OUP user menu

Absence of the BRAF Mutation in HBME1+ and CK19+ Atypical Cell Clusters in Hashimoto Thyroiditis
Supportive Evidence Against Preneoplastic Change

Michel R. Nasr MD, Sanjay Mukhopadhyay MD, Shengle Zhang MD, Anna-Luise A. Katzenstein MD
DOI: http://dx.doi.org/10.1309/AJCPCGCZZ1OYF0IC 906-912 First published online: 1 December 2009

Abstract

An association between Hashimoto thyroiditis and papillary thyroid carcinoma has been postulated for decades. We undertook this study to identify potential precursors of papillary thyroid carcinoma in Hashimoto thyroiditis using a combination of morphologic, immunohistochemical, and molecular techniques. For the study, samples from 59 cases of Hashimoto thyroiditis were stained with antibodies to HBME1 and cytokeratin (CK)19. Tiny HBME1+ and CK19+ atypical cell clusters were identified and analyzed for the BRAF mutation by the colorimetric Mutector assay and allele-specific polymerase chain reaction. HBME1+ and CK19+ atypical cell clusters were identified in 12 (20%) of 59 cases. The minute size (<1 mm) of the clusters and the incomplete nuclear changes precluded a diagnosis of papillary microcarcinoma. The atypical cell clusters from all 12 cases were negative for BRAF. The absence of the BRAF mutation in these atypical cell clusters suggests that they may not be preneoplastic. Caution should be exercised in interpreting positive HBME1 or CK19 staining in Hashimoto thyroiditis.

Key Words:
  • Hashimoto thyroiditis
  • HBME1
  • Cytokeratin 19
  • BRAF mutation

An association between Hashimoto thyroiditis and papillary thyroid carcinoma has been postulated for decades. A meta-analysis of 2,552 thyroidectomies showed a 2.7 times increased rate of papillary thyroid carcinoma in Hashimoto thyroiditis compared with the rate in control subjects and a 1.9 times increased rate of Hashimoto thyroiditis in papillary thyroid carcinoma compared with other thyroid cancers.1 Several studies have attempted to identify premalignant lesions in Hashimoto thyroiditis. Overlapping morphologic and immunohistochemical features with papillary thyroid carcinoma have been noted in small clusters of cells in Hashimoto thyroiditis.25 Molecular studies have added additional evidence for precursor lesions in that RET/PTC aberrations have been noted in greater than 90% of thyroids with Hashimoto thyroiditis,6,7 although this finding was not confirmed by others.8 Hunt et al9 observed loss-of-heterozygosity mutations of tumor suppressor genes in small areas with atypical nuclei in chronic lymphocytic thyroiditis, suggesting that these atypical areas are neoplastic rather than reactive.

HBME1 and cytokeratin (CK)19 have been shown to be useful immunohistochemical markers for the diagnosis of papillary thyroid carcinoma,10,11 and, therefore, we postulated that these markers might help identify putative precursor lesions that do not show full-fledged morphologic features of papillary thyroid carcinoma. Another potential marker of thyroid neoplasia is the BRAF mutation, which is fairly sensitive and highly specific for papillary thyroid carcinoma. A review of the literature showed BRAF mutations in 44% (range, 29%–83%) of 1,856 reported cases of papillary thyroid carcinoma but no BRAF mutations in 165 follicular carcinomas, 65 medullary carcinomas, or 542 benign thyroid neoplasms.12 The high specificity of BRAF mutations for papillary thyroid carcinoma makes it an ideal molecular marker for the detection of neoplastic and/or preneoplastic lesions in patients with Hashimoto thyroiditis.

We undertook this study to identify potential precursors of papillary carcinoma in Hashimoto thyroiditis by using a combination of morphologic features, immunohistochemical staining for HBME1 and CK19, and molecular testing for the BRAF mutation.

Materials and Methods

For the study, 59 cases of Hashimoto thyroiditis were retrieved from the archives of the Department of Pathology, State University of New York Upstate Medical University, Syracuse. Hashimoto thyroiditis was defined by the presence of diffuse lymphocytic thyroiditis with germinal center formation and Hürthle cell change of the follicular epithelium.

All cases were stained with antibodies to HBME1 (M3505, dilution 1:50; DAKO, Carpinteria, CA) and CK19 (Ks19.1, dilution 1:50; NeoMarkers, Fremont, CA). Immunohistochemical analysis was performed on formalinfixed, paraffin-embedded (4-μm thick) sections using a standard technique (streptavidin-biotin-peroxidase technique) with appropriate positive and negative control samples. For both antibodies, immunoreactivity was considered positive if more than 10% of follicular epithelial cells stained. Membrane staining along lateral and abluminal surfaces (basolateral staining) was considered positive for HBME1, while cytoplasmic staining was considered positive for CK19.

H&E-stained slides on all cases were reviewed to identify tiny clusters of atypical epithelial cells showing incomplete nuclear features of papillary thyroid carcinoma within areas of Hashimoto thyroiditis. These atypical cell clusters were included for study if they showed positivity for HBME1 and CK19. The clusters were then microdissected and analyzed for the BRAF mutation using 2 assays: the colorimetric Mutector Assay (TrimGen, Sparks, MD) and allele-specific polymerase chain reaction (PCR). Testing for the BRAF mutation was also performed on 5 cases each of papillary thyroid carcinoma and papillary microcarcinoma (positive control cases) and 4 cases of Hashimoto thyroiditis without atypical epithelium (negative control cases). The 5 cases of papillary thyroid carcinoma in Hashimoto thyroiditis tested for the BRAF mutation only (positive control cases) were selected from our previous study cases that demonstrated positivity for CK19 and HBME1.10

All authors reviewed the H&E-stained sections and interpreted the immunohistochemical staining results. Disagreements in the histopathologic diagnosis or differences in interpreting the immunostaining results were resolved by consensus.

Genomic DNA Extraction

Atypical cell clusters were microdissected from HBME1-stained slides and processed for DNA extraction and PCR. Microdissection was performed with a disposable scalpel blade under microscopic observation. Total DNA was extracted using the Wax Free Paraffin DNA extraction kit (TrimGen). All samples were extracted per the manufacturer’s instructions.

Allele-Specific PCR

DNA extracted from HBME1-stained slides was used for PCR amplification with allele-specific BRAF primer sets Table 1 as previously described by Jin et al.13 These primer sets included a forward primer to cover the wild-type allele and another forward primer to cover the mutant allele. The same reverse primer was used with both forward primers. PCR was performed in a 50-μL reaction mixture using 1 μL of extracted DNA, 25 μL of HotStar Taq Master Mix (Qiagen, Valencia, CA), and 0.6 μmol/L primers (Integrated DNA Technologies, Coralville, IA), the remaining volume being made up by Molecular Biology Grade Water (Sigma-Aldrich, St Louis, MO). The reaction mixture was subjected to an initial activation step at 95°C for 15 minutes, initial denaturation at 95°C for 1 minute, 40 cycles of denaturation at 95°C for 5 seconds, annealing at 66°C for 5 seconds, extension at 72°C for 6 seconds, and final extension at 72°C for 10 minutes on the GeneAMP PCR system 9700 (PE Applied Biosystems, Foster City, CA). Water was used as a negative control sample. These PCR products were then electrophoresed on Ready Gels of 10% TBE (Tris, boric acid, and EDTA) for polyacrylamide electrophoresis (Bio-Rad, Hercules, CA), stained with ethidium bromide, and visualized and photographed under UV light. Bands for mutated and wild-type alleles were expected to be 149 base pairs (bp).

View this table:
Table 1

Colorimetric Mutector Assay

The colorimetric Mutector assay was performed on DNA extracted from HBME1-stained slides. BRAF mutation analysis was performed using the Mutector Dual Well Test Kit (TrimGen). The assay was performed per manufacturer’s instructions. Extracted DNA was first amplified using the kit’s primer mix. PCR was performed in a 50-μL reaction mixture containing 1 μL of extracted DNA, 25 μL of HotStar Taq Master Mix, 6 μL of primer mix and Molecular Biology Grade Water. The reaction mixture was subjected to an initial activation step at 95°C for 1 minute, 40 cycles of denaturation at 94°C for 30 seconds, annealing at 56°C for 1 minute, extension at 72°C for 1 minute, and final extension at 72°C for 10 minutes on the GeneAMP PCR system 9700. The PCR products were then subjected to hybridization with specific detection primers, primer extension, and 2 different enzymatic reactions testing for both wild-type and mutant alleles. After color development, samples were placed on a Vmax microplate reader at a wavelength of 405 nm (Molecular Devices, Sunnyvale, CA) to measure the color intensity. The evaluation of Mutector assay results was performed according to the manufacturer’s instructions: cases in which the ratio of intensity of mutant type to wild-type was 0.3 or more were interpreted as having the mutation genotype, as described previously.13

Results

Histologic and Immunohistochemical Findings

Of 59 cases of Hashimoto thyroiditis, 19 contained papillary microcarcinomas and 40 did not. Atypical cell clusters were identified in 12 (20%) of 59 cases within areas of Hashimoto thyroiditis, including 8 (42%) of 19 with papillary microcarcinoma elsewhere and 4 (10%) of 40 without papillary microcarcinoma. The clusters were tiny (<1 mm) and distributed in multiple foci throughout the gland. These atypical clusters showed some but not all of the nuclear changes of papillary thyroid carcinoma. The nuclei showed variable degrees of overlapping, slight clearing and few grooves Image 1A . Nuclear pseudoinclusions were not identified. The cells lacked the abundant eosinophilic cytoplasm of typical Hürthle cells, and the nuclear/cytoplasmic ratio was high. Not only the minute size of the clusters, but also the inconclusive nuclear changes and the pattern of distribution precluded a diagnosis of papillary microcarcinoma. All 12 cases with atypical cell clusters were positive for HBME1 Image 1B and CK19 Image 1C .

The remaining epithelium in Hashimoto thyroiditis showed focal cytoplasmic HBME1 staining in the Hürthle cells without membrane staining. This purely cytoplasmic staining pattern was considered negative. CK19+ foci, away from the atypical cell clusters noted on H&E, were observed in 56 (95%) of 59 cases. These areas did not correlate with nuclear features of papillary thyroid carcinoma or with HBME1 positivity. The intensity of CK19 expression was variable, with most of the epithelium staining weakly.

Molecular Findings

The atypical cell clusters from all 12 cases were negative for the BRAF mutation by the colorimetric Mutector assay and by allele-specific PCR. Both assays detected BRAF mutations in 4 of 5 papillary thyroid carcinomas and 2 of 5 papillary microcarcinomas. No BRAF mutations were detected in 4 cases of uncomplicated Hashimoto thyroiditis.

The results of the 2 assays are shown in Table 2 . The results of allele-specific PCR assay correlated with the Mutector assay results in all cases. By the Mutector assay, the ratio of intensity of mutant type to wild type in the mutation-positive cases ranged from 0.31 to 0.8. In mutation-negative cases, this ratio was less than 0.05. Image 2 demonstrates examples of 4 representative cases analyzed for the BRAF mutation by allele-specific PCR. Cases of papillary microcarcinoma and papillary thyroid carcinoma with the BRAF mutation showed 149-bp bands for mutant and wild-type alleles, while uncomplicated Hashimoto thyroiditis and atypical cell clusters in Hashimoto thyroiditis showed a 149-bp band for the wild-type allele but not the mutant allele, indicating absence of the BRAF mutation.

Discussion

The increased incidence of papillary thyroid carcinoma in Hashimoto thyroiditis has led to the hypothesis that Hashimoto thyroiditis may contain a precursor to papillary thyroid carcinoma. Attention has been focused on the follicular epithelium in Hashimoto thyroiditis, which frequently shows atypical nuclear changes that fall short of the criteria for a definitive diagnosis of papillary carcinoma. The incidence of finding such atypical cell clusters in Hashimoto thyroiditis on routine microscopic examination has varied dramatically among different studies, ranging from 11% to 100%, which could be explained by the differences in sampling and the number of cases studied.24

Our finding of an HBME1+, CK19+ immunophenotype typical of papillary thyroid carcinoma in the atypical cell clusters lends credence to the hypothesis that they are preneoplastic. HBME1 staining in these clusters is significant because the observed membrane staining pattern differs from the cytoplasmic staining of background Hürthle cells lacking atypical features. The CK19 staining is less specific, however, because a similar pattern of cytoplasmic positivity is also seen in Hürthle cells without atypical nuclear features. These immunohistochemical results are similar to other studies. Prasad et al4 detected atypical clusters with papillary thyroid carcinoma–like nuclear alterations in all 23 (100%) cases of Hashimoto thyroiditis. HBME1 and CK19 staining was observed only in these atypical clusters in 26% and 43%, respectively. Similarly, Gasbarri et al3 found atypical follicular epithelial cells in 15 (11.3%) of 133 cases of Hashimoto thyroiditis. In their study, HBME1 staining was performed on 12 of 15 cases with atypical foci and revealed 100% HBME1 positivity in all atypical areas identified.

Image 1

Atypical cell clusters in Hashimoto thyroiditis. A, Tiny clusters of atypical epithelial cells showing nuclear overlapping and nuclear grooves (H&E, ×400). B, Strong membrane staining for HBME1 (×400). C, Cytoplasmic and membrane positivity for cytokeratin 19 (×400).

The presence of precursor lesions in Hashimoto thyroiditis has been supported at the molecular level by demonstrating papillary carcinoma–associated oncogenes, RET/PTC1 and RET/PTC3, in greater than 90% of Hashimoto thyroiditis cases, without microscopically detected papillary thyroid carcinoma.6,7 RET/PTC oncoprotein expression has also been detected in atypical cellular nodules associated with Hashimoto thyroiditis with some cytologic features of papillary thyroid carcinoma.14 Rhoden et al15 detected low-level RET/PTC rearrangements in nonneoplastic follicular cells in Hashimoto thyroiditis using 2 detection techniques, fluorescence in situ hybridization and real-time reverse transcriptase–PCR. However, the biologic significance of this low level of RET/PTC with respect to neoplastic transformation remains unknown. In their study, the authors used a quite low cutoff level of 3.5% positive cells to be considered positive for RET/PTC by fluorescence in situ hybridization. However, if a more widely accepted cutoff level of 10% was used, almost all cases of Hashimoto thyroiditis would be considered negative for RET/PTC.16 In addition, other studies did not detect RET/PTC rearrangements in Hashimoto thyroiditis.8

View this table:
Table 2

Despite the proven association of RET/PTC with papillary thyroid carcinoma, the specificity of this association has been challenged by several studies that reported the detection of RET/PTC in benign thyroid lesions.7,1719 In contrast, the BRAF mutation has been consistently reported as the most specific molecular alteration in papillary thyroid carcinoma. It is detected in papillary thyroid carcinoma from 29% to 83% of cases.12 A meta-analysis of 1,168 cases with papillary thyroid carcinomas showed that the tall-cell variant of papillary thyroid carcinomas demonstrated a higher incidence of BRAF mutations (79%), whereas the follicular variant of papillary thyroid carcinomas harbored the lowest percentage (17%).20 BRAF is one of the 3 isoforms of RAF, which mediates cellular responses to growth and differentiation signals.13 The BRAF mutation has been found at all stages of progression of papillary thyroid carcinoma, including microcarcinomas, advanced papillary thyroid carcinoma, and poorly differentiated carcinomas arising from papillary thyroid carcinoma.21

Image 2

Primer sets: lanes 1–5, wild type; lanes 6–10, mutant; samples: lanes 1 and 6, papillary microcarcinoma; lanes 2 and 7, Hashimoto thyroiditis with atypical cell clusters; lanes 3 and 8, uncomplicated Hashimoto thyroiditis; lanes 4 and 9, papillary thyroid carcinoma; lanes 5 and 10: negative control. bp, base pairs; MW, molecular weight.

We tested for the BRAF mutation in the atypical cell clusters in Hashimoto thyroiditis because it is frequently present in papillary thyroid carcinoma and it is also highly specific for papillary thyroid carcinoma.12 The absence of the BRAF mutation in these clusters suggests that they may not be preneoplastic. Our findings are similar to those of Sargent et al,22 who found no BRAF mutations in 47 cases of atypical epithelium in chronic lymphocytic thyroiditis. Similarly, Kim et al23 could not demonstrate BRAF mutations in 27 cases of Hashimoto thyroiditis without papillary thyroid carcinoma. However, the same study found the BRAF mutation in 14% of Hashimoto thyroiditis cases with papillary carcinomas.23

The frequent presence of atypical cell clusters in Hashimoto thyroiditis expressing HBME1 and CK19 is a potential diagnostic pitfall, especially in fine-needle aspiration specimens. Although several studies have reported the usefulness of HBME124,25 and CK1926,27 in the cytologic diagnosis of papillary thyroid carcinoma, our findings suggest that caution must be exercised in the interpretation of immunohistochemical results in Hashimoto thyroiditis. Specifically, in the setting of Hashimoto thyroiditis, a diagnosis of papillary thyroid carcinoma or papillary microcarcinoma should not be made on tiny foci of atypical epithelium with inconclusive nuclear features, even in the presence of HBME1 and CK19 positivity.

Tiny atypical cell clusters in Hashimoto thyroiditis may exhibit incomplete nuclear changes of papillary thyroid carcinoma and may also show an HBME1+, CK19+ immunophenotype typical of papillary thyroid carcinoma. The absence of the BRAF mutation in these clusters suggests that they may not be preneoplastic. Larger studies are needed to further examine and confirm our findings. Caution should be exercised in interpreting positive HBME1 or CK19 staining in Hashimoto thyroiditis.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
View Abstract