Diagnostic Utility of p62 Expression in Intranuclear Inclusions in Thyroid Core Needle Biopsy Specimens

Discussion

FNAC is the standardized preoperative diagnostic modality for thyroid nodules (3,8-10). Thyroid CNB was introduced as an alternative method in South Korea in the 1990s and has since been widely used (3,11). Samples collected by CNB contain more abundant tissue and maintain the architectural characteristics of the nodules; therefore, non-diagnostic or inconclusive findings are less common than in FNAC, which generally provides scattered and structurally dissolved tissue samples (3,12). Although CNB sampling may not be the optimal choice for all thyroid nodules, it is widely accepted as the best method to diagnose samples for which previous FNAC results were inconclusive or atypia of undetermined significance (3-6). The most advantageous characteristic of CNB over FNAC is that it is optimized to distinguish specific cancers, such as PTC, medullary carcinoma, lymphoma, and metastatic carcinoma, through IHC (2). However, the NuVas that mimic NuIns in PTC are a major limitation of CNB. Jung et al. stated that NuIns in PTC are sharply delimited, round in shape, and nuclear membrane-bounded, whereas NuVas are irregular in shape and pale-looking (2). However, it is sometimes impossible to distinguish true NuIns from NuVas using only these parameters, since NuIns do not appear in all cancer cells and have different sizes, shapes, and numbers.

The mechanism of NuIn formation is yet to be established. Nuclear inclusions can also be called pseudodenuded holes, nuclear holes, and NuIns, and there can be one or two within a nucleus (13). In 2010, Fischer et al. reported the diagnostic importance of the nuclear envelope in cancer (14). They classified nuclear envelope structural changes into three groups: nuclear envelope changes associated with chromosomal instability (cell-to-cell variations in nuclear envelope size, shape, and deep infoldings); conserved nuclear envelope structural features within an unstable genetic population (increased nuclear/cytoplasmic ratio and fragile nuclear lamina of small cell carcinoma); and nuclear envelope changes directly associated with activation of specific oncogenes (long nuclear infolding or nuclear groove and intranuclear cytoplasmic inclusions induced by BRAF, RET/PTC, and TRK/PTC in PTC). They explained intranuclear cytoplasmic inclusion formation in the third group as a phenomenon of long inward folding of the nuclear envelope and spherical cytoplasmic invagination partially protruding into the nucleus (14).

Since it is difficult to discriminate NuVas from NuIns in hematoxylin and eosin-stained slides of PTC core needle biopsy specimens (Figure 2A), we compared different biomarkers of the PTC, including Galectin-3 (Figure 2B), CK19 (Figure 2C), and p62 (Figure 2D). There were very small numbers of NuIns in Galectin-3 staining (Figure 2B). For CK19, both NuIns and cytoplasmic staining intensity of tumor cells were overwhelming (Figure 2C), however, there were some NuIns, that were not stained (Figure 2C), though they were obvious from hematoxylin and eosin-stained slides (Figure 2A). With p62 staining (Figure 2D), more NuIns were highlighted and could be discriminated by their strong intensity with more contrast (Figure 2D). Based on the previous study of Fischer et al. (14), we hypothesized that the nuclear envelope changes should be directly associated with activation of specific oncogenes (BRAF, RET/PTC, and TRK/PTC) or biomolecular changes in PTC and that it is more reasonable to show a different aspect of staining patterns in the cytoplasm from NuIns. We thought that p62 best reflected the mechanism of NuIns formation, therefore, we chose p62 staining to evaluate NuIns in the CNB of PTC specimens.

Based on the guidelines suggested by previous studies, we numerically evaluated the extent to which NuIns filled the nucleus using 1/2 or 1/3 criteria. The diameters of each nucleus (A) and NuIn (B) were measured, and the number of p62-expressing NuIn-positive (p62In) cells was determined in each CNB specimen using the criteria of 1/2 (B/A) and 1/3 (B/A). The criterion of 1/3 includes NuIns larger than 1/3 and smaller than 1/2 of the nuclear diameter. The criteria of 1/2 includes NuIns larger than 1/2 of the nuclear diameter (Figure 1A). The sensitivity and specificity of each criterion were analyzed (Table I). The sensitivity and specificity in distinguishing FA from PTC using the 1/2 criterion were 0.88 and 1.00, respectively. The sensitivity and specificity in distinguishing FA from PTC using the 1/3 criterion were 0.92 and 0.90, respectively. Finally, the sensitivity and specificity to distinguish FA from follicular variant papillary carcinoma using both the 1/2 and 1/3 criteria were 1.00 and 1.00, respectively. Hence, the 1/2 criteria had higher specificity in distinguishing PTC from FA than the 1/3 criteria, and applying either criterion would be useful in distinguishing FA from follicular variant papillary carcinoma. Therefore, we agree with Schwertheim et al. (7) that p62 is a useful marker for NuIns. However, since NuIns have different sizes, shapes, and numbers, distinct criteria for evaluating NuIns are mandatory. We suggest only assessing NuIns above certain sizes using the 1/2 criteria based on specificity in distinguishing PTC from FA in our study.

There have been several attempts to distinguish NuIns from NuVas using IHC to diagnose PTC (7,15,16). Schwertheimet al. (7) examined the expression of several autophagy-associated proteins in NuIns, including LC3B, ubiquitin, cathepsin D, p62/sequestosome 1, and cathepsin B. They proved that NuIns are entirely closed on both two and three dimensions with no connection to the cytoplasm. This is why p62 is found only in the NuIn and not in the cytoplasm. In addition, p62 is not found in vacuoles because they are not produced by cytoplasmic invagination, which is the main source of p62 and its cargo. Finally, they concluded that the BRAF V600E mutation was significantly associated with the number of NuIns in PTC (p=0.042) and with IHC patterns of autophagy-associated proteins in NuIns (p≤0.035). They assert that the expression of autophagy-associated proteins, degenerated organelles, and lysosomal proteases within NuIns indicate that they contribute to autophagy and proteolysis.

Autophagy is a condition in which severely undernourished cells arrest their growth and cannibalize their cytoplasmic organelles, proteins, and membranes for energy (17). Some genes that promote autophagy are tumor suppressor genes, that is, loss of these genes enhances tumor growth. However, under severe nutrient-deficient conditions, tumor cells might use autophagy to enter a state of metabolic hibernation, allowing them to survive for increased periods of time or under stress. Therefore, autophagy can be protumorigenic depending on the controlling signaling pathways (17). P62, also known as sequestosome 1, is an autophagy receptor that forms autophagosomes that engulf and cannibalize damaged organelles, toxic proteins, and microorganisms in the cytoplasm. When autophagosomes are fused with lysosomes, the digested cargo and p62 are degraded (18,19). P62 recognizes, targets, and eliminates toxic cellular materials during autophagy. However, p62 is not always beneficial for normal cell homeostasis. Under normal conditions, basal autophagy constantly eliminates p62 and related cargo from the cytoplasm. When autophagy is insufficient, p62 and related cargo accumulate in the cytoplasm, which contributes to a pathological status (18). A previous study found that when p62 was removed, liver cell hypertrophy and lysosomal enzymes originating from oxidative stress were reduced (18). We designed this study based on previous reports that p62 accumulation is pathogenic. As Schwertheim et al. (7) concluded that p62 expression in NuIns indicates their contribution to autophagy, we hypothesized that p62In cells may be created due to accumulation of p62 not eliminated by autophagy, and that they drive pathological changes in the thyroid tissue in PTC. When stimulated by growth factors, normal cells generally increase their glutamine and glucose uptake as energy sources. However, in cancer, oncogenic mutations associated with growth factor signaling and other intermediate factors such as RAS, MYC, or BRAF constitutively activate metabolic pathways, resulting in tumor progression (17). The combination of deterioration of autophagy above a certain level of p62 accumulation and growth factor-associated oncogenic mutations, such as BRAF, may contribute to the formation of p62In cells in PTC.

To the best of our knowledge, this is the first report to propose distinctive size criteria for NuIns in CNB samples of PTC. The limitation of this study was the small number of cases. We only selected patients who had undergone both CNB and resection in our hospital. The association of BRAF status and p62Ins cells using the 1/2 criteria in thyroid cancers should be further evaluated in resected specimens to reveal the mechanism of NuIn formation in thyroid cancer. In conclusion, the diagnostic usefulness of p62Ins cells in CNB specimens of the thyroid was elucidated in this study. We recommend the 1/2 criteria for identifying p62In cells and differentiating PTC and FA.