Cell culture

Human primary melanoma cell lines MCM1, MCM1G, MCM1D and MCM1DLN were generated by Swoboda et al. [14]. WM793B, 1205Lu and WM35 were commercially acquired from ATCC (Manassas, VA). Melanoma cells were screened for mycoplasma contamination once a week and routinely cultured in MIM medium supplemented with 2% FCS (GE Healthcare, Little Chalfront, UK) containing: 80% MCDB153 (Sigma-Aldrich, St. Louis, MO), 20% Leibovitz’s l15 (Mediatech, Tewksbury, MA), 5 µg/ml Insulin, bovine (Sigma-Aldrich), 0.5 ng/ml Epidermal Growth Factor (EGF) (Sigma-Aldrich), 1.68 mM CaCl2 (Sigma-Aldrich), 100 IU/ml Penicillin (Sigma-Aldrich), 100 µg/ml Streptomycin (Sigma-Aldrich) and 2 µg/ml Ciprofloxacin (Sigma-Aldrich).

siRNA transfection

siRNA transfection was carried out using DharmaFECT (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s protocol. Before the day of transfection, melanoma cells were passaged in antibiotic-free 2% MIM and transfected as soon as the cells were 60–70% confluent. 20 µM siRNA (GE Healthcare, Little Chalfront UK) was diluted in 1x siRNA buffer (Thermo Fisher Scientific) to a final concentration of 5 µM. 5 µM siRNA was diluted 1:20 in serum free MIM. In a second reaction batch, DharmaFECT (Thermo Fisher Scientific) was diluted 1:50 in serum free MIM. Both reactions were incubated for 5 min at RT. Subsequently, the two reactions were mixed and incubated for 20 min at RT. Thereafter, the mixtures were added to the cells in a ratio of 1:5 in MIM medium to a final siRNA concentration of 25 nM. Cells were incubated for 24 h at 37 °C in 5% CO2. Hereafter, medium was replaced with 2% MIM. The following siRNAs were used: NLGN4X siRNA (GE Healthcare), VBP1 siRNA (GE Healthcare). Non-targeting siRNA (GE Healthcare) was used as a control in all experiments.

Plasmid transfection

NLGN4X plasmid (Addgene, Cambridge, MA) was mixed with the quadruple amount of PEI (Sigma-Aldrich) for 15 min. This mix was added to prior equally seeded cells in Opti-MEM, Reduced Serum Media (Gibco, Thermo Fisher Scientific, Waltham, MA) when they reached 60–70% confluence.

VBP1 overexpression

Before the day of transfection, melanoma cells were passaged in antibiotic-free 2% MIM and transfected as soon as the cells were 70–80% confluent. On the day of the transfection, the 2% MIM media was replaced with Opti-MEM, Reduced Serum Media and transfection mixes were prepared. These were assembled according to the manufacturer’s protocol using an equimolar mix of guide complex (CRISPR-Cas9 Synthetic tracrRNA : CRISPRmod CRISPRa crRNA Pool Human VBP1, 2.5 μM), 300 ng CRISPRa-EGFP-dCas9-VPR-mRNA (CAS12025) and 7.5 μL Dharmafect Duo (T-2010-01), diluted in 250 μL Opti-MEM. For the conditions receiving siRNA treatment, extra mixes were prepared containing siRNA (20 μM, Horizon) and 7.5 μL Dharmafect Duo diluted in the appropriate quantity of Opti-MEM. The transfection mixes were added to the respective wells and the cells were incubated for 24 h at 37 °C, 5% CO2 before media change. After another 24 h recovery in 2% MIM media, cells were lysed using Laemmli buffer supplemented with 1× Protease and Phosphatase Inhibitors, 1 mM PMSF and 0.5 mM EDTA (Thermo Fisher Scientific, Waltham, MA) and sonication.

Measurement of cell count, metabolic activity and ROS

To measure cell count and LDH, we seeded cells in triplicates in 6 well plates. siRNA transfection was performed directly in the 96 well plates. To measure cell counts of cell cultures, we used the CASY cell counting system (Schärfe System, Reutlingen, Germany). To test the metabolic activity of cells, Alamar Blue assay (Invitrogen, Carlsbad, CA) was carried out as stated in the manufacturer’s protocol. ROS was measured by incubating cells with 20 µM DCFHDA in PBS for 30 min. Subsequently, plates were measured at an excitation wavelength of 488 nm and an emission wavelength of 535 nm every 15 min. Treatment with 10 mM NAC was used as a control.

Spheroid invasion assay

For the spheroid invasion assay, cells were harvested and resuspended in MIM-methylcellulose (consisting of 20% methylcellulose and of 80% MIM). Thereafter, cells were distributed in drops of 100 µl in a round bottom 96 well plate at 2500 cells/drop and the plate was centrifuged at 300 g for 6 min. Spheroids were incubated at 37 °C (5% CO2, 95% humidity) for 96 h. Subsequently, spheroids were harvested and implanted into collagen gels consisting of 2 mg/ml collagen I (Corning, Corning, NY), 1.5% methylcellulose and with NaOH, used for neutralisation. Embedded spheroids were cultivated in 2% MIM and incubated at 37 °C (5% CO2, 95% humidity). Images were taken 24 h later and the invasive area was evaluated using Image J (National Institutes of Health, Bethesda, MD).

Transwell migration assay

For the migration assay, MCM1G and WM793B cells were transfected with NLGN4X and non-targeting siRNA as described. Additionally, subgroups of cells were treated with 5 µM YC-1 (Santa Cruz). 40000 cells in 100 µl serum-free MIM were pipetted into transwells (Sigma-Aldrich) with 8 µm pores placed in 24 well plates filled with 600 µl MIM with 10% serum content. Subsequently, cells were incubated for 6 h (WM793B) or 12 h (MCM1G) at 37 °C and 5% CO2. Next, cells were washed with PBS and fixated with 4% formaldehyde for 30 min at room temperature. Subsequently, cells were stained with 0.1% crystal violet (Sigma-Aldrich) for 10 min. For analysis, 5 photographs (middle, top, bottom, right and left) were taken of each transwell and cells were counted with ImageJ.

Quantitative real-time PCR (qRT-PCR)

Total RNA was isolated by directly adding RNA lysis buffer to the cells. The procedure was performed with peqGold total RNA kit (Peqlab, Erlangen, Germany) according to the manufacturer’s instructions. cDNA was synthesised from 5 µg total RNA using GoScript Reverse Transcription system (Promega, Fitchburg, WI) following manufacturer’s instructions. Quantitative Real-time PCR (SYBR Green) was performed using GoTaq® QPCR Master Mix (Promega), in a StepOnePlus Real-Time PCR Detection System (Thermo Fisher Scientific, Waltham, MA). Cycling conditions: 95 °C for 2 min, 40 cycles at 95 °C for 15 s and 65 °C for 1 min. For each reaction batch, 2 µl of the sample cDNA was mixed with 10 µl GoTaq qPCR Master Mix (Promega), 0.4 µl forward and reverse primer (Eurofins MWG Operon, Luxembourg) and 7.2 µl ddH2O. Primer sequences were as followed: NLGN4X: Forward 5′-CTA CTG CTC CCT GGA AAG CCC TAT- 3′, Reverse 5′-TGA ACA ACA AAG GAA GCC ATA GCA- 3′. VBP1: Forward 5′-TCA CGG AAT CCC GGC GGC-3′, Reverse 5′-TGC AGT CTC ATT CCC AGG CTG TTT C-3′. TXNIP: Forward 5′-CCA GCA ATT GGG GGA AAG AAG GC-3′, Reverse 5′-CTC CAA ATC GAG GAA ACC CCT TTG C-3′. HMOX1: Forward 5′-CCG CAG TCA GGC AGA GGG TG-3′, Reverse 5′-GAG CGG GTG TTG AGT GGG GG-3′. β-Actin: Forward 5′-CTA TCC AGG CTG TGC TAT CCC TGT-3′, Reverse 5′-CCT TAA TGT CAC GCA CGA TTT CC-3′. All experiments were done in triplicates. For normalisation, β-Actin was used and relative gene expression was calculated using the comparative Ct method (2-ΔΔCt). Reaction batches were done in duplicates.

Western blotting

For protein isolation, cells were directly lysed in whole cell extraction buffer RIPA (Cell Signalling, Danvers, MA) containing 1 mM PMSF (Sigma-Aldrich, St. Louis, MO) and 1% PIC (Sigma-Aldrich). Samples were centrifuged for 20 min at 16,000 rcf at 4 °C. Subsequently, supernatant was collected and stored at −80 °C. Protein concentration was measured using the BCA Protein Assay Kit (Cell Signalling) according to manufacturer’s instructions. 20–40 µg proteins per sample were denatured at 95 °C for 5 min and loaded onto a TGX Stain-FreeTM FastCastTM Acrylamide gel (Bio-Rad, Hercules, CA). Subsequently, the separated proteins were electroblotted onto an LF-PVDF or nitrocellulose membrane. The following primary antibodies were applied for immunodetection as follows: rabbit polyclonal anti-NLGN4X (Sigma-Aldrich, HPA001651, 1:1000, previously used by Davidson et al. [15]), rabbit polyclonal anti-VBP1 (Sigma-Aldrich, HPA023230, 1:1000), mouse monoclonal anti-HIF1A (Santa Cruz, Dallas, TX, sc53546, 1:1000), rabbit monoclonal anti-TXNIP (Abcam, Cambridge, UK, EPR14774, 1:1000), rabbit monoclonal anti-pH2AX (Cell Signalling, 9718, 1:1000), mouse monoclonal anti-HA-Tag (Cell Signalling, 2367, 1:1000). On the following day, membranes were incubated with HRP-linked heavy and light chain antibodies (Anti-Mouse IgG: Thermo Fisher Scientific, Waltham, MA, 31430, 1:5000, Anti-Rabbit IgG: Thermo Fisher Scientific, 31460, 1:5000). Pierce® ECL Western Blotting Substrate (Thermo Fisher Scientific) was used for detection. Total protein was used for normalisation.

Microarray analysis and GSEA of NLGN4X

Isolated mRNA samples were sent to the Core Facility Genomics (Medical University of Vienna, Vienna, Austria) where whole transcript expression profiling using GeneChip Human Gene 1.0 ST Arrays (Affymetrix, Santa Clara, CA) was conducted. RMA [16] was used for normalisation and limma [17] for inferring differential expression between groups. Expression Console and Transcriptome Analysis Console (Affymetrix) were used for differential gene expression analysis. GSEA was performed using the GSEA tool offered by the Broad Institute (Cambridge, MA). ‘Hallmark Gene Sets’, ‘Curated Gene Sets’, ‘GO Gene Sets’ and ‘Oncogenic Signatures’ were used as gene set universes. Permutation type was set to gene set and analysis was performed with 1000 permutations. The data generated by this study has been deposited in NCBI’s Gene Expression Omnibus [18] and is accessible through GEO Series accession number GSE96632.

Patient cohort and pathology

With institutional review board approval from the Medical University of Vienna (EK Nr: 1101/2015), tissue slides and microarrays were obtained from the Biobank Graz (Graz, Austria) and US Biomax (Rockville, MD). All samples were formalin-fixed less than 10 min after surgery, paraffin embedded and assembled in tissue arrays as cores of 1.5 mm in diameter or whole tissue sections. Tissue arrays and sections were quality controlled for melanoma tissue, representing different stages of disease progression. Each individual core was assigned to pathohistological characteristics and was reviewed by two board certified pathologists.

Generation of NLGN4X overexpression lines

Tumour lines were infected (MOI:15) with a lentivirus carrying the fusion construct NLGN4X-mGFP within their expression plasmid (Origene, RC222932L2). Control cell lines were infected (MOI:5) with a lentivirus carrying a mGFP control particle within their expression plasmid ORF (Origene, PS100071V). One week post infection cells were sorted for GFP expression (Sony, MA900 Multi-Application Cell Sorter). To ensure 100% GFP expression, ten individual clones were picked from each line and, after GFP signal testing, pooled again.

Organoid culture

Skin organoids were generated using human embryonic stem cells WA19 (WiCell CVCL_9780) of early passages (p16–19) after the protocol of Lee et al. [19]. The only modification to the protocol was using 5000 cells per well for the initial seeding.

Spheroid formation assay

Metastatic tumour lines (controls and re-expressing NLGN4X) were trypsinized and counted and a 10.5 mL cell suspension was prepared for a 96-well plate. This contained 150000 cells (1500 cells per well) in spheroid culture medium (80% vol DMEM with 5% FCS and 1% L-glutamine and 20% methycellulose 33 mM). 100 μL of this suspension was seeded in each well of a TC non-treated, round bottom 96-well plate (SPL Life Sciences). The plate was centrifuged for 10 min at 300 g and then placed in an incubator at 37 °C, 5% CO2, with no shaking. After 48 h tumour spheroids formed and the spheroid medium was replaced with OMM (organoid maturation medium containing 49.5% vol Advanced DMEM/F12, 49.5% Neurobasal medium, 1% Organoid Matrigel Corning, 100ug/mL Normocyn, 0.1 mM B-Mercaptoethanol, 0.5 × N2 supplement, 0.5 × B27-VitA supplement, 1 × Glutamine).

Spheroid transplantation

Human skin organoids that were 60 days old were placed on top of 72 h old tumour spheroids in 96 well U-bottom low-attachment plates (SPL Life Sciences). They remained in direct contact for 48 h in OMM at 37 °C, 5% CO2, with no shaking. After this time point, when the tumour spheroid attached, they were transferred to 24 well ultra-low attachment plates (Corning CLS3473-24EA) in OMM at 37 °C, 5% CO2, shaking (65 rpm) and cultivated for 12 days. Media change was performed every second day. The organoids with attached tumour spheroid are henceforth referred as tumour organoids.

Immunohistochemistry and Immunofluorescence

Tumour organoids were fixed in 4% paraformaldehyde, washed with 1 × PBS and embedded into 1% agarose molds, processed into paraffin blocks and sectioned. Melanoma tissue arrays and slides, containing paraffin-embedded samples, were melted for 20 min at 60 °C and rehydrated by subsequent incubation in Xylol, Isopropanol, 96% Ethanol, 70% Ethanol and 50% Ethanol. Then, arrays were washed and heated up to 120 °C in a pH 6.0 buffer or a pH 9.0 buffer (Dako, Glostrup, Denmark), depending on the antibody. After cooling to room temperature, samples were incubated with 1% H2O2 (Sigma-Aldrich, St. Louis, MO) for 10 min. Afterwards, samples were permeabilized with 0.1% TritonX-100 (Sigma-Aldrich) for 5 min. Then, sections were blocked with 2.5% horse sera (Vector Laboratories, Burlingame, CA) for at least 20 min at room temperature. Subsequently, sections were incubated overnight at 4 °C with the primary antibodies directed against anti-NLGN4X (Abcam, EPR13108, 1:2000), rabbit polyclonal anti-VBP1 (Sigma-Aldrich, HPA023230, 1:400) and S100β (Ab52642, Abcam, 1:350, pH 6). On the next day, slides were washed and either secondary biotinylated antibodies (Vector Laboratories), or corresponding fluorescent-labelled antibodies, were added for 45 min at room temperature. After a washing step, sections for immunohistochemistry were incubated for 30 min with Streptavidin-HRP (Leica, Wetzlar, Germany). For detection, arrays were incubated with AEC+ High Sensitivity Substrate Chromogen (Dako). Counterstaining with hematoxylin solution was performed according to Mayer (Carl Roth, Karlsruhe, Germany); arrays were mounted with Aquatex® (Merck Millipore, Billerica, MA).

Evaluation of immunohistochemical staining

Evaluation of tissue arrays and slides was performed by two independent researchers who were blinded regarding patient details. Immunostaining of the anti-NLGN4X antibody was scored on at least duplicate tissues by using the following arbitrary scale: no staining (0), low staining (1), medium staining (2) and high staining (3).

In addition to the observer-dependent scoring system, we used an computational-supported approach previously established by us [20]. In brief, tissue slides were scanned using a microscopic panoramic slide scanner (3DHistech, Budapest, Hungary). Image J was used to perform colour deconvolution to extract NLGN4X staining from haematoxylin staining and melanin pigment. On the base of these single colour images, automated analysis in ImageJ was used to assign each core an overall staining intensity score by calculating the percentage of stained area and normalising it to the total area of the core.

Bioinformatics

For mRNA microarray analyses of the MCM1 melanoma cell system data was generated by Swoboda et al. [14], accessible through GEO Series accession number GSE36035. For Kaplan Meier analyses on overall survival time, we analysed mRNA microarray files from Bogunovic et al. [21], accessible through GEO Series accession number GSE19234, and the TCGA consortium [22].

Statistics

For bar diagrams, standard error of the mean (SEM) and two-tailed P values were calculated by performing unpaired (independent) Student’s or Welch’s t test in SPSS v21 (IBM, Armonk, NY). Levene’s test with a threshold of 0.05 was performed to choose the appropriate t-test. For multiple comparisons, One-way ANOVA was calculated. Tukey HSD was used for post-hoc analysis. For line diagrams, standard deviation (SD) was calculated in GraphPad Prism 6 (GraphPad Prism Inc., La Jolla, CA). Two-tailed P values were calculated by performing unpaired (independent) Student’s or Welch’s t-test in SPSS to compare individual time points and the areas under the curve (AUC). Levene’s test with a threshold of 0.05 was performed to choose the appropriate t-test. For multiple comparisons, One-way ANOVA was calculated.

For scatter plots Pearson correlation analysis was applied to calculate corresponding R- and p values in SPSS. Kaplan Meier plots analysing effects of NLGN4X expression were generated by classifying the samples into groups of low and high target gene expression according to the median gene expression value (GSE19234) or the NLGN4X staining intensity score (Austrian cohort) in SPSS. Kaplan Meier plot analysing combined effects of NLGN4X and VBP1 expression and corresponding box plots were generated by classifying the samples into groups of low and high risk groups according to the median by the web-based tool SurvExpress [23] and edited in SPSS. To test for the significance of Kaplan Meier analysis, we performed the log-rank (Mantel-Cox) test. Two tailed P values for box plots were calculated by performing unpaired (independent) t-test. For analysis of immunohistological staining results, 2 × 2 contingency tables and Fisher’s exact test was applied to determine association of clinicopathological parameters with NLGN4X expression in SPSS.