Diagnostic and Prognostic Biomarkers of Adrenal Cortical Carcinoma
Ozgur Mete, MD, **¿ Hasan Gucer, MD, § Mehmet Kefeli, MD, | and Sylvia L. Asa, MD, PhD **¿
Abstract: The diagnosis of low-grade adrenal cortical carcinoma (ACC) confined to the adrenal gland can be challenging. Although there are diagnostic and prognostic molecular tests for ACC, they remain largely unutilized. We examined the diagnostic and prog- nostic value of altered reticulin framework and the immunoprofile of biomarkers including IGF-2, proteins involved in cell pro- liferation and mitotic spindle regulation (Ki67, p53, BUB1B, HURP, NEK2), DNA damage repair (PBK, y-H2AX), telomere regulation (DAX, ATRX), wnt-signaling pathway (beta-catenin) and PI3K signaling pathway (PTEN, phospho-mTOR) in a tissue microarray of 50 adenomas and 43 carcinomas that were characterized for angioinvasion as defined by strict criteria, Weiss score, and mitotic rate-based tumor grade. IGF-2 and proteins involved in cell proliferation and mitotic spindle regulation (Ki67, p53, BUB1B, HURP, NEK2), DNA damage proteins (PBK, y-H2AX), regulators of telomeres (DAXX, ATRX), and beta-catenin revealed characteristic expression profiles enabling the distinction of carcinomas from adenomas. Not all biomarkers were informative in all carcinomas. IGF-2 was the most useful biomarker of malignancy irrespective of tumor grade and cytomorphologic features, as juxtanuclear Golgi-pattern IGF-2 reactivity optimized for high specificity was identified in up to 80% of carcinomas and in no adenomas. Loss rather than qualitative alterations of the reticulin framework yielded statistical difference between carcinoma and adenoma. Angioinvasion defined as tumor cells invading through a vessel wall and intravascular tumor cells admixed with thrombus proved to be the best prognostic parameter, predicting adverse outcome in the entire cohort as well as within low-grade ACCs. Low mitotic tumor grade, Weiss
From the *Department of Pathology, University Health Network; ¡Department of Laboratory Medicine and Pathobiology, University of Toronto; ¿ Endocrine Oncology Site Group, The Princess Margaret Cancer Centre, Toronto, ON, Canada; §Department of Pathology, Recep Tayyip Erdogan University, Rize; and |Department of Pathology, Ondokuz Mayis University, Samsun, Turkey.
O.M .: concept and design; literature search and writing. O.M. and S.L.A .: critical reviews. O.M., S.L.A., H.G., and M.K .: data collection and/or processing. O.M. and H.G .: analysis and/or interpretation. S.L.A .: financial resource.
The preliminary results of this study were presented during the endocrine platform presentations at the USCAP meeting in 2014, San Diego, CA. Conflicts of Interest and Source of Funding: The authors have disclosed that they have no significant relationships with, or financial interest in, any commercial companies pertaining to this article.
Correspondence: Ozgur Mete, MD, FRCPC, Department of Pathology, University Health Network, 200 Elizabeth Street, 11th floor, Toronto, ON, Canada M5G 2C4 (e-mail: ozgur.mete2@uhn.ca).
Copyright @ 2017 Wolters Kluwer Health, Inc. All rights reserved.
score, global loss of DAXX expression, and high phospho-mTOR expression correlated with disease-free survival, but Weiss score and biomarkers failed to predict adverse outcome in low-grade disease. Our results underscore the importance of careful morpho- logic assessment coupled with ancillary diagnostic and prognostic biomarkers of ACC.
Key Words: angioinvasion, IGF-2, adrenal cortical carcinoma, reticulin, biomarkers, Ki67, phospho-mTOR, DAXX, tumor grade (Am J Surg Pathol 2017;00:000-000)
A drenal cortical carcinoma (ACC) is a rare endocrine malignancy with variable morbidity and mortality.1-6 ACCs with distant metastasis and/or invasive growth are easy to recognize. However, the distinction of noninvasive low-grade ACC from adrenal cortical adenoma (ACA) poses a diagnostic challenge6 that has resulted in the use of terms such as “atypical adenoma” or “adrenal cortical neoplasm of uncertain malignant potential.”1
The size and weight of an adrenal cortical neoplasm were initially thought to be useful parameters in the dis- tinction of malignancy; however, large tumor size and/or heavy tumor weight do not always indicate unequivocal malignancy. As a consequence, several algorithms and scoring schemes have been proposed to distinguish ACC from adenoma.7-13 The Weiss and modified Weiss scoring schemes have limitations as they fail to capture all ACCs.1,6,14,15 In addition, interobserver variation in the detection or application of certain morphologic criteria adds another level of complexity to the reproducibility of these scores.6,13,16 Although the recently proposed Helsinki score11,12 has been validated in a recent series, a reticulin algorithm9-10 that has shown promise in the di- agnosis of carcinoma remains to be validated in additional series.
To complicate the matter, intratumoral morphologic, proliferative, and molecular heterogeneity has been recognized in these neoplasms. Microscopic regions with low-grade pro- liferative features can be encountered in high-grade ACCs, and low-grade ACCs can contain areas indistinguishable from adenomas.1,5 Furthermore, recent observations also suggest the possibility of adenoma-carcinoma progression in some adrenal cortical neoplasms.1,6,17,18 These issues are clinically relevant with increasing detection of nonfunctioning adrenal cortical proliferations called “incidentalomas.”
The past decade has seen tremendous progress in our understanding of the molecular biology of adrenal cortical neoplasms.2-4,19-21 At the molecular level, around 90% of adult ACCs exhibit IGF-2 overexpression.2-4,22 Tran- scriptome profiling and the recently published TCGA network data demonstrated molecular alterations that can be applied as diagnostic and prognostic tests.2-4 The TCGA study expanded the 2 prognostic molecular clusters defined by transcriptome studies (cluster 1A and 1B) into 3 distinct prognostic molecular groups (cluster of clusters I-CoCI, CoCII, and CoCIII) based on DNA copy num- ber, mRNA expression, DNA methylation and miRNA expression that also showed correlation with cell pro- liferation and steroid phenotypes.4 However, these mo- lecular tests are still far from being adopted in routine diagnosis due to high cost and lack of accessibility.
In order to provide easy and affordable access to translational diagnostic, predictive, and prognostic bio- markers for the evaluation of surgical specimens with adrenal cortical neoplasms, we investigated the role of altered reticulin framework and immunoprofile of bio- markers of malignancy including IGF-2 as well as proteins involved in cell proliferation and mitotic spindle regu- lation (Ki67, p53, BUB1B, HURP, and NEK2), DNA damage repair (PBK and y-H2AX), telomere regulation (DAXX and ATRX), wnt-signaling pathway (beta-cat- enin) and the PI3K signaling pathway (PTEN and phos- pho-mTOR) in a surgical series of benign and malignant adrenal cortical neoplasms. We also evaluated the sig- nificance of vascular invasion as defined by rigid diag- nostic criteria23 as well as Weiss score, and the mitotic rate-based tumor grade as initially proposed by Weiss and subsequently endorsed by Giordano24 in the prognostica- tion of ACCs.
MATERIALS AND METHODS
Specimens and Clinicopathologic Parameters
After obtaining the approval of the institutional Research Ethics Board, a total of 93 surgical specimens (43 ACCs and 50 ACAs) were selected for the con- struction of tissue microarrays (TMA). Markers of adre- nal cortical differentiation (SF-1, Melan-A, calretinin, alpha-inhibin, and synaptophysin) were applied at the time of diagnostic workup of each neoplasm. All adrenal cortical neoplasms were classified according to the uni- versal diagnostic criteria endorsed by the WHO classi- fications including the modified Weiss criteria as well as the Lin-Weiss-Bisceglia criteria. All tumors were reviewed retrospectively by 3 pathologists to ensure the appropriate inclusion of the diagnostic categories (O.M., H.G., and S.L.A.). None of these cases had biomarkers used to make the diagnosis.
The mitotic grade was assessed based on mitotic count in 50 high power fields from high mitotic density areas in all specimens. ACCs displaying up to 20 mitotic figures per 50 high power fields were classified as low- grade carcinomas, whereas those exceeding 20 mitotic figures per 50 high power fields were recorded as high
2 | www.ajsp.com
grade carcinomas.24 The status of angioinvasion using rigid criteria23 characterized by tumor cells invading through a vessel wall and/or intravascular tumor cells admixed with thrombus was recorded in all carcinomas.
The available follow-up information was reviewed to determine the status of disease recurrence, distant meta- stasis and death of disease.
TMA Construction
Reference hematoxylin and eosin-stained sections of the donor tissue were reviewed and marked with at least 3 circles (each measuring 1 mm in diameter) to identify regions of tumor with no necrosis. Three cores of 1 mm diameter were taken from the donor paraffin block matching the circled regions, and were arrayed in a recipient paraffin block using a manual tissue arrayer.
Histochemistry
TMA blocks were subjected to Gordon-Sweet Silver histochemistry to assess the reticulin framework in all tumors. The extent of reticulin loss was scored as follows: score 1: no loss of reticulin framework; score 2: minimal loss (<25%) of reticulin framework; score 3: focal loss (25% to 50%) of reticulin framework; and score 4: obvious loss (> 50%) of reticulin framework. The mean reticulin score of each case was recorded. In addition, qualitative reticulin pattern changes (ie, pericellular reticulin pattern) were also documented.
Immunohistochemistry
Formalin-fixed paraffin-embedded sections (4 um) were dewaxed in 5 changes of xylene and rehydrated through graded alcohols. Antigen retrieval or unmasking procedures were applied as detailed in Table 1. Negative and positive control tissues were selected based on manufacturer recommendations as well as previous publications where these antibodies were applied. Multiple control experiments were undertaken to optimize each antibody. Endogenous peroxidase was blocked with 3% hydrogen peroxide. The detection system used was MACH 4 universal HRP polymer system (Intermedico; cat # BC-M4U534). Color development was performed with freshly prepared DAB (DAKO; cat # K3468). Sections were counterstained lightly with Mayer’s hematoxylin, dehydrated in alcohols, cleared in xylene and mounted with Permount mounting medium (Fisher; cat # SP15-500). The Ki67 (MIB-1) immunohistochemistry was performed in an automated stainer (Ventana Benchmark).
Individual cores were scored by multiplying the percent positive cells by intensity scores (1 to 3), and each tumor received an average score (max: 300) for IGF-2 (1/3000 and 1/6000), BUBIB, HURP, NEK2, PBK, PTEN, and phospho-mTOR. Complete (global) loss of DAXX, ATRX, and PTEN, and presence of nuclear beta-catenin expression was recorded. Nuclear labeling indices of p53, Ki-67 (MIB-1 antibody), and y-H2AX were determined by manual counting of the tumor cell nuclei in all 3 cores of each tumor sample.
Copyright @ 2017 Wolters Kluwer Health, Inc. All rights reserved.
| TABLE 1. Materials and Methods for Immunohistochemistry | ||||
|---|---|---|---|---|
| Antibody | Source | Catalog # | Pretreatment | Dilution and Incubation |
| PBK | Cell Signaling | 4942 | Citrate | 1/50 overnight |
| BUB1B | BD Transduction | 6102503 | Tris-EDTA 9.0 | 1/400 overnight |
| NEK2 | BD Transduction | 610593 | Tris-EDTA 9.0 | 1/600 overnight |
| MIB-1 (Ki-67) | Dako | M7240 | Ventana CC1 | 1/200 32 min |
| p53 | Leica | NCL-p53-D07 | Citrate | 1/1000 1 hour |
| Beta-catenin | BD Transduction | 610153 | Citrate+pepsin | 1/500 1 hour |
| IGF-2 | Abcam | Ab 9574 | Low-temperature ctitrate | 1/3000 overnight 1/6000 overnight |
| DAXX | Sigma | HPA008730 | Citrate | 1/250 1 h |
| ATRX | Sigma | HPA001906 | Citrate | 1/600 1 h |
| HURP (DLGAP) | Abcam | Ab84509 | Tris-EDTA | 1/700 overnight |
| PTEN | Cell Signaling | 9559 | Tris-EDTA | 1/50 overnight |
| Phospho-mTOR | Cell Signaling | 2976 | Citrate | 1/50 overnight |
| y-H2AX | Cell Signaling | 9718 | Cirate | 1/300 overnight |
Statistical Analysis
The reticulin score, biomarker expression scores, global loss of DAXX, ATRX, and PTEN expression, presence of nuclear beta-catenin expression, and nuclear labeling indices of p53, MIB-1, and y-H2AX were com- pared between carcinomas and adenomas. The biomarker profile, the status of vascular invasion, Weiss score, and
A
B
mitotic grade were correlated with disease-free survival and overall survival. The same variables were also as- sessed to determine correlation with disease-free survival and overall survival in low-grade ACCs. Statistical anal- yses were performed using an online t test module (www. graphpad.com).
RESULTS
Clinicopathologic Features of 43 ACCs
Selected clinicopathologic parameters were collected for patients with ACCs. This series consisted only of adult patients (patients’ age at the time of diagnosis ranged from 27 to 75 y). The mean tumor size was 12.8 cm (range: 4 to 30 cm). Thirty-two patients had complete clinical in- formation at the time of surgery but only 30 patients had follow-up information that ranged from 4 to 120 months (mean: 40.06 mo). Of 20 patients with an adverse outcome (67%), 13 died of disease or were terminally ill and 7 alive with disease at the time of study. Ten patients were alive without recurrence or distant metastasis (33%), constitut- ing the disease-free survival group (mean follow-up time: 59.6 mo; range: 27 to 120 mo). Overall survival in this cohort of 30 patients (ie, alive at the time of analysis) was 17 patients. All patients with adverse outcome had an- gioinvasive ACCs (Fig. 1).
There was no sarcomatoid ACC in this series. Six (14%) specimens had a predominant oncocytic cytomor- phology, which was represented in the TMA cores. ACCs enriched in oncocytic cells fulfilled the diagnostic criteria of malignancy in the Lin-Weiss-Bisceglia scheme. The Weiss score ranged from 4 to 9. The mean Weiss scores of deceased patients and alive patients were 6.42 and 6.14, respectively (P=0.6545). Disease-free patients had tumors with lower Weiss scores (mean: 5.40) when compared with those with adverse outcome (mean: 6.94) (P=0.0086).
The mean mitotic rate was 23.1 per 50 high power fields (range: 2 to 111). Twenty-five (58%) carcinomas had low-grade proliferative features and 18 (42%) carcinomas exhibited high-grade proliferative features. While 10% of disease-free patients had a high mitotic tumor grade, 50% of patients with adverse outcome had tumors with a high mitotic tumor grade (P=0.0305).
Four patients had histologic and/or imaging evi- dence of lymph node metastasis. Angioinvasion was noted in 24 of 32 (75%) patients with clinical information; this feature was determined during routine examination of the entire periphery of the tumors. Distant metastasis was recorded in 18 of 32 (56%) patients with available clinical information. Liver and lungs were the most common metastatic sites, followed by bone, skin, and mesentery. All 18 patients with distant metastasis had ACCs with angioinvasion characterized by intravascular tumor cells admixed with thrombus or tumor cells invading through a vessel wall associated with tumor-thrombus complexes. Ten of 18 (~55%) patients with distant metastasis had low- grade ACC with a mean mitotic rate of 12.3 per 50 high power fields (range: 4 to 19). The remaining eight patients (~45%) with distant metastatic disease presented with high grade ACC and displayed a mean mitotic rate of 45.6 per 50 high power fields (range: 26 to 60).
Gordon-Sweet Silver Histochemistry
Forty-three (86%) ACAs displayed an intact reticulin framework (mean score: 1.14) (Fig. 2A). Seven adenomas (~14%) had very focal minimal loss of reticulin (score 2) (Fig. 2B). A mesh-like reticulin pattern surrounding individual
tumor cells (pericellular pattern) was noted focally in 10 (20%) adenomas; among them, 2 displayed foci of a complete pericellular (entirely surrounding individual cells) pattern (2/10), whereas incomplete (partially surrounding individual cells) (4/10) and mixed (complete and incomplete) pericellular patterns (4/10) were more frequent in adenomas. Of note, 9 adenomas with focal pericellular pattern (incomplete pattern: 4, mixed pattern: 3, and complete pattern: 2) did not show quantitative loss of the reticulin framework, whereas one adenoma with a mixed pericellular reticulin pattern displayed only very focal reticulin loss (Figs. 2B, C).
All ACCs had an abnormal reticulin framework characterized by loss of the acinar reticulin pattern (mean score: 3.74) (Fig. 2D). Eleven carcinomas had a score of 3 and 32 carcinomas had a score of 4. In addition to these quantitative changes, 12 (~28%) carcinomas displayed variable qualitative changes characterized by the presence of an abnormal pericellular reticulin pattern (Figs. 2E, F). Eleven had individual tumor cells completely surrounded by reticulin (complete pericellular pattern); one had an incomplete pericellular pattern.
The mean reticulin scores were 3.74 and 1.14 carci- nomas and adenomas, respectively. The quantitative loss of reticulin framework was a distinctive finding between
A
B
D
E
F
4 | www.ajsp.com
carcinoma and adenoma (P=0.0001). However, the presence of any form of pericellular pattern did not yield a significant difference between adenoma and carcinoma (P=0.4739). In addition, all 43 carcinomas with altered reticulin had at least one of the following parameters: necrosis, angioinvasion or increased mitotic activity (>5/50 high power fields). Thus, all ACCs in this series would have been captured as malignant using the reticulin algorithm.9,10
Expression Profile of Biomarkers
The results of biomarkers are summarized in Tables 2-4.
Expression Profile of IGF-2
During the optimization of IGF-2 antibody, we observed that a high dilution was required to subtract the
| Biomarkers | Carcinomas | Adenomas | P |
|---|---|---|---|
| IGF-2, 1/3000 dilution | |||
| Overall positivity rate (%) | 95 | 50 | 0.0001 |
| Juxtanuclear Golgi pattern rate (%) | 78 | 0 | 0.0001 |
| Expression score (max: 300) | 215.73 | 12.20 | 0.0001 |
| IGF-2, 1/6000 dilution | |||
| Overall positivity rate (%) | 95 | 24 | 0.0001 |
| Juxtanuclear Golgi patternrate (%) | 80 | 0 | 0.0001 |
| Expression score (max: 300) | 171.34 | 4.60 | 0.0001 |
| Ki67 | |||
| Mean labeling index (%) | 9.20 | 1.50 | 0.004 |
| p53 | |||
| Mean labeling index (%) | 27.77 | 4.44 | 0.0001 |
| BUB1B | |||
| Overall positivity rate (%) | 62 | 4 | 0.0001 |
| Mean expression score (max: 300) | 27.10 | 0.12 | 0.0002 |
| Mean positive tumor cell percentage | 10.12 | 0.06 | 0.0001 |
| HURP | |||
| Overall positivity rate (%) | 12 | 0 | 0.0118 |
| Mean expression score (max: 300) | 6.31 | 0 | 0.0386 |
| Mean positive tumor cell percentage | 5.36 | 0 | 0.0567 |
| NEK2 | |||
| Overall positivity rate (%) | 30 | 2 | 0.0001 |
| Mean expression score (max: 300) | 7.91 | 0.04 | 0.0030 |
| Mean positive tumor cell percentage | 7.91 | 2 | 0.0030 |
| Y-H2AX | |||
| Mean percent of expression (%) | 25.47 | 7.95 | 0.0019 |
| PBK | |||
| Overall positivity rate (%) | 49 | 10 | 0.0001 |
| Mean expression score (max: 300) | 15.14 | 0.28 | 0.0016 |
| Mean positive tumor cell percentage | 6.39 | 0.12 | 0.0005 |
| DAXX | |||
| Rate of global loss (%) | 37 | 12 | 0.019 |
| Mean positive tumor cell percentage | 44.77 | 61.30 | 0.0347 |
| ATRX | |||
| Rate of global loss (%) | 48 | 16 | 0.0002 |
| Mean positive tumor cell percentage | 31.40 | 71.10 | 0.0001 |
| Beta-catenin | |||
| Rate of nuclear reactivity (%) | 14 | 0 | 0.0053 |
| PTEN | |||
| Rate of global loss (%) | 35 | 10 | 0.0038 |
| Mean expression score (max: 300) | 141.63 | 154.69 | 0.6021 |
| Phospho-mTOR | |||
| Overall positivity rate (%) | 37 | 40 | 0.8541 |
| Mean positive tumor cell percentage | 17.02 | 12.70 | 0.4336 |
| Mean expression score (max: 300) | 34.33 | 28.90 | 0.6995 |
| Biomarkers | Low Grade | High Grade | P |
|---|---|---|---|
| IGF-2, 1/3000 dilution | |||
| Juxtanuclear Golgi pattern rate (%) | 72 | 82 | 0.4517 |
| Expression score (max: 300) | 204.79 | 216.47 | 0.7102 |
| IGF-2, 1/6000 dilution | |||
| Juxtanuclear Golgi pattern rate (%) | 76 | 88 | 0.3336 |
| Expression score (max: 300) | 155.21 | 185.29 | 0.3573 |
| Ki67 | |||
| Mean labeling index (%) | 4.68 | 15.33 | 0.0177 |
| p53 | |||
| Mean labeling index (%) | 21.12 | 37 | 0.1324 |
| BUB1B | |||
| Mean expression score (max: 300) | 13.56 | 47 | 0.0276 |
| HURP | |||
| Mean expression score (max: 300) | 6.20 | 6.47 | 0.9683 |
| NEK2 | |||
| Mean expression score (max: 300) y-H2AX | 10.60 | 4.17 | 0.2594 |
| Mean percent of expression (%) | 26.16 | 24.47 | 0.8713 |
| PBK | |||
| Mean expression score (max: 300) | 5.76 | 28.17 | 0.0226 |
| DAXX | |||
| Rate of global loss (%) | 52 | 22 | 0.0503 |
| ATRX | |||
| Rate of global loss (%) | 52 | 50 | 0.8120 |
| Beta-catenin | |||
| Rate of nuclear reactivity (%) | 12 | 18 | 0.6180 |
| PTEN | |||
| Rate of global loss (%) | 28 | 44 | 0.6180 |
| Phospho-mTOR | |||
| Mean expression score (max: 300) | 52.40 | 7.33 | 0.0386 |
basal normal IGF-2 reactivity that can be identified in almost all normal cells; 2 dilutions (1/3000 and 1/6000) yielded satisfactory staining depending on the fixation status of the tumor. We therefore adopted the use of both IGF-2 dilutions and compared the performance of both dilutions. Two ACCs failed IGF-2 immunohistochemistry due to absence of tumor tissue cores in sections examined. Cytoplasmic granular IGF-2 expression at both 1/3000 and 1/6000 dilutions was noted in 39 (95%) carcinomas. Focal and weak cytoplasmic granular IGF-2 seen only at 1/3000 dilution was recorded in 25 (50%) ACAs (P=0.0001), whereas only 14 (28%) ACAs displayed focal and weak cytoplasmic positivity at the 1/6000 dilution (P=0.0001) (Fig. 3).
The mean IGF-2 scores at 1/3000 were 215.73 and 12.20 in ACCs and ACAs, respectively (P=0.0001). The extent of IGF-2 (1/3000) staining ranged from 20% to 100% of cells per carcinoma sample, whereas in positive adenomas, 10% to 80% of the cells expressed IGF-2.
The mean IGF-2 scores at 1/6000 were 171.34 and 4.60 in ACCs and ACAs, respectively (P=0.0001). The extent of IGF-2 (1/6000) staining ranged from 5% to 100% of cells per carcinoma sample, whereas 5% to 60% of the adenoma cells expressed IGF-2 per tumor sample.
Unlike ACAs, a distinct IGF-2 staining pattern characterized by a juxtanuclear Golgi pattern was noted in 32 (78%) and 33 (80%) ACCs using 1/3000 and 1/6000 dilutions, respectively (Fig. 3). As no adenomas exhibited
| Biomarkers | DFS | Adverse Outcome | P |
|---|---|---|---|
| IGF-2, 1/3000 dilution | |||
| Juxtanuclear Golgi pattern rate (%) | 80 | 68 | 0.5065 |
| Expression score (max: 300) | 225.50 | 199.55 | 0.4726 |
| IGF-2, 1/6000 dilution | |||
| Juxtanuclear Golgi pattern rate (%) | 80 | 77 | 0.8680 |
| Expression score (max: 300) | 143 | 154.32 | 0.7435 |
| Ki67 | |||
| Mean labeling index (%) | 3.10 | 14.91 | 0.0568 |
| p53 | |||
| Mean labeling index (%) | 26.10 | 29.18 | 0.8162 |
| BUB1B | |||
| Mean expression score (max: 300) | 14.40 | 37.45 | 0.2304 |
| HURP | |||
| Mean expression score (max: 300) | 6.50 | 9.09 | 0.7833 |
| NEK2 | |||
| Mean expression score (max: 300) | 14 | 8.64 | 0.5045 |
| Y-H2AX | |||
| Mean percent of expression (%) | 39.60 | 25.45 | 0.2886 |
| PBK | |||
| Mean expression score (max: 300) | 7.70 | 21.86 | 0.2965 |
| DAXX | |||
| Rate of global loss (%) | 60 | 18 | 0.0172 |
| ATRX | |||
| Rate of global loss (%) | 60 | 36 | 0.2245 |
| Beta-catenin | |||
| Rate of nuclear reactivity (%) | 10 | 23 | 0.4090 |
| PTEN | |||
| Rate of global loss (%) | 30 | 32 | 0.9213 |
| phospho-mTOR | |||
| Overall positivity rate (%) | 67.50 | 13.18 | 0.0170 |
| Frequency of vascular invasion (%) | 40 | 100 | 0.0001 |
| Mean Weiss score | 5.40 | 6.94 | 0.0086 |
| High mitotic tumor grade (%) (high grade: >20/50 HPF) | 10 | 50 | 0.0305 |
| Modified high mitotic tumor grade (%) (high grade: > 10/50 HPF) | 30 | 77 | 0.0093 |
DFS indicates disease-free survival; HPF, high power fields.
this pattern of staining, the identification of a juxtanuclear Golgi pattern of IGF-2 reactivity was identified as a distinctive diagnostic feature of carcinoma (P=0.0001). It is interesting that the higher 1/6000 dilution captured an additional carcinoma with this distinct staining pattern; however, this finding did not yield any statistically significant difference between dilutions (P=0.5990).
IGF-2 expression profile characteristics were not statistically different between low-grade and high-grade ACCs. This finding suggests that IGF-2 expression is not related to the mitotic grade of an ACC. There was also no difference with respect to cytomorphology of the tumor.
Expression Profile of Proteins Involved in Cell Proliferation and Mitotic Spindle Regulation
The mean Ki-67 labeling indices were 9.20% and 1.50% in ACCs and ACAs, respectively (P=0.004) (Figs. 4A, E). High grade ACCs had a significantly higher mean of 15.33% Ki-67 labeling index compared with low-grade ACCs with a mean of 4.68% (P=0.0177). The mean Ki-67 labeling indices
were 3.10% and 14.91% in tumors from patients with disease- free survival and adverse outcome, respectively (P=0.0568).
The mean percentages of p53-positive cells were 27.77% and 4.44% in carcinomas and adenomas, respectively (P=0.0001) (Figs. 4B, F). No statistical difference was noted between high grade (mean: 37%) and low-grade (mean: 21.12%) ACCs with respect to p53 staining (P=0.1324). The mean p53 labeling indices were 26.10% and 29.18% in tumors from patients with disease-free survival and adverse outcome, respectively (P=0.8162).
One ACC failed BUB1B immunohistochemistry due to absence of matching cores in sections examined. Twenty six of 42 (~62%) ACCs showed cytoplasmic BUB1B reactivity (mean positive tumor cell percentage: 10.12%, range: 5% to 70%), whereas only 2 (4%) ad- enomas exhibited a very focal cytoplasmic staining for BUB1B (mean positive tumor cell percentage: 0.06%; range: 1% to 2%) (P=0.001) (Figs. 4C, G). ACCs had a higher BUB1B mean expression percentage (P=0.0001). The mean BUB1B expression score was also significantly higher in carcinomas (mean score: 27.10, range: 10 to 210) when compared with adenomas (mean score: 0.12, range: 2 to 4) (P=0.0002). High grade ACCs were significantly associated with a higher BUB1B score (mean score: 47) when compared with low-grade carcinomas (mean score: 13.56) (P=0.0276). The mean BUB1B scores were 14.4 and 37.45 in tumors from patients with disease-free and adverse outcome, respectively (P=0.2304).
No ACA was positive for HURP. One ACC failed HURP immunohistochemistry due to absence of tumor tissue cores in sections examined. Five (~12%) ACCs showed HURP reactivity (mean positive tumor cell per- centage: 5.36%, range: 20% to 90%) (Figs. 4D, H). Four carcinomas displayed cytoplasmic reactivity whereas one had both cytoplasmic and nuclear staining. The mean HURP expression score was 6.31 in ACCs. Despite the low frequency of positivity in carcinomas, positivity for HURP and the mean HURP expression score were statistically significant when compared with adenomas (P=0.0118 and 0.0386, respectively). There was no statistical significance between low-grade (mean score: 6.20) and high-grade (mean score: 6.47) ACCs with respect to HURP expression (P=0.9683). The mean HURP expression scores were 6.50 and 9.09 in tumors from patients with disease-free survival and adverse outcome, respectively (P=0.7833).
Thirteen (~30%) ACCs showed cytoplasmic NEK2 reactivity (mean positive tumor cell percentage: 7.91%, range: 5% to 80%), whereas only one (2%) adenoma exhibited cytoplasmic expression in 2% of the tumor cells. The mean NEK2 expression score was significantly higher in carcinomas (mean score: 7.91, range: 5 to 80) when compared with adenomas (mean score: 0.04) (P=0.0030). There was no statistical significance between low-grade (mean score: 10.60) and high-grade (mean score: 4.17) ACCs with respect to NEK2 expression (P=0.2594). The mean NEK2 scores were 14.00 and 8.64 in tumors from patients with disease-free and adverse outcome, respectively (P=0.5045).
6 | www.ajsp.com
Copyright @ 2017 Wolters Kluwer Health, Inc. All rights reserved.
A
B
C
D
E
F
G
H
Expression Profile of Proteins Involved in DNA Damage Repair
Twenty-one (~49%) ACCs showed nuclear and cyto- plasmic PBK reactivity (mean positive tumor cell percent- age: 6.39%, range: 1% to 60%), whereas 5 (10%) adenomas exhibited nuclear and cytoplasmic expression (mean positive tumor cell percentage: 0.12%, range: 1% to 2%) (positivity rate, P=0.0001; mean positive percentage, P=0.0005). The mean PBK expression score was significantly higher in ACCs (mean score: 15.14, range: 2 to 180) when compared with adenomas (mean score: 0.28; range: 2 to 4) (P=0.0016). High grade carcinomas were significantly as- sociated with a higher PBK score (mean score: 28.17) when compared with low-grade carcinomas (mean score: 5.76) (P=0.0226). The mean PBK scores were 7.70 and 21.86 in tumors from patients with disease-free survival and adverse outcome, respectively (P=0.2965).
The mean y-H2AX percentages were significantly higher in ACCs (mean: 25.47; range: 1 to 100) when compared with ACAs (mean: 7.95; range: 0 to 90) (P=0.0019). There was no significant difference between high grade (mean: 24.47) and low-grade (mean: 26.16) ACCs with respect to y-H2AX expression (P=0.8713). The mean y-H2AX expressions were 39.60% and 25.45% in tumors from patients with disease-free survival and adverse outcome, respectively (P=0.2886).
Expression Profile of Proteins Involved in Telomere Regulation
Sixteen (~37%) ACCs showed global loss of nuclear DAXX expression. While 6 ACCs were diffusely positive, 21 carcinomas showed variable nuclear expression ranging
from 10% to 95% of the tumor nuclei. Six (12%) ACAs displayed global loss of nuclear DAXX expression while the nontumorous elements remained positive. Forty-three adenomas showed variable nuclear expression ranging from 10% to 95% of the tumor nuclei and 1 tumor was diffusely positive. The mean DAXX expression percentage of ACCs (mean: 44.77) was significantly lower than ACAs (mean: 61.30) (P=0.0347). In addition, global loss of DAXX was significantly higher in carcinomas than ad- enomas (P=0.019). The mean nuclear DAXX expression was slightly higher in high grade carcinomas (score: 57.78) when compared with low-grade carcinomas (score: 35.40), however, this finding did not yield statistical significance (P=0.0887). There was no significant difference in DAXX staining between low-grade and high-grade ACCs (P= 0.0503). While 60% of disease-free patients had tumors displaying global loss of DAXX expression, 18% of patients with adverse outcome had tumors displaying global loss of DAXX expression (P=0.0172).
Twenty-one (~48%) ACCs showed global loss of nuclear ATRX expression while the nontumorous elements remained positive. Although 1 ACC was diffusely positive for ATRX, 21 carcinomas showed variable nuclear expression ranging from 10% to 90% of the tumor nuclei. Eight (16%) ACAs displayed global loss of nuclear ATRX expression. Twenty adenomas showed variable nuclear expression ranging from 10% to 95% of the tumor nuclei. Twenty-two adenomas were diffusely positive. The mean ATRX expression percentage of ACCs (mean: 31.40) was significantly lower than ACAs (mean: 71.10) (P=0.0001). In addition, the global loss of ATRX was significantly higher in ACCs than adenomas (P=0.0002). The mean
A
B
C
D
E
F
G
H
nuclear ATRX expression was slightly higher in low-grade carcinomas (score: 32.60) than high grade carcinomas (score: 29.72). However, this finding did not yield any statistical significance (P=0.900). In addition, the global loss of nuclear ATRX expression was not significant between low-grade and high-grade ACCs (P=0.8120). While 60% of disease-free patients had tumors displaying global loss of ATRX expression, 36% of patients with adverse outcome had tumors displaying global loss of ATRX expression (P=0.2245).
Expression Profile of Beta-Catenin
With the exception of one ACC that failed beta- catenin immunohistochemistry due to absence of tumor tissue cores, all neoplasms were positive for beta-catenin. No nuclear beta-catenin expression was noted in ACAs, whereas 6 (~14%; 3 high grade and 3 low grade) ACCs showed diffuse nuclear and cytoplasmic beta-catenin staining. This finding was a distinctive feature of carci- noma (P=0.0053). There was no statistical difference between low-grade and high-grade carcinomas with re- spect to nuclear beta-catenin reactivity (P=0.6180). While 10% of disease-free patients had tumors displaying nuclear beta-catenin expression, 23% of patients with adverse outcome had tumors displaying nuclear beta-catenin ex- pression (P=0.490).
Expression Profile of PI3K Signaling Pathway Proteins
Fifteen (~35%) ACCs showed global loss of PTEN expression, whereas 5 (10%) adenomas revealed global loss of PTEN expression. The global loss of PTEN was statistically more frequent in carcinomas (P=0.0038). The
PTEN expression score of carcinomas (mean: 141.63; range: 20 to 300) was not significantly different than ad- enomas (mean: 154.69; range: 10 to 300) (P=0.6021). In addition, there was no significant difference between low- grade and high-grade carcinomas with respect to global PTEN loss (P=0.2752). While 30% of disease-free pa- tients had tumors with global loss of PTEN expression, 32% of patients with adverse outcome had tumors dis- playing global loss of PTEN expression (P=0.9213).
Variable cytoplasmic and/or mixed cytoplasmic and membranous phospho-mTOR expression was noted in 16 (~37%) and 20 (40%) ACCs and ACAs, respectively (P=0.8541). The mean phospho-mTOR expressing tumor cell percentage was 17.02% in carcinomas and 12.70% in adenomas (P=0.4336). There was no difference between carcinomas (mean score: 34.33) and adenomas (mean score: 28.90) with respect to phospho-mTOR scores (P=0.6995). Low-grade carcinomas showed higher phospho-mTOR score (mean: 52.40) when compared with high grade carcinomas (mean: 7.33) (P=0.0386). Tumors of disease-free patients had a higher phospho-mTOR (mean score: 67.50) expression than those with adverse outcome (mean score: 13.18) (P= 0.0170).
Prognostic Value of Biomarkers, Angioinvasion, Mitotic Grading, Weiss Score, and Tumor Size
The prognostic impact of each feature was assessed by correlating biomarkers with disease-free survival and overall survival in this series. Angioinvasion, mitotic tumor grade, Weiss score, global loss of DAXX expression (Fig. 5A), and phospho-mTOR expression (Fig. 5B) were significantly correlated with disease-free survival (P=0.0001, 0.0305, 0.0086, 0.0172, and 0.0170, respectively). Interestingly, the
8 | www.ajsp.com
A
®
B
Ki-67 and p53 labeling indices, nuclear beta-catenin expression and other biomarkers as well as tumor size did not show any significant correlation with disease-free in this series.
This study also investigated the value of vascular invasion, mitotic tumor grading, Weiss score, tumor size, and selected biomarkers (MIB-1 and p53 percentages, nuclear beta-catenin, global loss of DAXX, and ATRX) in the prediction of disease-free survival in patients with low-grade ACCs. Among these variables, the status of angioinvasion defined by rigid criteria was the only predictor of disease-free survival (P=0.0006) in patients with low-grade ACCs.
We examined whether the value of mitotic tumor grading could be improved by redefining the distinction between low-grade and high-grade carcinomas using 10 mitotic figures per 50 high power fields. This cutoff showed a better correlation with disease-free survival (P=0.0093) compared with conventional mitotic grading (P=0.0305).
DISCUSSION
In this study, altered reticulin framework, angioin- vasion, and juxtanuclear Golgi pattern of IGF-2 expression
Copyright @ 2017 Wolters Kluwer Health, Inc. All rights reserved.
were found to be the most useful diagnostic features of ACC. This series validated the usefulness of the recently proposed reticulin algorithm9,10 and also underscored that quantitative reticulin alterations rather than qualitative changes seem to be more valuable in the recognition of malignancy.
Some studies have implied a low impact of histo- pathologic examination in the prognostication of ACC. In contrast, we demonstrate the value of strict criteria for the recognition of vascular invasion in the prediction of ad- verse outcome in patients with ACCs. The criteria to di- agnose angioinvasion are controversial. In a previous study,23 we emphasized the importance of defining bio- logically significant vascular invasion in endocrine neo- plasms. Application of the rigid criteria predicts adverse outcome in several malignancies arising in endocrine organs.23-28 This definition is based on the biological evidence of tumor-induced extrinsic coagulation pathway activation that results in the formation of fibrin-thrombus which is thought to provide tumor cells with immune evasion, thus, facilitating tumor spread.23 Not surpris- ingly, this cohort also proves that the status of angioin- vasion, when defined using biologically relevant criteria, was the only prognostic biomarker in low-grade ACCs where other biomarkers failed to predict outcome. In contrast, the Weiss score and mitotic tumor grade showed significant correlations with disease free-survival in the entire series but not in low-grade disease. Our findings highlight angioinvasion as the most important prognostic biological parameter among all parameters and bio- markers examined in this series.
Considered to be a crucial mitogenic pathway in the development of the fetal adrenal cortex, the IGF pathway is known to stimulate basal and ACTH-driven steroidogenesis via IGF-1 and IGF-2 molecules in the normal adult cortex.29,30 At the molecular level, over- expression of IGF-2 due to alteration of IGF2/H19 locus on 11p15 is considered to be the hallmark of around 90% of ACCs.1-3,6,31,32 A recent study questioned the utility of IGF-2 as a serum biomarker in ACCs,33 as serum samples from patients with ACCs did not show any difference compared with those from patients with ACAs and healthy controls.33 Although it seems that IGF-2 overexpression is not a driver of ACC, it is an important biological event linked to ACC.34
The current study highlighted IGF-2 immuno- histochemistry as a reliable diagnostic biomarker of malignancy in adrenal cortical neoplasms. Regardless of the tumor proliferative grade or cytomorphology, ACCs displayed a characteristic juxtanuclear Golgi staining pattern which was absent in all ACAs. To our knowledge, 3 previous large-scale series investigated the role of IGF-2 immunohistochemistry in adrenal cortical neoplasms.35-37 In 2001, Erickson et al35 applied a monoclonal IGF-2 antibody (clone: S2F2, dilution: 1/100; Upstate Bio- technology, Lake Placid, NY) in a series of 67 ACCs and 64 ACAs. The authors reported variable cytoplasmic IGF-2 staining intensity in around 93% and 55% of carcinomas and adenomas, respectively.35 In 2014, Wang
et al37 studied IGF-2 immunohistochemistry (polyclonal sc-1415, dilution: 1/100; Santa Cruz Biotechnology Inc., CA) in 25 ACCs and 25 ACAs. They found IGF-2 reactivity in 64% and 28% of carcinomas and adenomas, respectively.37 In contrast to these studies, in 2006, Schmitt and colleagues investigated IGF-2 immuno- histochemistry (clone: S1F1, dilution: 1/400; Upstate Biotechnology) in 17 ACCs and 22 ACAs. In that study, IGF-2 expression was seen in around 77% of ACCs.36 Unlike carcinomas, they detected focal IGF-2 expression only in a single ACA; interestingly, that particular area with focal IGF-2 expression displayed cellular atypia and a MIB-1 labeling index of 10.7%36; in light of our current understanding of the pathogenesis of ACC, this area likely represents malignant transformation within an ACA. Nevertheless, one of the most important highlights from the series of Schmitt et al was the first observation of a distinct IGF-2 staining pattern characterized by Golgi pattern or juxtanuclear dot-like reactivity in all positive ACCs.36 This particular staining pattern was explained by the decreased secretion of mature IGF-2 and increased secretion of heavier precursor forms,38,39 suggesting altered IGF-2 processing in the Golgi apparatus of ACCs.36,40 Although the IGF-2 antibodies and method- ologies used in previous series are different, we have reproduced the initial observations of Schmitt et al36 in- dicating that the juxtanuclear Golgi reactivity attributable to abnormally translated IGF-2 is a distinctive feature of ACCs. Our results provide justification for the adoption of this biomarker.
ACC is a proliferation-driven malignancy and this study demonstrated significantly increased expression levels of biomarkers related to cell proliferation (Ki67, p53, BUB1B, HURP, and NEK2) in carcinomas com- pared with adenomas. However, there are carcinomas with low proliferative activity that develop metastases. This and previous series have demonstrated consistency with respect to the Ki67 labeling index often being > 5% in ACCs. 1,6,36,41,42
Staining for p53 has been well-known to serve as a biomarker of malignancy in adrenal cortical neoplasms.1-4,6 However, only around one quarter of ACCs have TP53 gene mutation resulting an altered p53 phenotype.1,4 As in previous studies,43-46 the ACAs in this series had sig- nificantly lower levels of p53 expression (mean: < 5%) than carcinomas (mean: > 25%). High grade ACCs had higher p53 expression than low grade, consistent with the enrich- ment of TP53 mutations in high grade carcinomas.1,44,47,48 However, p53 did not yield any statistical difference be- tween tumor grade and prognosis in this series, as in pre- vious studies. 44,45,49,50
Transcriptome studies showed that BUB1B was overexpressed in a subset of ACCs.2,3,51 BUB1B, a com- ponent of the spindle assembly checkpoint that regulates segregation of chromosomes during mitosis,52-56 correlates with aggressive behavior, increased proliferation, genomic instability and aneuploidy in several cancers.52-54,56-58 In our series, ACCs showed a statistically higher expression of BUB1B than ACAs. Giordano et al2 demonstrated that
BUB1B is overexpressed in cluster 1 (bad behavior) ACCs, whereas de Reyniès et al3 proposed a qualitative score based on difference between the expression values of BUB1B and PINK1 as a predictor of overall survival in ACC. A sub- sequent validation in a Brazilian cohort also underscored the predictive role of combined BUB1B-PINK1 expression profile in ACCs.51 BUB1B immunohistochemistry did not show a correlation with outcome in our series, however, its expression was more frequently encountered in high grade ACCs.
Previous data suggested that DLGAP5 is overex- pressed in a subset of ACCs and may predict disease-free survival.3,51 Given these findings, we investigated the expression of HURP (DLG7), the product of the DLGAP5 gene. HURP is a kinetochore protein that maintains mitotic spindle integrity and stabilizes microtubules.59,60 We found no HURP expression in ACAs, and only 5 ACCs expressed HURP in a mean of 5% of tumor cells. While patients with high-grade disease or adverse outcome had higher HURP expression, no statistical correlation was noted. Our results suggest that HURP imunohistochemistry has minimal value in the assessment of ACCs.
The expression of NEK2, a member of multiple kinases involving in spindle assembly and centrosome cycle,61-63 is largely unknown in adrenal cortical neo- plasms; however, NEK2 overexpression has been shown to result in mitotic errors due premature centrosome separation.64 Our findings suggest that higher expression may be a feature of ACC, but, NEK2 expression did not correlate with mitotic tumor grade or prognosis.
As one would have expected, the expression profile of Ki67, p53, BUB1B, HURP, and NEK2 suggest that carcinomas are more proliferative than adenomas. Although Ki67 and BUB1B showed significant correlation with mitotic grade, the lack of prognostic correlation for proliferative biomarkers may be partially due to the lim- itation of the study design, as intratumoral proliferative heterogeneity cannot be assessed in a TMA, particularly in a relatively small cohort with a slight predominance of low-grade ACCs and lack of very long-term follow-up in some patients.
Altered DNA damage repair is one of the charac- teristics of malignant neoplasms. H2AX protein, encoded by the H2AX gene located at 11q23, is an essential factor in DNA repair.65,66 Recent data suggested that y-H2AX is a reliable biomarker to detect early double-strand DNA breaks in malignant cells.65 It is also regarded as a potential therapeutic target for radiotherapy, as blocking y-H2AX by inhibiting upstream kinases or direct inhibition of H2AX function can promote cell death.65,67,68 We found no prog- nostic or mitotic grade correlation, but there was significantly higher y-H2AX expression in ACCs, providing a rationale for combined radiotherapy with synchronous blocking of H2AX and cytotoxic chemotherapy to enhance antitumor effect in ACCs. However, further studies are needed to expand the use of this biomarker in the management of patients with ACCs.
Recent data suggested that PBK (PDZ binding kinase) facilitates tumor growth via DNA damage repair
10 | www.ajsp.com
mechanisms as well as p38-mediated motility.69 Over- expression of PBK has been reported in several malignancies56,70-74 as well as in the transcriptome series of ACC.2 Consistent with this finding, we found statisti- cally significant differences in PBK expression. Although deceased patients and those with adverse outcomes had higher PBK scores, these observations did not yield any statistical significance. This finding may be of significance, as an antitumor effect of Ginsenoside Rh2 has been re- ported to act via inhibition of PBK in HCT116 colorectal cancer cells.75
Components of the DAXX/ATRX-chromatin re- modeling complexes facilitate deposition of the histone H3.376 and regulate telomere elongation; they are altered in several malignancies and have prognostic implications in other endocrine cancers.77-79 The current cohort of ACCs had frequent global loss of DAXX and ATRX; loss of DAXX was significantly more frequent in disease-free patients. The TCGA data identified shorter telomeres in adult ACCs than their matched normal samples.4 Inter- estingly, pediatric ACCs were reported to harbor longer telomeres.80 In contrast, the TCGA found an association between TERT expression and whole genome doubling (WGD), a molecular predictor of bad outcome in ACCs, as ACCs with WGD typically harbored shorter telomeres than non-WGD carcinomas. This important observation may elucidate the better prognosis of tumors with global loss of DAXX in our cohort.
CTNNB1 mutations do not seem to be specific to malignancy as they can occur in around 25% of sporadic ACAs and ACCs.81 In previous studies, molecular corre- lates of beta-catenin signaling were more frequent in aggressive tumors.2-4 In our series, diffuse nuclear beta- catenin expression was a feature of malignancy; while frequent in tumors with adverse outcome, it was not statistically significant, possibly due to the low frequency of nuclear beta-catenin expression, the slight predominance of low-grade disease, and the lack of very long-term follow-up.
Activation of the PI3K/AKT/PTEN/mTOR path- way is an uncommon finding in ACCs.82,83 A recent study highlighted mTOR pathway activation as an early step in tumorigenesis in a transgenic animal model of ACC.34 In addition, the TCGA data identified singleton gene-fusions including the EXOSC10-MTOR fusion resulting in high phospho-mTOR levels in ACCs.4 In the current series, ACCs and ACAs were no different with respect to PTEN and phospho-mTOR expression scores. Similarly, De Martino et al did not find any difference in mRNA levels of mTOR in ACCs and ACAs.83 Interestingly, higher phospho-mTOR expression scores significantly correlated with disease-free survival and low mitotic tumor grade in our series. Downregulation of phospho-mTOR signaling was reported in a subset of less differentiated aggressive ACCs83 and response to mTOR inhibitors was achieved only in a small subset of ACCs.83,84 A recent study also reported that phospho-mTOR was expressed in 32% of ACCs85 and everolimus inhibited in vitro cell growth in both mitotane-sensitive H295R and mitotane-resistant
SW13 ACC cell lines,85 however, combination of mito- tane with pasireotide and everolimus resulted in an an- tagonistic growth effect on mitotane-sensitive cell lines.85 Although these findings underscore the antagonistic drug interactions with mitotane, the authors also showed that oncocytic ACCs and ACCs with high Weiss score lacked phospho-mTOR expression.85 These observations may explain the limited response to everolimus in some pa- tients with advanced ACC,84 as well as the significant correlation of high phospho-mTOR expression with dis- ease-free survival and low mitotic tumor grade in the current study.
CONCLUSIONS
Our data show that a juxtanuclear Golgi pattern of IGF-2 reactivity in a highly sensitive assay provides a useful ancillary tool along with quantitative alteration of the reticulin framework in the diagnosis of ACC. Angioinvasion was the best prognostic parameter, corre- lating with adverse outcome in the entire cohort as well as in low-grade ACCs. Although transcriptome and TCGA studies have provided diagnostic and prognostic signatures of malignancy,2-4 they represent complex molecular analyses that are not applicable in routine diagnostics. Our data indicate that the application of careful morphological assessment with a small panel of ancillary biomarkers may prove valuable for the diag- nosis and prognosis of ACC.
REFERENCES
1. Else T, Kim AC, Sabolch A, et al. Adrenocortical carcinoma. Endocr Rev. 2014;35:282-326.
2. Giordano TJ, Kuick R, Else T, et al. Molecular classification and prognostication of adrenocortical tumors by transcriptome profiling. Clin Cancer Res. 2009;15:668-676.
3. de Reyniès A, Assié G, Rickman DS, et al. Gene expression profiling reveals a new classification of adrenocortical tumors and identifies molecular predictors of malignancy and survival. J Clin Oncol. 2009;27:1108-1115.
4. Zheng S, Cherniack AD, Dewal N, et al. Comprehensive pan- genomic characterization of adrenocortical carcinoma. Cancer Cell. 2016;29:723-736.
5. Duregon E, Volante M, Rapa I, et al. Dissecting morphological and molecular heterogeneity in adrenocortical carcinoma. Turk Patoloji Derg. 2015;31 (suppl 1):98-104.
6. Duan K, Giordano TJ, Mete O. Adrenal cortical proliferations. In: Mete O, Asa SL, eds. Endocrine Pathology. UK: Cambridge University Press; 2016:602-627.
7. Weiss LM, Medeiros LJ, Vickery AL Jr. Pathologic features of prognostic significance in adrenocortical carcinoma. Am J Surg Pathol. 1989;13:202-206.
8. Aubert S, Wacrenier A, Leroy X, et al. Weiss system revisited: a clinicopathologic and immunohistochemical study of 49 adrenocort- ical tumors. Am J Surg Pathol. 2002;26:1612-1619.
9. Volante M, Bollito E, Sperone P, et al. Clinicopathological study of a series of 92 adrenocortical carcinomas: from a proposal of simplified diagnostic algorithm to prognostic stratification. Histopathology. 2009;55:535-543.
10. Duregon E, Fassina A, Volante M, et al. The reticulin algorithm for adrenocortical tumor diagnosis: a multicentric validation study on 245 unpublished cases. Am J Surg Pathol. 2013;37:1433-1440.
11. Pennanen M, Heiskanen I, Sane T, et al. Helsinki score-a novel model for prediction of metastases in adrenocortical carcinomas. Hum Pathol. 2015;46:404 410.
12. Duregon E, Cappellesso R, Maffeis V, et al. Validation of the prognostic role of the “Helsinki Score” in 225 cases of adrenocortical carcinoma. Hum Pathol. 2017;62:1-7.
13. Papotti M, Libè R, Duregon E, et al. The Weiss score and beyond- histopathology for adrenocortical carcinoma. Horm Cancer. 2011;2:333-340.
14. Papotti M, Volante M, Duregon E, et al. Adrenocortical tumors with myxoid features: a distinct morphologic and phenotypical variant exhibiting malignant behavior. Am J Surg Pathol. 2010;34:973-983.
15. Giordano TJ. Adrenocortical tumors: an integrated clinical, patho- logic, and molecular approach at the University of Michigan. Arch Pathol Lab Med. 2010;134:1440-1443.
16. Tissier F, Aubert S, Leteurtre E, et al. Adrenocortical tumors: improving the practice of the Weiss system through virtual micro- scopy: a National Program of the French Network INCa-COMETE. Am J Surg Pathol. 2012;36:1194-1201.
17. Mete O, Asa SL. Precursor lesions of endocrine system neoplasms. Pathology. 2013;45:316-330.
18. Heaton JH, Wood MA, Kim AC, et al. Progression to adrenocortical tumorigenesis in mice and humans through insulin-like growth factor 2 and B-catenin. Am J Pathol. 2012;181:1017-1033.
19. Assie G, Giordano TJ, Bertherat J. Gene expression profiling in adrenocortical neoplasia. Mol Cell Endocrinol. 2012;351:111-117.
20. Assié G, Letouzé E, Fassnacht M, et al. Integrated genomic character- ization of adrenocortical carcinoma. Nat Genet. 2014;46:607-612.
21. Espiard S, Bertherat J. The genetics of adrenocortical tumors. Endocrinol Metab Clin North Am. 2015;44:311-334.
22. Ilvesmäki V, Kahri AI, Miettinen PJ, et al. Insulin-like growth factors (IGFs) and their receptors in adrenal tumors: high IGF-II expression in functional adrenocortical carcinomas. J Clin Endo- crinol Metab. 1993;77:852-858.
23. Mete O, Asa SL. Pathological definition and clinical significance of vascular invasion in thyroid carcinomas of follicular epithelial derivation. Mod Pathol. 2011;24:1545-1552.
24. Giordano TJ. The argument for mitotic rate-based grading for the prognostication of adrenocortical carcinoma. Am J Surg Pathol. 2011;35:471-473.
25. Erovic BM, Kim D, Cassol C, et al. Prognostic and predictive markers in medullary thyroid carcinoma. Endocr Pathol. 2012;23:232-242.
26. Erovic BM, Harris L, Jamali M, et al. Biomarkers of parathyroid carcinoma. Endocr Pathol. 2012;23:221-231.
27. Erovic BM, Goldstein DP, Kim D, et al. Parathyroid cancer: outcome analysis of 16 patients treated at the Princess Margaret Hospital. Head Neck. 2013;35:35-39.
28. Ozturk Sari S, Taskin OC, Gundogdu G, et al. The impact of phosphohistone-H3-assisted mitotic count and Ki67 score in the determination of tumor grade and prediction of distant metastasis in well-differentiated pancreatic neuroendocrine tumors. Endocr Pathol. 2016;27:162-170.
29. Voutilainen R, Miller WL. Coordinate tropic hormone regulation of mRNAs for insulin-like growth factor II and the cholesterol side- chain-cleavage enzyme, P450scc [corrected], in human steroidogenic tissues. Proc Natl Acad Sci U S A. 1987;84:1590-1594. Erratum in: Proc Natl Acad Sci U S A. 1987;84(17):6194.
30. Han VK, Lu F, Bassett N, et al. Insulin-like growth factor-II (IGF-II) messenger ribonucleic acid is expressed in steroidogenic cells of the developing ovine adrenal gland: evidence of an autocrine/paracrine role for IGF-II. Endocrinology. 1992;131:3100-3109.
31. Nielsen HM, How-Kit A, Guerin C, et al. Copy number variations alter methylation and parallel IGF2 overexpression in adrenal tumors. Endocr Relat Cancer. 2015;22:953-967.
32. Gaston V, Le Bouc Y, Soupre V, et al. Analysis of the methylation status of the KCNQ1OT and H19 genes in leukocyte DNA for the diagnosis and prognosis of Beckwith-Wiedemann syndrome. Eur J Hum Genet. 2001;9:409-418.
33. Patel D, Ellis R, Howard B, et al. Analysis of IGF and IGFBP as prognostic serum biomarkers for adrenocortical carcinoma. Ann Surg Oncol. 2014;21:3541-3547.
34. Batisse-Lignier M, Sahut-Barnola I, Tissier F, et al. P53/Rb inhibition induces metastatic adrenocortical carcinomas in a preclinical transgenic model. Oncogene. 2017;36:4445-4456.
35. Erickson LA, Jin L, Sebo TJ, et al. Pathologic features and expression of insulin-like growth factor-2 in adrenocortical neo- plasms. Endocr Pathol. 2001;12:429-435.
36. Schmitt A, Saremaslani P, Schmid S, et al. IGFII and MIB1 immunohistochemistry is helpful for the differentiation of benign from malignant adrenocortical tumours. Histopathology. 2006;49: 298-307.
37. Wang C, Sun Y, Wu H, et al. Distinguishing adrenal cortical carcinomas and adenomas: a study of clinicopathological features and biomarkers. Histopathology. 2014;64:567-576.
38. Boulle N, Logié A, Gicquel C, et al. Increased levels of insulin-like growth factor II (IGF-II) and IGF-binding protein-2 are associated with malignancy in sporadic adrenocortical tumors. J Clin Endo- crinol Metab. 1998;83:1713-1720.
39. Boulle N, Gicquel C, Logie A, et al. Fibroblast growth factor-2 inhibits the maturation of pro-insulin-like growth factor-II(Pro-IGF-II) and the expression of insulin-like growth factor binding protein-2 (IGFBP-2) in the human adrenocortical tumor cell line NCI-H295R. Endocrinology. 2000;141:3127-3136.
40. Duguay SJ, Jin Y, Stein J, et al. Post-translational processing of the insulin-like growth factor-2 precursor. Analysis of O-glycosylation and endoproteolysis. J Biol Chem. 1998;273:18443-18451.
41. Arola J, Salmenkivi K, Liu J, et al. p53 and Ki67 in adrenocortical tumors. Endocr Res. 2000;26:861-865.
42. Morimoto R, Satoh F, Murakami O, et al. Immunohistochemistry of a proliferation marker Ki67/MIB1 in adrenocortical carcinomas: Ki67/MIB1 labeling index is a predictor for recurrence of adreno- cortical carcinomas. Endocr J. 2008;55:49-55.
43. McNicol AM, Nolan CE, Struthers AJ, et al. Expression of p53 in adrenocortical tumours: clinicopathological correlations. J Pathol. 1997;181:146-152.
44. Reincke M, Karl M, Travis WH, et al. p53 mutations in human adrenocortical neoplasms: immunohistochemical and molecular studies. J Clin Endocrinol Metab. 1994;78:790-794.
45. Stojadinovic A, Brennan MF, Hoos A, et al. Adrenocortical adenoma and carcinoma: histopathological and molecular compa- rative analysis. Mod Pathol. 2003;16:742-751.
46. Das S, Sengupta M, Islam N, et al. Weineke criteria, Ki-67 index and p53 status to study pediatric adrenocortical tumors: is there a correlation? J Pediatr Surg. 2016;51:1795-1800.
47. Wasserman JD, Zambetti GP, Malkin D. Towards an understanding of the role of p53 in adrenocortical carcinogenesis. Mol Cell Endocrinol. 2012;351:101-110.
48. Ohgaki H, Kleihues P, Heitz PU. p53 mutations in sporadic adrenocortical tumors. Int J Cancer. 1993;54:408-410.
49. Stojadinovic A, Ghossein RA, Hoos A, et al. Adrenocortical carcinoma: clinical, morphologic, and molecular characterization. J Clin Oncol. 2002;20:941-950.
50. Edgren M, Eriksson B, Wilander E, et al. Biological characteristics of adrenocortical carcinoma: a study of p53, IGF, EGF-r, Ki-67 and PCNA in 17 adrenocortical carcinomas. Anticancer Res. 1997;17 (2B):1303-1309.
51. Fragoso MC, Almeida MQ, Mazzuco TL, et al. Combined expression of BUB1B, DLGAP5, and PINK1 as predictors of poor outcome in adrenocortical tumors: validation in a Brazilian cohort of adult and pediatric patients. Eur J Endocrinol. 2012;166:61-67.
52. Logarinho E, Bousbaa H. Kinetochore-microtubule interactions “in check” by Bubl, Bub3 and BubR1: the dual task of attaching and signalling. Cell Cycle. 2008;7:1763-1768.
53. Rizzardi C, Torelli L, Barresi E, et al. BUBR1 expression in oral squamous cell carcinoma and its relationship to tumor stage and survival. Head Neck. 2011;33:727-733.
54. Ando K, Kakeji Y, Kitao H, et al. High expression of BUBR1 is one of the factors for inducing DNA aneuploidy and progression in gastric cancer. Cancer Sci. 2010;101:639-645.
55. Lee YK, Choi E, Kim MA, et al. BubR1 as a prognostic marker for recurrence-free survival rates in epithelial ovarian cancers. Br J Cancer. 2009;101:504-510.
56. Stricker TP, Henriksen KJ, Tonsgard JH, et al. Expression profiling of 519 kinase genes in matched malignant peripheral nerve sheath tumor/plexiform neurofibroma samples is discriminatory and identifies
12 | www.ajsp.com
mitotic regulators BUB1B, PBK and NEK2 as overexpressed with transformation. Mod Pathol. 2013;26:930-943.
57. Pinto M, Vieira J, Ribeiro FR, et al. Overexpression of the mitotic checkpoint genes BUB1 and BUBR1 is associated with genomic complexity in clear cell kidney carcinomas. Cell Oncol. 2008;30: 389-395.
58. Yamamoto Y, Matsuyama H, Chochi Y, et al. Overexpression of BUBR1 is associated with chromosomal instability in bladder cancer. Cancer Genet Cytogenet. 2007;174:42-47.
59. Song L, Craney A, Rape M. Microtubule-dependent regulation of mitotic protein degradation. Mol Cell. 2014;53:179-192.
60. Wrighton KH. Mitosis: microtubules protect spindle assembly factors. Nat Rev Mol Cell Biol. 2014;15:150-151.
61. Barbagallo F, Paronetto MP, Franco R, et al. Increased expression and nuclear localization of the centrosomal kinase Nek2 in human testicular seminomas. J Pathol. 2009;217:431-441.
62. Fry AM. The Nek2 protein kinase: a novel regulator of centrosome structure. Oncogene. 2002;21:6184 6194.
63. Fletcher L, Cerniglia GJ, Nigg EA, et al. Inhibition of centrosome separation after DNA damage: a role for Nek2. Radiat Res. 2004; 162:128-135.
64. Hayward DG, Newbatt Y, Pickard L, et al. Identification by high- throughput screening of viridin analogs as biochemical and cell- based inhibitors of the cell cycle-regulated nek2 kinase. J Biomol Screen. 2010;15:918-927.
65. Kuo LJ, Yang LX. Gamma-H2AX-a novel biomarker for DNA double-strand breaks. In Vivo. 2008;22:305-309.
66. Pilch DR, Sedelnikova OA, Redon C, et al. Characteristics of gamma-H2AX foci at DNA double-strand breaks sites. Biochem Cell Biol. 2003;81:123-129.
67. Kao J, Milano MT, Javaheri A, et al. gamma-H2AX as a therapeutic target for improving the efficacy of radiation therapy. Curr Cancer Drug Targets. 2006;6:197-205.
68. Taneja N, Davis M, Choy JS, et al. Histone H2AX phosphorylation as a predictor of radiosensitivity and target for radiotherapy. J Biol Chem. 2004;279:2273-2280.
69. Ayllón V, O’connor R. PBK/TOPK promotes tumour cell prolifer- ation through p38 MAPK activity and regulation of the DNA damage response. Oncogene. 2007;26:3451-3461.
70. He F, Yan Q, Fan L, et al. PBK/TOPK in the differential diagnosis of cholangiocarcinoma from hepatocellular carcinoma and its involvement in prognosis of human cholangiocarcinoma. Hum Pathol. 2010;41:415-424.
71. Zhu F, Zykova TA, Kang BS, et al. Bidirectional signals transduced by TOPK-ERK interaction increase tumorigenesis of HCT116 colorectal cancer cells. Gastroenterology. 2007;133:219-231.
72. Park JH, Lin ML, Nishidate T, et al. PDZ-binding kinase/T-LAK cell-originated protein kinase, a putative cancer/testis antigen with an oncogenic activity in breast cancer. Cancer Res. 2006;66:9186-9195.
73. Ohashi T, Komatsu S, Ichikawa D, et al. Overexpression of PBK/ TOPK relates to tumour malignant potential and poor outcome of gastric carcinoma. Br J Cancer. 2017;116:218-226.
74. Wang MY, Lin ZR, Cao Y, et al. PDZ binding kinase (PBK) is a theranostic target for nasopharyngeal carcinoma: driving tumor growth via ROS signaling and correlating with patient survival. Oncotarget. 2016;7:26604-26616.
75. Yang J, Yuan D, Xing T, et al. Ginsenoside Rh2 inhibiting HCT116 colon cancer cell proliferation through blocking PDZ-binding kinase/T- LAK cell-originated protein kinase. J Ginseng Res. 2016;40:400-408.
76. Voon HP, Wong LH. New players in heterochromatin silencing: histone variant H3.3 and the ATRX/DAXX chaperone. Nucleic Acids Res. 2016;44:1496-1501.
77. Jiao Y, Shi C, Edil BH, et al. DAXX/ATRX, MEN1, and mTOR pathway genes are frequently altered in pancreatic neuroendocrine tumors. Science. 2011;331:1199-1203.
78. Scarpa A, Chang DK, Nones K, et al. Whole-genome landscape of pancreatic neuroendocrine tumours. Nature. 2017;543:65-71.
79. Heaphy CM, de Wilde RF, Jiao Y, et al. Altered telomeres in tumors with ATRX and DAXX mutations. Science. 2011;333:425.
80. Pinto EM, Chen X, Easton J, et al. Genomic landscape of paediatric adrenocortical tumours. Nat Commun. 2015;6:6302.
81. Tissier F, Cavard C, Groussin L, et al. Mutations of beta-catenin in adrenocortical tumors: activation of the Wnt signaling pathway is a frequent event in both benign and malignant adrenocortical tumors. Cancer Res. 2005;65:7622-7627.
82. Ross JS, Wang K, Rand JV, et al. Next-generation sequencing of adrenocortical carcinoma reveals new routes to targeted therapies. J Clin Pathol. 2014;67:968-973.
83. De Martino MC, Feelders RA, de Herder WW, et al. Character- ization of the mTOR pathway in human normal adrenal and adrenocortical tumors. Endocr Relat Cancer. 2014;21:601-613.
84. Fraenkel M, Gueorguiev M, Barak D, et al. Everolimus therapy for progressive adrenocortical cancer. Endocrine. 2013;44:187-192.
85. Germano A, Rapa I, Duregon E, et al. Tissue expression and pharmacological in vitro analyses of mTOR and SSTR pathways in adrenocortical carcinoma. Endocr Pathol. 2017;28:95-102.