Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses

Abstract

Background

Alzheimer's disease (AD) is a major neurodegenerative disorder with limited treatment options.

Objective

This study aimed to identify novel therapeutic targets for AD using proteome-wide Mendelian randomization (MR) and colocalization analyses.

Methods

We conducted a large-scale, proteome-wide MR analysis using data from two extensive genome-wide association studies (GWASs) of plasma proteins: the UK Biobank Pharma Proteomics Project (UKB-PPP) and the deCODE Health Study. We extracted genetic instruments for plasma proteins from these studies and utilized AD summary statistics from European Bioinformatics Institute GWAS Catalog. Colocalization analysis assessed whether identified associations were due to shared causal variants. Phenome-wide association studies and drug repurposing analyses were performed to assess potential side effects and identify existing drugs targeting the identified proteins.

Results

Our MR analysis identified significant associations between genetically predicted levels of 9 proteins in the deCODE dataset and 17 proteins in the UKB-PPP dataset with AD risk after Bonferroni correction. Four proteins (BCAM, CD55, CR1, and GRN) showed consistent associations across both datasets. Colocalization analysis provided strong evidence for shared causal variants between GRN, CR1, and AD. PheWAS revealed minimal potential side effects for CR1 but suggested possible pleiotropic effects for GRN. Drug repurposing analysis identified several FDA-approved drugs targeting CR1 and GRN with potential for AD treatment.

Conclusions

This study identifies GRN and CR1 as promising therapeutic targets for AD. These findings provide new directions for AD drug development, but further research and clinical trials are warranted to validate the therapeutic potential of these targets.

Keywords

Alzheimer's disease biomarkers colocalization pharmaceutical targets proteome-wide Mendelian randomization

Introduction

Alzheimer's disease (AD) is a major neurodegenerative disorder characterized by cognitive decline and dementia.¹ Current projections show that 6.5 million Americans aged 65 or above are diagnosed with AD, a statistic anticipated to increase twofold by 2060, reaching 13.8 million people.² Given its increased frequency with advancing age, AD represents an escalating public health threat, dovetailing with the swift global demographic ageing phenomenon.³ Over the last three decades, there has been an increasingly discernible surge in the prevalence of AD worldwide.³ This rapid expansion, coupled with its substantial societal and economic ramifications, underscores the imperative need for efficacious interventions and therapeutic modalities for AD.⁴

Although significant strides have been made in understanding the pathogenesis and diagnosis of AD, effective therapies to prevent, halt, or cure the condition remain elusive.⁵ In recent years, circulating proteins have emerged as a focal point for researchers, with numerous studies utilizing these proteins to uncover novel therapeutic targets for various diseases.^6,7 Some researchers hypothesize that circulating proteins not only contribute to disease risk but may also serve as promising targets for therapeutic interventions.⁸ This is due to their accessibility in the bloodstream, where they can be measured to provide early indicators of therapeutic efficacy.⁹ However, deciphering their role in disease causation presents challenges, as randomized controlled trials are often impractical without robust evidence of efficacy. Observational studies investigating circulating proteins in psychiatric disorders are particularly prone to bias. Confounding factors, such as socioeconomic variables,¹⁰ are often difficult to quantify and control in these studies. Additionally, reverse causation complicates findings, as psychiatric disorders may alter behaviors and biological pathways, which subsequently affect circulating protein levels.¹¹

A methodological approach to address these biases is Mendelian randomization (MR), which employs an instrumental variable framework to infer causal relationships while minimizing reverse causation.^8,12 MR leverages genetic variants as instruments to evaluate the causal effects of exposures, such as circulating proteins, on disease outcomes. Recent advancements in identifying genetic variants associated with circulating protein levels—known as protein quantitative trait loci (pQTLs)^13,14—in large-scale population studies have opened new avenues for pinpointing proteins with causal links to disease. Genetic variants located near protein-coding genes are particularly valuable as instruments because they likely exert a strong, direct influence on protein expression at the transcriptional or translational level, thereby reducing the risk of horizontal pleiotropy.¹⁵ Consequently, MR has been increasingly applied to investigate the causal roles of circulating biomarkers in complex diseases, including cancer,¹⁶ coronary heart disease¹⁷ and ischemic stroke.¹⁸

Several prior MR studies have investigated the relationship between blood proteins and AD. Building on these efforts, we performed a protein-wide MR analysis using data on a larger set of blood proteins derived from two extensive, large-scale studies. This analysis, complemented by colocalization methods, examined nearly 2000 plasma proteins to evaluate their potential associations with AD. By integrating these approaches, our study aims to provide novel perspectives on the underlying mechanisms of AD pathogenesis and to inform the development of more targeted and holistic therapeutic strategies for this complex disease.

Methods

Data sources for plasma proteins

Instrumental variables were selected as cis-single nucleotide polymorphism (cis-SNP) significantly associated with plasma protein levels at the genome-wide significance threshold (p < 5 × 10⁻⁸) from two large-scale genome-wide association studies (GWASs): the deCODE Health study¹⁹ and the UK Biobank Pharma Proteomics Project (UKB-PPP).²⁰ Cis-SNPs were defined as single nucleotide polymorphisms located within 1 Mb of the gene encoding the respective protein. Linkage disequilibrium (LD) was calculated using data from the 1000 Genomes European reference panel. Within a 1 Mb window, SNPs with LD values (r²) less than 0.001 were considered independent.

The UKB-PPP study performed proteomic profiling on blood plasma samples from 54,306 participants using the Olink platform, enabling the measurement of 1991 proteins.²⁰ For the two-sample MR analysis, index cis-SNPs were identified as instrumental variables for 1896 of these proteins. Similarly, the deCODE Health study identified index cis-SNPs for 1755 plasma proteins. This study utilized the SomaScan platform to measure 4907 aptamers in plasma samples collected from 35,559 Icelandic participants.¹⁹

Data sources for AD

The GWAS summary statistics data for AD were acquired from a published study in 2022, encompassing 111,326 AD patients and 677,663 controls.²¹ This dataset was sourced from the European Bioinformatics Institute GWAS Catalog (https://gwas.mrcieu.ac.uk/) with accession number GCST90027158.

Quantification and statistical analysis

MR analysis

To investigate potential causal relationships between circulating proteins identified in proteomic studies and AD, we applied a two-sample MR approach. MR is a powerful epidemiological tool that enhances causal inference by leveraging genetic variants as instrumental variables to proxy for the exposure of interest, such as circulating protein levels. This method relies on three core assumptions: (1) the genetic variants chosen as instrumental variables must be strongly associated with the exposure; (2) the genetic instruments must be independent of any confounding factors; and (3) the genetic instruments should affect the outcome solely through the exposure and not via alternative pathways. MR has been widely employed in prior research to examine links between plasma proteins and various health outcomes, demonstrating that these assumptions are generally valid when cis-variants within or near protein-coding genes are used as genetic instruments for the proteins.

For each protein, the deCODE and UKB-PPP cohorts identified an index cis-SNP as the variant with the lowest p-value among all SNPs associated with the protein.^11,22 MR analyses were conducted using the TwoSampleMR package in R. For proteins with only one genetic instrument, the Wald ratio test was applied to evaluate the potential causal relationship between each protein and AD. Protein-disease associations were deemed significant if they met a corrected p-value threshold (false discovery rate, FDR) of < 0.05. In the deCODE cohort, the corrected p-value threshold was 2.86 × 10⁻⁵, whereas in the UKB-PPP cohort, it was 2.63 × 10⁻⁵.

Colocalization analysis

A Bayesian model-informed colocalization assessment was performed to determine whether any correlations detected between proteins and AD were attributable to linkage disequilibrium. This evaluation was rooted in five nonoverlapping hypotheses: 1) lack of association with either characteristic; 2) exclusive correlation with the initial characteristic; 3) exclusive correlation with the subsequent characteristic; 4) simultaneous correlations with both characteristics, albeit propelled by distinct causal variants; and 5) simultaneous correlations with both characteristics, bolstered by a mutual causal variant. Each hypothesis (H0 through H4) was allotted a posterior probability. Significant evidence of colocalization between two signals was surmised if the posterior probability for shared causal variants (PH4) surpassed 0.8.²² This analysis was conducted employing the ‘coloc’ package within the R software infrastructure (version 4.4.1).²³

Phenome-wide association analysis

To further explore the horizontal pleiotropy of prospective drug targets and potential side effects, we conducted a phenome-wide association study (PheWAS) utilizing the AstraZeneca PheWAS Portal (https://azphewas.com/).²⁴ In this analysis, the correlations between rare protein-coding genetic variants and 18,780 traits were investigated within the UK Biobank cohort.²⁴ Multiple corrections were applied, and a threshold of 2E−8 was set, in line with the default setting in the AstraZeneca PheWAS Portal, to mitigate the risk of false-positives.

Drug repurposing

To obtain insights into the drug compounds targeting proteins identified in this study, we utilized the Comparative Toxicogenomics Database (CTD; http://ctdbase.org/)²⁵ (as of 25/3/2024). This database offers manually curated information on 50 million toxicogenomic relationships. We manually cross-referenced the identified drugs with the FDA's Approved Drug Products database (https://www.accessdata.fda.gov/scripts/cder/daf/) to determine their approval status. Our analysis focused primarily on the protein targets of each drug. Furthermore, we identified drugs aimed at these two proteins that are currently undergoing clinical trials using the ChEMBL²⁶ database (as of 24/3/2024).

Results

A schematic representation of the study design is presented in Figure 1. All analyses relied on the summary-level data delineated in Table 1.

Figure 1.

Study design overview.

Table 1.

Data sources for studied phenotypes.

Study	Phenotype	Cases	Controls	PMID	Adjustment
UKB-PPP	Plasma protein	54,306		Preprint	age, gender, batch, UKB center, UKB genetic array, time between blood sampling and measurement and the first 20 genetic principal components
deCODE	Plasma protein	35,559		34857953	age, gender, and sample age
European Bioinformatics Institute GWAS	AD	111,326	677,663	35379992

Associations between plasma proteins and AD

Figure 2 illustrates the summarized outcomes of the AD analyses. Genetically predicted levels of 9 proteins and 17 proteins exhibited significant associations with AD risk post-Bonferroni correction for multiple testing (p < 2.65 × 10⁻⁵) in the deCODE and UKB-PPP datasets, respectively (Figure 2A). For each standard deviation increase in genetically predicted protein levels, the odds ratio for AD varied from 0.74 (95% CI, 0.68–0.81) for progranulin protein (GRN) to 2.15 (95% CI, 1.90–2.44) for myc box-dependent-interacting protein 1 protein (BIN1) in the deCODE dataset and from 0.48 (95% CI, 0.44–0.53) for apolipoprotein E protein (APOE) to 2.06 (95% CI, 1.54–2.74) for B-cell linker protein (BLNK) in the UKB-PPP dataset (Figure 2B, Supplemental Tables 1 and 2). The identified associations demonstrated directional consistency between the discovery and replication studies for all proteins except basal cell adhesion molecule (BCAM) (Supplemental Table 3). Among the identified associations, four proteins were evaluated across two distinct profiling platforms and generally exhibited consistent associations with AD, including BCAM, CD55, complement_receptor_type_1 (CR1), and GRN (Supplemental Table 4). BCAM exhibited directionally discordant associations with AD across the deCODE and UKB-PPP cohorts. Elevated genetically predicted BCAM levels demonstrated a protective effect in the deCODE dataset (OR = 0.86, 95% CI: 0.81–0.92, p = 2.2 × 10⁻⁶), while paradoxically associating with elevated AD risk in the UKB-PPP cohort (OR = 1.63, 95% CI: 1.50–1.77, p = 8.1 × 10–26). The remaining studied proteins did not exhibit associations with AD risk (Supplemental Tables 1 and 2).

Figure 2.

Result summary of MR and colocalization analysis on the associations between plasma proteins and the risk of AD.

Among the four MR-identified proteins associated with AD, two proteins garnered substantial support from colocalization analysis (PH4 ≥ 0.8) (Supplemental Figures 1–4 and Supplemental Table 5), namely, GRN and CR1.

PheWAS

To comprehensively evaluate whether the potential drug target genes GRN and CR1 identified might have beneficial or adverse effects on other traits and to explore potential pleiotropy not captured by the MR‒Egger intercept test, 17,361 dichotomous phenotypes and 1419 quantitative phenotypes from the AstraZeneca PheWAS Portal database¹⁵ were used to conduct PheWAS at the gene level. PheWAS outcomes can be interpreted as the correlation between genetically determined protein expression and specific diseases or traits. As delineated in Supplemental Tables 6 and 7 and Supplemental Figures 5 and 6, CR1 exhibited no significant association with other traits at the gene level (p < 2E−8 for genomic association), indicating that the potential side effects of drugs targeting CR1 and the presence of horizontal pleiotropy in CR1 are likely minimal, further corroborating the reliability of the study's findings. Conversely, GRN was linked to diseases of the nervous system and mental and behavioral disorders, suggesting that drugs targeting the GRN gene may impact both traits and that MR analysis of the GRN gene might have a pleiotropic influence on the results.

Drug repurposing

The CTD database (https://ctdbase.org/) was queried for CR1 drug targets, revealing 9 FDA-approved drugs with potential for treating AD among 32 interacting chemicals (Supplemental Table 8). Furthermore, 34 FDA-approved drugs with potential for treating AD were among the 173 interacting chemicals targeting GRN (Supplemental Table 8). Notably, five drugs—acetaminophen, oestradiol, methotrexate, tretinoin, and valproic acid—interacted with both CR1 and GRN targets. Additionally, a search of the ChEMBL database revealed a drug targeting CR1 currently being tested in clinical trials (APT070).

Discussion

In this study, we combined proteome-wide MR and colocalization analyses to identify potential therapeutic targets for AD. Applying stringent Bonferroni correction for multiple testing (p < 2.86 × 10⁻⁵ for deCODE and p < 2.63 × 10⁻⁵ for UKB-PPP), our MR analysis revealed significant associations between genetically predicted levels of 9 proteins in the deCODE dataset and 17 proteins in the UKB-PPP dataset with AD risk. Among them, colocalization analysis supported the causal role of CR1 and GRN in the pathogenesis of AD. Remarkably, despite the distinct technologies and methodologies employed by the two proteomic profiling platforms, we observed consistent associations for BCAM, CD55, complement receptor type 1 (CR1), and GRN across both datasets. This consistency not only reduces the likelihood of platform-specific biases but also reinforces the validity of these proteins as potential therapeutic targets for AD, strengthening the foundation for further investigation into their roles in AD pathogenesis and their potential as drug targets. These findings offer new insights into the biological mechanisms underlying AD pathogenesis and highlight promising avenues for drug development.

Several researchers, including Fang et al.,²⁷ conducted their analyses using the UKB-PPP plasma proteomics database, a comprehensive resource for plasma protein quantification. Belbasis et al.²⁸ utilized both UKB-PPP and deCODE databases in a sequential manner, but primarily relied on cis-pQTLs from the UKB-PPP platform, even for proteins with instruments available in both databases, potentially limiting the utilization of complementary genetic instruments. Regarding tissue specificity, Hu et al.²⁹ extended their analysis to pQTL data from three tissues—brain, cerebrospinal fluid, and plasma—providing a multi-tissue perspective on protein expression relevant to AD. Ou et al.³⁰ uniquely concentrated on proteomics data from postmortem brain tissue samples and subsequently validated their findings using blood proteomics data, offering insights into potential translational biomarkers. In contrast, our study represents the only investigation that comprehensively integrates both UKB-PPP and deCODE large-scale plasma proteomics databases simultaneously in a unified analytical framework, thereby enhancing statistical power and robustness of our causal inference.

Our stringent analytical framework enhanced the clinical potential of identified targets like CR1 and GRN, while necessarily increasing the risk of overlooking potentially relevant proteins—notably APOE, a crucial component of plasma lipoproteins that transports lipids between cells across various organs and tissues. Our study found no significant causal relationship between plasma APOE levels and AD, despite extensive clinical evidence establishing APOE as a major AD risk factor.³¹ This apparent contradiction reflects a consistent pattern across methodologically diverse studies: UKB-PPP plasma-based investigations by Fang,²⁷ Belbasis,²⁸ and initially Zhan³² all reported a protective effect of circulating APOE against AD (ORs ranging from 0.57 to 0.616), while tissue-specific analyses revealed opposing relationships. Notably, Hu²⁹ found APOE to be a risk factor exclusively in brain tissue but not in blood or CSF, and Ou similarly identified APOE as significantly increasing AD risk in both brain (OR = 1.728) and blood (OR = 1.117).³⁰ These conflicting directions of effect likely stem from APOE's tissue-specific functions. Supporting this hypothesis, the recent large-scale cerebrospinal fluid (CSF) proteogenomic study by Western et al.³³ demonstrated that APOE exhibits markedly different regulatory patterns in CSF compared to plasma. They found that the APOE region on chromosome 19q13.32 was associated with substantially more protein quantitative trait loci (pQTLs) in CSF than in plasma (13.8% of all CSF pQTLs versus only 0.67% of all plasma pQTLs, p < 2.2 × 10⁻¹⁶).³³ Additionally, they identified APOE as the largest pleiotropic region in CSF, regulating 335 different proteins, a pattern not observed to the same extent in plasma. This tissue-specific regulation suggests that circulating APOE levels measured in plasma may inadequately represent the protein's pathological activity in brain tissue, where its contribution to AD pathogenesis is most pronounced.

Another puzzling result was the opposite associations of BCAM with AD risk in the deCODE and UKB-PPP datasets. Several factors could contribute to this discrepancy. One key factor is the difference in the proteomic platforms used: the deCODE study employed the SomaScan assay, while the UKB-PPP used the Olink platform. These platforms have distinct protein quantification methods, antibody specificity, and dynamic ranges, which could lead to divergent results for some proteins. Moreover, it is plausible that BCAM, like APOE, may have multiple isoforms or splice variants that could contribute to the inconsistent findings. Different isoforms of BCAM may have distinct functional properties or may be differentially expressed in various tissues, leading to heterogeneous effects on AD risk. The SomaScan and Olink platforms may have different sensitivities and specificities for detecting these isoforms, further compounding the discrepancies observed between the two datasets.

In addition to the proteomic platform differences, the populations studied differed between the two datasets (Icelandic population in deCODE vs. UK Biobank participants in UKB-PPP), and there may be underlying genetic, environmental, or demographic factors that modify the BCAM-AD association. Another important consideration is the difference in statistical power between the two studies. The deCODE dataset included 35,559 individuals, while the UKB-PPP had a larger sample size of 54,306 participants. The increased power in the UKB-PPP study may have allowed for the detection of a significant association in the opposite direction compared to the deCODE results.

These findings highlight the complexity and challenges of integrating genetic, proteomic, and phenotypic data to comprehensively understand disease biology, underscoring the importance of considering technical factors when interpreting and comparing results across different studies.

To further explore the potential safety and pleiotropy of the identified targets, we performed PheWAS. Progranulin, encoded by GRN, is a secreted protein that regulates lysosomal function, neuronal survival, and inflammation.³⁴ Decreased progranulin levels are associated with elevated AD risk,³⁵ and our study bolsters this causative link. Cognitive decline in AD correlates with synapse loss, and neuroinflammation, mediated by microglia and complement, is prominent in later stages.³⁶ CR1 has emerged as a potential risk factor for neuroinflammation.³⁷ PheWAS indicated that CR1-targeting drugs might have minimal side effects, while GRN might exhibit pleiotropy. Drug repurposing analysis identified five drugs interacting with both CR1 and GRN, showing promise for AD treatment. These findings provide new clues for biomarker and drug development, contribute to understanding AD pathogenesis, and lay the foundation for developing effective therapeutic strategies.

Building upon these insights, our study identified several medications, including acetaminophen, oestradiol, methotrexate, tretinoin, and valproic acid (VPA), that could potentially be repurposed for AD treatment. However, drugs targeting GRN may impact diseases of the nervous system and mental and behavioral disorders. VPA, a mood stabilizer and antiepileptic drug, shows neuroprotective potential by influencing PGRN expression, apoptosis-related proteins, neuron counts, and synaptic density.³⁸ Oestradiol supports neurogenesis, synaptic plasticity, and modulates inflammation, with combined VPA and 17β-oestradiol treatment showing promising results in AD model mice.^39,40 Protein aggregation is linked to various pathological conditions, including AD, and methotrexate shows promise for repurposing preformed α-chymotrypsinogen A aggregates.⁴¹ Tretinoin, a retinoid, maintains adult neuron differentiation, and preclinical studies advocate for retinoids to regulate Aβ formation and aggregation.⁴² New immunoneurological approaches focus on correcting immune surveillance deficiencies, with clinical trials exploring drugs that increase progranulin levels to modulate microglial function and APT070 (Mirococept), a complement inhibitor, showing promise in suppressing inflammation and providing neuroprotection.⁴³

Despite the strengths of our study, our study has several limitations. First, despite our efforts to minimize the risk of violating Mendelian randomization assumptions by using a single index cis-SNP as the instrumental variable for each protein, residual confounding or horizontal pleiotropy cannot be entirely ruled out. Second, while we focused on the subset of proteins consistently quantified across both the deCODE and UKB-PPP platforms to minimize heterogeneity, this approach may limit the number of proteins analyzed, because the deCODE and UKBPPP are limited and cannot cover all proteins. Third, our findings may not be generalizable to non-European populations, as the analysis was limited to individuals of European ancestry. Furthermore, as these associations were established through in silico analyses, they require further validation through animal studies and population-based research. Therefore. The next step of this study will involve experimental validation, such as evaluating the expression levels of CR1 and GRN in patients’ peripheral blood through quantitative reverse transcription (RT)-PCR, protein imprinting, and other methods.

Conclusion

In conclusion, our study combines proteome-wide MR and colocalization analyses to identify potential AD therapeutic targets. We leveraged large-scale genetic and proteomic data from two independent cohorts. Our findings revealed significant associations between genetically predicted levels of several plasma proteins and AD risk. The strongest evidence was found for a shared causal variant between GRN, CR1, and AD. These results not only deepen our understanding of AD pathogenesis but also provide clear directions for future drug development and clinical trials in AD treatment.

Supplemental Material

sj-docx-2-alz-10.1177_13872877251344572 - Supplemental material for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses

Supplemental material, sj-docx-2-alz-10.1177_13872877251344572 for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses by Kefu Yu, Ruiqi Jiang, Dabiao Zhou and Zhigang Zhao in Journal of Alzheimer's Disease

Supplemental Material

sj-xlsx-3-alz-10.1177_13872877251344572 - Supplemental material for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses

Supplemental material, sj-xlsx-3-alz-10.1177_13872877251344572 for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses by Kefu Yu, Ruiqi Jiang, Dabiao Zhou and Zhigang Zhao in Journal of Alzheimer's Disease

Supplemental Material

sj-xlsx-4-alz-10.1177_13872877251344572 - Supplemental material for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses

Supplemental material, sj-xlsx-4-alz-10.1177_13872877251344572 for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses by Kefu Yu, Ruiqi Jiang, Dabiao Zhou and Zhigang Zhao in Journal of Alzheimer's Disease

Supplemental Material

sj-xlsx-5-alz-10.1177_13872877251344572 - Supplemental material for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses

Supplemental material, sj-xlsx-5-alz-10.1177_13872877251344572 for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses by Kefu Yu, Ruiqi Jiang, Dabiao Zhou and Zhigang Zhao in Journal of Alzheimer's Disease

Supplemental Material

sj-xlsx-6-alz-10.1177_13872877251344572 - Supplemental material for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses

Supplemental material, sj-xlsx-6-alz-10.1177_13872877251344572 for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses by Kefu Yu, Ruiqi Jiang, Dabiao Zhou and Zhigang Zhao in Journal of Alzheimer's Disease

Supplemental Material

sj-xlsx-7-alz-10.1177_13872877251344572 - Supplemental material for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses

Supplemental material, sj-xlsx-7-alz-10.1177_13872877251344572 for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses by Kefu Yu, Ruiqi Jiang, Dabiao Zhou and Zhigang Zhao in Journal of Alzheimer's Disease

Supplemental Material

sj-xlsx-8-alz-10.1177_13872877251344572 - Supplemental material for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses

Supplemental material, sj-xlsx-8-alz-10.1177_13872877251344572 for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses by Kefu Yu, Ruiqi Jiang, Dabiao Zhou and Zhigang Zhao in Journal of Alzheimer's Disease

Supplemental Material

sj-xlsx-9-alz-10.1177_13872877251344572 - Supplemental material for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses

Supplemental material, sj-xlsx-9-alz-10.1177_13872877251344572 for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses by Kefu Yu, Ruiqi Jiang, Dabiao Zhou and Zhigang Zhao in Journal of Alzheimer's Disease

Supplemental Material

sj-xlsx-10-alz-10.1177_13872877251344572 - Supplemental material for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses

Supplemental material, sj-xlsx-10-alz-10.1177_13872877251344572 for Therapeutic targets for Alzheimer's disease: Proteome-wide Mendelian randomization and colocalization analyses by Kefu Yu, Ruiqi Jiang, Dabiao Zhou and Zhigang Zhao in Journal of Alzheimer's Disease

Footnotes

Acknowledgements

The authors thank the Genetic Investigation of deCODE Consortium, UK biobank, the MEGASTROKE GWAS dataset, and all concerned investigators for sharing GWAS summary statistics on plasma proteins and AD.

ORCID iDs

Ruiqi Jiang

Zhigang Zhao

Ethical considerations

There was no need to get informed consent or ethical approval for this study again because all data were taken from published sources.

Consent to participate

Informed consent and approval were received.

Consent for publication

All participants provided verbal consent for publication.

Author contributions

Kefu Yu (Conceptualization; Data curation; Formal analysis; Methodology; Software; Validation; Writing – original draft); Ruiqi Jiang (Data curation; Formal analysis; Investigation; Resources; Validation; Writing – original draft); Dabiao Zhou (Formal analysis; Funding acquisition; Supervision; Writing – original draft; Writing – review & editing); Zhigang Zhao (Conceptualization; Project administration; Supervision; Writing – original draft; Writing – review & editing).

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Beijing Hospitals Authority Clinical medicine Development of special funding support, code: ZLRK202508.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

The original contributions presented in the study are included in the article. Further inquiries can be directed to the corresponding author.

Supplemental material

Supplemental material for this article is available online.

References

Brookmeyer

Johnson

Ziegler-Graham

, et al. Forecasting the global burden of Alzheimer's disease. Alzheimers Dement 2007; 3: 186–191.

Alzheimer's Association. 2022 Alzheimer's disease facts and figures. Alzheimers Dement 2022; 18: 700–789.

GBD 2016 Dementia Collaborators. Global, regional, and national burden of Alzheimer's disease and other dementias, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol 2019; 18: 88–106.

Schwam

Abu-Shakra

del Valle

, et al. Health economics and the value of therapy in Alzheimer's disease. Alzheimers Dement 2007; 3: 143–151.

Ogos

Stary

Bajda

. Recent advances in the search for effective anti-Alzheimer's drugs. Int J Mol Sci 2024; 26: 157.

Chen

Ruan

, et al. Therapeutic targets for inflammatory bowel disease: proteome-wide Mendelian randomization and colocalization analyses. EBioMedicine 2023; 89: 104494.

Rasooly

Peloso

Pereira

, et al. Genome-wide association analysis and Mendelian randomization proteomics identify drug targets for heart failure. Nat Commun 2023; 14: 3826.

Schmidt

Finan

Gordillo-Marañón

, et al. Genetic drug target validation using Mendelian randomisation. Nat Commun 2020; 11: 3255.

Finan

Gaulton

Kruger

, et al. The druggable genome and support for target identification and validation in drug development. Sci Transl Med 2017; 9: eaag1166.

10.

Hemani

Bowden

and

Davey Smith

Evaluating the potential role of pleiotropy in Mendelian randomization studies

. Hum Mol Genet 2018; 27: R195–R208.

11.

Forgetta

Greenwood

CMT

, et al. Circulating proteins influencing psychiatric disease: a Mendelian randomization study. Biol Psychiatry 2023; 93: 82–91.

12.

Dohrenwend

Levav

Shrout

, et al. Socioeconomic status and psychiatric disorders: the causation-selection issue. Science 1992; 255: 946–952.

13.

Skrivankova

Richmond

Woolf

BAR

, et al. Strengthening the reporting of observational studies in epidemiology using Mendelian randomization: the STROBE-MR statement. JAMA 2021; 326: 1614–1621.

14.

Pietzner

Wheeler

Carrasco-Zanini

, et al. Mapping the proteo-genomic convergence of human diseases. Science 2021; 374: eabj1541.

15.

Sun

Maranville

Peters

, et al. Genomic atlas of the human plasma proteome. Nature 2018; 558: 73–79.

16.

Sun

Zhao

Jiang

, et al. Identification of novel protein biomarkers and drug targets for colorectal cancer by integrating human plasma proteome with genome. Genome Med 2023; 15: 75.

17.

Sun

Yun

Lin

, et al. Comprehensive Mendelian randomization analysis of plasma proteomics to identify new therapeutic targets for the treatment of coronary heart disease and myocardial infarction. J Transl Med 2024; 22: 404.

18.

Chong

Sjaarda

Pigeyre

, et al. Novel drug targets for ischemic stroke identified through Mendelian randomization analysis of the blood proteome. Circulation 2019; 140: 819–830.

19.

Ferkingstad

Sulem

Atlason

, et al. Large-scale integration of the plasma proteome with genetics and disease. Nat Genet 2021; 53: 1712–1721.

20.

Sun

Chiou

Traylor

, et al. Plasma proteomic associations with genetics and health in the UK Biobank. Nature 2023; 622: 329–338.

21.

Bellenguez

Küçükali

Jansen

, et al. New insights into the genetic etiology of Alzheimer's disease and related dementias. Nat Genet 2022; 54: 412–436.

22.

Yuan

, et al. Plasma proteins and onset of type 2 diabetes and diabetic complications: proteome-wide Mendelian randomization and colocalization analyses. Cell Rep Med 2023; 4: 101174.

23.

Foley

Staley

Breen

, et al. A fast and efficient colocalization algorithm for identifying shared genetic risk factors across multiple traits. Nat Commun 2021; 12: 764.

24.

Wang

Dhindsa

Carss

, et al. Rare variant contribution to human disease in 281,104 UK Biobank exomes. Nature 2021; 597: 527–532.

25.

Davis

Wiegers

Johnson

, et al. Comparative Toxicogenomics Database (CTD): update 2023. Nucleic Acids Res 2023; 51: D1257–D1262.

26.

Gaulton

Bellis

Bento

, et al. ChEMBL: a large-scale bioactivity database for drug discovery. Nucleic Acids Res 2012; 40: D1100–D1107.

27.

Fang

Xue

Lin

, et al. Multivariate proteome-wide association study to identify causal proteins for Alzheimer disease. Am J Hum Genet 2025; 112: 291–300.

28.

Belbasis

Morris

van Duijn

, et al. Mendelian randomization identifies proteins involved in neurodegenerative diseases. Brain. Epub ahead of print 3 March 2025: awaf018. DOI: https://doi.org/10.1093/brain/awaf018

29.

Liu

Dai

, et al. Proteome-wide association studies using summary pQTL data of three tissues identified 30 risk genes of Alzheimer's disease dementia. medRxiv 2024. Doi: https://doi.org/10.1101/2024.03.28.24305044 [Preprint]. Posted September 04, 2024.

30.

Yang

Deng

, et al. Identification of novel drug targets for Alzheimer's disease by integrating genetics and proteomes from brain and blood. Mol Psychiatry 2021; 26: 6065–6073.

31.

Raulin

Doss

Trottier

, et al. Apoe in Alzheimer's disease: pathophysiology and therapeutic strategies. Mol Neurodegener 2022; 17: 72.

32.

Zhan

Cammann

Cummings

, et al. Biomarker identification for Alzheimer's disease through integration of comprehensive Mendelian randomization and proteomics data. J Transl Med 2025; 23: 278.

33.

Western

Timsina

Wang

, et al. Proteogenomic analysis of human cerebrospinal fluid identifies neurologically relevant regulation and implicates causal proteins for Alzheimer's disease. Nat Genet 2024; 56: 2672–2684.

34.

Rhinn

Tatton

McCaughey

, et al. Progranulin as a therapeutic target in neurodegenerative diseases. Trends Pharmacol Sci 2022; 43: 641–652.

35.

Vardarajan

Reyes-Dumeyer

Piriz

, et al. Progranulin mutations in clinical and neuropathological Alzheimer's disease. Alzheimers Dement 2022; 18: 2458–2467.

36.

Hong

Beja-Glasser

Nfonoyim

, et al. Complement and microglia mediate early synapse loss in Alzheimer mouse models. Science 2016; 352: 712–716.

37.

Kiraly

Foss

Giordano

. Neuroinflammation, its role in Alzheimer's disease and therapeutic strategies. J Prev Alzheimers Dis 2023; 10: 686–698.

38.

Zhang

. Valproic acid as a promising agent to combat Alzheimer's disease. Brain Res Bull 2010; 81: 3–6.

39.

Liu

Zhang

, et al. Combined treatment with valproic acid and estrogen has neuroprotective effects in ovariectomized mice with Alzheimer's disease. Neural Regen Res 2021; 16: 2078–2085.

40.

Mishra

Davies

. The interaction between NF-κB and estrogen in Alzheimer's disease. Mol Neurobiol 2023; 60: 1515–1526.

41.

Ansari

Rais

Naeem

. Methotrexate for drug repurposing as an anti-aggregatory agent to mercuric treated α-chymotrypsinogen-A. Protein J 2024; 43: 362–374.

42.

Chakrabarti

McDonald

Will Reed

, et al. Molecular signaling mechanisms of natural and synthetic retinoids for inhibition of pathogenesis in Alzheimer's disease. J Alzheimers Dis 2016; 50: 335–352.

43.

Souza

Esser

Bradford

, et al. APT070 (Mirococept), a membrane-localised complement inhibitor, inhibits inflammatory responses that follow intestinal ischaemia and reperfusion injury. Br J Pharmacol 2005; 145: 1027–1034.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.83 MB

0.01 MB

0.10 MB

0.13 MB

0.01 MB

0.36 MB

0.27 MB