Intracellular signalling pathways and cytoskeletal functions converge on the psoriasis candidate gene CCHCR1 expressed at P-bodies and centrosomes

Intracellular signalling pathways and cytoskeletal functions converge on the psoriasis candidate... Background: CCHCR1 (Coiled-Coil α-Helical Rod protein 1) is a putative psoriasis candidate gene with the risk alleles CCHCR1*WWCC and *Iso3, the latter inhibiting the translation of isoform 1. CCHCR1 was recently shown to be a centrosomal protein, as well as a component of cytoplasmic processing bodies (P-bodies) that regulate mRNA turnover. The function of CCHCR1 has remained unsettled, partly because of the inconsistent findings; it has been shown to play a wide variety of roles in divergent processes, e.g., cell proliferation and steroidogenesis. Here we utilized RNA sequencing (RNAseq) using HEK293 cells overexpressing isoforms 1 or 3 (Iso1, Iso3 cells), in combination with the coding non-risk or risk (*WWCC) haplotype of CCHCR1. Our aim was to study the overall role of CCHCR1 and the effects of its variants. Results: The overexpression of CCHCR1 variants in HEK293 cells resulted in cell line-specific expression profiles though several similarities were observable. Overall the Iso1 and Iso3 cells showed a clear isoform-specific clustering as two separate groups, and the Non-risk and Risk cells often exhibited opposite effects. The RNAseq supported a role for CCHCR1 in the centrosomes and P-bodies; the most highlighted pathways included regulation of cytoskeleton, adherens and tight junctions, mRNA surveillance and RNA transport. Interestingly, both the RNAseq and immunofluorescent localization revealed variant-specific differences for CCHCR1 within the P-bodies. Conclusions: CCHCR1 influenced a wide variety of signaling pathways, which could reflect its active role in the P-bodies and centrosomes that both are linked to the cytoskeleton; as a centrosomal P-body protein CCHCR1 may regulate diverse cytoskeleton-mediated functions, such as cell adhesion and -division. The present findings may explain the previous inconsistent observations about the functions of CCHCR1. Keywords: CCHCR1, PSORS1, Centrosome, Cytoskeleton, P-body, Cell adhesion, RNAseq, Expression profiling * Correspondence: juha.kere@ki.se; outi.elomaa@helsinki.fi Folkhälsan Institute of Genetics, 00014 Helsinki, Finland Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Tervaniemi et al. BMC Genomics (2018) 19:432 Page 2 of 15 Background repression, and RNA-mediated gene silencing [16, 17]. Psoriasisisachronicmultifactorial dermatological disorder, There is a link between the P-bodies and centrosomes; characterized by abnormal proliferation and differentiation P-bodies transit along microtubule tracks to and from the of keratinocytes and infiltration of inflammatory cells. centrosome and furthermore, a pair of stationary P-bodies There are several susceptibility gene loci for psoriasis reside at the centrosome during the interphase [18, 19]. (PSORS1–15) (based on the Online Mendelian Inheritance Using stable HEK293 cells overexpressing isoforms in Man, OMIM); of which PSORS1 (6p21.3) has the stron- 1or3 with thenon-risk(*RRGS) or the risk haplo- gest risk effect [1]. Diverse psoriasis-associated alleles have type (*WWCC) we have shown that isoform 3 acceler- been identified within the region. However, a strong linkage ates cell proliferation and together with the *WWCC disequilibrium has made it difficult to distinguish their indi- allele apoptosis as well. Whereas isoform 1 lacks sig- vidual effects. Hence, the effector genes in psoriasis within nificant effects on cell proliferation or cell cycle pro- the 6p21.3 region are currently not fully understood. gression. Furthermore, the CCHCR1-HEK293 cell CCHCR1 (Coiled-Coil α-Helical Rod protein 1) is a putative lines show isoform- and haplotype-specific changes in PSORS1 candidate gene among others [2–4], and its cell size and shape and have alterations in the CCHCR1*WWCC allele is associated with psoriasis in sev- organization and expression of the cytoskeletal pro- eral populations [2, 3, 5]. WWCC stands for the amino teins actin, vimentin, and cytokeratins. We also dem- acids in the psoriasis risk haplotype, whereas in the onstrated that CCHCR1 may regulate EGF-induced non-risk haplotype the corresponding amino acids are STAT3 activation in an isoform-specific manner [6]. RRGS. We have previously described a novel form of Here we applied 5’end-targeted RNA sequencing (RNA- CCHCR1, isoform 1, where the N-terminal domain is lon- seq) on the previously established and characterized gerthaninisoform 3 [6]. The formation of isoform 1 is CCHCR1-HEK293 cell lines (Iso1Risk, Iso1Non-risk, dependent on a SNP (rs3130453) that results in either a Iso3Risk, and Iso3Non-risk) [6, 20]to examine isoform- longer open reading frame (allele *Iso1)or a stop codon and haplotype-specific effects on global gene expression (allele *Iso3), enabling only the translation of isoform 3. In profiles, and to identify the underlying mechanisms lead- addition, there are two transcription start sites, only one of ing to the previously observed differences between the them enabling the production of long isoform 1. Notably, CCHCR1 cell lines. The sensitivity of RNAseq allowed an we demonstrated that *Iso3 shows association with psoriasis in-depth assessment of altered gene expression in the cell − 7 (P <10 )[6]. lines, making it possible to get a better understanding of The CCHCR1 protein does not belong to any known the effects of CCHCR1 isoforms on the cellular pathways. protein family but is predicted to be a rod-like protein, with an alpha-helical coiled coil structure. Although its Results expression pattern in psoriatic skin differs from healthy CCHCR1-HEK293 cell lines exhibited isoform- and skin [7, 8] and its overexpression influences cell prolifer- haplotype-specific gene expression profiles ation [6, 9, 10], its role as a putative psoriasis effector In the present study we used stably transfected gene has remained unsettled, partially because it has CCHCR1 cell lines that were previously generated been suggested to function in several divergent bio- into human embryonic kidney (HEK) 293 cells. HEK293 is logical processes such as the regulation of steroidogene- an adenovirus-transformed cell line that is considered sis via mitochondrial steroidogenic activator protein tumorigenic. The establishment and characterization of (STAR) [8, 11] or muscle differentiation via RNA poly- the CCHCR1 overexpressing HEK293 clones have been merase II subunit 3 (RPB3) [12]. We have also demon- described previously [6]. Briefly, the CCHCR1 cell lines strated that in addition to its cytoplasmic location, both (Iso1Risk, Iso1Non-risk, Iso3Risk, and Iso3Non-risk) were CCHCR1 isoforms localize at the centrosome [6], a cell screened for the CCHCR1 expression by qPCR, fluores- organelle playing a crucial role in cell division. The cence microscopy, and western blotting. The clones used centrosomal localization of CCHCR1 has been verified in the present work were chosen based on the previous in other large-scale proteomics studies [13, 14]. studies; we selected clones with similar CCHCR1 expres- Furthermore, CCHCR1 has been shown to interact with sion levels and cell morphology. As most of the previous mitotic spindle proteins [15], providing evidence for its CCHCR1 studies have focused on isoform 3, we first possible involvement in cell division. A recent study performed gene expression profiling using Affymetrix demonstrated that, in addition to its location at the Human Gene ST 1.0 microarrays for the Iso1Non-risk centrosome, CCHCR1 is a component of processing and Iso1Risk CCHCR1 cell lines to increase knowledge of bodies (P-bodies) [15]. P-bodies are cytoplasmic ribonu- isoform 1. Wild type HEK293 and vector-transfected cells cleoprotein granules that regulate mRNA turnover in a were used as controls in the comparative expression post-transcriptional manner and are involved in mRNA analysis. The comparison of Iso1Non-risk and Iso1- degradation, surveillance, and transport, translational Risk with the vector control revealed 296 and 206 Tervaniemi et al. BMC Genomics (2018) 19:432 Page 3 of 15 differentially expressed genes (DEGs) (Cut-off fold lowest in Iso3Non-risk (2806) (up: FC > 1.5, down: change1.3-fold, p < 0.002), respectively (Additional file 1: FC < 0.75, FDR 0.25). All the CCHCR1 cell lines Table S1). Gene enrichment analyses (the Database shared 209 upregulated and 618 downregulated genes for Annotation, Visualization and Integrated Discovery (Venndiagram of DEGs showninFig. 1a). (DAVID) [21, 22], WebGestalt 2013, and WebGestaltR There were notable differences in the expression pro- 2017 [23]) of the DEGs (> 1.5-fold) from the files between the CCHCR1-HEK293 cell lines. We esti- Iso1Non-risk cells revealed such pathways and func- mated the dissimilarity between the cell lines by − 4 tions (p <1 ×10 )as focal adhesion, ECM-receptor principal component analysis (PCA) (Fig. 1b) that en- interaction,and regulation of actin cytoskeleton.The abled multiclass comparisons between samples [24]. The results of the gene enrichment analysis are shown in second PC (PC2) depicts the isoform-dependent differ- detail in Additional file 1:Table S1.Thoughthe ence, clustering isoform 1 (Iso1) and 3 (Iso3) as separate microarray study suggested differences in the gene ex- groups. The wild type and vector control samples clus- pression between Iso1Non-risk and Iso1Risk, the cell tered as an overlapping group between the Iso1 and Iso3 lines sharing only 50 genes, the sensitivity of the ex- samples. The PCA demonstrated that the replicates dif- pression profiling remained rather modest. Particularly fered to some extent, especially the Iso3Risk samples in the Iso1Risk cell line only a few pathways, such as (the outlier specimen 3Rc, Fig. 1b). The clustering of cell cycle, focal adhesion,and leukocyte migration Risk and Non-risk samples as separate groups was less showed enrichment of more than four genes when clear than observed with Iso1 and Iso3. WebGestalt was used. To get a higher sensitivity as- sessment of altered gene expression in all four Expression profiling of CCHCR1-HEK293 cell lines CCHCR1-HEK293 cell lines we performed RNAseq. highlighted cell adhesion and regulation of cytoskeleton RNA samples extracted from the wild type HEK293, Using the WebGestalt tool (WebGestalt 2013 and Web- the vector control, and the CCHCR1 cell lines (Iso1- GestaltR) for gene enrichment analysis, we analyzed the Non-risk, Iso1Risk, Iso3Non-risk, and Iso3Risk), were DEGs of four different comparisons (DEGs shown in subjected to RNAseq. The comparisons of transcripts Additional file 2: Table S2: Iso1Non-risk, Iso1Risk, from each CCHCR1-HEK293 cell line with the con- Iso3Non-risk, and Iso3Risk cells compared with wild trols (wild type and vector) revealed a large number type and vector controls). The KEGG pathway analysis of differentially expressed genes (DEGs) that are listed of up- and downregulated genes revealed several similarities in Additional file 2: Table S2. The number of upregu- between the CCHCR1-HEK293 cell lines (Additional file 3: lated genes was highest in Iso3Non-risk (6602) and Table S3 showing the results of gene enrichment analysis). lowest in Iso3Risk (1781), while the number of down- Adhesion-related pathways, such as focal adhesion, gap regulated genes was highest in Iso3Risk (6968) and junction, tight junction, adherens junction,and regulation of Fig. 1 Expression profiles of the CCHCR1-HEK293 cell lines. a Venn diagrams of differentially expressed genes (DEGs) of the CCHCR1-HEK293 cell lines; Iso1Non-risk (1 N), Iso1Risk (1R), Iso3Non-risk (3 N), and Iso3Risk (3R), when compared with vector and wild type controls (FC > 1.5 or FC < 0.75, FDR < 0.25). The most highlighted functions and pathways of the gene enrichment analysis of DEGs that were shared by the all four cell lines are shown. Up- and downregulated genes were analyzed separately (see in detail Additional file 3: Table S3). b Principal component analysis (PCA) demonstrates differences between CCHCR1-HEK293 cell lines. The two PCs depict the variation between samples. The PC2 illustrates the isoform-dependent difference, clustering isoforms 1 and 3 as separate groups. The wild type (wt) and vector (v) control samples cluster as an overlapping group between the isoform 1 and isoform 3 samples Tervaniemi et al. BMC Genomics (2018) 19:432 Page 4 of 15 actin cytoskeleton,as wellas Wnt signaling and calcium the gene expression were apparent, which is observable signaling, were highlighted in all the cell lines. In addition, from the Venn diagram (Fig. 1a) and the heatmaps. For MAPK, ErbB,and TGF-beta were among the top instance Wnt signaling (heatmap in Fig. 3a) was mainly pathways. Their expression profiles, however, showed vari- upregulated only in the Iso3Non-risk cells when com- ation between the control samples. The most highly up- pared with controls (Additional file 3: Table S3). While 8 − 7 and downregulated genes (1 × 10 <FC<1 × 10 )were several pathways were mainly downregulated in the enriched especially in functions and pathways that are Iso3Risk cells, these including regulation of actin cyto- involved in neuroactive ligand–receptor interaction, cell skeleton (heatmap Fig. 2a) and RIG-I-like receptor signal- adhesion, calcium signaling,and cytokine-cytokine receptor ing (heatmap in Fig. 3b; Additional file 3: Table S3). As interaction (Additional file 3: Table S3). CXCL8 (alias IL8) is involved in several signaling path- Although the same pathways were showing enrichment ways that were highlighted in our expression data, as in all the CCHCR1-HEK293 cell lines, their gene expres- well as related to psoriasis, we validated its expression in sion profiles differed; different genes were highlighted, and the CCHCR1-HEK293 cell lines by qPCR; IL8 showed the gene expression varied from down- to upregulation reduced expression in the Iso3 cell lines and upregula- between the cell lines. For example, focal adhesion and tion in the Iso1cells, especially in Iso1Non-risk (Fig. 3c), regulation of actin cytoskeleton were the top pathways in as observed in RNAseq. all the lines (Additional file 3: Table S3). However, the cor- To study the isoform and haplotype-specific effects of responding heatmaps showed differences between the cell CCHCR1 more specifically, and to get more statistical lines; the Non-risk and Risk cells often exhibited opposite power for the study, we analyzed the DEGs in four dif- effects (Fig. 2a showing heatmaps of focal adhesion and ferent groups; DEGs shared by the Iso1, Iso3, Non-risk, regulation of actin cytoskeleton). Several relevant genes of or Risk cell lines (Fig. 4: the Venn diagram of compari- focal adhesion and actin cytoskeleton were downregulated sons). The DAVID and WebGestalt (versions 2013 and particularly in the Iso3Risk cells. To verify this, we quanti- 2017) analyses revealed differences between the groups fied the expression of TLN1 (talin 1) and FN1 (fibronectin (Additional file 4: Table S4 showing gene enrich analysis 1) by qPCR (shown in Fig. 2b). They were highly downreg- in detail), though the number of gene counts in ulated in both isoform 3 expressing cell lines (Iso3), espe- highlighted pathways remained modest. Tight junction cially in Iso3Risk, whereas their expression was increased was highlighted in the shared DEGs of the Iso1 cells in the Iso1Non-risk cells similarly as was observed in (FDR = 0.06), progesterone mediated oocyte maturation RNAseq. We also examined the expression of SYT1 (syn- in the Iso3 cells (FDR < 0.05), mRNA surveillance aptotagmin 1) by qPCR (Fig. 2b); one of the most highly pathway in the Non-risk cells (FDR = 0.02), and antigen upregulated genes in the Iso3 cells. The qPCR verified that processing and presentation in the Risk cell lines (FDR < its expression was increased only in the Iso3 cells and was 0.05). The gene enrichment analysis of DEGs including unaltered in both isoform 1 expressing cells (Iso1). both the up-and downregulated genes that were shared Overall, the qPCR verification results agreed with the by all the CCHCR1-HEK293 cell lines (Fig. 4 Venn dia- RNAseq data. We have previously shown that our RNA- gram) highlighted only a few pathways with FDR < 0.1 seq protocol enables the accurate quantitation of gene ex- when analysed by WebGestaltR, these including calcium- pression [24]. and estrogen signaling (FDR = 0.05). WebGestalt highlighted To study the overall biological role of CCHCR1, we cytokine-cytokine receptor interaction, NOD-like receptor analyzed the DEGs that were shared by the four signaling,and adherens junction as well. CCHCR1-HEK293 cell lines (827 genes, 0.75 > FC > 1.5, As the Iso1 and Iso3 cell lines showed a clear FDR < 0.25). The analyses of shared DEGs (Additional isoform-specific clustering in the PCA (Fig. 1b), we pooled file 3: Table S3: Intersection up and down) identified the Iso1Non-risk and Risk samples and compared with such pathways and genes as: calcium signaling (PLCD1, the pooled vector and wildtype data (comb_Iso1), and PLCB2), NOD-like receptor signaling (NLRP3, CARD9, similarly we pooled and compared Iso3Non-risk and Risk and CXCL1), cytokine-cytokine receptor interaction vs vector and wildtype (comb_Iso3) to increase the num- (IL12A, IL22RA1, and VEGFC), cell adhesion (ITGAL, ber of replicates and to get more statistical power for the CNTN1, and CLDN7), and regulation of actin cytoskeleton study. The DEGs were re-extracted by a two class un- (PAK3, RRAS,and ITGAX), however, the gene counts of paired comparison using a higher permutation (n = 1000). the detected pathways were modest. (Additional file 5: Table S5 showing the DEGs and the re- sults of gene enrichment analysis for comb_Iso1 and com- Isoform- and haplotype-specific effects of CCHCR1 on cell b_iso3). We also pooled the data from Iso1Non-risk and signaling Iso3Non-risk (comb_Non-risk), and from Iso1Risk and Despite similarities between the CCHCR1-HEK293 cell Iso3Risk (comb_Risk) (Additional file 6: Table S6 showing lines, some isoform- and haplotype-specific effects on the DEGs and the results of gene enrichment analysis for Tervaniemi et al. BMC Genomics (2018) 19:432 Page 5 of 15 Fig. 2 Expression profiling of the CCHCR1-HEK293 cell lines highlighted cell adhesion and actin cytoskeleton. a Heatmaps showing gene expression of focal adhesion and regulation of actin cytoskeleton in the CCHCR1 Iso1Non-risk (1 N), Iso1Risk (1R), Iso3Non-risk (3 N), and Iso3Risk (3R) cell lines, and in the vector transfected cells (V), and wild type HEK293 (WT). Color key: red represents upregulated and blue downregulated expression (row Z-score). b The expression of FN1, TLN1,and SYT1 in the CCHCR1-HEK293 cell lines was validated with qPCR, and compared with the gene expression in the wild type cells. Error bars: standard deviation. Here shown normalization with HPRT1 Tervaniemi et al. BMC Genomics (2018) 19:432 Page 6 of 15 Fig. 3 Gene expression profiles of Wnt and RIG-I-like signaling in the CCHCR1-HEK293 cell lines. Both pathways were highlighted in the CCHCR1 cell lines. Heatmaps of Wnt (a) and RIG-I-like (RLR) (b) signaling in the Iso1Non-risk (IN), Iso1Risk (1R), Iso3Non-risk (3 N), and Iso3Risk (3R) CCHCR1 cell lines, vector transfected cells (v), and wild type HEK293 (wt) demonstrated that there were differences between cell lines. Wnt signaling is mainly upregulated only in the Iso3Non-risk cells while RLR signaling shows more downregulation in Iso3Risk than in other cell lines. Color key: red represents upregulated and blue downregulated expression (row Z-score). c The expression of CXCL8 in the CCHCR1-HEK293 cell lines was validated with qPCR, and compared with the gene expression in the wild type cells. Here shown normalization with HPRT1. Error bars: standard deviation Tervaniemi et al. BMC Genomics (2018) 19:432 Page 7 of 15 Fig. 4 Isoform- and haplotype-specific differences between the CCHCR1-HEK293 cell lines. The Venn diagram of the DEGs (FC > 1.5 or < 0.75, FDR < 0.25) from the CCHCR1-HEK293 cell lines Iso1Non-risk (1 N), Iso1Risk (1R), Iso3Non-risk (3 N), and Iso3Risk (3R) when compared with the control cells, and the DEGs from the psoriatic skin samples (Ps) [20]. The most highlighted gene enrichment pathways of DEGs (both the up- and down- regulated) that were shared by the Non-risk, Risk, Iso1, Iso3, or all the CCHCR1 cell lines are shown. The DEGs used for the gene enrichment analysis are shown in parentheses and indicated by color coded rectangles in the diagram. The DEGs (diff) and the results are shown in detail in the Additional file 4: Supplement Table S4 comb_Non-risk and comb_Risk), although the haplotypes presentation (FDR = 0.02). The latter and HLA-genes were showed less clustering than the isoforms. To control the highlighted also in the pathways analysis of comb_Risk specificity of the results we also generated mock gene lists which otherwise showed much less overrepresented path- by permuting the sample groupings in the comparison ways than the other groups. (Additional file 6: Table S6 showing the summary of gene enrichment analyses of the mock gene lists). Several path- Gene expression in the CCHCR1-HEK293 cell lines and ways were highlighted among the most mock gene lists. psoriatic skin There were, however, pathways that were highlighted spe- The comparison of transcriptomes from two completely dif- cifically among the CCHCR1-HEK293 clones; adherens ferent biological systems, such as the CCHCR1-HEK293 cell junction and tight junction in comb_Iso1, and RNA trans- lines and psoriatic skin, is challenging and the direct com- port and mRNA surveillance in comb_Iso3 (FDR < 0.002) parison of gene expression changes is unfeasible. Some simi- and comb_Non-Risk (FDR = 0.01) (Additional file 5:Table larities were observable when the current expression results S5; Additional file 6: Table S6), which agreed the results of CCHCR1 cells were compared to our previous RNAseq obtained with the cell line specific DEGs (Additional file data on psoriatic skin obtained from the comparison be- 3: Table S3). RIG-I-like receptor signaling, which was tween psoriatic lesional and non-sessional skin, and healthy highlighted in the pathway analyses of individual CCHCR1 control skin [20]. The comparison revealed some similarities clones (Additional file 3: Table S3), rose up in comb_Iso1 in the gene expression changes, such as upregulation of S100 (FDR = 0.06) as well. Progesterone mediated oocyte mat- and KRT (KRT6, KRT16,and KRT17) genes, and enrichment uration was highlighted in comb_Iso3 (FDR = 0.01) as was of the steroid-related pathways. Some genes of RIG-I-like re- also detected in the analysis of the shared Iso3 DEGs ceptor signaling (DDX58, CXCL8, and CASP8)showedal- (Additional file 4: Table S4; Additional file 5: Table S5). tered expression in the CCHCR1-HEK293 cells and psoriatic Mitochondria that play a role in the sythesisof steroids skin, as well. The results, however, should not be over inter- was highlighted among the comb_Iso3 DEGs as well. Sev- preted, as psoriasis is a complex disease and CCHCR1 is a eral HLA genes were overrepresented especially in com- susceptibility gene among several others, and as mentioned b_Iso3, highlighting pathways such as systemic lupus the direct comparison of CCHCR1 cells and skin samples is erythematosus (FDR = 0.0002) and antigen processing and not warranted. Tervaniemi et al. BMC Genomics (2018) 19:432 Page 8 of 15 We also examined the CCHCR1 genotype (*Iso1 or altered expression in Iso3Non-risk (the KEGG pathway of *Iso3, *Non-risk or *Risk allele) and the most well-known mRNA surveillance is shown in detail in Additional file 3: psoriasis haplotype HLA-Cw*06:02 in our psoriatic and Table S3: mRNA surveillance). healthy skin samples. The genotyping results are shown Here, we studied by immunofluorescence staining with in detail in Additional file 7: Table S7. We genotyped the the P-body markers EDC4 and DCP1A (Fig. 5b, c)[15] SNPs for rs3130453 (G/A) resulting *Iso1 or *Iso3 allele and the centrosomal marker γ-tubulin (Additional file 8: and rs130076 (C/T) corresponding to the second R/W Supplementary information Figure S2) whether the dif- amino acid change in CCHCR1*WWCC risk allele. Five ferent CCHCR1 isoforms localize with the P-bodies in out of seven psoriatic samples were homozygous for the the same manner. We observed differences in the stain- *Iso3 allele, whereas only two out of eight were homozy- ing pattern between the isoforms; in the Iso1 cells, espe- gous in the controls. None of the psoriatic samples had cially in Iso1Non-risk, the P-body marker colocalized the homozygous *Iso1 genotype. Only two of the rarely with CCHCR1, whereas in the Iso3 cells the psoriatic patients were heterozygous and therefore P-body staining colocalized with the CCHCR1 granules − 19 able to translate the longer isoform 1 as well. Five (P <4×10 ) (Additional file 8: Supplementary infor- out of seven psoriatic samples were heterozygous for mation describes the counting of colocalization of the Risk (*WWCC) haplotype whereas five out of CCHCR1 and P-body markers). In addition, we demon- eight controls had the homozygous Non-risk haplo- strated by γ-tubulin staining that in the Iso1 cells type. Most of the control samples were negative for P-body markers showed less accumulation at the centro- the psoriasis haplotype HLA-Cw*06:02 (six out of somes (Additional file 8: Figure S2). eight) but most of the psoriasis samples were hetero- zygous positive (five out of seven). The genotyping of psoriatic samples agrees the previous results [20], but Discussion the number of samples didn’t allow studying the cor- Our results show the power of RNAseq providing a large relation of CCHCR1 genotype and gene expression amount of detailed expression data on the HEK293 cells profiles. The systematic identification of all potentially expressing the CCHCR1 variants. The principal compo- causal variants in psoriasis in order to study the ef- nent and gene enrichment analyses demonstrated that fects of individual variants or their combinations the cell lines expressing different isoforms and haplo- would be an extremely laborious task to implement types of CCHCR1 have their distinct, specific expression in practice. The CCHCR1 and HLA-Cw6 genotypes profiles. The approach also strengthened the previous were illustrated in a PCA plot that clusters the sam- findings of the putative roles of CCHCR1 in the ple types (Additional file 7:FigureS2). regulation of cytoskeleton (Fig. 6) and steroid biosyn- thesis [6, 8]. Additionally, the RNAseq revealed that CCHCR1 affects P-body-related functions and shows CCHCR1 affected the cell signaling pathways mediated isoform-specific localization with P-bodies via the P-bodies, which provided evidence of its possible The recent study [15] that localized CCHCR1 at active role at the P-bodies. Furthermore, we observed P-bodies was carried out with the Iso3Non-risk form of isoform- and haplotype-specific differences in its effects CCHCR1. Interestingly in our RNAseq data, the on the P-body-related functions, as well as on the P-body-related pathways RNA transport, RNA degradation, localization with the P-bodies. The P-body localization and mRNA surveillance were highlighted in the Iso3 cells may also explain some of the putative interactions of − 15 (WebGestalt p <1 ×10 ) and especially in the Non-risk CCHCR1 that have been previously identified with such cells (FDR < 0.1) (Additional file 5: Table S5; Additional proteins as StAR, RPB3, and viral proteins (Fig. 6)[11, file 6: Table S6). Also the comparison Iso3Non-risk cells 12, 26]. We also compared the gene expression results versus controls revealed the enrichment of mRNA surveil- of the CCHCR1-HEK293 cell lines and psoriatic skin, − 9 lance (FDR = 1.7 × 10 ) (Additional file 3: Table S3: the though the direct comparison of transcriptomes of two gene enrichment results of 3 N); several SMG genes completely different biological sample types is unfeasible. (e.g. SMG1, SMG5,and SMG7) that code nonsense The approach was unable to give any final direct mediated mRNA decay factors of surveillance com- answers to the question what is the putative role of dif- plex [25], and CPSF genes encoding cleavage and ferent CCHCR1 variants. As an adenovirus-transformed polyadenylation specific factors were upregulated (1.5 cell line HEK293 cells have gone through structural al- < FC < 2.5). Heatmap of mRNA surveillance gene terations of genome. Therefore it is possible that we expression in the CCHCR1-HEK293 cell lines is have missed some gene expression data of the CCHCR1 shown in Fig. 5a. Although the gene expression cell lines. Furthermore the tumorigenic properties of changes were moderate, more than half of the genes HEK293 may have affected the gene regulation and -ex- acting in the mRNA surveillance pathway showed pression of the cell lines. Tervaniemi et al. BMC Genomics (2018) 19:432 Page 9 of 15 Fig. 5 CCHCR1 shows isoform-specific localization with centrosomal P-bodies and affects mRNA surveillance pathway. a Heatmap showing gene expression profiles of mRNA surveillance pathway in Iso1Non-risk (1 N), Iso1Risk (1R), Iso3Non-risk (3 N), and Iso3Risk (3R) CCHCR1-HEK293 cell lines, and in control cells; vector (V) and wildtype (WT). Color key: red represents upregulated and blue downregulated expression (row Z-score). b Immunofluorescent staining of the CCHCR1 (red) cell lines: Iso1Non-risk, Iso1Risk, Iso3Non-risk, and Iso3Risk, with the P-body markers (green) EDC4 and DCP1A. Panel b shows the overlap of red and green, and panel c shows the green channel images of the corresponding sites. The colocalization (orange) of CCHCR1 and P-body markers is absent in the Iso1Non-risk and Iso1Risk cells: only the CCHCR1 expression (red) is observable at the centrosomes and cytoplasm (b), whereas the antibodies against EDC4 and DCP1A fail to recognize these structures (c). In Iso3Non-risk and Iso3Risk, the colocalization (orange, marked with arrow heads in panel b) is observable with antibodies against the EDC4 and DCP1A (green in panel c). Nuclei are stained with DAPI (blue). Scale bar: 10 μm Tervaniemi et al. BMC Genomics (2018) 19:432 Page 10 of 15 Fig. 6 Model for CCHCR1 action in a cell and skin. CCHCR1 is expressed in a basal layer of keratinocytes in the epidermis of the healthy skin. In the thickened psoriatic epidermis, it is expressed in supra-basal layers as well. It is a dynamic protein that might play a role both in the cytoplasm at the centrosomes and P-bodies that occasionally colocalize, and possibly in the nucleus as well. Both centrosomes and P-bodies are physically linked to the cytoskeleton (cytokeratin (CK), actin, and microtubules (black lines)) that functions as a scaffold for the regulation of e.g. cell division and adhesion, mRNA transport and turnover, and transport of organelles like mitochondria. The previous results and the present RNAseq suggested that CCHCR1 regulates expression of cytoskeletal components and functions related to cytoskeleton, such as cell adhesion [6]. Furthermore, CCHCR1 may control the mRNA surveillance via the P-bodies, which may affect a wide variety of biological processes. Previous studies have identified several interaction partners for CCHCR1, these including the mitochondrial StAR protein that regulates steroidogenesis [8, 11], the RNA polymerase II subunit 3 (RPB3) that regulates transcription [12], some mitotic spindle proteins [15], and several viral proteins [30]. These interactions may result from the P-body localization of CCHCR1, or from the dynamic nature of CCHCR1 and its ability to regulate various processes via cytoskeleton. RNAseq identified several pathways for CCHCR1 which are relevant for the pathogenesis of psoriatic epidermis, epidermis development, and innate immunity. Interestingly, also the innate immunity including the RIG-like receptor (RLR) signaling is affected by the P-bodies CCHCR1 affects the expression of several genes adherens junctions specialized for cell-cell adhesion. The re- regulating cytoskeleton and cell-cell adhesion lated functions; regulation of actin cytoskeleton and focal Tight junction and adherens junction, were highlighted adhesion, were also highlighted in the CCHCR1-HEK293 cell especially in the Iso1 cells and showed gene enrichment in lines. They also showed enrichment among the mock genes. the Iso3 cells as well. Interestingly we have recently observed These two pathways consist of a large number of genes, CCHCR1 expression in the skin in the proximity of the cell hence their highlightment may result from several different membrane and desmosomes [6] that are like the tight and gene combinations and expression changes (up versus Tervaniemi et al. BMC Genomics (2018) 19:432 Page 11 of 15 down). Actin cytoskeleton and focal adhesion share several surveillance, the pathway analyses of the pooled data of genes with adherens and tight junctions that were not Non-risk and Iso3 highlighted RNA degradation and RNA highlighted among the mock genes. Furthermore, the func- transport. mRNA surveillance includes nonsense-mediated tions related to the cytoskeleton are supported by our previ- mRNA decay (NMD) that eliminates mRNAs containing ous microscopic observation; the actin cytoskeleton of the premature translation-termination codons [33]. In the Iso1 cells resembled that of wild type cells, but was disturbed RNAseq, several SMG genes encoding factors of NMD in the Iso3 cells. Furthermore, after disruption of microtu- surveillance complex, and CPSF genes encoding for cleav- bules, the Iso3Risk cells, especially, showed aberrant actin age and polyadenylation specific factors, were upregulated organization [6]. Here, we showed that several relevant in Iso3Non-risk. genes, including TLN1 (talin 1), FN1 (fibronectin 1), and P-bodies have been associated with several auto- FLNC (filamin C), were downregulated in the Iso3Risk cells, immune diseases, cancers, and neurological diseases which may disturb cytoskeleton organization. Talin [34, 35], but not with psoriasis. They are linked to vi- plays a role in the assembly of actin filaments and in- ruses and innate immunity as well. The P-bodies are teracts indirectly with extracellular fibronectin 1 [27], manipulated by the viruses for a productive infection andtheybothplayarolein adhesion, spreading, and to occur, but they also promote immune responses to migration of cells [28, 29]. Filamin crosslinks actin fil- the viral infection, e.g., via double-stranded RNA pro- aments into networks and anchor membrane proteins tein kinase (PKR alias EIF2AK2)[36]. Viruses, such as to the cytoskeleton, thereby affecting cell-cell and human papillomavirus HPV16 E6, induce in keratino- focal adhesion. Our previous results, as well as RNA- cytes the localization of PKR into the P-bodies sug- seq, also highlighted other cytoskeleton-related genes; gesting about the antiviral response. Interestingly, the several cytokeratin genes, e.g. KRT17 were upregu- interaction of CCHCR1 with HPV16 E2 induced a lated especially in the Iso1Non-risk cells [6]. One of massive redistribution of the virus protein into cyto- the most highly upregulated genes in the Iso3 cells was plasmic granules that were still co-localizing with SYT1. It encodes for synaptotagmin that interacts with CCHCR1 [30]. We suggest that these structures were tubulin, and may act as a microtubule-organizing cytoplasmic P-bodies playing an antiviral role. Inter- center-associated protein, thus playing a role in spindle estingly, RIG-I-like receptor signaling, which was organization [30, 31]. highlighted both in the CCHCR1-HEK293 cell lines and psoriatic skin, is part of the antiviral immune re- CCHCR1 shows isoform-specific effects on P-body sponse. The function of RIG-I-like receptors is linked localization and P-body-related functions at many levels to different RNA-granules including We have shown that in cultured cells CCHCR1 isoforms P-bodies [37]. 1 and 3 reside in the cytoplasm, as well as localize at the centrosomes [6, 8]. Interestingly, here we noticed a dif- Conclusions ference in their colocalization with the P-bodies in the The RNAseq highlighted pathways that support an ac- HEK293 cells. The immunostaining supported colocali- tive role for CCHCR1 in the centrosomes and P-bodies; zation with the P-bodies for isoform 3, especially at the the most significant pathways were related to the centrosome, while isoform 1 colocalized rarely with the cytoskeleton and cell adhesion, and furthermore, the P-bodies. The localization of P-bodies at the centro- mRNA surveillance pathway was highlighted in the somes seemed to be improved by the overexpression of Iso3Non-risk cells. The centrosomes and P-bodies are Iso3 CCHCR1. Recently it was shown that the physically linked to the cytoskeleton that functions as a N-terminus of Iso3Non-risk was necessary for the scaffold for the P-bodies (Fig. 6) and transported P-body localization [15], isoform 1 that is 89 amino acids mRNAs [17, 18], thus as a centrosomal P-body protein longer at its N-terminus than Isoform 3 was not along CCHCR1 might affect diverse cytoskeleton mediated in this study. The different N-terminus may explain the processes, such as cell division, cell adhesion, and trans- localization difference between isoforms 1 and 3. port of mRNA and mitochondria. The RNAseq revealed Previous studies have shown that there is a strong inter- clear isoform- and some haplotype-specific differences; action between CCHCR1 and EDC4 (enhancer of the Iso1 and Iso3 cells showed specific clustering as two mRNA-decapping 4 protein) [15] that is a central P-body separate groups, and the Non-risk and Risk cells often component regulating mRNA decay by decapping and exhibited opposite effects. Putative effects arising from 5′→ 3′ mRNA degradation. In our RNAseq, mRNA sur- the tumorigenic nature of HEK293 can’t be completely veillance pathway that is detecting and degrading abnor- ruled out. In summary, the present study strengthened mal mRNAs [32, 33] was highlighted in the HEK293 cells the previous observations about the effects and possible expressing Iso3Non-risk, the same isoform that was used interaction partners of CCHCR1 in the cell, as well as in the previous binding studies. In addition to mRNA provided new clues about its possible function. Tervaniemi et al. BMC Genomics (2018) 19:432 Page 12 of 15 Methods Microarray study with isoform 1-overexpressing cells Samples, RNA extraction, and RNAseq For the microarray study we used GeneChip® Human Generation of stable cell lines overexpressing the dif- Gene 1.0 ST Arrays. The total RNA samples of cells ferent CCHCR1 isoforms was described in our previ- were hybridized according to manufacturer’s protocols ous study [6]. Briefly, CCHCR1-pDsRed-constructs (Affymetrix Inc., Santa Clara, CA, USA). The arrays (Iso1Non-risk, Iso1Risk, Iso3Non-risk, or Iso3Risk) or were scanned with GeneChip scanner 3000 7G (Affyme- vector were stably transfected into HEK293 which is trix Inc.). The analyses of the microarray data were per- an easily transfectable tumorigenic cell line. Colonies formed using the statistical software R (http:// resistant to G418 antibiotics were screened for the www.R-project.org) using the Affy and Limma packages CCHCR1 expression by qPCR and fluorescence mi- [40, 41] similar to procedure earlier described [42]. In croscopy. The expression of selected clones was veri- short, we implemented the robust multiarray average fied by Western blotting [6]. The RNA extraction was method (RMA) [43] to normalize and calculate the Log2 performed at the same time as the quantification of expression values at gene level (annotated using hugen- clones without further passaging of cell lines. For the e10stv1.r3cdf_2.5.0). We did pair-wise comparison con- microarray and RNAseq experiments the total RNA trasts between the vector control and the different of the CCHCR1 cell lines and controls was extracted CCHCR1-HEK293 cell lines followed with t-statistics by aRNeasykit (Qiagen) and RNA qualitywas con- and log-odds of differential expression (B-value). The trolled by Bioanalyzer (RIN for all samples > 8.5). original microarray data compliance with MIAME pro- Total RNA samples, three replicates for each cedure has been submitted to ArrayExpress (the acces- CCHCR1-HEK293celllinewereusedfor theRNAseq sion number E-MTAB-6848). library preparation according to the previous single-cell tagged reverse transcription (STRT) proto- Quantitative real-time PCR col [38], which was adjusted for 10 ng samples as in We used qPCR to validate the expression of AREG, FN1, our previous studies [20, 24]. The libraries were se- TLN1, CXCL8, and SYT1 genes in the CCHCR1-HEK293 quenced using an Illumina HiSeq 2000 instrument. cell lines. cDNA synthesis was carried out with random Pre-processing of STRT reads, alignments and primers and SuperScript III First-Strand synthesis system per-gene quantitation were performed as previously (Invitrogen) according to manufacturer’s instructions. described [38]. The data normalization and differential ex- qPCR was carried out using an ABI PRISM 7900HT Se- pression analyses were performed using SAMstrt [39] quence Detection System with pre-designed TaqMan® (https://github.com/shka/R-SAMstrt/wiki), which is an Gene Expression Assays (Thermo Fisher Scientific) or enhancement of SAMseq [PMID: 22127579] for the spi- with self-designed primers by using Fast SYBR® Green kein-base normalization. Differentially expressed genes Master mix (Thermo Fisher Scientific) for detection, ac- were extracted by two class unpaired comparison; cording to manufacturer’s instructions. Lists of threshold of the significantly regulated genes was pre-designed TaqMan Gene Expression Assays and nu- FDR < 1%. We performed PCA with the scaling but cleotide sequences of self-designed qPCR primers are non-centering pre-process steps. A higher permuta- shown in Additional file 8. Each sample was run in three tion (n = 1000) was applied to the comparisons with replicates and the experiments were run twice. HPRT1 pooled samples (comb_Iso1, -Iso3, −Risk, and –Non-risk). was used as a reference gene for normalization and the The mock analysis with permutated sample groups was samples were compared against the wild type HEK293 performed to find possible false positive pathway results sample. Statistical comparisons were made with the Stu- and included six manual permutations per comparison dent’s t-test. group. The RNAseq of psoriatic skin samples was recently outlined elsewhere [20]. Briefly, all subjects involved in this study gave written informed consent and the study SNP genotyping followed the Declaration of Helsinki Guidelines. The We genotyped the CCHCR1 SNPs rs3130453 (CCH Institutional Review Board of the Helsinki University CR1*Iso1/3) and rs130076 (CCHCR1*WWCC/RRGS) Central Hospital had approved the study and the collec- with commercial allelic discrimination assays with tion of skin samples. The skin samples were harvested pre-designed probes and primers (TaqMan) described by a compressed air-driven dermatome (Zimmer®, previously [6] from DNA extracted from the skin sam- Warsaw, IN) to obtain an epidermal sample with min- ples [20]. Sample C.05 was not genotyped due to lack of imal dermis involvement. We collected lesional (PL) and sufficient amount of sample for DNA extraction. The non-lesional (PN) samples from six psoriatic patients HLA-Cw*06:02 genotype was also determined with com- and normal healthy skin from nine controls (C). mercial allelic discrimination assays (TaqMan) [44]. Tervaniemi et al. BMC Genomics (2018) 19:432 Page 13 of 15 Immunofluorescence The DEGs (comb_Iso1 and comb_Iso3) were analysed using the KEGG The CCHCR1-HEK293 cell lines, vector control, and pathway analysis of WebGestaltR. (XLSX 3054 kb) wild type HEK293 cells were grown on cover slips Additional file 6: Table S6. Haplotype specific gene enrichment analyses based on re-extracted DEGs of the CCHCR1-HEK293 cell lines. coated with rat tail collagen I (Gibco, Invitrogen) and The DEGs were obtained from the pooled data of Iso1Non-Risk and fixed with 4% paraformaldehyde solution. After fixation Iso3Non-Risk, and Iso1Risk and Iso3Risk compared to the controls (wild- cells were permeabilized with 0.1% Triton-×100 in PBS. type and vector). The DEGs (comb_Non-Risk, comb_Risk) were analysed using the KEGG pathway analysis of WebGestaltR. Summary of the gene The samples were incubated 1 h at room temperature enrichment results among the mock DEGs lists. (XLSX 2314 kb) with the antibodies against P-body markers EDC4 Additional file 7: Table S7 and Figure S1. CCHCR1 and HLA-Cw6 ge- (rabbit polyclonal, Cell Signaling) and DCP1A (mouse notypes of the skin samples. Figure S1. The CCHCR1 and HLA-Cw6 ge- monoclonal, Abnova), and centrosome marker γ-tubulin notypes illustrated in a PCA plot. (XLSX 79 kb) (mouse monoclonal, Sigma). Alexa Fluor 555 and 488 Additional file 8: Supplementary Information and Figure S2. Information about qPCR and co-localization of CCHCR1 with P-body conjugated IgGs (Invitrogen, Molecular Probes) were markers. Lists of pre-designed TaqMan Gene Expression Assays and used as secondary antibodies and the nuclei stained with nucleotide sequences of self-designed qPCR primers. Counting the DAPI (4′,6-diamidino-2-phenylindole, Sigma-Aldrich). colocalization of CCHCR1 with P-body markers in the CCHCR1-HEK293 cell lines and calculation of p-values for the comparison between cell lines. The pictures were taken with Zeiss LSM 5 Duo confocal Figure S2. γ-tubulin staining of the CCHCR1 cells. Antibody against microscope. Differences in localization of CCHCR1 iso- γ-tubulin was used as a marker for centrosomes. (PDF 1046 kb) forms with centrosomal P-bodies were determined by counting colocalized staining of CCHCR1 and EDC4 Abbreviations and DCP1A in each CCHCR1-overexpressing cell line *Iso1: Allele resulting in CCHCR1 isoform 1; *Iso3: Allele resulting in CCHCR1 isoform 3; *Non-risk: Non-risk haplotype (*RRGS)of CCHCR1; *Risk: Psoriasis- blindly (Additional file 8: Supplementary Information associated risk haplotype (*WWCC)of CCHCR1; DEG: Differentially expressed about qPCR and co-localization of CCHCR1 with gene; FC: Fold change; FDR: False discovery rate; Iso1 cells: CCHCR1 isoform P-body markers). 1 expressing cell lines; Iso3 cells: CCHCR1 isoform 3 expressing cell lines; PCA: Principal component analysis; STRT: Single-cell tagged reverse transcription Additional files Acknowledgements We thank Auli Saarinen for the technical support in the laboratory and Sten Linnarsson for help with RNA sequencing. Confocal imaging was done at Additional file 1: Table S1. Microarray data and gene enrichment Molecular Imaging Unit, Biomedicum Helsinki, University of Helsinki. This analysis of the CCHCR1 Iso1 cell lines. Gene expression profiling data of study was supported by Academy of Finland, Sigrid Jusélius Foundation, microarrays from the Isoform 1 CCHCR1-HEK293 cell lines. Gene enrich- Helsinki University Central Hospital Research Funds, and Folkhälsan Institute ment analysis of differentially expressed genes using the KEGG pathway of Genetics, and Swedish Research Council. The computation resources were analysis of WebGestaltR and functional cluster analysis of the DAVID. provided by SNIC through Uppsala Multidisciplinary Center for Advanced (XLSX 305 kb) Computational Science (UPPMAX) (b2010037). JK is a recipient of The Royal Additional file 2: Table S2. Differentially expressed genes of RNAseq Society Wolfson Research Excellence Award. from the comparisons between the CCHCR1-HEK293 cell lines and con- trols. Up- and downregulated DEGs of CCHCR1-overexpressing cell lines Funding Iso1Non-risk (1 N), Iso1Risk (1R), Iso3Non-risk (3 N), and Iso3Risk (3R) when This study was supported by Academy of Finland (130360; 1255560), Sigrid compared with controls (wild type and vector transfected cells). (XLSX Jusélius Foundation, Helsinki University Central Hospital Research Funds 1215 kb) (TYH2009233), and Folkhälsan Institute of Genetics, and Swedish Research Council. Additional file 3: Table S3. Gene enrichment analysis of DEGs from the CCHCR1-HEK293 cell lines based on RNAseq. Cell line-specific gene en- richment analyses of DEGs (Additional file 2: Table S2) from the CCHCR1- Availability of data and materials overexpressing cell lines Iso1Non-risk (1 N), Iso1Risk (1R), Iso3Non-risk All data generated during the study are included in this published article (3 N), and Iso3Risk (3R) when compared with wild type and vector con- and its supplementary information files. The microarray data will be available trols. The gene enrichment analysis of DEGs that were shared by the four on ArrayExpress (the accession number E-MTAB-6848). The processed STRT CCHCR1 cell lines (Intersection). Analyses were done using the KEGG reads supporting the results of this article are available in the European pathway analysis (WebGestaltR and WebGestalt), and GO and cluster ana- Nucleotide Archive (the accession number PRJEB20138). lyses from DAVID. Gene enrichment of Iso3Non-Risk DEGs in the mRNA surveillance pathway is shown (KEGG pathway figure and a list of genes). Authors’ contributions (XLSX 1183 kb) All authors (MHT, SK, TS, HAS, JV, KN, KT, SS, EK, JK, and OE) were involved in the study design. SS and JV recruited and clinically characterized the study Additional file 4: Table S4. Isoform- and haplotype-specific gene en- subjects in the psoriasis study, and sampled the skin specimen. HAS, KN, and richment with the shared DEGs of the CCHCR1-HEK293 cell lines. Gene TS prepared the samples for the experiments. KT analyzed the microarray enrichment analyses of DEGs shared by only the Non-risk (Diff N), Risk data. SK was responsible for the bioinformatics pipeline of the RNAseq data; (Diff R), isoform 1 (Diff iso1), or isoform 3 (Diff iso3) CCHCR1cell lines (see MHT and OE further analyzed the data; MHT was responsible for the in detail Fig. 4 Venn diagram). The DEGs shared by all the CCHCR1 cell immunofluorescent experiments. MHT, JK, and OE drafted the manuscript. All lines (Intersection) were analyzed as well. Analyses were done using the the authors contributed critical discussion and approved the final version of GO and cluster analyses from DAVID and KEGG pathway analysis from the manuscript. WebGestalt and WebGestaltR. (XLSX 307 kb) Additional file 5: Table S5. Isoform specific gene enrichment analyses Ethics approval and consent to participate based on re-extracted DEGs of the CCHCR1-HEK293 cell lines. The DEGs The collection of skin samples was approved by the Ethics Committee of the were obtained from the pooled data of Iso1Non-risk and Iso1Risk, and Hospital District of Helsinki and Uusimaa and by the Committee of Skin and Iso3Non-risk and Iso3Risk compared to the controls (wildtype and vector). Allergy Hospital, Helsinki University Central Hospital. All subjects involved in Tervaniemi et al. BMC Genomics (2018) 19:432 Page 14 of 15 this study gave written informed consent and the study followed the retained in the cytoplasm by its interaction with HCR, the psoriasis vulgaris Declaration of Helsinki Guidelines. candidate gene product. J Cell Sci. 2005;118(18):4253–60. 13. Andersen JS, Wilkinson CJ, Mayor T, Mortensen P, Nigg EA, Mann M. Competing interests Proteomic characterization of the human centrosome by protein correlation The authors declare that they have no competing interests. profiling. Nature. 2003;426(6966):570–4. 14. Jakobsen L, Vanselow K, Skogs M, Toyoda Y, Lundberg E, Poser I, Falkenby LG, Bennetzen M, Westendorf J, Nigg EA, et al. Novel asymmetrically Publisher’sNote localizing components of human centrosomes identified by complementary Springer Nature remains neutral with regard to jurisdictional claims in proteomics methods. EMBO J. 2011;30(8):1520–35. published maps and institutional affiliations. 15. Ling YH, Wong CC, Li KW, Chan KM, Boukamp P, Liu WK. CCHCR1 interacts with EDC4, suggesting its localization in P-bodies. Exp Cell Res. 2014;327(1): Author details 12–23. 1 2 Folkhälsan Institute of Genetics, 00014 Helsinki, Finland. Department of 16. Eulalio A, Behm-Ansmant I, Schweizer D, Izaurralde E. P-body formation is a Medical and Clinical Genetics, Medicum and Research Programs Unit, consequence, not the cause, of RNA-mediated gene silencing. Mol Cell Biol. Molecular Neurology, University of Helsinki, Helsinki, Finland. Department of 2007;27(11):3970–81. Biosciences and Nutrition, Karolinska Institutet, Huddinge, Sweden. Helsinki 17. Rajgor D, Shanahan CM. RNA granules and cytoskeletal links. Biochem Soc Burn Center, Department of Plastic Surgery, University of Helsinki and Trans. 2014;42(4):1206–10. Helsinki University Hospital, Helsinki, Finland. Department of Pharmacology, 18. Aizer A, Brody Y, Ler LW, Sonenberg N, Singer RH, Shav-Tal Y. The dynamics Medicum, University of Helsinki, Helsinki, Finland. Division of of mammalian P body transport, assembly, and disassembly in vivo. Mol Neuropsychiatry, Department of Women’s and Children’s Health, Center of Biol Cell. 2008;19(10):4154–66. Neurodevelopmental Disorders, Karolinska Institutet, Stockholm, Sweden. 19. Moser JJ, Fritzler MJ, Rattner JB. Repression of GW/P body components and Department of Dermatology, University of Helsinki and Helsinki University the RNAi microprocessor impacts primary ciliogenesis in human astrocytes. Hospital, Helsinki, Finland. School of Basic and Medical Biosciences, King’s BMC Cell Biol. 2011;12:37. College London, London, UK. 20. Tervaniemi MH, Katayama S, Skoog T, Siitonen HA, Vuola J, Nuutila K, Sormunen R, Johnsson A, Linnarsson S, Suomela S, et al. NOD-like receptor Received: 19 May 2017 Accepted: 21 May 2018 signaling and inflammasome-related pathways are highlighted in psoriatic epidermis. Sci Rep. 2016;6:22745. 21. Huang da W, Sherman BT, Lempicki RA. Systematic and integrative References analysis of large gene lists using DAVID bioinformatics resources. Nat 1. Harden JL, Krueger JG, Bowcock AM. The immunogenetics of psoriasis: a Protoc. 2009;4(1):44–57. comprehensive review. J Autoimmun. 2015;64:66–73. 22. Huang da W, Sherman BT, Lempicki RA. Bioinformatics enrichment tools: 2. Asumalahti K, Laitinen T, Itkonen-Vatjus R, Lokki ML, Suomela S, Snellman E, paths toward the comprehensive functional analysis of large gene lists. Saarialho-Kere U, Kere J. A candidate gene for psoriasis near HLA-C, HCR Nucleic Acids Res. 2009;37(1):1–13. (Pg8), is highly polymorphic with a disease-associated susceptibility allele. 23. Wang J, Vasaikar S, Shi Z, Greer M, Zhang B. WebGestalt 2017: a more Hum Mol Genet. 2000;9(10):1533–42. comprehensive, powerful, flexible and interactive gene set enrichment 3. Asumalahti K, Veal C, Laitinen T, Suomela S, Allen M, Elomaa O, Moser M, de analysis toolkit. Nucleic Acids Res. 2017;45(Web Server Issue):W130–7. Cid R, Ripatti S, Vorechovsky I, et al. Coding haplotype analysis supports 24. Katayama S, Skoog T, Jouhilahti EM, Siitonen HA, Nuutila K, Tervaniemi MH, HCR as the putative susceptibility gene for psoriasis at the MHC PSORS1 Vuola J, Johnsson A, Lonnerberg P, Linnarsson S, et al. Gene expression locus. Hum Mol Genet. 2002;11(5):589–97. analysis of skin grafts and cultured keratinocytes using synthetic RNA 4. Zhang XJ, Huang W, Yang S, Sun LD, Zhang FY, Zhu QX, Zhang FR, Zhang normalization reveals insights into differentiation and growth control. BMC C, Du WH, Pu XM, et al. Psoriasis genome-wide association study identifies Genomics. 2015;16:476. susceptibility variants within LCE gene cluster at 1q21. Nat Genet. 2009; 25. Yamashita A, Izumi N, Kashima I, Ohnishi T, Saari B, Katsuhata Y, Muramatsu 41(2):205–10. R, Morita T, Iwamatsu A, Hachiya T, et al. SMG-8 and SMG-9, two novel 5. Chang YT, Chou CT, Shiao YM, Lin MW, Yu CW, Chen CC, Huang CH, Lee DD, subunits of the SMG-1 complex, regulate remodeling of the mRNA Liu HN, Wang WJ, et al. Psoriasis vulgaris in Chinese individuals is associated surveillance complex during nonsense-mediated mRNA decay. Genes Dev. with PSORS1C3 and CDSN genes. Br J Dermatol. 2006;155(4):663–9. 2009;23(9):1091–105. 6. Tervaniemi MH, Siitonen HA, Soderhall C, Minhas G, Vuola J, Tiala I, 26. Muller M, Demeret C. CCHCR1 interacts specifically with the E2 protein of Sormunen R, Samuelsson L, Suomela S, Kere J, et al. Centrosomal human papillomavirus type 16 on a surface overlapping BRD4 binding. localization of the psoriasis candidate gene product, CCHCR1, supports a PLoS One. 2014;9(3):e92581. role in cytoskeletal organization. PLoS One. 2012;7(11):e49920. 27. Jiang G, Giannone G, Critchley DR, Fukumoto E, Sheetz MP. Two-piconewton 7. Suomela S, Elomaa O, Asumalahti K, Kariniemi AL, Karvonen SL, Peltonen J, slip bond between fibronectin and the cytoskeleton depends on Talin. Nature. Kere J, Saarialho-Kere U. HCR, a candidate gene for psoriasis, is expressed 2003;424(6946):334–7. differently in psoriasis and other hyperproliferative skin disorders and is 28. Ling K, Doughman RL, Firestone AJ, Bunce MW, Anderson RA. Type I downregulated by interferon-gamma in keratinocytes. J Invest Dermatol. gamma phosphatidylinositol phosphate kinase targets and regulates focal 2003;121(6):1360–4. adhesions. Nature. 2002;420(6911):89–93. 8. Tiala I, Suomela S, Huuhtanen J, Wakkinen J, Holtta-Vuori M, Kainu K, Ranta 29. Muro AF, Chauhan AK, Gajovic S, Iaconcig A, Porro F, Stanta G, Baralle FE. S, Turpeinen U, Hamalainen E, Jiao H, et al. The CCHCR1 (HCR) gene is Regulated splicing of the fibronectin EDA exon is essential for proper skin relevant for skin steroidogenesis and downregulated in cultured psoriatic wound healing and normal lifespan. J Cell Biol. 2003;162(1):149–60. keratinocytes. J Mol Med. 2007;85(6):589–601. 30. Honda A, Yamada M, Saisu H, Takahashi H, Mori KJ, Abe T. Direct, Ca2+ 9. Elomaa O, Majuri I, Suomela S, Asumalahti K, Jiao H, Mirzaei Z, Rozell B, −dependent interaction between tubulin and synaptotagmin I: a possible Dahlman-Wright K, Pispa J, Kere J, et al. Transgenic mouse models mechanism for attaching synaptic vesicles to microtubules. J Biol Chem. support HCR as an effector gene in the PSORS1 locus. Hum Mol Genet. 2002;277(23):20234–42. 2004;13(15):1551–61. 31. Zhu XL, Qi ST, Liu J, Chen L, Zhang C, Yang SW, Ouyang YC, Hou Y, Schatten H, 10. Tiala I, Wakkinen J, Suomela S, Puolakkainen P, Tammi R, Forsberg S, Song YL, et al. Synaptotagmin1 is required for spindle stability and metaphase- Rollman O, Kainu K, Rozell B, Kere J, et al. The PSORS1 locus gene CCHCR1 to-anaphase transition in mouse oocytes. Cell Cycle. 2012;11(4):818–26. affects keratinocyte proliferation in transgenic mice. Hum Mol Genet. 2008; 32. Chang YF, Imam JS, Wilkinson MF. The nonsense-mediated decay RNA 17(7):1043–51. surveillance pathway. Annu Rev Biochem. 2007;76:51–74. 11. Sugawara T, Shimizu H, Hoshi N, Nakajima A, Fujimoto S. Steroidogenic acute regulatory protein-binding protein cloned by a yeast two-hybrid 33. Kurosaki T, Maquat LE. Nonsense-mediated mRNA decay in humans at a system. J Biol Chem. 2003;278(43):42487–94. glance. J Cell Sci. 2016;129(3):461–7. 12. Corbi N, Bruno T, De Angelis R, Di Padova M, Libri V, Di Certo MG, Spinardi 34. Savas JN, Makusky A, Ottosen S, Baillat D, Then F, Krainc D, Shiekhattar R, L, Floridi A, Fanciulli M, Passananti C. RNA polymerase II subunit 3 is Markey SP, Tanese N. Huntington's disease protein contributes to RNA- Tervaniemi et al. BMC Genomics (2018) 19:432 Page 15 of 15 mediated gene silencing through association with Argonaute and P bodies. Proc Natl Acad Sci U S A. 2008;105(31):10820–5. 35. Anderson P, Kedersha N, Ivanov P. Stress granules, P-bodies and cancer. Biochim Biophys Acta. 2015;1849(7):861–70. 36. Dougherty JD, Reineke LC, Lloyd RE. mRNA decapping enzyme 1a (Dcp1a)- induced translational arrest through protein kinase R (PKR) activation requires the N-terminal enabled vasodilator-stimulated protein homology 1 (EVH1) domain. J Biol Chem. 2014;289(7):3936–49. 37. Lloyd RE. Regulation of stress granules and P-bodies during RNA virus infection. Wiley Interdiscip Rev RNA. 2013;4(3):317–31. 38. Islam S, Kjallquist U, Moliner A, Zajac P, Fan JB, Lonnerberg P, Linnarsson S. Highly multiplexed and strand-specific single-cell RNA 5′ end sequencing. Nat Protoc. 2012;7(5):813–28. 39. Katayama S, Töhönen V, Linnarsson S, Kere J. SAMstrt: statistical test for differential expression in single-cell transcriptome with spike-in normalization. Bioinformatics. 2013;29(22):2943–5. 40. Gautier L, Cope L, Bolstad BM, Irizarry RA. Affy–analysis of Affymetrix GeneChip data at the probe level. Bioinformatics. 2004;20(3):307–15. 41. Smyth GK. Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol. 2004;3:Article3. 42. Massinen S, Hokkanen ME, Matsson H, Tammimies K, Tapia-Paez I, Dahlstrom-Heuser V, Kuja-Panula J, Burghoorn J, Jeppsson KE, Swoboda P, et al. Increased expression of the dyslexia candidate gene DCDC2 affects length and signaling of primary cilia in neurons. PLoS One. 2011;6(6): e20580. 43. Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, Speed TP. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics. 2003;4(2):249–64. 44. Nikamo P, Stahle M. Cost-effective HLA-Cw06:02 typing in a Caucasian population. Exp Dermatol. 2012;21(3):221–3. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png BMC Genomics Springer Journals

Intracellular signalling pathways and cytoskeletal functions converge on the psoriasis candidate gene CCHCR1 expressed at P-bodies and centrosomes

Free
15 pages
Loading next page...
 
/lp/springer_journal/intracellular-signalling-pathways-and-cytoskeletal-functions-converge-Pm5KgyySyb
Publisher
BioMed Central
Copyright
Copyright © 2018 by The Author(s).
Subject
Life Sciences; Life Sciences, general; Microarrays; Proteomics; Animal Genetics and Genomics; Microbial Genetics and Genomics; Plant Genetics and Genomics
eISSN
1471-2164
D.O.I.
10.1186/s12864-018-4810-y
Publisher site
See Article on Publisher Site

Abstract

Background: CCHCR1 (Coiled-Coil α-Helical Rod protein 1) is a putative psoriasis candidate gene with the risk alleles CCHCR1*WWCC and *Iso3, the latter inhibiting the translation of isoform 1. CCHCR1 was recently shown to be a centrosomal protein, as well as a component of cytoplasmic processing bodies (P-bodies) that regulate mRNA turnover. The function of CCHCR1 has remained unsettled, partly because of the inconsistent findings; it has been shown to play a wide variety of roles in divergent processes, e.g., cell proliferation and steroidogenesis. Here we utilized RNA sequencing (RNAseq) using HEK293 cells overexpressing isoforms 1 or 3 (Iso1, Iso3 cells), in combination with the coding non-risk or risk (*WWCC) haplotype of CCHCR1. Our aim was to study the overall role of CCHCR1 and the effects of its variants. Results: The overexpression of CCHCR1 variants in HEK293 cells resulted in cell line-specific expression profiles though several similarities were observable. Overall the Iso1 and Iso3 cells showed a clear isoform-specific clustering as two separate groups, and the Non-risk and Risk cells often exhibited opposite effects. The RNAseq supported a role for CCHCR1 in the centrosomes and P-bodies; the most highlighted pathways included regulation of cytoskeleton, adherens and tight junctions, mRNA surveillance and RNA transport. Interestingly, both the RNAseq and immunofluorescent localization revealed variant-specific differences for CCHCR1 within the P-bodies. Conclusions: CCHCR1 influenced a wide variety of signaling pathways, which could reflect its active role in the P-bodies and centrosomes that both are linked to the cytoskeleton; as a centrosomal P-body protein CCHCR1 may regulate diverse cytoskeleton-mediated functions, such as cell adhesion and -division. The present findings may explain the previous inconsistent observations about the functions of CCHCR1. Keywords: CCHCR1, PSORS1, Centrosome, Cytoskeleton, P-body, Cell adhesion, RNAseq, Expression profiling * Correspondence: juha.kere@ki.se; outi.elomaa@helsinki.fi Folkhälsan Institute of Genetics, 00014 Helsinki, Finland Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Tervaniemi et al. BMC Genomics (2018) 19:432 Page 2 of 15 Background repression, and RNA-mediated gene silencing [16, 17]. Psoriasisisachronicmultifactorial dermatological disorder, There is a link between the P-bodies and centrosomes; characterized by abnormal proliferation and differentiation P-bodies transit along microtubule tracks to and from the of keratinocytes and infiltration of inflammatory cells. centrosome and furthermore, a pair of stationary P-bodies There are several susceptibility gene loci for psoriasis reside at the centrosome during the interphase [18, 19]. (PSORS1–15) (based on the Online Mendelian Inheritance Using stable HEK293 cells overexpressing isoforms in Man, OMIM); of which PSORS1 (6p21.3) has the stron- 1or3 with thenon-risk(*RRGS) or the risk haplo- gest risk effect [1]. Diverse psoriasis-associated alleles have type (*WWCC) we have shown that isoform 3 acceler- been identified within the region. However, a strong linkage ates cell proliferation and together with the *WWCC disequilibrium has made it difficult to distinguish their indi- allele apoptosis as well. Whereas isoform 1 lacks sig- vidual effects. Hence, the effector genes in psoriasis within nificant effects on cell proliferation or cell cycle pro- the 6p21.3 region are currently not fully understood. gression. Furthermore, the CCHCR1-HEK293 cell CCHCR1 (Coiled-Coil α-Helical Rod protein 1) is a putative lines show isoform- and haplotype-specific changes in PSORS1 candidate gene among others [2–4], and its cell size and shape and have alterations in the CCHCR1*WWCC allele is associated with psoriasis in sev- organization and expression of the cytoskeletal pro- eral populations [2, 3, 5]. WWCC stands for the amino teins actin, vimentin, and cytokeratins. We also dem- acids in the psoriasis risk haplotype, whereas in the onstrated that CCHCR1 may regulate EGF-induced non-risk haplotype the corresponding amino acids are STAT3 activation in an isoform-specific manner [6]. RRGS. We have previously described a novel form of Here we applied 5’end-targeted RNA sequencing (RNA- CCHCR1, isoform 1, where the N-terminal domain is lon- seq) on the previously established and characterized gerthaninisoform 3 [6]. The formation of isoform 1 is CCHCR1-HEK293 cell lines (Iso1Risk, Iso1Non-risk, dependent on a SNP (rs3130453) that results in either a Iso3Risk, and Iso3Non-risk) [6, 20]to examine isoform- longer open reading frame (allele *Iso1)or a stop codon and haplotype-specific effects on global gene expression (allele *Iso3), enabling only the translation of isoform 3. In profiles, and to identify the underlying mechanisms lead- addition, there are two transcription start sites, only one of ing to the previously observed differences between the them enabling the production of long isoform 1. Notably, CCHCR1 cell lines. The sensitivity of RNAseq allowed an we demonstrated that *Iso3 shows association with psoriasis in-depth assessment of altered gene expression in the cell − 7 (P <10 )[6]. lines, making it possible to get a better understanding of The CCHCR1 protein does not belong to any known the effects of CCHCR1 isoforms on the cellular pathways. protein family but is predicted to be a rod-like protein, with an alpha-helical coiled coil structure. Although its Results expression pattern in psoriatic skin differs from healthy CCHCR1-HEK293 cell lines exhibited isoform- and skin [7, 8] and its overexpression influences cell prolifer- haplotype-specific gene expression profiles ation [6, 9, 10], its role as a putative psoriasis effector In the present study we used stably transfected gene has remained unsettled, partially because it has CCHCR1 cell lines that were previously generated been suggested to function in several divergent bio- into human embryonic kidney (HEK) 293 cells. HEK293 is logical processes such as the regulation of steroidogene- an adenovirus-transformed cell line that is considered sis via mitochondrial steroidogenic activator protein tumorigenic. The establishment and characterization of (STAR) [8, 11] or muscle differentiation via RNA poly- the CCHCR1 overexpressing HEK293 clones have been merase II subunit 3 (RPB3) [12]. We have also demon- described previously [6]. Briefly, the CCHCR1 cell lines strated that in addition to its cytoplasmic location, both (Iso1Risk, Iso1Non-risk, Iso3Risk, and Iso3Non-risk) were CCHCR1 isoforms localize at the centrosome [6], a cell screened for the CCHCR1 expression by qPCR, fluores- organelle playing a crucial role in cell division. The cence microscopy, and western blotting. The clones used centrosomal localization of CCHCR1 has been verified in the present work were chosen based on the previous in other large-scale proteomics studies [13, 14]. studies; we selected clones with similar CCHCR1 expres- Furthermore, CCHCR1 has been shown to interact with sion levels and cell morphology. As most of the previous mitotic spindle proteins [15], providing evidence for its CCHCR1 studies have focused on isoform 3, we first possible involvement in cell division. A recent study performed gene expression profiling using Affymetrix demonstrated that, in addition to its location at the Human Gene ST 1.0 microarrays for the Iso1Non-risk centrosome, CCHCR1 is a component of processing and Iso1Risk CCHCR1 cell lines to increase knowledge of bodies (P-bodies) [15]. P-bodies are cytoplasmic ribonu- isoform 1. Wild type HEK293 and vector-transfected cells cleoprotein granules that regulate mRNA turnover in a were used as controls in the comparative expression post-transcriptional manner and are involved in mRNA analysis. The comparison of Iso1Non-risk and Iso1- degradation, surveillance, and transport, translational Risk with the vector control revealed 296 and 206 Tervaniemi et al. BMC Genomics (2018) 19:432 Page 3 of 15 differentially expressed genes (DEGs) (Cut-off fold lowest in Iso3Non-risk (2806) (up: FC > 1.5, down: change1.3-fold, p < 0.002), respectively (Additional file 1: FC < 0.75, FDR 0.25). All the CCHCR1 cell lines Table S1). Gene enrichment analyses (the Database shared 209 upregulated and 618 downregulated genes for Annotation, Visualization and Integrated Discovery (Venndiagram of DEGs showninFig. 1a). (DAVID) [21, 22], WebGestalt 2013, and WebGestaltR There were notable differences in the expression pro- 2017 [23]) of the DEGs (> 1.5-fold) from the files between the CCHCR1-HEK293 cell lines. We esti- Iso1Non-risk cells revealed such pathways and func- mated the dissimilarity between the cell lines by − 4 tions (p <1 ×10 )as focal adhesion, ECM-receptor principal component analysis (PCA) (Fig. 1b) that en- interaction,and regulation of actin cytoskeleton.The abled multiclass comparisons between samples [24]. The results of the gene enrichment analysis are shown in second PC (PC2) depicts the isoform-dependent differ- detail in Additional file 1:Table S1.Thoughthe ence, clustering isoform 1 (Iso1) and 3 (Iso3) as separate microarray study suggested differences in the gene ex- groups. The wild type and vector control samples clus- pression between Iso1Non-risk and Iso1Risk, the cell tered as an overlapping group between the Iso1 and Iso3 lines sharing only 50 genes, the sensitivity of the ex- samples. The PCA demonstrated that the replicates dif- pression profiling remained rather modest. Particularly fered to some extent, especially the Iso3Risk samples in the Iso1Risk cell line only a few pathways, such as (the outlier specimen 3Rc, Fig. 1b). The clustering of cell cycle, focal adhesion,and leukocyte migration Risk and Non-risk samples as separate groups was less showed enrichment of more than four genes when clear than observed with Iso1 and Iso3. WebGestalt was used. To get a higher sensitivity as- sessment of altered gene expression in all four Expression profiling of CCHCR1-HEK293 cell lines CCHCR1-HEK293 cell lines we performed RNAseq. highlighted cell adhesion and regulation of cytoskeleton RNA samples extracted from the wild type HEK293, Using the WebGestalt tool (WebGestalt 2013 and Web- the vector control, and the CCHCR1 cell lines (Iso1- GestaltR) for gene enrichment analysis, we analyzed the Non-risk, Iso1Risk, Iso3Non-risk, and Iso3Risk), were DEGs of four different comparisons (DEGs shown in subjected to RNAseq. The comparisons of transcripts Additional file 2: Table S2: Iso1Non-risk, Iso1Risk, from each CCHCR1-HEK293 cell line with the con- Iso3Non-risk, and Iso3Risk cells compared with wild trols (wild type and vector) revealed a large number type and vector controls). The KEGG pathway analysis of differentially expressed genes (DEGs) that are listed of up- and downregulated genes revealed several similarities in Additional file 2: Table S2. The number of upregu- between the CCHCR1-HEK293 cell lines (Additional file 3: lated genes was highest in Iso3Non-risk (6602) and Table S3 showing the results of gene enrichment analysis). lowest in Iso3Risk (1781), while the number of down- Adhesion-related pathways, such as focal adhesion, gap regulated genes was highest in Iso3Risk (6968) and junction, tight junction, adherens junction,and regulation of Fig. 1 Expression profiles of the CCHCR1-HEK293 cell lines. a Venn diagrams of differentially expressed genes (DEGs) of the CCHCR1-HEK293 cell lines; Iso1Non-risk (1 N), Iso1Risk (1R), Iso3Non-risk (3 N), and Iso3Risk (3R), when compared with vector and wild type controls (FC > 1.5 or FC < 0.75, FDR < 0.25). The most highlighted functions and pathways of the gene enrichment analysis of DEGs that were shared by the all four cell lines are shown. Up- and downregulated genes were analyzed separately (see in detail Additional file 3: Table S3). b Principal component analysis (PCA) demonstrates differences between CCHCR1-HEK293 cell lines. The two PCs depict the variation between samples. The PC2 illustrates the isoform-dependent difference, clustering isoforms 1 and 3 as separate groups. The wild type (wt) and vector (v) control samples cluster as an overlapping group between the isoform 1 and isoform 3 samples Tervaniemi et al. BMC Genomics (2018) 19:432 Page 4 of 15 actin cytoskeleton,as wellas Wnt signaling and calcium the gene expression were apparent, which is observable signaling, were highlighted in all the cell lines. In addition, from the Venn diagram (Fig. 1a) and the heatmaps. For MAPK, ErbB,and TGF-beta were among the top instance Wnt signaling (heatmap in Fig. 3a) was mainly pathways. Their expression profiles, however, showed vari- upregulated only in the Iso3Non-risk cells when com- ation between the control samples. The most highly up- pared with controls (Additional file 3: Table S3). While 8 − 7 and downregulated genes (1 × 10 <FC<1 × 10 )were several pathways were mainly downregulated in the enriched especially in functions and pathways that are Iso3Risk cells, these including regulation of actin cyto- involved in neuroactive ligand–receptor interaction, cell skeleton (heatmap Fig. 2a) and RIG-I-like receptor signal- adhesion, calcium signaling,and cytokine-cytokine receptor ing (heatmap in Fig. 3b; Additional file 3: Table S3). As interaction (Additional file 3: Table S3). CXCL8 (alias IL8) is involved in several signaling path- Although the same pathways were showing enrichment ways that were highlighted in our expression data, as in all the CCHCR1-HEK293 cell lines, their gene expres- well as related to psoriasis, we validated its expression in sion profiles differed; different genes were highlighted, and the CCHCR1-HEK293 cell lines by qPCR; IL8 showed the gene expression varied from down- to upregulation reduced expression in the Iso3 cell lines and upregula- between the cell lines. For example, focal adhesion and tion in the Iso1cells, especially in Iso1Non-risk (Fig. 3c), regulation of actin cytoskeleton were the top pathways in as observed in RNAseq. all the lines (Additional file 3: Table S3). However, the cor- To study the isoform and haplotype-specific effects of responding heatmaps showed differences between the cell CCHCR1 more specifically, and to get more statistical lines; the Non-risk and Risk cells often exhibited opposite power for the study, we analyzed the DEGs in four dif- effects (Fig. 2a showing heatmaps of focal adhesion and ferent groups; DEGs shared by the Iso1, Iso3, Non-risk, regulation of actin cytoskeleton). Several relevant genes of or Risk cell lines (Fig. 4: the Venn diagram of compari- focal adhesion and actin cytoskeleton were downregulated sons). The DAVID and WebGestalt (versions 2013 and particularly in the Iso3Risk cells. To verify this, we quanti- 2017) analyses revealed differences between the groups fied the expression of TLN1 (talin 1) and FN1 (fibronectin (Additional file 4: Table S4 showing gene enrich analysis 1) by qPCR (shown in Fig. 2b). They were highly downreg- in detail), though the number of gene counts in ulated in both isoform 3 expressing cell lines (Iso3), espe- highlighted pathways remained modest. Tight junction cially in Iso3Risk, whereas their expression was increased was highlighted in the shared DEGs of the Iso1 cells in the Iso1Non-risk cells similarly as was observed in (FDR = 0.06), progesterone mediated oocyte maturation RNAseq. We also examined the expression of SYT1 (syn- in the Iso3 cells (FDR < 0.05), mRNA surveillance aptotagmin 1) by qPCR (Fig. 2b); one of the most highly pathway in the Non-risk cells (FDR = 0.02), and antigen upregulated genes in the Iso3 cells. The qPCR verified that processing and presentation in the Risk cell lines (FDR < its expression was increased only in the Iso3 cells and was 0.05). The gene enrichment analysis of DEGs including unaltered in both isoform 1 expressing cells (Iso1). both the up-and downregulated genes that were shared Overall, the qPCR verification results agreed with the by all the CCHCR1-HEK293 cell lines (Fig. 4 Venn dia- RNAseq data. We have previously shown that our RNA- gram) highlighted only a few pathways with FDR < 0.1 seq protocol enables the accurate quantitation of gene ex- when analysed by WebGestaltR, these including calcium- pression [24]. and estrogen signaling (FDR = 0.05). WebGestalt highlighted To study the overall biological role of CCHCR1, we cytokine-cytokine receptor interaction, NOD-like receptor analyzed the DEGs that were shared by the four signaling,and adherens junction as well. CCHCR1-HEK293 cell lines (827 genes, 0.75 > FC > 1.5, As the Iso1 and Iso3 cell lines showed a clear FDR < 0.25). The analyses of shared DEGs (Additional isoform-specific clustering in the PCA (Fig. 1b), we pooled file 3: Table S3: Intersection up and down) identified the Iso1Non-risk and Risk samples and compared with such pathways and genes as: calcium signaling (PLCD1, the pooled vector and wildtype data (comb_Iso1), and PLCB2), NOD-like receptor signaling (NLRP3, CARD9, similarly we pooled and compared Iso3Non-risk and Risk and CXCL1), cytokine-cytokine receptor interaction vs vector and wildtype (comb_Iso3) to increase the num- (IL12A, IL22RA1, and VEGFC), cell adhesion (ITGAL, ber of replicates and to get more statistical power for the CNTN1, and CLDN7), and regulation of actin cytoskeleton study. The DEGs were re-extracted by a two class un- (PAK3, RRAS,and ITGAX), however, the gene counts of paired comparison using a higher permutation (n = 1000). the detected pathways were modest. (Additional file 5: Table S5 showing the DEGs and the re- sults of gene enrichment analysis for comb_Iso1 and com- Isoform- and haplotype-specific effects of CCHCR1 on cell b_iso3). We also pooled the data from Iso1Non-risk and signaling Iso3Non-risk (comb_Non-risk), and from Iso1Risk and Despite similarities between the CCHCR1-HEK293 cell Iso3Risk (comb_Risk) (Additional file 6: Table S6 showing lines, some isoform- and haplotype-specific effects on the DEGs and the results of gene enrichment analysis for Tervaniemi et al. BMC Genomics (2018) 19:432 Page 5 of 15 Fig. 2 Expression profiling of the CCHCR1-HEK293 cell lines highlighted cell adhesion and actin cytoskeleton. a Heatmaps showing gene expression of focal adhesion and regulation of actin cytoskeleton in the CCHCR1 Iso1Non-risk (1 N), Iso1Risk (1R), Iso3Non-risk (3 N), and Iso3Risk (3R) cell lines, and in the vector transfected cells (V), and wild type HEK293 (WT). Color key: red represents upregulated and blue downregulated expression (row Z-score). b The expression of FN1, TLN1,and SYT1 in the CCHCR1-HEK293 cell lines was validated with qPCR, and compared with the gene expression in the wild type cells. Error bars: standard deviation. Here shown normalization with HPRT1 Tervaniemi et al. BMC Genomics (2018) 19:432 Page 6 of 15 Fig. 3 Gene expression profiles of Wnt and RIG-I-like signaling in the CCHCR1-HEK293 cell lines. Both pathways were highlighted in the CCHCR1 cell lines. Heatmaps of Wnt (a) and RIG-I-like (RLR) (b) signaling in the Iso1Non-risk (IN), Iso1Risk (1R), Iso3Non-risk (3 N), and Iso3Risk (3R) CCHCR1 cell lines, vector transfected cells (v), and wild type HEK293 (wt) demonstrated that there were differences between cell lines. Wnt signaling is mainly upregulated only in the Iso3Non-risk cells while RLR signaling shows more downregulation in Iso3Risk than in other cell lines. Color key: red represents upregulated and blue downregulated expression (row Z-score). c The expression of CXCL8 in the CCHCR1-HEK293 cell lines was validated with qPCR, and compared with the gene expression in the wild type cells. Here shown normalization with HPRT1. Error bars: standard deviation Tervaniemi et al. BMC Genomics (2018) 19:432 Page 7 of 15 Fig. 4 Isoform- and haplotype-specific differences between the CCHCR1-HEK293 cell lines. The Venn diagram of the DEGs (FC > 1.5 or < 0.75, FDR < 0.25) from the CCHCR1-HEK293 cell lines Iso1Non-risk (1 N), Iso1Risk (1R), Iso3Non-risk (3 N), and Iso3Risk (3R) when compared with the control cells, and the DEGs from the psoriatic skin samples (Ps) [20]. The most highlighted gene enrichment pathways of DEGs (both the up- and down- regulated) that were shared by the Non-risk, Risk, Iso1, Iso3, or all the CCHCR1 cell lines are shown. The DEGs used for the gene enrichment analysis are shown in parentheses and indicated by color coded rectangles in the diagram. The DEGs (diff) and the results are shown in detail in the Additional file 4: Supplement Table S4 comb_Non-risk and comb_Risk), although the haplotypes presentation (FDR = 0.02). The latter and HLA-genes were showed less clustering than the isoforms. To control the highlighted also in the pathways analysis of comb_Risk specificity of the results we also generated mock gene lists which otherwise showed much less overrepresented path- by permuting the sample groupings in the comparison ways than the other groups. (Additional file 6: Table S6 showing the summary of gene enrichment analyses of the mock gene lists). Several path- Gene expression in the CCHCR1-HEK293 cell lines and ways were highlighted among the most mock gene lists. psoriatic skin There were, however, pathways that were highlighted spe- The comparison of transcriptomes from two completely dif- cifically among the CCHCR1-HEK293 clones; adherens ferent biological systems, such as the CCHCR1-HEK293 cell junction and tight junction in comb_Iso1, and RNA trans- lines and psoriatic skin, is challenging and the direct com- port and mRNA surveillance in comb_Iso3 (FDR < 0.002) parison of gene expression changes is unfeasible. Some simi- and comb_Non-Risk (FDR = 0.01) (Additional file 5:Table larities were observable when the current expression results S5; Additional file 6: Table S6), which agreed the results of CCHCR1 cells were compared to our previous RNAseq obtained with the cell line specific DEGs (Additional file data on psoriatic skin obtained from the comparison be- 3: Table S3). RIG-I-like receptor signaling, which was tween psoriatic lesional and non-sessional skin, and healthy highlighted in the pathway analyses of individual CCHCR1 control skin [20]. The comparison revealed some similarities clones (Additional file 3: Table S3), rose up in comb_Iso1 in the gene expression changes, such as upregulation of S100 (FDR = 0.06) as well. Progesterone mediated oocyte mat- and KRT (KRT6, KRT16,and KRT17) genes, and enrichment uration was highlighted in comb_Iso3 (FDR = 0.01) as was of the steroid-related pathways. Some genes of RIG-I-like re- also detected in the analysis of the shared Iso3 DEGs ceptor signaling (DDX58, CXCL8, and CASP8)showedal- (Additional file 4: Table S4; Additional file 5: Table S5). tered expression in the CCHCR1-HEK293 cells and psoriatic Mitochondria that play a role in the sythesisof steroids skin, as well. The results, however, should not be over inter- was highlighted among the comb_Iso3 DEGs as well. Sev- preted, as psoriasis is a complex disease and CCHCR1 is a eral HLA genes were overrepresented especially in com- susceptibility gene among several others, and as mentioned b_Iso3, highlighting pathways such as systemic lupus the direct comparison of CCHCR1 cells and skin samples is erythematosus (FDR = 0.0002) and antigen processing and not warranted. Tervaniemi et al. BMC Genomics (2018) 19:432 Page 8 of 15 We also examined the CCHCR1 genotype (*Iso1 or altered expression in Iso3Non-risk (the KEGG pathway of *Iso3, *Non-risk or *Risk allele) and the most well-known mRNA surveillance is shown in detail in Additional file 3: psoriasis haplotype HLA-Cw*06:02 in our psoriatic and Table S3: mRNA surveillance). healthy skin samples. The genotyping results are shown Here, we studied by immunofluorescence staining with in detail in Additional file 7: Table S7. We genotyped the the P-body markers EDC4 and DCP1A (Fig. 5b, c)[15] SNPs for rs3130453 (G/A) resulting *Iso1 or *Iso3 allele and the centrosomal marker γ-tubulin (Additional file 8: and rs130076 (C/T) corresponding to the second R/W Supplementary information Figure S2) whether the dif- amino acid change in CCHCR1*WWCC risk allele. Five ferent CCHCR1 isoforms localize with the P-bodies in out of seven psoriatic samples were homozygous for the the same manner. We observed differences in the stain- *Iso3 allele, whereas only two out of eight were homozy- ing pattern between the isoforms; in the Iso1 cells, espe- gous in the controls. None of the psoriatic samples had cially in Iso1Non-risk, the P-body marker colocalized the homozygous *Iso1 genotype. Only two of the rarely with CCHCR1, whereas in the Iso3 cells the psoriatic patients were heterozygous and therefore P-body staining colocalized with the CCHCR1 granules − 19 able to translate the longer isoform 1 as well. Five (P <4×10 ) (Additional file 8: Supplementary infor- out of seven psoriatic samples were heterozygous for mation describes the counting of colocalization of the Risk (*WWCC) haplotype whereas five out of CCHCR1 and P-body markers). In addition, we demon- eight controls had the homozygous Non-risk haplo- strated by γ-tubulin staining that in the Iso1 cells type. Most of the control samples were negative for P-body markers showed less accumulation at the centro- the psoriasis haplotype HLA-Cw*06:02 (six out of somes (Additional file 8: Figure S2). eight) but most of the psoriasis samples were hetero- zygous positive (five out of seven). The genotyping of psoriatic samples agrees the previous results [20], but Discussion the number of samples didn’t allow studying the cor- Our results show the power of RNAseq providing a large relation of CCHCR1 genotype and gene expression amount of detailed expression data on the HEK293 cells profiles. The systematic identification of all potentially expressing the CCHCR1 variants. The principal compo- causal variants in psoriasis in order to study the ef- nent and gene enrichment analyses demonstrated that fects of individual variants or their combinations the cell lines expressing different isoforms and haplo- would be an extremely laborious task to implement types of CCHCR1 have their distinct, specific expression in practice. The CCHCR1 and HLA-Cw6 genotypes profiles. The approach also strengthened the previous were illustrated in a PCA plot that clusters the sam- findings of the putative roles of CCHCR1 in the ple types (Additional file 7:FigureS2). regulation of cytoskeleton (Fig. 6) and steroid biosyn- thesis [6, 8]. Additionally, the RNAseq revealed that CCHCR1 affects P-body-related functions and shows CCHCR1 affected the cell signaling pathways mediated isoform-specific localization with P-bodies via the P-bodies, which provided evidence of its possible The recent study [15] that localized CCHCR1 at active role at the P-bodies. Furthermore, we observed P-bodies was carried out with the Iso3Non-risk form of isoform- and haplotype-specific differences in its effects CCHCR1. Interestingly in our RNAseq data, the on the P-body-related functions, as well as on the P-body-related pathways RNA transport, RNA degradation, localization with the P-bodies. The P-body localization and mRNA surveillance were highlighted in the Iso3 cells may also explain some of the putative interactions of − 15 (WebGestalt p <1 ×10 ) and especially in the Non-risk CCHCR1 that have been previously identified with such cells (FDR < 0.1) (Additional file 5: Table S5; Additional proteins as StAR, RPB3, and viral proteins (Fig. 6)[11, file 6: Table S6). Also the comparison Iso3Non-risk cells 12, 26]. We also compared the gene expression results versus controls revealed the enrichment of mRNA surveil- of the CCHCR1-HEK293 cell lines and psoriatic skin, − 9 lance (FDR = 1.7 × 10 ) (Additional file 3: Table S3: the though the direct comparison of transcriptomes of two gene enrichment results of 3 N); several SMG genes completely different biological sample types is unfeasible. (e.g. SMG1, SMG5,and SMG7) that code nonsense The approach was unable to give any final direct mediated mRNA decay factors of surveillance com- answers to the question what is the putative role of dif- plex [25], and CPSF genes encoding cleavage and ferent CCHCR1 variants. As an adenovirus-transformed polyadenylation specific factors were upregulated (1.5 cell line HEK293 cells have gone through structural al- < FC < 2.5). Heatmap of mRNA surveillance gene terations of genome. Therefore it is possible that we expression in the CCHCR1-HEK293 cell lines is have missed some gene expression data of the CCHCR1 shown in Fig. 5a. Although the gene expression cell lines. Furthermore the tumorigenic properties of changes were moderate, more than half of the genes HEK293 may have affected the gene regulation and -ex- acting in the mRNA surveillance pathway showed pression of the cell lines. Tervaniemi et al. BMC Genomics (2018) 19:432 Page 9 of 15 Fig. 5 CCHCR1 shows isoform-specific localization with centrosomal P-bodies and affects mRNA surveillance pathway. a Heatmap showing gene expression profiles of mRNA surveillance pathway in Iso1Non-risk (1 N), Iso1Risk (1R), Iso3Non-risk (3 N), and Iso3Risk (3R) CCHCR1-HEK293 cell lines, and in control cells; vector (V) and wildtype (WT). Color key: red represents upregulated and blue downregulated expression (row Z-score). b Immunofluorescent staining of the CCHCR1 (red) cell lines: Iso1Non-risk, Iso1Risk, Iso3Non-risk, and Iso3Risk, with the P-body markers (green) EDC4 and DCP1A. Panel b shows the overlap of red and green, and panel c shows the green channel images of the corresponding sites. The colocalization (orange) of CCHCR1 and P-body markers is absent in the Iso1Non-risk and Iso1Risk cells: only the CCHCR1 expression (red) is observable at the centrosomes and cytoplasm (b), whereas the antibodies against EDC4 and DCP1A fail to recognize these structures (c). In Iso3Non-risk and Iso3Risk, the colocalization (orange, marked with arrow heads in panel b) is observable with antibodies against the EDC4 and DCP1A (green in panel c). Nuclei are stained with DAPI (blue). Scale bar: 10 μm Tervaniemi et al. BMC Genomics (2018) 19:432 Page 10 of 15 Fig. 6 Model for CCHCR1 action in a cell and skin. CCHCR1 is expressed in a basal layer of keratinocytes in the epidermis of the healthy skin. In the thickened psoriatic epidermis, it is expressed in supra-basal layers as well. It is a dynamic protein that might play a role both in the cytoplasm at the centrosomes and P-bodies that occasionally colocalize, and possibly in the nucleus as well. Both centrosomes and P-bodies are physically linked to the cytoskeleton (cytokeratin (CK), actin, and microtubules (black lines)) that functions as a scaffold for the regulation of e.g. cell division and adhesion, mRNA transport and turnover, and transport of organelles like mitochondria. The previous results and the present RNAseq suggested that CCHCR1 regulates expression of cytoskeletal components and functions related to cytoskeleton, such as cell adhesion [6]. Furthermore, CCHCR1 may control the mRNA surveillance via the P-bodies, which may affect a wide variety of biological processes. Previous studies have identified several interaction partners for CCHCR1, these including the mitochondrial StAR protein that regulates steroidogenesis [8, 11], the RNA polymerase II subunit 3 (RPB3) that regulates transcription [12], some mitotic spindle proteins [15], and several viral proteins [30]. These interactions may result from the P-body localization of CCHCR1, or from the dynamic nature of CCHCR1 and its ability to regulate various processes via cytoskeleton. RNAseq identified several pathways for CCHCR1 which are relevant for the pathogenesis of psoriatic epidermis, epidermis development, and innate immunity. Interestingly, also the innate immunity including the RIG-like receptor (RLR) signaling is affected by the P-bodies CCHCR1 affects the expression of several genes adherens junctions specialized for cell-cell adhesion. The re- regulating cytoskeleton and cell-cell adhesion lated functions; regulation of actin cytoskeleton and focal Tight junction and adherens junction, were highlighted adhesion, were also highlighted in the CCHCR1-HEK293 cell especially in the Iso1 cells and showed gene enrichment in lines. They also showed enrichment among the mock genes. the Iso3 cells as well. Interestingly we have recently observed These two pathways consist of a large number of genes, CCHCR1 expression in the skin in the proximity of the cell hence their highlightment may result from several different membrane and desmosomes [6] that are like the tight and gene combinations and expression changes (up versus Tervaniemi et al. BMC Genomics (2018) 19:432 Page 11 of 15 down). Actin cytoskeleton and focal adhesion share several surveillance, the pathway analyses of the pooled data of genes with adherens and tight junctions that were not Non-risk and Iso3 highlighted RNA degradation and RNA highlighted among the mock genes. Furthermore, the func- transport. mRNA surveillance includes nonsense-mediated tions related to the cytoskeleton are supported by our previ- mRNA decay (NMD) that eliminates mRNAs containing ous microscopic observation; the actin cytoskeleton of the premature translation-termination codons [33]. In the Iso1 cells resembled that of wild type cells, but was disturbed RNAseq, several SMG genes encoding factors of NMD in the Iso3 cells. Furthermore, after disruption of microtu- surveillance complex, and CPSF genes encoding for cleav- bules, the Iso3Risk cells, especially, showed aberrant actin age and polyadenylation specific factors, were upregulated organization [6]. Here, we showed that several relevant in Iso3Non-risk. genes, including TLN1 (talin 1), FN1 (fibronectin 1), and P-bodies have been associated with several auto- FLNC (filamin C), were downregulated in the Iso3Risk cells, immune diseases, cancers, and neurological diseases which may disturb cytoskeleton organization. Talin [34, 35], but not with psoriasis. They are linked to vi- plays a role in the assembly of actin filaments and in- ruses and innate immunity as well. The P-bodies are teracts indirectly with extracellular fibronectin 1 [27], manipulated by the viruses for a productive infection andtheybothplayarolein adhesion, spreading, and to occur, but they also promote immune responses to migration of cells [28, 29]. Filamin crosslinks actin fil- the viral infection, e.g., via double-stranded RNA pro- aments into networks and anchor membrane proteins tein kinase (PKR alias EIF2AK2)[36]. Viruses, such as to the cytoskeleton, thereby affecting cell-cell and human papillomavirus HPV16 E6, induce in keratino- focal adhesion. Our previous results, as well as RNA- cytes the localization of PKR into the P-bodies sug- seq, also highlighted other cytoskeleton-related genes; gesting about the antiviral response. Interestingly, the several cytokeratin genes, e.g. KRT17 were upregu- interaction of CCHCR1 with HPV16 E2 induced a lated especially in the Iso1Non-risk cells [6]. One of massive redistribution of the virus protein into cyto- the most highly upregulated genes in the Iso3 cells was plasmic granules that were still co-localizing with SYT1. It encodes for synaptotagmin that interacts with CCHCR1 [30]. We suggest that these structures were tubulin, and may act as a microtubule-organizing cytoplasmic P-bodies playing an antiviral role. Inter- center-associated protein, thus playing a role in spindle estingly, RIG-I-like receptor signaling, which was organization [30, 31]. highlighted both in the CCHCR1-HEK293 cell lines and psoriatic skin, is part of the antiviral immune re- CCHCR1 shows isoform-specific effects on P-body sponse. The function of RIG-I-like receptors is linked localization and P-body-related functions at many levels to different RNA-granules including We have shown that in cultured cells CCHCR1 isoforms P-bodies [37]. 1 and 3 reside in the cytoplasm, as well as localize at the centrosomes [6, 8]. Interestingly, here we noticed a dif- Conclusions ference in their colocalization with the P-bodies in the The RNAseq highlighted pathways that support an ac- HEK293 cells. The immunostaining supported colocali- tive role for CCHCR1 in the centrosomes and P-bodies; zation with the P-bodies for isoform 3, especially at the the most significant pathways were related to the centrosome, while isoform 1 colocalized rarely with the cytoskeleton and cell adhesion, and furthermore, the P-bodies. The localization of P-bodies at the centro- mRNA surveillance pathway was highlighted in the somes seemed to be improved by the overexpression of Iso3Non-risk cells. The centrosomes and P-bodies are Iso3 CCHCR1. Recently it was shown that the physically linked to the cytoskeleton that functions as a N-terminus of Iso3Non-risk was necessary for the scaffold for the P-bodies (Fig. 6) and transported P-body localization [15], isoform 1 that is 89 amino acids mRNAs [17, 18], thus as a centrosomal P-body protein longer at its N-terminus than Isoform 3 was not along CCHCR1 might affect diverse cytoskeleton mediated in this study. The different N-terminus may explain the processes, such as cell division, cell adhesion, and trans- localization difference between isoforms 1 and 3. port of mRNA and mitochondria. The RNAseq revealed Previous studies have shown that there is a strong inter- clear isoform- and some haplotype-specific differences; action between CCHCR1 and EDC4 (enhancer of the Iso1 and Iso3 cells showed specific clustering as two mRNA-decapping 4 protein) [15] that is a central P-body separate groups, and the Non-risk and Risk cells often component regulating mRNA decay by decapping and exhibited opposite effects. Putative effects arising from 5′→ 3′ mRNA degradation. In our RNAseq, mRNA sur- the tumorigenic nature of HEK293 can’t be completely veillance pathway that is detecting and degrading abnor- ruled out. In summary, the present study strengthened mal mRNAs [32, 33] was highlighted in the HEK293 cells the previous observations about the effects and possible expressing Iso3Non-risk, the same isoform that was used interaction partners of CCHCR1 in the cell, as well as in the previous binding studies. In addition to mRNA provided new clues about its possible function. Tervaniemi et al. BMC Genomics (2018) 19:432 Page 12 of 15 Methods Microarray study with isoform 1-overexpressing cells Samples, RNA extraction, and RNAseq For the microarray study we used GeneChip® Human Generation of stable cell lines overexpressing the dif- Gene 1.0 ST Arrays. The total RNA samples of cells ferent CCHCR1 isoforms was described in our previ- were hybridized according to manufacturer’s protocols ous study [6]. Briefly, CCHCR1-pDsRed-constructs (Affymetrix Inc., Santa Clara, CA, USA). The arrays (Iso1Non-risk, Iso1Risk, Iso3Non-risk, or Iso3Risk) or were scanned with GeneChip scanner 3000 7G (Affyme- vector were stably transfected into HEK293 which is trix Inc.). The analyses of the microarray data were per- an easily transfectable tumorigenic cell line. Colonies formed using the statistical software R (http:// resistant to G418 antibiotics were screened for the www.R-project.org) using the Affy and Limma packages CCHCR1 expression by qPCR and fluorescence mi- [40, 41] similar to procedure earlier described [42]. In croscopy. The expression of selected clones was veri- short, we implemented the robust multiarray average fied by Western blotting [6]. The RNA extraction was method (RMA) [43] to normalize and calculate the Log2 performed at the same time as the quantification of expression values at gene level (annotated using hugen- clones without further passaging of cell lines. For the e10stv1.r3cdf_2.5.0). We did pair-wise comparison con- microarray and RNAseq experiments the total RNA trasts between the vector control and the different of the CCHCR1 cell lines and controls was extracted CCHCR1-HEK293 cell lines followed with t-statistics by aRNeasykit (Qiagen) and RNA qualitywas con- and log-odds of differential expression (B-value). The trolled by Bioanalyzer (RIN for all samples > 8.5). original microarray data compliance with MIAME pro- Total RNA samples, three replicates for each cedure has been submitted to ArrayExpress (the acces- CCHCR1-HEK293celllinewereusedfor theRNAseq sion number E-MTAB-6848). library preparation according to the previous single-cell tagged reverse transcription (STRT) proto- Quantitative real-time PCR col [38], which was adjusted for 10 ng samples as in We used qPCR to validate the expression of AREG, FN1, our previous studies [20, 24]. The libraries were se- TLN1, CXCL8, and SYT1 genes in the CCHCR1-HEK293 quenced using an Illumina HiSeq 2000 instrument. cell lines. cDNA synthesis was carried out with random Pre-processing of STRT reads, alignments and primers and SuperScript III First-Strand synthesis system per-gene quantitation were performed as previously (Invitrogen) according to manufacturer’s instructions. described [38]. The data normalization and differential ex- qPCR was carried out using an ABI PRISM 7900HT Se- pression analyses were performed using SAMstrt [39] quence Detection System with pre-designed TaqMan® (https://github.com/shka/R-SAMstrt/wiki), which is an Gene Expression Assays (Thermo Fisher Scientific) or enhancement of SAMseq [PMID: 22127579] for the spi- with self-designed primers by using Fast SYBR® Green kein-base normalization. Differentially expressed genes Master mix (Thermo Fisher Scientific) for detection, ac- were extracted by two class unpaired comparison; cording to manufacturer’s instructions. Lists of threshold of the significantly regulated genes was pre-designed TaqMan Gene Expression Assays and nu- FDR < 1%. We performed PCA with the scaling but cleotide sequences of self-designed qPCR primers are non-centering pre-process steps. A higher permuta- shown in Additional file 8. Each sample was run in three tion (n = 1000) was applied to the comparisons with replicates and the experiments were run twice. HPRT1 pooled samples (comb_Iso1, -Iso3, −Risk, and –Non-risk). was used as a reference gene for normalization and the The mock analysis with permutated sample groups was samples were compared against the wild type HEK293 performed to find possible false positive pathway results sample. Statistical comparisons were made with the Stu- and included six manual permutations per comparison dent’s t-test. group. The RNAseq of psoriatic skin samples was recently outlined elsewhere [20]. Briefly, all subjects involved in this study gave written informed consent and the study SNP genotyping followed the Declaration of Helsinki Guidelines. The We genotyped the CCHCR1 SNPs rs3130453 (CCH Institutional Review Board of the Helsinki University CR1*Iso1/3) and rs130076 (CCHCR1*WWCC/RRGS) Central Hospital had approved the study and the collec- with commercial allelic discrimination assays with tion of skin samples. The skin samples were harvested pre-designed probes and primers (TaqMan) described by a compressed air-driven dermatome (Zimmer®, previously [6] from DNA extracted from the skin sam- Warsaw, IN) to obtain an epidermal sample with min- ples [20]. Sample C.05 was not genotyped due to lack of imal dermis involvement. We collected lesional (PL) and sufficient amount of sample for DNA extraction. The non-lesional (PN) samples from six psoriatic patients HLA-Cw*06:02 genotype was also determined with com- and normal healthy skin from nine controls (C). mercial allelic discrimination assays (TaqMan) [44]. Tervaniemi et al. BMC Genomics (2018) 19:432 Page 13 of 15 Immunofluorescence The DEGs (comb_Iso1 and comb_Iso3) were analysed using the KEGG The CCHCR1-HEK293 cell lines, vector control, and pathway analysis of WebGestaltR. (XLSX 3054 kb) wild type HEK293 cells were grown on cover slips Additional file 6: Table S6. Haplotype specific gene enrichment analyses based on re-extracted DEGs of the CCHCR1-HEK293 cell lines. coated with rat tail collagen I (Gibco, Invitrogen) and The DEGs were obtained from the pooled data of Iso1Non-Risk and fixed with 4% paraformaldehyde solution. After fixation Iso3Non-Risk, and Iso1Risk and Iso3Risk compared to the controls (wild- cells were permeabilized with 0.1% Triton-×100 in PBS. type and vector). The DEGs (comb_Non-Risk, comb_Risk) were analysed using the KEGG pathway analysis of WebGestaltR. Summary of the gene The samples were incubated 1 h at room temperature enrichment results among the mock DEGs lists. (XLSX 2314 kb) with the antibodies against P-body markers EDC4 Additional file 7: Table S7 and Figure S1. CCHCR1 and HLA-Cw6 ge- (rabbit polyclonal, Cell Signaling) and DCP1A (mouse notypes of the skin samples. Figure S1. The CCHCR1 and HLA-Cw6 ge- monoclonal, Abnova), and centrosome marker γ-tubulin notypes illustrated in a PCA plot. (XLSX 79 kb) (mouse monoclonal, Sigma). Alexa Fluor 555 and 488 Additional file 8: Supplementary Information and Figure S2. Information about qPCR and co-localization of CCHCR1 with P-body conjugated IgGs (Invitrogen, Molecular Probes) were markers. Lists of pre-designed TaqMan Gene Expression Assays and used as secondary antibodies and the nuclei stained with nucleotide sequences of self-designed qPCR primers. Counting the DAPI (4′,6-diamidino-2-phenylindole, Sigma-Aldrich). colocalization of CCHCR1 with P-body markers in the CCHCR1-HEK293 cell lines and calculation of p-values for the comparison between cell lines. The pictures were taken with Zeiss LSM 5 Duo confocal Figure S2. γ-tubulin staining of the CCHCR1 cells. Antibody against microscope. Differences in localization of CCHCR1 iso- γ-tubulin was used as a marker for centrosomes. (PDF 1046 kb) forms with centrosomal P-bodies were determined by counting colocalized staining of CCHCR1 and EDC4 Abbreviations and DCP1A in each CCHCR1-overexpressing cell line *Iso1: Allele resulting in CCHCR1 isoform 1; *Iso3: Allele resulting in CCHCR1 isoform 3; *Non-risk: Non-risk haplotype (*RRGS)of CCHCR1; *Risk: Psoriasis- blindly (Additional file 8: Supplementary Information associated risk haplotype (*WWCC)of CCHCR1; DEG: Differentially expressed about qPCR and co-localization of CCHCR1 with gene; FC: Fold change; FDR: False discovery rate; Iso1 cells: CCHCR1 isoform P-body markers). 1 expressing cell lines; Iso3 cells: CCHCR1 isoform 3 expressing cell lines; PCA: Principal component analysis; STRT: Single-cell tagged reverse transcription Additional files Acknowledgements We thank Auli Saarinen for the technical support in the laboratory and Sten Linnarsson for help with RNA sequencing. Confocal imaging was done at Additional file 1: Table S1. Microarray data and gene enrichment Molecular Imaging Unit, Biomedicum Helsinki, University of Helsinki. This analysis of the CCHCR1 Iso1 cell lines. Gene expression profiling data of study was supported by Academy of Finland, Sigrid Jusélius Foundation, microarrays from the Isoform 1 CCHCR1-HEK293 cell lines. Gene enrich- Helsinki University Central Hospital Research Funds, and Folkhälsan Institute ment analysis of differentially expressed genes using the KEGG pathway of Genetics, and Swedish Research Council. The computation resources were analysis of WebGestaltR and functional cluster analysis of the DAVID. provided by SNIC through Uppsala Multidisciplinary Center for Advanced (XLSX 305 kb) Computational Science (UPPMAX) (b2010037). JK is a recipient of The Royal Additional file 2: Table S2. Differentially expressed genes of RNAseq Society Wolfson Research Excellence Award. from the comparisons between the CCHCR1-HEK293 cell lines and con- trols. Up- and downregulated DEGs of CCHCR1-overexpressing cell lines Funding Iso1Non-risk (1 N), Iso1Risk (1R), Iso3Non-risk (3 N), and Iso3Risk (3R) when This study was supported by Academy of Finland (130360; 1255560), Sigrid compared with controls (wild type and vector transfected cells). (XLSX Jusélius Foundation, Helsinki University Central Hospital Research Funds 1215 kb) (TYH2009233), and Folkhälsan Institute of Genetics, and Swedish Research Council. Additional file 3: Table S3. Gene enrichment analysis of DEGs from the CCHCR1-HEK293 cell lines based on RNAseq. Cell line-specific gene en- richment analyses of DEGs (Additional file 2: Table S2) from the CCHCR1- Availability of data and materials overexpressing cell lines Iso1Non-risk (1 N), Iso1Risk (1R), Iso3Non-risk All data generated during the study are included in this published article (3 N), and Iso3Risk (3R) when compared with wild type and vector con- and its supplementary information files. The microarray data will be available trols. The gene enrichment analysis of DEGs that were shared by the four on ArrayExpress (the accession number E-MTAB-6848). The processed STRT CCHCR1 cell lines (Intersection). Analyses were done using the KEGG reads supporting the results of this article are available in the European pathway analysis (WebGestaltR and WebGestalt), and GO and cluster ana- Nucleotide Archive (the accession number PRJEB20138). lyses from DAVID. Gene enrichment of Iso3Non-Risk DEGs in the mRNA surveillance pathway is shown (KEGG pathway figure and a list of genes). Authors’ contributions (XLSX 1183 kb) All authors (MHT, SK, TS, HAS, JV, KN, KT, SS, EK, JK, and OE) were involved in the study design. SS and JV recruited and clinically characterized the study Additional file 4: Table S4. Isoform- and haplotype-specific gene en- subjects in the psoriasis study, and sampled the skin specimen. HAS, KN, and richment with the shared DEGs of the CCHCR1-HEK293 cell lines. Gene TS prepared the samples for the experiments. KT analyzed the microarray enrichment analyses of DEGs shared by only the Non-risk (Diff N), Risk data. SK was responsible for the bioinformatics pipeline of the RNAseq data; (Diff R), isoform 1 (Diff iso1), or isoform 3 (Diff iso3) CCHCR1cell lines (see MHT and OE further analyzed the data; MHT was responsible for the in detail Fig. 4 Venn diagram). The DEGs shared by all the CCHCR1 cell immunofluorescent experiments. MHT, JK, and OE drafted the manuscript. All lines (Intersection) were analyzed as well. Analyses were done using the the authors contributed critical discussion and approved the final version of GO and cluster analyses from DAVID and KEGG pathway analysis from the manuscript. WebGestalt and WebGestaltR. (XLSX 307 kb) Additional file 5: Table S5. Isoform specific gene enrichment analyses Ethics approval and consent to participate based on re-extracted DEGs of the CCHCR1-HEK293 cell lines. The DEGs The collection of skin samples was approved by the Ethics Committee of the were obtained from the pooled data of Iso1Non-risk and Iso1Risk, and Hospital District of Helsinki and Uusimaa and by the Committee of Skin and Iso3Non-risk and Iso3Risk compared to the controls (wildtype and vector). Allergy Hospital, Helsinki University Central Hospital. All subjects involved in Tervaniemi et al. BMC Genomics (2018) 19:432 Page 14 of 15 this study gave written informed consent and the study followed the retained in the cytoplasm by its interaction with HCR, the psoriasis vulgaris Declaration of Helsinki Guidelines. candidate gene product. J Cell Sci. 2005;118(18):4253–60. 13. Andersen JS, Wilkinson CJ, Mayor T, Mortensen P, Nigg EA, Mann M. Competing interests Proteomic characterization of the human centrosome by protein correlation The authors declare that they have no competing interests. profiling. Nature. 2003;426(6966):570–4. 14. Jakobsen L, Vanselow K, Skogs M, Toyoda Y, Lundberg E, Poser I, Falkenby LG, Bennetzen M, Westendorf J, Nigg EA, et al. Novel asymmetrically Publisher’sNote localizing components of human centrosomes identified by complementary Springer Nature remains neutral with regard to jurisdictional claims in proteomics methods. EMBO J. 2011;30(8):1520–35. published maps and institutional affiliations. 15. Ling YH, Wong CC, Li KW, Chan KM, Boukamp P, Liu WK. CCHCR1 interacts with EDC4, suggesting its localization in P-bodies. Exp Cell Res. 2014;327(1): Author details 12–23. 1 2 Folkhälsan Institute of Genetics, 00014 Helsinki, Finland. Department of 16. Eulalio A, Behm-Ansmant I, Schweizer D, Izaurralde E. P-body formation is a Medical and Clinical Genetics, Medicum and Research Programs Unit, consequence, not the cause, of RNA-mediated gene silencing. Mol Cell Biol. Molecular Neurology, University of Helsinki, Helsinki, Finland. Department of 2007;27(11):3970–81. Biosciences and Nutrition, Karolinska Institutet, Huddinge, Sweden. Helsinki 17. Rajgor D, Shanahan CM. RNA granules and cytoskeletal links. Biochem Soc Burn Center, Department of Plastic Surgery, University of Helsinki and Trans. 2014;42(4):1206–10. Helsinki University Hospital, Helsinki, Finland. Department of Pharmacology, 18. Aizer A, Brody Y, Ler LW, Sonenberg N, Singer RH, Shav-Tal Y. The dynamics Medicum, University of Helsinki, Helsinki, Finland. Division of of mammalian P body transport, assembly, and disassembly in vivo. Mol Neuropsychiatry, Department of Women’s and Children’s Health, Center of Biol Cell. 2008;19(10):4154–66. Neurodevelopmental Disorders, Karolinska Institutet, Stockholm, Sweden. 19. Moser JJ, Fritzler MJ, Rattner JB. Repression of GW/P body components and Department of Dermatology, University of Helsinki and Helsinki University the RNAi microprocessor impacts primary ciliogenesis in human astrocytes. Hospital, Helsinki, Finland. School of Basic and Medical Biosciences, King’s BMC Cell Biol. 2011;12:37. College London, London, UK. 20. Tervaniemi MH, Katayama S, Skoog T, Siitonen HA, Vuola J, Nuutila K, Sormunen R, Johnsson A, Linnarsson S, Suomela S, et al. NOD-like receptor Received: 19 May 2017 Accepted: 21 May 2018 signaling and inflammasome-related pathways are highlighted in psoriatic epidermis. Sci Rep. 2016;6:22745. 21. Huang da W, Sherman BT, Lempicki RA. Systematic and integrative References analysis of large gene lists using DAVID bioinformatics resources. Nat 1. Harden JL, Krueger JG, Bowcock AM. The immunogenetics of psoriasis: a Protoc. 2009;4(1):44–57. comprehensive review. J Autoimmun. 2015;64:66–73. 22. Huang da W, Sherman BT, Lempicki RA. Bioinformatics enrichment tools: 2. Asumalahti K, Laitinen T, Itkonen-Vatjus R, Lokki ML, Suomela S, Snellman E, paths toward the comprehensive functional analysis of large gene lists. Saarialho-Kere U, Kere J. A candidate gene for psoriasis near HLA-C, HCR Nucleic Acids Res. 2009;37(1):1–13. (Pg8), is highly polymorphic with a disease-associated susceptibility allele. 23. Wang J, Vasaikar S, Shi Z, Greer M, Zhang B. WebGestalt 2017: a more Hum Mol Genet. 2000;9(10):1533–42. comprehensive, powerful, flexible and interactive gene set enrichment 3. Asumalahti K, Veal C, Laitinen T, Suomela S, Allen M, Elomaa O, Moser M, de analysis toolkit. Nucleic Acids Res. 2017;45(Web Server Issue):W130–7. Cid R, Ripatti S, Vorechovsky I, et al. Coding haplotype analysis supports 24. Katayama S, Skoog T, Jouhilahti EM, Siitonen HA, Nuutila K, Tervaniemi MH, HCR as the putative susceptibility gene for psoriasis at the MHC PSORS1 Vuola J, Johnsson A, Lonnerberg P, Linnarsson S, et al. Gene expression locus. Hum Mol Genet. 2002;11(5):589–97. analysis of skin grafts and cultured keratinocytes using synthetic RNA 4. Zhang XJ, Huang W, Yang S, Sun LD, Zhang FY, Zhu QX, Zhang FR, Zhang normalization reveals insights into differentiation and growth control. BMC C, Du WH, Pu XM, et al. Psoriasis genome-wide association study identifies Genomics. 2015;16:476. susceptibility variants within LCE gene cluster at 1q21. Nat Genet. 2009; 25. Yamashita A, Izumi N, Kashima I, Ohnishi T, Saari B, Katsuhata Y, Muramatsu 41(2):205–10. R, Morita T, Iwamatsu A, Hachiya T, et al. SMG-8 and SMG-9, two novel 5. Chang YT, Chou CT, Shiao YM, Lin MW, Yu CW, Chen CC, Huang CH, Lee DD, subunits of the SMG-1 complex, regulate remodeling of the mRNA Liu HN, Wang WJ, et al. Psoriasis vulgaris in Chinese individuals is associated surveillance complex during nonsense-mediated mRNA decay. Genes Dev. with PSORS1C3 and CDSN genes. Br J Dermatol. 2006;155(4):663–9. 2009;23(9):1091–105. 6. Tervaniemi MH, Siitonen HA, Soderhall C, Minhas G, Vuola J, Tiala I, 26. Muller M, Demeret C. CCHCR1 interacts specifically with the E2 protein of Sormunen R, Samuelsson L, Suomela S, Kere J, et al. Centrosomal human papillomavirus type 16 on a surface overlapping BRD4 binding. localization of the psoriasis candidate gene product, CCHCR1, supports a PLoS One. 2014;9(3):e92581. role in cytoskeletal organization. PLoS One. 2012;7(11):e49920. 27. Jiang G, Giannone G, Critchley DR, Fukumoto E, Sheetz MP. Two-piconewton 7. Suomela S, Elomaa O, Asumalahti K, Kariniemi AL, Karvonen SL, Peltonen J, slip bond between fibronectin and the cytoskeleton depends on Talin. Nature. Kere J, Saarialho-Kere U. HCR, a candidate gene for psoriasis, is expressed 2003;424(6946):334–7. differently in psoriasis and other hyperproliferative skin disorders and is 28. Ling K, Doughman RL, Firestone AJ, Bunce MW, Anderson RA. Type I downregulated by interferon-gamma in keratinocytes. J Invest Dermatol. gamma phosphatidylinositol phosphate kinase targets and regulates focal 2003;121(6):1360–4. adhesions. Nature. 2002;420(6911):89–93. 8. Tiala I, Suomela S, Huuhtanen J, Wakkinen J, Holtta-Vuori M, Kainu K, Ranta 29. Muro AF, Chauhan AK, Gajovic S, Iaconcig A, Porro F, Stanta G, Baralle FE. S, Turpeinen U, Hamalainen E, Jiao H, et al. The CCHCR1 (HCR) gene is Regulated splicing of the fibronectin EDA exon is essential for proper skin relevant for skin steroidogenesis and downregulated in cultured psoriatic wound healing and normal lifespan. J Cell Biol. 2003;162(1):149–60. keratinocytes. J Mol Med. 2007;85(6):589–601. 30. Honda A, Yamada M, Saisu H, Takahashi H, Mori KJ, Abe T. Direct, Ca2+ 9. Elomaa O, Majuri I, Suomela S, Asumalahti K, Jiao H, Mirzaei Z, Rozell B, −dependent interaction between tubulin and synaptotagmin I: a possible Dahlman-Wright K, Pispa J, Kere J, et al. Transgenic mouse models mechanism for attaching synaptic vesicles to microtubules. J Biol Chem. support HCR as an effector gene in the PSORS1 locus. Hum Mol Genet. 2002;277(23):20234–42. 2004;13(15):1551–61. 31. Zhu XL, Qi ST, Liu J, Chen L, Zhang C, Yang SW, Ouyang YC, Hou Y, Schatten H, 10. Tiala I, Wakkinen J, Suomela S, Puolakkainen P, Tammi R, Forsberg S, Song YL, et al. Synaptotagmin1 is required for spindle stability and metaphase- Rollman O, Kainu K, Rozell B, Kere J, et al. The PSORS1 locus gene CCHCR1 to-anaphase transition in mouse oocytes. Cell Cycle. 2012;11(4):818–26. affects keratinocyte proliferation in transgenic mice. Hum Mol Genet. 2008; 32. Chang YF, Imam JS, Wilkinson MF. The nonsense-mediated decay RNA 17(7):1043–51. surveillance pathway. Annu Rev Biochem. 2007;76:51–74. 11. Sugawara T, Shimizu H, Hoshi N, Nakajima A, Fujimoto S. Steroidogenic acute regulatory protein-binding protein cloned by a yeast two-hybrid 33. Kurosaki T, Maquat LE. Nonsense-mediated mRNA decay in humans at a system. J Biol Chem. 2003;278(43):42487–94. glance. J Cell Sci. 2016;129(3):461–7. 12. Corbi N, Bruno T, De Angelis R, Di Padova M, Libri V, Di Certo MG, Spinardi 34. Savas JN, Makusky A, Ottosen S, Baillat D, Then F, Krainc D, Shiekhattar R, L, Floridi A, Fanciulli M, Passananti C. RNA polymerase II subunit 3 is Markey SP, Tanese N. Huntington's disease protein contributes to RNA- Tervaniemi et al. BMC Genomics (2018) 19:432 Page 15 of 15 mediated gene silencing through association with Argonaute and P bodies. Proc Natl Acad Sci U S A. 2008;105(31):10820–5. 35. Anderson P, Kedersha N, Ivanov P. Stress granules, P-bodies and cancer. Biochim Biophys Acta. 2015;1849(7):861–70. 36. Dougherty JD, Reineke LC, Lloyd RE. mRNA decapping enzyme 1a (Dcp1a)- induced translational arrest through protein kinase R (PKR) activation requires the N-terminal enabled vasodilator-stimulated protein homology 1 (EVH1) domain. J Biol Chem. 2014;289(7):3936–49. 37. Lloyd RE. Regulation of stress granules and P-bodies during RNA virus infection. Wiley Interdiscip Rev RNA. 2013;4(3):317–31. 38. Islam S, Kjallquist U, Moliner A, Zajac P, Fan JB, Lonnerberg P, Linnarsson S. Highly multiplexed and strand-specific single-cell RNA 5′ end sequencing. Nat Protoc. 2012;7(5):813–28. 39. Katayama S, Töhönen V, Linnarsson S, Kere J. SAMstrt: statistical test for differential expression in single-cell transcriptome with spike-in normalization. Bioinformatics. 2013;29(22):2943–5. 40. Gautier L, Cope L, Bolstad BM, Irizarry RA. Affy–analysis of Affymetrix GeneChip data at the probe level. Bioinformatics. 2004;20(3):307–15. 41. Smyth GK. Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol. 2004;3:Article3. 42. Massinen S, Hokkanen ME, Matsson H, Tammimies K, Tapia-Paez I, Dahlstrom-Heuser V, Kuja-Panula J, Burghoorn J, Jeppsson KE, Swoboda P, et al. Increased expression of the dyslexia candidate gene DCDC2 affects length and signaling of primary cilia in neurons. PLoS One. 2011;6(6): e20580. 43. Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, Speed TP. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics. 2003;4(2):249–64. 44. Nikamo P, Stahle M. Cost-effective HLA-Cw06:02 typing in a Caucasian population. Exp Dermatol. 2012;21(3):221–3.

Journal

BMC GenomicsSpringer Journals

Published: Jun 4, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off