Synchronous Epidermodysplasia Verruciformis and Intraepithelial Lesion of the Vulva Is Caused by Coinfection With Alpha-Human Papillomavirus and Beta-Human Papillomavirus Genotypes and Facilitated by Mutations in Cell-Mediated Immunity Genes
Context.—
There have been exceedingly few reports of epidermodysplasia verruciformis (EV) or EV-like lesions in the vulva. We describe the first observation of vulvar lesions displaying synchronous EV-like histology and conventional high-grade squamous intraepithelial lesion (HSIL), a finding hitherto unreported in medical literature.
Objectives.—
To describe this novel vulvar lesion with hybrid features of HSIL and EV, attempt to confirm the hypothesis of coinfection with α and β human papillomavirus (α-HPV and β-HPV) genotypes, and describe relevant underlying genetic mutations.
Design.—
Cases were retrospectively selected from our institutional archive. Detailed review of clinical information, histologic examination, and whole genome sequencing (WGS) were performed.
Results.—
Five samples from 4 different patients were included. Three of 4 patients had a history of either iatrogenic immune suppression or prior immune deficiency, and all 3 featured classic HSIL and EV changes within the same lesion. One patient had no history of immune disorders, presented with EV-like changes and multinucleated atypia of the vulva, and was the sole patient without conventional HSIL. By WGS, several uniquely mappable reads pointed toward infection with multiple HPV genotypes, including both α-HPVs and β-HPVs. Mutations in genes implicated in cell-mediated immunity, such as DOCK8, CARMIL2, MST1, and others, were also found.
Conclusions.—
We provide the first description of vulvar lesions harboring simultaneous HSIL and EV features in the English-language literature, a phenomenon explained by coinfection with α-HPV and β-HPV genotypes. The finding of EV-like changes in a vulvar specimen should prompt assessment of the patient’s immune status.
Human papillomaviruses (HPVs) are nearly ubiquitous among the adult population, with more than 200 recognized genotypes across 5 genera.1 Infection by α-HPV subtypes (eg, HPV 16 and HPV 18) is a part of routine practice for many physicians because these viruses show tropism for both cutaneous and mucosal surfaces and are implicated in the pathogenesis of dysplastic and neoplastic lesions of several organs, including the cervix, vagina, and vulva.2–4
Contrarily, β-HPV subtypes display stricter cutaneous tropism and normally produce asymptomatic infections. Rare subsets of patients develop multiple flat wartlike lesions or hypopigmented/pityriasis versicolor-like macules, with increased risk of progression to cutaneous invasive squamous cell carcinoma, a condition called epidermodysplasia verruciformis (EV).
Originally described in 1922 by Lewandowsky and Lutz,5 the prototypical form of EV is a rare autosomal recessive genodermatosis, in which patients carrying biallelic amorphic mutations of EVER1/TMC6 and EVER2/TMC8 genes develop skin lesions after exposure to β-HPVs, especially HPV 5.1,6,7
Although reports of immunosuppressed renal transplantation patients showing EV-like skin changes emerged in the 1980s,8 it was only in 2009 that Rogers et al9 established the concept of acquired EV in the medical literature by describing 2 cases of patients displaying EV-like cutaneous lesions in the context of human immunodeficiency virus (HIV) infection.
The discovery of additional genes implicated in inherited EV (such as RHOH, MST-1, and CORO1A) has further expanded our current understanding of the disease. This led to the proposal of an updated classification, in which the genetic form of EV can be further stratified as classic and nonclassic and the acquired form of EV can be subclassified as HIV-related and iatrogenic.6
Little is known regarding the clinicopathologic presentation of EV or β-HPV infections in the gynecologic tract, which appears to be a rare phenomenon. Only 3 cases of EV-related vulvar lesions have been reported thus far, 2 of them occurring in immunosuppressed patients10 and the most recent in a patient with the inherited form of EV.11
We have recently examined a case displaying simultaneous usual type high-grade vulvar intraepithelial lesion (VIN 3) and EV-like changes. We hypothesized that this presentation could be caused by coinfection with both α-HPV and β-HPV subtypes. With the intent to investigate this unusual occurrence, we reviewed all cases of vulvar lesions labeled with “epidermodysplasia verruciformis” in our institution, and performed whole genome sequencing (WGS) on DNA retrieved from formalin-fixed, paraffin-embedded tissue blocks, with emphasis on HPV DNA detection and previously described EV susceptibility genes.
MATERIALS AND METHODS
Case Series
After approval by the Women & Infants Hospital of Rhode Island Institutional Review Board (Providence, Rhode Island), we located 5 samples from 4 different patients on our electronic records that consisted of vulvar specimens labeled with the terms “epidermodysplasia verruciformis” or “epidermodysplasia verruciformis-like.” A detailed histologic review of previously mentioned samples and all other relevant material (vulvar, vaginal, and cervical biopsies) from the same patients was performed by 2 of the authors. Clinical history, treatment, and follow-up information was collected. All procedures were performed in compliance with relevant laws and institutional guidelines.
Immunohistochemistry
Immunohistochemical stains performed at the time of original diagnosis were also reviewed. Dako’s ER PharmDx Kit (catalog No. SK310) was used on the Dako Autostainer Link 48 instrument. Tissue sections (4 µm thick) were dewaxed (60°C in oven for 1 hour) and rehydrated. Antigen retrieval in a Dako Pascal pressure cooker (pH 6.1) was followed by peroxidase blocking (×5 minutes) incubation with primary antibody (Dako’s ER clone ER-2-123 + 1D5), secondary antibody (both for 30 minutes at room temperature) and diaminobenzidine chromogen-substrate system (×10 minutes). The following antibodies were employed: p16INK4a antigen (BioCare Medical, Pacheco, California); Ki-67 (MIB-1, Dako, Santa Clara, California); and p53 (DO-7, Dako).
HPV DNA Testing
HPV DNA polymerase chain reaction (PCR) testing was available for patients 2, 3, and 4. In our institution, this testing is performed in tandem with cervical cytology screening using the Cobas 6800 System (Roche Diagnostics, Rotkreuz, Switzerland).
Whole Genome Sequencing
DNA fragmentation, library preparation, and sequencing were conducted by Genewiz Inc. Sequencing configuration was Illumina MiSeq, 2 × 150 bp per lane with estimated data output of ∼10 M raw paired-end reads per flow cell and quality guarantee more than 80% of bases Q30 or higher. Bioinformatics analysis was assisted by GeneVia Technologies (Tampere, Finland). The genomic DNAs were fragmented and the libraries were prepared using Illumina Compatible Paired-End Sequencing kit followed by 150-bp paired-end sequencing on an Illumina HiSeq sequencing machine. The short-read variant calling was performed using the Dragen variant calling pipeline.12 Furthermore, the called short variants were annotated using Variant Effect Predictor.13 HPV subtypes were identified by aligning the reads against a composite genome of human genome (version hg38) and a collection of HPV reference sequences and quantifying the number of unique reads mapping to each HPV subtype. The reads were aligned and quantified using BWA-mem14 and Samtools,15 respectively. The HPV reference sequences were obtained from the Papillomavirus Episteme (PAVE) database.16
RESULTS
A total of 22 samples from 4 patients were reviewed, including vulvar wide local excisions or biopsies and cervical excisions or biopsies. EV changes were seen in 5 of 22 samples, all of which were from the vulva. Patient age ranged from 29 to 63 years. Clinicopathologic findings of all 4 patients are summarized in Table 1.
A total of 3 of 4 patients (75%) presented with multiple lesions, whereas 1 patient (25%) presented with a single lesion. All 3 patients with multiple lesions had a history of either immune suppression (2 patients) or immune deficiency (1 patient). The single patient with a unifocal lesion had no history of immune suppression. Lesions were located on the inner thighs, labia minora, labia majora, buttocks, perianal region, right mons/labia junction, and left fourchette. All patients were treated with wide local excisions.
Histologically, the vulvar wide local excision specimens from the 3 patients with multiple lesions showed a mixture of high-grade squamous VIN 3 and EV features, albeit with morphologic nuance between them.
One patient displayed a full-thickness proliferation of keratinocytes with high-grade nuclear atypia, and abundant mitotic and apoptotic figures, consistent with VIN 3 (Figure 1, A). Classic EV features, such as keratinocytes with blue-gray cytoplasm, coarse keratohyalin granules, and abnormally enlarged nuclei with minute nucleoli, were present (Figure 1, B and C). Some areas showed a hybrid of VIN 3, koilocytotic atypia, and partial EV features (Figure 1, D).
Citation: Archives of Pathology & Laboratory Medicine 148, 9; 10.5858/arpa.2023-0193-OA
One patient showed a usual-type high-grade vulvar intraepithelial lesion (Figure 2, A) and extensive classic EV features. These changes were quite extensive when compared to other cases in the series, and they were present not only in superficial keratinocytes but also in deeper layers of the epithelium (Figure 2, B). Perhaps because of the degree of EV histopathologic changes, unusual features were observed. In certain areas the bluish gray cytoplasm had a vacuolated or “bubbly” appearance (Figure 2, C). In smaller foci the vacuoles seemed to coalesce into larger vesicles that displaced the nuclei (Figure 2, D). The same patient already showed synchronous VIN 3 and EV features in a wide local excision performed 3 years prior (Figure 3).
Citation: Archives of Pathology & Laboratory Medicine 148, 9; 10.5858/arpa.2023-0193-OA
Citation: Archives of Pathology & Laboratory Medicine 148, 9; 10.5858/arpa.2023-0193-OA
Another patient displayed a strikingly unusual pattern of intraepithelial lesion. The atypical keratinocytes had a prominent basaloid phenotype, abundant atypical mitotic figures, and apoptotic bodies, and were seen scattered throughout the epithelium (Figure 4, A). Only in some areas did the atypia reach full-thickness involvement, where a more conventional phenotype was seen (Figure 4, B). Small foci of partial EV features were seen (Figure 4, C), as well as areas of conventional koilocytotic atypia (Figure 4, D).
Citation: Archives of Pathology & Laboratory Medicine 148, 9; 10.5858/arpa.2023-0193-OA
Alternatively, the patient with a single lesion and no history of immunodeficiency showed marked hyperkeratosis and epidermal squamous hyperplasia (Figure 5, A). The keratinocytes showed occasional nuclear atypia, with binucleated forms (Figure 5, B). Classic EV features were seen in 2 small foci, notable for the more basal location of affected cells (Figure 5, C). Multinucleated keratinocytes were seen, reminiscent of what has been previously described as “multinucleated atypia of the vulva” (Figure 5, D). By immunohistochemistry the lesion was negative for p16 and showed wild-type p53 expression and habitual basal Ki-67 expression.
Citation: Archives of Pathology & Laboratory Medicine 148, 9; 10.5858/arpa.2023-0193-OA
The follow-up period ranged from 7 to 168 months. The 3 patients with multiple lesions developed recurrences, 1 reporting a spontaneous “ebb and flow” of similar vulvar lesions that would spontaneously resolve and recur, and the other 2 treated with wide local excision. The patient with a single lesion did not have a recurrence as of last follow-up. Two patients in our series had a long history of cervical dysplasia, but none of the available cervical specimens for review displayed EV features.
WGS: Coinfection by Multiple HPV Subtypes
Uniquely mappable reads to the HPV reference genomes of both α-HPV and β-HPV subtypes were found in all 5 samples. Figure 6 summarizes the HPV mapping statistics of the cohort using a “heat map” distribution.
Citation: Archives of Pathology & Laboratory Medicine 148, 9; 10.5858/arpa.2023-0193-OA
Among the β-HPV genotypes most represented in the cohort are β-HPV 5 (found in 4 samples from 3 patients) and β-HPV 47 (found in 2 samples from 2 patients). The α-HPV genotype most represented in the cohort is α-HPV 53 (found in 3 samples from 3 patients).
WGS: Short Variants in EV Susceptibility Genes
Three of the four patients harbored likely-pathogenic short variants in genes previously described as implicated in EV. Table 2 presents a summary of these findings.
Likely-pathogenic frameshift variants in TMC6 were found in 2 samples from 2 different patients, including 1 featuring EV and VIN 3 and the other showing multinucleated atypia of the vulva (MAV) and EV. One of these 2 patients also displayed a likely-pathogenic frameshift variant in CORO1A.
Likely-pathogenic frameshift variants in LCK, DOCK8, MST1, and CARMIL2 were found in the 2 samples originating from a single patient.
Further details regarding the aforementioned mutations are provided in the Supplemental Table (see supplemental digital content at https://meridian.allenpress.com/aplm in the September 2024 table of contents).
DISCUSSION
To the best of our knowledge, this is the first case series reporting synchronous conventional vulvar intraepithelial lesion and acquired EV of the vulva. Previous reports in which classic histologic features of EV were present in vulvar specimens include 2 cases reported by Pohthipornthawat et al10 and 1 case by Singh et al.11
In the former study both patients displayed exclusively EV histologic changes, with no concomitant classic vulvar intraepithelial lesion.10 Presumably, this can be attributed to iatrogenic acquired EV, considering 1 patient had received a stem cell transplant and the other was status post–multiorgan transplant prior to the vulvar biopsies. Both patients tested positive for β-HPV 5 by PCR performed with DNA extracted from the tissue specimens.
Alternatively, the latter study11 presents a young patient with the inherited form of EV who received a diagnosis of high-grade vulvar intraepithelial lesion and simultaneous EV changes plus high-grade squamous intraepithelial lesion of the uterine cervix. The cervical biopsy showed no EV features. This is consistent with the findings reported in the present study, because 2 of the 4 patients had a long history of cervical dysplasia, but none of the cervical specimens reviewed showed the EV features seen in their respective vulvar biopsies. This could be explained by the fact that β-HPV subtypes show stricter cutaneous tropism when compared with α-HPV subtypes that can infect both skin and mucosal surfaces.1,4 In fact, the patient reported by Singh et al11 was positive for HPV 16 and other high-risk α-HPV subtypes by PCR testing on the cervical sample. These would putatively be responsible for the patient’s concurrent cervical and vulvar high-grade intraepithelial lesion. No testing for β-HPVs was performed in that study. In our cohort, both patients were also positive for other high-risk α-HPV subtypes by PCR testing on cervical samples.
Although reports of EV histologic features on vulvar biopsies are recent and scarce, DNA of EV-related HPV subtypes was isolated from vulvar specimens as early as 1982,7 a time in which many HPV genotypes and their causal relationship to human cancer were yet to be characterized. In that study, DNA from an HPV subtype tentatively named “HPV-EV” was isolated from the skin lesions in the hands and legs of a patient with known EV. Homologous “HPV-EV” DNA was extracted from a Bowenoid carcinoma “in situ” of the vulva in the same patient. There was no detailed histologic characterization of this tumor.
Four years later, de Villiers et al17 screened 268 biopsies of a diverse cohort of genital tumors in search of HPV DNA. Three condylomata in that study were found to have unexpected HPV genotypes: 2 tested positive for HPV 3–related DNA and 1 tested positive for EV-related HPV DNA. Once again, no morphologic description was provided in the article.
The wide temporal gap between the observations regarding EV and gynecologic pathology likely originates from the fact that α-HPV subtypes, especially high-risk genotypes such as HPV 16 and HPV 18, have been the drivers of research and advancements in the field.2,3 In current clinical practice, routine testing understandably evaluates almost exclusively these high-risk subtypes.
By WGS, all 4 patients in our study demonstrated DNA reads mapping uniquely to both α-HPV and β-HPV reference genomes. Although several HPV genotypes could be found, most reads clustered in between 2 and 5 genotypes, which are more likely to represent evidence of actual infection. The “heatmap” distribution used in Figure 6 displays most of the HPV subtypes found but emphasizes the greater likelihood of infection by the genotypes from which most reads come. The lower number of reads from other HPV subtypes could be attributed to infection with a low number of viral copies or attributed to noise, considering how similar the DNA sequences between each subtype can be. Nevertheless, in this clinicopathologic setting, the WGS findings confirm our hypothesis of coinfection leading to the dual phenotype of conventional high-grade dysplasia (VIN 3) and simultaneous EV histologic changes.
HPV 53 was the most prevalent α-HPV genotype in our cohort. Interestingly, this genotype has been linked to infection by multiple HPV genotypes, which in turn has been associated with persistence of infection and higher likelihood of developing high-grade intraepithelial lesions in different studies.18,19 The reason why HPV 53 in particular appears to have a greater role in multiple HPV infection, as opposed to single HPV infection, and its potential synergistic interactions with other HPV genotypes is currently unknown, and should merit scientific inquiry.
The presence of EV histologic features and DNA from EV-associated β-HPVs in vulvar biopsy samples opens new lanes of investigation. The 1 patient with a single lesion displaying EV features with no clinical history of immune deficiency or the inherited form of the disease represents an illustrative case in terms of uncharted territories. By histology alone, a hyperkeratotic lesion of the vulva displaying focal yet classic EV features and MAV is a puzzling finding. The patient had no signs of recurrence during the follow-up period, which may indicate that immunocompetent patients could present with transient or self-limiting, focal EV lesions.
The current understanding of MAV has remained mostly unchanged since the original report was published by McLachlin et al.20 MAV is thought to be a reactive change to the keratinocytes in a background of chronic irritation, and previous attempts to detect HPV DNA or evidence of infection by other viruses in such lesions have been unsuccessful.20–22 The 1 patient with MAV in our study demonstrated, by WGS, roughly 55% of reads mapping uniquely to the reference genomes of β-HPVs, namely HPVs 24, 5, and 47, all of which have been previously implicated in EV. It is entirely possible that these findings are coincidental, considering the hyperkeratotic/irritated nature of the patient’s lesion and the fact that cutaneous infection by β-HPVs in immunocompetent humans is thought to be mostly asymptomatic. However, the possibility of EV-related β-HPV subtypes being implicated in MAV needs to be explored.
Furthermore, other lesions previously regarded as HPV independent may be related to infection by β-HPV subtypes for which testing is not frequently performed. Vulvar vestibulitis or vulvitis, a disputed lesion with relatively vague clinicopathological characteristics,23 has been revisited in a 2008 study that reported persistent detection of novel EV-related HPV genotypes.24 A possible association between EV-related β-HPV subtypes and vulvovaginal melanomas has also been proposed.25
At this time, the relationship between β-HPV subtypes and vulvar cancer remains unclear. Is EV another risk factor for cancers of the vulva, at least in the population harboring some form of immune deficiency? What are the implications of coinfection with both α-HPV and β-HPV subtypes in terms of the pathogenesis and progression of high-grade vulvar intraepithelial lesion? Could EV-related HPV subtypes be implicated in the genesis of poorly understood vulvar lesions, such as MAV and vulvitis?
Our study also appears to be the first to report DNA sequencing in patients presenting clinically with the acquired form of EV. Frameshift variants in genes previously reported to be implicated in both classic inherited EV (TMC6)6 and nonclassic inherited EV (MST1 and CORO1A)6 were found in 3 patients in this series.
Additionally, mutations in LCK and DOCK8, found in 1 patient in this cohort, have been more recently described in patients who appear to harbor nonclassic inherited EV via cell-mediated immune deficiency.26,27
The same patient also harbored a frameshift variant in CARMIL2, a gene that has never been associated directly with EV skin lesions but is linked to a genetic immune deficiency syndrome that is implicated in myriad clinical findings, including Epstein-Barr virus–associated smooth muscle tumors, dermatitis, viral and fungal skin infections, and even early-onset inflammatory bowel disease.28–32
Unfortunately, because of the lack of nonlesion DNA samples from our patients, we cannot establish with absolute certainty which mutations are germ-line variants and which are somatic mutations. However, apart from the frameshift variant in DOCK8, all other variants found in our study have extremely low frequencies or seem to be “nonexistent” in population databases, which could indicate that they might be somatic. The DOCK8 variant found in 2 samples from a single patient, despite failing the quality filters in 1 of the samples, seems to harbor a higher likelihood of representing a germ-line variant. Additionally, the different mutational profiles found between these samples may represent evidence that the molecular background in nonclassic EV is not static over time.
The WGS findings in our study are intriguing because they provide initial insight into the possible mechanism for the development of acquired EV. We propose that mutations in genes related to cell-mediated immunity, whether germ line or somatic, are either a precondition or a putative risk factor for the development of acquired EV in the setting of an external immune suppressant (whether iatrogenic or infectious). This may partially explain the rarity of the disease, despite an ever-increasing lifespan and absolute number of patients under some form of immune suppression. Further studies are necessary to elucidate the pathogenesis of acquired EV, preferably with long term follow-up and sequencing performed on lesional samples as well as controls.
In summary, we report the first cohort in which histologic features of EV are seen in tandem with vulvar intraepithelial lesion, coexisting within the same surgical specimen. This phenomenon is caused by coinfection with both α-HPV and β-HPV genotypes and appears to be facilitated by immunosuppression and/or alterations in genes related to cell-mediated immunity. For clinical practitioners and pathologists, we propose that the finding of EV features in vulvar biopsies should prompt investigation of the patient’s immune status.
Hematoxylin-eosin–stained (A through D) photomicrographs displaying conventional high-grade squamous vulvar intraepithelial neoplasia (A). Classic histologic features of epidermodysplasia verruciformis were present (B and C), including bluish gray cytoplasmic color, coarse keratohyalin granules, and enlarged pale nuclei with small nucleoli. Focal koilocytotic atypia was also present (D) (original magnifications ×100 [A, C, and D] and ×200 [B]).
Hematoxylin-eosin–stained (A through D) photomicrographs displaying an abrupt transition from high-grade intraepithelial neoplasia (A, left) to areas of epidermodysplasia verruciformis (A, right). The epidermodysplasia verruciformis (EV) changes in this specimen were quite extensive, virtually present on the entire thickness of the epithelium (B). Focal vacuolization imparts a “bubbly” appearance to the otherwise characteristic bluish cytoplasm (C). In some areas the vacuoles appeared to coalesce, displacing the nuclei (D) (original magnifications ×40 [A], ×100 [B], and ×200 [C and D]).
Hematoxylin-eosin–stained photomicrograph displaying both changes of epidermodysplasia verruciformis (left) and high-grade vulvar intraepithelial neoplasia (VIN 3, right) (original magnification ×100).
Hematoxylin-eosin–stained (A through D) photomicrographs displaying a striking pattern of basaloid atypical cells scattered throughout the epithelium (A). Conventional areas of high-grade vulvar intraepithelial neoplasia (B). Partial features of epidermodysplasia verruciformis with mild bluing of the cytoplasm and enlarged, pale nuclei (C). Koilocytotic atypia was also noted (D) (original magnifications ×40 [A and B] and ×100 [C and D]).
Hematoxylin-eosin–stained (A through D) photomicrographs displaying a hyperkeratotic hyperplastic squamous proliferation (A). Mild spongiosis, nuclear atypia, and binucleated forms were present (B). Focal classic features of epidermodysplasia verruciformis were seen (C). Notice the bluish cytoplasmic hue, coarse keratohyalin granules, and the enlarged pale nuclei with small nucleoli. Multinucleated cells in the intermediary layers were present, consistent with multinucleated atypia of the vulva (D, arrow) (original magnifications ×40 [A], ×400 [B and C], and ×200 [D]).
Human papillomavirus (HPV) genotype “heatmap” distribution per sample, according to whole genome sequencing reads that map uniquely to each HPV genotype reference genome.
Contributor Notes
Supplemental digital content is available for this article at https://meridian.allenpress.com/aplm in the September 2024 table of contents.
Ribeiro e Ribeiro is currently with the Laboratório Cicap, Hospital Alemão Oswaldo Cruz, and the Departamento de Patologia, Instituto do Câncer do Estado de São Paulo (ICESP), both located in São Paulo, Brazil
The authors acknowledge the Luke Charitable Foundation (Pawtucket, Rhode Island) for assistance with funding.
The authors have no relevant financial interest in the products or companies described in this article.