C-Kit SCF receptor (CD117) expression and<i>KIT</i>gene mutation in conjunctival pigmented lesions

Abstract

Purpose: To investigate the presence of KIT gene mutations and immunoreactivity in 85 conjunctival melanocytic tumours and to clarify the role of KIT as a potential therapeutic target in this group of patients. Methods: Eighty-five conjunctival pigmented tumours [27 melanomas, 12 primary acquired melanosis (PAMs) and 46 nevi] were immunostained for KIT. Intensity and pattern of expression were evaluated. Molecular analysis to identify KIT mutations was performed in 15 selected cases (tumour-rich areas >50%). KIT immunostaining score and pattern were statistically related to patients’ age, sex, diagnostic category, presence of relapse, disease-free survival, presence of metastases, metastasis-free survival, limbal versus nonlimbal tumour location and thickness of melanomas. Results: KIT stains were documented in 48% of melanomas, 50% of PAMs and 24% of nevi. The mean score of KIT staining in the melanomas/PAMs group was significantly different from nevi (p = 0.0076). No statistically significant differences were detected between either c-kit immunostaining score or pattern and each of the other clinico-pathologic parameters considered. No KIT gene mutations were detected in melanomas and nevi. A silent mutation/polymorphism in KIT exon 13 was found in one PAM. Conclusions: Despite the high level of KIT immunostains in PAMs and melanomas, this parameter seems not to be a good predictor of the presence of molecular mutations. KIT-activating mutations should be considered an uncommon event in this tumour. Conjunctival melanocytic proliferations are considered the most common ocular surface tumours accounting for up to 53% of all conjunctival neoplasms in a clinical survey of 1643 consecutive conjunctival lesions in an ocular oncology referral practice (Shields et al. 2004). These lesions include a broad spectrum of benign, premalignant and malignant conditions, whose classification encompasses melanocytic nevi, conjunctival melanosis and invasive melanomas. Conjunctival melanoma, nevi and primary acquired melanosis (PAM) are biologically strictly interrelated, considering that conjunctival melanoma may arise from PAM (74%), from nevus (7%) or de novo (19%; Shields et al. 2011). Occurring in the anatomical context of the conjunctiva, conjunctival melanocytic proliferations produce unique histological patterns, often different from those in the skin, reflecting the biological difference of these lesions compared with the cutaneous ones (Zembowicz et al. 2010). Although many clinicopathological studies of conjunctival melanocytic proliferations have been reported, there are no studies specifically focused on the immunohistochemical expression and gene mutations of c-KIT, which is notoriously known as a key factor in the melanocyte migration, differentiation and tumorigenesis (Terada 2012). c-Kit (or CD117) is a 145 kD immunoglobulin superfamily member, also known as stem cell factor receptor (SCFR). It is a transmembrane tyrosine kinase receptor that binds the c-Kit ligand (also known as steel factor, stem cell factor and mast cell growth factor). CD117 is expressed on hematopoietic stem cells (including multipotent hematopoietic stem cells, progenitors committed to myeloid and/or erythroid lineages, and T- and B-cell precursors), mast cells, interstitial cells of Cajal in the digestive tract and melanocytes. CD117 is a proto-oncogene whose activating mutations are associated with gastrointestinal stromal tumours, testicular seminoma, mast cell disease, melanoma and acute myeloid leukaemia. Melanomas with oncogenic mutations in KIT have been reported in several recent studies focused on KIT role in melanocyte transformation (Went et al. 2004; Willmore-Payne et al. 2005a,b; Curtin et al. 2006; Antonescu et al. 2007). Because of most of the reported mutations of KIT are considered to be sensitive to KIT kinase inhibitors, recent successes in the treatment for gastrointestinal stromal tumours (GIST) with imatinib (Gleevec; Novartis Pharma, Basel, Switzerland) have led to speculate that KIT-mutant melanoma might also be managed using competitive tyrosine kinase inhibitors. The aim of this study is to investigate the presence of KIT gene mutations and immunostains in a consecutive series of conjunctival melanocytic lesions to clarify (i) the role of KIT in conjunctival melanocyte transformation and (ii) the existence of a potential therapeutic target in these groups of patients. This study complied with the tenets of the Declaration of Helsinki and was approved by the IRB of our Institutions. To be included in this pilot, institutional, retrospective, nonconsecutive case series, each patient was recruited from those referred to our institutions because affected by conjunctival melanocytic proliferations from December 1996 to February 2010 and treated by surgical excision as a first-line treatment. Included patients were randomly selected by this cohort of patients representing a proportion of 5% of the total amount of cases for each included subgroup (melanoma, PAM, nevi). Exclusion criteria were as follows: any coexisting chronic corneo-conjunctival disease (excluding dry eye) and any other medical or surgical treatment previously performed for the conjunctival lesion. Each patient underwent surgical tumour excision, performed by an expert ocular oncologist, using standard oncologic criteria. Briefly, separate instrument sets were used for removal and reconstruction phases. The surgical no-touch technique was employed. Ocular surface reconstruction was performed at the same time of excision by conjunctival sliding or amniotic membrane implantation, depending on the amount of excided tissue. Adjuvant treatments (postsurgical second-line treatments) were reserved for PAMs and melanomas. Adjuvant cycles of topical mitomycin-C (MMC) were used for each case of PAM (one droop, four times a day for 2 weeks: one cycle). The number of cycles was based on clinical evolution, histopathological grade of dysplasia and amount of residual pigmentation after surgery. Two cycles of adjuvant topical MMC were also used for each conjunctival melanoma (completely excided), until the introduction in our protocols of brachytherapy as a standard adjuvant treatment for these patients (early 2009). Microscopically incompletely excised melanoma underwent additional surgical tumour removal until free surgical margins were obtained. Clinical, histopathological and follow-up data are summarized in Table 1. All tissues were fixed in 10% buffered formalin and embedded in paraffin wax, and 5-micron sections were stained with haematoxylin and eosin. PAM was histologically classified according to Folberg et al. (1985) criteria into: PAM without atypia (defined as hyperpigmentation without melanocytic hyperplasia and/or melanocytic hyperplasia without cytologic atypia) and PAM with atypia (characterized by the presence of cytologic atypia). PAM with atypia was further characterized reporting two selected histological factors at higher risk of subsequent progression to melanoma: the presence of epithelioid melanocytes and the distribution of melanocytes in a pattern other than basilar. PAM with atypia featuring at least one of these two characteristics was considered at high risk of transformation. From each of the 85 tissue blocks, 5-μm sections were cut for immunohistochemical determinations. Sections were pretreated in a microwave oven (750 W) for 20 min in a citrate buffer (10 mm, pH 6.0). Sections were then stained with the CD117 antibody (c-kit polyclonal antibody, clone A4502, 1:250 dilution; DAKO, Glostrup, Denmark) using the Envision ChemMate Kit on the Dako Autostainer. Colour was carried out using 3,3′-diaminobenzidine (DAB; DAKO) for 4 min. Sections were counterstained with Meyer’s haematoxylin. Positive and negative controls were added to each automated immunohistochemistry run: negative controls had primary antibody replaced by phosphate-buffered saline (PBS) solution, and positive controls were sections of gastrointestinal stromal tumours. C-kit immunostaining was graded in a semiquantitative manner as follows: 0 (no staining), 1+ (1–10% staining), 2+ (11–29% staining), 3+ (30–50% staining) or 4+ (>50% staining). C-kit staining pattern was assessed in the cytoplasm and in cytoplasm membrane. A total of 25 formalin-fixed paraffin-embedded tissue samples from 11 primary conjunctival melanomas, 3 PAM and 11 conjunctival nevi were selected for the molecular analysis as tumour-rich areas (>50%) samples. KIT mutations were addressed by analysing exons 9, 11, 13 and 17 because considered the regions where the majority of mutations are located (Wallander et al. 2011). Genomic DNA was isolated using the QIAamp DNA Mini Kit (Qiagen, Milan, Italy) and subjected to PCR amplification and direct sequencing of exons 9, 11, 13 and 17 of the KIT gene using specific primer pairs (Corless et al. 2002). PCR was carried out in a total volume of 25 μl, using 100 ng of DNA, 1 U AmpliTaq Gold DNA Polymerase (Applied Biosystems, Foster City, CA, USA) and 10 pmol/camp of each primer. PCR products were purified and sequenced by Big Dye Terminator v.1.1 Cycle sequencing kit using ABI Prism 310 genetic Analyser (Applied Biosystem). When appropriate, the following statistical tests have been applied considering a nonparametric approach, given a non-normal distribution of the data: Mann–Whitney two-sample rank sum test, Kruskal–Wallis rank sum test and Spearman’s rank correlation coefficient. Survival analysis was performed applying the Cox model in the univariate modality. In case of proportions, the binomial exact 95% confidence interval (CI) was calculated. The statistical analysis was carried out using the stata, statistical package, release 8.1, for Windows. A p-value of 0.05 was considered significant. For statistical purposes, patients with a diagnosis of melanoma and PAM were grouped together (39/85; 46%) and compared with patients with a diagnosis of nevus (46/85; 54%). C-kit immunostaining score and pattern of expression were statistically related to patients’ age, sex, diagnostic category (melanoma/PAM versus nevus), limbal versus nonlimbal tumour location and thickness (only for melanomas), frequency of relapse, disease-free survival, presence of nodal metastases and metastases-free survival. Eighty-five cases of conjunctival melanocytic lesions were included in this study. Forty-six cases were histopathologically classified as nevi (54%), 12 as PAM (14) and 27 as melanoma (32%). Examined PAMs were all characterized by the presence of atypia and were subcategorized as low risk (5/12; 42%) and high risk of melanoma development (7/12; 58%), respectively. Among patients with conjunctival melanoma, male/female ratio was 11:16, with a mean age at diagnosis of 62.6 years (range, 30–79 years) and a mean follow-up time of 51.7 months (range, 2–142 months). Patients with melanoma were staged according to 2009 WHO TNM classification of conjunctival melanomas: five cases (5/27; 18%) were in situ melanoma pTis; 1 case (1/27; 4%) was PT1a stage; 10 cases (10/27; 37%) were pT1c stage; 3 cases (3/27; 11%) were pT2a stage; 1 case (1/27; 4%) was pT2b stage, and the remaining 7 cases (7/27; 26%) were pT2c stage. Grouping together patients with conjunctival melanoma and PAM, male/female ratio was 14:25, with a mean age at diagnosis of 59.7 years (range, 16–94 years) and a mean follow-up of 47.3 months (range, 2–142 months). Thirteen of 27 conjunctival melanomas (48%; 95% CI 28–68) showed immunoreactivity for CD117 (c-kit), with both cytoplasmic and membranous predominant staining pattern (11/13; 85%; 95% CI 54–98). Two cases (15%; 95% CI 1–24) expressed a membranous pattern alone (Fig. 1). Immunoreactivity for c-kit staining was also found in six of twelve PAMs (50%; 95% CI 21–78). Of these, immunoreactivity was present in three low-risk (60%; 95% CI 14–94) and in three high-risk (42%; 95% CI 9.8–81) PAM, respectively. C-kit expression showed a membranous distribution pattern in three cases (50%; 95% CI 11–88), whereas both cytoplasmic and membranous predominant reactivity pattern was documented in the remaining three cases (50%). Among the 46 conjunctival nevi assessed, eleven cases (24%; 95% CI 12–38) demonstrated immunoreactivity for CD117 (c-kit). Of these, six cases had both a cytoplasmic and membranous staining pattern, whereas five cases had a membranous staining pattern alone. c-kit scores for each type of lesion are summarized in Table 2. The mean score of c-kit staining in the melanoma/PAM group was significantly different from the mean score among patients with nevi (p = 0.0076; Mann–Whitney two-sample rank sum test). No statistical difference between the two groups was found in c-kit immunohistochemical pattern of expression. No statistical significance was detected between either c-kit score or pattern and each of the other clinico-pathologic parameters considered: age (p = 0.52; Spearman’s rank correlation), sex (p = 0.10; Mann–Whitney two-sample rank sum test), presence of relapse (p = 0.82; Mann–Whitney two-sample rank sum test), disease-free survival (p = 0.37; Cox model; HR = 0.98; 95% CI 96–101), presence of metastases (p = 0.96; Mann–Whitney two-sample rank sum test), metastasis-free survival (p = 0.71; Cox model; HR = 0.97; 95% CI 87–112), limbal versus nonlimbal tumour location (p = 0.59; Kruskal–Wallis rank sum test) and thickness (only for melanomas; p = 0.17; Spearman’s rank correlation). A case of pigmented conjunctival melanoma (A) [haematoxylin and eosin (H&E), magnification 100×] showing a strong and diffuse immunoreaction for CD117 (c-Kit) with a membranous distribution pattern (c-kit immunostaining in red, magnification 100×) (B). A second case of conjunctival melanoma, epithelioid type (C) [haematoxylin and eosin (H&E), magnification 200×], showing a less intense focal positivity for CD117 (c-Kit) immunoreaction (c-kit immunostaining in red, magnification 200×) (D). Molecular analysis could be performed in 15/25 cases (60%; 95% CI 38–78; 8/11 melanomas, 2/3 PAM and 5/11 nevi). Mutation status could be determined only for KIT exon 11 and 13 in 8/25 (32%; 95% CI 14–53) cases (2/11 melanomas, 1/3 PAM and 5/11 nevi) due to the low amount of DNA and in 2/25 (8%; 95% CI 0.9–26) of remaining cases (1/11 melanoma and 1/11 nevus) due to the low DNA quality. No KIT gene mutations were detected in any of the conjunctival melanoma and nevi and in 2/3 PAM. A silent mutation/polymorphism in KIT exon 13 (TAC646TAT; p.Y646Y) was found in a single case of PAM (4%; 95% CI 0.1–20). Among the 15 lesions analysed, 6/8 (75%; 95% CI 35–96) melanomas, 1/2 PAM (50%; 95% CI 12–98) and 3/5 nevi (60%; 95% CI 14–94) showed positive immunostaining for c-kit. The single case with a silent mutation/polymorphism had negative c-kit immunostaining expression. c-kit activation becomes independent from its ligand mainly through oncogenic mutations, leading to a persistent activation of the signal transduction cascades, influencing tumour proliferation rate as well as reducing apoptosis and increasing transformed cells’ chemotaxis ability (Wehrle-Haller 2003; Alexeev & Yoon 2006; Grichnik 2006). Recent studies have focused on melanomas with oncogenic mutations in KIT. Although KIT mutations have been largely described in cutaneous melanomas, these mutations seem to be more common in mucosal and acral ones (Went et al. 2004; Willmore-Payne et al. 2005a,b; Curtin et al. 2006; Antonescu et al. 2007). Interestingly, most of the reported mutations are predicted to be sensitive to KIT kinase inhibitors (Willmore-Payne et al. 2005a,b; Curtin et al. 2006; Antonescu et al. 2007; Rivera et al. 2008). Nevertheless, few studies have attempted to investigate KIT mutational status and/or copy number as well as CD117 immunohistochemical expression in conjunctival melanomas (Beadling et al. 2008; Wallander et al. 2011). To the best of our knowledge, conjunctival melanomas have been included (as part of larger series) only in two different studies (Beadling et al. 2008; Wallander et al. 2011): the former analysed 14 conjunctival melanomas detecting a mutation in a single case; the latter screened five conjunctival melanomas for mutational status alone and did not find any mutations despite the presence of one immunopositive case. A further case of conjunctival melanoma has recently been analysed, showing positive c-kit immunostaining, but no KIT mutations (Terada 2012). In the present study, we enrolled 85 patients affected by conjunctival pigmented lesions, including 27 melanomas, 12 PAM and 46 nevi. C-kit immunohistochemical expression could be evaluated for all lesions, whereas molecular analysis was feasible in 25 cases (8 melanomas and 3 PAMs) as they only were tumour-rich samples. Thus, this is mainly an immunohistochemical study, and as such a study of localization of immunoreaction rather than gene expression. Nevertheless, to the best of our knowledge, the present series comprises the largest number of conjunctival melanomas tested for KIT immunostains and molecular status ever reported. Furthermore, we have also included benign and premalignant melanocytic conjunctival lesions because of the presumed interrelation between PAMs, nevi and melanomas (Shields et al. 2011). No KIT gene mutations were detected in any of the 25 tissue samples analysed. Only a silent mutation/polymorphism in KIT exon 13 (TAC646TAT; p.Y646Y), not influencing protein function and therefore unhelpful for therapeutic purposes, was found in a single case of PAM. Therefore, KIT-activating mutations should be considered an uncommon event in these lesions. Moreover, immunohistochemical scores and pattern of expression were correlated with clinicopathological data and outcome of all patients: the only parameter that showed statistical significance was c-kit mean staining score significantly different between melanoma/PAM compared with nevi (p = 0.0076). Thus, our results are consistent with overall immunohistochemical and molecular tests performed in a total of 20 conjunctival melanomas already published (Beadling et al. 2008; Wallander et al. 2011; Terada 2012). Only a single case of a KIT-activating mutation was recently described in conjunctival melanoma by Beadling et al. In the present study, we did not find kit-activating mutation, considering both immunopositive and immunonegative melanomas. Given the high percentage of immunopositive melanomas (48%) and PAM (50%), the usefulness of CD117 staining as a screening method to identify conjunctival melanomas that would benefit from imatinib therapy is addressed. Besides KIT-activating mutations, additional mechanisms may drive CD117 overexpression. Constitutive phosphorylation of the receptor protein (Antonescu et al. 2007), a decrease in protein turnover resulting in accumulation and thus overexpression of C-kit in malignant cells or stem cell factor/KIT autocrine loop alterations (highlighted in uveal melanoma cells; Lefevre et al. 2004), could contribute to tumorigenic mechanisms also in conjunctival melanoma. In conclusion, KIT-activating mutation appears a rare event in conjunctival melanoma, and CD117 has not proved to be a good predictor of its presence. Nevertheless, performing the molecular screening to identify Kit molecular alteration in conjunctival melanoma patients may theoretically allow to offer a tailored treatment in those patients that could benefit from imatinib therapy irrespective of CD117 expression.