| Background:Influenza A(H1N1)has continued to spread between humans and animals since its emergence in 1918,which led to a catastrophic global epidemic.Under long-term genetic selection pressure,the influenza A(H1N1)virus began to mutate and produce new religands.The mutated virus can evade host immunity and quickly adapt to environmental changes.This change not only increases population susceptibility,but also causes vaccine resistance,posing a continuing threat to human health and public health.Therefore,it is very important to strengthen the traditional surveillance of influenza,to develop the prediction model of influenza incidence and to analyze the mutation status of influenza virus gene sequence in real time.Objectives:Based on the influenza surveillance data of national sentinel hospitals in Jiangsu Province from 2017 to 2020,the epidemiological characteristics of influenza cases in Jiangsu Province from 2017 to 2020 were analyzed to understand the current epidemic trend of influenza.The meteorological data of 208 weeks in Jiangsu Province from 2017 to 2020 were collected and used as independent variables to construct 8 statistical models to predict the incidence of influenza A(H1N1)in Jiangsu province,providing scientific basis for precise prevention and control of influenza A(H1N1)in Jiangsu Province.The evolutionary characteristics of influenza A(H1N1)NA gene in Jiangsu Province from 2017 to 2020 were analyzed to understand the variation of NA gene fragments,which laid a foundation for the search of vaccine development sites.Monoclonal antibody against H1N1 NA was prepared by hybridoma technology,and its specificity and cross-binding property were preliminarily explored,and a sandwich ELISA system of double antibody was established,providing a new direction for detection and treatment of influenza A H1N1 virus.Methods:1.Statistical methods:The 2017-2020 influenza A(H1N1)surveillance data were obtained from the influenza surveillance network in Jiangsu Province.The epidemiological characteristics of influenza cases in Jiangsu Province were analyzed by using IBM SPSS Statistics 22.0 software and epidemiological method.Count data were described using the number of cases and percentage[N(%)].The difference of demographic information was tested by Pearson Chi-square test,with test standardα=0.05.Seven machine learning models(decision tree,support vector machine,nearest Neighbor node,Adaboost,gradient lifting decision tree,Bagging,random forest)and one generalized linear model were constructed by using Python 3.9software.70%of the weekly incidence data of influenza A(H1N1)in Jiangsu Province from 2017 to 2019 were taken as the training set,30%as the test set,and the weekly incidence data of influenza A(H1N1)in Jiangsu Province in 2020 were taken as the external validation set.The determination coefficient,mean square error,root mean square error and mean absolute error were selected as the evaluation indexes of the model.2.Genetic evolutionary:48 representative strains of influenza A(H1N1)from2017 to 2020 in Jiangsu Province were selected and their NA fragments were amplified and sequenced.MEGA 6.0 software was used to construct the evolutionary tree,analyze the main epidemic branches,and explore the NA sequence homology,enzyme activity center and surrounding related sites,drug resistance sites,and antigen-determining cluster sites.Potential glycosylation sites were also analyzed using Net NGlyc 1.0 Server(http://www.cbs.dtu.dk)software.3.Preparation and functional exploration of monoclonal antibody:The recombinant H1N1 NA protein was used as immunogen to immunize mice.After 4times of immunization,the spleen cells with the highest titer were fused with SP2/0cells,and the monoclonal antibody against NA was obtained by subclonal technique.The specificity and cross binding of antibodies were verified by indirect ELISA.The double-antibody sandwich ELISA system was established with the monoclonal antibody with high binding activity as the capturing antibody,and the optimal experimental conditions,specificity,and sensitivity of the system were further explored.Results:1.The overall distribution of influenza A(H1N1)in Jiangsu Province from 2017to 2020 shows an obvious winter epidemic pattern;There was no difference in the annual mean positive rate between males and females(χ~2=0.047,P=0.829).The mean annual positive rate of all age groups ranged from 2.5%to 3.7%,and the difference was statistically significant(χ~2=45.385,P<0.001);The annual average positive rate of A(H1N1)was the highest in the service industry and teachers(4.4%),followed by the workers(3.9%),and the lowest in the cadres(2.6%),the difference was statistically significant(χ~2=63.655,P<0.001);The mean annual positive rates in southern Jiangsu,central Jiangsu,and northern Jiangsu were 3.1%,2.9%and 3.3%,respectively,with significant differences(χ~2=14.052,P=0.001).2.The distribution of influenza A(H1N1)in Jiangsu Province in each year has different phenomenon.In terms of gender,the positive rate of A(H1N1)in males(5.8%)and females(5.9%)reached the highest level in 2018,and the lowest level in males(0.1%)and females(0.2%)in 2020.In terms of age,the highest positive rate was found in 5-14 years old in 2017(2.2%),25-44 years old in 2018(6.8%),25-44years old and 45-59 years old in 2019 and 2020.In terms of occupational population,the positive rate in most occupational groups(cadres,workers,retirees,students,preschool children,medical personnel and other groups)was higher in 2018 than that in other years.Regionally,the positive rate of influenza A(H1N1)was shown that central Jiangsu>southern Jiangsu>northern Jiangsu in 2017,northern Jiangsu>southern Jiangsu>central Jiangsu in 2018,northern Jiangsu>central Jiangsu in 2019,and southern Jiangsu>central Jiangsu and southern Jiangsu>northern Jiangsu in 2020.3.Among the eight prediction models,the decision tree model based on CART algorithm has higher model evaluation indexes than the other 7 models in Predicting the number of cases of influenza A(H1N1)in Jiangsu Province.4.Among the 8 prediction models about the number of cases of influenza A(H1N1)in Jiangsu Province,the four evaluation indexes of the models(determination coefficient,mean square error,root mean square error and mean absolute error)of the decision tree model based on CART algorithm were 0.74,774.89,27.84 and 6.31,respectively,which were higher than the other 7 models.5.The molecular characteristics of genes showed that there were many mutation sites in Jiangsu isolates from 2017 to 2020,with 9,32,25,and 9 mutations occurring in each year,respectively.The mutation sites involved several important regions,including E119K mutation at the enzyme activity site and the drug resistance site,K386N and E432K mutation at the antigenic cluster site,and large variation at sites50,386,and 450 of glycosylation sites.In addition to the key location,mutations were frequent at 77,81,188,and 449 in isolates from 2017-2018 and 2018-2019,and significant at 13,51,and 74 in isolates from 2018 to 2020.416 has been mutated repeatedly from 2017 to 2020.6.A monoclonal antibody 5D3 of H1N1 NA was successfully prepared,which could bind to both NA recombinant protein and whole H1N1 virus with certain specificity and no cross binding.7.Double antibody sandwich ELISA system was established with 5D3monoclonal antibody as the best capture antibody and mouse anti-NA PAb-HRP as detection antibody.The best antibody dilution of the system was 1:500,and the best blocking solution was 1%BSA plus 3%sucrose,which had certain specificity but sensitivityneeded to be improved.Conclusion:1.From 2017 to 2020,there was an obvious winter epidemic pattern of influenza A(H1N1)in Jiangsu Province.The age of onset increased from 5 to 14years old to 25 to 59 years old,and the service industry,teachers and workers were the most susceptible groups.In the future,Prevention and control of winter influenza should be carried out well,and surveillance should be strengthened among the 25-59age group and susceptible population.2.The decision tree model based on CART algorithm shows good performance in predicting the number of cases of influenza A(H1N1)in Jiangsu Province.On the basis of traditional influenza surveillance and machine learning methods,a new prevention and control mode combining normality and prediction will be more conducive to accurate observation of influenza change level and activity rules.3.From 2017 to 2020,there were many amino acid mutation sites encoded by NA of influenza A(H1N1)in Jiangsu Province,involving many important regions such as enzyme activity sites,drug resistance sites and glycosylation sites.There were persistent mutations in many noncritical sites,so etiological monitoring of mutation sites should be strengthened in the future.4.In the preparation of mouse monoclonal antibody,the recombinant protein of H1N1 NA was first designed and expressed,and immunized mice with it as an immunogen,which improved the probability of specific screening to a certain extent.Through a series of experiments,a stable NA monoclonal antibody 5D3 was obtained,which had certain specificity and no cross binding.Double antibody sandwich ELISA system was established with 5D3 monoclonal antibody as the best capture antibody and mouse anti-NA PAb-HRP as detection antibody.The system has certain specificity but sensitivity needs to be improved.This is of great significance for the preparation and application of NA-targeted monoclonal antibodies,and further exploration is needed in the future. |
PDF Full Text Request