LABORATORY SCIENCES Differential Expression Profile Prioritization of Positional Candidate Glaucoma Genes The GLC1C Locus Frank W. Rozsa, PhD; Kathleen M. Scott, BS; Hemant Pawar, PhD; John R. Samples, MD; Mary K. Wirtz, PhD; Julia E. Richards, PhD Objectives: To develop and apply a model for priori- est because of moderate expression and changes in tization of candidate glaucoma genes. expression. Transcription factor ZBTB38 emerges as an interesting candidate gene because of the overall expres- Methods: This Affymetrix GeneChip (Affymetrix, Santa sion level, differential expression, and function. Clara, Calif) study of gene expression in primary cul- ture human trabecular meshwork cells uses a positional Conclusions: Only1geneintheGLC1C interval fits our differential expression profile model for prioritization of model for differential expression under multiple glau- candidate genes within the GLC1C genetic inclusion in- coma risk conditions. The use of multiple prioritization terval. models resulted in filtering 7 candidate genes of higher interest out of the 41 known genes in the region. Results: Sixteen genes were expressed under all condi- tions within the GLC1C interval. TMEM22 was the only Clinical Relevance: This study identified a small sub- gene within the interval with differential expression in set of genes that are most likely to harbor mutations that the same direction under both conditions tested. Two cause glaucoma linked to GLC1C. genes, ATP1B3 and COPB2, are of interest in the con- text of a protein-misfolding model for candidate selec- tion. SLC25A36, PCCB, and FNDC6 are of lesser inter- Arch Ophthalmol. 2007;125:117-127 IGH PREVALENCE AND PO- identification of additional GLC1C fami- tential for severe out- lies7,18-20 who provide optimal samples for come combine to make screening candidate genes for muta- adult-onset primary tions.7,18,20 The existence of 2 distinct open-angle glaucoma GLC1C haplotypes suggests that muta- (POAG) a significant public health prob- tions will not be limited to rare descen- H1 lem. Genetic components to glaucoma are dants of a single founder, so GLC1C mu- suggested by high concordance for POAG tations may be found in multiple current among older monozygotic twins2 as well glaucoma genetics study populations. as mapping of a large number of glau- Here we describe expression data for 41 coma risk factor loci3,4 plus 12 POAG loci, genes within the GLC1C genetic inclu- called GLC1 loci.5-14 No unifying cellular sion interval and 7 high-priority candi- or biochemical themes have emerged from date genes selected based on overall or dif- the identification of the first 3 GLC1 genes: ferential expression. These data should Author Affiliations: myocilin (MYOC; Online Mendelian In- help to inform the work of many groups Departments of Ophthalmology heritance in Man [OMIM] 601652) at the in the field who are looking for genes that and Visual Sciences (Drs Rozsa, GLC1A locus,15,16 optineurin (OPTN; cause glaucoma. Pawar, and Richards and OMIM 602432) at the GLC1E locus,17 and Ms Scott) and Epidemiology WD-repeat36 (WDR36)atGLC1G11; these METHODS (Dr Richards), University of Michigan, Ann Arbor; and genes do not account for most of the fa- Department of Ophthalmology, milial cases of POAG. CELL LINES Oregon Health and Science The GLC1C locus is of special interest University, Portland because the gene itself has not been found For this institutional review board–approved (Drs Samples and Wirtz). but the locus has been confirmed through study, eyes were obtained from the Midwest (REPRINTED) ARCH OPHTHALMOL / VOL 125, JAN 2007 WWW.ARCHOPHTHALMOL.COM 117 ©2007 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/01/2021 Table 1. Primers Used in Quantitative Polymerase Chain Reaction Genomic cDNA Gene Forward Primer Reverse Primer Size, bp Size, bp GLC1C-Region Genes Not on the U133A GeneChip, 5؅ to 3؅ ACPL2 GTGCATGTGTTCATTCGCCAC TCAGCAGCTGACCGTTCTGCA 12 894 271 DZIP1L CAGCTAGAAGCTCCAGCAAAG GCAGCCTCTTCTGTTGAAGTG 1991 341 FAM62C ATCAGAGTGCACTTGCTGGAG GTGGCCACATTGTGCACAAAG 5777 654 FNDC6 CCCTGTTTGCCTTTGTTGGCT TGTCCCCTCATGGTTTCCATG 14 852 296 GRK7 GTTGCTGGACGAACACCATTC GGTTCAATTAGGCCAGCTTCC 9048 254 KY ATGACATTGCAGCTGCTCAGG ACTTGAGCGTGTACTCCAGCA 15 001 708 MGC40579 TTCTCTAGCTGGCCTTGCTGT TCCCTGAGCTTCATCTCATCTG 4796 242 NMNAT3 GCAAACTGCTCAGATCTCCAC AAGGCTCGCCTGATGTATGTG 12 427 322 RBP2 AAGGGACCAGAATGGAACCTG CTTAACATGCCGGTTATCCAG 14 343 247 SPSB4 ATGAGGGCACACTCAGCTTCA GCTCACTGGTACTGCAGATAG 80 617 274 TRIM42 AGCAACACTGACAAGAAGGCC CTGAGGTTGTTTCCTGCTCTG 12 694 574 TXNDC6 TCAGAGAGGACCTGTACCTTG CAACCTTGGCTGTGTACTGGA 43 915 489 XRN1 CTGGGAAGCCCTTCCATCATA TGGCTAGTCTGAACTGGAGTG 965 181 Non-GLC1C Genes for Normalization and Regression, 5؅ to 3؅ ACTA2 ATGGAGTCTGCTGGCATCCAT CCGGCTTCATCGTATTCCTGT 2987 293 FST GGGATTTCAAGGTTGGGAGAG GGAAAGCTGTAGTCCTGGTCT 961 238 G1P2 GCAGCGAACTCATCTTTGCCA CAGAGGTTCGTCGCATTTGTC 718 311 GAPDH AAGGTGAAGGTCGGAGTCAAC CCTGGAAGATGGTGATGGGAT 1955 229 GREM2 AGACCAAACTTAGACCCCGCT TGCCATCTCTCCGAGTTGTTG 118 656 249 HGF GGGAAGGTGACTCTGAATGAG GTGGGTGCTTCAGACACACTT 2906 274 KLF4 CACCCACACTTGTGATTACGC GACTCAGTTGGGAACTTGACC 1493 392 LIF ACGCCACCCATGTCACAACAA TGGGGTTGAGGATCTTCTGGT 971 278 MFGE8 AAACGCGGTGCATGTCAACCT CGTTGAAGTTGCCCTGCTTGT 4287 266 MYOC TATCTCAGGAGTGGAGAGGGA CTGGCTGATGAGGTCATACTC 2103 216 PTGS1 CAGGAACATGGACCACCACAT TCCGGAGAACAGATGGGATTC 2161 318 SULF1 TGGAGCTCAGAAGCTGTCAAG CATAGTGACTCTTCAGCAGTG 19 907 295 TNFAIP6 AGGAGTGTGGTGGCGTCTTTA CCAGCTGTCACTGAAGCATCA 9390 316 Abbreviations: bp, base pairs; cDNA, complementary DNA. Eye Banks (Ann Arbor, Mich), which obtained informed con- test, with PϽ.005 considered statistically significant. Scatter- sent and confirmed that no donors had been diagnosed with plots were drawn using Spotfire DecisionSite 8.2 software (Spot- glaucoma. Fifth-passage primary cultures of human trabecu- fire, Inc, Cambridge, Mass). lar meshwork (TM) cells used in age experiments came from Of the genes in the GLC1C interval between D3S3637 and 12- and 60-year-old females and 16- and 74-year-old males. D3S3694 obtained from the University of California, Santa Cruz, Cells were grown to confluence and maintained for 1 week Genome Browser (http://genome.ucsc.edu),18 28 were assayed before isolating RNA as previously described.21 In dexametha- by the U133A GeneChip and 13 were assayed by quantitative sone studies, fifth-passage TM cells were derived from young polymerase chain reaction (qPCR) using primers (Table 1) donors aged 12 and 16 years (as described in this article ear- in iQ SYBR Green Supermix reactions (BioRad Laboratories, lier) and a 17-year-old girl. The TM cells were grown with and Hercules, Calif) as previously described.21 We did not per- without a 21-day course of 100nM dexamethasone as previ- form qPCR confirmation of all of the GeneChip values be- ously described.21 cause results from previous experiments show GeneChip data to be adequate in the signal range being evaluated here. MICROARRAY ANALYSIS QUANTITATIVE PCR Labeling, hybridization to Affymetrix U133A GeneChips (Af- fymetrix, Santa Clara, Calif), and data extraction were done as Data from qPCR were assigned an estimated Affymetrix signal 21 previously described. The RNA was prepared from 2 sepa- value (EASV) using linear regression. The log2 values of Af- rate flasks for each culture, and conditions and data from bio- fymetrix signals from 9 control genes (Table 1) were plotted logical replicates were compared as previously described.21 against their corresponding qPCR threshold values normal- Image analysis was performed using Affymetrix Microar- ized to glyceraldehyde-3-phosphate dehydrogenase (Figure 1). ray Analysis Suite version 5.1 with samples scaled to 1500. Sig- These data are from the same RNA and GeneChips used for nals below 300 were considered absent and signals between 300 the other aspects of this study. The plot had an R2 value of and 750 were considered marginal. Values above 750 were con- 0.7459, and imputed values were calculated using the equa- ϫ ϩ [(C t −0.3866) 21.517] sidered present. tion EASV=2 , where Ct is the threshold value Fold change for each probe was calculated by dividing the for qPCR. The validity of the regression was examined by com- mean “treated” signal by the mean “untreated” signal in the dex- paring the actual and imputed signal values using qPCR data amethasone experiments and the mean “old” signal by the mean from 3 genes that were not used to construct the curve (Table 2). “young” signal in the age experiments. The inverse function When comparing data for the 3 control genes under 4 condi- was applied to values below 1 to indicate decreased expres- tions, we found that 8 of 12 data points showed less than a 2.5- sion. The significance of fold change was evaluated using the t fold difference between the imputed and actual Affymetrix val- (REPRINTED) ARCH OPHTHALMOL / VOL 125, JAN 2007 WWW.ARCHOPHTHALMOL.COM 118 ©2007 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/01/2021 ues. In a separate experiment in which data from a different RESULTS RNA sample tested on the Affymetrix U133Av2 version 2 chips were examined for the 6 experimental genes for which we need to impute values, we found that 5 of 6 measured data points The 7 879 854–base pair GLC1C interval between fell within 2.5-fold of the imputed values. D3S3637 and D3S3694 contains 41 genes (Table 3). We The EASV gives us an approximate value for comparing data evaluated gene expression under