing to bacterial overgrowth. In these patients, a decontamination therapy with rifamixin improved gastrointestinal symptoms often associated with hypothyroidism (80). In patients affected by small intestinal overgrowth, it was found that the strongest contributor to this condition is levothyroxine use (82). However, evidence for potential bacterial contribution to the onset and progression of AITDs (including Hashimoto thyroiditis and Graves disease) is based on retrospective studies measuring bacterial antibodies (especially toward Yersinia enterocolitica, Helicobacter pylori, and Borrelia burgdorferi) (83). Hyperthyroidism is associated with a decrease of Bifidobacterium and Lactobacillus and an increase of Enterococcus, compared with the control group (84).

• Duodenum and cecum play the major role in the adsorption of dietary selenium depending on its chemical form (85), but common GM components such as Escherichia coli, clostridia, and enterobacteria possess selenoprotein-encoding genes and can compete with the host for selenium uptake (86). The quantity of selenium not absorbed in the small intestine may be actively taken up in the colon where it is metabolized by the resident microbial community, suggesting that a competition that potentially decreases selenium bioaccessibility likely exists (87).

• Studies on rats showed that deiodinases are present and active in diverse tissues, including intestinal mucosa (88), and these enzymatic activities may be inhibited by the resident microflora (89). The human intestinal tract retains a relevant deiodinase activity and, owing to its large surface, gives a noteworthy contribution to the whole-body T3 pool. Additionally, diluted human and rat fecal suspensions were able to hydrolyze significant amounts of iodothyronine conjugates due to the presence of obligate anaerobic bacteria with glucuronidase activity (90, 91). The glucuronidase activity in fecal content indicates the presence of enterohepatic circulation for iodothyronines as suggested by data showing that GM allows the reabsorption of native T3 following the hydrolysis of conjugated forms of the hormone (92). The fraction of reabsorbed T3 escaping from liver extraction may reenter in the general circulation, contributing to the systemic pool of iodothyronines. The observation that plasma reabsorption of radiolabeled T3 is abolished in germ-free animals (92) supports the key role of GM on the thyroid homeostasis through this enterohepatic cycle.

As far as we know, no studies have been performed evaluating directly the correlation between age-related thyroid physiological changes and age-related GM remodeling, despite a variety of studies on the correlation between GM changes and thyroid diseases such as Hashimoto thyroiditis and Graves disease.

An interesting and unexpected novelty is that in centenarians a longevity-specific (common to different ethnicities) GM remodeling and signature and peculiar hypothyroidism (see “Thyroid aging in centenarians and their offspring” below) are concomitantly present, potentially paving the way to develop a GM/thyroid-based biomarker for healthy aging/longevity even if the link between these two phenotypes is unclear at present.

“…centenarians can be considered as the most successfully remodeled people….”

Finally, the role of anaerobic enteric bacteria in humans is not yet supported by direct experiments, even if GM seems to be a further regulator of thyroid homeostasis acting directly through its metabolic enzymes as well as by modulating the chemical bioavailability of iodothyronines for reabsorption in the blood (93). Further studies, particularly in old subjects, on the bidirectional, lifelong crosstalk between host thyroid function and GM are urgently needed.

THs, epigenetic changes, and aging

Epigenetic changes involving histone modifications, noncoding RNAs and DNA methylation have a role in the modulation of aging and AADs (94–97). Although the role of THs to influence these processes remains poorly elucidated, few pieces of evidence exist regarding the capacity of THs to modulate epigenetic profile (98, 99).

The available data on this topic are summarized as follows:

• T3 treatment caused different effects in adult C57BL/6 mice (histone modifications involved in regulating transcription in liver and no significant changes in the DNA methylation) (100) and in postnatal day 6 C57BL/6J mice [increased transcription of de novo DNA methyltransferase 3a (DNMT3a) gene in the brain] (101). T3 increased DNMT3a mRNA expression during metamorphosis also in Xenopus tadpoles (101).

• A recent paper showed lower global DNA methylation levels and DNMT1 expression in T and B lymphocytes of patients with Graves disease compared with age-matched controls, and these parameters were restored by treating the hyperthyroidism (102).

• A mild maternal hypothyroxinemia during pregnancy in mice induced in the offspring damage in learning and memory in adulthood, probably due to a persistent DNA hypermethylation in the promoter region of the brain-derived neurotrophic factor (BDNF) gene in the hippocampus, capable of suppressing BDNF expression and thus promoting cognitive disorders in adult offspring (103). Although it is difficult to establish a cause–effect relationship between these events in the absence of specific human data, several epidemiological studies reported that children born from mothers with hypothyroxinemia showed cognitive and psychomotor deficits (104–107). Such studies focused on intellectual development during school age, and it is unknown whether epigenetic mechanisms related to maternal hypothyroidism during pregnancy predict an age-related change in cognitive ability in the offspring in their adulthood and through old age.

• We have characterized the DNA methylation profiles from peripheral leukocytes of female centenarians, their female offspring, and female offspring of both non–long-lived parents. Importantly, note that centenarians’ offspring have a consistently healthier phenotype than do their age-matched control born from non–long-lived parents (4). Several genes involved in DNA/RNA synthesis, metabolism, and cellular signaling were differently methylated between centenarians’ offspring and controls. More recently, in an epigenome-wide association study [using a 450 BeadChip array capable of assessing the methylation of more than half a million cytosine phosphate guanosines (CpGs)] involving an Italian cohort of semisupercentenarians (≥105 years of age), their offspring, and age-matched controls, we reported that those ≥105 years of age and their offspring are biologically younger than their chronological age. We showed that according to the Horvath “epigenetic clock” (which considers 353 CpGs in the entire genome), centenarians and their offspring are younger than expected based on their chronological age (8.7 years and 5.2 years, respectively) (108). Although the cause–effect relationship is difficult to verify in such a cross-sectional study, we cannot exclude that slower cell growth/metabolism and better control in signal transmission through epigenetic mechanisms may be involved in the process of longevity (109). These aspects reflect previous observations on the potential benefits of mild hypothyroidism in the elderly through its lowering effects on basal metabolic rate and oxidative metabolism, with a consequent reduction of ROS-induced DNA damage (110). Indeed, high levels of TSH (111) and low levels of FT4 (112) are associated with a better survival in elderly subjects. A mild thyroid failure may suppress processes involved in nucleotide biosynthesis and DNA replication (113, 114), suggesting that a specific TH-related epigenetic modulation of genes involved in DNA metabolism and control of signal transmission may contribute to the longer lifespan and healthy aging of centenarians’ offspring. Additionally, epigenetic mechanisms may also influence the thyroid landscape during aging.

• The expression of TRβ in peripheral blood mononuclear cells obtained from healthy elderly and long-lived individuals was significantly lower than in young individuals, and it was likely related to the increased methylation of the CpG island located within the TRβ promoter (115).

• Maternal exposure to iodine excess of Wistar rats induces hypothyroidism in their adult male offspring, morphological alterations in thyroid follicles, increased thyroid oxidative stress, and decreased expression of thyroid differentiation markers and transcription factors (116). Increased DNA methylation and DNA methyltransferases expression, hypermethylation of histone H3, hypoacetylation of histones H3 and H4, increased expression/activity of histone deacetylases, and decreased expression/activity of histone acetyltransferases are involved in the repression of thyroid gene expression (116). Overall, these epigenetic changes appear to be a kind of adaptive phenomenon to protect the offspring’s thyroid from the deleterious effects of iodine excess.

• Finally, much evidence suggests a main role of the epigenetic network in the control of the expression of D1 in chickens (117), D2 in chickens (118) and rats (119), and D3 in mice (120, 121), neonatal goats (122), and Siberian hamsters (123). Within this scenario and in the absence of specific data in humans, we cannot exclude that a modulation of the local levels of THs and metabolites occurs in aging through epigenetic mechanisms.

Overall, available data suggest that epigenetic modifications may represent another mechanism of the pervasive influence of THs on the aging process, but many mechanistic details are lacking.

Thyroid Aging and the Lesson From Centenarians

Thyroid aging in the oldest old

Evaluation of the TRH–TSH–T4/T3 axis during aging presents many problems due to the concomitant presence of several confounding variables and difficulties in the correct definition of “healthy elderly” (124), and results of studies examining the influence of age on the hypothalamic–pituitary–thyroid axis also in the absence of thyroid disease remain controversial, as reported in Table 1 (125–142).

Table 1.

Summary of Results Obtained in a Number of Studies Evaluating the TRH–TSH–T4/T3 Axis During Aging in Thyroid Disease–Free Populations (n > 300)

Summary of Results	Country (Ethnicity)	Participants	Reference	Year
TSH and the prevalence of antithyroid antibodies are greater in females, increase with age, and are greater in whites and Mexican Americans than in blacks	United States (white 35%, black 32%, Mexican American 29%, other 4%)	Thyroid disease–free population of 13,344 people (≥12 y of age)a	(125)	2002
Serum TSH gradually decreases with age, whereas after age 60, serum FT4 increases, possibly because of the development of thyroid autonomy after longstanding borderline sufficient iodine intake	Netherlands (96% white)	Population of 5167 individuals (≥18 y of age) selected by excluding those at risk for thyroid diseasea	(126)	2006
TSH distribution progressively shifts toward higher concentrations with age	United States (white, black, Mexican American, other)	Thyroid disease–free population of 14,376 people (≥12 y of age)a	(127)	2007
FT3 and TSH decreased with age; the TSH response to TRH was blunted in older subjects, especially in male individuals	Germany	387 thyroid disease–free population (13–100 y of age; mean age, 39.5 y)a	(128)	2008
Age was not associated with serum TSH levels	China (Asian)	2237 individuals of reference population (>13 y of age)a	(129)	2008
A shift to higher TSH with aging occurred in black and white subgroups	United States (black or African American 33%, white 15%, Hispanic 5%, unknown 47%)	Population of 22,116 people (>10 y of age; median age, 44 y) without clinical evidence of thyroid diseasea	(130)	2009
The TSH 2.5th, 50th, and 97.5th percentiles increased with age, with the most significant effects seen at the 97.5th percentile, which increased by 0.3 mIU/L with each 10-y increase in a subject’s age	United States (white 35%, black 31%, Mexican American 29%, unknown 5%)	Thyroid disease–free population of 13,344 people (>12 y of age)a	(131)	2011
A statistically significant increase in TSH (+12%) and FT4 (+2.5%), and a decrease in T3 (−13%) during the 13-y period	United States	Thyroid disease–free population of 533 participants (≥75 y of age)b	(132)	2012
Aging was associated with increased serum TSH concentrations (+21%), with no change in FT4 during 13 y of follow-up, suggesting an age-related alteration in TSH set point or reduced TSH bioactivity rather than occult thyroid disease	Australia (predominantly white)	Thyroid disease–free population of 908 participants (mean age at baseline 45.5 y)b	(133)	2012
An increasing age is associated with an increase in the median and 97.5th percentile for TSH	Australia (predominantly white)	Thyroid disease–free population of 148,938 people (1–106 y of age; mean age, 48.2 y in women and 53.8 y in men)a	(134)	2012
During the 11-y follow-up, mean TSH increased significantly (+8.7%), particularly in the area with the highest iodine intake	Denmark	2203 participants (18–65 y of age) with no previous thyroid diseaseb	(135)	2012
An increase in median and 97.5th percentile TSH with increasing age	United Kingdom	Thyroid disease–free population of 153,127 people (≥18 y of age)a	(136)	2013
TSH and FT3 were inversely associated with age	Germany	Thyroid disease–free population of 1002 individuals (45–83 y of age)a	(137)	2014
TSH (+8.9%) and FT4 (+9.3%) values increased during the 11-y period, particularly from the age of 50 y; no significant change was observed for FT3	Spain	Thyroid disease-free population of 552 participants (18–65 y of age; mean age at baseline, 41.7 y)b	(138)	2015
FT4 increased and TSH decreased at follow-up evaluation	Korea	A total of 313 euthyroid participants (5-y follow-up evaluation)b	(139)	2015
Until age 40, for each increase in TSH quartile, FT3 and the FT3/FT4 ratio increased and FT4 decreased significantly; in older age groups, increasing TSH was not associated with increased FT3/FT4 ratio; this could reflect a decrease in deiodinase activity and/or the development of TSH resistance with aging	Israel	Thyroid disease–free population of 27,940 people (1–80 y of age)a	(140)	2016
TSH levels did not change over time, irrespective of age; FT4 levels increased over time, most prominently in those >65 y of age	Netherlands (predominantly white)	9402 participants (≥45 y of age; mean age, 65.1 y) from the Rotterdam Study not taking thyroid medication (longitudinal for 1225 people with a follow-up of 6.5 y)b	(141)	2016
FT3 and FT4 decreased throughout life, whereas TSH declined until age 50 y and then increased slightly; FT4 declined, among females more than among males until middle age; after 60 y of age, FT4 levels mildly increased only in females	Israel	Thyroid disease–free population of 27,940 people (≥1 y of age)a	(142)	2017

A review of the literature was conducted using the PubMed database and “thyroid” and “aging” as key words. The search included articles published in English between January 2000 and February 2019.

^a

Cross-sectional study.

^b

Longitudinal study.

As previously argued, the oldest old will represent a considerable percentage of the elderly as they are the segment of the elderly population that is increasing fastest, suggesting that particular attention should be devoted to the thyroid status of this segment of the population as a rational prerequisite for any type of possible intervention. The problem is that the oldest old are very heterogeneous regarding their overall phenotype and health status, a result of possible successful or unsuccessful adaptation, according to the remodeling theory of aging (143). For this reason, the oldest old necessitate a careful and specific study, and available data on their thyroid status will be reviewed and critically evaluated. Subjects >85 years of age have been only rarely included in most of the studies exploring the effect of thyroid function on mortality, and available results are conflicting (40), as reported in Table 2 (112, 144–159), focused on the association between thyroid function and all-cause mortality in euthyroid individuals.

Table 2.

Summary of the Results Obtained in Different Studies Evaluating the Association Between Thyroid Function and All-Cause Mortality in Euthyroid Individuals (n > 300)

TH Changes Significantly Associated With Increased All-Cause Mortality	Main Measures of Thyroid Function	Follow-Up Period (y)	Country	Participants	Reference	Year
↑FT4	TSH, FT4, TT4, T3, rT3, and T4-binding globulin	4	Netherlands	403 male participants (73 to 94 y of age) of the Zoetermeer Study	(112)	2005
No association	FT4, FT3, and TSH	8.5	Germany	3651 individuals of the Study of Health in Pomerania (20–79 y of age)	(144)	2010
No association	TSH and FT4	8.3	United States	1387 euthyroid men of the Osteoporotic Fractures in Men (MrOS) study (mean age, 73.6 y)	(145)	2012
↓TSH	TSH	4.5	Israel	42,149 subjects (≥40 y of age)	(146)	2012
↑FT4	TSH and FT4	6.4	Australian	3885 euthyroid men (≥65 y of age)	(147)	2013
↓FT4	FT4, FT3, and TSH	4.3	South Korea	212,456 middle-aged (40.2 y of age) euthyroid participants of the Kangbuk Samsung Health Study	(148)	2014
↑FT4, ↑TSH	TSH, FT4, and peroxidase antibodies	9.4	Netherlands	493 participants (≥80 y of age) of the Nijmegen Biomedical Study	(149)	2014
↓TSH, ↑FT4	TSH, T3, and FT4	>17	United States	2843 participants (74.5 ± 5.1 y of age)	(150)	2015
↓TSH	TSH, FT3, and FT4	9	Italy	815 euthyroid participants of Aging in the Chianti Area (InCHIANTI Study)	(151)	2016
↓↑TSH (U-shaped association)	TSH and FT4	19.1	United States	12,584 adults ≥20 y of age	(152)	2016
↑rT3	rT3, FT3, FT4, and TSH	9	United Kingdom	645 participants (85 y of age) of the Newcastle 85-Plus Study	(153)	2016
↓FT3/FT4 ratio, ↓FT3, ↑FT4	TSH, FT3, and FT4	5 and 3.8	Netherlands	805 nonagenarians from the Leiden Longevity Study (median age, 91 y) and 259 nonagenarians from the Leiden 85-Plus Study (median age, 94 y)	(154)	2017
↓TSH and ↑FT4 in men; ↓FT3 in women	TSH, FT3, and FT4	10	Italy	933 participants (324 men and 609 women) of Milan Geriatrics 75-Plus Cohort Study with normal TSH (81.6 ± 4.6 y of age)	(155)	2017
↓FT3/FT4 ratio	FT3 and FT4	2.5	Italy	643 geriatric patients (83.8 ± 7.4 y of age)	(156)	2018
↑TSH in women	TSH, FT3, and FT4	7.7	Turkey	614 hospitalized patients (40–79 y of age)	(157)	2018
No association	TSH, FT3, and FT4	13	Netherlands	2431 participants of the PREVEND cohort, 28–75 y of age	(158)	2017
↓FT3/FT4 ratio	TSH, FT3, and FT4	1	China	953 euthyroid patients with acute myocardial infarction	(159)	2018

A review of the literature was conducted using the PubMed database and “thyroid,” “mortality,” and “euthyroid” as key words. The search included articles published in English between January 2000 and February 2019.

Arrows indicate increases (↑) or decreases (↓) of TSH and/or THs, but always within the euthyroid range.

Abbreviation: TT4, total T4.

Centenarians and their offspring as a model of longevity in humans

Studies on centenarians could help to better understand the role of THs in healthy aging and longevity as they reach the extreme limits of life and have escaped or delayed the onset of major AADs (160). Therefore, centenarians can be considered as the most successfully remodeled people, presenting a complex and adaptive phenotype but also characterized by a precarious homeostatic balance (5). Centenarians are even more interesting considering that they are a sort of gold standard of Homo sapiens, that is, people who exploited the maximum living capacity of the species. Indeed, despite that the number of centenarians is rapidly increasing worldwide, no one has been reported to live longer than ∼120 years, suggesting that the lifespan of H. sapiens has a biological limit that cannot be overcome unless within the uncanny nightmare of changing its basic genetics/biology (161). Table 3 (162–179) highlights the main features of centenarians taking into account the seven pillars of geroscience. This conceptual framework applies also to possible intervention regarding alterations in the thyroid functioning in the oldest old, as illustrated in “Summary and Perspectives” below.

Table 3.

The Seven Pillars of Aging in Centenarians

Metabolism	Preserved glucose tolerance and insulin sensitivity and lower levels of serum IGF-1 in centenarians with respect to elderly controls	(162–165)
Inflammation	The increased plasma levels of inflammatory molecules such as IL-6, IL-18, IL-15, CRP, serum amyloid A, fibrinogen, von Willebrand factor, resistin, and leukotrienes are counterbalanced by a concomitant large quantity of anti-inflammatory molecules (i.e., adiponectin, TGF-β1, IL-1 receptor antagonist, cortisol, anti-inflammatory arachidonic acid compounds)	(166–173)
Epigenetics	According to the “epigenetic clock,” centenarians are younger (8.6 y) than expected based on their chronological age	(174)
Adaptation to stress	Higher plasma levels of cortisol, ACTH, and CRH than in young subjects	(170)
Stem cells and regeneration	The basal hematopoietic potential (capability of CD34+ cells to respond to hemopoietic cytokines and to form erythroid, granulocyte, macrophage, and mixed colonies) is well preserved in healthy centenarians	(175)
Proteostasis	Cultures of fibroblasts derived from healthy centenarians have a functional proteasome	(176)
Macromolecular damage	Lymphocyte cell lines from centenarians preserve their capability of priming the mechanism of repair after H2O2 oxidative damage and in poly(ADP-ribosyl)ation capacity	(177–179)
	Differences in BRCA1 genotype frequencies between the centenarians and controls
	No difference in the number of spontaneous chromatid breaks in lymphocytes from healthy centenarians and controls, but centenarians’ cells show a higher sensitivity (DNA breaks per cell) to the radiomimetic agent bleomycin

A correlated model of healthy aging is represented by centenarians’ offspring who can overcome some limitations inherent in the study of centenarians (rarity, lack of an age-matched control group, and presence of frailty related to their extreme age) (165, 180). Centenarians’ offspring can be compared with age-matched controls born from non–long-living parents (180), and this comparison showed that they are healthier (4, 180), biologically younger (174), and have a higher probability to become long-lived than do members of the same demographic cohorts (180, 181).

Thyroid aging in centenarians and their offspring

Aging is associated with a decreased volume of the thyroid gland and decreased levels of THs (182). As suggested by the remodeling theory of aging (143), such a situation may represent an adaptive phenomenon to attain “successful” aging and to prevent excessive catabolism in the elderly through a reduction in basal metabolic rate and, consequently, in the production of ROS and DNA damage. Despite conflicting results in the literature, most studies reported that higher TSH and/or lower FT4 concentrations within the euthyroid range are associated with lower mortality in old subjects (Table 2). Studies on the relationship between thyroid function and longevity also produced conflicting results [Table 4 (183–191)].

Table 4.

Summary of Results Obtained in Different Studies Evaluating Thyroid Function in Centenarians

Summary of Results	Main Outcome Measures	Population	Participants	Reference	Year
The prevalence of thyroid autoantibodies increased with age until ninth decade of life. The prevalence of thyroid autoantibodies in centenarians was not significantly different from that in controls <50 y of age	Serum thyroid autoantibodies	Italian	34 healthy centenarians (100–108 y of age), 549 control subjects (7–85 y of age)	(183)	1992
FT3 and TSH decreased with age. FT4 did not change with age. rT3 was significantly higher in centenarians than in elderly and adult subjects. The prevalence of serum anti-Tg and anti-TPO antibodies was low and did not differ among centenarians, elderly, and adult subjects	Serum anti-Tg and anti-TPO antibodies, FT4, FT3, rT3, and TSH	Italian	Healthy centenarians (100–110 y of age), 33 healthy elderly subjects (65–80 y of age), 98 healthy adults (20–64 y of age), and 52 patients with miscellaneous nonthyroidal illness (28–82 y of age)	(184)	1993
All parameters were within the normal range, with the exception of TT4 values, which were reduced in 60% of centenarians	Total T3, total T4, FT3, FT4, TSH, anti-Tg, and antimicrosomal antibodies	Italian	20 healthy centenarians (100–108 y of age), 40 healthy elderly subjects (70–84 y of age), and 50 healthy adults (38–62 y of age)	(185)	1997
TSH decreased significantly, whereas rT3 slightly but significantly increased with age. The FT3/FT4 ratio decreased with age, suggesting a decline of the 5′-deiodinase activity. The incidence of thyroid autoantibodies was lower in centenarians than in elderly subjects	TSH, FT3, FT4, rT3, anti-Tg, anti-TPO antibodies, and nutritional markers	Italian	24 healthy centenarian women (100–106 y of age), 24 healthy elderly women (71–93 y of age), and 20 healthy young subjects (22–33 y of age)	(186)	2002
TSH did not differ significantly among centenarians, elderly, and young women. T3 was significantly lower in centenarian women than in elderly and young women	T3, T4, glucose and lipid profiles, plasma leptin, neuropeptide Y, insulin, TSH, GH, prolactin, LH, FSH, and cortisol	Polish	78 centenarian women (100–115 y of age), 21 early elderly women (64–67 y of age), 21 postmenopausal women (50–60 y of age), and 35 younger women (20–50 y of age)	(187)	2007
TSH was significantly lower in centenarians than in healthy elderly and young controls. The FT3/FT4 ratio was significantly lower in elderly subjects and centenarians when compared with young controls. rT3 was higher in centenarians compared with both elderly and young controls	Serum cortisol, dehydroepiandrosterone-sulfate, FT3, FT4, rT3 and TSH, urinary free cortisol, and 6-hydroxymelatonin sulfate	Italian	59 centenarians (100–107 y of age), 24 healthy elderly subjects (mean age, 85 y), and 20 young controls (mean age, 28 y)	(188)	2008
TSH was significantly higher in centenarians than in controls. The TSH frequency distribution curve of centenarians shifted significantly to higher TSH values compared with controls. FT4 was similar in centenarians and controls, and there was a significant inverse correlation between FT4 and TSH in both groups	TSH, FT4, and TSH frequency distribution curves	North American (Ashkenazi Jewish and US NHANES)	Ashkenazi Jew centenarians (median age, 98 y), Ashkenazi controls (median age, 72 y), healthy NHANES controls (median age, 68 y)	(189)	2009
TSH increased with age. T3, FT3, and the FT3/FT4 ratio decreased with age. T4 and FT4 did not change with age. A significant association was found between TSH and FT3 levels of centenarians and those of their offspring, suggesting that TSH and FT3 concentrations may be considered a heritable phenotype.	TSH, T3, FT3, T4, and FT4	Chinese	61 centenarians (mean age, 103 y), 63 centenarians’ offspring (mean age, 62 y), 47 spouses of the offspring (mean age, 60 y), 25 centenarians’ second-generation offspring (mean age, 32 y), and 10 spouses of second-generation offspring (mean age, 31 y)	(190)	2015
FT3 level and the FT3/FT4 ratio decrease with age whereas FT4 and TSH increase with age. In CENT/105+, higher FT4 levels and a lower FT3/FT4 ratio are associated with an impaired functional status and increased mortality. Cluster analysis identified three clusters of CENT/105+ based on their FT3, FT4, and TSH levels. Cluster 3, characterized by lower FT3 and TSH and higher FT4, shows the worst health status and the shortest survival. A group of CENT/105+ showed a thyroid profile suggestive of nonthyroidal illness syndrome and are characterized by a worse functional and cognitive status and an increased mortality with respect to CENT/105+ without non–thyroidal illness syndrome.	TSH, FT3, and FT4	Italian	672 well-characterized Italian subjects (age range, 52–113 y), including 144 centenarians (mean age, 100 y), 70 semisupercentenarians (mean age, 105.9 y), as well as 308 centenarians’ offspring (mean age, 71 y) and 150 age-matched elderly (mean age, 70 y)	(191)	2018

A review of the literature was conducted using PubMed database and “thyroid” and “centenarians” as key words. The search included articles published in English between January 1990 and February 2019.

Abbreviations: 105+, ≥105 y of age (semisupercentenarians); anti-Tg, anti-thyroglobulin; anti-TPO, anti–thyroid peroxidase; CENT, centenarians; NHANES, National Health and Nutrition Examination Survey 1998–2002; TT4, total T4.

Mariotti et al. (184) reported that healthy centenarians had lower serum TSH and FT3 levels and higher serum reverse T3 (rT3) levels compared with those observed in other age control groups. In this centenarian population, the prevalence of thyroid autoantibodies was not significantly different from that observed in controls <50 years of age, notwithstanding the age-related increase in the prevalence of thyroid autoantibodies observed with aging (183). These data were also confirmed by Magri and colleagues (186, 188) who found in centenarians lower TSH levels, higher rT3 levels, and lower thyroid autoantibody positivity as compared with 70- to 80-year-old subjects. In another Italian population of centenarians, total T4 values were lower than the normal range in 60% of examined subjects (185). In Polish centenarians, Baranowska et al. (187) found that serum TSH and T4 concentrations were comparable with those observed in younger women, whereas serum T3 levels were lower compared with the other groups. Atzmon et al. (189) demonstrated that Ashkenazi Jewish centenarians have significantly higher median serum TSH concentrations compared with younger Ashkenazi controls and with a population of thyroid disease–free individuals. An inverse correlation between FT4 and TSH levels in centenarians and Ashkenazi controls has been observed (189), and this phenotype appears to be heritable (192). Also in Chinese centenarians’ families, a decline in thyroid function (high TSH and low FT3 concentrations) appears to be associated with age, and this phenotype is heritable and likely contributes to longevity (190). In relatives of Italian centenarians (offspring or nieces and nephews), lower comorbidities and FT3, FT4, and TSH levels have been reported compared with age-matched controls (193). Lower plasma levels of FT4 in centenarians’ offspring compared with age-matched controls were confirmed in another Italian population (180). Rozing et al. (194) in the Leiden Longevity Study, showed that when compared with their partners, the group of offspring of nonagenarian siblings showed a trend toward higher serum levels of TSH together with lower FT3 and FT4 levels. Lower mortality in the parents of nonagenarian siblings was associated with higher serum TSH levels and lower serum FT3 and FT4 levels in the nonagenarian siblings (195).

We have recently characterized a thyroid function profile in an Italian cohort of 672 subjects consisting of centenarians, semisupercentenarians (i.e., persons who are ≥105 years of age), centenarians’ offspring, and elderly subjects age-matched with centenarians’ offspring. We have found an age-dependent decrease in FT3 level and FT3/FT4 ratio, whereas FT4 and TSH levels increased. In long-lived individuals, a higher FT4 level and lower FT3/FT4 ratio were associated with an impaired functional status and increased mortality. From this analysis, we excluded subjects with a thyroid profile suggestive of nonthyroidal illness syndrome. These results indicated that the age-related decrease in the FT3/FT4 ratio could be due to a decline in 5′-deiodinase activity. Centenarians and semisupercentenarians with a relatively high FT3/FT4 ratio are probably able to preserve D1 activity, likely maintaining a good hormonal negative feedback (191). This phenomenon could be relevant for preserving a good functional capability and survival optimization. During aging, a decline in serum T3 levels could be balanced by a compensatory increase in D1 activity. This adaptive ability, aimed at maintaining an adequate local production of T3, could allow preservation of TH signaling and counteraction of the aging-associated metabolic disturbances. Such interpretation is also supported by a recent prospective study showing that the FT3/FT4 ratio represents an independent marker of frailty and survival in a population of euthyroid older patients hospitalized for an acute event (156).

In conclusion, the mild and progressive decrease in thyroid function observed with aging could be part of adaptive strategies also involving the endocrine system (182) that the body uses to survive in the last decades of life and that likely contribute to attaining the extreme limit of human life, having largely avoided or postponed most AADs. In particular, better preservation of local T3 concentration through a suitable peripheral T4 to T3 conversion may have a relevant role in assuring a remarkable longevity and healthy aging.

“Thyroid Biography” and Thyroid Aging Within a Lifelong Perspective

Aging and longevity are complex traits, where each individual follows a different personal trajectory. The result is the high phenotypic heterogeneity that characterizes elderly subjects and that increases with age (196), mirrored by the large individual variability in the levels of serum TSH, T4, and T3, which in turn explain the difficulty encountered in the diagnosis of thyroid dysfunctions in the elderly and in the oldest old. This heterogeneity is a complex biomedical and public health problem, as all the symptoms that characterize thyroid diseases diminish working capacity and the quality of life. Moreover, thyroid pathologies in the elderly have systemic effects and can cause hypertension, cardiac insufficiency, an adverse lipid profile, insulin resistance, and endothelial dysfunction, among others, in turn posing an increased risk for atherosclerosis, cardiovascular disease, type 2 diabetes mellitus, cognitive impairment, depression, and mortality (197).

Thus, the crucial question becomes where does this heterogeneity come from. Starting from the period of life spent in utero, each individual is exposed to a unique combination of stimuli, including hormonal stimuli, that can affect all organs of the body, including the thyroid. We recently proposed the concept of “immunobiography” (2, 196) to grasp the large heterogeneity of immune system aging and responsiveness (immunosenescence and inflammaging) in different individuals (196). Similarly, we propose here to adopt the concept of “thyroid biography” to better understand thyroid aging at the individual level (Fig. 2). The basic idea is to systematically and accurately collect and store data capable of reconstructing in each individual or patient the unique, lifelong combination of variables such as age, sex, place and geography of birth and of living, type of work, socioeconomic and psychological status, lifestyle habits (nutrition and physical activity), diseases (with particular attention to the endocrinological ones, including those of the parents), comorbidities, results of blood examination performed lifelong (including endocrinological data), and drugs used, among others, that can have long- and/or short-term effects on thyroid function and physiopathology. In particular, environmental factors, including the exposure to endocrine-disrupting compounds as well as smoking and other conditions, could influence thyroid homeostasis lifelong (198).

Figure 2.

The new concept of “thyroid biography” is proposed to better understand the heterogeneity of thyroid aging in each individual or patient as a consequence of the unique combination of variables impinging lifelong on thyroid function.

Summary and Perspectives

The main message of this review is that to fully understand the aging of the thyroid in humans, we have to follow the suggestions that have emerged in the field of aging research, and particularly those conceptualized and proposed by geroscience, which are summarized as follows:

• There is an urgent need to investigate in depth at the thyroid level the main molecular and cellular mechanisms identified as key to understanding the aging process in animal models and in humans. As far as we know, this approach, which has started to be systematically applied in many organs and systems in experimental animals and humans, represents a novelty in the studies on thyroid aging. Available data on these mechanisms or pillars have been scrutinized, and the emerging scenario is that the knowledge on these critical points is still scarce, suggesting that data on fundamental mechanisms such as the accumulation of senescent cells in the thyroid, the relationship between GM dysfunction and thyroid function, the role of inflammaging and its propagation within the thyroid and, systemically, thyroid cell DNA methylation are still largely unexplored.

• Derangements in these basic mechanisms of aging can help in explaining the pathogenesis of age-related thyroid pathologies as well to better understand the role of thyroid aging on the aging of other organs and systems. In this regard, a rather neglected but very interesting topic that deserves much more attention both from a basic and clinical perspective is the contribution of thyroid function abnormalities to the onset of chronic AADs.

• Thyroid status and function in the oldest old are particularly complex and heterogeneous, and this topic also needs further investigations.

• A variety of environmental factors and lifestyle habits have the capability to deeply interfere with thyroid function. The knowledge on this point of public health importance, particularly for the next generations, is still scarce and an effort for improved understanding is urgently needed.

• The complex history behind thyroid status in the elderly, as well as their physiological and clinical heterogeneity regarding thyroidal function, could be better understood by adopting the comprehensive concept of thyroid biography and its inherent capability to grasp the combination of factors impinging lifelong on the thyroid at the individual level (Fig. 2). We envisage the difficulties in the present scenario of medical services and organizations to realize, starting from birth, such a lifelong collection of data on thyroid function, thus implementing a sort of “thyroid passport” for each citizen. At the same time, we consider our proposal a suggestion that we hope is useful for colleagues working in the public health sector to start building a personalized medicine as a prerequisite for personalized aging.

Finally, besides such clarifications, we have the impression that a change of paradigm is emerging regarding the thyroid age-related changes occurring physiologically, that is, in the absence of overt clinically relevant pathologies. These changes, as well as many others concomitantly occurring in the immune system, such as immunosenescence and inflammaging (5), are no longer considered simply detrimental but can be conceptualized as part of the systemic, adaptive remodeling that helps humans survive in relatively good shape until the limit of human life. This new perspective could help practitioners in making decisions regarding TH treatment of the elderly and oldest old, which is at present a controversial issue.

Acknowledgments

Financial Support: This work was supported by Fondazione Cassa di Risparmio delle Province Lombarde (CARIPLO) Contract 2015-0564, the European Union (EU) Horizon 2020 Project PROPAG-AGEING (Grant 634821), the EU JPND ADAGE project, the EU HUMAN project (Grant 602757), and the Ministry of Education and Science of the Russian Federation Agreement No. 075-15-2019-871 (all to C.F.).

Disclosure Summary: The authors have nothing to disclose.

Abbreviations:

Abbreviations:
 
• AAD

  age-associated disease

 
• AITD

  autoimmune thyroiditis

 
• CRP

  C-reactive protein

 
• D2

  D2 deiodinase

 
• FT3

  free T3

 
• FT4

  free T4

 
• GM

  gut microbiota

 
• PGC-1α

  peroxisome proliferator–activated receptor-γ coactivator-1α

 
• ROS

  reactive oxygen species

 
• rT3

  reverse T3

 
• TH

  thyroid hormone

 
• TR

  thyroid hormone receptor

 
• UCP1

  uncoupling protein 1

Author notes

(*C.F. and R.O. contributed equally to this study.)

(‡D.M. and G.V. are co–senior authors.)