Ovarian Failure Preceding Head and Neck Squamous Cell Carcinoma Identifies an Adult-Onset Cancer-Prone Syndrome Caused by FANCM Mutations

Head and neck squamous cell carcinoma (HNSCC) is considered environmentally induced with tobacco, alcohol, and human papillomavirus implicated in most cases. However, there is some evidence that genetic predisposition plays a role as evidenced by the increased incidence in two hereditary syndromes. Patients with short telomere syndromes have a 50-fold increased risk of HNSCC; this risk is particularly high in males.^1,2 By contrast, patients with Fanconi anemia have a up to a 500-fold increased risk, and the risk is higher in females.^3,4 In Fanconi anemia, HNSCC is diagnosed most often in patients with a history of bone marrow failure. The median age of diagnosis of HNSCC in Fanconi anemia is 30 years, and nearly all cases are diagnosed before the age of 40.⁵ In both of these DNA repair syndromes, the most common anatomic site is the oral cavity with the majority involving the anterior tongue.^2,5 Whether germline mutations in telomere-related and Fanconi anemia genes influence HNSCC susceptibility in older adults is not fully known.

A 59-year-old woman presented with T4aN2b squamous cell carcinoma (SCC) of the retromolar trigone of the oral cavity. Her clinical history was significant for biopsy-proven ovarian failure diagnosed at age 19, which showed absent ovarian follicles. She also reported history of toe melanoma diagnosed at age 50. Family history was negative for cancer and hematologic disease (Fig 1A), and she had unremarkable tobacco and alcohol exposures. Blood counts at presentation were normal. Given the high risk of recurrence after the SCC resection at presentation, a plan for chemoradiation was implemented. However, after the initial dose of cisplatin, the patient developed severe, prolonged pancytopenia and mucositis; this subsequently recurred even after dose reduction and both episodes required hospitalization (Fig 1B). The chemotoxicity prompted evaluation for a genetic susceptibility syndrome. Telomere length measurement fell between the 10th and 50th age-adjusted percentiles in lymphocytes (Fig 1C) and granulocytes (not shown). This range is rarely seen in short telomere syndrome patients who present in this patient's age group. Since Fanconi anemia causes sensitivity to platinum compounds, diepoxybutane (DEB) testing was ordered and showed abnormally high levels of chromosome breakage with a mean of 3.92 breaks per metaphase (positive range: 1.12-9.5 breaks/cell v 0-0.08 breaks/cell in controls; Fig 1D). Germline genetic evaluation identified two variants in FANCM (Fanconi anemia complementation group M), a gene involved in meiosis and DNA interstrand crosslink (ICL) repair^6-8: c.1972C>T p.Arg658* and c.5494-5495delGA p.Glu1832Serfs*19. The c.1972C>T allele was rare with an allele frequency of 8 × 10⁻⁵ among more than 282,000 individuals in gnomAD, and the c.5494-5395delGA has not been previously reported. Both variants, if stable, were predicted to lack conserved C-terminus FANCM domains (Fig 1E). An asymptomatic brother who carried the p.Arg658* variant showed no sensitivity to DEB. If this second variant did not arise de novo, based on the recombination rate in the region spanning the two variants (29 Mb), the probability of the patient's mutations being biallelic was 99.97%.

FIG 1.

Clinical and molecular characterization of the index patient showing susceptibility to DNA interstrand crosslink repair. (A) Pedigree of proband with clinical features and age of onset shown below. (B) Timeline of leukopenia and neutropenia relative to cisplatin dosing as indicated by arrows and in the key above. (C) Telomere length in lymphocytes relative to age-adjusted nomogram as measured by flow cytometry and fluorescence in situ hybridization. (D) Metaphase spread showing chromosome breaks (red arrow) and triradial chromosomes (blue arrow). The mean number of abnormalities per metaphase is shown in the lower right. (E) FANCM protein scheme showing mutations above. (F) Example immunoblot showing FANCD2 levels pre- and post-MMC exposure. FANCD2-Ub refers to the monoubiquitinated protein. Protein lysates were derived from lymphoblastoid cell lines. (G) Quantification of FANCD2-Ub relative to total FANCD2 levels with each dot showing quantification from a single gel and from an independent protein extraction. Mean is shown with SEM. (H) Surviving fraction of LCLs after MMC exposure; each timepoint reflects mean of 3 replicates with error bars representing SEM. ANC, absolute neutrophil count; MMC, mitomycin C; SCC, squamous cell carcinoma.

To test the functional significance of the FANCM variants, we generated lymphoblastoid cell lines (LCLs) from the patient and her heterozygous brother. We then tested the cells' capacity to signal recognition of ICLs generated by mitomycin C. FANCM^+/Arg658* cells behaved similar to controls with respect to FANCD2 monoubiquitination (FANCD2-Ub) at baseline and postchallenge. They also showed no mitomycin C sensitivity in vitro (Fig 1H). However, baseline FANCM^{Arg658*/Glu1832fs} LCLs showed a lower baseline FANCD2-Ub fraction, although this defect was not as severe as FANCA biallelic mutant cells (Fig 1F). After mitomycin C exposure, FANCM^{Arg658*/Glu1832fs} LCLs also had lower FANCD2-Ub levels, nearly half of controls, although in this setting, this decrease was intermediate relative to FANCA-mutant cells (Figs 1F and 1G). We also examined the magnitude of the ICL defect by assessing cell survival in response to mitomycin C and found a dose-dependent effect in FANCM^{Arg658*/Glu1832fs} cells. In this assay, the cellular survival defect was similar to FANCA-mutant LCLs (Fig 1H). The Appendix lists the experimental methods.

We queried the experience of the authors at Johns Hopkins with HNSCC patients who developed severe cisplatin chemosensitivity but identified none among 500, consistent with the rarity of this presentation. Given the challenges in managing advanced cancer in patients with Fanconi anemia who cannot tolerate cytotoxic therapies, we examined whether the patient's tumors may have signatures of immune responsiveness. Whole-exome sequencing of the metastatic melanoma and oral cavity SCC showed an overall low tumor burden of 0.8/Mb and 2/Mb, respectively (Fig 2A). The number of missense variants per exome was 36 and 94 (Fig 2A), comparable to sun-shielded melanoma and HNSCC, which have means of 9 and 94 mutations, respectively.^9,10 However, although both tumors had negative programmed death-ligand 1 (PDL-1) staining, they showed a brisk T-cell immune infiltrate, which was CD8+ T-cell rich (Fig 2B).

FIG 2.

Pathologic and genomic landscape of melanoma and OCSCC derived from FANCM mutation carrier. (A) Summary of single nucleotide variants identified by whole-exome sequencing of tumors. The melanoma sequenced was obtained from resected metastasis to the adrenal gland. (B) Representative photomicrographs of patient-derived tumors (all 200×) with panels as labeled. CD3+ and CD8+ T cells are quantified in the lower right of those respective images and the quantification was based on all assessable tumor areas. PDL-1 staining was interpreted as negative given lack of membranous staining as per widely accepted criteria.¹⁹ H&E, hematoxylin and eosin; OCSCC, oral cavity squamous cell carcinoma; PDL-1, programmed death-ligand 1; TMB, tumor mutation burden.

We report here what we believe to be one of the oldest case presentations of Fanconi anemia and biallelic deleterious mutations in FANCM. The patient had an antecedent history of ovarian failure, which was also seen as an initial presentation of compound heterozygous FANCM mutations in the series that defined this entity in 2018.^7,8 The clinical phenotype is consistent with the known role of FANCM in meiosis and is consistent with the male and female infertility seen in animal models.^6,11 The history of primary ovarian failure, or the milder presentation of early menopause, provide a clinical clue that should increase the index of suspicion of a FANCM-related genetic susceptibility in patients with HNSCC and melanoma. Seven patients with cancer have been reported in the literature with FANCM biallelic mutations and their most common cancers include breast (N = 5; age range, 29-62 years) followed by HNSCC (N = 3), endometrial (N = 1), and acute lymphocytic leukemia (N = 1).^7,8 Aside from the shared susceptibility to HNSCC, the FANCM-mediated cancer spectrum is distinct clinically from that seen in classic Fanconi anemia where myelodysplastic syndromes or acute myeloid leukemia and liver tumors predominate. The patient continues to be cancer-free 3 years after resection of her oral cavity tumor and 8 years since resection and immunotherapy for metastatic melanoma. She is currently under active surveillance.

Bone marrow failure is diagnosed in 80% of classic Fanconi anemia patients before the age 40,¹² but the patient we studied, like other patients with FANCM biallelic mutations reported to date, had no antecedent bone marrow failure but developed severe persistent cytopenias only after platinum exposure. Despite the long interval since treatment, the patient's blood counts have not fully recovered with a residual mild thrombocytopenia. The basis for the intact bone marrow function despite the in vitro ICL sensitivity is unclear, but in our analysis, FANCM biallelic mutant cells showed an intermediate FANCD2 monoubiquitination defect relative to FANCA mutant cells, suggesting these mutations are hypomorphic or possibly other compensatory mechanisms play a role in signaling ICL repair.¹³ All the heretofore symptomatic FANCM biallelic mutations described are truncating, and except for the minimally truncated homozygous p.Arg1931*, all have been found to show in vitro ICL agent sensitivity.⁸ This observation suggests a potential genotype-phenotype correlation with early-onset disease potentially caused by more severe FANCM protein truncations.

The treatment of HNSCC in patients with classic Fanconi anemia is a clinical challenge, given their exquisite platinum sensitivity and often aggressive nature of the disease.⁵ We analyzed the somatic landscape of two cancers from the same patient and found no difference in mutation burden relative to sporadic disease. Nonetheless, despite the PDL-1-negative status of both the tumor and immune cells in the tumors, there was a brisk T-cell infiltrate. Ongoing efforts continue to investigate the role of immunotherapy in Fanconi anemia–related HNSCC. The long-term disease-free follow-up in the patient we report postimmunotherapy suggests that it will be important to assess this approach. In summary, our report identifies a distinct but rare clinical presentation of adult-onset genetic susceptibility to HNSCC that may be identifiable clinically by the presence of premature ovarian failure and/or history of infertility.

APPENDIX. MATERIALS AND METHODS

Subjects and Consent

The study was approved by the Johns Hopkins Medicine Institutional Review Board, and participants provided written informed consent to participate in the study and for publication of the data.

Chromosome and Genetic Studies

Telomere length was measured in the Johns Hopkins Pathology Laboratories by flow cytometry and fluorescence in situ hybridization as previously described.¹⁴ Chromosomal breakage analysis was performed in the Johns Hopkins Cytogenetics Laboratory on peripheral blood using diepoxybutane (DEB) (0.1 µg/ml) using standard protocols. Variant allele frequencies were assessed in gnomAD v.2 in December 2020.

Cell Lines

Lymphoblastoid cell lines (LCLs) were derived as described.¹⁵ FANCA-mutant cells were obtained from Corriell Institute (Camden): GM13022 is an Ebstein Barr Virus–immortalized LCL from a classic Fanconi anemia patient where FANCA deficiency was established by complementation.¹⁶

FANCD2 Monoubiquitination

Cells were treated with mitomycin C (MMC) at 500 ng/mL for 24 hours. Protein extraction and immunoblot were performed as previously described¹⁷ using rabbit monoclonal anti-FANCD2 antibody (RRID: ab108928; 1:1,000, Abcam) and antitubulin (rabbit polyclonal, RRID: ab6046, 1:4,000, Abcam). Blots were visualized using secondary fluorescent conjugated antibodies and an Odyssey scanner.¹⁸

Mitomycin Sensitivity Assay

MMC (Sigma) sensitivity was performed on LCLs by plating 1 × 10⁴ to 1 × 10⁵ cells in 96-well plates in 200 µL of complete culture medium. Cells were exposed to MMC (0-100 ng/mL). After a 3-day incubation, the Cell Counting Kit-8 (CCK-8, Dojindo Molecular Technologies) was used to quantify cell number. For each condition, cells were plated in triplicate.

Immunohistochemistry

Immunohistochemistry was performed on 5-µm paraffin-embedded sections. Immunostaining was performed at the Johns Hopkins Oncology Tissue Services using standard methods. Immunostaining was performed using anti-CD3 (polyclonal, 1:100; Dako, Carpinteria, CA), anti-CD8 (clone C8144B; prediluted CellMarque/Sigma-Aldrich, St Louis, MO), and anti-PD-L1 (clone 22C3, prediluted, Dako). Signals were visualized on a Ventana Benchmark XT autostainer in the presence of appropriate controls for each stain. Slides were scored manually by a pathologist (L.R), and membranous PDL-1 staining was considered positive as per standard criteria.¹⁹

Whole-Exome Sequencing

Whole-exome sequencing was performed at Personal Genome Diagnostics (Baltimore, MD) using CancerXOME-R, a proprietary Illumina-based platform. Tumor DNA was obtained from microdissected formalin-fixed paraffin-embedded specimens, and tumor purity was estimated at 80% and 40% for the melanoma and oral cavity SCC, respectively. Genomic DNA was blood-derived. The coding regions of > 20,000 genes were sequenced and the average depth of coverage was 157× and 183× for the melanoma and oral cavity squamous cell cancer, respectively. Target regions had 10 or more reads in at least 89% of bases. Variants were annotated to reference genome hg19.

AUTHOR CONTRIBUTIONS

Conception and design: Paz J. Vellanki, Hyunseok Kang, Mary Armanios

Provision of study materials or patients: Lynn Schuchter, Ranee Mehra, Hyunseok Kang

Collection and assembly of data: Paz J. Vellanki, Emily A. DeBoy, M. Malek Bawadkji, Lynn Schuchter, Hyunseok Kang, Mary Armanios

Data analysis and interpretation: Paz J. Vellanki, Emily A. DeBoy, Lisa Rooper, Ranee Mehra, Hyunseok Kang, Mary Armanios

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/po/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).