aging of the human adrenal cortex - sage kesageke.sciencemag.org/cgi/reprint/2004/35/re6.pdf ·...

http://sageke.sciencemag.org/cgi/content/full/2004/35/re6

The most striking age-related change in the humanadrenal cortex is the decline in secretion of dehy-droepiandrosterone and its sulfate, steroids synthesizedby the inner zone of the cortex, the zona reticularis. Be-cause these steroids are of essentially unknown function,the importance of this age-related change is the subject ofconsiderable debate. It is likely that the age-related changein these steroids results from loss of zona reticularis cellsor impairment of their function. During aging, cumulativedamage to the zona reticularis could occur through is-chemia-related infarcts and other causes of cell death. Cel-lular senescence could contribute to a loss of the ability ofthe tissue to replace lost cells. In contrast, feedback mech-anisms that regulate adrenocortical growth cause com-pensatory local tissue hyperplasias called nodules. The ef-fect of imperfect repair of damage combined with compen-satory overgrowth in the form of nodules leads to an in-creasingly abnormal tissue architecture.

Introduction: Structure and Function of the AdrenalCortexThe adrenal glands are paired organs that overlie the cranial polesof the kidneys. Embryologically, the glands arise from two dis-tinct precursor cell types: mesodermal cells that become the outerpart of the gland, the cortex, and ectodermal cells that become theinner part of the gland, the medulla. The cortex secretes steroids,whereas the medulla secretes catecholamines. This Review con-cerns only the cortex and its changes during aging.

The adult human adrenal cortex, as in other mammals, hasthree morphologically defined zones. In most species, thesezones are an outer zona glomerulosa (ZG), a middle zona fasci-culata (ZF), and an inner zona reticularis (ZR) (Fig. 1). Thesetraditional divisions have been recognized in mammals general-ly since the middle of the 19th century. In all mammals, the ZGsynthesizes mineralocorticoids (corticosteroids such as aldos-terone that primarily function in the regulation of electrolytebalance in the kidney) and the ZF synthesizes glucocorticoids(corticosteroids such as cortisol that function in carbohydratemetabolism and have a variety of other roles). However, it islikely that the ZR has a unique function in humans and a fewother primates (1, 2). The morphologically recognizable ZR inother species does not functionally resemble the zone in the hu-man cortex. In humans, this zone secretes the unusual steroiddehydroepiandrosterone and its sulfate [DHEA and DHEAS,referred to here collectively as DHEA(S)]. The function of these

steroids is essentially unknown, despite much speculation. Adrenocortical cells divide at a slow rate throughout the life

span. The generally accepted model for cell renewal is that cellsin the outer part of the adrenal cortex, the ZG and the outer ZF,divide at a higher rate than those located in the more central ar-eas of the cortex, the inner ZF and the ZR. Cells are therebypushed by the pressure of cell division from the outer part of thecortex to the inner part. A ZG cell may therefore find itselfpushed down into the ZF, and it is assumed that, once past theboundary between the zones, cells that were originally ZG cellsnow become ZF cells. Similarly, a ZF cell that is pushed intothe ZR is assumed to become a ZR cell. There is evidence forthese transformations from both cell culture and in vivo experi-ments (3). The inner ZR appears to be a site of a higher rate ofcell death than the rest of the cortex (3). Thus, over time there iscell turnover without a net change in tissue size.

The question of whether there is a stem cell compartment inthe adrenal cortex has not yet received a definitive answer. Typi-cally, stem cells are defined by their ability to repopulate an organ and generate various differentiated cells. By this defini-tion, many or most adrenocortical cells might be described asstem cells; the adrenal cortex can regenerate from very smallnumbers of cells, after most of the organ has been destroyed (3),and cell transplantation experiments have shown that ~25% of clones of bovine adrenocortical cells can regenerate function-al tissue (4). However, genes involved in Wnt signaling(http://stke.sciencemag.org/cgi/cm/stkecm%3BCMP_5533)—WNT4 and DKK3—are expressed at higher levels in the ZGthan in the rest of the cortex (5, 6). By analogy with stem cellsystems in other organs, the higher levels of Wnt signalingmolecules in the ZG might suggest that it is a stem-cell zone(7), but more studies are needed to decide these issues.

Age-Related Changes in the Structure and Functionof the Human Adrenal CortexChanges in zonationAs shown in Fig. 2, the major age-related change in the humanadrenal cortex is a striking decrease in the biosynthesis ofDHEA(S) (8-11). In contrast, cortisol biosynthesis is relativelyunaffected by aging, but there may be a subtle hypercortisolismassociated with nodules and adenomas of the adrenal cortex.The simplest hypothesis for the decline in DHEA(S) is that itresults from a decline in the number of functional ZR cells, butthis has not been firmly established. Direct measurements ofthe width of the zones show a decrease in the size of the ZRduring aging (12, 13), but this is a relatively modest decreaseand does not account quantitatively for the decline in DHEA(S).An older study emphasized that the boundary between the ZFand the ZR becomes increasingly irregular with age (14). Theworking hypothesis that I adopt here is that measurements of

Aging of the Human Adrenal Cortex

Peter J. Hornsby

(Published 1 September 2004)

(Corrected 22 December 2004)

R E V I E W

SAGE KESCIENCE OF AGING KNOWLEDGE ENVIRONMENT

The author is in the Department of Physiology and Sam and AnnBarshop Center for Longevity and Aging Studies, University ofTexas Health Science Center, San Antonio, TX 78245, USA. E-mail:[email protected]


ZR-

width underestimate the decline in the number of functional ZRcells and that the increasing irregularity of the boundary be-tween the ZF and the ZR makes it difficult to use morphologi-cal criteria alone to estimate the number of ZR cells. A more ac-curate count of ZR cells at different ages requires the use ofmolecular markers to enable recognition of ZR cells. Suchmarkers include the lack of expression of type II 3β-hydroxys-teroid dehydrogenase and higher levels of expression of DHEAsulfotransferase, cytochrome b5, and major histocompatibility

complex (MHC) class II molecules (15-17). Direct proof thatthe decline in DHEA(S) abundance results from a decrease inthe number of ZR cells will be difficult to obtain, because it de-pends on observations on postmortem material with little op-portunity to study functional/morphological correlations.

Neoplastic and preneoplastic changesApart from disorders of zonation, disruption of the normal ar-chitecture of the adrenal cortex also occurs because of an age-

R E V I E W


Medulla

Zona reticularis[DHEA(S)]

Zona glomerulosa(Aldosterone)

Zona fasciculata(Cortisol)

A

B C

Fig. 1. (A) The zones of the human adrenal cortex and the steroids they secrete. (B) The structure of the threezones (C, capsule; ZG, zona glomerulosa; ZF, zona fasciculata; ZR, zona reticularis). (C) The blood vessels of theadrenal cortex; the direction of blood flow is indicated (C, capsule; G, zona glomerulosa; F, zona fasciculata; R, zonareticularis; M, medulla; SA, adrenal artery; CP, capsule plexus; AP, medullary artery; Cap, capillaries; V, veins; MV,medullary veins; SV, adrenal vein; IC, smooth muscle in the wall of the adrenal vein). [(B) and (C) are reproducedfrom (80) with permission from Springer-Verlag]


related increase in the frequency of neoplastic and nonneoplas-tic lesions. These most commonly comprise nodules, somewhatless commonly adenomas, and more rarely carcinomas. Theage-related accumulation of small focal hyperplasias termednodules is so frequent as make it more or less a part of normalaging (18-21) (Fig. 3). In one study, mild to distinct nodularitywas observed in 65% of adrenal glands from 113 consecutiveautopsies in adults (18). Multiple small nodules are usually pre-sent, but relatively frequently one nodule enlarges, concurrentlywith atrophy of the rest of the adrenal cortex (22).

Benign tumors (adenomas) are larger than nodules (in terms ofweight, up to ~100 grams) and also occur relatively often. The in-creasingly routine use of patient imaging [computed tomography(CT, in which multiple x-ray images are incorporated into a cross-sectional image) and magnetic resonance imaging(http://sageke.sciencemag.org/cgi/content/full/2003/4/pe2) (MRI)] in diagnostic procedures has made thediscovery of suspicious f indings in the adrenalglands a very common occurrence. The great majori-ty of the suspicious lesions are in the cortex, ratherthan the medulla, and of these the great majority areadenomas or nodules (23-26). Over the past 20years, the common designation for these incidentallydiscovered adenomas has been “adrenal incidentalo-mas.” In one autopsy series, 8.7% of the populationover 65 was shown to have lesions of the adrenal cor-tex >1 cm in diameter (27). In a recent review, the au-thors comment, “With the widespread escalation in theuse of these imaging modalities and marked improve-ments in image resolution, the probability of findingan incidentaloma on cross-section imaging is rapidlyapproaching the 6% prevalence of previously reportedautopsy studies” (26). Grumbach et al., reporting onthe conclusions of a National Institutes of Health con-sensus conference, state “Improvements in abdominalimaging techniques have increased detection of adrenalincidentalomas, and because the prevalence of thesemasses increases with age, appropriate management ofadrenal tumors will be a growing challenge in our ag-ing society” (25). Continued improvements in technol-ogy will make the discovery of adrenal incidentalomasan even greater problem for the future, but even at pre-sent the adrenal incidentaloma has been termed a major publichealth problem (23-26).

Conventionally, most incidentalomas have been thought tobe “non-functioning adenomas.” This term is a misnomer, inthat in vitro studies have consistently shown that isolated cellsfrom these tumors synthesize glucocorticoids and/or mineralo-corticoids (20, 28, 29). The term refers to the concept that thecirculating plasma concentrations of adrenocortical steroidswere not thought to be elevated above normal in individualswith incidentalomas. However, recent studies have challengedthis idea. Subtle elevations in the concentrations of circulatingadrenocortical steroids, specifically cortisol, might be a causeof the metabolic syndrome (http://sageke.sciencemag.org/cgi/content/full/2003/41/ns6) (syndrome X), an increasingly preva-lent age-related disorder characterized by obesity, dyslipidemia(altered levels of lipoproteins and triglycerides), insulin resis-tance, impaired glucose tolerance, and hypertension (30, 31).

Whereas benign lesions (nodules and adenomas) are a fre-quent occurrence in the elderly, adrenocortical carcinomas are

rare. Carcinomas are larger (50 to 400 grams) than the benignlesions and most often discovered because of the symptomscaused by increased growth (22, 32). The most straightforwardinterpretation of the available clinical data is that adenomas arethe precursors of carcinomas but that the transition of adenomato carcinoma is rare. This interpretation is consistent with theconcept of multistep tumorigenesis as it occurs in other organs,such as the colon. Few adenomatous polyps in the colonprogress to colon cancer (33). However, as better diagnostictests are developed, increasingly smaller, earlier premalignantlesions will be detected (as observed with colonoscopy andCT/MRI scanning, for the colon and the adrenal cortex respec-tively), thereby contributing to the clinical dilemma of whetherfurther tests and treatment are necessary.

Possible Molecular Mechanisms for Age-RelatedChangesTo summarize the age-related changes in the human adrenalcortex, there is a reduction in the size of the ZR and a loss ofnormal zonation, which may account for the steep decline inDHEA(S) synthesis and an increase in disruptions of the nor-mal architecture—nodules, adenomas, and carcinomas, in or-der of decreasing frequency. Here, I propose that all or someof the following pathological processes may contribute tothese age-related changes: (i) infarcts (areas of tissue death)caused by temporary ischemia, and other causes of cell death,leading to disruption of tissue architecture; (ii) cellular senes-cence (telomere shortening), leading to failure of replacementof cells; (iii) DNA mutations or chromosome aberrations, ini-tiating and promoting neoplastic processes; and (iv) possibly,the production of senescence-related enzymes and cytokinesby adrenocortical cells or by fibroblasts within the tissue,leading to disruption of architecture and/or the promotion ofneoplasia.

R E V I E W


A

10 20 30 40 50 60 70 80

B

C D

E

F

Fig. 2. Diagrammatic representation of the plasma concentrations of DHEASover the life span in humans. Before birth (A) the fetal zone of the adrenalcortex secretes large amounts of DHEA(S); following birth (B), the fetal zonerapidly involutes (C). By about 6 to 7 years of age, ZR cells have developed inthe adrenal cortex and plasma DHEAS concentrations begin to rise (adre-narche) (D). The achievement of the peak concentration of plasma DHEAS inyoung adulthood (E) is followed by a progressive decline in adrenal secretionof DHEA(S), so that plasma DHEAS concentrations are often very low by age70 and beyond (F). The ordinate represents age in years; the scale is expand-ed before 10 years of age.


Ischemia/reperfusionA role for temporary ischemia (abnormally low blood flow) inage-associated changes in the adrenal cortex has been suspectedfor many years (18, 21), although direct evidence is still lack-ing. In this laboratory, we first became interested in this ques-tion when we assessed expression of the cell-cycle inhibitor p21in the human adrenal cortex. In adrenocortical cells in glandsfrom some organ donors (trauma victims), evidence of DNAdamage was accompanied by the presence of immunoreactivep53 and p21 (34), which normally would not be present in thesecells. DNA damage, present as single-strand breaks, caused ac-tivation of the tumor suppressor p53, which tran-scriptionally induced p21 (35). This process wasaccompanied by some apoptosis, presumably alsomediated by p53 (36).

We hypothesized that damage to the adrenalglands in human trauma victims would involve pe-riods of low blood flow alternating with periods ofhigher blood flow, as observed in clinical practice(21, 37-39). We observed that DNA damage andincreases in immunoreactive p53 and p21 could beinduced experimentally in the rat adrenal gland byischemia/reperfusion and by sepsis (high levels ofbacteria in the blood) (34, 40). A likely commonelement in injury of the adrenal gland is damageto the microvasculature. The function of capillary endothelial cells is altered by cytokines(http://sageke.sciencemag.org/cgi/content/full/2002/29/re3#SEC5) that circulate at high concen-trations during shock and sepsis. As a result, theendothelium (the layer of cells that lines the bloodvessels) becomes leaky, and red blood cells andleukocytes escape from the vessels into the tissue(41). Although there is no direct evidence for thismechanism, it would account for DNA damage byexposure of the cells in the tissues to oxygen-radi-cal generating systems that are normally se-questered in the blood, such as granulocytes(http://sageke.sciencemag.org/cgi/content/full/2002/29/re3#SEC7) or transition metal/proteincomplexes (42-44). Oxygen radicals (http://sageke.sciencemag.org/cgi/content/full/sageke;2001/1/oa5) and their products cause extensive strandbreaks in target cell DNA (45). The hypothesis Ipresent here is that over the life span of an individ-ual many such episodes of damage could occur,caused by temporary ischemia or by other kinds ofdamage to the capillary bed in the adrenal gland.Each such episode may result in a small infarct, thedeath of a few cells or perhaps a segment of thecortex. In fact, many years ago pathologists had al-ready noted evidence for such infarcts and their re-lation to ischemia (18, 21). For several reasons, di-rect evidence for this hypothesis is difficult to obtain. The rate ofcell death in the human adrenal cortex under “normal” circum-stances is quite uncertain. Adrenal glands available for study arealmost never normal; either they have been surgically removedfrom patients with serious illnesses or they have been obtainedfrom cadaver organ donors, who have often been maintained forseveral days on life-support systems before death.

It seems unlikely that cell death per se, in the absence of

other changes, could result in permanent alterations in thestructure of the adrenal cortex. In the absence of other changes,cell loss would be expected to be balanced by cell division torestore the original size of the tissue. Indeed, this is what is as-sumed to occur under normal conditions in the young healthyadult, where cell death in the ZR is balanced by cell division inthe outer cortex. Although the normal mode by which the sizeof the adrenal cortex is controlled is only partly understood, itis known to involve feedback via the hypothalamus and pitu-itary (3). Loss of adrenocortical cells (ZF cells) leads to somedecrease in secreted steroids (cortisol). By feedback at the lev-

el of the hypothalamus and pituitary, this decrease leads to in-creased secretion of adrenocorticotrophic hormone (ACTH) bythe pituitary gland, which stimulates the growth of the adrenalcortex (3). A dramatic example of the operation of this feed-back cycle is seen when one of the enzymes of steroidogenesisis missing because of a genetic defect; this absence causes con-genital adrenal hyperplasia, a greatly enlarged adrenal cortex(46). Suppression of ACTH secretion leads to a reversal of the

R E V I E W


Fig. 3. Diagrammatic representation of age-related nodule formation in the hu-man adrenal cortex. [Reproduced from (18) with permission from John Wileyand Sons Ltd]


excess growth. Therefore, in adult life the loss of some adreno-cortical cells would be expected to be compensated by feed-back. For cell death to create a permanent defect in structureand function, there must be some interference with the normalmode of cell replacement.

Several factors could lead to imperfect feedback and therebycontribute to making the defect permanent. First, loss of ZRcells, as opposed to ZF cells, does not lead to a decrease in cor-tisol secretion and therefore does not increase ACTH secretionby the hypothalamus and pituitary; so far as is known, a de-crease in DHEA(S) levels does not exert a feedback action. Sec-ond, the cells could lose the ability to divide, perhaps as a resultof cellular senescence (see below). Third, it may be that infarctscaused by ischemia cannot be repaired with complete efficien-cy. The loss of an entire capillary (or one segment of it), togeth-er with the adrenocortical cells that are dependent on the vessel,might be permanent. The loss of such a unit may well increasethe stimulus to growth acting on the rest of the cortex, but re-generation of the local deficit in the vicinity of an infarct mightnot occur. Over time, this process may result in the loss of seg-ments of the cortex, together with compensatory hyperprolifera-tion of other parts of the cortex.

Specific susceptibility of ZR cells to apoptosisGiven these potential mechanisms for the loss of adrenocorticalcells, why are ZR cells especially vulnerable to being lost and notreplaced? One possibility is that, because the ZR is on the venousend of the adrenocortical capillary bed (3), it is more liable to bedamaged by ischemia. That possibility would also be consistentwith the increasing irregularity of the ZF/ZR border—infarcts oc-curring on the venous end of the capillary bed could produce smallareas of damage in the ZR. Additionally, it is possible that differ-ences in gene expression between ZF cells and ZR cells cause ZRcells to be primed to undergo apoptosis and/or to be eliminated bythe immune system. ZR cells, but not ZF cells, express MHC classII antigens on their cell surface (47, 48). MHC class II moleculescan be expressed by a wide variety of normal tissues in vivo, aswell as in tumors and tissues affected by abnormal conditions suchas autoimmune diseases (49). Cells expressing class II genes do notgenerally act as antigen-presenting cells because they lack the co-factors for antigen presentation (49). MHC class II antigen expres-sion on ZR cells may result from the effects of cytokines releasedby macrophages or other immune cells within the adrenal gland(50). However, unexpectedly, MHC class II expression was report-ed to persist in ZR cells in culture, even though they were mixedwith ZF cells that did not become class II–positive (47). Whateverthe cause, expression of MHC class II molecules could make ZRcells susceptible to immune attack. Fas protein (a cell-surface pro-tein that mediates apoptosis, also called CD95) is also expressed ata greater level on ZR cells (51, 52). Cells expressing Fas proteinmay be killed by Fas ligand or by anti-Fas antibodies (53).

Cellular senescenceA second factor contributing to the age-related loss of ZR cellsmay be loss of replicative capacity. In culture, normal humancells exhibit a limited proliferative capacity that results fromshortening of telomeres, which in turn results from lack oftelomerase activity (54) (see Hornsby Perspective at http://sageke.sciencemag.org/cgi/content/full/2003/30/pe21 and “MoreThan a Sum of Our Cells” at http://sageke.sciencemag.org/cgi/content/full/2001/1/oa4). We investigated the relationship be-

tween telomere biology and replicative senescence in humanadrenocortical cells by measuring replicative capacity andtelomere length as a function of donor age (55). Cells culturedfrom adrenal tissue from donors of different ages were found toshow a strong age-related decline in total replicative capacity,falling from ~50 population doublings for fetal cells to an almosttotal lack of division in culture for cells from older donors (>70years). Telomere restriction fragment (TRF) length was analyzedin the same sets of cells and was found to decrease from a valueof ~12 kb in fetal cells to ~7 kb in cells from older donors. Thelatter value is consistent with that in f ibroblasts that havereached replicative senescence. Furthermore, there was a goodcorrelation in individual donor samples between TRF length andreplicative capacity in culture. These data show that adrenocorti-cal cells undergo continuous telomere shortening over the lifespan and that cells in glands from older donors have relativelyshort telomeres.

Telomere shortening has been observed in a variety of humantissues as a function of donor age (56). Studies on various nonfi-broblast cells have often shown striking decreases in prolifera-tive capacity as donor age increases (56). However, whethertelomere shortening in vivo actually results in cells that can nolonger divide in situ has never been directly determined. Thiscomplex question has been discussed in detail elsewhere (57).

Age-Related Disruptions of Tissue ArchitectureThese factors—ischemia, apoptosis, and deficits in the capacityfor cellular replacement—may also contribute to the developmentof nodules and neoplastic lesions of the cortex. As suggested byGeorge Martin, one consequence of loss of replicative capacitymay be compensatory local hyperplasia (58). This scenario cer-tainly fits the observed age-related changes in the adrenal cortex.The cumulative effect of cell loss in the adrenal cortex is an in-crease in circulating ACTH concentrations, mediated by feedbackto the hypothalamus and pituitary. When part of the cortex isdamaged and is unable to undergo compensatory growth, theremay be an overcompensation of growth of other parts of the cor-tex, perhaps leading to nodule formation. The pattern of capillar-ies in nodules is in agreement with this possibility (59); the nor-mal capillary pattern is discernible both in nodules and in the ad-jacent tissue but is distorted by expansion of the tissue within thenodule and by compression of the tissue adjacent to the nodule(59). Evidence that ACTH can drive nodule formation comesfrom a rare cause of Cushing’s disease, in which excess cortisol issecreted by a hyperplastic nodular cortex that results from the ac-tion of abnormally high concentrations of ACTH on adrenocorti-cal cell proliferation (60). Moreover, the hyperplastic adrenalglands in congenital adrenal hyperplasia often show nodules, andthese diminish in size when the excess ACTH that occurs in thissyndrome is normalized (61).

Nodules appear to be distortions of the architecture of the cor-tex rather than clonal neoplastic growths. On the other hand, thereis substantial evidence that adenomas and carcinomas are clonaland that their chromosomes have been altered by mutations andrearrangements (32, 62), but a detailed consideration of the ge-netic and other factors that lead to adenomas and carcinomas ofthe adrenal cortex is beyond the scope of this Review (32, 62).

Possible effects of senescent cellsSenescent cells may be formed by telomere shortening and alsoby a variety of other forms of cellular stress and damage (54). If

R E V I E W



cells within the adrenal cortex do become senescent during ag-ing, and if they persist in the tissue rather than dying or beingeliminated by the immune system, they could have substantial ef-fects on neighboring cells (63). Senescent cells undergo a wealthof changes in gene expression. This finding has been well docu-mented in fibroblasts but is less well established for other celltypes, including adrenocortical cells. Enzymes, extracellular ma-trix proteins, and cytokines produced by senescent cells could af-fect the surrounding tissue by (i) promoting the growth of neo-plastic cells (adenomas or carcinomas), (ii) disrupting tissue ar-chitecture, and (iii) causing alterations in capillary maintenanceand growth, for example (63). However, the frequency of the oc-currence of senescent cells within aging tissues is still not wellestablished, and therefore the relative importance of this mecha-nism for age-related changes in tissues is uncertain (57).

Functional Consequences of Age-Related ChangesIn humans, ZR cells are responsible for the synthesis ofDHEA(S). The function of DHEA(S) is not known. As I have dis-cussed elsewhere (1, 2), these steroids are usually and correctlytermed “adrenal androgens,” because, although they lack directandrogen activity, they are readily converted into active androgensin many tissues, but whether this is their major function is notclear. Low serum concentrations of DHEAS have been associatedwith frailty (see Walston Perspective at http://sageke.sciencemag.org/cgi/content/full/2004/4/pe4), whereas higher than normalserum concentrations have been observed in rhesus monkeys un-dergoing calorie restriction (http://sageke.sciencemag.org/cgi/content/full/2003/8/re2) and in long-lived human males (see“Monkey in the Middle” at http://sageke.sciencemag.org/cgi/con-tent/abstract/2002/31/nw108). DHEA has been reported to have avariety of effects when administered to human subjects (64, 65),but these studies have not shed light on the normal functions ofthese steroids. The lack of suitable animal models has severely im-peded progress in this area.

The human adrenal cortex has three functional zones only inthe adult, from prepuberty to old age (66). ZR cells do not appearin substantial numbers in the adrenal cortex until about 6 to 7years of age, when plasma DHEAS concentrations begin to rise(67-69). This process is termed “adrenarche.” Adrenarche is ob-served only in species closely related to humans, such as the goril-la and chimpanzee (67, 70). Primate species outside this group,such as the rhesus and cynomolgus monkeys, have measurableconcentrations of DHEAS in plasma but do not undergo adre-narche and do not maintain the high adult concentrations ofDHEAS found in humans (71-74). Evidently, therefore, thebiosynthesis of DHEA(S) by the adult human adrenal cortex hasevolved recently. It is an inescapable conclusion that these steroidsmust have some important function, despite our ignorance of itsnature. That function might not be evident under the present-dayliving conditions of the human species, however (1, 2).

What factors at the time of adrenarche cause the differentia-tion of a new cell type within the adrenal cortex? Presumably,although this is not unequivocally established, ZR cells derivefrom existing ZF cells. As the adrenal cortex grows in thick-ness, ZR cells may differentiate in response to signals existingon the venous end of the capillary bed, perhaps as a result of thepresence of a certain concentration of a morphogen originatingin the outer cortex. The signal for the differentiation of ZR cellsappears to be of adrenal origin, but its biosynthesis might be in-fluenced by pituitary factors. Adrenarche does not take place in

individuals with congenital unresponsiveness to ACTH (75).Some cases of precocious adrenarche are associated withgrowth hormone and prolactin excess (76). A dependence onadrenal growth for the occurrence of adrenarche can be inferredfrom data on patients treated with glucocorticoids for congeni-tal adrenal hyperplasia resulting from genetic deficiency of 21-hydroxylase, one of the enzymes of steroid biosynthesis (77,78). This treatment was associated in adult life with extremelylow levels of DHEAS, even when the patients received nosteroids for a period of 6 weeks (79). The authors suggested thatsteroid treatment interferes with adrenarche by suppressinggrowth of the cortex so that it does not achieve a width appro-priate for the development of the ZR.

At adrenarche, the development of a certain number of ZRcells sets a certain level of DHEA(S) biosynthesis. The surge ofDHEA(S) biosynthesis at adrenarche can be viewed as fulfillingan undetermined biological requirement for DHEA(S) at thisphase of the life span, where crossing some threshold for theplasma concentrations of these steroids is needed but preciseregulation is not required (1, 2). The high variability of peakadrenal androgen levels in young adulthood and the variabilityin the age-related decline is consistent with this concept. Solv-ing the problem of the importance of the age-related decline inDHEA(S) biosynthesis requires an understanding of the func-tion of these steroids in young adult life.

Summary and Future QuestionsAn attractive model for the age-related changes in the humanadrenal cortex can be made that fits together the consequencesof cellular senescence resulting from telomere shortening, tem-porary ischemia leading to small infarcts, and compensatorygrowth resulting in increased nodularity and zonal irregularity.A consequence of these changes might be the gradual loss ofZR cells and the loss of DHEA(S) biosynthesis. However, thismodel is speculative and needs experimental verification. Alarge problem here is that animal models may be unreliable forseveral reasons. First, most mammalian species that are avail-able for experimental study lack a functional DHEA(S)-secret-ing ZR. Second, age-related cellular processes in rodents coulddiffer substantially from those in humans and other long-livedmammals (see Hornsby Perspective at http://sageke.sci-encemag.org/cgi/content/full/2003/30/pe21). In my laboratory,we have used cell transplantation techniques to approach thesequestions (4), but evidently a variety of creative approaches isneeded to solve the difficult problems of human aging.

References1. P. J. Hornsby, Biosynthesis of DHEAS by the human adrenal cortex and it-

sage-related decline. Ann. N.Y. Acad. Sci. 774, 29-46 (1995).2. P. J. Hornsby, DHEA: A biologist’s perspective. J. Am. Geriatrics Soc. 45,

1395-1401 (1997).3. P. J. Hornsby, in Adrenal Cortex, D.C. Anderson, J. S. D. Winter, Eds. (But-

terworth, London, 1985), pp. 1-31.4 P. J. Hornsby, Transplantation of adrenocortical cells. Rev. Endocr. Metab.

Disord. 2, 313-321 (2001).5. M. Heikkila, H. Peltoketo, J. Leppaluoto, M. Ilves, O. Vuolteenaho, S. Vainio,

Wnt-4 deficiency alters mouse adrenal cortex function, reducing aldos-terone production. Endocrinology 143, 4358-4365 (2002).

6. T. Suwa, M. Chen, C. L. Hawks, P. J. Hornsby, Zonal expression of dickkopf-3 and components of the Wnt signalling pathways in the human adrenalcortex. J. Endocrinol. 178, 149-158 (2003).

7. W. J. Nelson, R. Nusse, Convergence of Wnt, beta-catenin, and cadherinpathways. Science 303, 1483-1487 (2004).

8. N. Orentreich, J. L. Brind, R. L. Rizer, J. H. Vogelman, Age changes andsex differences in serum dehydroepiandrosterone sulfate concentrationsthroughout adulthood. J. Clin. Endocrinol. Metab. 59, 551-555 (1984).

R E V I E W



9. F. Labrie, A. Belanger, L. Cusan, J. L. Gomez, B. Candas, Marked declinein serum concentrations of adrenal C19 sex steroid precursors and conju-gated androgen metabolites during aging. J. Clin. Endocrinol. Metab. 82,2396-2402 (1997).

10. J. E. Morley, F. Kaiser, W. J. Raum, H. M. Perry III, J. F. Flood, J. Jensen, A.J. Silver, E. Roberts, Potentially predictive and manipulable blood serumcorrelates of aging in the healthy human male: Progressive decreases inbioavailable testosterone, dehydroepiandrosterone sulfate, and the ratio ofinsulin-like growth factor 1 to growth hormone. Proc. Natl. Acad. Sci. U.S.A.94, 7537-7542 (1997).

11. C. R. Parker Jr., S. M. Slayden, R. Azziz, S. L. Crabbe, G. A. Hines, L. R.Boots, S. Bae, Effects of aging on adrenal function in the human: Respon-siveness and sensitivity of adrenal androgens and cortisol to adrenocorti-cotropin in premenopausal and postmenopausal women. J. Clin. En-docrinol. 85, 48-54 (2000).

12. C. R. Parker Jr., R. L. Mixon, R. M. Brissie, W. E. Grizzle, Aging alters zona-tion in the adrenal cortex of men. J. Clin. Endocrinol. Metab. 82, 3898-3901(1997).

13. C. R. Parker Jr., Dehydroepiandrosterone and dehydroepiandrosterone sul-fate production in the human adrenal during development and aging.Steroids 64, 640-647 (1999).

14. E. Kreiner, G. Dhom, Altersveranderungen der menschlichen Nebenniere.Zbl. Allg. Pathol. Pathol. Anat. 123, 351-356 (1979).

15. H. Sasano, J. I. Mason, N. Sasano, H. Nagura, Immunolocalization of 3β-hydroxysteroid dehydrogenase in human adrenal cortex and its disorders.Endocrine Pathol. 1, 94-101 (1990).

16. C. Marx, S. R. Bornstein, G. W. Wolkersdorfer, M. Peter, W. G. Sippell, W. A.Scherbaum, Relevance of major histocompatibility complex class II expres-sion as a hallmark for the cellular differentiation in the human adrenal cor-tex. J. Clin. Endocrinol. Metab. 82, 3136-3140 (1997).

17. S. Mapes, C. J. Corbin, A. Tarantal, A. Conley, The primate adrenal zonareticularis is defined by expression of cytochrome b5, 17alpha-hydroxy-lase/17,20-lyase cytochrome P450 (P450c17) and NADPH-cytochromeP450 reductase (reductase) but not 3beta-hydroxysteroid dehydroge-nase/delta5-4 isomerase (3beta-HSD). J. Clin. Endocrinol. Metab. 84,3382-3385 (1999).

18. J. W. Dobbie, Adrenocortical nodular hyperplasia: The ageing adrenal. J.Pathol. 99, 1-18 (1969).

19. A. M. Neville, The nodular adrenal. Invest. Cell Pathol. 1, 99-111 (1978).20. A. M. Neville, M. J. O’Hare, The Human Adrenal Cortex. Pathology and Bi-

ology: An Integrated Approach (Springer-Verlag, Berlin, 1982).21. F. P. Kuhajda, G. M. Hutchins, Adrenal cortico-medullary junction necrosis:

A morphological marker for hypotension. Am. Heart J. 98, 294-297 (1979).22. E. E. Lack, Atlas of Tumor Pathology. Third Series. Fascicle 19. Tumors of

the Adrenal Gland and Extra-Adrenal Paraganglia (Armed Forces Instituteof Pathology, Washington, D.C., 1997).

23. S. R. Bornstein, C. A. Stratakis, G. P. Chrousos, Adrenocortical tumors: re-cent advances in basic concepts and clinical management. Ann. Intern.Med. 130, 759-771 (1999).

24. D. C. Aron, The adrenal incidentaloma: Disease of modern technology andpublic health problem. Rev. Endocr. Metab. Disord. 2, 335-342 (2001).

25. M. M. Grumbach, B. M. Biller, G. D. Braunstein, K. K. Campbell, J. A. Car-ney, P. A. Godley, E. L. Harris, J. K. Lee, Y. C. Oertel, M. C. Posner et al.,Management of the clinically inapparent adrenal mass (“incidentaloma”).Ann. Intern. Med. 138, 424-429 (2003).

26. G. B. Thompson, W. F. Young Jr., Adrenal incidentaloma. Curr. Opin. Oncol.15, 84-90 (2003).

27. H. Hedeland, G. Ostberg, B. Hokfelt, On the prevalence of adrenocorticaladenomas in an autopsy material in relation to hypertension and diabetes.Acta Med. Scand. 184, 211-214 (1968).

28. M. J. O’Hare, P. Monaghan, A. M. Neville, The pathology of adrenocorticalneoplasia: A correlated structural and functional approach to the diagnosisof malignant disease. Hum. Pathol. 10, 137-154 (1979).

29. K. Matsuo, K. Kawai, H. Tsuchiyama, Nonfunctioning adrenocortical adeno-ma in culture: Quantitative and morphological observations. Acta Pathol.Jpn. 35, 871-884 (1985).

30. R. Rossi, L. Tauchmanova, A. Luciano, M. Di Martino, C. Battista, L. Del Vis-covo, V. Nuzzo, G. Lombardi, Subclinical Cushing’s syndrome in patientswith adrenal incidentaloma: clinical and biochemical features. J. Clin. En-docrinol. Metab. 85, 1440-1448 (2000).

31. M. Terzolo, A. Pia, A. Ali, G. Osella, G. Reimondo, S. Bovio, F. Daffara, M.Procopio, P. Paccotti, G. Borretta, A. Angeli, Adrenal incidentaloma: A newcause of the metabolic syndrome? J. Clin. Endocrinol. Metab. 87, 998-1003(2002).

32. C. A. Stratakis, G. P. Chrousos, Adrenal cancer. Endocrinol. Metab. Clin.North. Am. 29, 15-25 (2000).

33. B. Vogelstein, K. W. Kinzler, The multistep nature of cancer. Trends Genet.

9, 138-141 (1993).34. V. V. Didenko, X. Wang, L.Yang, P.J. Hornsby, Expression of p21WAF1/CIP1/SDI1

and p53 in apoptotic cells in the adrenal cor tex and induction byischemia/reperfusion injury. J. Clin. Invest. 97, 1723-1731 (1996).

35. W. S. El-Deiry, T. Tokino, V. E. Velculescu, D. B. Levy, R. Parsons, J. M. Trent,D. Lin, W. E. Mercer, K. W. Kinzler, B. Vogelstein, WAF1, a potential media-tor of p53 tumor suppression. Cell 75, 817-825 (1993).

36. J. D. Weber, G. P. Zambetti, Renewing the debate over the p53 apoptoticresponse. Cell Death Differ. 10, 409-412 (2003).

37. L. M. Siegel, S. K. Grinspoon, G. J. Garvey, J. P. Bilezikian, Sepsis andadrenal function. Trends Endocrinol. Metab. 5, 324-328 (1994).

38. B. F. Painter, Infectious causes of adrenal insufficiency. Infections Med. 11,515-520 (1994).

39. R. H. Rao, Bilateral massive adrenal hemorrhage. Med. Clin. N. Am. 79,107-129 (1995).

40. V. V. Didenko, X. Wang, L. Yang, P. J. Hornsby, DNA damage andp21WAF1/CIP1/SDI1 in experimental injury of the rat adrenal cortex and trauma-associated damage of the human adrenal cortex. J. Pathol. 189, 119-126(1999).

41. L. B. Hinshaw, Sepsis/septic shock: Participation of the microcirculation: Anabbreviated review. Crit. Care Med. 24, 1072-1078 (1996).

42. B. W. Griffin, in Oxygen and Oxy-Radicals in Chemistry and Biology, M. A.J. Rogers, E. L. Powers, Eds. (Academic Press, New York, 1981), pp. 141-146.

43. R. L. Baehner, L. A. Boxer, L. M. Ingraham, in Free Radicals in Biology, Vol.5, W. A. Pryor, Eds. (Academic Press, New York, 1982), pp. 91-114.

44. B. Halliwell, J. M. C. Gutteridge, Iron and free radical reactions: Two aspectsof antioxidant protection. Trends Biochem. Sci. 11, 372-375 (1986).

45. J. A. Imlay, S. Linn, DNA damage and oxygen radical toxicity. Science 240,1302-1309 (1988).

46. M. I. New, Inborn errors of adrenal steroidogenesis. Mol. Cell. Endocrinol.211, 75-83 (2003).

47. E. L. Khoury, J. S. Greenspan, F. S. Greenspan, Adrenocortical cells of thezona reticularis normally express HLA-DR antigenic determinants. Am. J.Pathol. 127, 580-591 (1987).

48. E. L. Khoury, J. W. Berline, HLA-DR expression by adrenocortical cells ofthe zona reticularis: Structural and allotypic characterization. Tissue Anti-gens 31, 191-203 (1988).

49. J. Guardiola, A. Maffei, Control of MHC class II gene expression in autoim-mune, infectious, and neoplastic diseases. Crit. Rev. Immunol. 13, 247-68(1993).

50. M. Ehrhart-Bornstein, S. R. Bornstein, W. A. Scherbaum, Sympathoadrenalsystem and immune system in the regulation of adrenocortical function.Eur. J. Endocrinol. 135, 19-26 (1996).

51. G. W. Wolkersdorfer, M. Ehrhart-Bornstein, S. Brauer, C. Marx, W. A.Scherbaum, S. R. Bornstein, Differential regulation of apoptosis in the nor-mal human adrenal gland. J. Clin. Endocrinol. Metab. 81, 4129-4136(1996).

52. G. W. Wolkersdorfer, C. Marx, J. W. Brown, W. A. Scherbaum, S. R. Born-stein, Evaluation of apoptotic parameters in normal and neoplastic humanadrenal. Endocrine Res. 22, 411-419 (1996).

53. S. Nagata, P. Golstein, The Fas death factor. Science 267, 1449-56 (1995).54. W. E. Wright, J. W. Shay, Historical claims and current interpretations of

replicative aging. Nature Biotechnol. 20, 682-688 (2002).55. L. Yang, T. Suwa, W. E. Wright, J. W. Shay, P. J. Hornsby, Telomere shorten-

ing and decline in replicative potential as a function of donor age in humanadrenocortical cells. Mech. Ageing Dev. 122, 1685-1694 (2001).

56. P. J. Hornsby, in Handbook of the Biology of Aging (Fifth Edition), E. J. Ma-soro, S. N. Austad, Eds. (Academic Press, San Diego, 2001), pp. 207-266.

57. P. J. Hornsby, Cellular senescence and tissue aging in vivo. J. Gerontol. ABiol. Sci. Med. Sci. 57, B251-B256 (2002).

58. G. M. Martin, Clonal attenuation: Causes and consequences. J. Gerontol. ABiol. Sci. Med. Sci. 48, B171-B172 (1993).

59. D. P. Magennis, A. M. McNicol, Vascular patterns in the normal and patho-logical human adrenal cortex. Virchows Archv 433, 69-73 (1998).

60. M. Boscaro, L. Barzon, N. Sonino, The diagnosis of Cushing’s syndrome:atypical presentations and laboratory shortcomings. Arch. Intern. Med. 160,3045-3053 (2000).

61. L. R. Giacaglia, B. B. Mendonca, G. Madureira, K. F. Melo, C. A. Suslik, I. J.Arnhold, T. A. Bachega, Adrenal nodules in patients with congenitaladrenal hyperplasia due to 21-hydroxylase deficiency: Regression after ad-equate hormonal control. J Pediatr. Endocrinol. Metab. 14, 415-419 (2001).

62. C. Gicquel, J. Bertherat, Y. Le Bouc, X. Bertagna, Pathogenesis of adreno-cortical incidentalomas and genetic syndromes associated with adrenocor-tical neoplasms. Endocrinol. Metab. Clin. North Am. 29, 1-13 (2000).

63. J. Campisi, Cancer and ageing: Rival demons? Nature Rev. Cancer 3, 339-349 (2003).

R E V I E W



64. F. Labrie, V. Luu-The, C. Labrie, J. Simard, DHEA and its transformation in-to androgens and estrogens in peripheral target tissues: Intracrinology.Front. Neuroendocrinol. 22, 185-212 (2001).

65. P. Celec, L. Starka, Dehydroepiandrosterone: Is the fountain of youth dryingout? Physiol. Res. 52, 397-407 (2003).

66. G. Dhom, The prepuberal and puberal growth of the adrenal (adrenarche).Beitr. Path. 150, 357-377 (1973).

67. G. B. Cutler Jr., M. Glenn, M. Bush, G. D. Hodgen, C. E. Graham, D. L. Lori-aux, Adrenarche: A survey of rodents, domestic animals, and primates. En-docrinology 103, 2112-2118 (1978).

68. B. H. Rich, R. L. Rosenfield, A. W. Lucky, J. C. Helke, P. Otto, Adrenarche:Changing adrenal response to adrenocorticotropin. J. Clin. Endocrinol.Metab. 52, 1129-1136 (1981).

69. R. L. Rosenfield, Normal and almost normal precocious variations in puber-tal development: Premature pubarche and premature thelarche revisited.Horm. Res. 41 Suppl. 2, 7-13 (1994).

70. P. J. Smail, C. Faiman, W. C. Hobson, G. B. Fuller, J. S. D. Winter, Furtherstudies on adrenarche in nonhuman primates. Endocrinology 111, 844-848(1982).

71. D. R. Koritnik, R. F. Laherty, D. Rotten, R. B. Jaffe, A radioimmunoassay fordehydroepiandrosterone sulfate in the circulation of rhesus monkeys.Steroids 42, 653-667 (1983).

72. N. Meusy-Dessolle, D. C. Dang, Plasma concentrations of testosterone, di-hydrotestosterone, delta4-androstenedione, dehydroepiandrosterone, andoestradiol-17β in the crab-eating monkey (Macaca fascicularis) from birth

to adulthood. J. Reprod. Fertil. 74, 347-359 (1985).73. C. Guillemette, D. W. Hum, A. Belanger, Levels of plasma C19 steroids and

5α-reduced C19 steroid glucuronides in primates, rodents, and domesticanimals. Am. J. Physiol. 271, E348-E353 (1996).

74. M. A. Lane, D. K. Ingram, S. S. Ball, G. S. Roth, Dehydroepiandrosteronesulfate: A biomarker of primate aging slowed by calorie restriction. J. Clin.Endocrinol. Metab. 82, 2093-2096 (1997).

75. A. Weber, A. J. L. Clark, L. A. Perry, J. W. Honour, M. O. Savage, Diminishedadrenal androgen secretion in familial glucocorticoid deficiency implicatesa significant role for ACTH in the induction of adrenarche. Clin. Endocrinol.46, 431-437 (1997).

76. N. Iwatani, M. Kodama, H. Seto, A child with pituitary gigantism and preco-cious adrenarche: Does GH and/or PRL advance the onset of adrenarche?Endocrinol. Japon. 39, 251-257 (1992).

77. P. Dewis, D. C. Anderson, in Adrenal Cortex, D. C. Anderson, J. S. D. Winter,Eds. (Butterworth, London, 1985), pp. 96-116.

78. I. Rezvani, L. R. Garibaldi, A. M. Digeorge, H. G. Artman, Disproportionatesuppression of dehydroepiandrosterone sulfate (DHEAS) in treated pa-tients with congenital adrenal hyperplasia due to 21-hydroxylase deficien-cy. Pediatr. Res. 17, 131-134 (1983).

79. D. I. Conway, D. C. Anderson, D. E. Bu’Lock, The steroid response to con-trolled adrenal stimulation in congenital adrenal hyperplasia.

80. R. V. Krstic, Ed., Illustrated encycolpedia of Human Histology, (Springer-Verlag, Berlin, 1984), pp. 9-10.

R E V I E W


aging of the human adrenal cortex - sage kesageke.sciencemag.org/cgi/reprint/2004/35/re6.pdf ·...

Documents