Germ-line mutations in WDR77 predispose to familial papillary thyroid cancer

Significance

Familial nonmedullary thyroid cancer (FNMTC) accounts for up to 15% of all thyroid tumors. Here, we report two germ-line loss-of-function variants in WDR77 in two unrelated families with NMTC. WDR77 is expressed in diverse human tissues. It forms a complex with the protein arginine methyltransferase PRMT5 and mediates H4R3me2 in frogs and mammals. Functional analyses show that WDR77 variants impair formation of the complex of WDR77–PRMT5, resulting in reduced H4R3me2 in patients. Knockdown of WDR77 results in increased growth of thyroid cancer cells. Taken together, our findings reveal a predisposing gene of FNMTC and expand the variant profile predisposing to FNMTC. Additionally, this report presents germ-line variants in WDR77 associated with human disease.

Abstract

The inheritance of predisposition to nonsyndromic familial nonmedullary thyroid cancer (FNMTC) remains unclear. Here, we report six individuals with papillary thyroid cancer (PTC) in two unrelated nonsyndromic FNMTC families. Whole-exome sequencing revealed two germ-line loss-of-function variants occurring within a 28-bp fragment of WDR77, which encodes a core member of a transmethylase complex formed with the protein arginine methyltransferase PRMT5 that is responsible for histone H4 arginine 3 dimethylation (H4R3me2) in frogs and mammals. To date, the association of WDR77 with susceptibility to cancer in humans is unknown. A very rare heterozygous missense mutation (R198H) in WDR77 exon 6 was identified in one family of three affected siblings. A heterozygous splice-site mutation (c.619+1G > C) at the 5′ end of intron 6 is present in three affected members from another family. The R198H variant impairs the interaction of WDR77 with PRMT5, and the splice-site mutation causes exon 6 skipping and results in a marked decrease in mutant messenger RNA, accompanied by obviously reduced H4R3me2 levels in mutation carriers. Knockdown of WDR77 results in increased growth of thyroid cancer cells. Whole-transcriptome analysis of WDR77 mutant patient-derived thyroid tissue showed changes in pathways enriched in the processes of cell cycle promotion and apoptosis inhibition. In summary, we report WDR77 mutations predisposing patients to nonsyndromic familial PTC and link germ-line WDR77 variants to human malignant disease.

The incidence of thyroid cancer has increased rapidly (1, 2). Approximately 95% of all cases of thyroid cancer are nonmedullary thyroid cancer (NMTC) of follicular cell origin. Several large population-based studies have shown a high familial risk of NMTC, which strongly suggests that hereditary factors are involved in the disease (3–7). However, the genes involved in predisposition to NMTC are poorly understood. Familial NMTC (FNMTC) accounts for up to 15% of all cases of thyroid tumors and includes syndromic and nonsyndromic tumors (8, 9). The genetic events in syndromic FNMTC are known (8, 9), while the genetic basis of nonsyndromic FNMTC remains unclear, although several chromosomal loci (1p13.2-1q22, 2q21, 14q32, 19p13.2, and 8p23.1-p22) (10–14) and low-penetrance mutations in SRGAP1, NKX2.1, FOXE1, HABP2, and CHEK2 have been identified to confer susceptibility to nonsyndromic FNMTC (8, 13, 15–17).

WDR77 (also known as MEP50) is located on human chromosome 1p13.2. WDR77 is a core component of a transmethylase complex with the protein arginine methyltransferase PRMT5. WDR77 contains seven WD40-repeat (WDR) domains, the majority of which bind to the surface of PRMT5 and form a heterooctameric PRMT5 4:WDR77 4 complex (18). It recruits the substrate to PRMT5 and activates PRMT5 to methylate a specific arginine to dimethylarginines in target proteins such as histone H4 (18, 19).

Here, we reported six affected individuals with thyroid cancer from two unrelated Chinese families (F3 and F4) (Fig. 1A and Materials and Methods). Tumor tissue samples from four affected family members were analyzed with hematoxylin–eosin staining, which provided pathological confirmation of papillary thyroid cancer (PTC) (Fig. 1B). We subsequently performed whole-exome sequencing (WES) in the proband and in patients from the two families in order to identify new predisposing genes of thyroid cancer and expand our current understanding of the genetic basis of the disease.

Fig. 1.

Identification of WDR77 mutations in two affected families. (A) Pedigree of two affected families (family 3 and family 4) and DNA sequence electropherograms showing mutations with the disease phenotype. Arrows indicate probands. No clinical information regarding the thyroid of F4_II.2 is available. Thyroid cancer in the affected family members with the splice-site mutation in family 4 occurred earlier (mean diagnosis age of 39 y) than in those with the R198H variant in family 3 (mean diagnosis age of 50.3 y), although this difference did not reach statistical significance; P = 0.07 by Student’s t test. (B) Hematoxylin and eosin staining performed on tumor sections from F3_Patient III.2, F3_Patient III.7, F4_Patient III.2, and F4_Patient III.3. (Scale bars, 100 μm.)

Results

In this study, WES data were generated using peripheral-blood DNA from the affected individuals of two families with PTC (SI Appendix, Supplementary Materials and Methods). Utilizing filtering criteria (SI Appendix, Table S1), we identified two heterozygous variants, a missense variant, and a splice-site mutation located in exon 6 and at the boundary between exon 6 and intron 6 of WDR77, respectively, both of which occur within a DNA fragment of 28 bp (Fig. 2A). The two variants were then validated in the corresponding affected family members by Sanger sequencing, indicating an autosomal dominant inheritance pattern of PTC associated with mutations in WDR77 (Fig. 1A and SI Appendix, Table S2).

Fig. 2.

The effect of WDR77 mutations on histone H4R3 methylation. (A) Molecular genetic findings in patients with WDR77 variants from these two families. Schematic of the exon–intron structure of WDR77 showing the positions of one missense (c.593G > A, p.R198H) and one splice-site variant (c.619+1G > C). Conservation of the WD40 repeat (green rectangle indicated by a red arrow) and R198 across species are illustrated. RBBP4, RB Binding Protein 4. RBBP7, RB Binding Protein 7. GRWD1, Glutamate Rich WD Repeat Containing 1. (B, Upper Panel) The model of the complex of WDR77 (orange) and PRMT5 (cyan) (Protein Data Bank, 4gqb). (Lower Panel) Substitution of arginine to histidine at position 198 is predicted to cause loss of the hydrogen bond network. The hydrogen bond between Asp196 and Arg198 is highlighted. (C) The R198H mutation disrupts the interaction between WDR77 and PRMT5. 293T cells were transiently transfected with vectors expressing FLAG-tagged wild-type (lane 2) or R198H (lane 3) WDR77 protein alone or with HA-PRMT5 (lane 4 through lane 9). Immunoprecipitation was performed using an HA antibody, and immunoblotting was performed with anti-FLAG or anti-HA antibodies. (D) Immunoblotting shows the expression of histone H4 methylation in six samples taken from each F3 family member–derived peripheral WBCs (two unaffected family members without R198H mutation, F3_IV.2 and F3_IV.3; one unaffected member with R198H mutation, F3_IV.1; three affected members, F3_Patient III.2, III.3, and III.7). Total histone H4 served as loading control. Quantification of immunoblotting of H4R3me2 in F3 family members was performed. H4R3me2 intensity values were normalized against the levels of total histone H4. Experiments were repeated three times with similar results (see SI Appendix, Fig. S7A for details). (E) Immunoblotting shows the expression of histone H4 methylation in the normal thyroid tissues from the affected family member (F3_Patient III.2) and sporadic PTC patients without the mutation (BJH355 and BJH365). Total histone H4 served as loading control. Quantification of immunoblotting of H4R3me2 in the index patients was performed. H4R3me2 intensity values were normalized against the levels of total histone H4. Experiments were repeated three times with similar results (see SI Appendix, Fig. S7C for details). (F) PCR analysis shows the expression of WDR77 mRNAs with the c.619+1G > C variant and wild-type WDR77 mRNAs in F4 family member–derived peripheral WBCs (F4_Patient III.3 and an unaffected family member, F4_II.1). Control, WDR77 mRNA expression levels in peripheral WBCs obtained from a sporadic case without the mutation. The EIF4H mRNA bands were indicated as a loading control. Marker, DNA marker. Sequence chromatograms of Sanger sequencing of WDR77 mRNAs from F4_Patient III.3. (G and H) Immunoblotting shows the expression of WDR77 protein (G) and H4R3me2 (H) in F4 family member–derived peripheral WBCs (F4_Patients III.3 and F4_II.1). GAPDH and total histone H4 served as loading controls. Experiments were repeated three times with similar results (see SI Appendix, Fig. S7 D and E for details).

The missense variant that was identified in family 3 results in a G-to-A substitution in WDR77 (Chr.1:111,986,515, rs768526334, c.593G > A), which was present in the proband (F3_Patient III.7) and two other affected family members (F3_Patients III.2 and III.3) but was not in an unaffected sibling (F3_III.5) (Fig. 1A). The variant was also detected in the proband’s 78-y-old mother (F3_II.5), who was found to have small thyroid nodules (4 to 5 mm) on imaging. Additionally, the variant was absent in 2,827 high-coverage exome sequences in a cancer-free in-house Chinese cohort (SI Appendix, Supplementary Materials and Methods), absent in 62,784 high-coverage genome sequences in TOPMed (Trans-Omics for Precision Medicine) (20) and in 141,431 low-coverage Chinese genome sequences (Chinese Millionome Database, CMDB) (21), and was only detected in 2 out of 125,722 persons in gnomAD (Genome Aggregation Database) version 2.1.1 (22). The G-to-A substitution converted an arginine to a histidine at position 198 of WDR77 (NM_024102, c.593G > A, p.R198H). The R198 residue is located in the fourth WDR domain and is well conserved in WDR77. Moreover, high conservation of the residue was also observed in RBBP4 and RBBP7, two of which are chromatin remodeling factors involved in chromatin assembly and cell proliferation by directly binding to the RB1 protein (23, 24) (Fig. 2A). The R198H variant was consistently predicted to be a disruptive mutation by SIFT (25), Polyphen2 (26), CADD (27), MutationAssessor (28) and fathmm (29) (Table 1).

Table 1.

Genetic and clinical characteristics of the individuals with WDR77 mutations identified in this study

Patient number	F3_Patient III.2	F3_Patient III.3	F3_Patient III.7	F4_Patient III.1	F4_Patient III.2*	F4_Patient III.3
Sex	Female	Male	Female	Male	Female	Female
Base change	c.593G > A	c.593G > A	c.593G > A	c.619+1G > C	c.619+1G > C	c.619+1G > C
Amino acid change	p.R198H	p.R198H	p.R198H	—	—	—
Position	Exon 6	Exon 6	Exon 6	Donor splice site of intron6	Donor splice site of intron6	Donor splice site of intron6
MAF in Novo in-house cohort (n = 2,827)	Absence	Absence	Absence	Absence	Absence	Absence
MAF in CMDB (n = 141,431)	Absence	Absence	Absence	Absence	Absence	Absence
MAF in TOPMed (n = 62,789)	Absence	Absence	Absence	Absence	Absence	Absence
MAF in gnomAD (n =125,722)	0. 0008%	0. 0008%	0. 0008%	Absence	Absence	Absence
SIFT	Deleterious	Deleterious	Deleterious	—	—	—
Polyphen2	Probably damaging	Probably damaging	Probably damaging	—	—	—
MutationAssessor	Deleterious	Deleterious	Deleterious	Deleterious	Deleterious	Deleterious
CADD	35	35	35	26.1	26.1	26.1
fathmm	Pathogenic (0.979)	Pathogenic (0.979)	Pathogenic (0.979)	Pathogenic (0.983)	Pathogenic (0.983)	Pathogenic (0.983)
GERP++	5.94	5.94	5.94	5.94	5.94	5.94
Age at diagnosis (year)	57	49	45	33	43	41
Histologic type of cancer	PTC	PTC	PTC	PTC	PTC	PTC
Stage of cancer (TNM)†	T1aN0M0	T1bN1aM0	T1aN1aM0	NA	T1aN0M0	T1bN0M0
Radioiodine ablation	No	No	No	No	No	No
Vital status	Alive	Alive	Alive	Alive	Alive	Alive
Germ-line mutations tested for PTEN‡, APC‡, DICER1‡, WRN‡, PRKAR1A‡, STK11‡, SPGAP1§, NKX2.1§, FOXE1§, HABP2§, CHEK2§	WT	WT	WT	WT	WT	WT
Somatic mutations tested for WDR77 mutations	c.6G > A, p.R2 =	NA	c.129C > T, p.L43= ; c.427G > A, p.G143D	NA	Negative	Negative
WDR77 CNAs	No change	NA	No change	NA	No change	No change
BRAF V600E	Negative	NA	Negative	NA	Positive	Positive
KRAS, NRAS, HRAS, TERT c. 124C > T and c. 146C > T	Negative	NA	Negative	NA	Negative	Negative

NA, not available; MAF, mutant allele frequency.

^*F4_Patient III.2 was also diagnosed with invasive ductal carcinomas of the breast.

^† Staging was evaluated on the tumor-node-metastasis (TNM) classification of the American Joint Committee on Cancer and the Union for International Cancer Control, Eighth Edition.

^‡ Known syndromic FNMTC genes.

^§ Known predisposing genes to nonsyndromic FNMTC.

To explore whether the WDR77 R198H mutation was likely to interfere with protein function, we first examined the expression of the mutated WDR77 transcript in the normal thyroid tissue and the matched tumor from the mutation carrier (F3_Patient III.2) by RT-PCR. Sequencing of the resulting RT-PCR products showed comparable expression abundance between wild-type and mutant alleles in thyroid tissues from the affected family member (SI Appendix, Fig. S1A). Similarly, no notable change in WDR77 protein level was detected between the tumor and the adjacent normal tissue in the mutation carriers (SI Appendix, Fig. S1 B and C). We also observed similar levels of WDR77 proteins between the mutation carriers and mutation noncarriers (SI Appendix, Fig. S2A). Since the R198 residue is very close to the nuclear localization sequence (144 to 197 residues) of WDR77 (30), we subsequently tested the localization of WDR77 in HEK293T cells overexpressing the wild-type or the R198H variant. Little difference in the localization of WDR77 was observed between wild-type– and WDR77–mutant HEK293T cells (SI Appendix, Fig. S1D).

When we modeled the R198H variant on the structure of the WDR77 protein, we found that the substitution to histidine at the R198 residue disrupted the formation of three backbone hydrogen bonds of R198 with D196, which was predicted to affect the stability of the β-turn that encompasses these two residues (Fig. 2B). This suggested destabilizing impact on the WDR domain likely interfered with the binding of WDR77 to PRMT5. Next, we established HEK293T cells coexpressing HA-PRMT5 with Flag-tagged wild-type WDR77 or with the WDR77 R198H variant. We found that the interaction between the WDR77 R198H mutant and PRMT5 was markedly decreased compared to that of wild-type WDR77 and PRMT5, as determined by coimmunoprecipitation assays (Fig. 2C).

Furthermore, peripheral white blood cells (WBCs) that were derived from four family members harboring R198H showed less dimethylation at arginine 3 of histone H4 (H4R3me2) than WBCs from two family members with wild-type WDR77 (Fig. 2D), while no difference in WDR77 protein abundance was detected in any of these six individuals (SI Appendix, Fig. S2A). Similarly, reduced histone H4 methylation was also seen in the normal thyroid tissue from the mutation carrier (F3_Patient III.2) compared to those of mutation noncarriers (Fig. 2E and SI Appendix, Fig. S2B). Together, these results indicated that the R198H mutation impaired the binding of WDR77 with PRMT5, leading to a decrease in histone H4 methylation.

On the other hand, the splice-site variant in WDR77 identified in family 4 (Chr.1:111,986,488, NM_024102, c.619+1G > C) was present in the proband (F4_Patient III.2) and her two affected siblings (F4_Patients III.1 and III.3) but not in her unaffected father (F4_ II.1) (Fig. 1A). In addition, the variant was absent from all public DNA sequence databases including TOPMed, gnomAD, and CMDB. The proband’s mother (F4_II.2) died from kidney failure at the age of 58. She had noncalcified pulmonary nodules, but no medical information related to her thyroid was available.

This heterozygous c.619+1G > C mutation destroys a canonical splice donor site in intron 6. The effect of the splice-site mutation was determined using RT-PCR followed by Sanger sequencing of WBCs derived from the affected family member (F4_Patient III.3) and her unaffected father (F4_II.1) (Fig. 2F). The mutation caused exon 6 skipping, consequently leading to a frameshift starting with an alanine at amino acid 189 and creating a premature stop codon at position 43 of the new reading frame (SI Appendix, Fig. S3). The shorter transcript from the mutant allele was observed at very low levels (Fig. 2F), suggesting that the mutation triggers the nonsense-mediated messenger RNA (mRNA) decay (NMD) process. Consistently, WDR77 protein levels were evidently decreased in the mutation carrier (F4_Patient III.3) compared to the mutation noncarrier (F4_ II.1) (Fig. 2G). Similar to the functional consequence of the R198H variant in WDR77, the c.619+1G > C mutation led to a lower level of histone H4 methylation in the mutation carrier (F4_Patient III.3) than in the mutation noncarrier (F4_II.1) (Fig. 2H). Together, these data indicated that both the missense variant and the splice-site mutation led to the loss of function of WDR77 in the affected family members.

To further evaluate the putative biological effect of loss of function of WDR77, we conducted the knockdown of WDR77 in two human thyroid cancer cell lines (TPC-1 and FTC-133) and a human embryonic kidney cell line (HEK293 cells) using two different small interfering RNAs (siRNAs)—siWDR77#1 and siWDR77#2 (31). The three cell lines carry no WDR77 protein-coding or splice-site mutation. We observed that each of the two siRNAs was able to effectively inhibit WDR77 protein expression in these cell lines (Fig. 3A and SI Appendix, Fig. S4A). Cell proliferation assays showed that siRNA-mediated WDR77 knockdown significantly increased the growth of TPC-1, FTC-133, and HEK293 cells (Fig. 3A and SI Appendix, Fig. S4A), suggesting that a decrease in WDR77 expression resulted in an increase in the growth of thyroid cancer cells.

Fig. 3.

Biological effect of loss of function of WDR77. (A) CellTiter-Glo assay was used to determine the growth of TPC-1, FTC-133, or HEK293 cells treated with the control siRNA and siWDR77 (siWDR77#1, siWDR77#2), respectively, over a four-point time course. The data of representative experiments resulted from two independent experiments with four replicates per group. (Inset) Protein blot showing knockdown of WDR77 protein in the corresponding cell lines. Quantification relative to the levels in cells treated with the control siRNA of the total intensity of the WDR77 band relative to the alpha-tubulin band is shown above. Experiments were repeated with similar results (see SI Appendix, Fig. S4A for details). (B) Differentially expressed genes associated with the processes of the cell cycle and apoptosis by RNA sequencing of normal thyroid tissues from the affected family member (F3_Patient III.2) and sporadic PTC patients without the mutation (BJH324, BJH355, and BJH372). (C) Summary model of the biological effects of WDR77 mutations in this study. (Upper Panel) The formation of the PRMT5:WDR77 complex is a key step for H4R3me2 modification. (Middle and Lower Panels) WDR77 mutations diminish binding with PRMT5 or reduce the amount of the complex, resulting in a decrease in H4R3me2 levels, which might eventually regulate the expression of cancer hallmark–associated genes involved in promoting the cell cycle and inhibiting apoptosis. **P < 0.01 and ***P < 0.001 by two-tailed Student’s t test, mean ± SD; n = 4.

As both WDR77 R198H and the inactivated splice-site mutation altered histone modification status, we conducted RNA sequencing on normal thyroid tissues from an affected individual (F3_Patient III.2) and three sporadic PTC patients without the mutations (SI Appendix, Fig. S4B) to analyze the associated gene expression changes. We observed that 147 genes showed more than twofold up-regulation and that 116 genes had more than twofold down-regulation in the mutation carrier compared to those of three mutation noncarriers (SI Appendix, Fig. S4C). Analysis of cancer hallmark-associated genes in the gene expression profiles showed an increase in expression of genes for promoting cell cycle and suppressing cell apoptosis and a decrease in expression of genes for suppressing cell cycle and promoting cell apoptosis in the mutation carrier compared with controls (Fig. 3B and SI Appendix, Table S3).

Moreover, to examine the somatic mutation status of WDR77 in thyroid tumors from the affected family members, we sequenced the whole coding regions and splicing sites of WDR77 using genomic DNAs available from the formalin-fixed paraffin-embedded tumor specimens of four affected individuals (F3_Patient III.2, F3_Patient III.7, F4_Patient III.2, and F4_Patient III.3) (SI Appendix, Fig. S5). Three somatic single-nucleotide variants (two silent mutations and one missense mutation) were found in F3_Patient III.2 and F3_Patient III.7 (SI Appendix, Fig. S5B and Table 1). Furthermore, copy number analysis in the tumor samples showed no copy number alterations of WDR77 compared with the matched blood DNA (Table 1). We also found no known somatic mutations in KRAS, NRAS, HRAS, or TRET, except BRAF in the tumor samples (Table 1).

Discussion

In summary, we reported that individuals with nonsyndromic FPTC harbored damaging germ-line mutations in WDR77, which resulted in impairment of the association of WDR77 with PRMT5. WDR77 in complex with PRMT5 is required for activation of PRMT5 to modify histone H4 (18, 19, 32). The disruption of the interaction between WDR77 and PRMT5 suppressed the methylation of histone H4 (33). Accordingly, we observed a reduction in the methylation of histone H4 in our affected family members. In engineered mouse models, mice fully lacking wdr77 die as early-stage embryos. Wdr77^+/− heterozygous mice exhibited excessive epithelial cell proliferation that resulted in prostatic intraepithelial neoplasia and increased testis size (34, 35), suggesting that lacking one copy of the WDR77 allele was sufficient to cause a propensity of proliferation of normal cells and that haploinsufficiency may be a mechanistic explanation for the lesion due to the low WDR77 protein level and, consequently, decreased methyltransferase activity of PRMT5. Likewise, a reduction of the activation of PRMT5 in human hematopoietic progenitor cells by using short hairpin RNA (shRNA) led to an increase in colony formation of the cells (36). Nevertheless, WDR77 plays roles in cell proliferation or differentiation depending on the cell context (31, 34–38). Our functional experiments showed that the reduced WDR77 expression promoted human thyroid cancer cell proliferation. Our expression profile data of the normal thyroid tissues from mutation carriers and noncarriers indicated that differentially expressed genes were enriched in the processes of cell cycle promotion and apoptosis suppression, providing insight into the possible mechanism for the variants of WDR77 predisposing to PTC (Fig. 3C). Moreover, we cannot rule out the possibility that the identified mutations could alter some currently unknown functions of WDR77 that may contribute to predisposition.

Moreover, we analyzed two large population cohorts: the gnomAD (version 2.1.1 and version 3.1.1) datasets, which contain 125,748 exomes and 76,156 whole genomes from unrelated noncancer individuals and 132,345 deeply sequenced unrelated genomes from the TOPMed (freeze 8) dataset. We found 30 high-confidence predicted loss-of-function variants of WDR77 in 55 gnomAD individuals and 41 TOPMed individual carriers, with a combined carrier frequency of 0.03% (SI Appendix, Fig. S6 and Dataset S1). Furthermore, we found 17 tumor samples with 14 somatic truncating variants, including 1 frameshift, 4 splicing, and 9 stop-gain mutations, in the Catalogue of Somatic Mutations in Cancer (39) and cBioPortal (40) databases (SI Appendix, Fig. S6B and Dataset S1). Notably, the search revealed a 29-y-old female thyroid cancer patient with a stop-gain mutation in WDR77 (THYROID-CN-WZ040T, WDR p.C208*, c.624C > A), which could result in the loss of function of WDR77, similar to the consequence of the c.619+1G > C mutation identified in our study. Additionally, one WDR77 somatic missense mutation combined with a deletion mutation in the gene was found in each of seven tumors (SI Appendix, Fig. S6B and Dataset S1). All these data further supported the association of heterozygous loss-of-function variants in WDR77 with human disease.

Taken together, our findings identified two loss-of-function variants in WDR77 predisposed to FNMPTC and established a link between germ-line WDR77 mutations and human malignant disease.

Materials and Methods

Study Patients and Oversight.

This study was approved by the Research Ethics Board of the Beijing Hospital, Ministry of Health. All participants provided written informed consent before participation. Family members in two families (F3_Patient III.2, F3_Patient III.3, F3_Patient III.7, F4_Patient III.2, and F4_Patient III.3) were admitted to our hospital during 2018 to 2020 due to thyroid neoplasms detected with ultrasound-guided fine-needle aspiration biopsy. The proband in family 3 (F3_Patient III.7), a 48-y-old woman, was the youngest of four children. Her mother (F3_II.5) is 78 y old and was found to have small thyroid nodules (4 to 5 mm) upon neck ultrasonography. F3_IV.1, a 26-y-old man, had undergone neck ultrasonography, and no thyroid nodule was found when undergoing evaluation in our study. The proband in family 4 (F4_Patient III.2), a 44-y-old woman, was diagnosed with breast cancer at age 36 and thyroid cancer at age 43. Her mother (F4_II.2) died from kidney failure at the age of 58. She had noncalcified pulmonary nodules, but no medical information related to her thyroid was available. All five patients (F3_Patient III.2, F3_Patient III.3, F3_Patient III.7, F4_Patient III.2, and F4_Patient III.3) were diagnosed with papillary thyroid carcinoma and received a thyroidectomy. None of the family members, except one member (F4_Patient III.2), had a history of other primary cancers. No other benign tumors were detected in the two families.

Genetic Testing and Function Analyses.

We performed WES on Patient III.2 and Patient III.7 from family 3 and the Patient III.1 and Patient III.3 from family 4. Function analyses of the variants were performed. The detailed methods are available in SI Appendix.

Data Availability

All study data are included in the article and/or supporting information.