Analysis of hedgehog signaling in periocular sebaceous carcinoma

This study obtained national ethics committee approval (REC reference 14/NW/1080) and adhered to the tenets of the Declaration of Helsinki. The research was designed as a non-paired observational, case control study whereby SC (the case) is compared to nBCC (the Hh tumor control). Fifteen patients with SC were included for comparison against 15 nBCC patients with a mean age of 68 and 73, years respectively. All of the SC patients had non-metastatic disease, ten nodular and five pagetoid, requiring wide excision or exenteration as treatment. Only clearly defined SC or BCC on histology were included using the Royal College of Pathologists, UK dataset [27, 28]. The presence of any non-nodular BCC subtype, and poorly differentiated SC where it was difficult to identify adnexal sebaceous lineage, were also excluded from the study. In addition, exclusion criteria included SC or BCC individuals with metastasis and/or recurrence. Subjects were seen at one center, Moorfields Eye Hospital.

Hematoxylin and eosin (H&E) staining was performed using standard methods. 3–3′-diaminobenzidine (DAB) immunostaining of 5 μm formalin-fixed, paraffin-embedded (FFPE) tissue sections was performed using a panel of commercially available antibodies for the Hh pathway proteins including PTCH1, SMO, Gli1, and Gli2 (Abcam, Cambridge, UK). Conditions for individual antibodies are summarized in Supplementary Table S2. Tissue sections were mounted on slides, deparaffinized in xylene and then alcohol, followed by treatment with the DAKO EnVision™FLEX+ System (Dako, Glostrup, Denmark) [29]. Briefly, antigen retrieval was performed using the DAKO PT LINK machine at 97°C for 20 min under alkaline (pH 9.0) or acidic (pH 6.0) conditions (Table S1). Endogenous peroxidase activity was blocked using EnVision™FLEX peroxidase-blocking reagent. Tissue was incubated with primary antibody for 30 min, followed by incubation with a secondary antibody (Envision™FLEX/HRP LINKER; 20 min) and then visualization of the primary-secondary reaction with EnVision™FLEX substrate working solution containing buffer and DAB (10 min). Sections were sealed and covered using the automated Thermo Clearview Coverslipper (Thermo Fisher Scientific Inc., Waltham, MA, USA). Positive control tissue included breast, brain, testes, and intestine for PTCH1, SMO, Gli1, and Gli2, respectively. Negative controls involved all steps excluding the primary antibody. Finally, sections were counterstained with hematoxylin using the Gemini autostainer (Thermo Fisher Scientific Inc., Waltham, MA, USA).

Immunofluorescence was used for semi-quantitative analysis of Hh pathway expression in SC in comparison to nBCC. FFPE sections were cut (5 μm) and antigen retrieval carried out as described above. Blocking of the sections occurred with 5% goat serum for 1 h followed by primary antibody staining as mentioned. Secondary antibody staining utilized AlexaFluor-568 (Invitrogen™, Life Technologies, Thermo Fisher Scientific Inc., Waltham, MA, USA). Nuclei were counterstained with 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI; Sigma-Aldrich® MO, USA) at 1:1000 dilution to highlight the number of cells present and aid accurate semi-quantification. Coverslips were placed using VECTASHIELD® mounting medium (Vector Laboratories Inc., Burlingame, CA, USA). Negative controls omitted the primary antibody.

Immunofluorescence sections were examined using the Zeiss LSM710 Meta confocal laser microscope (Carl Zeiss Microscopy GmbH, Jena, Germany) and a ZEN configuration tool (Carl Zeiss Microscopy GmbH, Jena, Germany), which is digital image-processing software that produces an output TIFF image. Negative control slides were used to remove background fluorescence and set up the acquisition conditions that subsequently remained constant throughout the imaging process. Images were taken at 200× magnification. Output TIFF images were analyzed using ImageJ (http//imagej.nih.​gov/​ij) to quantify antibody expression. For semi-quantification of antibody expression, fluorescence intensity was determined in regions of interest (ROIs), ensuring a standardized area size whilst containing the same number of nuclei (averaging 18 cells) as determined by the DAPI staining. Tumor was defined as within the bulk of the tumor and at least 1 mm from the edge. Stroma was defined as non-cancerous surrounding tissue within 2 mm of the tumor edge. Furthermore, three separate ROIs were taken for each area in each sample (i.e., three tumor ROIs and three stroma ROIs in nBCC or SC samples) to obtain a mean signal for comparison.

Hh pathway antibodies were validated against protein lysates from LnCaP cells ectopically expressing Gli1 with resulting upregulation of the key effectors of Hh signaling [30]. Negative control tissue was human wild-type fibroblasts. Pre-cast Any kD™SDS-polyacrylamide gel (Bio-Rad, Boston, MA, USA) was used to separate proteins. Twenty micrograms of protein was heated for 5 min at 95°C and mixed with 5X loading dye (100 mM Tris-GCL pH 6.8, 4% SDS; 200 mM dithiothreitol, 0.2% bromophenol blue, and 20% glycerol). Exact volume varied in order to get 20 μg of protein per lane. DualColor Prestained Protein Ladder (Bio-Rad) was used for the identification of protein size. The gel was placed into a mini-PROTEAN tetra cell system (Bio-Rad) and run for 1 h and 30 min at 60 V in 1× Nupage MES SDS running buffer (Life Technologies, Thermo Fisher Scientific Inc., Waltham, MA, USA). Protein was transferred onto an Immun-Blot™ PVDF membrane (Bio-Rad) using a Trans-Blot SD semi-dry transfer cell (Bio-Rad) and Nupage Transfer Buffer (Life Technologies) at ~8 V for 1 h. Subsequently, the membrane was blocked overnight at 4 °C in blocking solution (0.4% PBS/Tween (PBST), 5% dry milk). The following morning, the membrane was washed five times in PBST and then incubated for 2 h at room temperature with the antibody of interest (1:500 anti-PTCH, 1:200 anti-SMO, 1:1000 anti-Gli1, and 1:500 anti-Gli2). The membrane was then washed five times in PBST for 5 min, and incubated with 1:10,000 conjugated secondary anti-rabbit IgG horseradish peroxidase (HRP) (Dako, Glostrup, Denmark) for 1 h at room temperature. A further five washes of PBST for 5 min before the use of an ECL™ Prime Western blotting detection system kit (GE Healthcare, Buckinghamshire, UK). All blots were imaged under the same conditions; signal was detected using the ChemiDoc MP Imaging system (Bio-Rad) and ImageLab software (Bio-Rad). Membranes were stripped of primary antibody using Restore™ Western Blot Stripping Buffer (Thermo Fisher Scientific, Inc) at 55 °C for 30 min, washed five times for 5 min in PBST, and reprobed with 1:5000 polyclonal anti-β-actin antibody (Sigma-Aldrich, MO, USA) as a loading control for each sample.

Recorded immunofluorescence signal data for each sample is expressed as a mean ± SEM. Comparison between SC, nBCC, and normal groups was made using the Mann–Whitney U test. Critical values were identified for a p value of less than 0.05 and 0.01 and a U value less than these were considered to be significant.