|
|
||||||||
1 Plan II, The University of Texas at Austin.
2 Department of Cell Biology, University of Texas Southwestern Medical Center, Dallas.
3 Department of Biopathology and Diagnostic Imaging, Tor Vergata University, Rome, Italy.
4 Department of Psychology, The University of Texas at Austin.
Address correspondence to Marc Lewis, PhD, Department of Psychology, The University of Texas at Austin, Austin, Texas, 78703. E-mail: lewis{at}psy.utexas.edu
| Abstract |
|---|
|
|
|---|
|
| PROGERIAS WITH SYMPTOMS IN COMMON |
|---|
|
|
|---|
|
Nail atrophy can occur for secondary reasons such as shoe trauma or scarring from the disease lichen planus, and as the occasional side effect of some peripheral vascular disorders (e.g., leprosy), but it is almost never a primary disease symptom. In fact, we searched more than 7000 diseases and conditions (see "Scope of Literature Review") and found nail atrophy mentioned as a symptom in just five. Four of those five diseases are progerias. The fifth disease, Cronkhite-Canada Syndrome (CCS), has not fully been studied or classified, but it has much in common with the progerias, and it may represent an unrecognized addition to that family of diseases. We will discuss CCS separately a little later.
Nail atrophy is not only a rare symptom, but it also co-occurs with two other less rare, but relatively uncommon symptoms, alopecia and osteoporosis. The four progerias in the third column are, in fact, the only diseases in the literature that share these three symptoms. It is improbable that the association is coincidental, and we will argue that a single underlying process unites them. We begin with a discussion of fingernails.
| THE FINGERNAIL SYSTEM |
|---|
|
|
|---|
|
Although nail keratinocytes escape photodamage, no cycling cell can escape the dangers of replication. Fingernail TACs replicate frequently and, like all replicating cells, each time they do they lose an average of 50100 bp of DNA because of the end replication problem (5,6) and subsequent processing of the resulting 3' overhang. The loss is not crucial in early life because humans are born with about 15 kb of disposable TTAGGG tandem repeats (telomeres) at the ends of their chromosomes. But with replication those protective telomeres shorten. The danger is not so much in the shortening itself, but rather in the loss of the second protective role of telomeres, that of capping chromosomes and preventing the cell from mistakenly treating their ends as double strand breaks. The telomere achieves that protected state by, among other things, looping back and tucking its 3' overhang into the double stranded upstream telomeric DNA (7). The result is a large telomere loop, ending in a smaller displacement loop, which is held in place by critical proteins such as TRF2 (reviewed in 8). As telomeres shorten, this end protection is lost, possibly because critical proteins are unable to maintain the loop. The result is either senescence, or in some instances, end-to-end fusions that create a fusion-breakage-bridge cycle (9).
Some highly replicating cells mitigate the problems caused by shortening telomeres by replacing some lost repeats using the cellular reverse transcriptase enzyme, telomerase. Hair keratinocyte stem cells, for example, have weak telomerase activity, and there is strong activity in their rapidly dividing TAC progeny (10). The amount of telomerase present is not, however, sufficient to fully maintain telomeres and so keratinocyte telomeres get progressively shorter with age. Telomerase has not been studied in nails, but it is found in highly replicative cells of skin (11), gastrointestinal tract-stomach-intestine (12), bone marrow (13), testes (14), and generally in all systems affected in the telomerase disorder, Dyskeratosis Congenita.
In some cases, such as Werner Syndrome, the capping loop may be lost before the telomeres become excessively short. As a result, Werner Syndrome cells senesce more quickly than normal cells, but possibly with longer telomeres (15). That scenario is not fully proven, but an emerging concept is that short-telomere problems are probably due to loop (capping) malfunctions (e.g., 16). Nevertheless, short telomeres have been established in certain of the progerias, whereas capping problems have not. For that reason, we will refer to those progerias as having a problem of short telomeres with the understanding that the final problem may be loss of capping.
It is generally believed that with each replication, the cell must transiently untie the capping loop, copy its DNA, replace some of the lost telomere repeats (if active telomerase is present), and retie the loop (reviewed in 17). If lost repeats are not replaced (as happens in Dyskeratosis Congenita), or if the mechanism that closes the displacement loop is dysfunctional, as may happen in Werner Syndrome (16), telomeres progressively shorten and eventually cause the cell to enter senescence or apoptosis. This loss of replicative ability is the likely cause of fingernail atrophy. The pattern of diseases in Figure 1 supports the idea that nail atrophy is telomere driven. Figure 1 divides the progerias into the six that have short telomeresDyskeratosis Congenita (18), HGP (19), Werner Syndrome (15), Ataxia Telangiectasia (20), NBS (21)and the five that do not. Nail atrophy occurs only in progerias that have short telomeres.
<1?twb=.25w>Nail atrophy does not, however, occur in every short-telomere progeria. The two exceptions, however, Ataxia Telangiectasia and NBS, strengthen the evidence for telomere involvement with nails. Unlike those in the other progerias in this category, the telomeres in Ataxia Telangiectasia and NBS do not shorten because of increased DNA replication problems. Instead, in both diseases, cells with DNA double strand breaks fail to halt for repairs at the S-phase checkpoint; this failure is a problem that eventually leads to a fusion-bridge-breakage cycle and short telomeres. Because of the repair problem in these two diseases, chromosomes that acquire normal DNA damage should have short telomeres. But, as we have already explained, there is little DNA damage in the nail system. Thus, in nails, Ataxia Telangiectasia and NBS difficulties in damage repair should not create a problem. With little damage to repair, telomeres in Ataxia Telangiectasia and NBS nail stem cell keratinocytes should not shorten, and hence, nail atrophy should not, and does not, occur.
The short-telomere group contains one provisional entryRTS. Two-thirds of RTS cases are caused by mutations in the RECQ4 helicase (22,23); the remaining one-third of cases have an unknown cause. RTS has the nail-alopecia-osteoporosis triad seen in other short-telomere diseases, but its telomeres have not been studied. The symptom pattern of the disease suggests that short telomeres might occur in RTS, especially in bone progenitors where the RECQ4 protein is highly expressed. Against this prediction weighs the fact that life span is shortened in all of the progerias in which short telomeres have been established, but life span is not shortened in RTS. It is possible, then, that RTS either does not have short telomeres or its effect on telomeres is more subtle, as is the case in Werner Syndrome. We are currently testing various aspects of these predictions.
The fifth disease with nail atrophy is CCS, Cronkhite Canada Syndrome. The age of diagnosis (range 3186, mean 55) of CCS is later than that of most progeroid syndromes, but similar to that of Werner Syndrome, and many of its symptoms indicate accelerated aging. Alopecia, hyperpigmentation, vitiligo, anemia, gastrointestinal carcinoma, cataracts, loss of taste, and malabsorption (50% die of malnutrition) are common. There are more than 100 case reports of CCS in the literature, but few research studies, and the disease remains rare and mysterious.
An interesting implication of the telomere explanation for nail atrophy is that the velocity at which nails grow should be a non-invasive marker of telomere status. That is, the shorter the telomeres, the more senescent cells there should be in the system, and the more slowly the nail should grow. Moreover, nail growth velocity should be a marker that is not sensitive to the confounding effects of most DNA damage and damage repair. In the next section we elaborate on that idea.
| FINGERNAIL ATROPHY MIRRORS NORMAL AGING |
|---|
|
|
|---|
Orentreich and Sharp (24) demonstrated that fingernail growth velocity changes predictably with age. In one cross-sectional study of 257 human subjects, they found that nail growth velocity increased linearly to a peak of 0.83 mm per week at about 25 years old, then decreased linearly thereafter by 0.5% per year. Orentreich and colleagues followed that study with a twelve-year longitudinal study of 27 beagles (25). Nail growth velocity in the dogs increased linearly until age 3, then decreased by 3% per year through age 15. The beagle result is notable for its parallel to the human data. Beagles live about one fifth the life span of humans; correspondingly, the dogs' nail growth velocity rose and declined 5 times faster than the human decline.
When Orentreich looked at a contemporaneous study of rat nail growth velocity (26), he found something unexpected. The velocity of nail growth did not decline in rats over their three-year life span. Why should rat nails act differently from beagle and human nails? The explanation, so elusive in its own time, is less of a puzzle now. Many species of laboratory rat have very long telomeres (e.g., 27), and telomerase is active in many of their cells. In fact, their telomeres are so long (about four times longer than those of humans) that there is not enough shortening in the rat life span to make a difference in many aging symptoms, including, as Lavelle showed, nail growth. It is difficult to explain why rat nails differ from those of other animals by any means other than the link between nail atrophy and telomere shortening.
Oddly, fingernail growth velocity may be more than a marker of telomere length; it might also be a marker of length of life. About a decade after Orentreich's studies, Williams and colleagues (28) and Short and colleagues (29) reported a relationship between nail growth velocity and aging in pigtailed macaques. The finding inspired Coe and Ershler to measure nail growth a year later in their colony of rhesus monkeys, but that team did not immediately report their result. Thirteen years later, however, while looking at the relationship between natural killer cells and aging, the two researchers recalled the nail data. So much time had passed that 12 of their original monkeys had died and the authors were able to compare their single measurement of nail growth velocity, taken more than a decade earlier, with the number of years beyond the experiment that each animal had survived and with the age of each animal at death (30). Fingernail growth velocity was a surprisingly good predictor, correlating significantly with both years of survival (r =.67, p <.02) and age at death (r =.57, p <.05). Given the other links that have been established between telomeres and mortality (e.g., 31), the relationship of nail growth rate to survival is interesting.
The conclusion that follows from the forgoing discussion is that fingernail growth may be a noninvasive marker of telomere status. If so, it is a telomere marker that is not sensitive to the confounding effects of DNA damage and damage repair problems. Measuring nail growth velocity could provide information about telomere status in a remarkably short time. For example, Orentreich and colleagues (25) described a method for measuring nail growth velocity to accuracies of 0.1 µm over periods as short as 15 minutes using a split-image range finder adapted to a trinocular microscope and a fixed reference object cemented to the nail fold. Such a non-invasive measure could have many uses. For example, the longitudinal nature of caloric restriction studies makes short-term measures of efficacy difficult to obtain. Measurement of nail growth velocity using modern equivalent of the split-image range finder might produce a non-invasive assessment of treatment effects within a few weeks after the start of a caloric restriction experiment. The simpler but slower procedure of marking the nail and measuring nail growth with calipers might also provide a continuous way to monitor telomere changes.
| ALOPECIA |
|---|
|
|
|---|
| HAIR DIAMETER IS REDUCED IN AGING AND IN SHORT-TELOMERE PROGERIAS |
|---|
|
|
|---|
The role of keratinocyte TACs has already been discussed. The role of hair diameter is a little more complex. Hair diameter is determined by the FP: the larger the FP, the thicker the hair shaft (34). The size of the FP is determined at the start of each anagen hair cycle when specialized fibroblasts migrate to the interior of the FP. The more those fibroblasts replicate during this growth period, the larger the final size of the FP (35), and the wider the subsequent hair shaft. Thus, hair diameter changes with each hair cycle.
With age, there is a decline in both the keratinocyte and fibroblast progenitor populations. The decline in keratinocyte progenitors slows hair growth; the decline in fibroblast progenitors and subsequent reduced hair diameter means that fewer keratinocytes are needed to lengthen the hair, so hair grows faster. Because of the simultaneous contribution of these two opposing factors, hair growth velocity varies unpredictably with age.
Although hair diameter and keratinocyte activity interact with regard to hair growth velocity, hair diameter is independent of the number of keratinocyte TACs; increasing the number of active keratinocytes does not increase the diameter of the hair. Hair diameter is, therefore, a measure of the replicative ability of the FP fibroblast progenitors.
On the basis of our results with nails, we expect that with age, shortening telomeres will lead to fewer active fibroblast progenitors in the FP and, therefore, hair diameter will decrease. Studies of the matter confirm this expectation; both the number of active fibroblast progenitors in the FP (36) and hair diameter (e.g., 3740) decrease with age. The effect is, in fact, so strong that it consistently appears despite influences of hormones (which also affect hair diameter) and other factors (40).
Our earlier nail result suggests that hair diameter should decrease in the four short-telomere progerias with replication problems. There are no formal studies of hair diameter in any progeria, but the clinical literature often refers to hair in those four progerias as "fine" (41,42). With the exception of the progeroid form of Ehlers-Danlos Syndrome, in which hair is described as curly and fine, no other progeria is described by that term. (The curliness in Ehlers-Danlos Syndrome suggests that the hair is being flattened as it passes through the follicle.) The single histological study looking at hair in the progerias (an RTS model mouse) (43) found that FPs were significantly reduced in the RTS mice versus controls, thereby confirming the clinical impression that hair in the short-telomere progerias is "fine." Thus, hair diameter, like nail growth velocity, may reflect telomere shortening both in natural aging and in some progerias.
| OSTEOPOROSIS IN SOME PROGERIAS MAY DEPEND ON TELOMERASE |
|---|
|
|
|---|
Bone remodeling is carried out by the Basic Multicellular Unit (BMU). At the front of the BMU is a team of large, multinucleated osteoclasts which dissolve existing bone in a sealed compartment on their ruffled undersides. At the back of the BMU is a trailing group of individual osteoblasts, each rebuilding some of the bone lost during the passage of the osteoclasts (Figure 2C). Osteoporosis occurs when osteoclasts resorb more bone than osteoblasts resupply (reviewed in 44). There are many ways for that to happen, but osteoporosis in old age occurs because fewer BMUs are made and because the osteoblasts secrete less osteoid, the collagen-rich bone-rebuilding material. Thus, there are fewer BMUs working to remodel bone, and the passage of each BMU produces a net loss. The result is a slow, steady erosion of bone and loss of bone microarchitecture.
Although osteoporosis in the progerias may be due to shortened telomeres (see, for example, 45 and 46), some researchers believe that telomerase itself is the problem. Telomerase is required for progenitor cells to differentiate into osteoblasts in mice (47) and possibly in humans as well. If so, then the telomerase problems in Dyskeratosis Congenita would explain osteoporosis in that disease. Problems with differentiation are more likely to affect osteoblasts than osteoclasts because the pool of immediate osteoclast progenitors is 2500 times larger than the pool of osteoblast precursors (48). Osteoblasts are, therefore, unlikely to ever be rate limiting (48). Because osteoblasts regulate entry of osteoclasts into the bone-remodeling system, it is reasonable to expect that fewer BMUs will be produced; the same situation is seen in normal aging.
The foregoing scenario is unlikely to be true for Rothmund-Thomson Syndrome, which has no obvious connection to telomerase, or for Werner Syndrome, which has a pattern of osteoporosis very unlike that seen in normal aging, including a 60%+ incidence, severe involvement of the limbs, and symmetric osteosclerosis of the hands (49,50) and again, no obvious link to telomerase. In HGP the picture is more interesting. It is our observation that symptoms in that disease are almost entirely attributable to the progeny of mesenchymal stem cells in bone marrow (which make adipocytes, myoblasts, chondrocytes, ligament cells, and osteoblasts). It is possible then that HGP is caused by a failure of mesenchymal stem cells to differentiate resulting in fewer osteoblasts, hence fewer BMUs, and osteoporosis. This is an idea that deserves further investigation, especially into (a) the question of telomerase in differentiation of mesenchymal stem cells and (b) the potential role of the mechanical weakening of the inner nuclear wall (5154) that occurs in HGP.
Although there is no obvious connection between short telomeres and osteoporosis, the grouping of this symptom with nail atrophy and alopecia suggests that there is some link, probably not because of a common pathway, but rather, because of a single process that independently affects the telomeres of progenitor cells in each of the three systems. The same seems to be true of gray hair as well.
| GRAY HAIR IS ASSOCIATED WITH SHORT-TELOMERE PROGERIAS |
|---|
|
|
|---|
How might short telomeres lead to gray hair? An analysis of the situation calls for a brief digression into the physiology of hair pigmentation and graying. Hair is pigmented by melanocytes located among the keratinocytes on the FPs. Each time the growth part of the hair cycle begins (anagen), a new population of melanocytes migrates to the FP from a stem cell reservoir located in the outer root sheath at about the midpoint of the hair follicle (59). Each melanocyte supplies pigment to between one and five follicular hair keratinocytes (60) throughout the entire growth period (up to 10 years) (59). That pigment, melanin, is manufactured inside the melanocyte within membrane-bounded organelles called melanosomes. The melanin, still in its melanosome, is transferred to keratinocytes via dendritic connections. Once inside the keratinocyte, the package is degraded and the pigment absorbed.
Gray hair can occur when any part of the pigmentation process is disturbed (for example, if melanogenesis fails, or if transfer of melanosomes to keratinocytes does not go smoothly). There are other ways that gray hair can occur, but oddly, the one cause not listed among them is cellular senescence. Senescence does occur in melanocytes, but senescent melanocytes produce more, not less, melanin (61,62). In skin, for example, moles are thought to be senescent melanocyte clones (62,63). Consequently, we cannot draw the otherwise obvious conclusion that short telomeres cause senescent melanocytes, less melanin, and hence, gray hair.
If telomere-driven senescence does not cause gray hair, then what does? And why are Ataxia Telangiectasia and NBS involved with gray hair when they were not involved with the nail atrophy and alopecia symptoms? The answer may lie with another suspected agent of aging, reactive oxygen species (ROS).
The most common cause of gray hair is a reduction in the active melanocyte population (61). Estimates vary, but after age 30 the number of melanocytes that supply pigment to hair drops between 10% and 20% each decade (60,64), thereby providing the basis for the rule that 50% of the population is 50% gray by age 50 (9). A widely held theory (e.g., 9,64) is that this loss of melanocytes is caused by ROS, and indeed melanogenesis produces hydrogen peroxide and reactive quinone intermediates (65,66) which can cause cross-links, single strand breaks, and other types of DNA damage (67). Existing evidence suggests ROS damage to nuclear and mitochondrial DNA leads to mutations in active melanocytes (59).
The link between ROS theory and the short-telomere progerias lies in the fact that ROSs are preferentially attracted to the GGG sequences at the 5' end on the telomere (68). In particular, hydrogen peroxide preferentially cleaves 5'-RGGG-3' (69). It is possible, therefore, that long telomeres protect against ROS damage occurring elsewhere in the DNA. As telomeres shorten, this protection is lost and stress-induced premature senescence, or apoptosis, results. According to this theory, gray hair is based entirely on the shortness of telomeres, not on why they are short. More specifically, persons with Werner Syndrome or one of the other three replication-based progerias have gray hair because their telomeres shorten due to inability to restore lost repeats, whereas those persons with Ataxia Telangiectasia or NBS have short telomeres because of breakage. But persons with any of the six progerias have short telomeres and hence are vulnerable to gray hair.
Hair melanocytes are particularly vulnerable to these problems because of their increased exposure to ROS and their long life spans. This link between ROSs and short telomeres is attractive, and a recent serendipitous finding supports it. Chronic myeloid leukemia is a hematopoietic stem cell malignancy in which the Abelson oncogene (which encodes a tyrosine kinase) transfers from chromosome 9 to the BCR region of chromosome 22. The fused gene encodes a protein thought responsible for the disease. In 1999, a drug aimed at treating the disease, imatinib mesylate, entered phase 2 clinical trials. Imatinib is a selective tyrosine kinase inhibitor that blocks phosphorylation of an ATP binding site of the fused gene product. But imatinib also inhibits the c-Kit tyrosine kinase receptor, which is on a pathway that activates promoter for the tyrosinase pigmentation gene (70). Mutations in the homologous c-kit gene in mice produce a white coat; mutations in c-Kit in humans produce piebaldism (71). It was, therefore, predicted that blocking the c-Kit receptor with imatinib might have an unwelcome side effectgray hair. The two-year study showed that imatinib was indeed an effective treatment for chronic myeloid leukemia (72) and that it did indeed have a side effect. It repigmented gray hair. The result was published in the New England Journal of Medicine as a serendipitous finding and an interesting puzzle (73).
Shortly thereafter, Brummendorf and colleagues (74) independently reported a second peculiar finding; peripheral blood leukocytes in people who had responded to the imatinib had longer telomeres than did those in people who had not responded. It is possible that selective cloning occurred, but the authors provided convincing evidence against that possibility. It looks very much as if imatinib restored telomere length to migrating melanocyte TACs arriving at the start of a new hair cycle, and that that restoration led to repigmentation of gray hair. An interesting, albeit impractical test of this idea would be the administration of imatinib to people with Ataxia Telangiectasia. Immortalization of Ataxia Telangiectasia cells by ectopic expression of hTERT rescues the short-telomere phenotype without affecting the DNA repair/breakage problem (75). It follows, then, that if imatinib restored pigment to gray hair in Ataxia Telangiectasia, the original cause was likely to have been short telomeres, rather than the breakage problem.
| DO SHORT-TELOMERE PROGEROID SYNDROMES RESEMBLE NORMATIVE AGING? |
|---|
|
|
|---|
|
Differences in the order in which symptoms appear in aging and progeria are also predictable. For example, with the quick shortening of telomeres in progerias, the first hairs to thin should be the ones that cycle most rapidly rather than the ones most vulnerable to hair loss. It follows that in normal aging the vulnerable scalp hair should thin before eyebrows and lashes, but in the progerias, the faster cycling eyebrows and lashes should thin before scalp hair.
The situation with osteoporosis is less clear. Only Dyskeratosis Congenita resembles normal aging in its rate of onset and areas affected (76). Werner Syndrome has a much higher incidence of osteoporosis than is seen in normal aging [estimates range from 60% (49) to almost 100% (50) of all cases], and it includes osteosclerosis (abnormal bone density), a symptom rarely seen in normal aging. Rothmund-Thomson Syndrome has an onset rate more in line with normal aging, but it too includes osteosclerosis and extensive bone remodeling. Osteoporosis in HGP is also different from normal aging, in that the underlying problem is bone development, not bone loss. Thus, although the co-occurrence of osteoporosis in only the short-telomere progerias is interesting, in view of the dissimilarity of the overall picture to normal aging in all but Dyskeratosis Congenita, the relationship between short telomeres and osteoporosis remains conjectural.
Finally, we note that although the short-telomere progerias have very different causes and a wide variety of symptoms, they are very similar in the particular hair loss, graying, and nail symptoms that occur (Table 2). For example, in the four progerias with short telomeres due to replication problems, nails are short and dystrophic with beading and longitudinal ridges, progressing to atrophy. Pictures depicting nail changes in the various progerias (e.g., 76,77) sometimes also show great cracking, splitting, and separation of the nail plate from the nail bed (a consequence of nail atrophy), but no other dystrophies are common. All of these symptoms can be attributed to telomere shortening. Conversely, symptoms that are not attributable to telomeres are not reported. Nail pitting, nail plate softening, spooning, clubbing, claw-like changes, white nails (leukonychia), and the many other types of symptoms that are attributable to problems with keratin, enucleation, nail plate construction, or any processes other than telomere shortening, do not normally occur. The same is true in hair. Thus, although all nail and hair symptoms increase with age, those that we have identified as belonging to the short-telomere progerias are most common, which suggests that the short telomeres are a reasonable model for hair and nail symptoms in aging.
| IF THE PROGERIAS ARE SEGMENTAL, IS AGING MODULAR? |
|---|
|
|
|---|
Our approach of looking at different properties of progerias in different systems works because each progeria operating within a system of the body can be considered a natural experiment. Other natural experiments may be available. The limbus of the cornea and the pregermative zone in the lens are two populations with unique propertiesthe lens is especially interesting because it captures changes in a permanent sediment of cells for which developmental history is known. Cataracts are seen in several progerias but were not included in our analysis because the scenario is too complex and the symptom appears in too many diseases to consider fully. Such cross-system analyses may, however, uncover insights into aging that are not readily available from single-system studies.
| SCOPE OF LITERATURE REVIEW |
|---|
|
|
|---|
We supplemented those searches with numerous medical texts, in part, because diagnosis of fingernail problems is not standardized and many sources had to be consulted to get an understanding of the range and degree of the nail problems for each progeria. Major medical texts (1,7984) and one research monograph (85) were included. Other texts were consulted on a limited basis. We obtained additional information on systems, symptoms, and diseases from several dozen major review articles and numerous original research publications. We regret that, in many cases, overlapping information and space limitation forced us to choose which of several equally incisive articles to cite.
| Acknowledgments |
|---|
A. C. Hofer, R. T. Tran, and O. Z. Aziz contributed equally to this article.
Figure 2 was drawn by Susanna Douglas, Electronic Specialist in the Department of Psychology at The University of Texas at Austin.
| Footnotes |
|---|
Received June 21, 2004
Accepted August 26, 2004
| References |
|---|
|
|
|---|