BRAFV600E Mutant Allele Frequency (MAF) Influences Melanoma Clinicopathologic Characteristics

Simple Summary The mutational load of BRAFV600E in melanomas has been described as a possible prognostic biomarker but there is no information about the mutant allele frequency (MAF) variability of BRAFV600E within cutaneous melanomas and its potential prognostic implications. Our study suggests that the variation degree of BRAFV600E MAF within primary cutaneous melanoma could act as a determinant in the location of primary melanomas as well as their first metastases and could influence prognostic indicators such as Breslow and mitotic indexes. BRAFV600E MAF variation is also related to the neoplastic cell phenotype and tumor lymphocytic infiltrate of primary tumors. For all these reasons, detection of BRAFV600E MAF variation could be a useful prognostic biomarker. It is worth exploring the role of BRAFV600E MAF variation with regards to the response of melanoma to targeted therapies. Abstract Background: Cutaneous melanoma shows high variability regarding clinicopathological presentation, evolution and prognosis. Methods: Next generation sequencing was performed to analyze hotspot mutations in different areas of primary melanomas (MMp) and their paired metastases. Clinicopathological features were evaluated depending on the degree of variation of the BRAFV600E mutant allele frequency (MAF) in MMp. Results: In our cohort of 14 superficial spreading, 10 nodular melanomas and 52 metastases, 17/24 (71%) melanomas had a BRAFV600E mutation and 5/24 (21%) had a NRASQ61 mutation. We observed a high variation of BRAFV600E MAF (H-BRAFV600E) in 7/17 (41%) MMp. The H-BRAFV600E MMp were all located on the trunk, had lower Breslow and mitotic indexes and predominantly, a first nodal metastasis. Regions with spindled tumor cells (Spin) and high lymphocytic infiltrate (HInf) were more frequent in the H-BRAFV600E patients (4/7; 57%), whereas regions with epithelial tumor cells (Epit) and low lymphocytic infiltrate (LInf) were predominant (6/10; 60%) and exclusive in the low BRAFV600E MAF variation tumors (L-BRAFV600E). The H-BRAFV600E/Spin/HInf MMp patients had better prognostic features and nodal first metastasis. Conclusions: The H-BRAFV600E MMp were located on the trunk, had better prognostic characteristics, such as lower Breslow and mitotic indexes as well as high lymphocytic infiltrate.

Simple Summary: The mutational load of BRAF V600E in melanomas has been described as a possible prognostic biomarker but there is no information about the mutant allele frequency (MAF) variability of BRAF V600E within cutaneous melanomas and its potential prognostic implications. Our study suggests that the variation degree of BRAF V600E MAF within primary cutaneous melanoma could act as a determinant in the location of primary melanomas as well as their first metastases and could influence prognostic indicators such as Breslow and mitotic indexes. BRAF V600E MAF variation is also related to the neoplastic cell phenotype and tumor lymphocytic infiltrate of primary tumors. For all these reasons, detection of BRAF V600E MAF variation could be a useful prognostic biomarker. It is worth exploring the role of BRAF V600E MAF variation with regards to the response of melanoma to targeted therapies. Abstract: Background: Cutaneous melanoma shows high variability regarding clinicopathological presentation, evolution and prognosis. Methods: Next generation sequencing was performed to analyze hotspot mutations in different areas of primary melanomas (MMp) and their paired metastases. Clinicopathological features were evaluated depending on the degree of variation of the BRAF V600E mutant allele frequency (MAF) in MMp. Results: In our cohort of 14 superficial spreading, 10 nodular melanomas and 52 metastases, 17/24 (71%) melanomas had a BRAF V600E mutation and 5/24 (21%) had a NRAS Q61 mutation. We observed a high variation of BRAF V600E MAF (H-BRAF V600E ) in 7/17 (41%) MMp. The H-BRAF V600E MMp were all located on the trunk, had lower Breslow and mitotic indexes and predominantly, a first nodal metastasis. Regions with spindled tumor cells (Spin) and high lymphocytic infiltrate (HInf) were more frequent in the H-BRAF V600E patients (4/7; 57%), whereas regions with epithelial tumor cells (Epit) and low lymphocytic infiltrate (LInf) were predominant (6/10; 60%) and exclusive in the low BRAF V600E MAF variation tumors (L-BRAF V600E ). The H-BRAF V600E /Spin/HInf MMp patients had better prognostic features and nodal first metastasis.

Introduction
Cutaneous melanoma (CMM) is the most aggressive skin malignancy. Although CMM can be cured with wide local excision when it is diagnosed in the early stages, advanced melanoma is associated with poor outcomes [1]. Nevertheless, knowledge advances in melanoma pathogenesis and the subsequent development of targeted and immunobased therapies have improved metastatic melanoma treatment [2].
The MAPK-ERK pathway plays a major role in the development and progression of melanoma [3]. Overall, about 35-50% of melanomas of all clinical types have mutations in the V600 codon of BRAF [4]. Approximately 90% of BRAF mutations result in the substitution of valine (V) for glutamic acid (E) at position 600 (BRAF V600E ) [5][6][7][8][9]. The BRAF V600E mutation confers to this kinase the ability to activate MEK [4,10]. This ligand-independent pathway activation is essential in melanoma development and progression [11,12]. Combinations of BRAF and MEK inhibitors have emerged as new treatment options for patients with BRAF-mutated melanoma with a good response rate [13,14]. Unfortunately, most patients that initially show an extremely good response to BRAF inhibitors develop resistance to those drugs within a relatively short period of time (<1 year) [10,14].
Besides the usefulness of the BRAF V600E driver mutation as a therapeutic target, a recent study has suggested that the mutational load of BRAF V600E is a possible prognostic biomarker [15], expanding the classical histopathological prognostic criteria such as ulceration, Breslow, Clark or mitotic indexes [1,[16][17][18][19][20][21][22]. Moreover, some studies in recent years have attempted to relate the mutational load to the response in order to target therapies, but controversial results have been obtained [23][24][25][26].
Next-generation sequencing (NGS) technologies have played an essential role in understanding the altered genetic pathways involved in human cancer. It is a high-throughput method which allows for the sequencing of multiple samples and targeted genes in the same run. NGS can quantitate the proportion of reads for a given mutation, also known as mutant allele frequency (MAF), which represents the percentage of tumor cells that harbor a specific mutation in a neoplastic tissue. The sensitivity of NGS is higher than that observed with Sanger sequencing (2-10% versus 15-25% allele frequency, respectively) [27,28].
It is widely accepted that melanomas are compound by tumor cells, with different histophenotypic characteristics (morphology or pigmentation), and different non-tumor cells as inflammatory cells. These characteristics can also be different between the primary tumor and its metastases [29]. Therefore, understanding MAF using NGS technology could provide invaluable information regarding melanomagenesis and dissemination.
In this work, we used NGS to study the variation of BRAF V600E MAF between different regions of 24 primary melanomas and their paired metastases as well as its relation to their histophenotypic and clinical characteristics, in an attempt to identify the predictors of disease outcome.

Patient Selection and Data Collection
The Pathology Department archive was retrospectively reviewed from 1 January 2000 to 30 June 2020. Patients who had paired primary cutaneous melanoma (MMp) and metastases (MMx) were selected. In order to avoid excessive genetic diversity, the study was restricted to superficial spreading melanoma (SSM) and nodular melanoma (NM), according to the WHO classifications [8].
Gender, age, MMp subtype, Breslow and Clark indexes, presence of ulceration, number of mitoses, MMp location, number and location of the MMx, disease free survival (DFS) and overall survival (OS) were recorded.

Diagnostic Criteria of Histological Features
A detailed microscopic analysis of routine H&E sections for the identification of areas of different cell morphology (spindled, epithelioid and mixed) and different degree of pigmentation (absent/low, moderate and high) was performed. A selection of 3 and 2 different areas of the MMp and the MMx, respectively, was done ( Figure 1a).

DNA Extraction and NGS Analysis
In total, 191 areas selected were microdissected from the formalin-fixed paraffinembedded blocs. DNA extracted by Maxwell ® FFPE Plus DNA Kit (Promega, Madison, WI, USA) were quantified by Qubit fluorometer (Thermo Fisher Scientific, Waltham, MA, USA).
We performed NGS with the Ampliseq Cancer HotSpot panel v2 (Illumina, San Diego, CA, USA) to detect genetic mutations from 50 oncogenes and tumor suppressor genes. Genomic DNA (10 ng) was used to construct the amplicon libraries according to the manufacturer's protocol and quality control of the libraries was done with Agilent 2100 Bioanalyzer. Resulting data files were analyzed by DNA amplicon analysis workflow to perform alignment and variant calling. We excluded non-coding variants and synonymous changes of each polymorphism. Single nucleotide variants with a variant allele frequency of at least 10% and a minimum depth coverage of 150 reads were used as cut-off values for variant identification. (Figure 1a).

TMA Elaboration and Analysis
Four tissue microarrays (TMAs) containing the 191 areas, were constructed using a tissue arrayer device (Beecher Instrument, Sun Prairie, WI, USA). Histopathological tumor characteristics (cell morphology, degree of pigmentation and lymphocytic infiltrate) were qualitatively analyzed within each TMA spot. Cell shape was visually graded as epithelioid or spindled when the majority of the tumor cells expressed the same phenotype (>80) and mixed when both phenotypes were observed and none of them were >80%.
Pigmentation was scored on a three-point scale (high, moderate and low/absent). Melanophages were not considered. Lymphocytic infiltrate was scored on a three-point scale (high, moderate and low/absent). The dissected areas contained at least 50% of tumor nuclei. (Figure 1a).

Statistical Analysis
The Fisher exact test and T-test were used to examine associations between high or low intratumoral BRAF V600E MAF variation (H-BRAF V600E or L-BRAF V600E ) and demographic, clinical and MMp histopathological variables. Overall survival (OS) and disease-free survival (DFS) curves were calculated using the Kaplan-Meier method with log-rank test. The Mann-Whitney test was used to assess immunohistochemical characteristics. The p-values are indicated by asterisks * p < 0.05; ** p < 0.01; *** p < 0.001. All statistical analyses and figures were performed using SPSS Statistics software and Graphpad Prism software.

Clinicopathological Characteristics
Our series comprised 29 patients with 29 MMp and 52 paired MMx, which supposed a total number of 87 MMp and 104 MMx spots. Due to low quality samples, 5 patients were rejected and 24 patients were elected for further analysis ( Figure S1). This cohort was composed of 16 (67%) men and 8 (33%) women, with a median age of 63 years (24-87). The MM subtype was SSM in 14 (58%) and NM in 10 (42%) biopsies. In total, 12 (50%) MMp were located on the trunk, 6 (25%) on the lower extremities, 3 (12.5%) on the upper extremities and 3 (12.5%) on the head and neck. The mean MMp Breslow and Clark indexes were 4.22 mm (1.04-13.50) and 4 (3-5), respectively. Ulceration was present in 14 (58%) cases and the mean mitotic index was 5/cm 2 (1-11). Eight patients (33.33%) were alive at the end of the study with a mean survival time of 60 months (median 40.5 (9-177)) (Tables 1 and S1).  (Figure 1a). We did not observe significant differences in the distribution of pathological characteristics between the MMp and MMx groups (Table S2).  (Figure 1b). The BRAF V600E mutation was present in 17/24 (71%) patients, NRAS Q61 in 5/24 (21%) and 2/24 (8%) MMp were of the wild type for both genes. We compared the clinicopathological characteristics of BRAF V600E and NRAS Q61 patients and no statistical differences were detected. The BRAF V600E and NRAS Q61 mutations where highly preserved in the MMp and their paired MMx. Only 2/17 (11.8%) BRAF V600E mutated patients presented an absence of BRAF V600E mutation in one of the three MMp spots (intratumor heterogeneity). Additionally, in 2/17 (11.8%) patients whose three MMp spots were BRAF V600E mutated, both spots of their paired MMx were the BRAF wild type (intertumor heterogeneity) [32,33]. Across the cohort, 11.8% intratumoral and intertumoral heterogeneity was detected in the BRAF V600E mutation.
Striking differences in the MAF of BRAF V600E between spots of the same MMp were detected. Therefore, we decided to quantify such variations. First, to obtain the real BRAF V600E MAF we took into account the percentage (%) of tumoral cells present in each spot by three investigator-blinded assessments on H&E staining. Discrepant cases were discussed until an agreement was reached. Only spots composed of at least 50% tumor cells were considered. In all cases, the mean BRAF V600E MAF obtained by NGS was recalculated depending on the percentage of tumor cells observed in each spot from the TMA created. We defined H-BRAF V600E primary tumors as those that contained one spot with a twofold increase MAF value regarding at least one of the other MMp spots. Samples that lacked this diversity were considered L-BRAF V600E . In our cohort of 17 MMp with the BRAF V600E mutation, 7 (41.18%) showed H-BRAF V600E while the remaining 10 (58.82%) had L-BRAF V600E (Figure 2a).  Immunohistochemical examples of primary melanoma tumor with high immunohistochemical histoscore variation of BRAF V600E between its spots and primary melanoma tumor with low immunohistochemical histoscore variation of BRAF V600E between its spots.

Immunohistochemical Analysis: There Is Variation in BRAF V600E Immunoexpression in Primary Melanomas
High and low variation in BRAF V600E MAF was validated by an immunohistochemical study. First, histoscores of BRAF V600E were obtained by Qupath analysis from all of the TMA spots. As expected, NRAS mutated MMp spots presented a low histoscore (<20). Then, we considered a histoscore of ≥20 as the cut-off for the immunohistochemical histoscore scale for BRAF V600E positivity with a sensitivity of 87.5% and a specificity of 100% [34][35][36].
Then, we defined tumors with high immunohistochemical variation of BRAF V600E as those primary tumors that contained one spot with a two-fold increase histoscore value regarding at least one of the other MMp spots. Samples that lacked this diversity were considered to show a low immunohistochemical variation of BRAF V600E (Figure 2b). We observed that immunohistochemical evaluation of BRAF V600E showed significant consistency with the NGS analysis, as it was able to identify patients with H-BRAF V600E with a sensitivity of 80% and a specificity of 77.78%.

H-BRAF V600E Correlation with Demographic, Clinical and Tumor Characteristics
First, we analyzed all clinicopathological characteristics of H-BRAF V600E patients compared with those of the L-BRAF V600E group ( Figure 3A).
H-BRAF V600E patients had statistically lower Breslow and mitotic indexes than L-BRAF V600E (Mean 2.617 ± 0.4426 vs. 5.280 ± 0.8373 and Mean 2.571 ± 0.5714 vs. 6.000 ± 1.202). These results indicated that the H-BRAF V600E MMp had better prognostic features than the L-BRAF V600E since their Breslow tended to be <4 mm and they had a lower mitotic index [1,20,22] (Figure 3B,C). No statistical differences were observed regarding age and gender, MMp subtype, Clark or ulceration.
Moreover, we observed important differences in the first MMx location. Most H-BRAF V600E patients presented a nodal first MMx (6/7; 86%), while the first MMx in L-BRAF V600E patients was mostly cutaneous/subcutaneous (7/10; 70%) ( Figure 3E). Interestingly, in H-BRAF V600E patients the nodal MMx was detected in the sentinel lymph node biopsy in 4/6 (67%) cases, whereas in 4/7 (57%), skin MMx of the L-BRAF V600E patients were already present at diagnosis of the primary tumor as a satellitoses or in-transit metastases. No statistical differences were observed in OS and DFS between the groups ( Figure S2a).
All these results indicate that H-BRAF V600E seems to correlate with the development of the MMp on the trunk, better prognostic features such as lower Breslow and mitotic indexes, and nodal first MMx.
All these results suggest that the spindle phenotype and high/moderate lymphocytic infiltrate correlate with the H-BRAF V600E samples. This confers better clinicopathological prognostic characteristics for these MM patients, such as lower Breslow and mitotic indexes [1,20,22].

Discussion
Melanoma can show different morphological characteristics such as different cytological subtypes or degrees of pigmentation either between different MMp or even within the same MMp (intratumoral heterogeneity) and between the MMp and paired MMx (intertumoral heterogeneity).
This study attempts to investigate the relationship between the diverse histophenotypic features present in different regions of MMp compared with the genetic characteristics found by NGS assay in a series of 24 patients with primary SSM or NM.
BRAF and NRAS mutation frequencies differ between cutaneous melanoma subtypes. BRAF mutation occurs more frequently in SSM (20-65%) and in second place in NM (20-43%) [38][39][40][41][42][43], while NM is the most frequent MM subtype with NRAS mutations (21-27%) [41,44]. Our series only included SSM and NM, which harbored the BRAF V600E mutation in 71% and NRAS Q61 in 21% of MMp. These findings were similar to those previously described, although the presence of BRAF V600E was slightly higher. As the number of SSM and NM in our series are quite balanced (14 and 10 cases, respectively) differences in sequencing methodology could perhaps explain this fact, since a greater sensibility of NGS has been described in the detection of BRAF mutations compared to allele specific PCR [37,45].
We found that the BRAF V600E and NRAS Q61 mutations where highly preserved in MMp and their paired MMx. Only 2/17 (11.8%) cases presented intratumor heterogeneity and 2/17 (11.6%) cases presented intertumor heterogeneity regarding BRAF V600E mutation. These findings support the idea that driver mutations occur early during the development of malignant melanoma [23,38,39,46]. Then, melanoma can show intratumoral and intertumoral BRAF V600E mutation heterogeneity, although its frequency is low.
In recent years, controversial results have been obtained by studies which related the BRAF V600E MAF of melanoma tumors with clinicopathological and prognostic characteristics as well as the melanoma response, in order to target therapies [15,[23][24][25][26]. All these studies analyzed the MAF value of one specific region for each tumor, so we wanted to know whether the variation degree of BRAF V600E MAF within different regions of the same tumor could influence these characteristics.
Interestingly, our study suggests that variation of BRAF V600E MAF within MMp is related to the clinical features of the MMp, histopathological prognostic factors and progression behavior, although all the patients in our series developed metastasis at some point (since it was one of the inclusion criteria). Thus, H-BRAF V600E MMp would settle more frequently on the trunk and their first MMx would more frequently be nodal, while the L-BRAF V600E melanomas would be located more frequently on the extremities and develop most often cutaneous/subcutaneous, regional first MMx. The preferential location of the first MMx on the skin in L-BRAF V600E MMp cases could be explained because all were satellite or in-transit MMx, and this type of MMx are more frequently associated with MMp located on the extremities [47,48]. In addition, the H-BRAF V600E patients seem to have better histopathological prognostic indicators such as lower Breslow and mitotic indexes [1]. These associations were even more evident when we analyzed the H-BRAF V600E /Spin/Hinf and L-BRAF V600E /Epit/Linf subgroups. We identified an even stronger predilection of the MMp and first MMx for the trunk and nodal regions, respectively, with lower Breslow and mitotic indexes in the H-BRAF V600E /Spin/Hinf patients, while the L-BRAF V600E /Epit/Linf MMp were preferentially located on the extremities and their first MMx was significantly more frequently cutaneous, showing higher Breslow and mitotic indexes. Our findings could be explained in part by the suggestion of some authors that the presence of tumor-infiltrating lymphocytes (TILs) in melanoma is associated with a lower Breslow and mitotic rate [49][50][51]. Moreover, with regards to cytological features, a recent study showed that melanoma cells with rounded behavior had a higher capacity for tumor formation, tumor progression and metastasis than elongated cells [52], indicating that an increase in epithelioid cells in MMp could mean a higher proliferative and invasive behavior.
Even though we are aware of the modest sample size, our results suggest that variation of BRAF V600E MAF can constitute a potential biomarker that is worth exploring further. Moreover, these findings could serve as initial step for the future development of therapies focused on the presence of worse genetic and histopathological characteristics, such as L-BRAF V600E , epithelioid cell compound and low lymphocytic infiltrate in MMp.

Conclusions
The attained results suggest that MAF variation of BRAF V600E within primary cutaneous melanomas (MMp) could influence their histopathological characteristics, as well as their anatomic location and that of their first metastases (MMx). Here we described for the first time in the biomedical literature, that H-BRAF V600E could correlate with a spindle phenotype, high/moderate lymphocytic infiltrate and the development of the MMp on the trunk and nodal first MMx. On the contrary, L-BRAF V600E seems to correlate with an epithelioid phenotype, low lymphocytic infiltrate, the development of the MMp predominantly on the extremities and a cutaneous first MMx.
In addition, we suggest a possible new link between H-BRAF V600E and better classical prognostic indicators such as lower Breslow and mitotic indexes.

Institutional Review Board Statement:
The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Hospital Universitary Arnau de Vilanova Ethics Committee (CEIm (ref 1465)). Samples were obtained with the support of the Xarxa de Bancs de Tumors de Catalunya sponsored by Pla Director d'Oncología de Catalunya (XBTC)", IRBLleida Biobank (B.0000682) and PLATAFORMA BIOBANCOS (PT17/0015/0027 and PT20/00021).

Informed Consent Statement:
Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available in this article (and Supplementary material).

Conflicts of Interest:
The authors declare no conflict of interest.