Single Nucleotide Polymorphism Micro-array Analysis On Copy-number-variation In Whole Genome Of Human Glioma Cells

Glioma is a malignant tumor of the central nerves system （CNS）, which developsfrom neural ectoderm. As the most prevalent malignant tumor of the central nervessystem （occurrence3-10/100000）, And it accounts for about46%of the encephalictumors. So far, surgery or surgery plus radiotherapy and chemotherapy is the maintreatment for glioma. After surgery, the mean survival time span become only8-11months with the recurrence rate of nearly100%. Once recurred, the mean survivaltime span decreases to only3-5months.Genetic research on glioma has undergone courses from studying largechromosomal segments change to genetic change and from studying single genefunction to multiple gene function. Comparative genomics hybridization （CGH） is oneof the conventional technologies which can detect deletions and duplications at thechromosomal level. However, it has a relatively low resolution in the range of1to5Mb. Fluorescence in situ hybridization （FISH） was better with a resolution of0.5to1Mb, but it can only detect the effect of a single gene. In addition, the above methodsshow a high false positive rate and are experimentally laborious. Therefore, it’snecessary to find a new way to observe and detect the genome-wide changes inglioma, as well as the effects of multiple interacting genes.With the help of genome-wide technologies, many researches have performed onshort tandem repeats （STR）, single nucleotide polymorphism （SNP） and copy numbervariations （CNV） which broadly exist in human genes. STRs or microsatellite DNAscontain a repeat unit of2-6bp and the repeat number varies between10-60times,forming a segment of100-300bp. Using STR as markers, researchers can investigatethe loss of the loss of heterozygosity in glioma cells can be investigated. But theresolution is only1Mb and STRs are not very stable （with high mutation rate）. Single nucleotide polymorphism （SNP） is caused by mutation of single basewidely spread in the genome, and covers more than90%of the genomepolymorphism. SNPs are widely spread in the genome with a total number of3millions or more, with an average of1SNP every500-1000bp. SNPs are proved to beassociated with many diseases and thought as the third generation molecular markerfor detecting copy number variation thanks to its high resolution and stability.The role of copy number variation （CNV） in glioma has gained more and moreattention with the development of genomics. CNVs were first comprehensivelyreported by Iafrate and Sebat et al. in2004（Iafrate A J, Sebat J et al.,2004） andinclude insertion, deletion, duplication and complex structural changes that span1kb-100kb. In2006, Redon et al. observed1447CNV regions that cover about12%of the human genome, and generated the human CNV map in which many CNVsreside in disease related loci. Since then, many researchers reported the prevalenceand importance of CNV （Iafrate et al.2004, Sebat et al.2004, Tuzun et al.2005, conradet al.2006, McCarroll et al.2006, Lei et al.2006）. In general, CNVs involve incomplex diseases through two mechanisms: change of gene structure and dosageeffect. Furthermore, these CNVs may also alert long-region regulatory control. Thus,it is reasonable to use CNV as a breakthrough point in research on glioma.After the generation of first human CNV map in2006（Redon etal.,2006）, apaper in Nature reported a great improvement in high-throughput genomictechonology and pointed out that it was feasible to perform cancer genome project,and both CNV and DNA methylation had important effect in glioma （L. chin and M.Meyerson. et al.2008）. Subsequently, some researchers reported duplications onchromosomes1,2,3,4,7,8and12（Chan, J Y.,2011） and deletions on chromosomes1,3,5,7,9.10and17（Chi A S et al.,2012）. These authors adopted genome-wideSNP microarray technology which has a higher resolution and better accuracycompared with earlier methods and can simultaneously analyze the CNV of manyspecific loci in the genome. So far, there is no report on Chinese glioma genomicchanges, especially on the CNV in different stages of glioma in brain. In this study, we analyzed human glioma sample pairs from18Han Chinesepatients with Illumina human370microarray in order to understand the geneticbackground of Han Chinese glioma, particularly the association between CNV anddifferent stages of glioma. The Illumina370chip contains370000probes in total,with52000CNV specific proves and a minimum and average distance of5kb and7.7Kb, respectively. In this experiment, we detected4types of CNVs: loss ofhomozygosity, loss of heterozygosity （with or without copy number change）, andsegmental duplication. We use the software from illumine to get SNP genotypes andCNV bodundaries. The association of genes localized in the identified CNV reigionwith different stage of glioma was calculated and the data based on KEGG and GOdatabases are further analyzed to find specific changes and provide new insights intoglioma etiology.And this disertation will show that18Chinses glioma patiens can be divided into2subtypes: concordant subtype with concordant CNV between glioma and bloodDNA; and disconcordant subtype for which the glioma CNV and blood CNV aredifferent. We found that the differential stage of glioma was associated with the totalnumber of CNV and the types of genes in these CNV, and CNVs also occurred innoncoding regions. In highly differentiated glioma cells,10genes in MARK signalingpathway were all duplicated. In low differentiated glioma cells,4genes in calcium signaltransduction pathway were duplicated. In total, there are442CNV genes in highlydifferentiated glioma, mainly enriched in organelle organization, transcription regulation,translation regulation, cell metabolism etc. Moreover, there are112CNV genes in lowdifferentiated gilima, mainly enrich in signal transmission, neuronal synapsis, catalyticactivity, metabolism, DNA repair, reproduction regulation and so on.Results are as follows:Part1. Comparison between glioma genome and blood PBMC genome1.11of12brain glioma patients showed concordant CNVs between glioma genomeand blood PBMC genome, which was a loss of homozygosity on chromosome3qbetween the noncoding regions163-165Mb. Thirteen brain glioma patients showed same CNVs between glioma genome andblood PBMC genome, which was a loss of homozygosity on chromosome4qbetween69.8-69Mb. There are3functional genes in this region: UGT2B17（NCBI Reference Sequence: NM₀01077.3）, YTHDC1（YT521B, transcriptvariant1NCBI Reference Sequence: NM₀01031732.2）and TMPRSS11E（NCBIReference Sequence: NM₀14058.3）.These results suggested that:1、 The CNVs on chromosomes3q and4q are associated with the genesis of glioma.2、 The CNV on chromosome3q is noncoding.3、 The CNV on chromosome3q is in coding region.4、 The gene UGT2B17in4q region is a member of the UGT gene family. Thehomozygosity loss of UGT2B17is associated with susceptibility to prostatecancer and lung cancer.5、 The TMPRSS11E in4q region encodes a protein that belongs to type II serineprotease family, the homozygosity loss of which is associated with squamous cellcarcinoma of the head and neck.2. Of the18sample pairs, we also found some occasional CNVs. However,there is no evidence to show these changes correlated with giloma.The concordant CNVs between glioma genome and blood PBMC genomesuggested that there might be some CNVs with genetic susceptibility to gliomaand the genes in these CNVs may be candidate genes for glioma.Part2. Comparison of somatic CNVs between high and low differentiatedglioma genome.In addition to the concordant CNVs in both glioma and blood genomes, wefound some somatic CNVs that only exist in glioma but not in the blood geneme. Wecompared these somatic CNVs between high and low differentiated glioma genomeand found the following results:1、 There are hybrid loss of heterozygosity with copy number losses in both highand low differentiated glioma genomes. They spanned cross the genome on chromosomes1q,2q,5q,6q,13q and19q.2、 There are also loss of heterozygosity without copy number change both highand low differentiated glioma genomes, They spanned cross the genomeexcept the regions on chromosomes13p,14p,15p,19p,21p and22p.3、 Somatic duplications are also found in both high and low differentiatedglioma genomes. The regions with occasional somatic duplications include2p,2q,3p,3q,4p,8p,9q,11q,12p,12q,13q,16q,18p,18q,19p,19q,20pand20q.4、 Somatic amplification is also found in both high and low differentiatedglioma genomes.Of the9highly differentiated glioma patients,5showed CNV amplicaiton. Oneof the amplification was on chromosome1（187.85-187.8Mb）, another was onchromosome19（62.7-6335Mb）, including Zinc finger proteins, the last one was onchromosome20（25.6-26.2Mb）, including4functional genes: MIR663，FAM182A，C20orf191and FAM182B.The above results suggested that:1. The CNVs on chromosomes1q,19q and20may be associated with thesomatic changes in glioma.2. The CNV on chromosome1q is noncoding.3. The CNVs on chromosomes19q and20p are in coding region.4. The Zinc finger protein gene in19may encode a protein that functions as atranscript factor to regulate the expression of its target gene.5. The MIR663code in20p encodes a micro-RNA, this kind of RNA isassociated with gastric cancer.6. By far, there is no report on the function of proteins encoded by FAM182A，C20orf191and FAM182B. Part3. The pathway of cell singnaling involved in the genes of CNV relevantto brain gliomaIn total, there are442CNV genes in highly differentiated glioma and112CNVgenes in low differentiated glioma.1.442CNV genes in highly differentiated glioma were enriched in15signalingpathways, for example, mitogen-activated pro-teinkinase （MAPK） pathway, in which10genes were significantly duplicated, suggesting association with glioma.2.112CNV genes in low differentiated glioma were enriched in8signalingpathways, for example, the calcium ion signaling pathway, in which4genes weresignificantly duplicated, suggesting association with glioma cancerization.Part4.Glioma CNV genes that involved in cell biologyGO analysis of442and112CNV genes in high and low differentiated gliomasrevealed that:1.442CNV genes in highly differentiated glioma were summarized into18gene clusters. Encoding proteins includes important physical andpathologic processes such as organelle organization, transcriptionregulation, translation regulation and cell metabolism.2.112CNV genes in low differentiated gilima, mainly enriched in29geneclusters including signal transmission, neuronal synapsis, catalytic activity,metabolism, DNA repair, reproduction regulation and so on.the Following conclusions drawn from results above:1. Loss of homozygosity, including UGT2B17, YTHDC （YT521B） andTMPRSS11E （DESC1） genes, were found in both glioma and blood gemome,suggesting susceptibility to glioma.2. The zinc finger protein gene, MIR663，FAM182A，C20orf191and FAM182Bthat were significantly amplicated in highly differentiated glioma may be associatedwith glioma genesis while the function of amplication in noncoding reging on1q needfurther investigation. 3. The genomes of low differentiated glioma showed complex, violent,disordered and variable CNVs, which are suitable for genetic research.4. The functional analysis based on KEGG suggested that highly differentiatedglioma was associated with MAPK signaling pathway while low differentiated gliomamight be associated with calcium ion signaling pathway.In summary, we analyzed the CNV of18tumor-blood sample pairs usingIllumina human370microarray, and found that homozygous deletions in ChineseCNVs may be associated with susceptibility of glioma. We also found that highdifferential glioma non-genetic zinc lipoprotein family gene and the amplification ofMIR663copys may caused by the mutation under the pressure of externalenvironment. The study and exploration of the phenomenon of human genome CNVhelp identify susceptibility genes in glioma and human cell genome gene mutationoccurred in the role of external risk factors. Further exploration can be made in theimpact of the gene mutation as well as the mechanism of the synergistic effect of themulti-gene in multi-channel.