Genetically determined levels of mTOR-dependent circulating proteins and risk of multiple sclerosis

doi:10.21203/rs.3.rs-2474348/v1

Download PDF

Research Article

Genetically determined levels of mTOR-dependent circulating proteins and risk of multiple sclerosis

https://doi.org/10.21203/rs.3.rs-2474348/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

Observational studies have reported an association between circulating levels of mammalian target of rapamycin (mTOR)-dependent circulating proteins and multiple sclerosis (MS). However, the casual association has not been fully elucidated. Mendelian randomization (MR) is used to overcome limitations inherent to observational studies and assess the causal association.

Methods

To explore the causal association between mTOR-dependent proteins (AKT, RP-S6K, eIF4E-BP, eIF4A, eIF4E, eIF4G, and PKC-α) and MS, summary statistics were obtained from GWAS meta-analysis of the International Multiple Sclerosis Genetics Consortium (47429 patients and 68374 controls) and the INTERVAL study (genetic associations with 2994 plasma proteins from 3301 healthy individuals). MR analysis and sensitivity analyses were conducted.

Results

Among seven selected mTOR-dependent proteins, the circulating level of PKC-α (OR = 0.90, 95%CI 0.82–0.98, P = 0.017) and RP-S6K (OR = 1.12, 95%CI 1.00-1.25, P = 0.045) were associated with MS risk, while no significant causation was found between other proteins (AKT, eIF4E-BP, eIF4A, eIF4E, eIF4G) and MS.

Conclusion

Molecules in the mTOR signaling pathway may bidirectionally regulate the occurrence and development of MS. PKC-α is a protective factor, while RP-S6K is a risk factor. They might be used as future therapeutic targets for screening high-risk individuals.

Multiple Sclerosis

Mendelian Randomization Analysis

mTOR Associated Protein

Molecular Epidemiology

Multiple sclerosis (MS) is an inflammatory-neurodegenerative disease of the central nervous system (CNS) which causes a range of symptoms such as blurred vision, weak limbs, tingling sensations, dizziness, and fatigue[1]. Discrete episodes of neurologic dysfunction may be relapses, flares, attacks, or exacerbations, which may be pretty devastating to the brainstem and spinal cord, resulting in poor recovery. Even in the advanced stages, patients’ symptoms could be vague, although axonal injury has been caused[2]. Thus, detecting the condition as early as possible is very important.

The mechanistic target of rapamycin (mTOR) is a 289-kDa serine/threonine protein kinase that serves as a downstream effector of the PI3K/AKT pathway[3]. The mTOR is a primary component of two distinct protein complexes: rapamycin-sensitive mTORC1 and rapamycin-insensitive mTORC2. mTORC1 is generally activated in response to various extracellular and intracellular stimuli, including hormones, growth factors, cytokines, oxidants, and amino acids, which occur on lysosomal membranes[4]. It regulates protein translation via the phosphorylation of ribosomal protein S6 kinase (RP-S6K) and eukaryotic initiation factor 4E binding protein (eIF4E-BP)[5]. eIF4E-BP inhibits translation initiation by sequestering translation factor eIF4E from the eIF4F complex (eIF4E, eIF4G and eIF4A)[6]. mTORC2 is activated by PI3K-mediated growth factors signaling, and it phosphorylates AGC kinases such as protein kinase B (AKT), protein kinase C-alpha (PKC-α), serum and glucocorticoid-induced protein kinase1 (SGK1) to promote cell growth and cell proliferation [7](Fig. 1).

Upstream and downstream regulation of mTOR-related proteins. The core components of the regulatory network could regulate apoptosis, autophagy, protein synthesis, and proliferation, resulting in neuronal axon loss and demyelination of the central nervous system by complex mechanisms.

The associations between MS and the mTOR drawn from epidemiological studies have recently attracted much attention[3]. However, the role of mTOR in the progression of MS is still controversial. For instance, Mohammadarian Akbari et al.[8] found that patients with MS had a lower expression of mTOR-associated lncRNA, such as SNHG5 and DANCR. SNHG5 has been shown to modulate a variety of immune-related microRNAs, such as miR-205-5p, miR-154-5p, and miR-26a-5p[9], while DANCR can induce the expression of IL-6 and TNF-α in monocytes, which involved in the autoimmune response of MS. In contrast, Fahimeh Akbarian et al.[10] reported a higher expression of mTOR-related signals (mTOR, RPS6KB1, and EIF4EBP1) in relapsing-remitting MS.

It is still uncertain whether there is a causal relationship between MS and mTOR. However, single nucleotide polymorphism (SNP) is a powerful tool for predicting the relationships between two variables. Alleles are both independently segregated and randomly assigned at meiosis, and are independent from environmental or lifestyle factors, thus producing less potential confounding and biases. Based on the widely researched genome-wide association studies (GWAS), SNPs are chosen to stimulate the effects of exposures[11]. In this study, we use GWAS summary statistics to perform mendelian randomization (MR) analysis to estimate the impact of mTOR signals on MS risk.

2.1 Study Design

The current study was a two-sample MR analysis based on genome-wide association studies (GWAS) data, and SNPs were chosen to stimulate the effects of exposures[11]. Several assumptions for accuracy were made here. First, the genetic variants selected as instrumental variables (IVs) should be significantly associated with mTOR-dependent circulating proteins. Second, IVs should not be associated with any confounding factors. Third, IVs should only affect the risk of MS via mTOR-dependent proteins[12] (Fig. 2).

Directed acyclic graph (DAG) representing casual associations between mTOR-related proteins and MS risk. A node represents a variable, and an arrow represents a direct causal effect. IVs: instrumental variables; MS: multiple sclerosis.

2.2 Data Sources

Effects for SNPs associated with MS were derived from GWAS meta-analysis of the International Multiple Sclerosis Genetics Consortium (47429 patients and 68374 controls of European descent), revealing 200 non-major histocompatibility complex (MHC) loci associated with MS[13]. Genetic datasets of genetic predictors of mTOR-dependent circulating proteins (AKT, RP-S6K, eIF4E-BP, eIF4A, eIF4E, eIF4G, and PKC-α) were found in the INTERVAL study, which contains genetic associations with 2994 plasma proteins from 3301 healthy individuals of European descent (2,481 and 820 in the two sub-cohorts)[14].

2.3 Selection of instrumental variables

An initial significance threshold was set using independent variants in linkage disequilibrium (r² < 0.001) at the standard (p < 1e^− 5). Considering multiple comparison correction, a Bonferroni correction was applied and the low p < 1e^− 5 was set. SNPs with effect allele frequencies less than 0.01 (MAF < 0.01) were removed. Moreover, the F-statistic for each SNP and overall was examined to measure the strength of instrumental variables, where F is equal to R²(N-2)/(1-R²), R² refers to the variance explained by selected SNPs, and N refers to the number of sample size[15]. SNPs with F-statistic < 10 were excluded from further analyses to remove the weak-instrument bias. We used the PhenoScanner V2 (http://www.phenoscanner.medschl.cam.ac.uk/phenoscanner) to assess whether the instruments were associated with other phenotypes. The circulating level of vitamin D[16], sunlight exposure, obesity[17], age, sex, drinking, and smoking [18] were considered the confounding factors of MS, and any relevant instrument was excluded.

2.4 Statistical methods

All analyses were implemented in R (version 4.0.5) using TwoSampleMR packages (version 0.5.6) and MRPRESSO packages (version 1.0). The effect allele of each genetic variant was aligned and matched so that each sequence read could only match with e allele. Summary statistics’ beta coefficients (standard errors) were converted into Odds ratios with 95% confidence intervals(95%CI) for further analysis. Causal estimates were presented using three models: inverse variance weighted (IVW), weighted median estimator (WME), and MR-Egger regression methods (MR-ER). Sensitivity analyses were then performed. The Cochrane’s Q test was applied to quantify the degree of heterogeneity, and a p-value < 0.05 was considered statistically significant. A funnel plot was inspected to provide a visual inspection, whose asymmetry implied potential publication bias. The associations between mTOR-dependent circulating proteins and MS were depicted using scatter plots. The pleiotropic effect was detected by testing whether the y-intercept from MR-Egger analysis is zero, where p < 0.05 indicated horizontal pleiotropy[19]. The leave-one-out sensitivity analysis was performed to determine whether any instrument disproportionately affected the pooled results[20]. The MR Pleiotropy Residual Sum and Outlier (MR-PRESSO) could detect and correct the outliers in IVW linear regression[21, 22].

All SNPs strongly and independently predicted exposures and outcomes were extracted and initially quality-controlled. After harmonization, the number of SNPs that remained for the MR analyses ranged from 8 to 12 (Supplement Table1-7 and Fig. 3). The MR estimates using conventional MR analysis were presented (Fig. 4), and the results of sensitivity analysis (Fig. 5, 6, and Table 1).

Scatter plots for mTOR-related proteins (AKT, RP-S6K, eIF4E-BP, eIF4A, eIF4E, eIF4G, and PKC-α) on MS. The pleiotropic effect was detected by testing whether the y-intercept from MR-Egger analysis was zero.

Table 1

Heterogeneity and pleiotropy tests for multiple sclerosis associations with genetic predictors of mTOR-related proteins.
Exposure	Q-value (IVW)	P _Q (IVW)	Q-value (MR-ER)	P _Q (MR-ER)	Intercept	P _Intercept
AKT	9.256	0.235	9.256	0.160	5.36e^− 6	0.999
RP-S6K	8.995	0.253	8.030	0.236	-0.034	0.871
eIF4E-BP	17.185	0.046	12.492	0.131	0.079	0.121
eIF4A	2.633	0.917	2.472	0.872	0.005	0.702
eIF4E	6.635	0.675	6.606	0.580	0.002	0.884
eIF4G	13.751	0.131	7.761	0.457	-0.067	0.040
PKC-α	10.657	0.472	10.268	0.417	-0.012	0.552

Q-value: the statistics of Cochrane's Q test; P _Q: p-value corresponding to Cochrane's Q test; P_intercept: p-value corresponding to MR-Egger intercept test.

As presented, PKC-α circulating level was associated with a lower risk of MS with an odds ratio (OR) of 0.90, 95%CI (0.82, 0.98, P = 0.017) in IVW. However, the estimates from WME and MR-ER showed no significance, with an odds ratio (OR) of 0.95, 95%CI (0.84, 1.07, P = 0.385), and odds ratio (OR) of 0.97, 95%CI (0.74, 1.28, P = 0.840), respectively. Moreover, heterogeneity in sensitivities was not observed (Q = 13.75, P _Q = 0.131). No horizontal pleiotropy was detected (egger regression intercept = -0.0117, P = 0.552). The results from the leave-one-out analysis revealed that no single SNP strongly or reversely influenced the overall effect of exposure on the outcome (P༞0.05, Supplementary figure).

For RP-S6K, MR estimates were recalculated after removing two SNPs (rs482759 and rs62143197) that were identified as outliers using MR-PRESSO. The IVW method showed that RP-S6K circulating level was causally associated with the MS with an odds ratio (OR) of 1.12, 95%CI (1.00, 1.25, P = 0.045). No heterogeneity (Q = 9.00, P _Q = 0.253) or horizontal pleiotropy (egger regression intercept = -0.0039, P = 0.871) was detected.

For eIF4G, Egger regression indicated significant pleiotropy (egger regression intercept = -0.0667, P = 0.040). eIF4G circulating level was associated with a higher risk of MS with an odds ratio (OR) of 1.67, 95%CI (1.19, 2.36, P = 0.018) using MR-ER, and no heterogeneity (Q = 7.076, P _Q = 0.457). However, no significant causation existed using IVW or WME.

The MR analyses however showed no evidence of significant causal association between other four targets and MS, including AKT (IVW: OR = 1.06, 95%CI = 0.96–1.18, P = 1.177), eIF4E-BP (IVW: OR = 0.04, 95%CI = 0.90–1.20, P = 0.593), eIF4A (IVW: OR = 1.02, 95%CI = 0.95–1.10, P = 0.591), and eIF4E (IVW: OR = 0.99, 95%CI = 0.92–1.07, P = 0.786).

In this MR study, summary-level data for the associations of mTOR-related SNPs with the outcomes were obtained from large-scale GWAS or genetic consortia. Specifically, PKC-α and RP-S6K were investigated as relevant factors for MS risk, while AKT, eIF4E-BP, eIF4A, eIF4E, and eIF4G were of no significance.

The present results might offer strong evidence to strengthen the inference between MS and mTOR. Compared with the conventional observational epidemiological studies, the influence of confounders and reverse causality were minimized, and less potential confounding and biases were produced. Meanwhile, we used summary statistics from GWAS studies which have been used in several MR studies[23, 24]. The large sample size and high citations of two selected samples result in higher credibility.

The role of mTOR in the progression of MS is still controversial. Accordingly, we found that mTOR bidirectionally regulates the risk of MS. Some researchers revealed that inhibition of mTOR would curtail protein synthesis, and rapamycin targeting of mTOR could elicit responses resembling starved cells and block cells from processing beyond the G1 phase[25, 26]. Thus, mTOR is associated with the regulation of axonogenesis and cellular morphogenesis during neuronal differentiation[27], which can resist the procession of MS. Among these mTOR-related proteins, PKC-α, a downstream signaling molecule of TREM2 and PLCγ2, is a growth-associated protein after nerve damage, which can modulate cytoskeletal reorganization and affect the function of the microglia and macrophages[28]. According to these features, PKC modulators are used as a novel therapy for neuroinflammatory and neurodegenerative diseases[29]. This is consistent with our findings. In our study, PKC-α was demonstrated to be a protective factor for MS.

PI3K/AKT/mTOR pathway is associated with autophagy and apoptosis, and autophagy has significant effects on neurons and astrocytes[30]. RP-S6K, the upstream regulator of mTORC1, has been shown to affect the CNS by mediating autophagy and apoptosis[3], during which reactive oxygen (ROS) involves in oxidative stress to lead to neuronal axon loss and demyelination during MS[31]. In this study, we consider RP-S6K as a risk factor for MS, which is consistent with other research. For instance, rapamycin can inhibit multiple sclerosis by increasing Teff and decreasing Tregs[30].

Our findings demonstrate for the first time a causal relationship between mTOR and MS. However, several limitations existed. First, although sensitivity analyses were conducted, the possibility of residual pleiotropy could not be excluded completely. Pleiotropy was detected in the analysis between eIF4G and MS, and the outliers could not be identified by MR-PRESSO, which might be due to the potential confounding factors[22]. Thus, we believe the causation between eIF4G and MS was not meaningful. Second, data sources limit the number of mTOR-related proteins, and only seven proteins in the mTOR family were enrolled in the analysis. Third, since a few SNPs were selected using a threshold (p < 5e^− 8) according to Bonferroni-adjustment, we relaxed it to p < 1e^− 5, which could avoid the high false-negative rate[32], as portrayed in other MR study[33]. Besides, the GWAS data are from European ancestry, and further study for people of non-European ancestry was required.

Although it is apparent that mTOR-related proteins are not effective prevention strategies for MS, they may be used as an objective basis for individuals with strong MS family history to assess the potential risk of disease[3, 8, 10].

In conclusion, the causal relationship between mTOR related targets (PKC-α and RP-S6K) and MS risk was demonstrated. To some extent, mTOR has diagnostic value in differentiating MS patients. However, more detailed studies regarding the mechanism of MS were required.

Funding

This work was supported by the National Natural Science Foundation of China (grant number No. 82001740) and the National Social Science Fund of China (grant number 21BTQ050).

Conflict of interest

The authors have no potential conflicts of interest to declare.

Ethics approval

We followed the STROBE-MR guidelines to perform the present study. Ethics approval and informed consent were not required for the current study, as they were obtained in the original studies. The original studies were conducted in compliance with the Declaration of Helsinki.

Author contributions

Yao-Chen Zhang and Ke-Yi Fan: Conceptualization (lead); writing-original draft (lead); formal analysis (lead); writing-review and editing (equal). Qi Wang: Software (lead); writing-review and editing (equal). Jing-Xi Hu: Methodology (lead); Qian Wang: Visualization; He-Yi Zhang, Rong Zhao, Shan Song and Jun Qiao: writing–review and editing (equal). Sheng-Xiao Zhang: Conceptualization (supporting); Writing – original draft (supporting); Writing – review and editing (equal).

Data Availability Statement

These data were derived from the following resources available in the public websites: the exposure data for mTOR related proteins (https://www.phpc.cam.ac.uk/ceu/proteins/), and the outcome data for MS (https://gwas.mrcieu.ac.uk/datasets/)

Acknowledgements

We are grateful to the IMSGC (International Multiple Sclerosis Genetics Consortium) for providing the GWAS summary statistics. The Figure1 was partly generated using Servier Medical Art, provided by Servier, licensed under a Creative Commons Attribution 3.0 unported license.

Rodriguez Murua S, Farez MF, Quintana FJ (2022) The Immune Response in Multiple Sclerosis. Annu Rev Pathol 17:121–139
Howard J, Trevick S, Younger DS (2016) Epidemiology of Multiple Sclerosis. Neurol Clin 34(4):919–939
Maiese K (2021) Novel Insights for Multiple Sclerosis and Demyelinating Disorders with Apoptosis, Autophagy, FoxO, and mTOR. Curr Neurovasc Res 18(2):169–171
Kim E et al (2008) Regulation of TORC1 by Rag GTPases in nutrient response. Nat Cell Biol 10(8):935–945
Csibi A et al (2014) The mTORC1/S6K1 pathway regulates glutamine metabolism through the eIF4B-dependent control of c-Myc translation. Curr Biol 24(19):2274–2280
Petroulakis E et al (2009) p53-dependent translational control of senescence and transformation via 4E-BPs. Cancer Cell, 16(5): p. 439 – 46
Laplante M, Sabatini DM (2012) mTOR signaling in growth control and disease. Cell 149(2):274–293
Akbari M et al (2022) Expression analysis of mTOR-associated lncRNAs in multiple sclerosis. Metab Brain Dis 37(6):2061–2066
Han W et al (2020) Latest Advances of Long Non-Coding RNA SNHG5 in Human Cancers. Onco Targets Ther 13:6393–6403
Akbarian F et al (2020) Upregulation of MTOR, RPS6KB1, and EIF4EBP1 in the whole blood samples of Iranian patients with multiple sclerosis compared to healthy controls. Metab Brain Dis 35(8):1309–1316
Visscher PM et al (2012) Five years of GWAS discovery. Am J Hum Genet 90(1):7–24
Skrivankova VW et al (2021) Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization: The STROBE-MR Statement. JAMA 326(16):1614–1621
International Multiple Sclerosis Genetics (2019) C., Multiple sclerosis genomic map implicates peripheral immune cells and microglia in susceptibility.Science, 365(6460)
Sun BB et al (2018) Genomic atlas of the human plasma proteome. Nature 558(7708):73–79
Burgess S, Thompson SG, Collaboration CCG (2011) Avoiding bias from weak instruments in Mendelian randomization studies. Int J Epidemiol 40(3):755–764
Mokry LE et al (2015) Vitamin D and Risk of Multiple Sclerosis: A Mendelian Randomization Study. PLoS Med 12(8):e1001866
Ascherio A, Munger KL (2016) Weighing Evidence from Mendelian Randomization-Early-Life Obesity as a Causal Factor in Multiple Sclerosis? PLoS Med 13(6):e1002054
Jiang X, Olsson T, Alfredsson L (2018) Age at Menarche and Risk of Multiple Sclerosis: Current Progress From Epidemiological Investigations. Front Immunol 9:2600
Egger M et al (1997) Bias in meta-analysis detected by a simple, graphical test. BMJ 315(7109):629–634
Burgess S, Thompson SG (2017) Interpreting findings from Mendelian randomization using the MR-Egger method. Eur J Epidemiol 32(5):377–389
Wu F et al (2020) Mendelian randomization study of inflammatory bowel disease and bone mineral density. BMC Med 18(1):312
Verbanck M et al (2018) Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet 50(5):693–698
Yang Y et al (2021) Investigating the shared genetic architecture between multiple sclerosis and inflammatory bowel diseases. Nat Commun 12(1):5641
Mokry LE et al (2016) Obesity and Multiple Sclerosis: A Mendelian Randomization Study. PLoS Med 13(6):e1002053
Barbet NC et al (1996) TOR controls translation initiation and early G1 progression in yeast. Mol Biol Cell 7(1):25–42
Noda T, Ohsumi Y (1998) Tor, a phosphatidylinositol kinase homologue, controls autophagy in yeast. J Biol Chem 273(7):3963–3966
Kumar S et al (2021) Genome-Wide Scanning of Potential Hotspots for Adenosine Methylation: A Potential Path to Neuronal Development.Life (Basel), 11(11)
Wu DY et al (2003) PKC isozymes in the enhanced regrowth of retinal neurites after optic nerve injury. Invest Ophthalmol Vis Sci 44(6):2783–2790
Kim PM, Kornberg MD (2022) Targeting PKC in microglia to promote remyelination and repair in the CNS. Curr Opin Pharmacol 62:103–108
Vakrakou AG et al (2022) The mTOR Signaling Pathway in Multiple Sclerosis; from Animal Models to Human Data.Int J Mol Sci, 23(15)
Ohl K, Tenbrock K, Kipp M (2016) Oxidative stress in multiple sclerosis: Central and peripheral mode of action. Exp Neurol 277:58–67
Menyhart O, Weltz B, Gyorffy B (2021) MultipleTesting.com: A tool for life science researchers for multiple hypothesis testing correction. PLoS ONE 16(6):e0245824
Chande AT et al (2020) Ancestry effects on type 2 diabetes genetic risk inference in Hispanic/Latino populations. BMC Med Genet 21(Suppl 2):132

Download PDF

Editor assigned by journal
16 Jan, 2023
First submitted to journal
12 Jan, 2023

You are reading this latest preprint version

Genetically determined levels of mTOR-dependent circulating proteins and risk of multiple sclerosis

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

1. Introduction

2. Methods

2.1 Study Design

2.2 Data Sources

2.3 Selection of instrumental variables

2.4 Statistical methods

3. Results

4. Discussion

5. Conclusion

Declarations

References

Supplementary Files

Status:

Version 1