Retroviral Expression Vectors and Infection of Keratinocytes

The generation of the pLXSN-based retroviral vectors coding for HPV8 E1, E2, E6, E7, or E6E7 has been previously described (Akgül et al., 2005, 2007; Leverrier et al., 2007). Production of retroviruses and retroviral transduction of N/TERT and PM1 keratinocytes were performed as previously described (Hufbauer et al., 2013). Briefly, keratinocytes were seeded out in 6 cm dishes. Next day, retroviral supernatants were mixed with an equal volume of DMEM in the presence of 5 μg/mL of hexadimethrine bromide (polybrene) and added to the keratinocytes. Spin infection was made by centrifugation for 1 h at 300 × g. Cells were then washed with PBS and cultured further for 2 days. Then, cells were selected against G418 (500 μg/mL) until only infected keratinocytes had survived. In subsequent experiments we used pooled stable cell populations to minimize possible variations due to the randomness of the viral integration site in the cellular chromosomes.

RTqPCR and PrimePCR

To quantify mRNA levels of cellular genes, quantitative reverse transcription-PCR (RTqPCR) using the LightCycler system (Roche, Mannheim, Germany) was performed as previously described (Hufbauer et al., 2015). The primers used for this study were:

ATF3-fw: TTTGCCATCCAGAACAAGC;

ATF3-rev: CATCTTCTTCAGGGGCTACCT;

GADD34-fw: GCTTCTGGCAGACCGAAC;

GADD34-rev: GTAGCCTGATGGGGTGCTT;

GDF15-fw: CCCGGGACCCTCAGAGTT;

GDF15-rev: CCGCAGCCTGGTTAGCA;

HPRT1-fw: TGACACTGGCAAAACAATGCA;

HPRT1-rev: GGTCCTTTTCACCAGCAAGCT.

The RTqPCR based PrimePCR (Fibronectin-Binding Integrin in Cell Motility H96, BIORAD, Feldkirchen, Germany) was performed in quadruplicate using 96-well plates with pre-spotted primers in each well of the plate. The plates were designed to be analyzed using a Roche LC 480 light cycler, using the SsoAdvanced^TM Universal SYBR^® Green Supermix. Analyses of these plates were carried out following manufacturer’s instructions.

DNA Microarrays

The cDNA microarray analyses performed in the Akgül group used total RNA extracted from monolayer cultures of PHAKs in total passage 3 (passage 2 after transduction). The RNA was analyzed in biological triplicates from cells containing either the empty vector control pLXSN or pLXSN-HPV8-E7. Changes in gene expression were measured using the Human Genome U133A arrays according to Affymetrix protocols. Microarray data were analyzed using Transcriptome Analysis Console software (Thermo Fisher Scientific, v4.0.1.36). For analysis of differential gene expression, we applied the following conditions to determine statistically relevant genes: (a) ANOVA test: ebayes, (b) a fold change cut-off of < –2 or > 2, and (c) determination of statistical significance with p < 0.05. Gene ontology analysis was carried out using the PANTHER (Protein ANalysis THrough Evolutionary Relationships) database (v14.1). These microarray data files have been uploaded to the Gene Expression Omnibus (GEO) database (GSE133813) (Kirschberg et al., 2019).

The cDNA microarray analyses performed in the Alonso group at the DKFZ were performed with RNA isolated from primary keratinocytes also transduced with pLXSN or pLXSN-HPV8-E7 coding retroviruses. The Affymetrix chip-based experiments and data analyses were performed at the core facility of the German Cancer Center (DKFZ) (Supplementary Table 1). Differentially expressed genes (DEG) from both independent studies were overlapped to minimize bias of the data in respect to genetic variations among keratinocyte donors, cell culture media effects, thus making the results more robust.

ELISA

The Human GDF15 Quantikine ELISA Kit (RnD systems) was used to measure GDF15 in cell culture supernatants. For this N/TERT-8E7 were cultured on fibronectin and cell culture supernatants were harvested after 4 days. Collected supernatants were treated with protease/phosphatase inhibitors (Protease/Phosphatase Inhibitor Cocktail, Cell Signaling). The ELISA was performed according to the manufacturer’s protocol.

Western Blotting and Phospho-Kinase Array Analysis

Western blotting was performed as previously described (Marx et al., 2018). Briefly, cell pellets were resuspended in radioimmunoprecipitation assay (RIPA) buffer containing protease inhibitors, incubated on ice for 30 min, sonicated and then centrifuged at 15,000 × g at 4°C for 15 min. Protein concentrations were measured with the Pierce^TM BCA Protein Assay Kit (Thermo Fisher Scientific, Dreieich, Germany). SDS gels were transferred upon completion to nitrocellulose membranes, which were then blocked with either 5% bovine serum albumin or 5% dry milk in TBST [10 mM Tris/HCl (pH 8.0), 150 mM NaCl, 0.05% Tween 20]. The primary antibodies used targeted 14-3-3β (sc-59419, clone 60C10, Santa Cruz), ERK1/2 (#4695, clone 137F5, Cell Signaling), Fyn (#4023, Cell Signaling), p-Tyr⁵³⁰-Fyn (#PA5-36644, Invitrogen), Lyn (#4576, clone 5G2, Cell Signaling), p-Tyr³⁹⁷-Lyn (#orb6340, Biobyrt, Cambridge, United Kingdom), GAPDH-Alexa680 (sc-365062, cloneG9, Santa Cruz). HuT78 whole cell lysates (sc-2208, Santa Cruz) were used as positive controls. Antibodies were incubated for 2 h at RT or overnight at 4°C. Following washing, membranes were incubated with corresponding secondary antibodies (IRDye, LI-COR, Lincoln, Nebraska, United States). Western blots were visualized by employing the LI-COR ODYSSEY FC Imaging system with the Image Studio Ver 5.2. software.

The Human Phospho-Kinase Array Kit (RnD Systems, Minneapolis, United States), spotted with antibodies targeting phospho-sites in kinases associated with malignant processes, was used to identify potentially deregulated kinases in HPV8-E7 expressing keratinocytes grown for 96 h on fibronectin coated tissue culture plates. Cells were rinsed with PBS and then lysed using the provided buffer supplemented with 10 μg/mL Aprotinin, 10 μg/mL Leupeptin, and 10 μg/mL Pepstatin. The lysates were gently incubated at 4°C for 30 min and then centrifuged at 14,000×g for 5 min, before transferring the supernatants into a clean test tube. Samples were then further processed according to the manufacturer’s protocol. Visualization of the membranes and evaluation of individual spot densities were carried out employing the LI-COR ODYSSEY FC Imaging system with the Image Studio Ver 5.2. software.

Proteomics

Proteomic and phospho-proteomic analyses were performed with total cell extracts from HPV8-E7 expressing N/TERT keratinocytes (N/TERT-8E7) grown for 96 h on fibronectin coated tissue culture plates which were lysed using 8M urea supplemented with protease and phosphatase inhibitors. The whole cell lysates were prepared in triplicates in two separate experiments (n = 6) were analyzed at the proteomics core facility of the Cologne Excellence Cluster on Cellular Stress Responses in Aging-Associated Diseases (CECAD) by means of “stable isotope labeling with amino acids in cell culture” (SILAC)-based mass-spectrometry. The protein profile was submitted to the Proteomics Identifications Database (PRIDE) and is available via ProteomeXchange with the identifier PXD026099. Briefly, samples were reduced with DTT (100 mM) and alkylated with 550 mM Iodoacetic acid (IAA) prior to digestion with trypsin and Lys-C overnight at 37°C. Generated peptides were extracted by incubation with increasing amount of acetonitrile, concentrated in a speed-vac and primed prior to LC-MS/MS analysis by the STAGE tip technique (Rappsilber et al., 2003). For liquid chromatography and tandem mass spectrometry (LC-MS/MS), an easy nLC 1,000 (Thermo Fisher Scientific) was coupled to the quadrupole-based Q Exactive Plus (Thermo Fisher Scientific) instrument by a nano-spray ionization source. Peptides were separated on a 50 cm in-house-packed column by a two-solvent buffer system: buffer A (0.1% formic acid) and B (0.1% formic acid in acetonitrile). The content of buffer B was increased from 7 to 23% within 40 min and followed by an increase to 45% in 5 min and a washing and re-equilibration step before continuing with the injection of the next sample. Proteins were considered differentially expressed or phosphorylated under the following criteria: (a) log₂fold change < –2.5 or > 2.5, (b) p-value of p < 0.05, and (c) same phosphorylation status in all data sets. Data analysis was performed using the Perseus software suite, KinasePhos2.0 and ShinyGO (v0.61). The differentially phosphorylated protein targets of E7 were imported into the STRING database (Szklarczyk et al., 2019) to identify relevant protein-protein interactions and functional enrichment networks. STRING holds physical as well as functional interaction information derived from high-throughput experimental data, published literature, and computationally predicted co-expression studies.

For secretomic analyses the supernatants of N/TERT-8E7^wt, N/TERT-8E7^L23A as well as empty vector control grown for 96 h on fibronectin coated tissue culture plates (n = 6) were harvested and treated with protease/phosphatase inhibitors (Protease/Phosphatase Inhibitor Cocktail, Cell Signaling). Afterward, the supernatants were concentrated using Amicon ultra-15 centrifugation 3K tubes (Merck Millipore, Darmstadt, Germany). Concentrated supernatants as well as a media-only control were then further processed at the CECAD proteomics facility and subjected to mass-spectrometric analysis. The profile of secreted proteins was submitted to the PRIDE Database and is available via ProteomeXchange with the identifier PXD026100.

HPV8-E7 Mediated Alterations in Fibronectin Associated Signaling Pathways

Since we already knew that the ECM protein fibronectin is key in the regulation of cell fate and also has an important effect on invasive behavior of HPV8-E7 positive keratinocytes (Heuser et al., 2016), we first investigated how HPV8-E7 may alter the secretome of keratinocytes relevant for a crosstalk between epidermal keratinocytes and the ECM. Here, we cultivated both the N/TERT-8E7wt, N/TERT-8E7^L23A (which codes for the invasion deficient mutant) as well as the empty vector control on a fibronectin matrix and collected the cell culture supernatants after 96 h. In the mass-spectrometric analysis we unveiled 11 proteins that were exclusively differentially secreted in N/TERT-8E7 cells compared to control and N/TERT-8E7^L23A (Figures 2A,B and Supplementary Table 2). A further GO enrichment analysis pointed toward most of the identified proteins (except for GALNT2 and UGP2) to be involved in processes associated with cell growth and proliferation (Figure 2C). Interestingly, 14-3-3 family members, which are known to be scaffolds that facilitate a multitude of protein-protein interactions, were present in these pathways and may therefore represent an important regulatory axis. In order to check if the reduced quantity of 14-3-3 in the secretome was the result of decreased protein expression we next performed Western blots with cell extracts of the same cells. Here, we could again show that the presence of HPV8-E7 had a suppressive effect on 14-3-3 β expression (Figure 2D).

Open in a separate window

FIGURE 2

Secretome analysis of N/TERT-8E7^wt or N/TERT-8E7^L23A grown on fibronectin. (A) Venn diagram depicting a three set comparison of differentially secreted proteins taken from N/TERT-pLXSN, N/TERT8-E7^wt, or N/TERT8-E7^L23A. The starred number indicates proteins targeted only by HPV8-E7^wt. (B) Description of differentially proteins exclusively secreted by HPV8-E7^wt positive keratinocytes as depicted in subfigure A The log2-fold changes of the secretome results from N/TERT-8E7 and N/TERT-pLXSN were compared to calculate the difference. (C) Gene ontology analysis of differentially secreted proteins using ShinyGO v0.61. (D) Representative immunoblots of RIPA extracts from N/TERTs expressing HPV8-E7 tested for 14-3-3β protein expression (n = 3). HuT78 whole cell lysate served as a positive control. Equal loading was confirmed by immunoblotting for GAPDH.

To narrow down on gene expression changes of known fibronectin-associated genes in HPV8-E7 positive keratinocytes we next performed PrimePCRs using the Fibronectin-Binding Integrin in Cell Motility H96 assay. These RTqPCRs were performed in quadruplicate, with RNA derived from both control as well as HPV8-E7 positive N/TERT grown on fibronectin coated cell culture dishes for 96 h. The PrimePCR revealed four genes that were more than 15-fold up- or down regulated, namely PIK3R2 (phosphoinositide-3-kinase regulatory subunit 2), Fyn as well as RhoA (ras homolog family member A) and Cfl1 (cofilin 1), respectively (Figure 3A). Since Src kinase family members, like Fyn, are known to be key modulators of cancer cell invasion and metastasis including the skin (Serrels et al., 2009), we next performed Western blots and could confirm higher protein levels and observed four major splice variants. Interestingly, Fyn phosphorylation at pTyr530, which is a negative regulatory tyrosine residue for this kinase (Qayyum et al., 2012) was weaker in E7 cells compared to the empty vector control (Figure 3B).

Open in a separate window

FIGURE 3

Analysis of fibronectin associated gene expression and protein kinase phosphorylation changes in HPV8-E7 positive keratinocytes. (A) PrimePCR results for N/TERT-8E7 grown on fibronectin for 96 h (n = 4). (B) Representative immunoblots of RIPA extracts from N/TERT-8E7 tested for total Fyn and p^Tyr530-Fyn (n = 3). HuT78 whole cell lysates served as a positive control. Equal loading was confirmed by immunoblotting for GAPDH. (C) Protein kinase phosphorylation patterns in N/TERT-8E7 cells cultured on fibronectin for 96 h were analyzed using the Proteome Profiler Human Phospho- Kinase Array Kit (RnD Systems). The cut-off value for up- and down-regulated genes was set to fivefold. The dashed line indicates the 15-fold change. (D) Representative immunoblots of RIPA extracts of N/TERT-8E7 tested for total Lyn and p^Tyr397-Lyn (n = 3). HuT78 whole cell lysates served as a positive control. Equal loading was confirmed by immunoblotting for GAPDH.

This observation sparked the idea to have a deeper look at differentially phosphorylated kinases in HPV8-E7 positive keratinocytes. We therefore used a phosphokinase array kit (Proteome Profiler^TM Array) with total cell extracts from N/TERT-8E7 grown on fibronectin-coated tissue culture plates for 96 h. Based on the obtained experimental data we discovered that now particularly the Src kinase family member Lyn seemed to be strongly hyperphosphorylated in HPV8-E7 positive cells (Figure 3C). Subsequent Western blot analyses proved that Lyn was indeed hyperphosphorylated at residue Tyr397 in HPV8-E7 positive keratinocytes whilst total protein levels remained unchanged (Figure 3D). Taken together, our findings show that Src-kinases are deregulated in HPV8-E7 positive keratinocytes grown on fibronectin.

N/TERT keratinocytes were kindly provided by the James Rheinwald, Harvard Medical School, Boston, MA, United States.