Evaluation of GARP immunohistochemical expression in papillary thyroid carcinoma

Abstract

Background

Glycoprotein A repetitions predominant (GARP) is a novel transmembrane protein highly expressed on the surface of regulatory T cells (Tregs), which are a subset of immunosuppressive T lymphocytes that play a major role in inhibiting the antitumor immune response. Many studies documented increased GARP expression in various tumors, which is related to a poorer prognosis, and only one single paper investigated its expression in thyroid tumors.

Aim

To evaluate the immunohistochemical expression of GARP in differentiated thyroid carcinomas and their tumor-infiltrating lymphocytes (TILs) in comparison to its expression in other benign and low-risk lesions.

Methods

Sixty-nine cases of different thyroid lesions were subgrouped into 37 cases of malignant thyroid neoplasms, 25 cases of benign thyroid lesions, and 7 cases of low-risk neoplasms collected from the Pathology Department Laboratories of Ain Shams University Hospitals during the period from January 2017 to December 2021 and stained immunohistochemically for GARP. Immunohistochemical (IHC) results were evaluated in thyroid epithelial cells and TILs. The expression of GARP was correlated with the different clinicopathological parameters.

Results

GARP expression discloses a significant statistical difference between the three studied groups (P < 0.001). High GARP expression was detected in 89.19% of the malignant cases and in 28.57% of low-risk neoplasms, while all benign lesions exhibited low GARP expression. High GARP expression of TILs was detected in 60% of the malignant cases. Synchronous high GARP expression in tumor tissue and in the surrounding TILs was detected in 63.16% of the malignant cases, yet these results did not reach statistical significance.

Conclusion

GARP is a marker of Tregs, whose high expression is increased in malignant over benign and low-risk lesions. It might be a potential novel target for anticancer immunotherapy.

Background

Thyroid cancer is the most common endocrine malignancy, and its incidence has been rapidly increasing all over the world, representing about 2.1% of all malignant tumors [1]. In Egypt, it represents about 1.5%, being responsible for 30% of all endocrine neoplasms [2]. The most common thyroid malignancy is papillary thyroid carcinoma (PTC), accounting for 80–90% of thyroid cancer cases, followed by follicular thyroid carcinoma (FTC), accounting for 10–15% [3].

Although 5-year survival rates for differentiated thyroid tumors reach up to 98%, the clinical outcomes for patients with recurrent and/or metastatic diseases are often poor, having a less than 50% 5-year survival rate despite conventional treatment modalities [4]. So, research on new therapeutic agents has been performed; one of them is cancer immunotherapy, which is anticipated now and soon to be a promising therapeutic tool for oncologists [1].

The thyroid cancer microenvironment is rich in immune-reactive cells, making it a good candidate for immunotherapy [5].

Tumor-infiltrating lymphocytes (TILs) are important members of the tumor microenvironment (TME); they play a vital role in the host’s immune response against malignant tumors. The type and intensity of TILs in TME are important [6]. One of the most significant immune-suppressive T-lymphocytes that is essential for stifling the antitumor immune response is the regulatory T cells (Tregs) [7].

Decreased Treg numbers or impaired function than normal stand at the basis of autoimmune and inflammatory diseases, by contrast; Tregs number is extensively increased in nearly all malignant tumors as a part of tumor immune escape phenomena, promoting tumor growth, invasion, and treatment failure. Suppression of Tregs function is considered to be an effective strategy for cancer immunotherapy, but its application is still limited [8].

Ollendorff et al. first discovered a new protein, glycoprotein A repetition predominant (GARP) encoded by the LRRC32 gene located in the chromosomal 11q13–14 regions, a chromosomic region frequently altered in many human cancers [9]. Until 2006, GARP attracted attention because of its gene amplification in aggressive forms of human metastatic carcinomas; after that, it was identified as a latent transforming growth factor-β (TGF-β) receptor expressed on Tregs. Over the past 3 years, some studies have found a substantial association between GARP and cancer by describing the pro-tumorigenic function of this protein in several human malignancies [10]. GARP is a type 1 transmembrane glycoprotein, highly expressed on the surface of activated CD4 + FOXP3 + Tregs. It induces FOXP3 expression and activates latent TGF-β on the surface of Tregs, promoting their immunosuppressive function [11].

Many previous studies documented GARP overexpression in many tumor cells, e.g., breast, colon, gastric, lung, brain, bone sarcomas, and primary malignant melanomas, and its overexpression is linked to a less favorable prognosis [12]. Some experimental studies consider GARP to serve as an immune therapeutic target [5, 13,14,15].

To our knowledge, only Zhang et al. investigated GARP expression in thyroid lesions [16], and no previous studies investigated GARP expression in TILs of thyroid tumors. Thus, we were motivated to evaluate the IHC expression of GARP in differentiated thyroid carcinomas and their TILs in comparison to its expression in other benign and low-risk lesions.

Methods

Study design

This is a comparative cross-sectional study including 69 cases of surgically excised thyroid samples received at the Pathology Department Laboratories of Ain Shams University Hospitals, during the period from January 2017 to December 2021.

Material and data collection

Clinicopathological data were retrieved from an electronic database system, and cases were retrieved anonymously using the coded archive number in a randomized way. Cases with insufficient clinicopathological data or associated with other malignancies were excluded from the study. Hematoxylin and eosin-stained slides were re-examined and re-evaluated according to the 2022 WHO classification of endocrine tumors for confirmation of the final diagnosis. Other histologic parameters (tumor focality, extrathyroid extension, lymphovascular invasion, lymph node metastasis, and TILs) were assessed. Accordingly, cases were classified into three groups: Group A: malignant thyroid neoplasms (37 cases); they included 12 cases of FTC and 25 cases of PTC; the latter included 19 cases of classic PTC and 6 cases of follicular variant of papillary thyroid carcinoma (FVPTC); Group B: benign thyroid lesions (25 cases); they included 8 cases of nodular hyperplasia (NH), 8 cases of Hashimoto thyroiditis (HT), and 9 cases of follicular adenoma (FA); and Group C: low-risk neoplasms (7 cases); they included 4 cases of non-invasive follicular tumor with papillary-like nuclear features (NIFTP) and 3 cases of well-differentiated tumors of uncertain malignant potential (WDT-UMP).

Immunohistochemical staining

Tissue sections of 4-μm thickness were obtained from formalin-fixed paraffin-embedded blocks, and IHC procedures using an automated Benchmark GX (Ventana, CA, USA) immunostainer were performed based on the manufacturers’ recommendations. Concentrated polyclonal GARP (anti-LRRC32) antibody from Abcam Company, UK (catalog number: ab231210), was diluted at 1:150 and applied to the sections, then incubated for 32 min at 37 °C. Positive and negative controls were included in each run. Sections from human colonic adenocarcinoma served as a positive control, while omission of the primary antibody served as a negative control. All IHC-stained sections were examined independently by two investigators who were blinded to clinical and histopathological data using Olympus BX41 light microscopy in at least five areas (× 400 magnification).

Scoring of GARP IHC expression in thyroid epithelial cells

Cytoplasmic and/or nuclear staining is considered positive. According to Carrillo-Gálvez et al., the percentage of stained cells was scored as follows: (1) no positive staining, (2) positive staining of < 10% of tumor cells, (3) positive staining of 10 to 50% of tumor cells, and (4) positive staining of > 50% of tumor cells. The staining intensity was scored as follows: (0) no expression, (1) low intensity, and (2) high intensity [5].

Then, a final combined score was calculated for each tumor by multiplying both percentage and intensity scores, producing a range from 0 to 8. The cases were further categorized into high and low GARP expressions using the mean value of the combined score as a cutoff point (the calculated cutoff point for our study was 3.95) [5].

Scoring of GARP IHC expression in TILs

The cutoff for lymphocytic infiltration in the tumor tissue was determined at 1%; tumors with less than 1% of stromal infiltrating lymphocytes were considered negative for TILs [17]. Membranous and/or cytoplasmic staining of TILs was considered positive and evaluated in terms of both staining intensity and percentage collectively, as follows: 0, negative; 1, weak staining of < 10% of cells; 2, moderate staining of 10–90% of cells; and 3, strong staining of > 90% of cells. Scores of (0 and 1) were designated as low expression, while scores of (2 and 3) were designated as high expression [18].

Data management and analysis

All the collected data was revised, coded, tabulated, and introduced to a PC using the statistical package for social science (SPSS 25), and suitable analysis was done according to the type of data obtained for each parameter as follows:

Descriptive statistics

Parametric numerical data were expressed as mean, standard deviation (SD), and range, whereas nonparametric numerical data were expressed as median and interquartile range (IQR). Nonnumerical data were expressed as frequency and percentage.

Analytical statistics

The ANOVA test was used to assess the statistical significance of the difference between more than two study group means. The Kruskal–Wallis test was used to assess the statistical significance of the difference between more than two study group ordinal variables. Post hoc test was used for comparisons of all possible pairs of group means. Mann–Whitney U test was used to assess the statistical significance of differences in nonparametric variables between the two study groups. Fisher’s exact test was used to examine the relationship between two qualitative variables when the expected count was less than 5 in more than 20% of cells. The diagnostic value indices (sensitivity, specificity, and diagnostic accuracy) were calculated for GARP. Receiver operating characteristic (ROC) curves were used to assess the diagnostic indices of quantitative data and calculate the best cutoff point. Statistical significance was set at P < 0.05. To assess the agreement between the two investigational methods, Kappa statistics were used. Kappa’s over 0.75 is excellent, 0.40 to 0.75 is fair to good, and below 0.40 is poor.

Results

Clinicopathological parameters

The mean age of the patients in the malignant, benign, and low-risk groups was 45.35 ± 17.22, 35.2 ± 9.99, and 33.71 ± 13.72 years, with the female predominance being 59.46%, 88%, and 71.43% years, respectively. The mean tumor size of the malignant, benign, and low-risk groups was 4.04 ± 2.3, 4.23 ± 1.59, and 2.84 ± 1.41 cm, respectively. Most of the malignant tumors were unifocal (70.27%). Extrathyroid extension was proven in 29.7% of the cases. A neck dissection procedure was performed in 23 cases, and 12 out of them revealed positive cervical lymph node metastasis (52.17%).

GARP IHC expression

GARP expression in the studied groups

• Group A: Malignant thyroid neoplasms. High GARP expression was detected in 89.19% of all malignant cases, representing 91.67% of FTC cases and 88% of PTC cases. All cases showed positivity of ≥ 10% of the tumor cells, and 75.68% showed positivity of > 50% of the tumor cells. High staining intensity was detected in 86.49% of the cases. The mean combined score was 7.08 ± 1.68 (Table 1) (Fig. 1A, B, C).

• Group B: Benign thyroid lesions. All cases exhibited low GARP expression, with a mean combined score of 1.44. No or low staining intensity of less than 10% of neoplastic cells was detected in 84% of the cases (Table 1) (Fig. 2B, C, D).

• Group C: Low-risk neoplasms. GARP expression was low in five out of seven low-risk cases (75% of NIFTP and 66.67% of WDT-UMP), with a mean combined score of 3.42 ± 1.67 (Table 1) (Figs. 1D and 2A).

Table 1 The IHC expression of GARP in different studied groups

Fig. 1

A A representative case of classic PTC shows well-formed neoplastic papillae (red arrows) and reveals high GARP expression, combined score 8, diffuse positive cytoplasmic and nuclear staining with high intensity (IHC ×100). B A representative case of FVPTC illustrates characteristic PTC nuclear features (red arrows) and reveals high GARP expression, combined score 8, and diffuse positive cytoplasmic staining with high intensity (IHC ×100). C A representative case of FTC with capsular invasion (red arrow) reveals high GARP expression, combined score 8, diffuse positive cytoplasmic and nuclear staining with high intensity (IHC ×40). D A representative case of WDT-UMP reveals high GARP expression, combined score 6, and focal positive cytoplasmic staining with high intensity (IHC ×200) (N.B.: All inset pictures are IHC ×400)

Fig. 2

A A representative case of NIFTP reveals low GARP expression, combined score 2, and focal positive cytoplasmic staining with low intensity (IHC ×200). B A representative case of FA shows an intact tumor capsule (red arrow) and reveals low GARP expression, combined score 0, and negative staining (IHC ×40). C A representative case of NH reveals low GARP expression, combined score 2, focal cytoplasmic, and nuclear staining with low intensity (IHC ×100). D A representative case of HT reveals low GARP expression of both thyroid epithelial cells (yellow arrow) (combined score 0) and surrounding lymphocytes (red arrow) (score 0) (IHC ×100). (N.B.: All inset pictures are IHC ×400)

GARP IHC expression in correlation with clinicopathological parameters

An inverse statistically significant relation was detected between GARP expression and tumor size in the low-risk neoplasms group. Low expression was detected in tumors of larger size. No other statistically significant difference was detected between GARP expression and any of the other parameters (Table 2).

Table 2 GARP IHC expression in correlation with clinicopathological parameters

The correlation between GARP IHC expression in different studied groups

The expression of GARP was significantly different among the three studied groups (p < 0.001). Further paired comparisons using post hoc tests revealed significant differences between malignant versus benign and low-risk groups and between low-risk versus benign groups (Table 3). No statistically significant difference was detected regarding GARP IHC expression between malignant tumor group subtypes (Table 4).

Table 3 The correlation between GARP IHC expression in different studied groups

Table 4 The correlation of GARP IHC expression between malignant tumor group subtypes

The diagnostic indices of GARP IHC expression and combined score between different thyroid lesions

GARP IHC expression in the malignant group compared with the benign group showed 89.2% sensitivity, 100% specificity, and 93.5% diagnostic accuracy. The specificity and diagnostic accuracy decreased when the malignant group was compared to the low-risk group, being 71.4% and 86.4%, respectively. The sensitivity was significantly decreased to 28.6% while comparing low-risk to benign groups, whereas the specificity was 100% and the diagnostic accuracy was 84.4% (Table 5).

The obtained cutoff points of the GARP combined score using the ROC curve were significant in diagnosing malignant versus benign groups; malignant versus low-risk groups; and low-risk versus benign groups (P < 0.001), with cutoffs > 3, > 6, and > 2 and diagnostic accuracy of 99.1, 92.7, and 80.6%, respectively (Table 5).

Table 5 The diagnostic indices of GARP IHC expression and combined score

GARP expression in tumor-infiltrating lymphocytes

Group A: Malignant thyroid neoplasms. Twenty malignant cases were positive for TILs; 18 of them were PTC. High GARP expression was detected in 60% of the malignant cases, with no significant difference in staining between PTC and FTC cases (Table 6).

Table 6 GARP IHC expression in tumor-infiltrating lymphocytes

Synchronous high GARP expression of thyroid epithelial cells and surrounding TILs was recorded in 63.16% of malignant cases, yet these results were statistically non-significant and verified poor agreement between the two studied parameters (Kappa < 0.40) (Table 7 and Figs. 3 and 4A).

Table 7 Agreement between GARP expression in tumor tissue and surrounding TILs

Fig. 3

Agreement between GARP expression in tumor tissue and TILs

Fig. 4

A A representative case of classic PTC with positive TILs reveals high GARP expression of both tumor cells (yellow arrow) (combined score 8) and TILs (red arrow) (score 3) (IHC ×100) (inset picture is IHC ×400). B A representative case of HT reveals low GARP expression of both thyroid epithelial cells (yellow arrow) (combined score 0) and surrounding lymphocytes (red arrow) (score 1) (IHC ×400)

No statistically significant difference was detected regarding GARP scoring or GARP expression of TILs between classic PTC cases with and without cervical lymph node metastasis showing positive TILs (Table 8).

Table 8 GARP IHC comparison in TILs between classic PTC cases with and without cervical lymph node metastasis

Group B and C: Benign and low-risk thyroid lesions. All NH and HT cases exhibited low GARP expression of their lymphocytic infiltrate (Fig. 4B). All FA and low-risk cases exhibited no notable lymphocytic infiltrate.

Discussion

The current study investigated GARP expression in different thyroid lesions and TILs. A significant difference in GARP expression and the corresponding combined score was found between the malignant and benign groups. These results are similar to Zhang et al. and are also supported by others, Jiang et al., Zimmer et al., and Metelli et al., who investigated GARP expression in other different malignancies of other organs [12, 14, 16, 19]. Thus, GARP might have a substantial utility in differentiating malignant thyroid lesions from potential benign histologic mimics.

GARP expression was not significantly different between our studied malignant thyroid subtypes (classic PTC, FVPTC, and FTC); high GARP expression was detected in all of them. The significance of GARP in differentiating the subtypes of malignant thyroid lesions needs further verification by testing in larger-scale studies.

The development and progression of thyroid cancer result from the accumulation of genetic and epigenetic changes, as documented by the Cancer Genome Atlas Network. These changes result in the disturbance of the two major signaling pathways: the mitogen-activated pathway (MAPK) and phosphoinositide-3-kinase (PI3K) [20]. David and Massagué reported that TGF-β can also activate MAPK and PI3K oncogenic pathways, and since GARP acts as a cell surface docking receptor for latent TGF-β, thus promoting secretion and activation of TGF-β, this mechanism can particularly explain GARP increase in thyroid carcinoma [21]. We presume that GARP increase leads to activation of TGF-β; the latter, with the accumulation of other mutations, activates MAPK and PI3K pathways, resulting in the development of thyroid neoplasm. Such assumptions shed light on the need for studies investigating the relationship between GARP expression and other molecular pathways of thyroid tumor development.

While GARP has been described as a transmembrane protein on the surface of Tregs, it exhibits cytoplasmic and/or nuclear localization in tumor cells. Zimmer et al. explained the possible dynamic interaction of GARP with other proteins in the cell [19], a hint for a second independent, unrecognized pathway for GARP to exert its tumor immunosuppressing function that is not related to the canonical TGF-β-dependent mechanism. Nevertheless, detailed information about these pathways is still mysterious and can be an interesting topic for further fruitful studies.

Our preliminary results, together with previous studies that revealed increased GARP expression in malignant tumors of other organs, conclude that GARP is involved in the process of oncogenesis and is a putative marker in thyroid malignancy.

The present study reported a subtle increase in GARP expression in classic PTC cases with positive cervical lymph node metastasis compared to cases with negative lymph nodes, but these results did not reach a significant statistical correlation. This is in agreement with Zhang et al. and is near to Jiang et al., who reported a statistically significant increase in GARP expression detected by the fluorescence-based (mIHC) technique in gastric carcinoma cases with advanced lymph node metastasis over cases with free ones, this can be attributed to the different tumor tissue and staining technique used, also, the limited number of cases cannot validate such results [12, 16]. Thus, more studies with a larger sample size, more studied clinicopathological parameters (e.g., other high-grade and more aggressive types of thyroid tumors, local tumor recurrence, and distant metastasis), and patients’ survival status are advisable to assess the correlation between GARP expression and thyroid tumor aggressiveness.

In the current study, GARP expression was statistically different in comparing low-risk groups versus both malignant and benign ones, but to our knowledge, no previous studies investigated GARP expression in low-risk neoplasms. Our preliminary results suggest that WDT-UMP and NIFTP represent a distinct group and may be regarded as precursor lesions of FVPTC, wherein a progressive transformation to a malignant phenotype may occur. Thus, strict, close follow-up for low-risk cases with high GARP expression is highly recommended.

Tregs are a subset of inhibitory T-lymphocytes. Decreased Treg numbers or impaired function than normal stands at the basis of autoimmune and inflammatory diseases; by contrast, Tregs number is extensively increased in nearly all malignant tumors as a part of tumor immune escape phenomena, promoting tumor growth, invasion, and treatment failure. It is clear that there have been few research studies on the expression of GARP in the tumor tissue and surrounding TILs, and our study is innovative in assessing GARP expression in thyroid TILs. Our findings showed high GARP expression of TILs in 60% of the malignant cases, with no statistical difference between positivity in PTC and FTC cases. These results are in concordance with Li et al. and Jin et al., who explored GARP expression of TILs in gastric and lung malignancies, respectively [18, 22].

Synchronous high GARP expression in tumor tissue and in the surrounding TILs was detected in 63.16% of malignant cases, but our results did not reach statistical significance, yet these findings were near to those of Jiang et al., who documented the same finding in gastric carcinoma [12].

No statistically significant difference regarding GARP scoring or GARP expression of TILs between classic PTC cases with and without cervical lymph node metastasis showing positive TILs was detected in our work; on the contrary, Jin et al. reported a positive association between increased GARP-positive Tregs in tumor tissue of lung cancer patients having lymph node metastasis, distant metastasis, and clinical stage [22]. All this confirms the complexity of the immunological host reaction to different cancer types and warrants further studies on a larger scale.

All our NH and HT cases exhibited low GARP expression of their lymphocytic infiltrate; this is in concordance with Yu et al., Liu et al., González-Amaro and Marazuela, and Chen et al. [23,24,25,26].

In a study of Treg numbers and activity in papillary thyroid carcinoma with and without HT, Zhao et al. reported that the number of Tregs was significantly decreased in PTC cases associated with HT compared to PTC-only cases [27], and they, in addition to Lai et al. and Dvorkin et al., reported that PTC cases developed in a background of HT, usually having a smaller tumor size, less lymph node metastasis, low recurrence, and a low mortality rate, and concluded that although chronic inflammation of HT is a risk factor for developing PTC, these cases are usually associated with a more favorable prognosis [28, 29]. These findings may be explained by the decreased number and function of inhibitory Tregs as a part of the HT disease mechanism, which is more favorable for body immunity against PTC. Unfortunately, none of our PTC cases was associated with HT to assess this conflict as an interesting point of research.

A lot of experimental studies consider GARP could serve as an immune therapeutic target [5, 13,14,15], but none of them investigated it as thyroid immunotherapy, and further studies at this point could be of help.

Conclusion

In conclusion, GARP is a potentially valuable marker, seems to be involved in the mechanism of oncogenesis, and could help as a useful IHC marker in differentiating malignant tumors from benign and low-risk counterparts with overlapped histological features. GARP is suggested to be introduced as a novel therapeutic target for thyroid carcinoma cases unresponsive to conventional treatment modalities.