Genetic And Functional Study Of Neurofilament Coding Genes In Chinese Sporadic Amyotrophic Lateral Sclerosis

Objectives1.To construct variants profile of neurofilaments coding genes(NEFH/NEFM/NEFL)clarified the variants burden in Chinese sporadic amyotrophic lateral sclerosis(s ALS)and control populations.2.To analyze the relationship between variants and the clinical phenotype on the level of gene and mutation site,and to construct the survival prediction model by machine learning after the enrichment of rare gene mutations was used to find the importance of variants.3.The experiment of high-risk rare variant effect in cells indicated that high-risk variant leads to changes in protein function.Furtherly,we clarified the relationship between neurofilament protein-coding genes and s ALS.Methods1.Using PCR-Sanger DNA sequencing technology,all exons and the intron and exon junction area of the three NEFH/NEFM/NEFL genes of 371 s ALS and 711 normal healthy controls in multiple centers were sequenced.The detected variants are searched and compared according to the genetic database.A case-control analysis was performed on allele frequencies of common,low-frequency and rare variants.To construct protein mutation maps of three gene mutations based on the sequence.An independent Chinese healthy elderly population control cohort of 776 people was used for validation.2.We analyze the clinical characteristics of the detected rare variants carrying patients.Then,we synthesize the information of standard clinical dimensions with neurofilament gene mutations to construct a clinical practical survival prediction model through Logistic regression,Support vector machines,Decision trees,Random forest,and Random survival forest.3.The plasmid expressing recurring high-risk variant was constructed for exploring its effect in protein function and potential mechanism through CCK8 assay,LDH release assay,ATP level detection and immunofluorescence.Results1.In 371 cases with s ALS and 711 normal controls,A total of 92 variants were identified,including 36 rare heterozygous variants in NEFH,27 in NEFM,and 16 in NEFL.In 5 patients with s ALS(n=371),a missense mutation in exon 4 of NEFH gene(rs568759161)was found,which was detected in 5 s ALS,1 in the healthy control group(n=711)and 1 in the independent elderly healthy control group(n=776)(OR=OR:10.08,95% CI:1.95-52.07,p=0.005).2.Patients with the high-risk variant rs568759161 had clinical onset of limbs(80%),primary in males(80%),and their clinical characteristics were not statistically different from the non-carrying group.3.Using the standard clinical and genetic mutation information to predict survival,Random forest was more efficient,with an accuracy rate of 85.00% and AUC of 0.90 in long survival.While the Random survival forest was more efficient in short survival.However,NEFH/NEFM/NEFL were not crucial in prediction models.4.High-frequency NEFH variant(p.Ser787Arg)showed probable loss of function effect of NEFH protein with an unknown mechanism in vitro.Conclusions:1.In this study,we performed the screening of NEFH/NEFM/NEFL gene mutations in patients with sporadic amyotrophic lateral sclerosis.The variant rs568759161 belongs to high-risk sites in s ALS.The variant was verified to the probable loss of function type in a cellular assay.It was provided an essential basis for further elucidating the occurrence of ALS disease.2.The types and frequency in variants of NEFH/NEFM/NEFL are different in Chinese and Caucasian patients,and they are an essential supplement to s ALS genetics.This data implied the genetic heterogeneity of s ALS.3.This study proposes a simple Random forest and Random survival forest prediction model constructed by clinical information and genetic variants.It provided further evidence for the medical application of artificial intelligence.