Structural Variants of Midnolin, a Genetic Risk Factor for Parkinson’s Disease, in a Yamagata Cohort

Abstract

Parkinson’s disease (PD) is a common neurodegenerative disease. We previously identified Midnolin (MIDN) to be a genetic risk factor for PD in both Yamagata (Japan) and British populations. However, the scale of our previous study was not sufficient to identify MIDN structural variants in the ascertained control of Yamagata Prefecture. We, therefore, reanalyzed MIDN variants in 3021 individuals from Yamagata Prefecture to compare with that in our previous British cohort study. MIDN copy number loss was only found in two cases (0.0662%), which was a lower frequency than that (1.64%) of the previously studied British cohort. Between the Yamagata and British groups, there was significant difference for rs3746106, located in the 5′-UTR of MIDN mRNA (p = 0.0003344, odds ratio 1.143), and for rs3746107, which corresponds to Ala34 (p < 2.2 × 10⁻¹⁶, odds ratio 5.89401). This study indicates that MIDN loss is relatively rare in the general Japanese population. Considering our previous studies that the frequency of MIDN loss is high among patients with PD (10.5 and 6.55% in Yamagata and Britain, respectively), the MIDN variants are much higher genetic risk factors for PD in a Japanese population than in a British population.

INTRODUCTION

Parkinson’s disease (PD) is the second most common neurodegenerative disease. More than twenty causative genes have been identified in patients with familial PD, including SNCA, Parkin, PINK1 and LRRK2.¹⁾ Additionally, genetic risk variants have been reported in genes such as GBA and INPP5F.^1,2) However, the majority of PD cases are sporadic and the mechanism of disease onset remains unclear.

Midnolin (MIDN) was originally discovered in mouse embryonic stem cells in 2000 by Tsukahara³⁾ and was named after its abundant expression in the developing midbrain and the intracellular localization of MIDN in the nucleus and nucleolus. The pathophysiological roles of MIDN are unclear. The N-terminal ubiquitin-like domain of MIDN binds to glucokinase and MIDN plays an inhibitory role in insulin secretion from MIN6 cells.⁴⁾ We have recently demonstrated that insulin promotes MIDN gene expression in human SH-SY5Y neuroblastoma cells,⁵⁾ indicating involvement in glucose metabolism. A MIDN variant is one of the candidates of genetic risk factors for autism spectrum disorder among Caucasian girls.⁶⁾ Additionally, duplications in 19p13.3, which includes the MIDN locus, are associated with male infertility⁷⁾ and MIDN is involved in progression of hepatocellular carcinomas.⁸⁾

Previously, we identified MIDN as a genetic risk factor for PD by array-based genome-wide association studies of cohorts from Yamagata Prefecture in north-east Japan and Great Britain.^9,10) In the Yamagata Prefecture cohort, MIDN copy number loss was observed in 10.5% of sporadic PD patients, whereas none was observed in 100 controls. In a large British cohort study, 142 of 2168 patients with PD (6.55%) and 47 of 2860 controls (1.64%) possessed structural variants spanning the MIDN locus. Furthermore, we have demonstrated that neurite outgrowth induced by nerve growth factor was completely inhibited and expression of Parkin E3 ubiquitin ligase was largely suppressed following knockout of MIDN in PC12 cells,⁹⁾ indicating strong association between MIDN loss and PD. However, Billingsley et al. showed that no PD-associated deletions within the MIDN locus were identified by whole genome sequencing and no common deletions were detected in general populations.¹¹⁾ We assume that discrepancies between our conclusions and those of Billingsley et al. reflect the different methodologies used in the studies.¹²⁾ We identified no MIDN loss in healthy controls in the Yamagata cohort study, possibly because of Japanese genetic characteristics or because our previous Yamagata cohort study was not large enough to identify MIDN structural variants unlike in the large British population. Therefore, we reanalyzed MIDN variants in 3021 subjects of the Yamagata Prefecture general population in more detail.

MATERIALS AND METHODS

This cohort study included 1621 subjects (Male 727, Female 894) from the general population from the town of Takahata in Yamagata Prefecture, and 1400 general population subjects (Male 694, Female 706) from Yamagata City in Yamagata Prefecture. The mean ages of participants from Takahata and Yamagata City were 61.28 (age range 40 to 87) and 60.71 (age range 39 to 77) years, respectively, at the time they were recruited in 2004–2006 (Takahata) and 2010–2012 (Yamagata). DNA was purified from peripheral blood samples and genotyped using Illumina Human660W-Quad (660 K) (for the Takahata samples) and Human OmniExpressExome (1.0 M) (Illumina, San Diego, CA, U.S.A.) (for the Yamagata samples). Copy number variation (CNV) was calculated using the R-based GenoCN program and the computational resources of the Yamagata University School of Medicine YuM-HPC high performance computing cluster. Both the B-allele frequency data and the signal intensity data (LogR Ratio) were used after normalization of intra- and inter-chip data, and CNVs were identified with at least three markers. Single nucleotide polymorphisms (SNPs) were genotyped; however, because the probe for rs3746107 is not on the Human OmniExpressExome array, the theoretical value for rs3746107 in 1400 subjects from Yamagata City was calculated by imputation using 1000 genome reference data (https://www.internationalgenome.org/category/imputation/). The statistical significance of differences among CNVs and SNPs were analyzed by Fisher’s exact test using R4.1.3 software. The study was performed in accordance with a protocol approved by the Ethics Committee of Yamagata University (Approval No. 2021-338).

RESULTS

Under our criteria for CNV detection, we identified two individuals (0.123% in the Takahata group and 0.0662% in all subjects) with CN loss (CN = 1). All other subjects had CN = 2 (normal) and no participant had CN = 0 (loss) or CN = 3 or 4 (multiplication) unlike in the British cohort.¹⁰⁾ The genomic locations of deletions encompassing MIDN are shown (Fig. 1). Individuals #1 (ID#1, Female) and #2 (ID#2, Male) were 51 and 63 years old in 2004, respectively, and their deletions were approximately 68.9 and 63.4 kb, respectively, both spanning the entire MIDN gene. The mean length of MIDN deletions in 47 individuals of the British control group was 30.3 kb, with a median of 32.9 kb.¹⁰⁾ Although only limited clinical information of their current conditions is available, individual #1 was healthy and individual #2 had brain infarction, corneal transplantation, apnea syndrome, hypertension and pyorrhoea alveolaris at the time of last investigation in 2011. The percentage of CN loss was significantly smaller compared with the British cohort (1.64%), (p = 7.848 × 10⁻¹³, odds ratio 0.03966, CI₉₅ = 0.004681–0.1517). Together with our previous result of no individual among 100 Yamagata controls having CN loss,⁹⁾ the results of the current study suggest MIDN structural variants are relatively rare in the general Japanese population.

Fig. 1. Genomic Location of MIDN Deletions Found in Two Individuals

The green and blue bars represent genes and the deletions, respectively. We identified two individuals [ID#1 (Female, 51 years old) and#2 (Male, 63 years old)] (0.123% in the Takahata group and 0.0662% in all subjects) with CN loss (CN = 1). All other subjects had CN = 2 (normal) and no participant had CN = 0 (loss) or CN = 3 or 4 (multiplication).

We also analyzed two relatively frequent SNPs (rs3746106 and rs3746107) in the MIDN gene (Table 1). Probes for these two SNP are contained in the microarray chips used for both this and the previous studies of British populations, and these SNP frequencies can be compared among Yamagata and British populations. Between the Yamagata (Takahata and Yamagata) and British groups, there was significant difference for rs3746106, located in the 5′-UTR of MIDN mRNA (p = 0.0003344, odds ratio 1.143, CI₉₅ = 1.062–1.230), and for rs3746107, which corresponds to Ala34 (p < 2.2 × 10⁻¹⁶, odds ratio 5.89401, CI₉₅ = 4.802–7.288)¹⁰⁾ (Table 1). The minor allele frequencies (MAFs) for rs3746106 and rs3746107 were 0.4588 and 0.1079, respectively, in the Yamagata Prefecture cohort, which were comparable with those in East Asian subjects (0.5799 and 0.1093 for rs3746106 and rs3746107, respectively) (gnomAD browser, gnomAD v2.1.1 (non-cancer), https://gnomad.broadinstitute.org). These results indicate racial differences in the MAF, especially for rs3746107, between Japanese and British populations. rs3746106 and rs3746107 correspond to a substitution in the 5′-UTR and a synonymous substitution (Ala34Ala), respectively; therefore, the stability of the MIDN mRNA or translation efficiency may be affected by these SNPs. The functional significance of these SNPs, however, remains unclear.

Table 1. Two SNPs in the MIDN Gene in the General Population of Yamagata Prefecture, Japan

rs3746106 (19 : 1250109 C > A, 5′-UTR)
CC	CA	AA	N/A	MAF
894 (29.59%)	1482 (49.06%)	645 (21.35%)	0	0.4588
rs3746107 (19 : 1250397, C > T, 34Ala > 34Ala)
CC	CT	TT	N/A	MAF
2420 (80.11%)	550 (18.21%)	51 (1.688%)	0	0.1079

rs3746106 in the 5′-UTR region of MIDN and rs3746107, corresponding to Ala34, were genotyped. There were significant differences in the frequencies of these SNPs between Japanese and British general populations (p = 0.0003344 for rs3746106, and p < 2.2 × 10⁻¹⁶ for rs3746107).

DISCUSSION

We have demonstrated MIDN loss in 10.5% of Japanese patients with PD and in 6.55% of British PD patients,^9,10) suggesting that MIDN is a universal genetic risk factor for PD. Although various MIDN deletions were identified in the general British population (1.64%) and are archived in the Database of Genomic Variants (http://dgv.tcag.ca/dgv/app/home or UCSC Genome Browser; https://genome.ucsc.edu), MIDN deletions were only detected in two out of 3021 (0.0662%) individuals from the Takahata and Yamagata regions and in none of 100 individuals (0.00%) studied previously.⁹⁾ This frequency is very small compared with the British cohort, and MIDN structural variants are much higher genetic risk factors for PD in a Japanese population than in a British population (i.e., the odds ratio is 182 and 4.35 in Japanese and British populations, respectively). The reason(s) for this discrepancy is currently unclear. However, a possible technical reason is that the microarray platforms used in this and previous studies of British populations are not completely identical, with the probe number and position being different in each study, which can result in the detection of different numbers of CNVs. An alternative possibility is a diversity of genetic characteristics between Yamagata and British populations. A female preponderance of PD is observed in Japan unlike in Europe and the US¹³⁾; therefore, the stability of the MIDN locus may vary among these populations. However, the important point is that MIDN loss is a significant risk factor in both populations.

The MIDN deletion found in this and previous studies spans multiple other genes adjacent to MIDN (Fig. 1). Therefore, these genes may also be involved in PD pathogenesis. Although we have shown that suppression of MIDN expression caused PD-like phenotypes in PC12 cells,⁹⁾ the possibility cannot be excluded that deletion of other genes in addition to MIDN promotes PD progression. There was no CN = 0 individual in this and our previous studies, suggesting that entire loss of the MIDN gene may be lethal. Further studies examining family history, clinical conditions, MIDN functions in neuronal cells and pathophysiological roles of MIDN using knockout mice are necessary.

Acknowledgments

We thank the participants of the Yamagata Prefecture Cohort study and the staff of Yamagata University who collected and genotyped the DNAs. We also thank Jeremy Allen, Ph.D. for editing a draft of this manuscript. This work was supported in part by Grants-in-Aid from the Japan Society for the Promotion of Science (KAKENHI 21K06573 to Y.O.), Takeda Science Foundation (Y.O.) and 21st Century Centers of Excellence (COE) and the Global COE program of the Japan Society for the Promotion of Science (Yamagata University).

Conflict of Interest

Yutaro Obara received a research grant from Takeda Science Foundation. Hidenori Sato and Kuniaki Ishii have not conflict of interest.

Data Availability

The individual DNA data cannot be made available because of ethical restrictions on sharing data publicly.

REFERENCES

• 1) Lill CM. Genetics of Parkinson’s disease. Mol. Cell. Probes, 30, 386–396 (2016).
• 2) Mitsui J, Mizuta I, Toyoda A, Ashida R, Takahashi Y, Goto J, Fukuda Y, Date H, Iwata A, Yamamoto M, Hattori N, Murata M, Toda T, Tsuji S. Mutations for Gaucher disease confer high susceptibility to Parkinson disease. Arch. Neurol., 66, 571–576 (2009).
• 3) Tsukahara M, Suemori H, Noguchi S, Ji ZS, Tsunoo H. Novel nucleolar protein, midnolin, is expressed in the mesencephalon during mouse development. Gene, 254, 45–55 (2000).
• 4) Hofmeister-Brix A, Kollmann K, Langer S, Schultz J, Lenzen S, Baltrusch S. Identification of the ubiquitin-like domain of midnolin as a new glucokinase interaction partner. J. Biol. Chem., 288, 35824–35839 (2013).
• 5) Sagehashi N, Obara Y, Maruyama O, Nakagawa T, Hosoi T, Ishii K. Insulin enhances gene expression of midnolin, a novel genetic risk factor for Parkinson’s disease, via extracellular signal-regulated kinase, phosphoinositide 3-kinase and multiple transcription factors in SH-SY5Y cells. J. Pharmacol. Exp. Ther., 381, 68–78 (2022).
• 6) Butler MG, Rafi SK, Hossain W, Stephan DA, Manzardo AM. Whole exome sequencing in females with autism implicates novel and candidate genes. Int. J. Mol. Sci., 16, 1312–1335 (2015).
• 7) Singh V, Bala R, Chakraborty A, Rajender S, Trivedi S, Singh K. Duplications in 19p13.3 are associated with male infertility. J. Assist. Reprod. Genet., 36, 2171–2179 (2019).
• 8) Kweon SM, Kim G, Jeong Y, Huang W, Lee JS, Lai KKY. Midnolin regulates liver cancer cell growth in vitro and in vivo. Cancers, 14, 1421 (2022).
• 9) Obara Y, Imai T, Sato H, Takeda Y, Kato T, Ishii K. Midnolin is a novel regulator of parkin expression and is associated with Parkinson’s Disease. Sci. Rep., 7, 5885 (2017).
• 10) Obara Y, Sato H, Nakayama T, Kato T, Ishii K. Midnolin is a confirmed genetic risk factor for Parkinson’s disease. Ann. Clin. Transl. Neur., 6, 2205–2211 (2019).
• 11) Billingsley KJ, Bandres-Ciga S, Ding JH, Hernandez D, Gibbs JR, Blauwendraat C. MIDN locus structural variants and Parkinson’s Disease risk. Ann. Clin. Transl. Neurol., 7, 602–603 (2020).
• 12) Obara Y, Sato H, Nakayama T, Kato T, Ishii K. Reply to: MIDN locus structural variants and Parkinson’s disease risk. Ann. Clin. Transl. Neurol., 7, 604–605 (2020).
• 13) Kimura H, Kurimura M, Wada M, Kawanami T, Kurita K, Suzuki Y, Katagiri T, Daimon M, Kayama T, Kato T. Female preponderance of Parkinson’s disease in Japan. Neuroepidemiology, 21, 292–296 (2002).