Reduced PIN1 expression in neocortical and limbic brain regions in female Alzheimer ’ s patients correlates with cognitive and neuropathological phenotypes

Women have a higher incidence of Alzheimer ’ s disease (AD), even after adjusting for increased longevity. Thus, there is an urgent need to identify genes that underpin sex-associated risk of AD. PIN1 is a key regulator of the tau phosphorylation signaling pathway; however, potential differences in PIN1 expression, in males and females, are still unknown. We analyzed brain transcriptomic datasets focusing on sex differences in PIN1 mRNA levels in an aging and AD cohort, which revealed reduced PIN1 levels primarily within females. We validated this observation in an independent dataset (ROS/MAP)


Introduction
Alzheimer's disease (AD) is the most prevalent neurodegenerative disorder worldwide, characterized clinically by impairments in cognition.Its pathological features include extracellular amyloid-β peptide (Aβ) plaques, intraneuronal tangles comprised of hyperphosphorylated tau, and synaptic and neuronal loss (DeTure and Dickson, 2019).Given the aging global population, the burden of AD is expected to dramatically increase over the next few decades, with an estimated prevalence of 20 million individuals in the US alone by 2050 (Association, 2018).
Unfortunately, current treatment strategies offer limited effectiveness in preventing, treating, or managing AD.
Moreover, sex differences have been reported in various aspects of AD, including epidemiological, risk factors, and biomarker studies (Zhu et al., 2021).Women exhibit a higher incidence and prevalence of AD than men (Rajan et al., 2021), even after accounting for increased longevity (Ferretti et al., 2018;Mielke et al., 2014;Rocca, 2017;Zhu et al., 2021).Thus, there is an unmet need to identify the molecular networks governing cognitive processes and contributing to the sex-related risk of AD.
Phosphorylation of the amyloid precursor protein (APP) and hyperphosphorylation of tau protein, both increased in AD brains, lead to the formation of disease-specific Aβ accumulation and neurofibrillary tangles (NFTs), respectively (Busche and Hyman, 2020).These observations underscore the critical role of phosphorylation events in understanding AD pathogenesis and treatment.One key protein in phosphorylation pathways is peptidyl-prolyl cis/trans isomerase (PIN1), which catalyzes the isomerization of the peptide binding bond between phosphorylated Ser/Thr-Pro in proteins, thereby regulating protein function post-phosphorylation (Liou et al., 2003;Wang et al., 2023).Various reports show that PIN1 expression is dysregulated in AD, indicating its potential significance in disease pathogenesis (Balastik et al., 2007;Ramakrishnan et al., 2003;Xu et al., 2017).
Initial studies in PIN1-knock-out mice indicated normal development but defects in entering the cell cycle from G0 arrest (Fujimori et al., 1999).Additionally, PIN1-deficient mice exhibited lower production of Aβ40 and Aβ42 compared to wild-type mice, suggesting that PIN1 promotes Aβ production in the brain (Akiyama et al., 2005).Subsequent studies in germline mice have shown that deletion of the PIN1 gene can cause both tau and Aβ-related pathologies in an age-dependent manner (Liou et al., 2003;Pastorino et al., 2006).Homozygous PIN1 knockout (KO) mice develop a progressive phenotype of premature aging (Kondo et al., 2017) and show age-dependent neuropathy characterized by tau hyper-phosphorylation, tau filament formation, APP amyloidogenesis, intracellular Aβ42 accumulation, and neuronal degeneration (Liou et al., 2003;Pastorino et al., 2012Pastorino et al., , 2006)).However, germline PIN1 KO mice models present challenges in breeding and show adaptation during development that could affect the translatability of results.Findings from a floxed PIN1 mice model and neuronal cultures derived from them suggest that PIN1 can be inactivated early in the evolution of AD by Aβ42 signaling, directly contributing to spine pathology (Stallings et al., 2018).Finally, PIN1 has been shown to play a protective role for neurons under toxic in vitro conditions (Balastik et al., 2007;Kondo et al., 2017).This suggests that PIN1 may have a neuroprotective role, potentially by regulating critical cellular processes or signaling pathways that help neurons withstand damage or promote their survival in challenging environments.Collectively, these studies in both murine models and human participants illuminate the importance of investigating alterations in PIN1 expression for understanding the initiation and progression of AD.
Although previous work has shown that PIN1 expression levels are significantly reduced in AD human post-mortem brain tissue (Butterfield and Sultana, 2007;Sultana et al., 2006), it remains unclear whether PIN1 expression dysregulation occurs as a function of age or sex.Additionally, the specific classes of central nervous system (CNS) cells showing reduced expression are not known.Therefore, this study aimed to determine whether PIN1 expression is significantly altered with aging (20-90 years) and in pre-AD cases (amnestic mild cognitive impairment, MCI).Furthermore, AD expression changes were validated using three independent studies employing multiple expression platforms.The comprehensive findings suggest that PIN1 may serve as an informative molecule for elucidating the higher incidence of AD in females and as a marker of disease progression.(Berchtold et al., 2008;Deng et al., 2019;Mastroeni et al., 2017aMastroeni et al., , 2018aMastroeni et al., , 2017b;;Tan et al., 2018).Genomic data can be found in the NIAGADS database (accession number NG00057), and Gene Expression Omnibus (GEO) database, www.ncbi.nlm.nih.gov/geo(accession no.GSE11882).

Tissue homogenates and affymetrix arrays 2.2.1. Tissue collection
As previously published (Berchtold et al., 2013(Berchtold et al., , 2008(Berchtold et al., , 2014)), frozen unfixed tissue was available from two or more regions in 85 % of the cases, resulting in a total of 193 tissue samples (56 EC, 62 HIPP, and 75 SFG).SFG (Superior Frontal Gyrus, crest/superior surface at the genu of the corpus callosum), HIPP (Hippocampus, body of the HIPP at the level of the lateral geniculate nucleus), EC (Entorhinal Cortex, crest of the parahippocampal gyrus at the level of the anterior hippocampus).Group sizes were as follows: young (n = 61, 20-59 years, mean age 35.4 ± 10.5 years), aged controls (n = 73, age 69-99, mean age 84.2 ± 8.9 years), and AD cases (n = 59, ages 74-95 years, mean age 85.7 ± 6.5 years) with males and females similarly represented in each group.Refer to references (Berchtold et al., 2013(Berchtold et al., , 2014) ) for detailed sample information.Total RNA was extracted from the hippocampus as described previously.

Statistical analysis
All genes that did not meet the 50 % present call threshold were removed by Genespring G 7.3.1 Expression Analysis software.We preferentially selected those probe sets that were most specific, such that they were annotated with the smallest number of Ensembl gene IDs.After applying this criterion, if there remained multiple probes for any one gene, we excluded those probe sets expected to hybridize with targets in a non-specific fashion (i.e., those with "_x_" in the Affymetrix identifier).If there remained multiple probes for a given gene, we took the mean of the probe sets.
Select genes were investigated for statistical significance (p < 0.01).A two-tailed paired t-test, assuming equal variance (using multiple testing corrections, by Benjamini and Hochberg False Discovery Rate), was applied to locate genes that were significant in differentiating expression between young and aged controls and AD and aged controls.
Data have been deposited in the Gene Expression Omnibus database (www.ncbi.nlm.nih.gov/geo)accession number GSE11882.

PIN1 expression analysis in ROS/MAP dataset 2.3.1. Clinical and pathological assessment for the ROS/MAP dataset
The Religious Orders Study (ROS) and Rush Memory and Aging Project (MAP) are prospective, longitudinal clinical-pathologic cohort studies of aging and Alzheimer's disease (AD) (Bennett et al., 2018).Upon enrolment, participants without known dementia agree to annual clinical evaluation and brain donation.Demographic variables, including socioeconomic status, education, occupation, and early-life socioeconomic status indicators are collected for all participants.Syndromic clinical phenotypes including dementia (especially AD dementia), mild cognitive impairment (MCI), cerebrovascular disease, vascular cognitive impairment, Parkinson's disease, and depression are recorded.Quantitative clinical phenotypes include 21 cognitive performance tests and motor performance tests are performed annually.Various risk factors, including experiential factors, affect, and personality traits, are also measured.Upon autopsy, a detailed neuropathologic evaluation is performed.Comprehensive multiomic profiling is available on post-mortem brain tissue samples from a large contingent of decedents, including RNA-sequencing, proteomics, and DNA sequencing.
RNA sequencing data was obtained from the Accelerating Medicines Partnership -Alzheimer's Disease (AMP-AD) Knowledge Portal (ROS/ MAP: syn3505720).ROS and MAP cohort samples were collected from the dorsolateral prefrontal cortex (DLPFC) as previously described (Bennett et al., 2012a).Gene expression within the ROS and MAP cohorts was available as gene FPKM estimates.We log2 transformed PIN1 FPKM values after incrementing an offset of 0.5 to avoid the transformation of zero values.Participants were classified according to a clinical cognitive diagnosis summary as AC (Aged controls, no cognitive impairment), MCI (mild cognitive impairment, no other condition contributing to cognitive impairment), and AD (Alzheimer's disease dementia, no other condition contributing to cognitive impairment).Demographic, clinical, neuropathological, and technical variables for the participants that contributed to this data are summarized in Supplementary Table 1.

Immunohistochemical studies
Chromogenic immunohistochemical studies were completed on twenty-four human temporal neocortical samples, 12 AD and 12 ND samples.Samples were matched for PMI, age, and sex, among other covariates (Supp.Table 2).For detailed methods please see references (Mastroeni et al., 2013(Mastroeni et al., , 2015(Mastroeni et al., , 2009)).Briefly, 40µm free-floating sections were blocked in H 2 O 2 and bovine serum albumin.Following blocking steps, tissues were incubated in antibody raised against PIN1 (1:200 dilution, Sigma), overnight (ON) at 4 • C. Sections were washed and then incubated in species-specific secondary (1:1000, Vector) for two hours at room temperature.Sections were washed and incubated in 1:1000 avidin/biotin reagent, washed and incubated in DAB.All sections were reacted for the same amount of time, dried, taken through graded alcohols, cleared in xylene, and mounted using permount.Adjacent serial sections were stained with cresyl violet, or within sections with neutral red for structural visualization.Slides were imaged using an Olympus IX71.
For fluorescence microscopy, the sections were washed 3X in PBST, blocked with either 3 % normal goat serum or 3 % BSA, and incubated for 1 h.After further washing, sections were incubated in primary antibody [PIN1(mouse) antibody, 1:200 dilution, PS396 (rabbit) 1:1000, T231 (rb) 1:500] ON at 4 • C. Sections were washed 3X in PBST and incubated in species-specific, fluorophore-conjugated secondary antibodies (Molecular Probes).After a final wash, the sections were mounted, taken through Sudan Black to reduce autofluorescence, and coverslipped with Vectashield (Vector).Immunostained tissue sections were examined on Nikon Eclipse Ti2 confocal and Olympus IX70 microscopes equipped with epifluorescence illumination.The findings were documented photographically with an Olympus DP-71 color digital camera or, for confocal microscopy, by Nikon A1/A1R.Blocking peptides were used to confirm antibody specificity (Supp.Figure 1).

Western blot analysis
For Western blots, frozen temporal cortical blocks were lysed in a solution containing 20 mM Tris pH7.5, 0.5 % Nonident (Sigma), 1 mM EDTA (Sigma), 0.1 M NaCl (Sigma), 1 mM PMSF (Sigma), Sigma protease inhibitors 1, 2, and complete protease inhibitor cocktail (Roche).Protein concentrations were determined by BCA assay (Pierce) using bovine serum albumin as the standard.A total of 40µg of sample protein was combined with Laemmli sample buffer for separation by SDS-PAGE, followed by transfer to PVDF membrane (Bio-Rad).Membranes were blocked using 5 % non-fat dry milk and probed with anti-PIN1.After incubation with the primary antibody, membranes were washed, incubated with the secondary antibody, washed again, reacted with chemiluminescence substrate (Pierce), and imaged on Amersham imager 680 (GE).Exposure 1.5 min.

PIN1 mRNA levels are downregulated as a function of age
The most salient risk factor for Alzheimer's disease and other dementias is aging (Hou et al., 2019).To understand the effect of aging on the expression of PIN1, we analyzed three brain regions: HIPP, ENT, and SFG in an aging cohort: young control (YC) (n = 22, 20-59 years, mean age 35.4 ± 10.5 years) vs. aged control (AC) (n = 33, age 69-99, mean age 84.2 ± 8.9 years) (Mastroeni et al., 2017a(Mastroeni et al., , 2018b)).Significant downregulation of PIN1 was observed in AC compared to YC males and females in the HIPP, and SFG.Only females were significantly downregulated in ENT (Figure 1A).Linear regression analysis (Pearson's correlation coefficient (r)) revealed a significant correlation with PIN1 as a function of age in HIPP (P<0.001,F-ratio 8.48) (Figure 1B), ENT (P<0.0001,F-ratio 22.65) (Figure 1C), and SFG (P<0.001,F-ratio 11.20) (Figure 1D).These data show that PIN1 mRNA levels significantly decrease as a function of age.

PIN1 mRNA levels are reduced in Alzheimer's disease compared to age-matched controls
To characterize the effect of AD on PIN1 expression, we compared the same aged controls (AC, n = 33, age 69-99, mean age 84.2 ± 8.9 years), to clinically and neuropathologically confirmed AD cases (n = 26, ages 74-95 years, mean age 85.7 ± 6.5 years).As observed for advanced age (Figure1), PIN1 levels were significantly down-regulated in AD-hippocampus in AD vs. AC (p<0.0005,Log 2-Fold Change − 2.4).When separated by sex, both females (p<0.0001,F-ratio 11.79) and males (p<0.049,F-ratio 4.94) were significantly downregulated, but females were overall the most affected (Figure 2).In the SFG, AD vs. ND showed significant PIN1 downregulation (p<0.01,F-ratio 6.32) overall.When separated by sex, only females were significantly downregulated (p<0.01,F-ratio 8.12) (Figure2).Similarly, In the ENT, AD vs. ND showed an overall significant difference (p<0.001,F-ratio 10.19).When separated by sex, only females were significantly downregulated (p<0.001,F-ratio 8.42) (Figure 2).These findings indicate that PIN1 expression changes are largely driven by females in the HIPP and are solely responsible for the overall AD-associated expression changes in the SFG and ENT.

Impaired cognition and increased neuropathology severity are associated with reduced PIN1 expression in females
To further explore how the expression of PIN1 varies with age, sex, and clinicopathological markers of AD, we incorporated post-mortem brain RNA-sequence data from the Religious Orders Study ROS), and Memory and Aging Project (Bennett et al., 2012a(Bennett et al., , 2012b) (MAP) cohorts.These data together (ROS/MAP) comprised 581 brain tissue samples collected from the DLPFC (summarized in Supp.Table 1).We obtained normalized gene expression data and examined the abundance of PIN1 in samples obtained from participants with varying diagnoses, severity of neuropathology, sex, and cognitive scores.
Given the associations between PIN1 activity and synaptic function (Stallings et al., 2018), we examined correlations between PIN1 and a multi-regional measure of soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) proteins that have been assayed across the MAP cohort (N=258 participants).Mean SNARE protein immunodensities are based on the average of syntaxin-1, vesicle-associated membrane protein (VAMP), and synaptosomal-associated protein-25 (SNAP-25), aggregated across six different cortical regions (hippocampus, middle frontal gyrus, inferior frontal gyrus, calcarine cortex, ventromedial caudate, and posterior putamen) and converted to a Z-score across all participants (Honer et al., 2012).Mean SNARE protein density has been demonstrated as an indicator of pan-synaptic function, and predictive of cognitive function before death, with declines in SNARE density correlating with synaptic loss in a manner that appears independent from neuropathology-driven synaptic loss (Ramos-Miguel et al., 2018).We observed that PIN1 expression was positively correlated with mean SNARE protein density Z-score in both males and females (Mean SNARE protein immunodensity Z-score vs. PIN1 expression correlation, Female-only Rho: 0.31, FDR: 6.7e-4, Male-only Rho: 0.36, FDR: 7.2e-3, Figure 3E).
We also observed that PIN1 was positively correlated with global cognitive function in a sex-specific manner.Global cognitive function is comprised of a Z-score transformed composite from a battery of cognitive tests (Wilson et al., 2015) and was positively correlated with PIN1 expression, but only in females (Global Cognitive Function Z-Score vs. PIN1 expression correlation, Female-only Rho: 0.27, FDR: 2.4e-6, Male-only Rho: 0.05, FDR: 0.88, Figure 3F).
Moreover, to explore whether additional molecular players associated with PIN1, such as promyelocytic leukemia protein (PML), stratifin (SFN), and bridging integrator 1 (BIN1) exhibit sexually dimorphic patterns relevant to understanding the mechanisms underlying sex differences in AD, we analyzed their expression levels.We observed robust expression of these transcripts; however, there were no significant differences between AD, MCI, and AC participants, either when considering all participants (PML: Supp.

Axonal PIN1 protein levels are decreased in females
To investigate sex differences in PIN1 distribution, in AC and AD, we performed immunohistochemistry in the temporal neocortex  (Mastroeni et al., 2013(Mastroeni et al., , 2015)).These data may indicate that in normal aging the loss of PIN1 in females may be axonal, and in AD it appears to be both axonal and further, a redistribution of PIN1.

Colocalization with the early tau epitope pThr231
The phosphorylation of tau follows a specific pattern of reactivity, accumulating primarily in the entorhinal region in early Braak stages and subsequently progresses to the limbic system (e.g., HIPP), and neocortical regions (e.g., SFG) as disease progresses (Braak et al., 2006).The present study focused on the early tau marker pThr231 because of its association with early regional changes (Neddens et al., 2018).A comparative colocalization study was performed using two highly cited tau antibodies: the late-stage antibody PS396 (Figure 4H) vs. pThr231 an early-stage antibody (Figure 4K) and PIN1 (Figure 4G & J).The late tau marker (p-tau 396) shows a relatively weak association with PIN1 (6.9 % overlap) Figure 4I, and the early tau marker (p-tau 231) shows a strong association (88.2 % overlap) Figure 4L.Interestingly, the cytosolic punctate IR in AD does not colocalize with tangles (Figure 4I insert).It may be that the redistribution of PIN1 may predate the accumulation of phosphorylated tau.There were no significant differences between the number of overlapping cells and sex.

Discussion
Imaging (de Leon et al., 2007;Gonneaud et al., 2017), cognitive assessment (Riley et al., 2005;Snowdon et al., 1996), recent peripheral evidence (Palmqvist et al., 2020), and neuropathological data (Braak and Braak, 1997;Pletnikova et al., 2015) have established that Alzheimer's disease affects the brain for decades before clinical diagnosis (Bateman et al., 2012).Synaptic degeneration and aberrant phosphorylation are some of the earliest pathological features, and strongest correlates to cognitive decline in AD (D'Amelio et al., 2011;DeKosky and Scheff, 1990;Mondragon-Rodriguez et al., 2014, 2018;Perluigi et al., 2016;Prokopovich et al., 2017;Scheff et al., 2006;Selkoe, 2002;Terry et al., 1991).Although the precise etiology of synaptic degeneration and aberrant phosphorylation in AD is still unknown, evidence suggests that PIN1 may be involved in these processes (Ramakrishnan et al., 2003;Stallings et al., 2018;Xu et al., 2017) Here, we demonstrate that PIN1 is significantly down-regulated as a function of age (22-99 years) and further downregulated in AD limbic and neocortical brain regions in females more than in males.Additional analysis using the ROS/MAP dataset (N=581 participants) revealed a decrease in the expression of PIN1 in the transition to MCI, and a further decrease in the transition to AD, again, driven predominantly by female participants.We also observed that while PIN1 expression is negatively correlated with multiregional β-amyloid in both males and females, it is also negatively correlated with multiregional tangle density, and global cognitive function only in females.These results suggest the potential utility of therapeutic strategies to augment PIN1 function in females with preclinical AD and raise the possibility that monitoring PIN1 could represent an informative disease progression biomarker in female participants with MCI or AD.

PIN1 is a marker of disease progression
Over the past decade, researchers have described the importance of PIN1 in regulating phosphorylation events associated with AD pathogenesis (reviewed in (Balastik et al., 2007;Malter, 2023).Studies of AD animal models and human brain show that PIN1 co-localizes with phosphorylated tau (Kondo et al., 2017;Landrieu et al., 2011;Prokopovich et al., 2017) and shows an inverse relationship to the expression of tau (Holzer et al., 2002).Allen Brain Atlas' ISH data shows that PIN1 mRNA levels are the lowest in regions of the brain affected during the prodromal phase of AD including the hippocampus and the entorhinal cortex.Human studies have shown that allelic distributions of PIN1 single nucleotide polymorphisms (SNPs), in MCI participants, impact upon risk of developing AD.These findings indicate that polymorphisms of the PIN1 gene can predict the path to neurodegeneration (Arosio et al., 2007), but it is not clear if sex is a contributing factor.
Moreover, AD affects the brain decades before the clinical presentation (Pletnikova et al., 2015); our study indicates reduced PIN1 levels in amnestic MCI participants.Whether or not this is a cause or consequence of AD or general aging is unknown.However, it is known that significant PIN1 compensatory events, such as mitochondrial changes, are taking place in the transition state from MCI to disease (Mastroeni et al., 2017a).This idea of disease-associated compensation is not a new phenomenon; similar effects were observed in mitochondrial transcripts (Mastroeni et al., 2017a), as well as in another study in early AD (Reddy et al., 2012).These findings underscore the complexity of the relationship between PIN1, MCI, and AD, opening avenues for further exploration into the molecular mechanisms these associations.

PIN1, tau, and neurofibrillary tangles
Excessive post-translational modifications (PTMs) (e.g., phosphorylation) are known early events in AD, and studies have shown that some phosphorylation events may be sex-specific (Munoz-Mayorga et al., 2018).Although phosphorylation events are critical intermediates of "normal" protein function, excessive phosphorylation is pathogenic.Hyperphosphorylation of tau and amyloid triggers the formation of NFTs and toxic Aβ aggregates, both classical hallmarks of AD neuropathology.Recent efforts have identified PIN1 as a key regulator of the phosphorylation signaling pathway in tau (Ramakrishnan et al., 2003).
PIN1 regulates the dephosphorylation of tau and APP and promotes microtubule assembly by restoring pThr231-tau's ability to bind to them (reviewed in (Lu et al., 2007)).In distal axons, specifically, PIN1 stabilizes Collapsin Response Mediator Protein 2A (CRMP2A).Importantly, CRMP2A plays a role in translating upstream signaling cascades into axon growth (Balastik et al., 2015).Therefore, the downregulation of PIN1 may directly/indirectly affect microtubule assembly and axonal growth, and we provide evidence that this process is sex-specific.We found differences in the distribution of PIN1 protein levels, specifically in axons; AC females showed lower expression, compared to AC males.Interestingly, PIN1 was compactly wrapped around the nuclear compartment reminiscent of rough ER, in AC males.In AD, PIN1-IR appears to be redistributed outside the ER in the cytosolic fraction (in a similar manner in males and females), like what we have observed in other AD studies in our laboratory (Mastroeni et al., 2013(Mastroeni et al., , 2015)).
Our colocalization studies also show the strongest association with ptau 231 compared to the ghost tangle marker ps396, the latest tau marker (reviewed in (Moloney et al., 2021)).In a recent study by Ashton and colleagues, the p-tau 231 assay identifies the clinical stages of AD and neuropathology equally well as the earliest reported p-tau marker, p-tau 181.However, the changes in p-tau 231 are increased earlier, before the threshold for amyloid-β PET positivity, and in response to early brain tau deposition (Ashton et al., 2021).Although these data do not prove that PIN1 may solely be responsible for the initiation, it is clear there is an association, and it would be well to follow up in a larger cohort, especially in females.There is also evidence that tangles may be a compensatory mechanism to aggregate smaller, more toxic oligomeric species into less toxic inert forms (Kopeikina et al., 2012).Future studies focused on the more toxic oligomeric species of tau could also be warranted.

PIN1 expression changes are driven by female participants
Examination of multiple datasets, multiple brain regions, and multiple classes of cells in multiple disease states clearly shows that females express significantly lower levels of PIN1 as a function of age and AD.Data from 581 DLPFC samples from the ROS/MAP cohorts demonstrate heterogeneous associations between PIN1 expression and diverse clinicopathological traits associated with AD, including sex-specific associations.Consistent with our initial analysis, we observed a decline in PIN1 expression in AD, and subsequent stratification by sex revealed that this decline was driven entirely by female participants.While we did observe an inverse correlation of PIN1 with β-Amyloid plaque density and mean SNARE protein density in both males and females, correlations with multiregional tangle density and global cognitive function appear to be female-specific.Given the established literature linking PIN1 activity with synaptic function and NFT burden in AD generally, these findings may suggest an informative sexual dimorphism in PIN1 expression in aging females and AD, including impacts on molecular networks that mediate the impact of NFT burden on cognitive function.
Although these findings suggest that PIN1 may be involved in some of the earliest events in AD, the exact mechanism by which PIN1 is dysregulated requires further study.Moreover, the fact that another peptidyl-prolyl cis/trans isomerase (PIN4) is located on the X chromosome highlights the potential relevance of sex-linked genes in AD, more so than PIN1.However, the lack of observed changes in PIN4 expression levels in the ROS/MAP cohorts (Supp.Figure 3), suggests that while PIN4 may be located on the X chromosome like other potentially sexrelevant genes, its expression levels do not recapitulate expression changes observed in PIN1.Future studies regarding the potential involvement of estrogen or testosterone in the regulation of PIN1 will be important, while we cannot directly measure these levels in this cohort, previous research suggests that estrogen receptor binding can influence the isomerization of phosphorylated Ser/Thr-Pro motifs (Lucchetti et al., 2013).Further investigation into the role of hormones and other sexually dimorphic transcriptional elements near the PIN1 gene could provide valuable insights into the mechanisms underlying sex differences in AD.
In addition, future research is needed to better understand whether PIN1 ER staining patterns are affected by ER stress or not.Many factors contribute to the expression of genes, from transcription factors to epigenetic modifications (e.g., DNA methylation, microRNAs, histone modifications) to genetic variation.Although rare genetic variation in PIN1 was nominally enriched in 227 early-onset Alzheimer's disease cases (Supp.Table 3), further analysis of additional datasets would be beneficial.While further studies are warranted to understand the mechanism of PIN1-mediated regulation of tau phosphorylation (e.g., structural studies to determine the crystal structure of PIN1 in complex with tau and its phosphorylated forms), these results support the potential utility of therapeutic strategies to increase the function of PIN1 in preclinical AD and raise the possibility that monitoring PIN1 could be used for tracking the progression of healthy aging to MCI to AD.

Conclusion
Our study provides compelling evidence that PIN1 expression is significantly down-regulated with aging and further decreases in Alzheimer's Disease (AD), particularly among female participants.This sexspecific decline in PIN1 correlates with key pathological markers of AD, including multiregional β-amyloid and tau density, and is linked to cognitive decline.These findings underscore the potential of PIN1 as both a biomarker for disease progression and a target for therapeutic intervention, especially in the preclinical stages of AD in females.Future research should focus on elucidating the mechanisms underlying PIN1 dysregulation, the role of sex hormones, and the impact of genetic and epigenetic factors on PIN1 expression.Understanding these pathways may lead to novel strategies to mitigate AD progression and improve outcomes for affected individuals.In AD, PIN1 IR appears to be redistributed in the cytosolic fraction in both male and female AD (A & C insert).Both AD and control samples show axonal PIN1 IR but are less reactive in AD samples.Both control and AD samples show positive IR for neurofibrillary tangles (NFT), more abundant in AD than in controls.Cell count analysis shows that females are overall less reactive for PIN1 compared to males in both control (E) and AD samples (F).Colocalization studies in AD tissue using PIN1 (G & J) and late tau marker (p-tau 396, H), and early tau marker (p-tau 231, K) show the strongest association with early tau (p-tau 231) (I vs. L).Punctate IR in AD does not colocalize with tangles (I insert).*p<.05, **p<.001.
C. de Ávila et al.

Fig. 1 .
Fig. 1.PIN1 mRNA levels are downregulated as a function of age.To understand the effect of age and sex we analyzed three brain regions: hippocampus (HIPP), entorhinal cortex (ENT), and superior frontal gyrus (SFG) in an aging cohort: young (n = 22, 20-59 years, mean age 35.4 ± 10.5 years) vs. aged controls (n = 33, age 69-99, mean age 84.2 ± 8.9 years).1A) Black lines show group medians; white lines represent individual data points; polygons represent the estimated density of the data; dotted horizontal lines represent population mean normalized mRNA abundance.Mean expression levels within groups revealed a significant downregulation of PIN1 as a function of age in AC females and males.Female participants (YC and AC) showed lower PIN1 expression levels compared to their male counterparts in all regions except for ENT (1A).Females, however, did show a significant difference (1A).Linear regression analysis revealed a significant negative correlation of PIN1 with age in HIPP (B), ENT (C), and SFG (D) regardless of sex.(p<0.01*,p<0.001**, p<0.0001 ***, NS= not significant).

Fig. 2 .
Fig. 2. PIN1 mRNA levels are reduced in Alzheimer's disease compared to age-matched controls.To address the effect that AD has on the expression of PIN1, we compared the same aged controls (n = 33, age 69-99, mean age 84.2 ± 8.9 years), to clinically and neuropathologically confirmed AD cases (n = 26, ages 74-95 years, mean age 85.7 ± 6.5 years).Center lines show the medians; box limits indicate the 25th and 75th percentiles; whiskers extend to minimum and maximum values; crosses represent sample means; bars indicate 95 % confidence intervals of the means.Data show that females drive the overall significant changes in HIPP, SFG, and ENT.The only significant male-associated change was observed in the HIPP (p=0.03).Black bars are a comparison between AD and AC, these include both males and females.The blue bars are a comparison between AD males and AC males.The red bars are a comparison between AD females and AC females.*=p<0.01,**=p<0.001,***=p<0.0001,NS= Not Significant.

Fig. 3 .
Fig. 3. Impaired cognition and increased neuropathology severity are associated with reduced PIN1 expression in females.(A) PIN1 expression in DLPFC samples from 581 AD, MCI, and AC participants profiled within the ROS/MAP cohorts, (B) stratified by sex.Sex-stratified Spearman correlations between PIN1 expression and (C) multiregional neurofibrillary tangle density, (D) multiregional β-Amyloid density, (E) mean SNARE protein density, and (F) global cognitive function.

Fig. 4 .
Fig.4.Axonal PIN1 protein levels and colocalization with the early tau epitope pThr231.Representative photomicrographs of PIN1 immunoreactivity(IR) in the temporal neocortex in AD and matched control male(♂) and female (♀) samples (A-D).Both AD and control samples showed positive IR for PIN 1.Normal distribution of PIN1 in control samples were clumps of Nissl substance indicative of rough ER (red arrow, B & D insert).Punctate IR in ND extends into the axons (red arrowheads, B).In AD, PIN1 IR appears to be redistributed in the cytosolic fraction in both male and female AD (A & C insert).Both AD and control samples show axonal PIN1 IR but are less reactive in AD samples.Both control and AD samples show positive IR for neurofibrillary tangles (NFT), more abundant in AD than in controls.Cell count analysis shows that females are overall less reactive for PIN1 compared to males in both control (E) and AD samples (F).Colocalization studies in AD tissue using PIN1 (G & J) and late tau marker (p-tau 396, H), and early tau marker (p-tau 231, K) show the strongest association with early tau (p-tau 231) (I vs. L).Punctate IR in AD does not colocalize with tangles (I insert).*p<.05, **p<.001.
Tissue samples were obtained from seven well-established National Institute on Aging Alzheimer's Disease brain banks located at the University of California Irvine, Sun Health Research Institute, University of Rochester, Johns Hopkins University, the National Institute of Child Health and Human Development Brain and Tissue Bank for Developmental Disorders at the University of Maryland, University of Pennsylvania, and the University of Southern California.