Pin1 and telomere maintenance

Telomeres cap the ends of linear chromosomes and have many important functions, including prevention of degradation, fusion and recombination of chromosome ends and blocking the activation of DNA-damage checkpoints (Refs Reference Blackburn32, Reference Smogorzewska and de Lange33). Telomeres in most human somatic cells shorten during each cell division as a result of very low or absent activity of telomerase, a unique reverse transcriptase that extends telomeric sequences. Telomere shortening to a critical length and uncapping of the telomere lead to replicative senescence or crisis (Refs Reference Goldstein34, Reference Chin35). Indeed, telomere shortening is related to numerous age-associated diseases, including cancer, AD, hypertension, atherosclerosis and heart failure (or cardiovascular disease). Moreover, telomerase knockout in mice eventually causes telomere loss, limits cell proliferation and triggers premature ageing after four or five consecutive generations (Refs Reference Blasco36, Reference Lee37, Reference Rudolph38). In addition, telomere-associated proteins have been shown to have a central role in maintaining telomere homeostasis in human cells (Ref. Reference de Lange39). A key regulator in maintaining telomeres at reasonable length is the telomeric DNA-binding protein TRF1. TRF1 has been proposed to act in cis to progressively inhibit telomere elongation by telomerase because increased numbers of TRF1 molecules bind to telomeres (Refs Reference de Lange39, Reference van Steensel and de Lange40, Reference van Steensel, Smogorzewska and de Lange41, Reference Ancelin42). TRF1 is tightly regulated during the cell cycle and is an important signalling molecule in mitotic regulation (Refs Reference Shen43, Reference Kishi and Lu44, Reference Kishi45, Reference Kishi46). It has been shown that the F-box protein Fbx4 targets TRF1 and promotes its ubiquitin-mediated degradation (Refs Reference Lee47, Reference Chang, Dynek and Smith48).

TRF1 was also identified as Pin2 in a combined screen for mitotic regulators. A recent study demonstrates an essential role for Pin1 in the regulation of TRF1/Pin2 stability, telomere maintenance and ageing (Ref. Reference Lee20). Pin1 binds to TRF1 in vitro and in vivo after it has been phosphorylated on the conserved Thr149-Pro motif, only in mitotic cells, not in interphase cells (Ref. Reference Lee20). Furthermore, inhibition of endogenous Pin1 function by several approaches renders endogenous or expressed TRF1 fully resistant to protein degradation, increases TRF1 telomere binding and also causes gradual telomere shortening. However, inhibition of Pin1 fails to cause telomere shortening in cells where TRF1 is also knocked down, because telomere elongation is not suppressed by telomere-bound TRF1. Moreover, knockdown of TRF1 results in telomere elongation in mouse cells, and Pin1-deficient mice show elevated levels of TRF1, telomere loss and various premature ageing phenotypes (Ref. Reference Lee20). Within one generation, telomere loss and premature ageing phenotypes in Pin1-knockout mice (Ref. Reference Lee20) are similar to those in telomerase-knockout mice after four or five generations (Refs Reference Blasco36, Reference Lee37, Reference Rudolph38). These results demonstrate the essential and conserved role of Pin1 in the regulation of TRF1 protein stability and maintenance of telomeres, deregulation of which contributes to premature ageing. Given the well-known role of Pin1 in growth signalling, Pin1-catalysed prolyl isomerisation might help in coordinating cell growth and telomere maintenance.

Pin1 and oxidative stress

Oxidative stresses can occur when a balance between the formation and removal of reactive oxygen species (ROS) is disturbed. ROS are implicated in age-related diseases such as cancer and neurodegeneration (Refs Reference Calabrese49, Reference Bader and Grune50). A recent study shows that Pin1 negatively regulates expression of the cyclin-dependent kinase inhibitor p27 (CDKN1B) through binding to phosphorylated FOXO4, which controls mitochondrial and oxidative stress, thus contributing to the maintenance of proper cellular functions (Ref. Reference Brenkman51). FOXO4 is a member of the Forkhead boxO (FOXO) family. FOXO transcription factors have important roles in the suppression of ageing and tumours (Ref. Reference Greer and Brunet52). The FOXO orthologues DAF-16 and dFOXO also promote longevity in Caenorhabditis elegans and Drosophila (Ref. Reference Greer and Brunet52). Oxidative stress can induce binding of Pin1 and phosphorylated FOXO4 and inhibit its transcriptional activity by preventing its nuclear accumulation (Ref. Reference Brenkman51). One of the functional consequences of the association of Pin1 and FOXO4 is to negatively regulate expression of p27 (Ref. Reference Brenkman51). In addition, oxidative damage can induce binding to Pin1 and the 66 kDa isoform of the growth factor adaptor Shc (p66^Shc/SHC1; Src homology 2 domain-containing transforming protein 1) and mitochondrial import of p66^Shc (Ref. Reference Pinton53). p66^Shc is implicated in generation of ROS within mitochondria and translation of oxidative signals into cell death, indicating that it is involved in regulation of lifespan (Ref. Reference Migliaccio54). It is phosphorylated by protein kinase Cβ (PKCβ) in oxidative conditions and accumulates in mitochondria after it is recognised by Pin1, suggesting the involvement of Pin1 in oxidative stress and ageing (Ref. Reference Pinton53). Because FOXO4 is phosphorylated on several Ser/Thr-Pro sites and PKCβ is not a Pro-directed kinase, the Pin1 targeting sites in FOXO4 and p66^Shc remain to be determined.

Pin1 is also modified by oxidation, which causes its inactivation in the early stages of AD, suggesting that Pin1 has an important role in the response to oxidative stress (Refs Reference Lu and Zhou2, Reference Butterfield55, Reference Sultana56). A significant increase in Pin1 oxidation has been observed in the hippocampi of patients with mild cognitive impairment (MCI), and enzymatic activity of Pin1 is correspondingly decreased. Because MCI is an intermediate stage between normal cognitive ageing and early dementia or clinically probable AD, Pin1 is likely to be involved in the initial stages of development of AD, possibly through tangle formation and alteration of the cell cycle (Ref. Reference Butterfield55). Indeed, Pin1 is subjected to oxidation in the hippocampus of AD patients, which appears to decrease its isomerase activity (Ref. Reference Sultana56). However, identification of the oxidation sites in Pin1 remains a major challenge. Recently, other members of the PPIase family – cyclophilins – have been shown to secrete a factor that is induced by oxidative stress, suggesting that phosphorylation-independent prolyl cis–trans isomerisation has an important role in regulating oxidative stress (Ref. Reference Satoh, Nigro and Berk57).

Pin1 and the p53 family

The tumour suppressor p53 gene (TP53) family has an important role in maintaining genomic integrity, including cell cycle checkpoint and transcriptional activation (Refs Reference Slee, O'Connor and Lu58, Reference Harris and Levine59, Reference Toledo and Wahl60). Loss of p53 function increases genomic instability and transformation, and TP53 is mutated or deleted in more than 50% of human cancers (Ref. Reference Lutzker and Levine61). Pin1 has been shown to regulate the function of the p53 family of proteins in response to DNA damage (Refs Reference Zacchi62, Reference Zheng63, Reference Wulf64). Under normal conditions, p53 is targeted by MDM2 and degraded by the ubiquitin-mediated proteasome pathway. Pin1 can bind to DNA-damage-induced phosphorylation of p53 and prevent its binding to MDM2, which increases p53 protein stability (Refs Reference Zacchi62, Reference Zheng63, Reference Wulf64). In addition, Pin1 is recruited to chromatin by p53 and promotes p53 binding and acetylation by p300 (Ref. Reference Mantovani65). Pin1 is also required for dissociation of p53 from the relA-associated apoptosis inhibitor IASPP (PPP1R13L) (Ref. Reference Mantovani65), suggesting that Pin1 regulates p53-dependent transactivation in several ways, by affecting protein stability, protein activity and protein–protein interaction. The tumour suppressor p73 (TP73) also belongs to the p53 gene family and shows strong structural and functional overlaps with p53 (Refs Reference Melino, De Laurenzi and Vousden66, Reference Yang67). Pin1 is required for p73 protein stability because it promotes acetylation by p300 on DNA damage, and enhances the DNA-binding activity and transcriptional activity of p73 towards its target genes, including TP21 (Ref. Reference Mantovani68). Because Pin1 localises with p73 under normal conditions (Ref. Reference Mantovani68), it remains to be determined how Pin1 regulates p73 binding and function in both normal and stressed conditions.

Pin1 and its regulation in cancer

The connection between Pin1 and cancer was first suggested when Pin1 was found to be overexpressed in human cancer tissues (Refs Reference Wulf69, Reference Ryo70). These observations have been subsequently confirmed by demonstrations that Pin1 is significantly overexpressed in 38 out of 60 different human tumour types compared with expression in corresponding normal tissue (Ref. Reference Bao71). Pin1 is overexpressed in most prevalently encountered cancers, including breast, prostate, lung and colon cancer (Ref. Reference Bao71). Furthermore, increased Pin1 levels are positively correlated with a higher risk of, and a shorter period to, tumour recurrence following radical prostatectomy in a set of primary prostate cancers (Refs Reference Ayala72, Reference Sasaki73). These results suggest that Pin1 overexpression is a specific event and a potential prognostic marker in human cancers. Consistent with its overexpression in cancers, with a few exceptions such as Myc and G1/S-specific cyclin E1 (CCNE1), Pin1 activates a number of oncogenes or growth enhancers and also inactivates many tumour suppressors or growth inhibitors (Table 1; Fig. 2). In addition to its overexpression in cancer, Pin1 is predominantly phosphorylated in normal tissues and cells, but becomes hypophosphorylated in cancer tissues and cells (Refs Reference Wulf69, Reference Wiegand105), suggesting that Pin1 phosphorylation might have an important regulatory function in cancer. Indeed, death-associated protein kinase 1 (DAPK1), a known tumour suppressor, phosphorylates the Pin1 PPIase domain and inhibits its catalytic activity (Ref. Reference Lee106). Moreover, DAPK1 inhibits the ability of Pin1 to induce centrosome amplification, chromosome instability and cell transformation (Ref. Reference Lee106), illustrating an essential role of the catalytic activity for Pin1 function in cancer. Thus, Pin1 is a pivotal catalyst for tumourigenesis that acts by switching on or off numerous oncogenes or tumour suppressors, respectively, at several steps at the same time.

Figure 2.

Pin1 promotes oncogenesis by activating numerous oncogenes or growth enhancers and by inactivating tumour suppressors or growth inhibitors. Pin1 promotes oncogenesis by activating numerous oncogenes or growth enhancers and by inactivating a large number of tumour suppressors or growth inhibitors. Pin1-catalysed cis-to-trans prolyl isomerisation regulates its protein substrate functions, such as catalytic activity levels, phosphorylation status, protein interaction, subcellular localisation and protein stability. Thus Pin1 amplifies oncogenic pathways by acting on several targets through the positive- and negative-feedback mechanisms. The asterisk indicates that the result is based on the ability of knockout of Pin1 to prevent cancers in Tp53-null mice.

Table 1.

Pin1-interacting proteins that promote tumourigenesis

^aBased on the ability of Pin1 knockout to prevent cancers in Tp53-null mice.

Abbreviations: AIB1/SRC3, steroid receptor coactivator 3; Bax, Bcl-2-associated X protein; Btk, Bruton tyrosine kinase; C/EBP, CCAAT-enhancer-binding protein; Daxx, death-associated protein 6; FOXO4, Forkhead box protein; GRK2, G-protein-coupled receptor kinase 2; Hbx, hepatitis B virus X-protein; Mcl-1, myeloid cell leukaemia-1; NF-κB, nuclear factor-kappa B; p70S6K, ribosomal S6 kinases; Plk, polo-like kinase 1; PML, promyelocytic leukaemia protein; RARα, retinoic acid receptor alpha; SMRT, silencing mediator for retinoic acid and thyroid hormone receptor; STAT3, signal transducer and activator of transcription-3; Tax, T-cell leukaemia virus type 1; TRF1, telomeric repeat-binding factor 1.

Pin1 and oncogenic signalling pathways

Pin1 also increases cyclin D1 (CCND1) gene expression and protein stability by several mechanisms (Refs Reference Liou28, Reference Wulf69, Reference Ryo70, Reference Ryo89). First, Pin1 interacts with the pSer63/73-Pro motifs in Jun by the activation of Jun N-terminal kinases (JNKs) in response to growth-stimulating conditions and positively regulates Jun transcriptional activity towards its target genes, which include cyclin D1 (Ref. Reference Wulf69). Second, Pin1 directly binds to β-catenin phosphorylated on the Ser246-Pro motif and regulates its turnover and subcellular localisation by interfering with its interaction with adenomatous polyposis coli protein, thereby enhancing transcriptional activity of β-catenin towards its target genes, which again include cyclin D1 (Ref. Reference Ryo70). Third, upon cytokine treatment, Pin1 can target the pThr254-Pro motif in NF-κB (NFKB1) and p65, and inhibits p65 binding to its inhibitor IκB, resulting in increased nuclear accumulation, protein stability and transcriptional activity of NF-κB towards its target genes, which include cyclin D1 and IκB (Ref. Reference Ryo89). Furthermore, Pin1 can directly bind to the pThr286-Pro motif of cyclin D1, which stabilises the cyclin D1 protein by preventing its nuclear export and ubiquitin-mediated degradation (Ref. Reference Liou28). Moreover, deletion of Pin1 in the mouse results in a phenotype that resembles that of the cyclin-D1-knockout mouse in some organs (Ref. Reference Liou28). In addition, Pin1 is an E2F target gene, whose transcription is upregulated in response to activation of Her2/Neu or Ras (Ref. Reference Liou28), suggesting that a positive-feedback loop exists (Fig. 3). Therefore, Her2/Neu or Ras activation leads to an increase in Pin1 transcription by binding to the Pin1 promoter through the E2F family of transcription factors, and Pin1, in turn, positively regulates cyclin D1 at the transcriptional and post-translational level.

Figure 3.

Pin1 regulates several oncogenic signalling pathways at various levels. Pin1 can increase transcription of cyclin D1 by modulating at least three different signalling pathways. Pin1 targets the pSer63/73-Pro motifs in Jun, multiple pSer/Thr-Pro motifs in Fos, the pThr246-Pro motif in β-catenin and the pThr254-Pro motif in the p65/RelA subunit of nuclear factor (NF)-κB to increase their nuclear translocation, protein stability and/or activity, thereby increasing expression of cyclin D1. Pin1 also directly binds to the pThr286-Pro motifs in cyclin D1 to increase its nuclear translocation and protein stability. Pin1 is an E2F target gene, whose transcription is upregulated in response to oncogenic activation, suggesting that a positive-feedback loop exists. In addition, the tumour suppressor death-associated protein kinase 1 (DAPK1) phosphorylates Pin1 and inhibits its catalytic activity.

Subsequent studies using cell culture and mouse models have demonstrated the crucial role of Pin1 in cancer development in vivo. Pin1 overexpression not only leads to transformed properties on normal mammary epithelial cells, but also can enhance the transformed phenotypes of Neu- or Ras-induced mammary epithelial cells (Ref. Reference Liou28). By contrast, inhibition of Pin1 suppresses both cell proliferation and transformation induced by Neu- or Ras by cyclin D1 (Ref. Reference Wulf107). Pin1 can also inhibit a negative-feedback mechanism by promoting dephosphorylation of Raf, which activates Ras signalling (Ref. Reference Dougherty97). Pin1 activates numerous other oncogenes or growth enhancers, including NF-κB (Ref. Reference Ryo89), Fos (Ref. Reference Monje82), Stat3 (Ref. Reference Lufei103), Myb (Ref. Reference Pani87), Notch1 (Ref. Reference Rustighi90), AKT (Ref. Reference Liao75), AIB1 (Ref. Reference Yi74), Neu (Ref. Reference Lam88), Mcl-1 (Ref. Reference Ding85), Hbx (Ref. Reference Pang84), Tax (Ref. Reference Peloponese104), FAK (Ref. Reference Zheng81), p70S6K (Ref. Reference Lee91) and v-Rel (Ref. Reference Fan100), and also inactivates a large number of tumour suppressors or growth inhibitors, including FOXOs (Ref. Reference Brenkman51), PML (Ref. Reference Reineke96), Daxx (Ref. Reference Ryo80), Bcl2 (Ref. Reference Basu77), RARα (Ref. Reference Gianni98), SMRT (Ref. Reference Stanya102), Smad (Ref. Reference Nakano101), Bax (Ref. Reference Shen76), Btk (Ref. Reference Yu78), GRK2 (Ref. Reference Penela83) and Pin2/TRF1 (Ref. Reference Lee20) (Fig. 2). Finally, Pin1-knockout mice are almost completely resistant to cancer induced by overexpression of oncogenes, such as Neu or Ras (Ref. Reference Wulf107), or knockout of tumour suppressors, such as p53 (Ref. Reference Takahashi92).

Pin1 and centrosome amplification

A recent study determined the role of Pin1 in centrosome duplication during the cell cycle (Ref. Reference Suizu18). The centrosome is a principal microtubule-organising centre in animal cells that facilitates the formation of mitotic spindle poles that segregate duplicated chromosomes between dividing cells (Refs Reference Bornens108, Reference Doxsey, McCollum and Theurkauf109). Centrosome duplication is tightly regulated by the nuclear cell cycle and its deregulation can lead to abnormal centrosome numbers and aberrant mitosis, and eventual chromosome instability (Refs Reference Nigg4, Reference Doxsey110). In fact, most tumours contain cells with centrosome abnormalities or amplification, which can indicate a poor clinical outcome (Refs Reference Nigg4, Reference Doxsey110). Pin1 overexpression correlates with centrosome amplification in human cancer tissues (Ref. Reference Suizu18). Furthermore, Pin1 localises to, and copurifies with, centrosomes in interphase cells. Functionally, ablation or inhibition of Pin1 delays centrosome duplication during the S phase and suppresses centrosome amplification in S-phase-arrested cells. By contrast, Pin1 overexpression in S-phase-arrested cells induces centrosome amplification, aneuploidy and cell transformation. Moreover, mouse mammary tumour virus (MMTV)–Pin1 transgenic mice also develop centrosome amplification and eventually breast cancer (Ref. Reference Suizu18). Together with the findings that Pin1 expression is significantly increased at the G1–S transition and is also subjected to phosphorylation regulation during the cell cycle, these results indicate that Pin1 is crucial for coordinating centrosome duplication and DNA synthesis, and disruption of this process can induce centrosome amplification, chromosome instability and oncogenesis in vitro and in vivo.

Pin1 and the ER

Pin1 has also been shown to regulate both SRC-3 (the steroid receptor coactivator 3; also known as NCOA3) and SMRT (the corepressor silencing mediator for retinoic acid and thyroid hormone receptor; NCOR2) as downstream targets of Her2 to estrogen receptor (ER) signalling in breast cancer (Refs Reference Yi74, Reference Stanya102, Reference Ryo111). SRC-3 is a transcriptional activation factor that interacts with steroid receptors in a ligand-dependent manner (Ref. Reference Liao112). It is required for various physiological processes, such as cell proliferation, and is often overexpressed in breast and prostate cancers (Ref. Reference Liao112). Pin1 interacts with phosphorylated SRC-3 and activates nuclear-receptor-regulated transcription of SRC-3 towards its steroid receptor target genes (Ref. Reference Yi74). Pin1 also enhances interactions between SRC-3 and additional coregulator histone acetyltransferases such as CBP and p300 to recruit to the promoter-bound receptor coactivator complex (Ref. Reference Yi74). Moreover, Pin1 promotes SRC-3 cellular turnover, suggesting that Pin1 functions as a secondary coactivator for steroid receptors by modulating conformational changes of SRC-3 protein and turnover of the activated SRC-3 oncoprotein. A recent study shows a role for Pin1 as a target for preventing improper ER activation by modulating SMRT protein stability (Refs Reference Stanya102, Reference Ryo111). SMRT is a coregulatory protein that mediates the repressive function of nuclear hormone receptors including ER (Ref. Reference Xu, Glass and Rosenfeld113). Pin1 binds directly to phosphorylated SMRT and decreases SMRT protein stability, thereby affecting SMRT-dependent transcriptional repression (Ref. Reference Stanya102). Moreover, Pin1 is overexpressed in Her2-positive breast cancer and leads to the upregulation of erbB2 (receptor tyrosine protein kinase erbB-2, which is also known as Her2) (Ref. Reference Lam88). Consequently, Pin1 is a central cofactor for coordinated Her2 activity and it is speculated that there is regulatory crosstalk between Pin1 overexpression and the constitutive activation of Her2-mediated transcriptional activation in breast cancer. These results suggest a potential therapeutic role for Pin1 in the treatment of endocrine-resistant breast cancers.

Pin1 and tau pathology

The protein tau is critically involved in neurodegeneration. As a microtubule-binding protein, tau can regulate cytoskeleton structure by assembling microtubules; thus physiologically, tau serves as a controller of microtubule stability in the neuron. However, tau can become pathological when hyperphosphorylated (Ref. Reference Buee125). Under these conditions, tau detaches from the microtubules, leading to destabilisation of the cytoskeletal structure and to cell death (Ref. Reference Iqbal126). An additional pathological mechanism occurs when hyperphosphorylated tau precipitates into insoluble aggregates, forming paired helical filaments (PHFs), which comprise larger NFTs. This was also shown in animal models overexpressing the wild-type form of human tau (Ref. Reference Andorfer127). When in the form of aggregates, tau is sequestered from the environment and cannot be dephosphorylated and thus cannot return to microtubule binding. In this respect, when it occurs, tau phosphorylation is a pathological event that cannot be reversed, and once triggered it causes neuronal death. It is indeed intriguing that phosphorylation controls the activity of tau to maintain cell structure, particularly because the kinases that phosphorylate tau, such as Cdk5, JNK and GSK3β (Refs Reference Cruz and Tsai128, Reference Noble129, Reference Mandelkow130, Reference Sperber131), are active during neurodegeneration (Ref. Reference Vincent and Davies132). In this regard, it could be speculated that when neurodegeneration starts, tau is phosphorylated, and this will cause irreversible modifications on tau that will lead to the death of neurons. It was previously suggested that such modifications would change the structure and conformation of tau (Refs Reference Daly133, Reference Mondragon-Rodriguez134, Reference Jeganathan135, Reference Carmel136), affecting its capability to bind and to stabilise the microtubules.

Recently, it was shown that tau phosphorylation at Thr231 triggers conformational changes at this site that will initiate misfolding and will affect further changes in the conformation at different sites (Ref. Reference Luna-Munoz137). In this regard, among the many phosphorylatable sites present in tau, Thr231 seems to be key to determining the misfolding process and targeting this residue might be beneficial for prevention of wider tau misfolding and aggregation. Earlier studies have pointed to a possible role of Pin1 in tau pathology by regulation of phosphorylated Thr231. Pin1 can bind solely to Thr231 when phosphorylated and, in the presence of Pin1, hyperphosphorylated tau is able to bind once again to microtubules in vitro (Ref. Reference Lu117). This suggests that direct isomerisation at Thr231 is a post-phosphorylative mechanism to control the potential damage triggered by hyperphosphorylated tau, which is due to the activation of kinases during neurodegeneration. Moreover, a specific conformation of tau phosphorylated at Thr231 was found to colocalise with Pin1 in intracellular lesions. Using conformation-specific antibodies for phosphorylated tau at Thr231 in the brain of patients affected by AD, this form of tau was found to colocalise with Pin1 in intracellular granular aggregates (Ref. Reference Ramakrishnan, Dickson and Davies138) that are different from the granulovacuolar lesions found in AD (Ref. Reference Holzer139). Similarly, Pin1 was also found to colocalise with tau aggregates in frontotemporal dementia (FTD) tauopathies (Ref. Reference Thorpe140).

The presence of Pin1 in deposits containing tau suggests that Pin1 is somehow involved in their formation and that the cellular environment might be depleted of soluble Pin1. The consequence would be that less Pin1 is available to counteract the changes of conformation on newly phosphorylated tau at Thr231, and would probably sustain tauopathy. Interestingly, in AD samples, levels of Pin1 were decreased in brain preparations, which prevented the extraction of proteins from insoluble aggregates (Ref. Reference Lu117), suggesting that the pool of soluble, and thus functional, Pin1 is decreased. These findings suggest that Pin1 protects against neurodegeneration caused by hyperphosphorylated tau.

This hypothesis was subsequently substantiated by demonstrations that selective gene ablation of Pin1 in a mouse model expressing endogenous levels of native tau leads to an age-dependent neurodegenerative phenotype, characterised by hyperphosphorylation of tau at several sites and by the formation of PHFs (Ref. Reference Liou29). Moreover, Pin1 expression inversely correlated with the levels of NFTs, because hippocampal neurons that showed increased levels of NFTs were characterised by lower levels of soluble, cytosolic Pin1, whereas those neurons with little or no NFTs had strong Pin1 expression. These findings suggest that levels of functional Pin1 inversely correlate with the vulnerability of neurons to develop lesions associated with hyperphosphorylated tau, and point to a possible role of Pin1 in the early stages of tauopathy. Conformation-specific hyperphosphorylated tau appears only in older Pin1-knockout animals, suggesting that the lack of Pin1 leads to accumulation of such forms of tau, which are possibly linked to the degenerative phenotype. Other studies have shown a role for Pin1 in promoting dephosphorylation of tau in vitro (Ref. Reference Galas141).

The fact that selective gene ablation of Pin1 could result in pathological modification on its substrate tau confirmed that direct isomerisation of specific substrates is key to the regulation of neurodegeneration and that specific isomers of tau are strictly related to the disease. Moreover, the fact that the ageing phenotype observed in the Pin1-knockout mice recapitulated that observed in human AD supports the idea that Pin1 has a crucial role in protecting against the disease; thus, increasing Pin1 levels in the brain could be a potential therapeutic approach to the treatment of AD.

A recent study showed that overexpression of Pin1 in the brain can revert the age-dependent tauopathy observed in Pin1-knockout mice (Ref. Reference Lim142). As expected, overexpression of Pin1 in a transgenic animal model (with Pin1 expressed in postnatal neurons under the neuronal promoter Thy1) was protective from a phenotype linked to neurodegeneration. Pin1-transgenic mice showed a significant reduction in both stability of endogenous tau and phosphorylation of wild-type tau compared with levels in age-matched Pin1-knockout mice expressing tau at the same level. These data suggest that Pin1 reduces the availability of overall tau that can potentially be hyperphosphorylated in pathological conditions, and also regulates the equilibrium between physiological and pathological conformations of phosphorylated tau. Surprisingly, the ability of Pin1 to prevent a neurodegenerative phenotype associated with pathological tau was reverted when Pin1-knockout or Pin1-transgenic animals expressed the tau mutant P301L, which is associated with tauopathy observed in frontotemporal dementia with Parkinsonism linked to chromosome 17 (FTDP-17) (Refs Reference Mirra143, Reference Lewis144, Reference Lewis145). The presence of the P301L mutation stabilises tau levels, increases overall and conformation-specific tau phosphorylation in Pin1 transgenic mice, and causes peripheral neuropathy. Altogether, these data highlight the importance of direct isomerisation of proteins in their native form as a fundamental factor in the initiation and progression of the molecular changes that lead to neurodegeneration in sporadic AD, and point directly to Pin1 as a key player in this process. Moreover, they open new perspectives on the use and construction of proper animal models for the study of the disease. The P301L tau mutant has been considered to be a good model for the study of AD-related tauopathy, because these animals develop massive tau hyperphosphorylation and formation of NFTs along with cognitive decline (Ref. Reference Gotz146). Given the crucial role of Pin1 in regulating the toxicity of endogenous tau, in the P301L mutant model, it cannot be excluded that the presence of the mutation maintains tau in a toxic configuration that cannot be reverted by Pin1, which can then be sustained despite the attempts of Pin1 to isomerise the protein. In this respect, one could speculate that the massive tauopathy observed in the mutant P301L tau model might not be representative of what occurs in the human brain in AD, especially considering that neither this nor other mutations on tau have been associated with this disease. Thus, the use of the Pin1-knockout mouse model should be a good tool for the study of age-dependent neurodegeneration.

Pin1 and β-amyloid pathology

β-Amyloid peptides accumulate as aggressive oligomers (Ref. Reference Walsh147) that impact synaptic transmission (Refs Reference Walsh148, Reference Shankar149) in AD. These oligomers further aggregate, forming the core of the β-amyloid plaque, which is the pathological hallmark of AD. The insoluble peptides have molecular weights of approximately 4 kDa and are generated from the precursor APP by pathological amyloidogenic processing, which involves sequential cleavage by the β-secretase BACE and γ-secretase on internalisation of the full-length protein from the plasma membrane (Ref. Reference Thinakaran and Koo150). The activity of BACE on APP generates a soluble stub called β-APP and an intracellular fragment called C99, which still contains the intact Aβ sequence. The activity of γ-secretase on C99 releases the intact Aβ (ranging in size from Aβ39 to Aβ43, with Aβ42 being the most aggressive) and a short fragment called the APP intracellular domain. APP can also be processed through a nonamyloidogenic pathway that involves the activity of α-secretase at the plasma membrane. This pathway generates α-APPs, which have neurotrophic properties (Ref. Reference Mattson151), it also prevents the production of Aβ and protects against AD (Ref. Reference Young-Pearse152). Interestingly, the processing of APP can be regulated by phosphorylation at different sites (Ref. Reference Suzuki and Nakaya153), and phosphorylation at the Thr668-Pro residue by Cdk5 (Refs Reference Lee and Tsai5, Reference Lee154, Reference Cruz155) and GSK3β (Ref. Reference Phiel156) increases amyloid pathology in vivo and in vitro. Because the Thr668-Pro residue makes APP a suitable substrate for Pin1, it can be hypothesised that Pin1 binds to APP at this site and regulates the conformation of APP, influencing either the amyloidogenic or the nonamyloidogenic pathway (Ref. Reference Pastorino12). Pin1 was found to bind specifically to phosphorylated Thr668 APP in vitro, which regulates the conformation of APP to favour nonamyloidogenic processing (Ref. Reference Pastorino12). The Thr668-Pro motif in APP exists only in the trans conformation; however, the cis conformation forms after phosphorylation as a result of local hydrogen bond formation (Refs Reference Ramelot, Gentile and Nicholson157, Reference Ramelot and Nicholson158). Pin1 greatly accelerates the cis to trans isomerisation of the phosphorylated Thr668-Pro motif of APP and also increases the nonamyloidogenic processing in vivo and in vitro (Ref. Reference Pastorino12). These data suggest that phosphorylation at Thr668 changes the conformation of APP, making it toxic, and that Pin1 might protect from β-amyloid pathology by regulating the conformation of APP. These effects could depend on the capability of Pin1 to control the equilibrium between the cis and trans forms of the phosphorylated APP, and imply that the maintenance of such an equilibrium would protect against AD. Pin1-knockout animals were crossed with APPtg2576, an animal model characterised by age-dependent Aβ pathology and plaque deposition (Refs Reference Carlson159, Reference Kawarabayashi160), to study the processing of APP. The results showed an age-dependent shift in the processing of APP that would lead to an increase in the amyloidogenic versus nonamyloidogenic processing of APP. In particular, APP-Tg/Pin1-knockout mice at 6 months, but not at 2 months, had a massive increase in β-APPs and insoluble Aβ42 peptide paralleled with a decrease in α-APPs and the presence of aggregated Aβ42 in multivesicular bodies (Ref. Reference Pastorino12), a feature indicating early AD pathology (Ref. Reference Almeida, Takahashi and Gouras161). Similarly, the processing of APP and Aβ production seems to be affected when APP is expressed at endogenous levels in Pin1-knockout mice, although at an older age, because Aβ42 peptide accumulated only in 15-month-old mice (Ref. Reference Pastorino12). A role for Pin1 in modulating the processing of APP was confirmed in vitro by Pastorino and colleagues, who manipulated the expression of Pin1 in breast cancer cells and H4 human neuroglioma cells (Ref. Reference Pastorino12). However contrasting results have been reported, where ablation of the Pin1 gene protected against Aβ formation (Ref. Reference Akiyama162). In this study, Akiyama and colleagues did not measure the generation of soluble and insoluble Aβ in the mouse brain during ageing, but only at an unspecified age. Second, they do not show binding of Pin1 to the full-length APP, but only to the C-terminal fragment C99. Given that (1) full-length APP is particularly phosphorylated at the Pin1-binding site Thr668 in AD (Refs Reference Lee154, Reference Cruz155) and that (2) Aβ accumulates and forms insoluble aggregates as the animal ages, the study by Pastorino and co-workers (Ref. Reference Pastorino12) gives a more comprehensive approach to understanding the role of Pin1 in modulating the processing of APP, ultimately leading to the accumulation of Aβ in the ageing Pin1-knockout mice.

It could then be speculated that Pin1 acts as an early factor in the development of Aβ pathology, because the absence of the Pin1 gene causes increased pathological amyloidogenic processing of the precursor APP in APP transgenic mice at earlier stages than observed when expressing endogenous levels of Pin1, and before plaque formation. The fact that Pin1 might be involved in early stages of Aβ pathology is supported by Bulbarelli and colleagues (Ref. Reference Bulbarelli163), who propose a compensatory upregulation of Pin1 as a consequence of toxic insult. In particular, they show that Pin1 activity is sustained and tau phosphorylation is decreased when cultured neurons are exposed to Aβ42 oligomers, which are early aggregates of Aβs.

Thus, Pin1 might protect from Aβ pathology in an age-dependent manner, possibly through regulation of the equilibrium between the cis and trans conformation of APP. In particular, in physiological conditions, Pin1 would favour the trans conformation of APP, increasing nonamyloidogenic processing. By contrast, in conditions associated with loss of Pin1 function or absence of Pin1, the cis form would accumulate, because isomerisation of the two forms would be lost, favouring amyloidogenic processing (Fig. 4).

Figure 4.

Age-dependent loss of functional Pin1 leads to the pathology of AD. During ageing, the levels of active Pin1 decrease as a result of changes in the synthesis or degradation of the protein or post-translational modifications, such as protein oxidation. In the Pin1-knockout animal model, the lack of expression of functionally active Pin1 leads to the development of a neuropathological phenotype associated with motor deficits and neurodegeneration. From a molecular point of view, the lower levels of active Pin1 are not sufficient to maintain the equilibrium between the cis and trans conformation of Pin1 substrates, in particular APP and tau. This results in the accumulation of specific conformations that are potentially toxic, causing the accumulation of hyperphosphorylated tau and NFTs, and increasing the amyloidogenic processing of APP, and Aβ formation. By modifying the conformation, and thus the molecular properties of both tau and APP, Pin1 controls the mechanism that switches their function from physiological to pathological, contributing to protection from the disease in different ways. Images adapted with permission from Nature Publishing Group (Ref. 29). Abbreviations: Aβ, β-amyloid peptide; AD, Alzheimer disease; APP, amyloid precursor protein; NFT, neurofibrillary tangle.

This could be the case in a pathological scenario associated with AD: unlike physiological conditions characterised by low levels of phosphorylated APP at Thr668 and by physiological levels of Pin1, in AD pathology, levels of phosphorylated Thr668 are increased and Pin1 is possibly downregulated. Given the characteristics of Aβ pathology displayed by the Pin1-knockout model, it can be hypothesised that in AD, the equilibrium between the cis and trans conformation could shift towards the cis conformation, because of either an excessive level of APP phosphorylated at Thr668 or decreased levels of functional Pin1. More importantly, studies on Pin1, APP and tau (Refs Reference Pastorino12, Reference Liou29, Reference Lim142) open up a new perspective on the study of protein isomerisation as a crucial mechanism to control post-phosphorylative events that might be detrimental to the life of the cell.

Pin1 in the nervous system and degenerating human brain

An increasing body of evidence supports the direct involvement of Pin1 in neurodegenerative diseases in the human brain. First, it was shown that levels of soluble Pin1 are decreased in the brains of AD patients compared with age-matched control brains (Ref. Reference Lu117) and that Pin1 expression decreases in neurons that are vulnerable for tauopathy (Ref. Reference Liou29). These data pointed to a possible protective role offered by Pin1 and suggested that a decrease in Pin1 levels might account for neurodegeneration. This hypothesis was further strengthened by the work of other laboratories, which showed that the reduction of either the expression (Ref. Reference Wang164) or the activity of Pin1 (Refs Reference Butterfield14, Reference Butterfield55, Reference Sultana165) would associate with cognitive decline and AD. In particular, Wang and colleagues (Ref. Reference Wang164) showed that levels of the Pin1 protein, but not its mRNA, were decreased in AD and in patients affected by MCI, a condition that often precedes early stages of AD, and that Pin1 activity inversely correlated with both tau phosphorylation and clinical cognitive scores in both AD and MCI patients. These data suggest that reduced Pin1 levels have possible consequences on both tauopathy and dementia, and were further supported by the findings of Butterfield and colleagues (Ref. Reference Butterfield55), which demonstrate that Pin1 undergoes oxidation in the hippocampus of AD patients, paralleled by loss of function (Ref. Reference Sultana165).

Further evidence linked a possible reduction of the Pin1 protein level to the development of AD. Several SNPs were found to associate with AD and with decreased Pin1 levels in human cells (Ref. Reference Segat, Milanese and Crovella22), although it is under debate whether these genetic mutations on the PIN1 gene relate significantly to AD (Refs Reference Poli120, Reference Nowotny121, Reference Lambert166). However, the study of the genetics of PIN1 could be useful to deepen our knowledge on protein function in the human brain. Recently, Ma and colleagues (Ref. Reference Ma122) reported the existence of a polymorphism of Pin1 that reduces its own suppression because it inhibits binding to the brain-selective transcription factor AP-4, which would otherwise downregulate expression of Pin1. Because this polymorphism is associated with a delayed onset of sporadic AD in a Chinese population, this study opens new perspectives on mechanisms that could regulate and sustain Pin1 expression in AD and that could be potentially used for the treatment of this disease.

The presence of Pin1 in degenerating neurons has also been linked to amyotrophic lateral sclerosis (ALS) and to non-AD neurodegeneration associated with ageing, because Pin1 colocalises with hyperphosphorylated neurofilament protein in the spinal cord of ALS patients (Ref. Reference Kesavapany167) and with the aggregates of lipofuscin in elderly subjects (Ref. Reference Hashemzadeh-Bonehi168).

Ageing

Pin1 is essential for cell division in some organisms such as budding yeast, Aspergillus nidulans and Candida albicans (Ref. Reference Lu and Zhou2). However, Pin1 ablation is dispensable for cell growth in others, including fission yeast. Initial analyses of young Pin1-knockout mice have shown that they are viable and develop normally with mild phenotypes (Ref. Reference Fujimori169), suggesting that Pin1 is not required for cell division per se. However, further careful studies of adult Pin1-knockout mice have shown that they display a range of abnormalities similar to those of cyclin-D1-knockout mice (Ref. Reference Liou28). The body weight of adult Pin1-knockout mice is significantly smaller at 71% of body weight in wild-type mice (Ref. Reference Liou28). Moreover, Pin1-knockout mice showed retinal degeneration and Pin1-knockout pregnant females showed mammary gland proliferative impairment (Ref. Reference Liou28). Pin1-knockout male mice display primordial germ cell depletion during embryonic development, along with testis atrophy in the adult (Refs Reference Liou28, Reference Atchison, Capel and Means30). Furthermore, Pin1-knockout mice display acceleration of telomere shortening and obvious premature ageing phenotypes within a single generation, including osteoporosis, lordokyphosis and skin atrophy (Ref. Reference Lee20). In the brain, Pin1-knockout mice develop age-dependent neuropathy such as motor coordination and behavioural defects, as well as neuronal degeneration and loss (Refs Reference Pastorino12, Reference Liou29). These results suggest that Pin1 has a central role in regulating the ageing process in vivo.

Cancer

The critical roles of Pin1 in regulating cancer development have been studied in mouse models. Tissue-specific expression of Pin1 in mouse mammary gland induces centrosome amplification, eventually leading to mammary hyperplasia and malignant breast cancer with overamplified centrosomes (Ref. Reference Suizu18). Furthermore, Pin1 ablation in mice is highly effective in preventing oncogenic Neu or Ha-Ras, but not c-Myc, from inducing cyclin D1 and hence breast cancer (Ref. Reference Wulf107). Moreover, Pin1 ablation in p53-knockout mice completely prevents malignant tumour formation, although these mice exhibit accelerated thymic hyperplasia caused by a defective differentiation pathway (Ref. Reference Takahashi92). In addition, mice expressing p53 with the deletion of Pin1-binding sites in the crucial Pro-rich domain of p53 do not affect the ability of p53 to inhibit tumour development induced by oncogenes (Ref. Reference Toledo170), demonstrating an essential role for Pin1 in oncogenic signalling pathways.

Alzheimer disease

The existence of a common regulator of both tau and APP pathology has great relevance to the search for an appropriate therapeutic target to be used in AD, possibly at early stages of the disease. Although it has emerged that Aβ pathology occurs before tau pathology in vivo (Refs Reference Lewis144, Reference Gotz146), it should be noted that these studies were carried out using animal models overexpressing pathogenic mutants of both APP and tau proteins: thus they are far from representing the real scenario of the brain of patients affected by sporadic AD. The presence at the same time of a pathogenic mutant at high amounts pushes the biological system towards the exaggeration of toxic effects regulated by that protein. In particular, the presence of the mutation (tau P301L or APPswe) makes the protein very aggressive and capable of generating toxic species such as hyperphosphorylated tau and Aβs that can quickly aggregate into tangles and plaques, respectively, even at a young age. Given that AD develops late in life, these models do not fully represent the complex degenerative scenario associated with ageing that accompanies or causes AD in the human brain. However, it is important to point out that in AD, the levels of both tau and APP might be stabilised, for example, by an inefficient degradation system. The increased availability of the protein will certainly affect how much of the toxic species will be generated and this will influence progression of the disease. In this respect, animal models that are characterised by an accumulation or high levels of APP or tau are indeed useful for the study of AD.

Although Aβ seems to regulate tau hyperphosphorylation and aggregation, in vitro and in vivo (Refs Reference Lewis144, Reference Gotz146, Reference Rapoport171), it remains under debate whether an effective treatment of AD pathology should first tackle the APP pathway and Aβ generation, or rather tau hyperphosphorylation and tangle formation. The studies performed using Pin1-knockout mice have identified in these animals a model that recapitulates some of the phenotypes found in AD, such as the age-dependent tangle and β-amyloid pathology that occur specifically from adulthood on in the life of the animal, similarly to the development of the disease in the human brain (Refs Reference Pastorino12, Reference Liou29, Reference Lim142) (Fig. 5). Lack of Pin1 expression causes a degenerative phenotype related to tau and β-amyloid pathology when tau and APP are expressed at endogenous levels. In particular, it was observed that in the absence of Pin1 expression in the mouse brain, levels of endogenous tau are stabilised, with consequences on increased tau phosphorylation and PHF formation. Moreover, the Pin1-knockout mice also showed increased Aβ42 production. Thus, using the Pin1-knockout mouse model, it is possible to study the pathogenic phenotypes (i.e. tauopathy and AD) of proteins such as APP or tau expressed at endogenous levels, to allow a better understanding of the degenerative scenario associated with sporadic AD in the human brain.

Figure 5.

The Pin1-knockout mouse as an animal model for the study of ageing and neurodegeneration. Ablation of the Pin1 gene results in a phenotype that recapitulates the phenomena associated with ageing and neurodegeneration in AD, in the absence of transgenes such as mutant human APP or tau. This makes the Pin1-knockout mouse a unique animal model for the study of pathologies associated with ageing. Pin1-knockout mice are characterised by premature ageing due to changes in skeletal or muscular structure resulting in lordokyphosis and muscular atrophy. During ageing, these animals also develop massive tau phosphorylation and deposition in typical PHFs, as well as production and deposition of toxic Aβ species at an intracellular level. Pathological tau and Aβ lead to the neurodegenerative phenotype associated with lack of Pin1. Images adapted with permission from Nature Publishing Group (Refs 12, 20, 29). Abbreviations: Aβ, β-amyloid peptide; AD, Alzheimer disease; APP, amyloid precursor protein; PHF, paired helical filament.