2018 年 65 巻 5 号 p. 579-586
The pathogenesis of thyroid lymphoepithelial cysts is controversial, and two hypotheses have been proposed, namely derivation from branchial-derived remnants or from squamous metaplasia of the follicular cells. The aim of this study was to clarify the pathogenesis of thyroid lymphoepithelial cysts. We performed pathological and immunohistochemical examination of 21 thyroid lymphoepithelial cysts, 13 non-neoplastic squamous metaplasia samples without thyroid carcinoma, 13 solid cell nests, and 14 lateral cervical cysts. On ultrasound, half of thyroid lymphoepithelial cysts were interpreted as calcified nodules regardless of no calcification. Thyroid lymphoepithelial cysts and squamous metaplasia tended to be located in the central and lower portions of the thyroid, while solid cell nests were located in the upper and central portions (p < 0.05). In 95.2% of patients with thyroid lymphoepithelial cysts and all patients with squamous metaplasia, lesions were histologically associated with chronic thyroiditis forming lymph follicles. Hashimoto’s disease was serologically confirmed in 18 patients with lymphoepithelial cysts (85.7%) and 10 patients with squamous metaplasia (76.9%). Immunohistochemically, lymphoepithelial cysts showed nuclear positivity for PAX8, thyroid transcription factor 1, and p63. One lateral cervical cyst (7.1%) showed positive staining for PAX8, while solid cell nests were PAX8-negative. In three (14.3%) cases of thyroid lymphoepithelial cysts, squamous cells located on the superficial layer were focally and weakly positive for CEA. We concluded that thyroid lymphoepithelial cysts originate from follicular cells and are unrelated to solid cell nests and lateral cervical cysts arising from branchial-derived remnants.
LYMPHOEPITHELIAL CYSTS (LECs) are predominantly lined by attenuated or stratified squamous epithelium with dense aggregates of lymphoid tissue beneath the lining epithelium [1] and have been described at different sites, including the oral cavity [2, 3], salivary gland [4-6], parathyroid gland [7], pancreas [8], and thyroid [9-11]. Thyroid LECs were originally reported by Louis et al. in 1989 [9] and described as a lesion with histological similarities to lateral cervical cysts. The pathogenesis of thyroid LECs is controversial, and two hypotheses have been proposed. The first hypothesis is that LECs are branchial-derived remnants [9], such as the ultimobranchial bodies (solid cell nests) [10-12] and lateral cervical cysts [13]. It is suggested that the underlying pathologic process is squamous metaplasia and cystic degeneration of solid cell nests. The second hypothesis is that the pathogenesis of thyroid LECs derives from squamous metaplasia of follicular cells associated with Hashimoto’s disease [14]. Here, we performed pathological and immunohistochemical examination of 21 thyroid LECs and compared these to squamous metaplasia of the thyroid, solid cell nests, and lateral cervical cysts. The aim of this study was to clarify the pathogenesis of thyroid LECs.
We reviewed the pathology report database of 18,889 cases of thyroidectomy performed at Kuma Hospital between 2006 and 2017 and extracted the following thyroid pathology cases: 21 LECs, 13 squamous metaplasia, and 13 solid cell nests. Three patients exhibited both LEC and squamous metaplasia of the thyroid. LEC, squamous metaplasia, and solid cell nest were defined as a cyst composed of squamous cells, associated with dense lymphocytic infiltration at the periphery of the wall [1], non-neoplastic squamous cells without thyroid carcinomas [15-17], and an aggregation of interfollicular-located small nests composed of non-keratinized squamoid cells [18], respectively. All of LEC, solid cells nests, and squamous metaplasia were not the reason for the indication of thyroid resection. We also examined 14 cases of lateral cervical cyst in the lateral neck. Of these, 4 lateral cervical cysts were resected at Kuma Hospital during the same period, and the remaining 10 cases were referrals from Oita University Hospital. Clinical findings were obtained from the medical records of Kuma Hospital and Oita University.
Immunohistochemical staining was performed using 3-μm-thick, formalin-fixed, paraffin-embedded tissue. The primary antibodies used in the immunostaining and antigen retrieval methods are shown in Table 1. Staining was performed using the Leica Bondmax system (Leica Microsystems IL) and Bond refine kit (Leica Microsystems, IL) according to manufacturer recommendations. The positive decision criteria of immunostaining were defined as more than 10% of squamous or squamoid cells expressing moderate or strong positive staining [19]. The percentage of positive cells examined was determined as follows: 1+ (11–50%), 2+ (51–80%), and 3+ (>80%).
| Antibody | Clone | Vendor | Location | Antigen retrieval | Dilution |
|---|---|---|---|---|---|
| PAX8 | EPR13510 | Abcam | Cambridge, UK | Heat (pH9) | 1:250 |
| TTF-1 | 8G7G3/1 | Dako | Carpinteria, CA, USA | Heat (pH6) | 1:100 |
| p63 | 4A4 | Nichirei Biosciences | Tokyo, Japan | Heat (pH9) | ready to use |
| CEA | COL-1 | Nichirei Biosciences | Tokyo, Japan | (–) | 1:3 |
TTF-1, thyroid transcription factor-1; CEA, carcinoembryonic antigen
We assessed the statistical significance of the data between LEC and three other lesions using Fisher’s exact probability test and the independent t-test. P values < 0.05 were considered statistically significant.
Table 2 shows the clinical and pathological findings of the thyroid LECs, squamous metaplasia, solid cell nests, and lateral cervical cysts. The mean ages of patients with the four lesions were 61.8 years, 53.8 years, 55.1 years, and 35.1 years, respectively. Patients with lateral cervical cysts were significantly younger than those with LECs (p < 0.001). LECs and squamous metaplasia predominantly occurred in female patients (81.0% and 61.5%, respectively), while solid cell nests predominantly occurred in male patients (76.9%). There were no differences in the sex distribution of patients with lateral cervical cysts. Hashimoto’s disease was serologically confirmed in 18 patients with LECs (85.7%) and 10 patients with squamous metaplasia (76.9%). LEC, solid cell nest, and lateral cervical cyst were predominantly solitary lesions. All cases of squamous metaplasia involved multiple lesions. There was no significant difference in the laterality of the four lesions. On ultrasound, 15 (71.4%) out of 21 LECs were detected and their photographs were saved in medical record. They had been interpreted as benign in nine (60.0%), borderline in two (13.3%), malignant in one (6.7%), and no judgement due to calcification in three (20.0%). Five (33.3%) and eight (53.3%) LECs had been identified as cystic lesion and calcified lesion, respectively (Fig. 1). Fine needle aspiration cytology was performed for four (19.0%) LECs that included cyst fluid only in two, colloid goiter in one, and Hashimoto’s thyroiditis in one. Any case did not revealed squamous cells. LECs and squamous metaplasia tended to be located in the central and lower portions of the thyroid, and solid cell nests in the upper and central portions (p < 0.05). The size of LECs and lateral cervical cyst lesions were macroscopically measured (mean 12.0 mm and 42.4 mm, respectively) with a significant difference (p < 0.001).
| Lymphoepithelial cyst (21) |
Squamous metaplasia (13) |
Solid cell nest (13) |
Lateral cervical cyst (14) |
|
|---|---|---|---|---|
| Mean age (range) [years] | 61.8 (44–74) | 53.8 (31–70)* | 55.1 (28–70) | 35.1 (2–66)*** |
| Female:Male | 17:4 | 8:5 | 3:10** | 7:7 |
| TRAb >1.9 IU/L | 2/4 (50.0%) | 2/3 (66.7%) | 0/1 (0%) | — |
| TgAb >39.9 IU/mL or TPOAb >27.9 IU/mL | 18 (85.7%) | 10 (76.9%) | 0 (0%)*** | — |
| Solitary/Multiple | 17/4 | 0/13*** | 10/3 | 13/1 |
| Location | ||||
| Rt/Lt/Is | 9/12/0 | 5/7/1 | 6/7/0 | 6/8/– |
| Upper/Central/Lower | 1/12/8 | 2/4/7 | 5/7/1* | — |
| Mean size (range) [mm] | 12.0 (1–70) | <1 | <1 | 42.4 (21–68)*** |
| Unilocular/Multilocular | 17/4 | — | — | 14/0 |
| Histological findings | ||||
| Squamous or squamoid cell | 21 (100%) | 13 (100%) | 13 (100%) | 14 (100%) |
| With keratinization | 1 (4.8%) | 0 (0%) | 0 (0%) | 1 (7.1%) |
| Glandular cell | 0 (0%) | 0 (0%) | 8 (61.5%) | 1 (7.1%) |
| Goblet cell | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) |
| Ciliated cell | 0 (0%) | 0 (0%) | 0 (0%) | 1 (7.1%) |
| Intranuclear cytoplasmic inclusions | 0 (0%) | 1 (7.7%) | 4 (30.8%)* | 0 (0%) |
| Lymphocytic infiltration | ||||
| Peri-epithelial cuffing | 21 (100%) | 0 (0%) | 0 (0%) | 14 (100%) |
| With lymph follicles | 17 (81.0%) | — | — | 9 (64.3%) |
| Well-demarcated border | 3 (14.3%) | — | — | 14 (100%)*** |
| Intraepithelial permeation | 21 (100%) | 12 (92.3%) | 11 (84.6%) | 12 (85.7%) |
| Chronic thyroiditis | 20 (95.2%) | 13 (100%) | 0 (0%)*** | — |
*p < 0.05, **p < 0.01, ***p < 0.001
Statistical significance: between lymphoepithelial cysts and other three lesions.
TRAb, anti-thyroid stimulation hormone receptor antibodies; TgAb, anti-thyroglobulin antibodies; TPOAb, anti-thyroid peroxidase antibodies
Ultrasound of lymphoepithelial cysts are interpreted as cystic lesion (a) or calcified lesion (b).
Pathologically, LECs were unilocular in 81.0% and multilocular in 19.0% of cases. All lateral cervical cysts were unilocular. All of the four lesions examined were composed of non-keratinizing squamous or squamoid cells (Fig. 2). One LEC and one lateral cervical cyst were associated with focal keratinization. Out of 13 solid cell nests, 8 (61.5%) included glandular cells forming a small lumen with or without secretory materials (Fig. 3). In one lateral cervical cyst, focal glandular cells with cilia were observed. Goblet cells were not observed in all lesions examined. Intranuclear cytoplasmic inclusions were seen in one squamous metaplasia (7.7%) and 4 solid cell nests (30.8%). All the cases of LECs and lateral cervical cysts were characterized by cysts surrounded by dense lymphocytic infiltration (lymphocytic cuffing), and 81.0% of LECs and 64.3% of lateral cervical cysts were associated with lymph follicles, respectively. In all cases of lateral cervical cysts, the borders between the lymphocytic cuffing and the surrounding connective tissue were well demarcated. In 18 out of 21 cases of LECs (85.7%), the lymphocytic cuffing included the thyroid follicles, and the border between the cuffing and surrounding thyroid tissue was obscure. Intraepithelial permeation of lymphocytes was observed in 100% of cases with LECs, 92.3% of cases with squamous metaplasia, 84.6% of cases with solid cell nests, and 85.7% of cases with lateral cervical cysts (Fig. 2). Furthermore, 95.2% of patients with LECs and all patients with squamous metaplasia were histologically associated with chronic thyroiditis forming lymph follicles. There were no cases of chronic thyroiditis among patients with solid cell nests. All of the four lesions were not associated with calcification.
Histological findings of thyroid lymphoepithelial cyst (a), squamous metaplasia (b), solid cell nests (c), and lateral cervical cyst (d). (HE, ×40).
Solid cell nests showing small glandular lumens with proteinaceous, colloid-like materials (HE, ×40).
Table 3 shows immunohistochemical results of the four lesions. PAX8 showed nuclear positivity in all of LECs, and the incidence was significantly higher in LECs than that in the remaining three lesions. The immunoreactive nuclei were distributed in all layers of the squamous epithelia, and the reactivity was slightly stronger on the basal side of the squamous layer (Fig. 4a). Staining for PAX8 was positive in 76.9% of squamous metaplasia cases and the staining tended to be weaker than LEC staining. One lateral cervical cyst (7.1%) showed positive staining for PAX8. Solid cell nests were negative for the antibody. Thyroid transcription factor 1 (TTF-1) was positive in 57.1% of LECs, 76.9% of squamous metaplasia, 46.2% of solid cell nests, and 7.1% of lateral cervical cysts. The TTF-1 staining pattern was similar to that of PAX8 (Fig. 4b). Lateral cervical cyst testing positive for TTF-1 was also positive for PAX8. All of the four lesions tested positive for p63. The reactivity was almost strong, particularly on the basal side of the squamous layer or squamoid nests (Fig. 4c). In three (14.3%) out of the 21 LECs, squamous cells located at the superficial layer were focally and weakly positive for CEA (Fig. 4d). Two (15.4%) solid cell nests were focally positive for CEA. squamous metaplasia was negative for CEA. In 12 (85.7%) lateral cervical cysts, CEA was positive for squamous cells located on the superficial layer, and fluid content was also positive. The positive rate was significantly higher than that of LEC (p < 0.001).
| LEC (21) | SM (13) | SCN (13) | LCC (14) | |
|---|---|---|---|---|
| (1+/2+/3+) | (1+/2+/3+) | (1+/2+/3+) | (1+/2+/3+) | |
| PAX8 | 21 (100%) | 10 (76.9%)* | 0 (0%)*** | 1 (7.1%)*** |
| (0/2/19) | (2/5/3) | (0/0/0) | (0/0/1) | |
| TTF-1 | 12 (57.1%) | 10 (76.9%) | 6 (46.2%) | 1 (7.1%)** |
| (5/4/3) | (4/6/0) | (5/0/1) | (0/0/1) | |
| p63 | 21 (100%) | 13 (100%) | 13 (100%) | 14 (100%) |
| (0/0/21) | (0/1/12) | (0/0/13) | (0/0/14) | |
| CEA | 3 (14.3%) | 0 (0%) | 2 (15.4%) | 12 (85.7%)*** |
| (3/0/0) | (0/0/0) | (1/1/0) | (3/3/6) |
*p < 0.05, **p < 0.01, ***p < 0.001
Statistical significance: between lymphoepithelial cysts and other three lesions.
TTF-1, thyroid transcription factor-1; CEA, carcinoembryonic antigen; LEC, lymphoepithelial cyst; SM, squamous metaplasia; SCN, solid cell nest; LCC, lateral cervical cyst
Immunohistochemical findings of thyroid lymphoepithelial cyst (×40). Squamous epithelia are positive for PAX8 (a), thyroid transcription factor-1 (b), and p63 (c). Squamous cells located on the surface are weakly positive for carcinoembryonic antigen (d) (immunostaining, ×40).
The pathogenesis of thyroid LECs is controversial, and two hypotheses are proposed. Thyroid LECs are histologically lined by stratified squamous epithelium with dense aggregates of lymphoid tissue beneath the lining epithelium [1]. The morphology of LECs is similar to that of lateral cervical cysts, a branchial-derived remnant. Therefore, it has been suggested that thyroid LECs also arise from these branchial-derived remnants [13]. The lesions may originate from squamous metaplasia and cystic degeneration of solid cell nests and the enlargement may be related to the immunological mechanisms associated with autoimmune thyroiditis [13]. The lesion has been described as “branchial cleft-like cyst” [9, 11, 14, 20]. An alternative theory, based upon derivation from follicular cells, has also been proposed. Miyazaki et al. attributed the origin of LECs to squamous metaplasia of a cystic lesion which developed as a result of degeneration in a nodular goiter, or squamous metaplasia of the follicular cells associated with Hashimoto’s disease [14].
In this study, we demonstrated that LECs were predominant in female patients, were predominantly solitary, tended to be located in the central to lower portion of the thyroid, and were frequently associated with chronic thyroiditis. Concerning female predominance, we think that it is just due to female preponderance in thyroid lesions resected in our institute. Conversely, solid cell nests predominantly occurred in male subjects, were located in the upper to central portion of the thyroid, and were not associated with peri-epithelial cuffing of lymphocytes or chronic thyroiditis. Harach et al. and Mizukami et al. reported similar findings regarding the location of solid cell nests [18, 21]. Gender predilection varies according to different reports [18, 21, 22]. On the basis of these findings, in our view, LECs are not related to solid cell nests. Furthermore, this is supported by the finding that solid cell nests were not detected around LECs. Interestingly, on ultrasound, LECs were always cystic lesions. Half of LECs were interpreted as calcified lesions, nevertheless calcification is not demonstrated on histological examination. Here, we emphasize that calcified nodules seen in association with Hashimoto’s disease may be LEC.
LECs and squamous metaplasia shared a number of characteristics, such as female predilection, location in the central and lower portions of the thyroid, and the association with chronic thyroiditis. Three patients exhibited both LECs and squamous metaplasia in the thyroid. LECs appeared to be related to squamous metaplasia rather than solid cell nests. However, squamous metaplasia did not show peri-epithelial cuffing of lymphocytes. Furthermore, in all cases, squamous metaplasia consisted of multiple lesions, while 17 of 21 LECs were solitary. Apel et al. described six cases of LECs, all of which were associated with chronic thyroiditis, suggesting that enlargement may be related to immunological mechanisms associated with autoimmune thyroiditis [13]. Ryska et al. and Carter et al. reported LEC cases associated with chronic thyroiditis but without anti-thyroid autoantibodies, as in the three cases we describe [1, 12]. Therefore, the presence of autoimmune thyroiditis is not necessarily essential for the development of LECs.
Lateral cervical cysts are thought to originate from the branchial apparatus that did not completely obliterate during head and neck embryogenesis [23, 24]. Lateral cervical cysts and LECs are microscopically similar. Both types of cysts are composed of squamous cells surrounded by lymphocytic cuffing with lymph follicles. Lateral cervical cysts are clinically apparent in late childhood or early adulthood [23, 24]. However, LECs have been diagnosed in elderly patients [1, 11, 14]. In the cases described here, the mean age of patients with LECs was 61.8 years (range: 44 to 74 years). These findings do not easily fit a theory of developmental abnormalities.
Immunohistochemical studies using organ-specific antibodies are useful to confirm cell origin. For thyroid tissue, thyroglobulin, TTF-1, PAX8, and calcitonin have been generally used [25-27]. Thyroglobulin and TTF-1 react against mature follicular cells or well-differentiated cells derived from follicular cells [26]. PAX8 is expressed in cells from an earlier stage of thyroid development [28]. Then, anaplastic carcinoma, poorly differentiated carcinoma, and squamous cell carcinoma are positive for the antibody, and PAX8 immunostaining is useful in differentiating primary thyroid carcinoma from carcinoma originating from other organs [27]. In our study, both PAX8 and TTF-1, indicating follicular cell origin, were positive in 76.9% of cases of squamous metaplasia, while p63, indicating squamous differentiation, were positive in all cases of squamous metaplasia. LEC immunoreactivity for PAX8 and TTF-1 was 100% and 57.1%, respectively. Conversely, almost all cases of solid cell nests and lateral cervical cysts were negative for PAX8 and TTF-1. These findings confirm that LECs originated from follicular cells and are unrelated to solid cell nest and lateral cervical cyst pathogenesis.
Kang et al. and Haba et al. described the branchial origin of LECs based upon their CEA expression, which was comparable to that of lateral cervical cysts and solid cell nests [20, 29]. In our study, 14.3%, 15.4%, and 85.7% of LECs, solid cell nests, and lateral cervical cysts, respectively, were CEA positive. However, CEA expression has been observed on the cell membrane of squamous cells of the esophagus and in squamous cell carcinoma, which are unrelated to branchial-derived remnants [30]. Therefore, we conclude that the positivity for CEA in LECs does not indicate branchial origin.
In this study, we demonstrated that LECs were predominantly solitary, tended to be located in the central to lower portion of the thyroid, and were frequently associated with chronic thyroiditis. Immunohistochemically, the lesions were positive for PAX8, TTF-1, and p63. We conclude that LECs originate from follicular cells and are unrelated to solid cell nests and lateral cervical cysts arising from branchial-derived remnants. Therefore, we propose that thyroid LECs should not be described as “branchial cleft-like cysts.”
None of the authors have any potential conflicts of interest associated with this research.
