Progeroid syndromes (PSs) constitute a group of disorders characterized by clinical features mimicking physiological aging at an early age. In some of these syndromes, biological hallmarks of aging are also present, whereas in others, a link with physiological aging, if any, remains to be elucidated. These syndromes are clinically and genetically heterogeneous and most of them, including Werner syndrome and Hutchinson–Gilford progeria, are known as ‘segmental aging syndromes’, as they do not feature all aspects usually associated to physiological aging. However, all the characterized PSs enter in the field of rare monogenic disorders and several causative genes have been identified. These can be separated in subcategories corresponding to (i) genes encoding DNA repair factors, in particular, DNA helicases, and (ii) genes affecting the structure or post-translational maturation of lamin A, a major nuclear component. In addition, several animal models featuring premature aging have abnormal mitochondrial function or signal transduction between membrane receptors, nuclear regulatory proteins and mitochondria: no human pathological counterpart of these alterations has been found to date. In recent years, identification of mutations and their functional characterization have helped to unravel the cellular processes associated to segmental PSs. Recently, several studies allowed to establish a functional link between DNA repair and A-type lamins-associated syndromes, evidencing a relation between these syndromes, physiological aging and cancer. Here, we review recent data on molecular and cellular bases of PSs and discuss the mechanisms involved, with a special emphasis on lamin A-associated progeria and related disorders, for which therapeutic approaches have started to be developed.
Progeroid syndromes (PSs) are rare genetic disorders mimicking clinical and molecular features of aging. The interest in exploring age-related hereditary disorders is associated both to their severity, leading in many cases to life span shortening, and to the expectation that identifying causative genes could help understanding some of the mechanisms underlying aging. To date, most PSs, for which genes and pathophysiological mechanisms have been identified, are monogenic and fall into the category of segmental PS (1). Indeed, all the clinical and biological changes usually observed during natural aging are never totally recapitulated in PS. In this context, Werner syndrome (WS) (no. 277700) and Hutchinson–Gilford progeria syndrome (HGPS) (no. 176670) have been the most extensively studied, as they closely mimick natural aging.
Pathophysiologically, the known PSs are caused, basically, either by mutations in genes encoding DNA repair proteins, such as in WS, Bloom syndrome (BS), Rothmund–Thomson syndrome (RTS), Cockayne syndrome (CS) syndrome, xeroderma pigmentosum (XP) or trichothiodystrophy (TTD) (2,3), or by mutations in genes encoding lamins A/C or partners involved in their biological pathway, such as HGPS or restrictive dermopathy (RD) (4–7). However, in premature aging animal models, mitochondrial pathways have shown to be affected. This is of note, as the cause of other premature aging syndromes, as Hallerman Streiff or Wiedeman Rautenstrauch, remains to be elucidated.
Here, we will review the main categories of known PS with respect to their pathophysiological mechanisms, particularly emphasizing the most recently identified syndromes linked to the accumulation of mis-processed lamin A precursors in progeria and related disorders. Additionally, recent investigations will be discussed, which have succeeded in establishing functional relationships between DNA repair defects, lamin-associated PS, physiological senescence and cancer.
PSs AND DNA REPAIR DEFECTS
DNA repair is a complex mechanism dedicated to repair DNA damage, including single- and double-strand lesions, caused by exogenous or endogenous damaging agents. It is well documented that DNA repair defects may lead to age-related diseases such as cancer or cognitive impairment (8,9). In the last 10 years, explorations of premature aging syndromes have shed light on genomic instability and, more generally, DNA repair deficiency, as a key mechanism in several of these disorders, including typical WS. Besides, most DNA repair-associated phenotypes include a precocious senile appearance together with predisposition to neoplasias or CNS disorder (8–10).
Mutations in two major classes of proteins have been identified in DNA repair-associated PS: RecQ protein-like helicases (RECQLs) and nuclear excision repair (NER) proteins. We will review the literature data on WS as a model of RECQL-associated PS and on CS as an example of NER-associated PS.
RecQ helicases and WS
RecQ helicases are a family of conserved enzymes, initially observed in Escherichia coli (11), playing major roles in maintaining genome integrity and suppressing deleterious recombination events. In human, five RECQL genes have been characterized. Defects in RECQL2/WRN, RECQL3/BLM and RECQL4 are, respectively, associated to WS, BS and RTS, rare autosomal recessive PS whose clinical features have been extensively reported (8,12). Several features are shared by these three syndromes, including genomic instability and predisposition to cancer, and their associated cellular defects are distinct (13). WS has been widely explored as a model system of aging because of its close resemblance to natural aging. It was first identified by Werner in 1904 (14), reporting four siblings with premature aging. WS is a rare AR disease with about 1300 patients reported in the literature. The mean age at diagnosis is 24 and growth retardation results in short stature in adults. Clinical features include bilateral cataract, prematurely aged face with beaked nose, premature arteriosclerosis, skin atrophy with scleroderma-like lesions and sparse gray hair, lipodystrophy, type 2 diabetes and osteoporosis. Patients affected with WS are prone to develop malignancies. Patients typically die at a mean age of 47, because of cardiovascular disease or cancer. Mutations in WRN helicase, encoded by RECQL2, underlie this disorder. WRN is involved in several DNA repair pathways: double-strand break repair whose main known mechanisms include homologous recombination and non-homologous end joining (15) and single-strand break repair with NER being the main mechanism (16,17). Although the WRN function is only partially elucidated, its substrate specificity and interactions are relevant to WS pathophysiology. Indeed, WRN interacts with a recombination mediator protein (18) and part of the DNA/protein kinase complex regulating its exonuclease activity (19). Additionally, WRN interacts with components of the DNA replication complex (16,20–22). Its association with p53 suggests a further distinct role, discussed subsequently (23). Moreover, WRN is associated to telomeres, where it is recruited by a telomeric repeat binding factor essential for correct telomere maintenance and repair (24), activating WRN exonuclease activity (25). Alterations in DNA repair, replication and stability thus seem relevant to the molecular pathophysiology of WS.
Mutations observed in WS patients are located all along the RECQL2 coding sequence (26); most are predicted to cause the production of truncated proteins lacking the nuclear localization signal, thus preventing WRN from fulfilling its function in the nucleus. Furthermore, in several cases, mutated mRNAs have shown to have a shorter life span than wild-type ones (26). So, in patients affected with WS, mutations mostly lead to WRN loss of function, consistently with the rather mild phenotypic variations observed. A direct interaction between WRN and the BLM helicase, responsible for BS, was also identified (27). Furthermore, p53-dependent apoptosis was shown to be strongly reduced in both WS and BS cells (28,29) (Fig. 1).
NER proteins and CS
Another group of PS, more clinically heterogeneous, is linked to defects in the NER pathway. They include complex developmental defects and neurodegeneration, with or without cancer. Two main mechanisms initiate NER pathway: transcription-coupled repair (TCR) and global genomic repair, both leading to a process in which DNA is unwound, its damages are excised, replaced by a normal sequence and finally replicated (30–32). Alterations in any of these steps may lead to human aging disorders, including XP, CS and TTD, in distinct or combined forms. Several reviews detailed these complex syndromes (9,12,33). Recently, Cleaver (34) exposed new views on how DNA repair/replication deficiencies in XP involve most of the genome, whereas defects in CS are confined to actively transcribed genes.
CS is a rare autosomal recessive PS, classified as CSA (no. 216400) or CSB (no. 133540) according to the mutant gene involved. CSA is caused by mutations in the excision-repair cross-complementing gene 8 (ERCC8) encoding the CSA protein (35), whereas CSB is caused by mutations in ERCC6 encoding the CSB protein (36). Both are involved in TCR of DNA damage: by ubiquitinating RNA-polymerase II, they allow to carry out transcription (37–39).
CS clinical features include severe growth retardation with lipoatrophy, atrophic skin, sparse hair, neurodevelopmental abnormalities with calcium deposits, microcephaly, cataract, sensorineural hearing loss and sun-sensitivity, without significant cancer susceptibility (40). CSA age of onset (milder form) is 1–3 years with 20–40 years life span, whereas CSB symptoms (severe form) appear at birth, life span being 6–7 years. Death is usually due to severe nervous system deterioration and respiratory tract infections (12). CS can also be an allelic disorder to XP, both due to XPG (XP group-G/ERCC5) mutations (41). Although no obvious genotype–phenotype correlations have been observed, the comparison between these clinical forms can be very helpful. Nouspikel et al. (41) observed that in CS, severely truncated XPG proteins were produced, whereas in typical XP, the protein carried a missense mutation impairing its exonuclease function. XPG truncation thus probably abolished its enzymatic function and also a further specific interaction, leading to the severer CS phenotype.
Lamin A-related PSs
In 2003, we and others identified mutations in the LMNA gene as causing HGPS, one of the most emblematic segmental aging disorders (5,7). These studies revealed an unexpected role of lamins and global nuclear architecture in premature aging disorders and raised the crucial question of potential links between lamina, nuclear envelope associated proteins and natural aging (43,44).
Hutchinson–Gilford progeria syndrome
First described by Jonathan Hutchinson (45) and later by Hastings Gilford (46), progeria (HGPS) is an extremely rare, severe and fatal developmental disorder characterized by precocious onset of pathologies which are typical of advanced age. The clinical phenotype is characterized by severe growth retardation, usually associated to skeletal alterations (osteolyses, osteoporosis), marked amyotrophy, lipodystrophy, skin atrophy with sclerodermatous focal lesions (47) and alopecia. Affected children present with extremely severe atherosclerosis. The cognitive functions are fully preserved. As well, cancer incidence is not increased and some developmental processes are delayed (dentiton) or absent. Death occurs at a mean age of 13.5 years, mostly due to myocardial infarction [for clinical review, see Hennekam (48)].
Most typical HGPS cases are due to a recurrent, de novo, dominant point mutation (5,7) predicted to be a synonymous mutation (p.G608G). It is located in the part of the gene specifically encoding lamin A. In most mutated pre-mRNAs, a cryptic splice site inside exon 11 gains the upper hand on the normal consensus site, causing the in-frame deletion of the last 150 bp of the exon. These transcripts are translated into a truncated prelamin A precursor, also called progerin or LaminAΔ50 detected in WB experiments (6,7).
This protein lacks the second cleavage site recognized by the metalloprotease ZMPSTE24 (FACE1) during prelamin A post-translational processing: it cannot undergo complete maturation and maintains a farnesyl and a methyl moiety on its C-terminal cysteine residue. Lamins A and B maturation steps are summarized in Fig. 2.
Indirect immunofluorescence experiments with antibodies directed against lamins A/C or some of their molecular partners allow to observe nuclear blebs and herniations of the nuclear envelope, thickening of the nuclear lamina, loss of peripheral heterochromatin and clustering of nuclear pores (5,7,49) (Fig. 1). Alterations of nuclear morphology and composition increase with passages in cell culture and are correlated with an apparent intranuclear accumulation of progerin (49,50). Some of these nuclear alterations, in different extents and percentages of cells, are observed as well in other laminopathies such as FPLD, EDMD and MAD-A (51–53).
At least 16 other lamin A/C mutations have been reported as causing progeroid phenotypes, reviewed in (44) and collected in the free access UMD-LMNA database (www.umd.be:2000). Very recently, two new mutations at the composite heterozygous state were published by Verstraeten et al. (54). Most mutations are localized in the lamin A-specific C-terminal globular domain and in the N-terminal region, but only three have been reported to specifically alter lamin A splicing, leading to the production of truncated protein products (p.G608G, p.T623S and IVS11+1G>A) (6,7,55).
In 2003, Agarwal et al. (56) observed for the first time the involvement of ZMPSTE24/FACE-1 in severe mandibuloacral dysplasia (MAD-B), a milder PS.
In 2004 and then in 2005, a perinatal lethal genodermatosis, RD, mainly characterized by intrauterine growth retardation, tight and rigid skin, prominent superficial vessels, micrognathism, bone mineralization defects and multiple joint contractures [for clinical details, see Nijsten et al. (57), Smitt et al. (58) and Wesche et al. (59)], was linked to primary or secondary lamin A dysfunction (6).
One patient presenting with RD at birth and living up to 6 months of age, a much longer life span than is typical in RD patients, carried an heterozygous c.1968+1G>A (IVS11+1G>A) transversion, leading to the in-frame skipping of the whole exon 11 in transcripts and the production of a truncated protein (LaminAΔ90), detected in western blot studies (6).
A second patient, affected with another milder form of RD, carried the p.G608G mutation identified in most typical HGPS patients.
The functional consequences of LMNA-linked RD mutations are similar to HGPS, causing the appearance of deleted prelamin A precursors, which cannot undergo full post-translational maturation. These precursors accumulate inside patients' nuclei, where they exert toxic effects.
A similar pathophysiological mechanism is observed ZMPSTE24/FACE1-linked RD: in 10 reported cases, the absence of this fundamental prelamin A processing enzyme causes an accumulation of normal-length precursors. The commonest mutation was a frameshifting insertion (c.1085–1086insT) in exon 9, leading to a premature termination codon (p.Leu362PhefsX19), either at the homozygous or compound heterozygous state with another ‘null’ mutation on the opposite allele (4).
Western blot studies showed in six different patients that no ZMPSTE24 was detectable in different tissues, whereas untransformed prelamin A precursors, weighing ∼74 kDa, were evidenced, together with normal lamin C. As expected, mature lamin A was completely absent. RT–PCR studies of ZMPSTE24 transcripts were always possible, indicating that the genomic alterations did not affect transcript production or stability: the mutated proteins may be either not produced or rapidly degraded. Of note, the commonest c.1085_1086insT insertion, transmitted from parents in most cases, seems to occur in a mutational hotspot (4,6,56,60). Indeed, the thymine stretch located in ZMPSTE24 exon 9 (c.1077–c.1085) had already been reported as a specific target of microsatellite instability in colorectal tumors (61).
Immunocytochemical analyses performed with different antibodies directed against lamins and partners showed numerous and major nuclear alterations of morphology and composition in cells issued from patients affected with RD.
Pathophysiological mechanisms involved in PSs linked to A-type lamins defects
Mice inactivated for Zmpste24 (62,63) show striking phenotypic similarities with patients affected with lamin-related PSs, although a more severe phenotype is observed in humans. However, the involved pathophysiological mechanisms are presumably similar, involving the aberrant intranuclear accumulation of prelamin A precursors and the presence of nuclear abnormalities.
Heterozygous mice are indistinguishable from wild-type mice, indicating that half ZMPSTE24 gene dosage is sufficient to guarantee prelamin A processing, as observed in the parents of patients affected with RD (4,6). In this context, and as an additional proof of the toxic effects of prelamin A accumulation, it has been demonstrated that reducing prelamin A synthesis by half may spectacularly improve the phenotype in double knock-out mice carrying the Zmpste24−/− Lmna+/− genotype (64,65).
Conclusion from these studies and observations of data from human patients indicate the following. (i) A toxic prelamin A threshold exists, which, if trespassed, leads to the development of progeroid phenotypes; in this context, ZMPSTE24/FACE1-linked RD phenotypes represent one extreme form, with total enzyme inactivation and extreme prelamin A accumulation, whereas several MAD phenotypes resulting from partial ZMPSTE24/FACE1 inactivation represent milder forms (56,60). (ii) Interpersonal variations of the efficacy of splicing (i.e. mutated transcripts’ aberrant or correct splicing) and, consequently, of global amounts of progerin produced might be one factor modulating the phenotype severity in LMNA p.G608G patients [i.e. variation in the age of onset (nenonatal to 2 years) and of death (median of 13.5–45 years)].
A novel and fascinating view on the pathomechanism underlying prelamin A associated syndromes has been raised by the recent discovery that prelamin A accumulation impairs the recruitment of DNA repair factors at damage sites (66). An outstanding study from Varela et al. (65) demonstrated that accumulation of prelamin A in Zmpste24 deficient nuclei led to the activation of the p53 signaling pathway: many essential p53 targets are over-expressed, causing a senescent cellular phenotype and suggesting a key pathophysiological mechanism. As a proof, double knockout Zmpste24−/−, p53−/− mice showed a partially restored phenotype (65). This work correlated with the well-documented role of p53 in senescence and normal aging (67–71), indicating a potential link with pathophysiological mechanisms involved in natural aging. In this direction, recent studies have shown that LaminAΔ50 accumulates as well in nuclei of old donors, suggesting that the aberrant HGPS splice site is also used as normal cells age, providing a link between progeria and natural aging (72).
Lamins exert important metabolic and mechanical functions (44) not only in interphase nuclei but also during mitosis, together with nuclear matrix partners as NuMA (73–75), namely regarding mitotic spindle (76) and centrosome organization. These observations probably explain the perturbation of cell proliferation observed in diseases involving lamins.
Some HIV protease inhibitors also lead to an accelerated aging phenotype through side effects on cellular proteases: inhibition of ZMPSTE24 (77,78), inhibition of proteasomal (79,80) and mitochondrial AAA proteases (81) involved in mitochondrial importation of nuclear-encoded proteins. Nucleosides analogs also affected mitochondrial DNA and RNA (82).
Therapeutic approaches in progeria and related disorders
One therapeutic strategy focusses on restoring a correct lamin/prelamin ratio, by targeting the splicing defect observed in pogeria. To this end, pogerin levels have been experimentally reduced in cell culture by transfection of oligonucleotides targeted to the aberrant splicing site, leading to a spectacular reversion of the cellular phenotype (83).
A completely different therapeutic approach aims to modify the chemical composition of progerin, based on the hypothesis that its maintained farnesyl moiety mediates its toxicity: farnesyl-transferase inhibitors (FTIs) rescued nuclear morphology in cells issued from patients affected with RD or HGPS (84–89) or mouse models (87). Although studies present interesting potential therapeutic approaches, open questions remain to be solved, concerning a possible alternative prenylation of the truncated/wild-type lamin precursors, sidestepping the treatment with FTIs as in the Ras pathway, i.e. through geranyl-geranylation (90). Otherwise, FTIs seem to inhibit proteasome pathway leading to several secondary defects (91). In this respect, some other chemical substances with potential benefit should be considered, as statins, widely used as anticholesterol agents, which allow inhibition of lamin A maturation (92–94), or amino-bisphosphonates, currently used to treat osteoporosis and also known to inhibit cholesterol synthesis (95–97). Both act in the biological pathway leading to farnesyl group production and their particular action is summarized in Fig. 3. A last alternative therapeutic approach could be oriented to inhibiting p53 downstream targets (98).
MITOCHONDRIA, MISCELLANEOUS GENES AND AGING
Several animal models of premature aging or life span extension (99–101) opened new research fields and underscored the reciprocal relationships linking plasma membrane, nucleus and mitochondria. Although some of these models do not have human genetics diseases counterparts, they could explain some events observed in human PS as well as in age-related disorders.
The three main components of the mitochondrial theory of aging are (102–104): (i) increased production of reactive oxygen species (ROS) by complexes I and II (105), (ii) accumulation of mitochondrial DNA (mtDNA) damages and (iii) progressive respiratory chain dysfunction (100–103).
mtDNA is more sensitive to ROS-induced damage than nuclear DNA (106) probably because of the mitochondrial DNA polymerase, the only known enzyme involved in mitochondrial DNA replication and repair. mtDNA is thus a key target of oxidative damage (107). Two mice models expressing proof-reading defective DNA polymerase exhibit premature aging associated with mtDNA mutations (108,109). Aging in man is accompanied by a decline in mitochondrial function (110), mainly affecting the respiratory chain (111) and by high levels of mtDNA deletions in brain neurons during neurodegenerative diseases (112,113). The mitochondrial toxicity of ROS can be abolished by overexpression of a peroxisomal catalase targeted to mitochondria in a transgenic mice model, whose life span was increased (114). Signaling from mitochondria to the nucleus, during aging or ROS-induced diseases (115), is achieved by p32, a nuclear-encoded mitochondrial matrix protein imported into the nucleoplasm (116,117) where it controls mRNA transcription (118) and splicing (119). Interestingly, p32 also interacts with lamin B receptor, a major nuclear envelope component (120,121).
The backward signaling pathway, from nucleus to mitochondria, includes about 1500 nuclear-encoded mitochondrial proteins (122) (123). p53, already associated with premature aging in a mouse model (124), is involved in mitochondrial-induced apoptosis (125,126), controls mitochondrial functions both directly, through mtDNA, and indirectly, through nuclear genes’ regulation. p53 maintains mtDNA genetic stability through its interaction with mtDNA polymerase (127), controls mitochondrial respiration (128) and elicits both pro- and anti-oxidant effects through the transcriptional regulation of related nuclear genes (129,130).
Caloric restriction (CR) is another well-known circumstance of increasing rodent life span (131). It leads to the activation of sirtuin deacetylases, targetting both histones (with the lock of chromatin as a consequence) (132) and several other nuclear (DNA repair factors, transcription factors, p53, etc.), cytosolic (tubulin) and mitochondrial proteins (133–135). CR also promotes mitochondria biogenesis by inducing the expression of eNOS (136,137), thus resulting in a decreased production of ROS (138). Finally, CR inhibits both insulin and IGF signaling pathways (139,140) known to be involved in aging (141), as does the hormone klotho (142,143), whose mutations induce aging in mice (144). The adaptor protein p66shc is one of the signaling links between plasma membrane growth factor receptors, nucleus and mitochondria (145). p66shc−/− mice exhibit increased resistance to stress and increased life span (146). p66shc is a downstream target of p53 (147), localizes to mitochondria where it controls ROS production and metabolism (148,149). The deletion of p66shc was shown to inhibit cardiac stem cell aging and prevent heart failure induced by diabetes (150).
A first link can be argued, between the RecQL proteins family, NER proteins and ‘structural’ lamins, leading to nuclear abnormalities in PS. All PSs, whichever their genetic cause, seem to share defective/delayed DNA repair mechanics due to primary (DNA-repair genes-related syndromes) or secondary (LMNA/ZMPSTE24-related syndromes) alterations. However, these PSs also display fundamental differences. In DNA repair-related syndromes, a predisposition to cancer (or CNS disorders) is evidenced: a down-regulation of the p53 pathway having been observed in WS and BS cells, whereas in LMNA-related PS, there is no association between genomic instability and cancer susceptibility or CNS affection, and, oppositely, p53 pathway seems to be activated and associated with a senescent phenotype (10).
A second interesting point is the link between PS and natural aging. In LMNA-related PS, a global chromatin disorganization with heterochromatin loss was observed (49,151). Interestingly, interstitial chromatin alterations have been shown to cause persistent p53 activation and are involved in the radiation-induced senescence-like growth arrest (152), and heterochromatin loss has been observed concomitantly with nuclear shape changes during Caenorhabditis elegans aging (153).
Finally, various yeast or animal models of accelerated aging or of extended life span evidenced the role of proteins already involved in PS, acting in reciprocal interactions between nucleus, mitochondria and signaling pathways starting from plasma membrane receptors. Recently, much progress has been made in understanding pathomechanisms associated to human PS as well as in conceptualize potential therapeutic approaches in progeria and related disorders, pre-clinical and clinical trials will now be necessary to confirm the legitimacy of the hope raised by recent researches in this field.
NOTE ADDED IN PROOF
While this article was in press, Rusinol and Sinensky published an important article relating the mechanisms of prenylation in the context of their associated progeroid syndromes: Rusinol AE, Sinensky MS (2006) Farnesylated lamins, progeroid syndromes and farnesyl transferase inhibitors. J Cell Sci.119, 3265–3272.
We warmly acknowledge Dr Annachiara De Sandre-Giovannoli for critical reading, discussions and comments on this review. Our studies on lamins, progeria and restrictive dermopathy are supported by the Association Française contre les Myopathies (AFM) and the Institut National de la Santé et de la Recherche Médicale (INSERM). C.L.N. is supported by a fellowship grant from the AFM.
Conflict of Interest statement. None declared.