ELSEVIER
Original contribution
Pediatric adrenocortical tumors: morphological diagnostic criteria and immunohistochemical expression of matrix metalloproteinase type 2 and human leucocyte-associated antigen (HLA) class II antigens. Results from the Italian Pediatric Rare Tumor (TREP) Study project
Gaetano Magro PhDa,*, Giovanni Esposito MDb, Giovanni Cecchetto MDC, Patrizia Dall’Igna MDC, Raffaella Marcato MDb, Claudio Gambini MDd, Renata Boldrini MDe, Paola Collini MDf, Vittoria D’Onofrio MD9, Nunzio Salfi MDh, Emanuele d’Amore MDi, Andrea Ferrari MD, Gianni Bisogno MDk, Rita Alaggio MD’
ªDepartment G.F. Ingrassia, Anatomia Patologica, Università di Catania, 95123 Catania, Italy
Istituto Oncologico Veneto-IRCCS, 35128 Padova, Italy
“Department of Pediatrics, Section of Pediatric Surgery, Università, 35121 Padova, Italy
dServizio di Anatomia ed Istologia Patologica, Istituto Giannina Gaslini, 16147 Genova, Italy eServizio di Anatomia Patologica, Ospedale Pediatrico Bambin Gesù, 00165 Roma, Italy
‘Unità Operativa di Anatomia Patologica, Istituto Nazionale per lo Studio e la Cura dei Tumori (INT), 20133 Milano, Italy Servizio di Anatomia e Istologia Patologica, Azienda Ospedaliera di Rilievo Nazionale “Santobono Pausillipon,” 80129 Napoli, Italy “Unità Operativa di Anatomia e Istologia Patologica, Ospedale Sant’Orsola-Malpighi, 40138 Bologna, Italy ¡Unità Operativa di Anatomia e Cito-Istopatologia, Presidio Ospedaliero San Bortolo, 36100 Vicenza, Italy ¿Oncologia Pediatrica, Istituto Nazionale per lo Studio e la Cura dei Tumori (INT), 20133 Milano, Italy
kDivisione di Onco-Ematologia, Dipartimento di Pediatria, Università di Padova, 35121 Padova, Italy “Department of Pathology, Azienda Ospedaliera/Università, 35121 Padova, Italy
Received 23 January 2011; revised 19 April 2011; accepted 22 April 2011
Keywords:
Adrenocortical tumors; Pediatric; Histologic features; Matrix metalloproteinases; HLA class II antigens
Summary Pediatric adrenocortical tumors are neoplasms that only rarely occur in pediatric patients. Their clinical behavior is often unpredictable, and the histologic criteria of malignancy used in adults are not always useful in children. The aim of this study was to validate the prognostic value of the pathologic criteria of Wieneke et al and to evaluate the potential prognostic expression of matrix metalloproteinase 2 and human leucocyte-associated antigen (HLA) class II antigens in a series of 20 pediatric patients affected by adrenocortical tumors, who were enrolled in the Italian Pediatric Rare Tumor (TREP) Study between 2000 and 2007. The age range was 0 to 17.5 years (mean, 7.28 years) with a male-female ratio of 1:2. The mean follow-up was 64.4 months. The histologic diagnoses were
* This study was partially supported by a grant from “Fondazione Cassa Di Risparmio Di Padova e Rovigo,” Padova, Italy.
* Corresponding author.
E-mail address: g.magro@unict.it (G. Magro).
reviewed, and the cases were classified using the criteria for malignancy proposed by Wieneke et al. The immunohistochemical expression of matrix metalloproteinase 2 and HLA class II antigens was scored by semiquantitative analysis and compared with the clinicopathologic parameters and outcome. Based on the scoring system of Wieneke et al, 7 tumors were classified as malignant; 12 tumors, as benign; and only 1 tumor, with “unpredictable behavior.” In all cases, the clinical behavior was consistent with the pathologic criteria of Wieneke et al. Notably, areas of regressive myxoid changes, not included among the criteria of Wieneke et al, were observed in all but 1 case of malignant tumors and only in 2 cases of benign tumors. Matrix metalloproteinase 2 was focally to diffusely expressed in all malignant and in most benign tumors. HLA class II antigens immunoreactivity was absent in all benign tumors and restricted to rare isolated cells in most malignant tumors. Our findings confirm that the pathologic scoring system of Wieneke et al is a simple and reproducible diagnostic tool to predict prognosis in pediatric adrenocortical tumors. Unlike in their adult counterpart, the expression of matrix metalloproteinase 2 or the loss of HLA class II antigens does not discriminate between benign and malignant tumors in children. Although pediatric adrenocortical tumors seem to be similar histologically to their adult counterparts, it is likely that they have distinctive molecular features. @ 2012 Elsevier Inc. All rights reserved.
1. Introduction
Adrenocortical tumors (ACTs) represent about 0.2% of all pediatric malignancies and 5% to 6% of all adrenal tumors, with an incidence of 0.2 new cases per 1 million children per year. Their incidence varies across geographic regions, but it is remarkably high in southern Brazil, wherein it is about 12 to 18 times the incidence reported in the United States and Europe [1-3]. The male-female ratio is 2:1 [4-6]; and the age incidence curve is characterized by 2 peaks, the first being younger than 3 years and the second during adolescence [7,8]. The distinction of benign versus malig- nant ACTs in children is challenging because the established criteria useful in adulthood often lack sensitivity and specificity in pediatric age [1,8]. In this regard, children with tumors classified as malignant according to the pathologic criteria adopted in adults surprisingly experienced a better clinical outcome. Previous studies have suggested that clinical parameters, including disease stage, age at diagnosis, presenting signs of endocrine dysfunction, and tumor weight and volume, are important prognostic markers in pediatric ACTs [9,10]. In addition, a separate staging system for childhood ACTs has been proposed [2]; but it appears to be complex and difficult to apply, especially in consultation practice. In 2003, Wieneke et al [8] proposed the categorization of pediatric ACTs into 3 different prognostic groups, “benign,” “malignant,” and “indetermi- nate for malignancy” according to a scoring system based on a combined assessment of 9 macroscopic and microscopic parameters that were found to be the most statistically significant for a potential malignant behavior. However, it was emphasized that none of these adverse pathologic features can be used solely as a predictor of malignancy and that this classification, albeit useful in daily practice, is lacking specificity and sensitivity when evaluating a single patient’s diagnosis [8]. Nevertheless, the scoring system of Wieneke et al has not been tested in other pediatric series; and thus, its predictive value is not yet confirmed. Several
potential biomarkers, including p53, c-Erb-B2, Bcl-2, and Ki-67 have been assessed in pediatric ACTs with disap- pointing and conflicting results [9,11]. A study on a series of 30 pediatric ACTs has emphasized that histone messenger RNA labeling index (LI) may assist not only in differenti- ating adenomas from carcinomas but also seems to be a reliable prognostic parameter [11]. Most recently, West et al [12], based on gene expression profiling analysis, identified 6- to 8-fold lower levels of HLA-DRB1, HLA-DPB1, HLA- DRA, and HLA-DPA1 in pediatric adrenocortical carcino- mas as compared with adenomas. Although the immunohis- tochemical expression of some matrix metalloproteinases (MMPs) [13] and HLA class II antigens [14] has been studied in adult ACTs displaying promising results, to the best of our knowledge, similar studies have not been extended to pediatric tumors.
The aim of the present study was to validate the prognostic significance of the scoring system proposed by Wieneke et al [8] and to assess the potential prognostic role of the expression of MMP2 and HLA class II antigens in a series of 20 cases of pediatric ACTs.
2. Materials and methods
2.1. Case selection and clinical review
Among 31 cases of ACTs that were enrolled onto the Italian Pediatric Rare Tumor (TREP) Study between 2000 and 2007, only 20 patients could be included in the present study. This selection was based on the availability of paraffin blocks or unstained sections for histologic review. For each case, a wide sampling had been performed at the time of diagnosis, with at least 1 sample per centimeter; the number of sections reviewed ranged between 3 and 46, with an average of 11. The clinical features (endocrine symptoms, radiologic imaging, treatment, and follow-up data) were
obtained from the medical records. Pathologic features of the tumors, such as tumor size, weight, location, encapsulation, extra-adrenal soft tissue, and vena cava invasion, were obtained from the original pathologic reports of the referring pathologists, if not present in the revised slides. Histologic revision of all hematoxylin and eosin-stained slides was performed independently by 2 pathologists; and tumors were categorized into 3 categories, benign, malignant, or indeterminate for malignancy according to the grading score system proposed by Wieneke et al [8]. The following criteria were evaluated and recorded for each single tumor: (1) tumor weight more than 400 g, (2) tumor size more than 10.5 cm, (3) extension into periadrenal soft tissues and/or adjacent organs, (4) invasion into vena cava, (5) venous invasion, (6) capsular invasion, (7) tumor necrosis, (8) more than 15 mitoses per 20 high-power fields, and (9) presence of atypical mitotic figures. Tumors showing 2 or fewer of the aforementioned unfavorable criteria were classified as benign; tumors with 4 or more unfavorable criteria were regarded as malignant; and, lastly, tumors with 3 unfavor- able criteria were classified as indeterminate for malignancy [8] (Table 1).
Additional microscopic features, previously proposed for adult ACTs [15,16], were also evaluated: (1) clear cell component, (2) growth pattern (diffuse, trabecular, nested), (3) nuclear atypia (grades I-IV), (4) broad fibrous bands, (5) inflammatory infiltrate, and (6) myxoid changes (Table 2).
2.2. Immunohistochemistry
The immunohistochemical analyses were performed on 3- um-thick, formalin-fixed, paraffin-embedded sections from a single representative block for each primary tumor. The following primary antibodies were used: anti-HLA-DR (clone LN-3, mouse, dilution 1:300; Novocastra, Newcastle, UK) and anti-MMP2 (clone A-Gel-VC2, mouse, dilution 1:100; NeoMarkers, Fremont, CA). Antigens were detected by incubation with labeled polymer. In all cases, 3-3’ diaminobenzidine (DAB kit; Dakocytomation, Glostrup, Denmark) was used as chromogen substrate. Finally, the sections were slightly counterstained with Mayer hematox- ylin, dehydrated, cleared, and mounted. Sections of previ- ously tested positive tissues for both antigens were processed as positive controls. The specificity of each staining procedure was confirmed by replacing the primary anti- bodies with phosphate-buffered saline. The immunohisto- chemical results obtained with antibodies against MMP2 were scored according to previously published results in adult ACTs [13]. In this regard, the percentage of positively neoplastic stained cells was assessed by semiquantitative analysis according to a 3-tiered system (negative, 0% of positive cells; focal, ≤20% of positive cells; diffuse, >20% of positive cells). The expression of HLA class II antigens was evaluated separately on inflammatory/stromal cells and neoplastic cells. It was scored as negative (0% of positive
| Table 1 Correlation between the score system of Wieneke et al [8] and clinical outcome | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Case | Tumor size (cm) | Tumor weight (g) | Extension into periadrenal soft tissues and/or adjacent organs | Invasion into vena cava | invasion | Necrosis | Mitotic rate/atypical mitoses | Vascular invasion | Wieneke score/category | Follow-up |
| 1 | 3 | 25 | Absent | Absent | Absent | Absent | <15/absent | Absent | 0/benign | CR |
| 2 | 5 | 45 | Absent | Absent | Absent | Absent | <15/absent | Absent | 0/benign | CR |
| 3 | 5.5 | 50 | Absent | Absent | Absent | Absent | <15/absent | Absent | 0/benign | CR |
| 4 | 8 | 150 | Absent | Absent | Absent | Absent | <15/absent | Absent | 0/benign | CR |
| 5 | 5 | 40 | Absent | Absent | Absent | Absent | <15/absent | Absent | 0/benign | CR |
| 6 | 4.5 | 26 | Absent | Absent | Absent | Absent | <15/absent | Absent | 0/benign | CR |
| 7 | 3.5 | 15 | Absent | Absent | Present | Absent | <15/absent | Absent | 1/benign | CR |
| 8 | 7 | 50 | Absent | Present | Absent | Present | <15/absent | Absent | 2/benign | CR |
| 9 | 2.5 | 60 | Absent | Absent | Absent | Absent | <15/absent | Absent | 0/benign | CR |
| 10 | 5 | 30.5 | Absent | Absent | Absent | Absent | <15/present | Absent | 1/benign | CR |
| 11 | 3 | 6.2 | Absent | Absent | Absent | Absent | <15/absent | Absent | 0/benign | CR |
| 12 | 5.5 | 66.1 | Diffuse | Absent | Present | Absent | <15/absent | Absent | 2/benign | CR |
| 13 | 10 | 170 | Diffuse | Absent | Present | Absent | >15/absent | Present | 4/malignant | CR |
| 14 | 11 | 60.6 | Diffuse | Present | Present | Present | >15/present | Present | 7/malignant | DOD |
| 15 | 11 | a | Present | Present | Present | 30%-40% | >15/present | Absent | 6/malignant | DOD |
| 16 | 25 | 3750 | Present | Absent | Present | 40% | <15/present | Present | 6/malignant | DOD |
| 17 | 4.5 | 25 | Present | Absent | Present | Present | <15/absent | Present | 4/malignant | AWD |
| 18 | 2 | 1075 | Present | Absent | Present | Present | >15/present | Absent | 6/malignant | AWD |
| 19 | 11 | 1200 | Absent | Present | Present | Present | <15/present | Present | 6/malignant | DOD |
| 20 | 5 | 80 | Absent | Absent | Absent | Present | <15/present | Present | 3/indeterminate | CR |
Abbreviations: CR indicates complete remission; DOD, died of disease; AWD, alive with disease.
a Only biopsy was performed. Size was retrieved from pathologic and imaging data.
| Case | Clear cells | Growth pattern | Nuclear grade | Myxoid changes | Fibrous septa | Inflammatory infiltrate | Wieneke category | Follow-up |
|---|---|---|---|---|---|---|---|---|
| 1 | <30% | Trabecular/nested | G1 | Absent | Absent | Present | Benign | CR |
| 2 | Absent | Diffuse | G2 | Absent | Absent | Present | Benign | CR |
| 3 | Absent | Diffuse | G3 | Absent | Absent | Present | Benign | CR |
| 4 | Absent | Trabecular/nested | G2 | Present | Absent | Absent | Benign | CR |
| 5 | Absent | Diffuse | G2 | Absent | Absent | Absent | Benign | CR |
| 6 | <30% | Trabecular/nested, | G1 | Absent | Present | Present | Benign | CR |
| with pseudocystic spaces | ||||||||
| 7 | Absent | Diffuse | G2 | Absent | Absent | Absent | Benign | CR |
| 8 | Absent | Diffuse | G4 | Present | Present | Absent | Benign | CR |
| 9 | <30% | Nested | G2 | Absent | Absent | Present | Benign | CR |
| 10 | Absent | Nested | G2 | Absent | Absent | Absent | Benign | CR |
| 11 | Absent | Trabecular | G1 | Absent | Absent | Absent | Benign | CR |
| 12 | Absent | Diffuse | G2 | Absent | Absent | Present | Benign | CR |
| 13 | Absent | Diffuse | G4 | Absent | Absent | Absent | Malignant | CR |
| 14 | Absent | Diffuse | G3 | Present | Absent | Present | Malignant | DOD |
| 15 | >30% | Trabecular/nested | G2 | Present | Absent | Absent | Malignant | DOD |
| 16 | Absent | Diffuse | G3 | Present | Present | Absent | Malignant | DOD |
| 17 | Absent | Diffuse | G2 | Present | Absent | Absent | Malignant | AWD |
| 18 | Absent | Trabecular/nested | G3 | Present | Present | Absent | Malignant | AWD |
| 19 | Absent | Diffuse paraganglioma-like areas | G2 | Present | Present | Absent | Malignant | DOD |
| 20 | Absent | Diffuse | G3 | Absent | Absent | Absent | Indeterminate | CR |
cells), focal (rare positive cells), intermediate (multiple foci of positive cells), and diffuse (>50% of positive cells).
3. Results
3.1. Clinical findings
Patients’ age at diagnosis ranged from 2 to 210 months (median, 107 months): females were 12 and males 8. The mean follow-up was 64.4 months. Most of the patients presented endocrine symptoms, such as virilization (7 patients) and Cushing syndrome (5 patients). A tumor mass was the main clinical sign in 3 patients. Three patients had metastases at diagnosis: lung (1 case), lung and liver (1 case), and bone and bone marrow (1 case). The assigned postsurgical stage was the following: ST I (complete resection), 12 cases; ST II (microscopic residual, tumor volume >200 cm3, positive lymph nodes, high hormonal blood level after surgery), 4 cases; ST III (biopsy/ macroscopic residuals), 1 case; and ST IV (metastases), 3 cases. Thirteen patients (11 patients ST I, 2 patients ST II) obtained a complete remission after the surgical procedure, and they are alive without evidence of disease at 25 to 108 months from diagnosis. One patient (ST I) was lost at follow- up. Two patients, 1 registered as ST III and 1 ST II because of tumor size, are alive with disease after treatment: the first received chemotherapy only, whereas the second developed lung and liver metastases after surgery and he was treated
with a combination of chemotherapy (cisplatin, etoposide, doxorubucin) and mitotane; he also received radiotherapy and multiple surgical procedures to treat the metastases. Four patients died of disease from 2 to 13 months after diagnosis. One patient, considered ST II because of tumor size, was treated with standard chemotherapy (cisplatin, adriamycin, VP16) and mitotane after the occurrence of lung and liver metastases. The other patients, all at ST IV, died of progressive disease, despite chemotherapy (cisplatin, etopo- side, and doxorubucin and mitotane).
3.2. Pathologic findings
Pathologic findings are summarized in Tables 1 and 2. All tumors were mostly composed of polygonal cells with abundant eosinophilic cytoplasm (Fig. 1). Focal nests of clear cells, less than 30% of the entire tumor, could be identified in 4 cases (Table 2). According to the scoring system of Wieneke et al [8], 12 tumors were included in the group of benign lesions, showing only 2 or fewer unfavorable prognostic markers; 7 cases with 4 or more unfavorable criteria were regarded as malignant; and, lastly, only 1 case, with 3 unfavorable criteria, was classified as indeterminate for malignancy. Notably, this pathologic classification reflected the clinical findings in that all 12 patients with tumors classified as benign are in complete remission with no relapse, whereas 4 of 7 patients with tumors classified as malignant died of disease, 2 are alive with metastases and only 1 is in complete remission. The unique patient of our
A
?
B
C
D
series with a tumor classified as indeterminate for malignancy is having a benign clinical course.
Among the morphological features that are not included in the scoring system of Wieneke et al, we found that neoplastic cells exhibited a diffuse growth pattern in most cases (n = 11), both in benign (n = 6) and malignant (n = 4) tumors. Although high (G3-4) nuclear grade was seen in 5 of 7 malignant tumors and in the single indeterminated for malignancy tumor, it could be documented also in 2 benign neoplasms. Fibrous septa were seen in a minority of cases (5 cases), including both, benign (n = 2) and malignant (n = 3) cases. Notably, focal (<5% of tumor area) regressive myxoid stromal changes were seen in 1 or 2 slides in all but 1 malignant case and in 2 benign tumors (Fig. 2).
3.3. Immunohistochemical findings
Immunohistochemical results are summarized in Table 3. MMP2 was analyzed only in 14 of 20 cases, including
8 tumors classified as benign, 5 tumors classified as malignant, and 1 tumor as indeterminate for malignancy (Table 3). MMP2 was focally to diffusely expressed in most benign (6 of 8) and in all malignant tumors (Table 3) (Fig. 3A and B). The single case of tumor classified indeterminate for malignancy showed only a focal expression of this marker. Endothelial cells of intratumoral vascular network and the vascular component of the surrounding normal adrenal tissue strongly expressed MMP2 and served as internal control.
With regard to expression of HLA class II antigens, immunoreactivity was virtually lacking, being absent in all benign tumors (Fig. 3C) and restricted to rare isolated cells in most (5 of 7 cases) malignant tumors. Notably, in 5 of 8 benign tumors and in 4 of 7 malignant tumors, multiple foci of positive stromal cells were identified. These cells nicely decorated a trabecular network (Fig. 3D) similar to that seen in the adjacent normal adrenal cortex. No or only scattered positive stromal cells were found in the remaining 7 benign tumors, in the single case of tumor classified indeterminate for malignancy and in the 3 malignant tumors (Table 3).
A
B
4. Discussion
ACTs in children are rare and, interestingly, have a less aggressive clinical behavior when compared with their adult counterpart [8]. Their prognostic stratification is still challenging due to the difficult distinction between benign and malignant lesions, based solely on pathologic criteria [4-6,8]. Previous studies have emphasized the unpredictable behavior of these neoplasms, prompting researchers to identify potential morphological and biological prognostic markers. In the largest series of these tumors reported in the literature, the pathologic criteria proposed by Wieneke et al [8] were shown to be a useful tool in predicting the clinical behavior of such tumors. Although some adverse prognostic parameters (capsular and/or vascular invasion) are the same in both the grading systems of Weiss [15] and Wieneke et al, others are the same but have different cutoff values (eg, mitoses); and still others are different (eg, weight and size do not enter the system of Weiss; growth pattern and percentage of clear cells do not enter that of Wieneke et al).
To the best of our knowledge, there are no studies about the reliability of the scoring system of Wieneke et al in pediatric ACTs since it was originally proposed [8]. We applied this scoring system in 20 patients enrolled into the TREP Study between 2000 and 2007. In our series, 12 tumors were classified as benign, 7 tumors as malignant, and 1 tumor as indeterminate for malignancy, accordingly. Histologically, most of the ACTs, irrespective of being classified as benign or malignant, were composed of polygonal eosinophilic cells arranged in a predominant diffuse growth pattern. In the present series, the scoring system of Wieneke et al was shown to be highly predictive of the clinical outcome. This is confirmed by the evidence that all 12 patients with tumors classified as benign are in complete remission without relapsing, whereas 4 of 7 children affected by tumors classified as malignant died as a result of the disease, 2 are alive with metastases and only 1 is in complete remission. The unique patient of our series with a tumor classified as indeterminate for malignancy is having a benign clinical course.
Interestingly, capsular invasion and necrosis, which are both well-defined unfavorable criteria in the grading system of Weiss for adult ACTs, were concurrently present in almost all cases of our malignant tumors (7 of 8). Conversely, the absence of a clear cell component in most ACTs in the present series confirms the poor relevance of this parameter in pediatric patients as compared with adults [8,15]. Although tumor
| Table 3 Immunohistochemical results of MMP2 and HLA class II antigens in pediatric ACTs | |||||
|---|---|---|---|---|---|
| Case | MMP2 | HLA | Stromal cells | Wieneke category | Follow-up |
| Tumor cells | |||||
| 1 | 1 | 0 | 2 | Benign | CR |
| 2 | 0 | 0 | 1 | Benign | CR |
| 3 | 1 | 0 | 2 | Benign | CR |
| 4 | 2 | 0 | 1 | Benign | CR |
| 5 | 2 | 0 | 2 | Benign | CR |
| 6 | 2 | 0 | 2 | Benign | CR |
| 7 | 0 | 0 | 2 | Benign | CR |
| 8 | 1 | 0 | 1 | Benign | CR |
| 9 | NP | 0 | 1 | Benign | CR |
| 10 | NP | 0 | 1 | Benign | CR |
| 11 | NP | 0 | 0 | Benign | CR |
| 12 | NP | 0 | 1 | Benign | CR |
| 13 | NP | 0 | 2 | Malignant | CR |
| 14 | NP | 1 | 2 | Malignant | DOD |
| 15 | 1 | 1 | 1 | Malignant | DOD |
| 16 | 2 | 0 | 2 | Malignant | DOD |
| 17 | 1 | 1 | 2 | Malignant | AWD |
| 18 | 2 | 1 | 1 | Malignant | AWD |
| 19 | 2 | 1 | 1 | Malignant | DOD |
| 20 | 1 | 1 | 1 | Indeterminate | CR |
MMP2: 0, negative staining; 1, focal staining (≤20% of neoplastic cells); 2, diffuse staining (>20% of neoplastic cells). HLA class II antigens: 0, negative staining; 1, rare positive cells; 2, multiple foci of positive cells. Abbreviation: NP indicates not performed.
A
C
B
D
weight (>400 g) has been shown to be prognostically relevant in pediatric ACTs [8,16], we failed to confirm these results. In fact the weight of most malignant tumors (5 of 8) in our series ranged from 25 to 170 g, with only 3 cases that weighed more than 1000 g. Notably, among the morphological features assessed in this study and not including in the scoring system of Wieneke et al, the detection of focal myxoid stromal changes seems to be potentially promising in identifying malignant tumors. This finding could be seen in all but 1 malignant tumor and only in 2 of 12 benign tumors. However, whether the presence of myxoid stromal changes is an indicative of malignancy needs to be confirmed in a larger tumor series. In this regard, recently, Papotti et al [17] reported that a variable amount of myxoid stromal changes can be identified most frequently in adult malignant or borderline versus benign ACTs. Interestingly, focal myxoid changes, similar to those observed in our series, were associated with tumors that were otherwise reminiscent of malignancy, exhibiting large atypical eosinophilic cells as well as a diffuse growth pattern [17]. On the contrary, the presence of abundant myxoid component was
associated with tumors composed of small cells with only mild atypia and a predominant trabecular or microcystic growth pattern [17]. These latter findings were not encountered in our series. It is likely that the focal myxoid stromal changes found in adult tumors [17] and also in the present series are merely the result of focal intratumoral degenerative changes to which malignant neoplasms usually undergo [17].
Among the molecules involved in carcinogenesis, MMPs, a family of at least 26 members, have been investigated as promising prognostic biomarkers in several human carcino- mas [18]. It is commonly accepted that their overexpression may be crucial in the development of metastases through their proteolytic activity on extracellular matrix components, including collagen, laminin, elastin, and fibronectin [18,19]. Several studies have documented a correlation between MMPs up-regulation and occurrence of metastases in breast [20,21], colon [22,23], and lung [24]. Volante et al [13] investigated the immunohistochemical expression of MMP2 and MMP9 and their inhibitors in a series of adult ACTs, including 50 carcinomas and 50 adenomas classified
according to the scoring system of Weiss [15]. The most interesting finding was a restricted MMP2 expression to malignant tumors, with a high specificity (98.8%) but low sensitivity (74%). Furthermore, MMP2 expression in more than 20% of neoplastic cells correlated with a more aggressive clinical outcome [13]. Unfortunately, MMP2 was focally to diffusely expressed in both benign and malignant tumors of the present series. Accordingly, MMP2 does not appear to be a reliable marker of malignancy in pediatric ACTs.
With regard to HLA class II antigens, they are expressed by steroid-producing cells of the innermost adrenocortical zone (zona reticularis) of normal human gland. In fetal life and in the first 4 years of life, adrenocortical cells do not express HLA class II antigens [14,25]. It is not yet known when these antigens appear in the adrenal cortex, but they are probably involved in cellular differentiation. Loss of expression of HLA class antigens or deregulation has been reported in a great variety of tumors, including breast, cervix, larynx, colon, and pancreas cancers [26-32]. In a study of gene expression by microarray analysis, carried out on 24 pediatric ACTs (5 adenomas, 18 carcinomas, 1 tumor classified as indeterminate for malignancy), West et al [12] identified 52 probe sets for which significant differences were found when comparing adrenocortical adenomas versus carcinomas. Within this set, there was a marked decrease in the expression of major histocompatibility class II genes. Specifically, the median expressions of HLA-DRB1, HLA- DPB1, HLA-DRA, and HLA-DPA1 messenger RNA levels were 6- to 8-fold lower in adrenocortical carcinomas as compared with adenomas [12]. Such data seem to be supported by immunohistochemical studies on adult ACTs, showing the maintenance of class II antigens in majority of adrenocortical adenomas in contrast to their loss in carcinomas [14]. Accordingly, the immunohistochemical detection of HLA class II antigens in tumor cells has been proposed as a marker helpful in ruling out malignancy [14]. In our series, both benign and malignant neoplastic cells failed to express HLA class II antigens, with the exception of rare isolated tumor cells in some malignant tumors and in the single indeterminate for malignancy tumor. Interestingly, immunohistochemical analysis with HLA class II antigens displayed a subtle network of positive stromal cells in both benign (5 cases) and malignant (4 cases) tumors as similarly observed in the adjacent normal cortex. Notably, the absence of HLA class II antigens in pediatric ACTs may be explained by speculating their different histogenesis, probably origi- nating from embryonal cells, which are typically HLA class II antigen negative.
In conclusion, our study confirms that the scoring system of Wieneke et al is a simple and reproducible diagnostic tool to predict prognosis in pediatric ACTs. In addition, we found that malignant pediatric ACTs, in contrast to benign tumors, develop more frequently focal myxoid stromal changes. This promising morphological finding, if con- firmed in a larger series, could be included in the list of the
criteria of malignancy. Unlike in their adult counterparts, the expression of MMP2 or the loss of HLA-class II antigens does not discriminate between malignant and benign tumors. These findings further support the idea that although some childhood tumors seem to be similar histo- logically to their adult counterparts, they have distinctive molecular features [33].
References
[1] Mendoca BB, Lucon AM, Menezes CA, et al. Clinical hormonal and pathological findings in a comparative study of adrenocortical neoplasms in childhood and adulthood. J Urol 1995;154:2004-9.
[2] Sandrini R, Ribeiro RC, DeLacerda L. Childhood adrenocortical tumors. J Clin Endocrinol Metab 1997;82:2027-31.
[3] Pianovski MA, Maluf EM, de Carvalho DS, et al. Mortality rate of adrenocortical tumors in children under 15 years of age in Curitiba, Brazil. Pediatr Blood Cancer 2006;47:56-60.
[4] Humphrey GB, Pysher T, Holcombe JH, et al. Overview on the management of adrenocortical carcinoma. Cancer Treat Res 1983;17: 348-58.
[5] Stratakis A, Chrousos GP. Endocrine tumors. In: Pizzo A, Poplack D, editors. Principles and practice of pediatric oncology. Chapter 35, 3rd ed. Lippincott; 1997. p. 960-3.
[6] Plowman PN. Rare tumors. In: Pinkerton C, Plowman P, editors. Pediatric oncology clinical practice and controversies. Chapter 22, 2nd ed. Chapman & Hall; 1997. p. 569-70.
[7] Ribeiro RC, Sandrini Neto RS, Shell MJ, Lacerda I, Sambaio GA, Cat I. Adrenocortical carcinoma in children: a study of 40 cases. J Clin Oncol 1990;8:67-74.
[8] Wieneke JA, Thompson LDR, Heffess CS. Adrenal cortical neoplasms in the pediatric population: a clinicopathologic and immunophenotypic analysis of 83 patients. Am J Surg Pathol 2003;27:867-81.
[9] Sbragia L, Oliveira-Filho AG, Vassallo J, Pinto GA, Guerra-Junior G, Bustorff-Silva J. Adrenocortical tumors in Brazilian children: immunohistochemical markers and prognostic markers. Arch Pathol Lab Med 2005;129:1127-31.
[10] Michalkiewicz E, Sandrini R, Figueiredo B, et al. Clinical and outcome characteristics of children with adrenocortical tumors: a report from the International Pediatric Adrenocortical Tumor Registry. J Clin Oncol 2004;22:838-45.
[11] Orhan D, Kale G, Çağlar M, Göğüs S, Karaağaoğlu E. Histone mRNA in situ hybridization and Ki immunohistochemistry in pediatric adrenocortical tumors. Virchows Arch 2006;448:591-6.
[12] West AN, Neale GA, Pounds S, et al. Gene expression profile of childhood adrenocortical tumors. Cancer Res 2007;67:600-8.
[13] Volante M, Sperone P, Bollito E, et al. Matrix metalloproteinase type 2 expression in malignant adrenocortical tumor: diagnostic and prognostic significance in a series of 50 adrenocortical carcinomas. Mod Pathol 2006;19:1563-9.
[14] Marx C, Wolkersdörfer GW, Brown JW, Scherbaum WA, Bornstein SR. MHC class II expression-a new tool to assess dignity in adrenocortical tumours. J Clin Endocrinol Metab 1996;81:4488-91.
[15] Weiss LM. Comparative histologic study of 43 metastasizing and nonmetastasizing adrenocortical tumors. Am J Surg Pathol 1984;8: 163-9.
[16] Dehner LP, Hill DA. Adrenal cortical neoplasms in children: why so many carcinomas and yet so many survivors? Pediatr Dev Pathol 2009;12:284-91.
[17] Papotti M, Volante M, Duregon E, et al. Adrenocortical tumors with myxoid features: a distinct morphologic and phenotypical variant exhibiting malignant behaviour. Am J Surg Pathol 2010;34: 973-83.
[18] Egeblad M, Werb Z. New function for the matrix metalloproteinases in cancer progression. Nat Rev Cancer 2002;2:161-74.
[19] Deryugina EI, Quigley JP. Matrix metalloproteinases and tumor metastasis. Cancer Metastasis Rev 2006;25:9-34.
[20] Wolf C, Rovyer N. Stromelysin-3 belongs to a subgroup of proteinases expressed in breast carcinoma fibroblastic cells and possibly implicated in tumour progression. Proc Natl Acad Sci USA 1993;90: 1843-7.
[21] Freije JM, Díez-Itza I, Balbín M, et al. Molecular cloning and expression of collagenase-3, a novel human matrix metalloproteinase produced by breast carcinomas. J Biol Chem 1994;269:16766-73.
[22] Witty JP, McDonnell S, Newell KJ, et al. Modulation of matrilysin levels in colon carcinoma cell lines affects tumorigenicity in vivo. Cancer Res 1994;54:4805-12.
[23] Zeng ZS, Shu WP, Cohen AM, Guillem JG. Matrix metalloproteinase- 7 expression in colorectal cancer liver metastases: evidence for involvement of MMP-7 activation in human cancer metastases. Clin Cancer Res 2002;8:144-8.
[24] Muller D, Breathnach R, Engelmann A, et al. Expression of collagenase-related metalloproteinase genes in human lung or head and neck tumours. Int J Cancer 1991;48:550-6.
[25] Marx C, Bornstein SR, Wolkersdörfer GW, Peter M, Sippell WG, Scherbaum WA. Relevance of major histocompatibility complex class II expression as a hallmark for the cellular differentiation in the human adrenal cortex. J Clin Endocrinol Metab 1997;82:3136-40.
[26] Redondo M, García J, Villar E, et al. Major histocompatibility complex status in breast carcinogenesis and relationship to apoptosis. HUM PATHOL 2003;34:1283-9.
[27] Tamiolakis D, Venizelos I, Lambropoulou M, et al. Gains and losses of HLA class II (DR) and CD4 in atypical hyperplasia, carcinoma in situ and infiltrating ductal carcinoma of the breast. Acta Medica (Hradec Kralove) 2004;47:257-62.
[28] McDougall CJ, Ngoi SS, Goldman IS, et al. Reduced expression of HLA class I and II antigens in colon cancer. Cancer Res 1990;50:8023-7.
[29] Benevolo M, Mottolese M, Piperno G, et al. HLA-A, -B, -C expression in colon carcinoma mimics that of the normal colonic mucosa and is prognostically relevant. Am J Surg Pathol 2007;31:76-84.
[30] Esteban F, Ruiz-Cabello F, Concha A, Pérez-Ayala M, Sánchez-Rozas JA, Garrido F. HLA-DR expression is associated with excellent prognosis in squamous cell carcinoma of the larynx. Clin Exp Metastasis 1990;8:319-28.
[31] Scupoli MT, Sartoris S, Tosi G, et al. Expression of MHC class I and class II antigens in pancreatic adenocarcinomas. Tissue Antigens 1996;48(4 Pt 1):301-11.
[32] Hernández-Hernández DM, Cerda-Flores RM, Juárez-Cedillo T, et al. Human leukocyte antigens I and II haplotypes associated with human papillomavirus 16-positive invasive cervical cancer in Mexican women. Int J Gynecol Cancer 2009;19:1099-106.
[33] Nakamura M, Shimada K, Ishida E, et al. Molecular pathogenesis of pediatric astrocytic tumors. Neuro Oncol 2007;9:113-23.