Liprin-α1 modulates cancer cell signaling by transmembrane protein CD82 in adhesive membrane domains linked to cytoskeleton
PPFIA1 is located at the 11q13 region commonly amplified in cancer. The protein liprin-α1 encoded by PPF1A1 contributes to the adhesive and invasive structures of cytoskeletal elements and is located at the invadosomes in cancer cells. However, the precise mechanism of liprin-α1 function in cancer progression has remained elusive.
Invasion regulating activity of liprin-α1 was examined by analyzing the functions of squamous cell carcinoma of head and neck (HNSCC) cell lines in three-dimensional collagen I after RNAi mediated gene knockdown. Transcriptome profiling and Gene Set Enrichment Analysis from HNSCC and breast cancer cells were used to identify expression changes relevant to specific cellular localizations, biological processes and signaling pathways after PPFIA1 knockdown. The significance of the results was assessed by relevant statistical methods (Wald and Benjamini-Hochberg). Localization of proteins associated to liprin-α1 was studied by immunofluorescence in 2D and 3D conditions. The association of PPFIA1 amplification to HNSCC patient survival was explored using The Cancer Genome Atlas data.
In this study, we show that liprin-α1 regulates biological processes related to membrane microdomains in breast carcinoma, as well as protein trafficking, cell-cell and cell-substrate contacts in HNSCC cell lines cultured in three-dimensional matrix. Importantly, we show that in all these cancer cells liprin-α1 knockdown leads to the upregulation of transmembrane protein CD82, which is a suppressor of metastasis in several solid tumors.
Our results provide novel information regarding the function of liprin-α1 in biological processes essential in cancer progression. The results reveal liprin-α1 as a novel regulator of CD82, linking liprin-α1 to the cancer cell invasion and metastasis pathways.
KeywordsLiprin-α1 PPFIA1 CD82 Head and neck cancer Breast cancer Invasion RNA sequencing Three-dimensional cell culture
American Type Culture Collection
1,4-diazabicyclo (2.2.2) octane
(ELKS/RAB6-interacting/CAST family member 1)
False discovery rate
Gene Expression Omnibus
Gene set enrichment analysis
Head and neck squamous cell carcinoma
Normalized enrichment score
Oral squamous cell carcinoma
shScramble (small hairpin)
Sortilin-related receptor 1
The Cancer Genome Atlas
PPFIA1 is located at the 11q13 amplification region  which is related to poor prognosis of the patients in several cancers, including head and neck squamous cell carcinoma (HNSCC) and breast cancer [2, 3, 4]. PPFIA1 encodes liprin-α1 protein, which is a member of the liprin protein family of tyrosine phosphatase interacting proteins conserved in evolution [5, 6]. Liprin-α proteins have been studied extensively in neurons with reported involvement in synapse functions [7, 8, 9, 10].
In addition to the functions in neuronal cells, liprin-α1 has been associated to cancer metastases , cell migration and invasive growth [12, 13]. Of note, liprin-α1 affects cancer cell spreading, the distribution of cell surface β1-integrins , and regulates cell edge dynamics and focal adhesion assembly in motile epithelial cancer cells via proteins including vimentin, ERC1 (ELKS/RAB6-interacting/CAST family member 1) and β1-integrin [12, 15]. We have recently shown that in non-invasive cancer cells liprin-α1 locates to invadosome structures and promotes growth behavior with limited invasive capacity , whereas in invasive and motile cancer cells liprin-α1 is essential for mesenchymal cancer cell invasion and regulation of extracellular matrix degradation [12, 13]. Besides the cancer promoting functions, liprin-α1 has been recently implicated in recycling of active α5β1 in fibronectin polymerization-dependent vascular morphogenesis . These results suggest several important cellular functions of liprin-α1 in both neuronal and epithelial cancer cells.
In the present study, our aim was to explore the cellular liprin-α1 functions in three-dimensional (3D) collagen I matrix environment, and to identify genes and molecular mechanisms that are involved in liprin-α1 mediated regulation of cell invasive growth. Our results revealed a unique interplay between liprin-α1 and CD82 transmembrane protein in the invasion of HNSCC and breast cancer cells, thus providing mechanistic details of liprin-α1 function in cancer cell progression.
Cell lines and reagents
Two breast cancer cell lines MDA-MB-231 from metastatic breast adenocarcinoma and Hs578T cell line from breast carcinoma (ATCC, American Type Culture Collection, Manassas, MD, USA) were studied. HNSCC cell lines UT-SCC-42A from laryngeal cancer, UT-SCC-42B from corresponding neck metastasis, UT-SCC-19B from laryngeal persistent cancer and UT-SCC-24B from neck metastasis of tongue cancer were derived from clinical samples (Reidar Grénman, Department of Otorhinolaryngology – Head and Neck Surgery, Turku University Hospital, Finland). UT-SCC and MDA-MB-231 cell lines were cultured using Dulbecco’s Modified Eagle’s Medium (DMEM) (Lonza, Verviers, Belgium) with an added 2 mM of L-glutamine, 0.1 mM of non-essential amino acids (NEAA) (Lonza), penicillin/streptomycin antibiotics (100 U/ml) (Lonza) and 10% fetal bovine serum (FBS) (Gibco). The Hs578T cell line was cultured using RPMI-1640 medium (Lonza) with the same supplements added as with the DMEM.
Constructs and lentiviral transduction
Lentiviral particles were generated for shRNA constructs from the TRC1 library (Sigma-Aldrich, St. Louis, Missouri, U.S.) targeting PPFIA1 and VIM. Constructs for PPFIA1 were TRCN0000342514, TRCN0000380944, TRCN0000002969, and TRCN0000380097, named as shPPFIA1_14, shPPFIA1_44, shPPFIA1_69 and shPPFIA1_97, and for VIM they were TRCN0000029119 and TRCN0000029121 (Sigma-Aldrich, St. Louis, Missouri, U.S.). The shRNA was cloned into the pLKO.1 vector with the puromycin resistance gene. Lentiviral particles carrying the shRNA constructs for genes of interest and shScramble (shScr) used as a control were generated at the Biomedicum Functional Genomics Unit (FuGU, University of Helsinki, Finland). Transduction of each cell line was performed as briefly described: cells were counted at 100,000 cells/ml, seeded into 12-well plates, incubated overnight, and infected with the lentiviral particles. Selection for the infected shRNA cells was done by using 1 μg/ml puromycin (Sigma-Aldrich).
Primary antibodies used for immunofluorescence or western blotting were mouse monoclonal α-tubulin (Sigma-Aldrich, St. Louis, Missouri, U.S), mouse monoclonal vinculin (Sigma-Aldrich, St. Louis, Missouri, U.S.), rabbit polyclonal liprin-α1 (Proteintech, Manchester, U.K.), mouse monoclonal CD82 (Abcam, Cambridge, U.K.), rabbit monoclonal sortilin-related receptor 1 (SORL1) (Cell Signaling, Danvers, Massachusetts, U.S.), mouse monoclonal vimentin (Sigma-Aldrich, St. Louis, Missouri, U.S.), mouse monoclonal cytokeratin-13 and rabbit monoclonal cytokeratin-4 (Abcam, Cambridge, U.K.), rabbit polyclonal cytokeratin-13 and rabbit polyclonal cytokeratin-4 (Proteintech, Manchester, U.K.), mouse monoclonal cytokeratin 10 (Invitrogen, Thermo-Fisher Scientific, Rockford, U.S.), goat polyclonal actin (Abcam, Cambridge, U.K.), rabbit monoclonal EEA1 3288 (Cell Signaling), rabbit monoclonal caveolin-1 (D46G3) (Cell Signaling), rabbit polyclonal ZO-1 (Invitrogen). Alexa Fluor phalloidin 488 and 594 (Invitrogen, Thermo-Scientific, Eugene, OR, U.S.) were used to stain the actin cytoskeleton. As secondary antibodies, Alexa Fluor 488 or 594 F(ab’) 2 fragment of goat anti-mouse IgG (H + L) and Alexa Fluor 488 or 594 F(ab’) 2 fragment of goat anti-rabbit IgG (H + L) (Invitrogen, Eugene, OR, U.S.) were used for immunofluorescence microscopy. For immunoblotting, secondary antibodies horseradish peroxidase (HRP) conjugate goat anti-mouse IgG (H + L) (Life Technologies, Rockford, IL, U.S.), HRP conjugate goat anti-rabbit IgG (H + L) (Life Technologies, Rockford, IL, U.S.), and HRP conjugate rabbit anti-goat IgG (Abcam, Cambridge, U.K.) were used.
Cells were grown until they reached 70–80% confluency and lysed using lysis buffer (radioimmunoprecipitation buffer, RIPA) (Sigma-Aldrich, Rockford, IL, U.S.), including protease (complete protease inhibitor cocktail, Roche, Mannheim, Germany) and phosphatase inhibitors (phosphatase inhibitor cocktail, Roche). The lysates were agitated at + 4 °C for 10 min, incubated on ice for 30 min, resuspended gently and centrifuged at 14,000 g after which the supernatants were collected. Protein concentration was measured with an RC DC Protein Assay Kit (BioRad). Loading buffer with 5% β-mercaptoethanol and 2 x Laemmli buffer (BioRad) was added to each sample. Samples were denatured by boiling at + 95 °C for 5 min. For non-reducing conditions β-mercaptoethanol and boiling were excluded. Samples were loaded into the SDS gels (BioRad), from which proteins were transferred to polyvinylidene fluoride (PVDF) membranes (BioRad) using Transfer Blot Turbo (BioRad) equipment. Membranes were incubated at room temperature in a blocking solution consisting of 5% milk powder or 5% bovine serum albumin (BSA) in tris-buffered saline and Tween (TBST) followed by TBST washes. Membranes were incubated with primary antibodies for 1 h at room temperature or overnight at + 4 °C with shaking. After incubation, the membranes were washed with TBST followed by incubation with secondary antibody for 1 h and washing with TBST. Finally, the membranes were incubated with detection reagents (Millipore) for two minutes. Chemiluminescence signals were detected on X-ray films (Kodak) or by using Chemidoc Imaging System (BioRad). Quantification of western blot was measured by ImageLab (BioRad) or by ImageJ, and statistical significance was measured with student’s t-test and error bars as standard deviations. Constructs with efficient knockdown were used in quantification of protein levels.
Cell culture and immunofluorescence
For three-dimensional (3D) invasive growth assay, cells were grown on type I collagen (2.2 mg/ml, Sigma). Immunofluorescence for two-dimensional (2D) and 3D cultures and quantification of 3D cell cultures were performed as previously described [12, 17]. Alternatively, 40,000 cells/ml were counted and embedded into collagen I drops (Gibco, Life Technologies) in a 24-well plate and incubated for seven days changing medium every day. The cells were fixed using 4% paraformaldehyde in phosphate-buffered saline (PBS), followed by two washes with PBS. Permeabilization was done using 0.25% Triton-X, after which the cells were washed with PBS, incubated with 0.12% glycin, 0.25% Tx-100 in PBS and washed with 0.25% Tx-100 in PBS. After blocking the cells for 60 min in 3% bovine serum albumin (BSA) 0. 25% Tx-100 in PBS or in 15% FBS, 0.3% Tx-100 in PBS (2 h, RT), they were incubated in primary antibody diluted in 1:100 or 1:200 with 3% BSA, 0.25% Triton-X in PBS or in 15% FBS, 0.3% TX-100 in PBS 1 h-overnight at + 4 °C in shaking. The cells were then washed with 0.25–0.45% Triton-X in PBS followed by incubation in a secondary antibody using dilution of 1:400 in 3% BSA, 0.25% Triton-X in PBS or in 15% FBS, 0.3% Tx-100 in PBS for one to four hours with shaking in dark at RT. The cells were washed with 0.25–0.45% Triton-X in PBS minimum three times over several hours and milli-Q water two times with shaking followed by washes with PBS for one hour to overnight. Collagen I drops were mounted in Mowiol containing 1.4-diazabicyclo (2.2.2) octane (DABCO) and 4′6-diamidino-2-phenylindole (DAPI) (Sigma) for nucleus staining.
Confocal images were taken using a Zeiss Meta 780/880 laser scanning microscope with a Zeiss 40× or a 63×/1.4 N.A. plan-apochromat oil objective, and adjusted and background corrected with ZenLite (2.3 Lite), Adobe Photoshop CS6, Illustrator CS6 software, Corel and/or ImageJ. Brightness and contrast were linearly adjusted using ZEN 2012 (blue edition; Carl Zeiss), Adobe Photoshop CS6 or Photo-Paint X7 (Corel). Single optical sections were used for image display. Three-dimensional collagen contraction images were taken using light microscopy (Leica MZ FLIII). Image editing was done with Adobe Photoshop CS6 and Illustrator CS6 software.
For co-localization analysis, the confocal images were deconvolved, thresholded by the Costes method and analyzed by the Pearson’s correlation coefficient with Huygens Professional version 18.04 (Scientific, __http://svi.nl__).
Three-dimensional (3D) cell culture for RNA sequencing
A total of 300,000 cells/ml were counted and mixed with collagen I (Gibco, Life Technologies) diluted in 10 x PBS (Lonza), 1 M NaOH and sterile milliQ water according to manufacturer’s instructions (Gibco, Life Technologies). Collagen I / cell suspension was plated into the 24-well low adhesion plates for the final volume of 500 μl. The collagen I / cell suspension was incubated at + 37 °C for 30 min, after which medium was added on top and changed daily. Cells in collagen I gels were grown for five days after which they were further processed for RNA sequencing. For HNSCC cell lines, the plates were coated with poly(2-hydroxyethyl methacrylate)/poly-HEMA (Sigma-Aldrich) to avoid two-dimensional cell growth and attachment of the cells to the bottom of the plates.
Collagen contraction and spheroid formation assays
Collagen gel contraction model is a method to estimate cell mediated contracture of the ECM in vitro . MDA-MB-231 control and liprin-α1 knockdown cells were grown in collagen I using the same protocol than in 3D cell culture. Low adhesion 24-well plates were used as cell culture plates. Cell suspensions in collagen I were imaged using optical microscope (Leica MZ FLIII) on days 1, 2, 5 and 6. Relative cell area was calculated using ImageJ software. For spheroid formation assay, 10,000 cells were counted and plated on U bottom 96-well low-adhesion plates and spheroid formation was monitored using light microscope.
RNA isolation and removal of ribosomal RNA
One milliliter of Trizol (Life Technologies) was added to cell/collagen I mixture which was further mixed with the Precellys ceramic beads followed by homogenization with a homogenisator (Precellys). 0.2 ml of chloroform (Fisher Scientific, Hampton, New Hampshire, US) was added to the samples followed by shaking the samples manually for 15 s. After 2 min incubation in room temperature, the samples were centrifuged at 11,300 rpm for 15 min at + 4 °C. Upper bright phase was kept, 0.5 ml of isopropanol (Fisher Scientific) was added and the solution was mixed. The samples were incubated for 10 min in room temperature and centrifuged at 11,300 rpm for 10 min at + 4 °C after which 1 ml of 75% ethanol (Etax) was added. The samples were mixed and centrifuged at 8800 rpm for 5 min at + 4 °C. Pellets were dissolved to Rnase-free water (Lonza AccuGENE Molecular biology Grade Water) at + 55 °C for 10 min. Isolated RNA was purified using Qiagen RNeasy Mini kit according to Qiagen protocol (Qiagen, Hilden, Germany). Ribosomal RNA was removed by using RiboZero Complete Gold Human kit (Illumina, San Diego, California, US) according to manufacturer’s instructions. Similarly, the second strand cDNA synthesis (NEBNext, Illumina, San Diego, California, US) and purification of double-stranded cDNA by 1.8X Agencourt AMPure XP beads were carried out according to manufacturers’ instructions (Illumina, San Diego, California, US).
For library preparation, NEBNext Directional RNA library kit was used according to manufacturer’s instructions (Illumina, San Diego, California, US). The PCR amplified library was purified using Agencourt AMPure XP beads according to manufacturer’s instructions (Illumina, San Diego, California, US). The quality of the sequencing library was assessed using Bioanalyzer (Agilent High Sensitivity Chip) and concentrations were measured by Qubit. Library denaturation and dilution were made according to NextSeq 500 system guide (Illumina). Samples were loaded onto the Illumina NextSeq 500 High 75 bp reagent cartridge for RNA sequencing.
RNA sequencing data analysis
RNA-seq data was pre-processed using QualiMap, FastQC, STAR aligner, and trimmomatic software. Differential expression analysis was performed by DESeq2 package R v3.2.3 and results underwent logarithmic transformation. Statistical significance of the gene expression changes between liprin-α1 knockdown (shPPFIA1) and control cells (shScr) was evaluated by Wald test. Significantly differentially expressed genes were considered those with corrected p-values padj < 0.05 (Benjamini-Hochberg). Three different biological replicates per cell line (UT-SCC-42A, UT-SCC-42B, MDA-MB-231) were studied. The RNA sequencing data have been deposited in The National Center for Biotechnology Information’s (NCBI’s) Gene Expression Omnibus (GEO, Series accession number GSE108392). In addition, we compared the current RNA sequencing data to our previously published UT-SCC-24B cell line hybridized on Affymetrix GeneChip Human Exon 1.0 ST gene expression microarrays (GEO Series accession number GSE75756) . Genes that were differentially expressed due to liprin-α1 silencing in different conditions and cell lines were compared using Venn diagrams by Venny 2.1 analysis program (http://bioinfogp.cnb.csic.es/tools/venny/index.html). Gene expression data for each sample and condition can be found in Gene Expression Omnibus under accession numbers described above.
Gene set enrichment analysis (GSEA) data analysis
Gene set enrichment analysis [19, 20] was used to study whether a priori defined set of genes showed statistically significant, analogous differences between control (shScr) and liprin-α1 (shPPFIA1) knockdown cells. The enrichment score (ES) was the result of a gene set enrichment analysis, and reflected how overrepresented the gene set was at the top or bottom of a ranked gene listing. Pathways were ranked by their normalized enrichment score (NES), and were considered significant if their q- and p-values were under 0.05. False discovery rate (FDR) signified the probability of a false positive finding and the nominal p-value stated the statistical significance of the ES for the gene set [19, 20].
Knock-down of CD82 by siRNAs
For knocking-down CD82 in shPPFIA1 HNSCC cells by using a siRNA pool for CD82 (Sigma-Aldrich), UT-SCC-42B cells were counted at 200000/ml, and seeded to six well plates. Cells were grown overnight and siRNA transfection was performed. Transfection reagents used were Lipofectamine RNAiMAX (Thermo-Fisher) and Opti-MEM (Thermo-Fisher). Transfection experiment was performed according to manufacturer’s instructions (Thermo-Fisher). Western blot was performed to verify the efficiency of knock-down. Control cells and CD82 knockdown cells were counted (40,000) and embedded to collagen I (Gibco), and were monitored and incubated for seven days, after which colonies were fixed with 4% PFA and stained with phalloidin. Colonies were quantified by ImageJ and average area of colonies from three replicates/condition was calculated. Error bars were calculated as standard deviation, and statistical significance was measured by unpaired student’s t-test with three different replicates/condition.
Liprin-α1 promotes invasive cellular growth in metastatic head and neck squamous cell carcinoma cells
In control cells, liprin-α1 was enriched at the cell colony edge in F-actin and/or vinculin containing structures protruding into the collagen matrix (Fig. 1f-g, Additional file 1: Figure S1). Especially in the lumen-forming UT-SCC-42B cell colonies, liprin-α1 was concentrated to the outer surface of the colony (Fig. 1f, Additional file 1: Figure S1).
Liprin-α1 knockdown resulted in differential expression of 102 shared genes in HNSCC and breast cancer cells
Gene set enrichment analysis (GSEA) revealed involvement of liprin-α1 in regulation of protein trafficking, cell-cell junctions and membrane microdomains
Enriched gene sets in GSEA analyses in HNSCC and breast cancer cell lines after liprin-α1 knockdown cultivated in 3D collagen I
Gene Set: HNSCC
GO cell substrate junction
GO establishment of protein localization to endoplasmic reticulum
GO protein localization to endoplasmic reticulum
GO anchoring junction
GO protein targeting to membrane
GO establishment of protein localization to organelle
GO regulation of vacuolar transport
GO phosphatase complex
GO establishment of protein localization to membrane
GENE SET: Breast cancer
GO negative regulation of hydrolase activity
GO regulation of proteolysis
GO negative regulation of peptidase activity
GO regulation of peptidase activity
GO membrane microdomain
GO positive regulation of cell death
GO morphogenesis of an epithelium
Liprin-α1 suppresses tetraspanin CD82 expression
Modulation of cytoskeletal elements by liprin-α1 in cancer cells
Our previous results from metastatic UT-SCC-24B HNSCC cell line grown in 2D cell culture  showed vimentin upregulation in shPPFIA1 versus shScr cells. In the present study, HNSCC cell lines cultured in 3D matrix VIM was upregulated in both the RNA ((log2) fold change = 1.7, p = 0.0033, Additional file 6: Figure S3A, Additional file 2: Table S1) and protein level (Additional file 10: Figure S7A) in shPPFIA1 cells, which is in conjunction with our previous results. Our previous findings by microarrays showed liprin-α1 knockdown to repress keratin intermediate filaments in metastatic UT-SCC-24B cell line , which suggests that keratin network compensates low endogenous vimentin expression in certain types of HNSCC (Additional file 10: Figure S7B-C). Next, we evaluated whether vimentin knockdown affects liprin-α1 expression or localization. After knocking down vimentin in MDA-MB-231 and Hs578T breast cancer cells, which highly express endogenous vimentin, liprin-α1 was not localized near the cell edge or after the leading edge as in control cells, although there were no significant changes at the protein expression levels in these breast cancer cell lines (Additional file 10: Figure S7D-E). Vimentin knockdown had effect on focal adhesion localization exemplified by cytoplasmic staining of vinculin after vimentin knockdown, and lack of liprin-α1 localization at the vicinity of focal adhesions (Additional file 10: Figure S7E). Vimentin knockdown in these cells was not compensated by upregulating keratin intermediate filaments. On the contrary, keratin 13 expression was decreasing after knockdown of vimentin in breast cancer cell lines (Additional file 10: Figure S7D). These results support our previous data on the effect of liprin-α1 knockdown to adhesion and intermediate filament signaling. Furthermore, breast cancer cells with high endogenous vimentin expression were incapable for mesenchymal invasion inside three dimensional collagen I matrix after liprin-α1 knockdown due to disorganized vimentin network .
Significance of PPFIA1 amplification to survival of clinical HNSCC and breast cancer patients
Due to our results showing the association between liprin-α1 and CD82, we explored the significance of PPFIA1 amplification in survival of clinical HNSCC and breast cancer patients from The Cancer Genome Atlas (TCGA) data. Survival data showed shorter survival for patients with PPFIA1 amplification when compared to patients with no alteration in both the HNSCC (35.81 versus 64.78 months, p = 0.0152) and breast cancer data (125.3 versus 164.3 months, p = 0.00998) (cbioportal, TCGA-data, Additional file 11: Figure S8) [21, 22].
PPFIA1 encoding liprin-α1 shows high correlation between copy number and gene expression in different cancer types and is related to metastatic potential and cell invasive growth in breast carcinoma [1, 11, 12, 23]. The amplification of 11q13 region, where PPFIA1 is located, is associated to the presence of metastases also in oral squamous cell carcinoma (OSCC) . In the present study, we demonstrated that liprin-α1 enhances invasive potential in metastatic HNSCC cells cultured in 3D collagen I matrix, corroborating our previous findings on the effect of liprin-α1 on cell invasion or invasive growth of motile cells. In this study, we further investigated the mechanism behind the function of liprin-α1 by performing RNA sequencing of breast cancer and HNSCC cells cultured in 3D collagen I environment. When comparing UT-SCC cell lines to metastatic breast cancer after liprin-α1 knockdown, 6.8% of differentially expressed genes were shared between HNSCC and breast cancer, which underlines the heterogeneity of the cancer cell lines. GSEA data analysis demonstrated liprin-α1 involvement in protein trafficking, modulation of membrane composition, as well as cell-cell and cell-substrate signaling. These associations suggest that liprin-α1 is an important player in a variety of different processes related to cell-cell communication and intracellular protein transport in HNSCC and breast cancer cells in 3D environment.
Liprin-α1 knockdown led to upregulation of CD82 cell surface protein, which inhibits formation of invasive structures called microprotrusions during cancer cell invasion . CD82 overexpression correlates with reduced cell growth, migration, invasion and xenograft tumor growth in OSCC , and CD82 downregulation is associated with poorer survival in OSCC . Moreover, CD82 has been proposed to decrease cellular dissemination of cancerous cells from primary tumor  and invasive potential of different solid tumors (reviewed in ). CD82 downregulation associates to increased cell spreading in dendritic cells  supporting previous findings on association of liprin-α1 expression to cell edge dynamics and enhanced cell spreading . When we knocked down liprin-α1 in HNSCC and breast cancer cells grown in two-dimensional cell culture, CD82 was located at the vesicle-like structures and showed partial co-localization with caveolin-1. In 3D environment liprin-α1 was localized to adhesion or invadosome-like structures or near cell edge, whereas CD82 was located at the cell edge and cell-cell contacts after liprin-α1 silencing suggesting that the localization of liprin-α1 or CD82 are dependent on environment, matrix availability and cell-cell contacts. Liprin-α1 is a part of β1-integrin signaling and recycling [15, 16], while CD82 modulates β1-integrin maturation and integrin-related cell adhesion in cancer cells [31, 32, 33], and is regulated by endocytosis and palmitoylation [34, 35]. In line with this data, caveolin-1, an important scaffolding protein involved in endocytosis and membrane trafficking, is required for CD82 mediated EGFR signaling . Most importantly, we showed in the present work that downregulation of CD82 by siRNAs partially enhanced invasive capabilities of metastatic shPPFIA1 HNSCC cells in 3D collagen providing evidence on the interplay between liprin-α1 and CD82 which is most likely due to the regulative role of liprin-α1 in cell-cell and cell-extracellular matrix contacts.
Gene set enrichment analysis provided several clues for the overall regulatory patterns and genes that are linked to liprin-α1. Interestingly, sortilin-related receptor 1 (SORL1), which likely plays a role in endocytosis and sorting and it has association to Alzheimer’s disease [37, 38], was upregulated in liprin-α1 knockdown cells. SORL1 sorting and quality control occurs in the endoplasmic reticulum (ER)-Golgi complex, and mutations in SORL1 may prevent trafficking of the protein to the cell membrane due to defects in ER quality control . Interestingly, our gene expression and gene set enrichment analysis data revealed that liprin-α1 is involved in the regulation of the protein signaling related to endoplasmic reticulum. In addition, ARRDC3, which is a tumor suppressor and involved in integrin trafficking , was upregulated in breast cancer, whereas FN1 was downregulated after liprin-α1 knockdown. FN1 has been correlated to poorer prognosis  and metastasis formation  while a recent publication shows the involvement of liprin-α1 in fibronectin secretion in endothelial cells . In HNSCC, ASAP3 (alias ACAP4) with a potential role in focal adhesions, β1 integrin recycling and migration [43, 44], was downregulated in liprin-α1 knockdown cells. Another gene related to adhesion, integrin inhibition and suppression of focal adhesion formation, SMURF1 , was upregulated in liprin-α1 knockdown cells, which is also likely to have anti-tumor activity . Furthermore, ANXA2, whose expression is diminished in dysplastic HNSCC  was upregulated in liprin-α1 knockdown cells from HNSCC data.
Our previous data show regulative role of liprin-α1 on vimentin  and indeed VIM was upregulated in HNSCC cells after liprin-α1 knockdown both in 2D and 3D cell culture platforms. Changes in the proteins related to adhesion and protein recycling underlines the important role of liprin-α1 in cell signaling in 3D environment. Intriguingly, both liprin-α1 and vimentin have important roles in invadosome function in cancer cells [12, 13, 48, 49], vimentin being important in stabilizing and in elongation of mature invadopodia . There is evidence that invadosomes are formed in vivo, and the shape of invadosomes vary depending on the cellular microenvironment . We showed that in 3D environment liprin-α1 localized to adhesion- or invadosome-like structures. Taken together, our GSEA data underlined the role of liprin-α1 in the cell edge functions of cancer cells during cell invasion, shown by modifications in CD82 expression as well as by rearrangements in actin cytoskeleton after liprin-α1 silencing.
Studying the translational modifications following liprin-α1 silencing provided us with better understanding on liprin-α1 association to metastatic progression of the cells. Therefore, we next explored how PPFIA1 amplification influences the survival of the patients from HNSCC and breast cancer in clinical data. Interestingly, in The Cancer Genome Atlas data PPFIA1 alteration was identified in 189 (36%) of 528 sequenced head and neck squamous cell carcinoma patients, whereas in breast cancer data set PPFIA1 amplification was identified in 439 (17%) of 2509 sequenced patients [21, 22]. Survival data showed shorter survival for altered patients when compared to patients with no alteration in both the HNSCC and in breast cancer data [21, 22]. Clinical data emphasizes the importance of PPFIA1 amplification in cancer cell progression, and as a potential prognostic marker for HNSCC and breast cancer.
Our study provides novel information regarding liprin-α1 function in cancer cell signaling and strengthens the role of liprin-α1 in regulating cell edge functions during cell invasion. We present novel insight into liprin-α1 dependent regulation of different biological processes and of CD82 which links liprin-α1 to metastatic progression of cancer cells. In conclusion, we show that liprin-α1 controls cell edge protrusions during invasion and metastasis underlining importance of liprin-α1 in cancer cell dissemination and metastatic process of cancer cells.
Biomedicum Functional Genomics Unit (https://www.helsinki.fi/en/infrastructures/genome-analysis/infrastructures/biomedicum-functional-genomicsunit), Biomedicum Imaging Unit (https://www.helsinki.fi/en/infrastructures/light-microscopy/biomedicum-imaging-unit), Genome Biology Unit (https://www.helsinki.fi/en/infrastructures/genomeediting-function-and-stem-cell-platform/infrastructures/genome-biology-unit), University of Helsinki, Finland are thanked for excellent technical support. “The results for PPFIA1 alterations and survival data in HNSCC and breast cancer presented here are based upon data generated by the TCGA Research Network: http://cancergenome.nih.gov/.” Venn diagrams used in this work has been done by Venny, Oliveros J.C. (2007-2015). An interactive tool for comparing lists with Venn Diagrams (http://bioinfogp.cnb.csic.es/tools/venny/index.html). Mr. Mikko Kivento is thanked for performing the RNA sequencing and GSEA data analysis. Mrs. Hanna Pesonen, Mrs. Marika Kuisma, Mrs. Maiju Merisalo-Soikkeli are thanked for excellent technical support.
Academy of Finland, Finnish Cancer Organizations, Sigrid Juselius Foundation, Ida Montin Foundation and Jane and Aatos Erkko Foundation are thanked for their financial support to this work.
Availability of data and materials
RNA-sequencing data is deposited at the Gene Expression Omnibus with a series number GSE108392. The data supporting the conclusions of this study is included within the article and supplementary files.
HP, PvN, KL and OM participated in designing the study. HP, ML and AF generated knockdown cell lines. HP and ML performed RNA-seq experiments, western blots and immunofluorescence stainings. PvN and HP performed invasion assays and immunofluorescence analysis from three dimensional cell culture. RG has provided cell lines and clinical expertise for the study. HP wrote the first draft of the manuscript and OM participated in writing. All authors reviewed and commented on the manuscript, and approved its final submission.
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 1.Jarvinen AK, Autio R, Haapa-Paananen S, Wolf M, Saarela M, Grenman R, Leivo I, Kallioniemi O, Makitie AA, Monni O. Identification of target genes in laryngeal squamous cell carcinoma by high-resolution copy number and gene expression microarray analyses. Oncogene. 2006;25(52):6997–7008.CrossRefPubMedGoogle Scholar
- 3.Akervall JA, Jin Y, Wennerberg JP, Zatterstrom UK, Kjellen E, Mertens F, Willen R, Mandahl N, Heim S, Mitelman F. Chromosomal abnormalities involving 11q13 are associated with poor prognosis in patients with squamous cell carcinoma of the head and neck. Cancer. 1995;76(5):853–9.CrossRefPubMedGoogle Scholar
- 4.Schuuring E, Verhoeven E, Vantinteren H, Peterse JL, Nunnink B, Thunnissen FBJM, Devilee P, Cornelisse CJ, Vandevijver MJ, Mooi WJ, et al. Amplification of genes within the chromosome-11q13 region is indicative of poor prognosis in patients with operable breast-Cancer. Cancer Res. 1992;52(19):5229–34.PubMedGoogle Scholar
- 10.Scholz CJ, Weber H, Jungwirth S, Danielczyk W, Reif A, Tragl KH, Fischer P, Riederer P, Deckert J, Grunblatt E. Explorative results from multistep screening for potential genetic risk loci of Alzheimer's disease in the longitudinal VITA study cohort. J Neural Transm (Vienna). 2018;125(1):77–87.CrossRefGoogle Scholar
- 16.Mana G, Clapero F, Panieri E, Panero V, Bottcher RT, Tseng HY, Saltarin F, Astanina E, Wolanska KI, Morgan MR, et al. PPFIA1 drives active alpha5beta1 integrin recycling and controls fibronectin fibrillogenesis and vascular morphogenesis. Nat Commun. 2016;7:13546.CrossRefPubMedPubMedCentralGoogle Scholar
- 19.Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102(43):15545–50.CrossRefPubMedPubMedCentralGoogle Scholar
- 20.Mootha VK, Lindgren CM, Eriksson KF, Subramanian A, Sihag S, Lehar J, Puigserver P, Carlsson E, Ridderstrale M, Laurila E, et al. PGC-1 alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet. 2003;34(3):267–73.CrossRefPubMedGoogle Scholar
- 21.Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, Jacobsen A, Byrne CJ, Heuer ML, Larsson E et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov 2012;2(5):401–404.Google Scholar
- 24.Xu C, Liu Y, Wang P, Fan W, Rue TC, Upton MP, Houck JR, Lohavanichbutr P, Doody DR, Futran ND, et al. Integrative analysis of DNA copy number and gene expression in metastatic oral squamous cell carcinoma identifies genes associated with poor survival. Mol Cancer. 2010;9:143.CrossRefPubMedPubMedCentralGoogle Scholar
- 30.Jones EL, Wee JL, Demaria MC, Blakeley J, Ho PK, Vega-Ramos J, Villadangos JA, van Spriel AB, Hickey MJ, Hammerling GJ, et al. Dendritic cell migration and antigen presentation are coordinated by the opposing functions of the Tetraspanins CD82 and CD37. J Immunol. 2016;196(3):978–87.CrossRefPubMedGoogle Scholar
- 41.Ioachim E, Charchanti A, Briasoulis E, Karavasilis V, Tsanou H, Arvanitis DL, Agnantis NJ, Pavlidis N. Immunohistochemical expression of extracellular matrix components tenascin, fibronectin, collagen type IV and laminin in breast cancer: their prognostic value and role in tumour invasion and progression. Eur J Cancer. 2002;38(18):2362–70.CrossRefPubMedGoogle Scholar
- 44.Yu X, Wang F, Liu H, Adams G, Aikhionbare F, Liu D, Cao X, Fan L, Hu G, Chen Y, et al. ACAP4 protein cooperates with Grb2 protein to orchestrate epidermal growth factor-stimulated integrin beta1 recycling in cell migration. J Biol Chem. 2011;286(51):43735–47.CrossRefPubMedPubMedCentralGoogle Scholar
- 48.Sala K, Raimondi A, Tonoli D, Tacchetti C, de Curtis I. Identification of a membrane-less compartment regulating invadosome function and motility. Sci Rep 2018;8(1):1164.Google Scholar
Open AccessThis 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.