SDHD promoter mutations are rare events in cutaneous melanomas but SDHD protein expression is downregulated in advanced cutaneous melanoma

Background SDHD promoter mutations were reported in 4–10% of cutaneous melanomas. The advanced clinico-pathological and patient survival association with SDHD mutation and/or expression in cutaneous melanoma remains controversial. Objectives To evaluate the presence of SDHD promoter mutations and SDHD protein expression in a melanoma series and its possible association with prognosis and survival of the patients. Methods We assessed SDHD promoter status in cutaneous melanomas (CM), ocular melanomas (OM) and melanoma cell lines, and the expression of SDHD protein by immunohistochemistry in CM and OM, and by western blot in melanoma cell lines. We explored the putative association between SDHD protein expression and clinico-pathological and prognostic parameters of melanoma. Results We detected 2% of SDHD promoter mutations in CM, but none in OM and cell lines. SDHD protein expression was present in all CM, in OM and in all CM and OM derived cell lines analysed. A significant association between lower SDHD mean protein expression and presence of ulceration and higher pT stage was found. Conclusions SDHD promoter mutation seems to be a rare event in CM but SDHD lower expression might associate with worst prognostic features in CM.


Objectives
To evaluate the presence of SDHD promoter mutations and SDHD protein expression in a melanoma series and its possible association with prognosis and survival of the patients.

Methods
We assessed SDHD promoter status in cutaneous melanomas (CM), ocular melanomas (OM) and melanoma cell lines, and the expression of SDHD protein by immunohistochemistry in CM and OM, and by western blot in melanoma cell lines. We explored the putative association between SDHD protein expression and clinico-pathological and prognostic parameters of melanoma.

Results
We detected 2% of SDHD promoter mutations in CM, but none in OM and cell lines. SDHD protein expression was present in all CM, in OM and in all CM and OM derived cell lines analysed. A significant association between lower SDHD mean protein expression and presence of ulceration and higher pT stage was found.

Sample selection, clinical-pathological and prognostic parameters
Formalin-fixed, paraffin-embedded tissues from 107 CM and 35 OM (29 uveal and 6 conjunctival melanomas) were retrieved from the Department of Anatomic Pathology of the Hospital S. João, Porto, and of Hospital S. Marcos, Braga. Clinico-pathological (Tables 1 and 2). Followup data were obtained from the patients' records, the Oncology Registries of Hospital S. João and of Hospital S. Marcos, and from RORENO (Oncology Registry of North Region). All cases were revised and staged according to the 7 th edition of AJCC [11]. In CM, follow-up data included time of recurrences and metastases (disease-free survival; DFS) (n = 96) and death due to melanoma (overall survival; OS) (n = 105). The mean follow-up time of the patients for DFS was 52 months (SE±3.94, range 1-195) and for OS was 56 months (SE±3.82, range 1-207). This work was approved by the Local Ethical Committee (CES) and was in accordance with the National ethical rules and Helsinki declaration.

DNA extraction
Extraction of DNA from tumours smaller than 5mm was performed after microdissection with PALM MicroLaser Systems (PALM, Germany) and using the Quiamp DNA micro kit (Quiagen, Hilden). In tumours larger than 5mm, DNA extraction was done by manual dissection of 10μm whole sections of paraffin-embedded tissue using the Invisorb spin tissue mini kit (Invitek, Berlin). DNA extraction from the cell lines was also performed with the Invisorb spin tissue mini kit (Invitek, Berlin).

Mutation analysis
Fragments encompassing SDHD promoter region were amplified by polymerase chain reaction (PCR) of the tumour samples with the sets of primers: Fwd: 5'-CTCCGCCATTGTTCG CCTCA-3', Rev: 5'-TTCCTGAGGGCTCAAGGTCAT-3'. Genomic DNA (25-100 ng) was amplified by PCR using the following cycling conditions: 30s at 94˚C, 90s at 59˚C and 60s at 72˚C for 35 cycles. Products were enzymatically purified and directly sequenced in an ABI Prism 3130 xl Automatic sequencer (Perkin-Elmer, Foster City, CA) using the BigDye Terminator Sequencing kit (Perkin-Elmer). Cases with mutations were confirmed by an independent amplification. BRAF, NRAS, GNAQ and TERT mutational analysis in the series was previously reported [16,28,32].

Immunohistochemical analysis
Paraffin sections were deparaffinised and rehydrated, followed by a microwave antigen retrieval procedure with 10 mM sodium citrate buffer pH 6.0. The sections were incubated overnight at 4˚C in a humidified chamber with the primary antibody SDHD (polyclonal, rabbit, 1:100), from Santa Cruz Biotechnology. The antibody was validated by the manufacturer to ensure the antibody specificity to the target protein and IHC procedures were already published [38]. The detectionwas obtained with the alkaline phosphatase method (APAAP), with the EXPOSE Mouse and Rabbit Specific AP (red) Detection IHC Kit (ab94734; Abcam; Cambridge, UK), and the colour was developed with fast red chromogen, or with a streptavidinbiotin immunoperoxidase detection system with the Ultravision Quanto Detection System HRP (Thermo Scientific, Fremont, USA), and the immunohistochemical staining was developed with AEC substrate HIGHDEF1 Red IHC (Enzo Life Sciences, Inc., New York, USA). The slides were counterstained with haematoxylin, and then mounted using a water-miscible mounting medium. A pancreatic endocrine tumour case, previously tested, was used as negative (omission of primary antibody) and positive control. pERKs and TERT expression in the series has been previously reported [16,32].

Western blot analysis and antibodies
Cells were lysed for 15 min at 4˚C using RIPA buffer (1% NP-40 in 150 mM NaCl, 50 mM Tris [pH 7.5], 2 mM EDTA) containing phosphatase and protease inhibitors. Proteins were quantified using a modified Bradford assay (Biorad). Protein samples (50 μg) were separated in 10% SDS/PAGE gels and electroblotted to Hybond ECL membrane (Amersham Biosciences). SDHD (polyclonal, rabbit, 1:200, Santa Cruz Biotechnology) was used. Secondary antibody was conjugated with peroxidase (Santa Cruz Biotechnology) and visualized by the ECL detection solution. Membrane was re-stained with a goat polyclonal anti-actin (Santa Cruz Biotechnology) for loading protein control. XTC1, a hurthle cell thyroid cell line, was used as positive control.

Statistical analysis
Statistical analysis was performed using STAT VIEW-J 5.0 (SAS Institute, Inc., Cary, NC). The relationship between the average expression level (score) of the immunohistochemistry markers and clinical-pathological parameters was evaluated by ANOVA. When appropriate, multiple comparison corrections were performed using the post hoc Bonferroni or Tamhane tests. The correlation between the immunoreactivity score of the different markers was assessed using the Fisher's exact test. The Kaplan-Meier method and log-rank test were used to evaluate the melanoma survival data. Univariate analyses were performed to determine the prognostic value of covariates regarding OS and DFS using the Cox regression model. OS and DFS were calculated from the time of diagnosis until death due to disease or metastasis, respectively, or censored at the time of the latest follow-up or death unrelated to the disease. A p value <0.05 was considered statistically significant.

SDHD promoter mutations analysis in CM, OM and melanoma cell lines
86 CM were analysed for SDHD promoter mutations. We found a chr. 11:111,957,523 (TTCC>TTTC) SDHD alteration [one of the mutations reported by Weinhold et al [8]] in two CM (2%) (Fig 1). One mutated case was a superficial spreading melanoma diagnosed in 1999, that did not display BRAF, NRAS and TERT promoter mutations, and the patient was alive at the last follow-up (180 months). The other mutated case was an acral melanoma diagnosed in 2009, displayed a -124:G>A TERT promoter mutation, but not a BRAF mutation; the patient died 24 months after diagnosis, in line with the poor prognosis of CM harbouring TERT promoter mutations. [4,5,16,39]. None of the OM cases (26),cutaneous (n = 4) and ocular (n = 7) melanoma cell lines studied harboured any alteration in the promoter of the SDHD gene. Due to the lower number of mutated cases, we could not assess any association between the presence of the mutation and the clinico-pathological and prognostic parameters of CM.

Expression of SDHD protein in CM, OM and melanoma cell lines
SDHD protein expression, evaluated in 107 CM, was cytoplasmic and present in all cases, including those (two cases) with SDHD promoter mutations (Fig 2). SDHD was expressed not only in melanocytes/melanoma cells, but also in keratinocytes and in the cells of sebaceous glands and hair follicles. Low staining score (score 2) was observed in 41% and moderate/ high staining score (score >2) was observed in 59% of melanomas.
SDHD protein expression was also evaluated in 33 OM. Cytoplasmic immunoreactivity was observed in 67% of conjunctival melanomas and 30% of uveal melanoma (Fig 2). SDHD was expressed also in adjacent non-tumour eye structures, mostly iris, retina and ciliary pigment epitheliums. Low staining score (score 2) was observed in all the positive conjunctival melanomas and 19% of uveal melanoma, and moderate/high staining score (score >2) was observed in 11% of uveal melanoma.
No association was found between the expression of SDHD and CM histological subtypes or type of sunlight/UV exposure. Concerning the prognostic factors of CM, a significant association was found between lower mean SDHD protein expression and the presence of ulceration (p<0.01) and higher pT stage (p<0.01) ( Table 3). No significant association was found between SDHD protein expression and other clinico-pathological parameters (age, sex, mitotic index and tumour thickness), although the same tendency, lower mean SDHD protein expression, was observed with higher thickness and mitotic rate. To evaluate whether SDHD protein expression correlates with disease-free (DFS) and overall (OS) survival in CM, Kaplan Meier curves and univariate Cox regression were performed. No significant association between SDHD protein expression and DFS and OS of the patients was found, although it was observed in the Kaplan Meier curves that patients with low expression of SDHD displayed non-significant reduced DFS and OS compared with patients with high expression of SDHD (S1 Fig). Regarding OM, the low number of conjunctival and uveal cases did not allow the statistical analysis. We also evaluated if the expression of SDHD protein was related to the activation of the MAPK pathway, and the presence of TERT promoter mutation and TERT protein expression. No association was found between the expression of SDHD and BRAF/NRAS mutations, pERK expression (the readout of MAPK pathway activation) or with TERT promoter mutations and TERT protein expression. Due to the low number of mutated cases, we cannot infer if there is a reduction in the protein expression associated with the presence of SDHD promoter mutation, but moderate/high staining scores (6 and 9) were found in the two cases with SDHD promoter mutations. All the melanoma cell lines studied expressed SDHD protein (S2 Fig).

Discussion
In this work we established, for the first time, that SDHD protein expression associates with prognostic features of CM. SDH enzyme has a central role in mitochondrial metabolism [2]. SDHD alterations result in the disruption of the SDH complex and loss of SDH enzymatic activity [40,41]. In paragangliomas, SDHB alterations were linked to malignancy and SDHD alterations are more frequent in head-and-neck localized tumours [42,43]. Recently, SDHD promoter mutations were reported in melanomas, associated with reduced gene expression and reduced patient survival [8,9].
At variance with Weinhold et al [8], who reported frequent noncoding alterations in SDHD promoter in melanomas, we only found two cases (2%) with SDHD alteration in our cutaneous melanoma series, similar to the 4% reported by Scholz et al [9]. Several factors may explain these discrepant results when compared with Weinhold et al work which was exclusively based on data mining, without sequencing validation, namely the use of different methodologies with different sensitivities, and differences in the cutaneous melanoma subtypes and melanomas staging. Although C>T alterations are generally considered a marker of UV-exposure, we found a 523C>T SDHD mutation in an acral melanoma case that occur in skin without sun exposure, at variance with Scholz et al [9] that did not find SDHD mutations in acral melanomas. We did not find also any mutations in uveal melanomas, that are considered not associated with exposure to sunlight/UV radiation [19]. No alterations were found by us in cutaneous and ocular melanoma cell lines; these results reinforce the possibility that these alterations are rare in melanomas. The rarity of SDHD mutations detected in our series does not allow us to validate the association between the presence of SDHD promoter mutation and a reduction in the expression of SDHD gene or the association between this mutation and prognostic parameters of CM, as reported by Weinhold et al [8].
SDHD protein was expressed in all melanoma cell lines analysed and in all CM cases, including the two cases with SDHD mutation. We found a significant association between low mean SDHD protein expression and the presence of ulceration and high pT stage. Our results indicate an association between the reduction of SDHD protein expression and worst prognosis, but without a significant relation with survival of patients with CM. Although the molecular and cellular mechanisms linking SDH inactivation and tumorigenesis it is not completely understood, SDHD mutation/inactivation results in a loss of electron transport chain complex II activity and in the activation of the hypoxia-angiogenic pathway, namely an increase of EPAS-1, HIF-1 and VEGF expression, which may be the involved mechanism in tumorigenesis [40,44]. The presence of pigmentation is also linked to increase HIF-1 expression and shorter DFS and OS [45,46], however in our series, no relation between SDHD expression and pigmentation status was found.
Diverse etiopathogenic mechanisms may operate in the development of cutaneous, conjunctiva and uveal melanomas and it seems that conjunctiva melanomas share more comparable pathogenesis with cutaneous melanomas than with uveal melanomas [47,48]. We still dispute the biological meaning of absence of SDHD expression in OM, as no SDHD promoter mutations were found in the OM cases analysed.
In conclusion, our results indicate that SDHD promoter mutation is a rare event in CM and is absent in OM. Yet, SDHD expression might have prognostic relevance in CM. Larger studies are necessary to validate if low expression of SDHD (related or not with the presence of the promoter mutation) might associate with worst prognostic features of CM. Importantly, the metabolic reshape created by SDHD alteration may open a possible therapeutic window that can benefit CM patients, through drugs that can revert this metabolism shift.