Characterization of Single Nucleotide Variants of OPN3 Gene in Melanocytic Nevi and Melanoma.

In this study, we examined single nucleotide variants (SNVs) of the OPN3 gene in malignant melanoma and melanocytic nevi. A total of 20 variants of SNVs were detected. Of these variants, five nonsynonymous mutations of OPN3 were identified, including c.T152C, c.T401C, c.G547A, c.G768A, and c.G992A. Three prediction tools, MutationTaster2, Polymorphism Phenotyping version 2, and PROVEAN (Protein Variation Effect Analyzer), which predict possible impact of an amino acid substitution, suggested that the mutations could be deleterious. Nine SNVs occurred in 3' untranslated regions, whereas two were observed in 5' untranslated regions. In all cases, four intronic variants were identified. In addition, we identified nine 3' untranslated region SNVs in OPN3; one of them (OPN3[NM_014322:c.∗83C>T]) is predicted to disrupt a conserved microRNA (has-miR-376c-3p) target site, located in position 86-93 of OPN3 3' untranslated region. Our findings suggest that there is a strong possibility that OPN3 SNVs play a role in the pathogenesis of melanocytic tumors via prediction of functional phenotype.

Introduction

First described by Blackshaw and Snyder (1999), OPN3 was identified as a mammalian extraocular opsin. It is highly conserved throughout vertebrates as an ancestral type of opsin (Fischer et al., 2013). On the genomic level, OPN3 is composed of four exons on human chromosome 1q43, encompassing two flanking genes, CHML and KMO. CHML, a single exon gene, resides in intron 1 of OPN3, whereas KMO and OPN3 overlap with their 3′-untranslated regions (UTRs), which are transcribed in opposite directions (Halford et al., 2001). Recently, we and others have found that OPN3 is highly expressed in human epidermal melanocytes compared with other opsins (Ozdeslik et al., 2019; Regazzetti et al., 2018; Wang et al., 2020). These studies further demonstrated that OPN3 mediates nonvisual functions such as melanogenesis and apoptosis in melanocytes. In addition, previous studies of the relationship between OPN3 and human diseases showed that OPN3 is identified as an asthma susceptibility gene (White et al., 2008). Miyanaga et al. (2020) recently found that OPN3 is upregulated in pulmonary carcinoid tumors that developed postsurgical metastasis. With The Cancer Genome Atlas data analysis, we also observed that OPN3 expression is upregulated in human melanoma compared with normal skin (NS) (Figure 1). Although its role in benign and malignant melanocytic lesions remains uncharacterized, based on these studies, it is likely that OPN3 modulates proliferation, pigmentation, and apoptosis of malignant melanocytes in tumor initiation and progression.

Figure 1

The height of each bar represents the median expression. The expression of OPN3 in SKCM is higher than that in normal tissues. SKCM, skin cutaneous melanoma; TCGA, The Cancer Genome Atlas.

Results

Here, we performed single nucleotide variant (SNV) analysis of OPN3 in 68 malignant melanoma (MM), 166 melanocytic nevi (MNs), and 42 NS tissues from a Chinese population (Table 1). This study of formalin-fixed, paraffin-embedded tissues was approved by the ethics committee of Affiliated Hospital of Guizhou Medical University (Guiyang, China). Five nonsynonymous mutations were identified in the melanocytic lesions, including c.T152C, c.T401C, c.G547A, c.G768A, and c.G992A (Table 2). The c.T152C mutation was detected in one case of acral lentiginous melanoma. For the c.T401C mutation, MM mutations were much more common than MN and NS mutations, which were not associated with sex, age, localization, or histological subtype of samples. In addition, no significant differences were seen for SNVs in OPN3 in MM, MN, and NS (Table 2). In the c.G768A mutation of MM and MN, all five MM lesions displayed acral lentiginous melanoma. Of them, three MNs were found to be junctional nevi, and two were compound nevi. Overall, 3 of 68 MMs (4.41%) and 14 of 166 MNs (8.43%) revealed simultaneous OPN3 and BRAF V600E mutations. In 2 of 40 acral lentiginous melanoma cases (5%) with hotspot mutant BRAF V600E, the c.G768A mutation in OPN3 was also identified. In MM, the remaining one case with two mutations was metastatic melanoma (Table 3).

Table 1

Participants’ Characteristics of the Study from Chinese Patients with MN, MM, and NS

Subject	MN		MM		NS
	Total (n)	BRAFV600E (n)	Total (n)	BRAFV600E (n)	Total (n)
Number	166	81	68	7	42
Sex
Male	67	27	29	4	18
Female	99	54	39	3	26
Age, median (IQR)	30.0 (18.0–47.0)	33.00 (23.00–47.00)	65.0 (53.0–72.0)	42.0 (39.0–57.0)	44.0 (35.0–60.0)
Anatomical site
Head	16	10	1	0	3
Face	78	54	4	0	13
Neck	4	3	0	0	4
Trunk	15	9	3	2	11
Limbs	24	1	6	2	6
Hand	9	0	10	1	3
Foot	20	4	39	2	2
Brain	—	—	2	0	#
Lymph node	—	—	3	0	#
Subtype
Junctional nevus	28	3	—	—	—
Compound nevus	52	29	—	—	—
Intradermal nevus	53	49	—	—	—
Blue nevus	33	0	—	—	—
Superficial spreading melanoma	—	—	10	1	—
Nodular melanoma	—	—	7	2	—
Lentigo maligna melanoma	—	—	4	0	—
Acral lentiginous melanoma	—	—	40	2	—

Abbreviations: IQR, interquartile range; MM, malignant melanoma; MN, melanocytic nevus; NS, normal skin.

BRAF status was previously assessed by direct sequencing (Sanger) in all cases.

Table 2

Analysis of OPN3 SNVs in MM, MN, and NS Groups

OPN3 SNVs	MM (n = 68), %	MN (n = 166), %	NS (n = 42), %	P-Value
Exon
Exon 1 c.T152C A>G
AG	1.47 (1/68)	0	0	0.2106
AA	97.06 (66/68)	100 (166/166)	100 (42/42)
NA	1.47 (1/68)	0	0
Exon 2 c.T401C A>G
AG	4.41 (3/68)	1.20 (2/166)	2.38 (1/42)	0.1122
AA	92.65 (63/68)	98.80 (164/166)	97.6 (41/42)
GG	1.47 (1/68)	0	0
NA	1.47 (1/68)	0	0
Exon 2 c.G547A C>T
CT	10.29 (7/68)	9.04 (15/166)	14.28 (6/42)	0.6810
CC	93.93 (60/66)	90.36 (150/166)	85.71 (36/42)
TT	0	0.60 (1/166)	0
NA	1.47 (1/68)	0	0
Exon 3 c.G768A C>T
CT	7.35 (5/68)	3.01 (5/166)	9.52 (4/42)	0.1440
CC	92.66 (63/68)	95.78 (159/166)	90.48 (38/42)
NA	0	1.20 (2/166)	0
Exon 4 c.G992A C>T
CT	0	1.81 (3/166)	2.38 (1/42)	0.4948
CC	98.53 (67/68)	97.60 (162/166)	97.6 (41/42)
NA	1.47 (1/68)	0.60 (1/166)	0
3′-UTR
KMO(c.∗1253G>A), OPN3(c.∗1022C>T) rs371271799
GA	1.47 (1/68)	0	0	0.2663
GG	66.18 (45/68)	47.59 (79/166)	100 (42/42)
NA	32.35 (22/68)	52.41 (87/166)	0
KMO(c.∗1498A>G), OPN3(c.∗777T>C) rs144936606
AG	1.47 (1/68)	0.60 (1/166)	0	0.6418
AA	97.06 (66/68)	99.40 (165/166)	100 (42/42)
NA	1.47 (1/68)	0	0
KMO(c.∗1932G>C), OPN3(c.∗343C>G)
GC	1.47 (1/68)	0	0	0.2086
GG	94.12 (64/68)	96.99 (161/166)	100 (42/42)
NA	4.41 (3/68)	3.01 (5/166)	0
KMO(c.∗1955_∗1956insA),OPN3(c.∗319_∗320insT)
A/ins_A	38.24 (26/68)	37.35 (62/166)	28.57 (12/42)	0.5028
A/A	60.29 (41/68)	62.05 (103/166)	71.43 (30/42)
NA	1.47 (1/68)	0.60 (1/166)	0
KMO(c.∗2094G>T), OPN3(c.∗181C>A) rs3765809
GT	13.24 (9/68)	15.06 (25/166)	19.05 (8/42)	0.7272
TT	0	0	4.76 (2/42)	0.04042
GG	85.29 (58/68)	84.34 (140/166)	76.19 (32/42)
NA	1.47 (1/68)	0.60 (1/166)	0
KMO(c.∗2146C>T), OPN3(c.∗129G>A)
CT	0	1.20 (2/166)	0	0.5140
CC	98.53 (67/68)	98.19 (163/166)	100 (42/42)
NA	1.47 (1/68)	0.60 (1/166)	0
KMO(c.∗2192G>A), OPN3(c.∗83C>T)
GA	0	0.60 (1/166)	0	0.7171
GG	100 (68/68)	99.40 (165/166)	100 (42/42)
NA	0	0	0
KMO(c.∗2258G>A), OPN3(c.∗17C>T) rs199779503
GA	1.47 (1/68)	0	0	0.2122
GG	97.06 (66/68)	99.40 (165/166)	100 (42/42)
NA	1.47 (1/68)	0.60 (1/166)	0
KMO(c.∗2267C>G), OPN3(c.∗8G>C) rs201495076
CG	1.47 (1/68)	1.20 (2/166)	0	0.7469
CC	97.06 (66/68)	98.19 (163/166)	100 (42/42)
NA	1.47 (1/68)	0.60 (1/166)	0
5′-UTR
OPN3(c.-80A>G) rs7513575
TC	1.47 (1/68)	0.60 (1/166)	0	—
CC	1.47 (1/68)	0	100 (42/42)
NA	97.06 (66/68)	99.40 (165/166)	0
OPN3(c.-102T>C) rs7513451
AG	1.47 (1/68)	0.60 (1/166)	0	—
GG	1.47 (1/68)	0	100 (42/42)
NA	97.06 (66/68)	99.40 (165/166)	0
Intron
rs45572340 T>C
TC	8.82 (6/68)	9.04 (15/166)	11.90 (5/42)	0.8494
CC	2.94 (2/68)	0.60 (1/166)	0	0.2244
TT	85.29 (58/68)	89.76 (149/166)	88.10 (37/42)
NA	2.94 (2/68)	0.60 (1/166)	0
OPN3 241767624_53 T>G
TG	1.47 (1/68)	0	0	0.2051
TT	92.65 (63/68)	96.39 (160/166)	0
NA	5.8 (4/68)	3.61 (6/166)	100 (42/42)
rs140858921 A>G
AG	4.41 (3/68)	4.22 (7/166)	0	0.3931
GG	0	0.60 (1/166)	0	0.7185
AA	89.71 (61/68)	91.57 (152/166)	100 (42/42)
NA	5.8 (4/68)	3.61 (6/166)	0
rs632966 G>A
GA	30.88 (21/68)	33.73 (56/166)	21.43 (9/42)	0.3058
AA	64.71 (44/68)	62.05 (103/166)	76.19 (32/42)	0.2297
GG	4.41 (3/68)	4.22 (7/166)	2.38 (1/42)
NA	0	0	0

Abbreviations: MM, malignant melanoma; MN, melanocytic nevus; NA, not applicable; NS, normal skin; SNV, single nucleotide variant; UTR, untranslated region.

1Two groups were compared by Fisher’s exact test.

P < 0.05.

Table 3

OPN3 Nonsynonymous SNVs in MM, MN, and NS

OPN3 Mutation	Base Substitution	Amino Acid Substitution	MM (n = 68)	MS (n = 166)	NS (n = 42)	Concomitant BRAF V600E Mutation
						MM	MN
Exon 1	c.T152CA>G	Ile>Thr	1	0	0	0	—
Exon 2	c.T401CA>G	Val>Ala	4	2	1	0	2
Exon 2	c.G547AC>T	Val>Ile	7	15	6	1	9
Exon 3	c.G768AC>T	Met>Ile	5	5	4	2	1
Exon 4	c.G992AC>T	Cys>Tyr	0	3	1	—	2

Abbreviations: Ala, alanine; Cys, cysteine; Ile, isoleucine; Met, methionine; MM, malignant melanoma; MN, melanocytic nevus; NS, normal skin; SNV, single nucleotide variant; Thr, threonine; Val, valine.

We predicted the functional consequences of amino acid substitutions via three prediction tools (Polymorphism Phenotyping version 2, PROVEAN [Protein Variation Effect Analyzer], MutationTaster2) to evaluate the pathogenic potential of these nonsynonymous mutations (Table 4) (Schwarz et al., 2014). The prediction results of the c.G768A variant by the three tools was the most consistent, suggesting that it has deleterious effects. The results of two prediction tools showed that the c.T152C variant was also probably damaging. Other mutations were predicted to be deleterious by one of the three prediction tools. Moreover, mutant protein three-dimensional models were generated via the remote homology detection method of Phyre2 (Kelley et al., 2015) (Figure 2).

Table 4

Prediction Evaluation for OPN3 Nonsynonymous SNVs with Three Kinds of Bioinformatics Software (PolyPhen-2, PROVEAN, MutationTaster2)

OPN3 Mutation	Base Substitution	NCBI dbSNPReference ID	PolyPhen-2			PROVEAN		MutationTaster2
			HumDiv (0–1)	HumVar (0–1)	Prediction	Score (–14 to 14)	Prediction (Cutoff = –2.5)	PhyloP (–14 to 6)	PhastCons (0–1)	Prediction
Exon 1	c.T152CA>G	rs201734451	0.180	0.025	Benign	–7.63	Deleterious	1.404	0.991	Disease causing
Exon 2	c.T401CA>G	rs117720055	0.116	0.057	Benign	–9.15	Deleterious	0.693	0.627	Polymorphism
Exon 2	c.G547AC>T	rs2273712	0.001	0.002	Benign	–8.83	Deleterious	–0.604	0	Polymorphism
Exon 3	c.G768AC>T	rs78202695	0.985	0.977	Probably damaging	–4.76	Deleterious	5.232	1	Disease causing
Exon 4	c.G992AC>T	rs180909883	0.146	0.024	Benign	–2.10	Neutral	1.4	1	Disease causing

Abbreviations: dbSNP, SNP database; Div, division; Hum, human; ID, identification; NCBI, National Center for Biotechnology Information; PolyPhen-2, Polymorphism Phenotyping version 2; PROVEAN, Protein Variation Effect Analyzer; SNV, single nucleotide variant; Var, variant.

Figure 2

Mutant 3D models for all five nsSNVs in 3D, three-dimensional; Ala, alanine; Cys, cysteine; Ile, isoleucine; Met, methionine; nsSNV, nonsynonymous single nucleotide variant; Thr, threonine; Tyr, tyrosine; Val, valine.

In addition, nine SNVs in the 3′-UTR and two SNVs in the 5′-UTR of OPN3 were found in all MN and MM samples, including four novel SNVs of 3′-UTR: KMO(c.∗1932G>C), OPN3(c.∗343C>G); KMO(c.∗1955_∗1956insA), OPN3(c.∗319_∗320insT); KMO(c.∗2146C>T), OPN3(c.∗129G>A); and KMO(c.∗2192G>A), OPN3(c.∗83C>T) (Table 2). Two SNVs (OPN3[c.-80A>G] and OPN3[c.-102T>C]) in the 5′-UTR occurred in the same cases. Only two SNVs in the 3′-UTR (OPN3[c.∗319_∗320insT], rs3765809) were detected in the NS control group. The homozygote (TT) variant of rs3765809 had significant differences between the MN and NS groups (P < 0.05). Furthermore, we predicted the effects of SNVs in the 3′-UTR via miRDB and TargetscanHuman7.2 (Peng et al., 2020), which predicted microRNA (miRNA) targets in mammals. In 3′-UTR variants, as biological targets of miRNAs, seven miRNAs (has-miR-1272, has-miR-1267, has-miR-376c-3p, has-miR-6507-3p, has-miR-10399-5p, has-miR-137-5p, and has-miR-376c-3p) were predicted, one (has-miR-376c-3p) of them with a conserved 8mer (monomeric unit) site. This 3′-UTR SNV was detected in one intradermal nevus. In addition, four intronic SNVs were detected in all samples, including rs45572340, OPN3 241767624_53 T>G, rs140858921, and rs632966 (Table 2).

Discussion

In this study, we comprehensively identified OPN3 genetic variants in patients with MM or MN. Five nonsynonymous SNVs of OPN3 were detected in our results, and predictions of functional phenotypes provided evidence that these SNVs altered OPN3 conservation, especially for c.G768A, suggesting that this missense variant is deleterious. Crystal structure of OPN3 has been not resolved; we constructed the mutant three-dimensional structures of OPN3 by homology modeling based on OPN2. Although we could not demonstrate a significant statistical difference between MM and MN in OPN3 SNVs, which may be due to the small sample size and therefore lack of statistical power, some implications can be suggested. OPN3 is a cell surface receptor of the G-protein coupled receptor family that plays a vital role in the regulation of proliferation, migration, and survival (Bar-Shavit et al., 2016). In addition, OPN3 expression influences melanocyte apoptosis (Wang et al., 2020). Therefore, c.G768A or other nonsynonymous SNVs could alter OPN3 protein structure and function and have an impact on melanocytic proliferation and apoptosis in MM and MN. BRAF mutations are low frequency in nature (about 15%) (Nakamura and Fujisawa, 2018). In all 40 samples of acral lentiginous melanoma, only two cases showed the BRAF V600E mutation. Both of them also contained the c.G768A variant in OPN3. This suggested that the products of multiple gene mutations may affect melanocytic proliferation and tumor formation. In addition, the 3′-UTR SNV (OPN3[NM_014322:c.∗83C>T]) is predicted to disrupt a conserved miRNA (has-miR-376c-3p) target site, located in position 86–93 of OPN3 3′-UTR. It is possible that this UTR SNV might influence RNA stability or posttranscriptional regulation of OPN3 (Mayr, 2017), which may restrain melanocytic proliferation and malignant transformation by downregulation of OPN3 expression.

In this study, because we focused on overall OPN3 SNVs, additional investigation will be necessary to further elucidate whether or not these variants affect the formation and growth of melanocytic lesions.

Materials and Methods

Study population and data collection

All subjects with MMs and MNs were collected at Affiliated Hospital of Guizhou Medical University from January 2015 to December 2019 (Table 1). The control skin samples were obtained from normal adjacent nevi tissues. H&E-stained sections were reviewed by an experienced pathologist, and cases fulfilling criteria for the appropriate diagnoses (MM and MN) were selected for study. The study was approved by the Ethics Committees of our institution (Affiliated Hospital of Guizhou Medical University) and was performed according to the Declaration of Helsinki.

DNA extraction

DNA extraction from formalin-fixed, paraffin-embedded tissue was performed using an FFPE DNA Extraction Kit (AmoyDx, Xiamen, China), following the manufacturer’s instructions. We measured the concentration of DNA using a Qubit 2.0 (Thermo Fisher Scientific, Waltham, MA) to ensure that adequate amounts of high-quality genomic DNA had been extracted.

Multiplex PCR and sequencing as described by Wei et al. (2020)

Library preparation was performed by two-step PCR. The first round PCR reaction was set up as follows: DNA (10 ng/μl) 2 μl; amplicon PCR forward primer mix (10 μM) 1 μl; amplicon PCR reverse primer mix (10 μM) 1 μl; and 2× PCR Ready Mix 15 μl (total 25 μl) (Kapa HiFi Ready Mix). The plate was sealed and PCR performed in a thermal instrument (T100TM, Bio-Rad, Hercules, CA) using the following program: one cycle of denaturing at 98 °C for 5 minutes, eight cycles of denaturing at 98 °C for 30 seconds, annealing at 50 °C for 30 seconds, elongation at 72 °C for 30 seconds, 25 cycles of denaturing at 98 °C for 30 seconds, annealing at 66 °C for 30 seconds, elongation at 72 °C for 30 seconds, and a final extension at 72 °C for 5 minutes with a final hold at 4 °C. The PCR products were checked using electrophoresis in 1 % (w/v) agarose gels in Tris, boric acid, and EDTA buffer stained with ethidium bromide and visualized under UV light. Then we used AMPure XP beads to purify the amplicon product. After that, the second round of PCR was performed. PCR reaction was set up as follows: DNA (10 ng/μl) 2 μl; universal P7 primer with barcode (10 μM) 1 μl; universal P5 primer (10 μM) 1 μl; and ×2 PCR Ready Mix 15 μl (total 30 μl) (Kapa HiFi Ready Mix). The plate was sealed, and PCR performed in a thermal instrument (T100TM, Bio-Rad) using the following program: one cycle of denaturing at 98 °C for 3 minutes, five cycles of denaturing at 94 °C for 30 seconds, annealing at 55 °C for 20 seconds, elongation at 72 °C for 30 seconds, and a final extension at 72 °C for 5 minutes. Then we used AMPure XP beads to purify the amplicon product. The libraries were then quantified and pooled. Paired-end sequencing of the library was performed on the HiSeq XTen sequencers (Illumina, San Diego, CA).

Data quality control and SNV calling as described by Wei et al. (2020)

Raw reads were filtered according to two steps: (i) removing adaptor sequence if reads contain by cutadapt (version 1.2.1) and (ii) removing low quality bases from reads 3′–5′ (Q < 20) by PRINSEQ-lite (version 0.20.3). The remaining clean data were mapped to the reference genome by BWA (version 0.7.13-r1126) with default parameters. SAMtools (version: 0.1.18) was used to calculate each genotype of target site. ANNOVAR (16 April 2018) was used to detect genetic variants.

Prediction of amino acid substitution

We evaluated the functional consequences of amino acid substitutions via three prediction tools, Polymorphism Phenotyping version 2 (http://genetics.bwh.harvard.edu/pph2/), PROVEAN (http://provean.jcvi.org/index.php), and MutationTaster2 (http://www.mutationtaster.org/), and compared predictions of the three tools on OPN3 nonsynonymous SNVs. The web versions of the three prediction tools are used to predict the pathogenic potential of DNA sequence alterations. In addition, we built three-dimensional models of OPN3 to analyze the effect of SNVs on the protein sequence using SWISS-MODEL (https://swissmodel.expasy.org/). We also predicted the effects of SNVs in the 3′-UTR via miRDB (http://mirdb.org/) and TargetscanHuman7.2 (http://www.targetscan.org/vert_72/), which predicted miRNA targets in mammals. Details about the methods and further statistics followed their websites and previous reports.

Statistical analyses

All data were entered into GraphPad Prism (version 8.0) for statistical analysis. Categorical data were analyzed using Fisher’s exact test. A two-tailed P < 0.05 was considered statistically significant.

Data availability statement

No large datasets were generated or analyzed during this study. Minimal datasets necessary to interpret and/or replicate data in this paper are available on request to the corresponding author.

Ethics Statement

This study of formalin-fixed, paraffin-embedded tissues was approved by the ethics committee of Affiliated Hospital of Guizhou Medical University (Guiyang, China). Under Chinese law, written consent from the patients was not required because the material used had been collected for diagnostic and therapeutic purposes in the archives of the Institute for Pathology, Affiliated Hospital of Guizhou Medical University (Guiyang, China) and used for this study in pseudonymized form.

ORCIDs

Wei Zhang: http://orcid.org/0000-0003-1796-0182

Jianglong Feng: http://orcid.org/0000-0002-3667-415X

Wen Zeng: https://orcid.org/0000-0002-8251-4624

Zhixu Zhou: http://ocid.org/0000-0002-9633-9521

Yu Wang: http://orcid.org/0000-0002-8043-9165

Hongguang Lu: http://orcid.org/0000-0002-5002-4276

Author Contributions

Conceptualization: HL, YW, WZh; Data Curation: JF, WZe; Formal Analysis: WZh, WZe, ZZ; Funding Acquisition: HL; Writing - Original Draft Preparation: WZh, HL; Writing - Review and Editing: HL