Oncocytic glioblastoma: a glioblastoma showing oncocytic changes and increased mitochondrial DNA copy number

Human Pathology (2013) 44, 1867–1876

www.elsevier.com/locate/humpath

Original contribution

Oncocytic glioblastoma: a glioblastoma showing oncocytic changes and increased mitochondrial DNA copy number☆,☆☆ Gianluca Marucci MD, PhD a , Alessandra Maresca PhD b,c , Leonardo Caporali BSc b,c , Anna Farnedi BSc, PhD a , Christine Margaret Betts BSc, MD d , Luca Morandi BSc, PhD a , Dario de Biase BSc, PhD a , Serenella Cerasoli MD e , Maria Pia Foschini MD a , Elena Bonora BSc, PhD f , Michele Vidone PhD f , Giovanni Romeo MD f , Elena Perli PhD g , Carla Giordano MD, PhD g , Giulia d'Amati MD, PhD g , Giuseppe Gasparre PhD f , Agostino Baruzzi MD b,c , Valerio Carelli MD, PhD b,c , Vincenzo Eusebi MD, FRCPath a,⁎ a

Department of Biomedical and NeuroMotor Sciences (DiBiNeM), University of Bologna, Section of Pathology, “M. Malpighi”, Bellaria Hospital, via Altura 3, 40139 Bologna, Italy b IRCCS “Istituto delle Scienze Neurologiche”, via Ugo Foscolo 7, 40123 Bologna, Italy c Department of Biomedical and NeuroMotor Sciences (DiBiNeM), University of Bologna, Section of Neurology, Neurosurgery and Infantile Neuropsychiatry, via Ugo Foscolo 7, 40123 Bologna, Italy d Dip. Medicina Specialistica, Diagnostica e Sperimentale (DISMES), University of Bologna, Via S.Giacomo 14, 40126 Bologna, Italy e Section of Anatomic Pathology, Bufalini Hospital, V. le Ghirotti 286, 47521 Cesena, Italy f Department of Medical and Surgical Sciences (DIMES), University of Bologna, via Massarenti 9, 40138 Bologna, Italy g Department of Internal Medicine and Medical Specialties, “La Sapienza” University, via del Policlinico 155, 00161 Roma, Italy Received 21 December 2012; revised 23 February 2013; accepted 28 February 2013

Keywords: Glioblastoma; mtDNA; Oncocytic features

Summary Ten cases of glioblastomas showing oncocytic changes are described. The tumors showed mononuclear to multinuclear cells and abundant, granular, eosinophilic cytoplasm. The cytoplasm of these same cells was filled by strongly immunoreactive mitochondria. At ultrastructure, numerous mitochondria, some of which were large, were evidenced in the cytoplasm of neoplastic cells. Finally, 9 of 10 of these cases had a significantly high mitochondrial DNA content compared with control tissue (P b .01). It seems that, for these tumors, the designation of oncocytic glioblastoma is appropriate. To the best of our knowledge, oncocytic changes have not been previously reported in such neoplasms. Oncocytic glioblastomas have to be added to the long list of various tumors that can manifest “unexpected” oncocytic changes in different organs. Albeit failing to show statistical significance (log-

☆

Funding: This study was funded by Progetto Emiliano-Romagnolo di Neuro-Oncologia (Emilia-Romagna region project). The work was partly supported by Associazione Italiana Ricerca sul Cancro (IG8810) to G. R., by the Italian Ministry of University FIRB Futuro in Ricerca ‘TRANSMIT’ and by Fondazione Umberto Veronesi ‘DISCO-TRIP’ to G. G.; M. V. was supported by Associazione Italiana Tumori Cerebrali. ⁎ Corresponding author. Sezione di Anatomia Patologica “M. Malpighi”, Ospedale Bellaria, 40139 Bologna, Italy. E-mail address: [email protected] (V. Eusebi). ☆☆

0046-8177/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.humpath.2013.02.014

1868

G. Marucci et al. rank test, P = .597; Wilcoxon test, P = .233), we observed a trend for longer median survival in oncocytic glioblastomas, when compared with “ordinary” glioblastomas (median survival of 16 versus 8.7 months). Thus, it seems that the definition of neoplasms showing oncocytic changes, currently based on classic morphological parameters (ie, histology, ultrastructure, and immunohistochemistry), can be expanded by including the quantitative assessment of mitochondrial DNA content. © 2013 Elsevier Inc. All rights reserved.

1. Introduction Glioblastoma multiforme (GBM) is a malignant astrocytic primary tumor of the brain [1] with poor outcome [2,3] to such extent that patients who survive more than 36 months are referred to as long-term survivors. In the last few years, a new standard treatment regimen that combines the widest surgical tumor resection, temozolomide (TMZ) and local field radiotherapy (RT), has been found to improve median overall survival from 12.1 to 14.6 months [4]. Markedly atypical astrocytic tumor cells, numerous mitoses, prominent microvascular proliferation, and necrosis are the features that characterize the garden variety of GBMs. Several histologic subtypes of GBMs have been reported that include small cell, giant cell, rhabdoid, lipid-rich epithelioid, and oligodendroglial cell rich [1] as well as gliosarcoma. None of these different patterns have consistent prognostic impact. The present article aims to describe a series of GBMs showing cells with oncocytic (OC) features, a special/additional type of cell differentiation that might have prognostic value. Oncocyte means “swollen cell” from ancient Greek “Ογχοΰσθαι.” The term was coined in salivary glands by Hamperl [5] in 1931 to describe a distinct type of large epithelial cell showing deeply acidophilic granular cytoplasm. Electron microscopy (EM) subsequently evidenced that OC cells were enlarged due to an increase in the number and size of mitochondria, which also imparted the cytoplasmic granularity [6]. Oncocytes were distinguished from mitochondrion-rich cells in thyroid follicles because their numerous mitochondria were evenly distributed throughout the cytoplasm, whereas in mitochondrion-rich cells, mitochondria were mainly localized at the poles of follicular cells [7]. According to Ghadially [8], 60% of the total cytoplasmic volume loaded by mitochondria has become the generally accepted cutoff value for the designation of oncocyte. Oncocytomas are considered those neoplasms that contain 70% or more of immunohistochemically proven oncocytes [9]. Mitochondria of oncocytes appear to be nonfunctional, and their aberrant proliferation is probably due to a compensatory effect triggered by a bioenergetic deficit [10]. Oncocytomas have been observed in numerous sites of the body, although thyroid, kidney, and breast are the most frequent. OC changes have been reported in some primary intracranial tumors, such as pituitary adenomas [11], meningiomas [12], choroid plexus carcinomas [13], ependymomas [14], and subependymal giant cell astrocytomas [15]. Mammalian cells contain numerous mitochondria, and the mitochondrial genome (mitochondrial DNA [mtDNA]),

a double-stranded circular molecule of 16 569 base pairs, is present in multiple copies that vary in number according to the cell type [16]. mtDNA is very susceptible to reactive oxygen species damage, and the mutation rate of mtDNA is consistently higher than that seen in genomic DNA [11]. Qualitative [17] and quantitative [18] mtDNA alterations have been described in neurodegenerative, metabolic diseases and some types of cancer [19–21], including tumors showing OC features [22]. Furthermore, mtDNA copy number correlates with the number and size of mitochondria [23]. The mitochondrial genome may hence play an important role in cellular sensitivity to cancer treatment [24], and mitochondrial metabolism might be related to chemoresistance in gliomas [25,26]. Analyses of mtDNA copy number and of mtDNA mutations have already been performed in low-grade gliomas [27], whereas, to the best of our knowledge, mtDNA content has not been evaluated in a population of patients with GBM. In the present study, mtDNA copy number was analyzed in a series of GBMs to evidence any correlation with morphology and survival. We identified 10 cases of GBM where most of the cells showed typical OC features at histologic, immunohistochemical, and ultrastructural levels. The OC cells also contained, with exception of 1 case, a significant increase of mtDNA copy number. We defined these tumors as OC-GBMs. These same tumors showed a trend for longer median survival when compared with non–OC-GBMs.

2. Materials and methods 2.1. Ethics statement The study was approved by the Ethics Committee of Azienda Sanitaria Locale di Bologna (no. of study 08075, protocol number 139/CE of February 5, 2009, Bologna, Italy). All patients signed a written consent for molecular analysis and for anonymous data publication in scientific studies, and all information regarding the human material used in this study was managed adopting anonymous numerical codes.

2.2. Patients Ten cases of GBM with histologic OC features were collected from the files of the Section of Pathology of the University of Bologna at Bellaria Hospital (Bologna, Italy) and of the Section of Pathology of the Bufalini Hospital (Cesena, Italy). These cases are defined propositus cases. For

Oncocytic glioblastoma comparison, 18 cases of clinically diagnosed primary GBMs were consecutively recruited. Patients had not undergone neoadjuvant therapy before surgery. All the above patients had been enlisted within the Progetto Emiliano-Romagnolo di Neuro-Oncologia project where all cases of brain tumors from the region are registered. All specimens from each case had been immediately snap frozen, and a cryostat section was evaluated by a pathologist (G. M. or S. C.) to evidence presence of “high-grade glioma.” From each case, a portion of tissue was stored at −80°C, whereas the remaining tissue was formalin fixed and paraffin embedded for routine histologic diagnosis. Hematoxylin and eosin (H&E) sections were reviewed by 2 pathologists (G. M. and V. E.) and were all confirmed to be GBMs according the 2007 World Health Organization criteria [1].

2.3. Immunohistochemistry Serial, 3-μm-thick, paraffin sections mounted on precoated slides were processed by standardized automated procedures using prediluted antibodies glial fibrillary acidic protein (GFAP), cytokeratin CAM5.2, Mart-1, and Ki-67; Ventana-Benchmark, Tucson, AZ), with the exception of the antimitochondrion antibody (Ab-2, clone MTC02; LabVision, Freemont, CA). For the latter, immunohistochemistry was manually performed using a polymer as detection system (Ultravision LP Detection System HRP Polymer; LabVision), and the optimal dilution of the antimitochondrion antibody (1:2000) had been established in a previous study of OC invasive breast carcinomas [9]. In brief, dewaxing/deparaffinization and antigen retrieval were achieved by pretreatment with W-CAP TEC buffer solution pH 8 (Bio-Optica, Milan, Italy) at 98°C for 25 minutes, and inhibition of the endogenous peroxidases was obtained using 3% H2O2. After rinsing the slides in bidistilled water and in phosphate-buffered saline solution, the sections were incubated in a humid chamber at room temperature for 5 minutes with Ultra V Block solution (Ultravision LP; Thermo Fisher Scientific, Fremont, CA). The slides were then incubated with antimitochondrion primary antibody for 1 hour at room temperature in a humid chamber. The slides were then rinsed in buffer solution and incubated with a primary antibody enhancer solution (Ultravision LP; Thermo Fisher Scientific) for 20 minutes at room temperature in a humid chamber. Following several washes in buffer solution, the sections were finally incubated for 30 minutes with horseradish peroxidase polymer solution (Ultravision LP; Thermo Fisher Scientific). The reaction was revealed with 3,3′-diaminobenzidine solution for 3 minutes. The immunopositivity for mitochondria was scored using the semiquantitative scale previously used for OC breast carcinomas as 0 (no reactivity), 1+ (b10% positive neoplastic cells), 2+ (11%-69% positive), and 3+ (≥70% positive) [9]. For the statistical analysis, scores 0 and 1+ were considered negative, score 2+ indicated mitochondrion-rich tumors, whereas score 3+ indicated tumors showing OC changes [9].

1869 In 2 selected cases (cases 1 and 5), sections were simultaneously immunostained for mitochondria and GFAP to investigate whether cells with 3+ mitochondria score (brown) also stained for GFAP (red). The double staining protocol required 2 steps, the first manual and the second in a Ventana-Benchmark automatic stainer. Staining and scores were assessed independently by 2 observers (G. M. and V. E.). When discrepancies occurred, the case was discussed on a multihead microscope, and a consensus was reached.

2.4. Ultrastructural study Ultrastructural analysis was performed on 2 cases (cases 1 and 5) that had 3+ score for mitochondria. Small samples were microdissected from paraffin-embedded blocks, placed in buffered glutaraldehyde, postfixed in OsO4, and embedded in Epon 812 (Fort Washington, PA). Thin sections were stained with uranyl acetate and Reynold's lead citrate and observed in a Philips CM 10 transmission EM (Eindhoven, the Netherlands). 2.4.1. O6-methylguanine DNA-methyltransferase methylation status assessment Formalin-fixed and paraffin-embedded tissue blocks were selected for DNA extraction from all 28 cases. Tumor material was manually dissected under H&E guidance from the corresponding 10-μm sections, to ensure greater than 90% content of neoplastic cells. O6-methylguanine DNA-methyltransferase (MGMT) methylation status was assessed by the methylation-sensitive quantitative locked nucleic acid polymerase chain reaction (PCR) as previously described [28]. 2.4.2. Isocitrate dehydrogenase 1 mutation analysis Isocitrate dehydrogenase 1 (IDH1) analysis for the p.R132H mutation was performed on paraffin sections, using allele-specific locked nucleic acid quantitative PCR, according to the previously described protocol [29] from 22 cases that included all propositus cases, as material was not available in 6 cases. The forward primers used for the analysis were designed for the wild-type allele (5′-TTGATCCCCATAAGCATGA [+C]-3′) and for the p.R132H allele (5′-TTGATCCCCATAAGCATGA[+T]-3′) (LNA bases are preceded by +); the reverse primer (5′- GTGGCACGGTCTTCAGAGA-3′) allowed amplification of a segment 104 base pairs long. The allele-specific locked nucleic acid quantitative PCR IDH1 results were confirmed by sequencing using the 454 GS-Junior Next Generation sequencer (Roche, Branford, CT).

2.5. mtDNA quantification DNA was extracted from frozen tissues using High Pure PCR Template Preparation kit (Roche Applied Science, Mannheim, Germany) according to the manufacturer's instructions and quantified by QuantiT TM dsDNA BR kit (Invitrogen, Carlsbad, CA) from all 28 samples.

1870 Of the DNA, 10 to 20 ng was used for the mtDNA content evaluation by a real-time PCR assay based on hydrolysis probe chemistry. Briefly, an mtDNA fragment (MTND2 gene) and a nuclear DNA fragment (FASLG gene) were coamplified by multiplex PCR, and their concentration was determined by absolute quantification. Primers, probes, and conditions have been previously published [30].

2.6. Control cases To verify the incidence of morphological and immunohistochemical OC features in a case series, 50 consecutive paraffin-embedded GBMs from the surgical files of the Section of Pathology at Bellaria Hospital were studied. These cases were manually immunostained for mitochondrion antibody and scored according to the procedure described above. To compare the mtDNA quantification in “normal” tissues, we used a sex- and age-matched control group of patients from which 10 samples of substantia nigra and 18 samples of frontoparietal cortical gray matter had been collected postmortem. Tissue had been immediately frozen at −80°C, within 24 hours from death. In all individuals, any central nervous system disease was clinically and pathologically excluded as found at postmortem.

2.7. Statistical analysis Overall survival from surgery was calculated by GraphPad Prism v.5 tool (GraphPad Software Inc., La Jolla, CA, USA). Kaplan-Maier curves were compared using log-rank and Wilcoxon tests. P b .05 was considered statistically significant. Features of the 2 groups (OC-GBM and non– OC-GBM) were compared using Fisher exact test (P b .05). Mean age was compared using the Student t test (P b .05). All P values were 2 tailed; 95% confidence intervals (CIs) were adopted. mtDNA copy number was compared using the 1-way analysis of variance on ranks test, and P b .05 was considered statistically significant.

3. Results 3.1. Clinical features The clinical features of the 28 cases (including 10 propositus) are summarized in Table 1. The age ranged from 41 to 75 years (mean, 60.1 years) at time of surgery. Fifteen patients (54%) were female, and 13 (46%) were male. The age at time of surgery of the propositus cases was not different from that of all other patients, including the 50 consecutive paraffin-embedded cases. Accordingly, it ranged from 39 to 76 years, with mean of 60.5 years.

G. Marucci et al.

3.2. Pathologic features (histology and immunohistochemistry) All 28 cases (including 10 propositus) were composed of cells showing marked nuclear atypia. Mitoses were numerous, and necrotic spots as well as microvascular proliferation were prominent. In all tumors, numerous neoplastic elements were positive for GFAP and Ki-67, the latter ranging from 10% to 35% in nuclear positivity. On the contrary, all cases were negative for cytokeratin CAM5.2 and Mart-1. The 10 propositus cases (case nos. 1-10), in addition to the above features, displayed 4 easily recognizable types of cells consisting of: 1. numerous mononuclear elements with abundant eosinophilic granular cytoplasm (Figs. 1A and 2). 2. multinucleated cells with large eosinophilic granular cytoplasm (Fig. 1A and B) observed in no less than 6 cases. These latter cells ranged from 3% to 5% in different cases and showed irregular nuclei that ranged from 2 to 4 in number. Nuclei appeared peripherally located close to the cell membrane. Occasional mononuclear or multinucleated cells showed ground glass cytoplasm reminiscent of gemistocytes. 3. the cytoplasms of a minority of both mononuclear and multinucleated cells were strongly eosinophilic and had a targetoid appearance with a central glassy core surrounded by a ring of coarse granules (Fig. 2). 4. cells with foamy clear cytoplasm suggestive of a degranulating process. These were seen in approximately 5% of the total neoplastic proliferation (Fig. 3). All the cells with abundant granular eosinophilic cytoplasm, both of mononuclear or multinuclear variety, showed 3+ strong mitochondrial immunohistochemical positivity in all 10 cases (Fig. 1B). The number of elements with 3+ cytoplasmic positivity for mitochondria ranged from 70% to 90% of the tissue examined. The cells with targetoid features showed most of their mitochondria condensed at the periphery of the cytoplasm, whereas in the center, a negative spot was observed (Fig.4). These features were better evidenced in cases 1 and 5, which were simultaneously stained for mitochondria and GFAP, which showed a heterogeneous cell population. Numerous cells with abundant cytoplasm were exclusively positive for mitochondria. Smaller cells contained GFAP only. Finally, several neoplastic elements were positive for both GFAP and mitochondria at the same time (Fig. 5). The cells that displayed targetoid features presented a central cytoplasmic core immunoreactive for GFAP, surrounded by a crown of mitochondria (Fig. 4). On the contrary, cells located in adjacent areas, not showing eosinophilic cytoplasmic granularity at H&E level, displayed at the most 2+ positivity. The 18 cases of GBM not composed by a significant number of cells with strong eosinophilic granular cytoplasm

Oncocytic glioblastoma Table 1 Case a

1 2 3 4 5a 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

1871

Clinical-pathologic and biomolecular features of enrolled GBM Age

Sex

H&E

IHC

H/L mtDNA

MGMT

IDH1

OS (mo)

CH

RT

41 74 68 69 61 72 63 60 72 77 50 75 75 53 51 49 68 60 45 78 70 65 75 55 58 41 59 68

F F F M F M M M F M F F F M M F F M M M F F M F M M F F

OC OC OC OC OC OC OC OC OC OC Not Not Not Not Not Not Not Not Not Not Not Not Not Not Not Not Not Not

3+ 3+ 3+ 3+ 3+ 3+ 3+ 3+ 3+ 3+ 1+ 1+ 2+ 2+ 2+ 2+ 1+ 1+ 1+ 1+ 1+ 2+ 1+ 2+ 1+ 1+ 2+ 1+

H H H H H H H H H L H H L L L L L L L L L L L L L L L L

Met U U U U U U Met Met Met U Met Met Met U U U U U U Met U U Met U Met U Met

W W W W W W W W W W W W W NA W W W W W NA W W W W NA NA NA NA

23.1 7.2 Alive 13.8 3.9 16.8 Alive 16 Alive 4.3 4.1 15.8 8.8 2.4 8.7 Alive 7.1 5.1 Alive 1.8 7.3 6.9 Alive Alive Alive 16.3 25.2 2.9

Y N N N Y Y Y Y N N N N N Y Y N Y N Y N Y N Y Y Y Y Y Y

Y N N N Y Y Y Y N N Y N Y Y Y N Y N Y N Y N Y Y Y Y Y Y

OC OC OC OC OC OC OC OC OC OC OC OC OC OC OC OC OC OC

12.4 mo after surgery

17.7 mo after surgery 14.7 mo after surgery

17.4 mo after surgery

20.1 mo after surgery

13.1 mo after surgery 27.3 mo after surgery 21.9 mo after surgery

Abbreviations: H&E, hematoxylin and eosin; OS, overall survival; RT, radiotheraphy; F, female; M, male; IHC, immunohistochemistry; H, high; L, low; Met, methylated; NA, material not available; U, unmethylated; CH, chemotherapy; Y, yes; N, no; W, wild type. a Analyzed with EM and double immunohistochemistry.

were weakly positive (1+) for mitochondria in 11 cases (61.1%) and moderately positive (2+) in 7 cases (38.9%).

3.3. Control cases In the consecutive routine series of 50 GBMs, histologic and immunohistochemical (3+) OC features were observed in 2 cases (4%).

3.4. Ultrastructural findings The 2 cases (1 and 5) investigated by EM displayed cells with prominent nuclei, occasionally multiple, with dusty scattered chromatin; 1 to 2 small nucleoli; and a neat, linear nuclear membrane. Numerous mitochondria were scattered throughout the cytoplasm, without evidence of polar condensation. Mitochondria varied in shape from round to ovoid, and their cristae were flattened to ballooned. No cytoplasmic secretory granules were observed. Rare lysosomes were apparent. The large gemistocytic-like elements often showed a central cytoplasmic core constituted by thin filaments and by numerous mitochondria (Fig. 6).

3.4.1. MGMT methylation status assessment and IDH1 mutation analysis The MGMT promoter was found to be methylated by methylation-sensitive quantitative locked nucleic acid PCR in 11 samples (39%) and unmethylated in 17 samples (61%). All analyzed cases (22/28) did not harbor IDH1 mutations.

3.5. mtDNA content The mtDNA copy number evaluation of the 28 GBM cases revealed a consistent association between immunohistochemical observations and the mtDNA content. In fact, 9 of the 10 propositus cases (case nos. 1-9), characterized by 3+ positivity for mitochondria, showed a mtDNA copy number significantly higher than that seen in 16 of 18 cases (case nos. 13-28), where immunohistochemical score varied from 1+ to 2+ (P = .009) as well as in the grey matter of control samples (P b .01) (Fig. 7A and B). Thus, immunohistochemical results appeared strongly consistent with mtDNA copy number (P = .0005). These 9 cases represent the OC-GBMs. In only 3 cases, the value of mtDNA copy number did not correspond to immunohistochemical features: case 10 had

1872

G. Marucci et al.

Fig. 3 Propositus case number 3 (OC-GBM): most of the cells have abundant foamy cytoplasm, reminiscent of degranulated elements.

low mtDNA content in spite of a high reactivity (3+) to the antimitochondria antibody; on the contrary, cases 11 and 12 had high mtDNA content, although they did not show strong reactivity to the antimitochondrion antibody.

3.6. Follow-up

Fig. 1 A, Propositus case number 8 (OC-GBM): numerous neoplastic elements with abundant eosinophilic cytoplasm and vascular proliferation. Two giant cells are indicated by arrows. B, Most of the cells of the same case show cytoplasms filled with powdery to coarse granules positive with antimitochondrion antibody. A giant cell is indicated by arrow.

Fig. 2 Propositus case number 7 (OC-GBM): neoplastic elements showing abundant, granular eosinophilic cytoplasm. Nuclei are irregular. Three cells show targetoid features (arrows).

The standard treatment regimen (TMZ concurrent with and adjuvant to RT) was administered in 16 patients (57%), whereas 2 patients (7%) were treated with RT alone, and the remaining 10 patients (36%) did not receive any adjuvant therapy.

Fig. 4 Propositus case number 2 (OC-GBM): most of the cells show granules stained for mitochondria in numerous cells in a ring-like fashion. Other cells show all cytoplasm loaded with immunoreactive mitochondria. The inset in the lower left corner shows 1 multinucleated cell (from case 1) evidencing ringlike immunoreactivity for mitochondria. The central portion of the cytoplasm is unstained. The inset in the upper right corner (from case 1) highlights a cell with targetoid features, where a central GFAP-positive cytoplasmic core is surrounded by a crown of mitochondria.

Oncocytic glioblastoma

1873 Although there was no statistically significant difference (log-rank test, P = .597; Wilcoxon test, P = .233) in survival rate between the 2 groups, a trend for longer median survival was observed in the OC-GBM patients (median survival of 16 versus 8.7 months). The median survival benefit was 7.3 months.

4. Discussion

Fig. 5 Propositus case number 5 (OC-GBM): GFAP is stained in red, and mitochondria in brown. The neoplasm shows a remarkable heterogeneity with cells that are purely oncocytes; a minority of small elements are uniquely GFAP positive, whereas numerous neoplastic cells display mitochondria and GFAP positivity at the same time.

The median survival of the 28 patients was 13 months (95% CI, 2.88-24.71). Clinical features of the OC versus the non-OC group side by side are summarized in Table 2. Neither statistically significant differences nor better trend were observed for age (t test, P = .548), MGMT status (Fisher exact test, P = 1.000), and treatment with TMZ + RT (Fisher exact test, P = .689). Survival curves for OC versus the non-OC group are reported in Fig. 8 from which it appears that the unadjusted hazard ratio for death in the 9 OC-GBMs as compared with the other 19 cases was 0.777 (95% CI, 0.306-1.975).

Fig. 6 EM, propositus case number 1 (OC-GBM): a cell showing several mitochondria of different size irregularly scattered through the cytoplasm. No secretory-like granules are visible.

In the present study we report, for the first time, the occurrence of OC changes in GBMs documented by histology, immunohistochemistry, ultrastructure, and high levels of mtDNA copy number. The 10 propositus cases (cases 1-10) described here were composed mostly of mononuclear to multinuclear neoplastic cells having abundant eosinophilic granular cytoplasm. Some cells were foamy and clear reminiscent of degranulated elements. Neoplastic cells showed cytoplasmic granules intensely and consistently positive for mitochondria at immunohistochemistry in 70% or more of the total neoplastic proliferation. With the double-staining immunohistochemical method, it was evident that, in different areas, the neoplastic proliferation was heterogeneous, but there were numerous cells with abundant cytoplasm positive for mitochondria. Other cells of small size were purely immunoreactive for GFAP. Several neoplastic elements were positive with both antisera, and this was mostly evident in cells showing a cytoplasmic targetoid appearance. The remarkable presence of a high number of mitochondria was also demonstrated in 2 cases by EM: mitochondria were also of different sizes. All these features led to the interpretion of the cells composing these cases as OC neoplastic elements according to the currently established morphological criteria as described in various sites [9]. Of the 10 propositus cases, 9 were also found to have high levels of mtDNA copy number. The significantly increased amount of mtDNA in these 9 cases, with respect to control gray matter tissue, is in keeping with the OC nature of these cells. Therefore, the quantitative mtDNA assessment, in addition to the classical morphological features of OC, may constitute a useful integrated approach to diagnose OC tumors. Thus, we consider it appropriate to expand the definition of oncocytoma, currently based on classic morphological parameters (ie, histology, EM, and immunohistochemistry), by including the quantitative assessment of mtDNA content, whenever possible. To the best of our knowledge, OC changes have not been previously reported in GBM, and we propose to adopt the designation of OCGBM for these neoplastic lesions. Three cases were discordant between classic morphological parameters and mtDNA quantitative assessment. Such discordance is probably related to the different areas analyzed and depends on the tissue heterogeneity typical of most tumors, including oncocytomas.

1874

G. Marucci et al.

Fig. 7 Mitochondrial DNA quantification. A, mtDNA copy number shown for single cases analyzed. B, mtDNA copy number in OC cases, non-OC cases, and 2 sex- and age-matched control groups. Bars represent mean and SD of the different groups. Asterisk indicates statistically significant difference between OC and non-OC group and between OC and substantia nigra controls.

Admittedly, 10 cases displayed consistent H&E and immunohistochemical (3+) OC features, a somewhat high number for a not previously documented phenotype. This is consequent to the fact that we have been investigating tumors

showing “unexpected” OC changes for many years [9,12], and we had the opportunity to recruit the present cases from a large database offered by the Progetto Emiliano-Romagnolo di

Table 2 Summary of the clinical features of the OC versus the non-OC group side by side

Age (y), mean (range) Sex F M MGMT Met U TMZ + RT Y N Median survival (mo)

OC group

Non-OC group

64.4 (41-74)

61.6 (41-77)

55.6% 44.4%

52.6% 47.4%

33.3% 66.7%

42.1% 57.9%

55.6% 44.4% 16

57.9% 42.1% 8.7

Abbreviations: OC, ovarian carcinoma; F, female; M, male; Met, methylated; U, unmethylated; TMZ, temozolomide; RT, radiotheraphy; Y, yes; N, no.

Fig. 8 Survival curves for each group. There was no statistically significant difference (log-rank test, P = .597; Wilcoxon test, P = .233) in survival rate between GBM with OC features and GBM without OC features in this series, but a trend for longer median survival has been evidenced in patients with OC-GBM (median survival of 16 versus 8.7 months).

Oncocytic glioblastoma Neuro-Oncologia regional project. Thus, we aimed to verify the incidence of morphologic and immunohistochemical OC changes by re-examining a consecutive routine series of 50 “ordinary” GBMs and thereby found OC features in 2 cases (4%). Although this percentage is not high, it may indicate a similar situation to what is seen in other anatomical districts where OC changes often go undetected, especially when the changes are focal. The OC nature of the tumors hereby analyzed has been revealed by the assessment of the mtDNA content, which paralleled the immunohistochemistry and EM, indicating that routine histology may benefit from the enforcement by molecular diagnostics procedures. Overall, the present findings highlight the importance of integrating molecular analyses in supporting the final diagnosis of OC. Recent studies have shown that OC changes in most of the cases are due to accumulation of mtDNA mutations, frequently occurring in genes encoding complex I subunits that are functionally affected [31,32]. The mitochondrial proliferative phenotype is interpreted as a compensatory mechanism to severe impairment of oxidative phosphorylation consequent to mtDNA mutations, by analogy with similar features that characterize patients with mitochondrial encephalomyopathies [33]. Interestingly, albeit not reaching statistical significance, a trend for a longer survival rate for OC-GBMs was observed in our sample set, which deserves a further, appropriately designed, investigation on a larger cohort of cases. This trend is consistent with the less aggressive and less proliferative characteristics of most OC tumors that carry an identifiable mitochondrial dysfunction. Although not all OC tumors per se are less aggressive or malignant than their non-OC counterparts, this holds true, for instance, in the case of kidney, parotid, and pituitary OC neoplasms, whereas in the thyroid, other criteria than the OC phenotype are used to ascertain malignancy [34]. It has been demonstrated that a heavy impairment of the oxidative phosphorylation due to disruptive mtDNA mutations, the main hallmark of the OC phenotype, has a counterproductive effect on the tumor progression to malignancy [35], in keeping with the survival trend observed in the present small series of OC-GBMs. This may be mainly consequent to the pseudonormoxia the cells sense when the Krebs cycle is slowed down and αketoglutarate accumulates due to lack of respiratory complex I [35]. In fact, lack of the hypoxia-inducible factor 1a, the main mediator of progression to malignancy, and transcriptional activator of the vascular endothelial growth factor A, following α-ketoglutarate accumulation, explains both the paucity of vascularization of OC tumors as well as their low invasive potential. In conclusion, molecular investigations of the several mtDNA mutations in OC-GBMs may further clarify the mechanisms underlying the OC differentiation of GBMs and shed light on the biological and prognostic features of these neoplasms.

1875

References [1] Louis DN, Ohgaki H, Wiestler OD, Cavenee WK. World Health Organization Classification of Tumours of the Central Nervous System. Lyon: IARC Press; 2007. [2] Henriksson R, Asklund T, Poulsen HS. Impact of therapy on quality of life, neurocognitive function and their correlates in glioblastoma multiforme: a review. J Neurooncol 2011;104:639-46. [3] Deorah S, Lynch CF, Sibenaller ZA, Ryken TC. Trends in brain cancer incidence and survival in the United States: surveillance, epidemiology, and end results program, 1973 to 2001. Neurosurg Focus 2006; 20(4):E1. [4] Stupp R, Mason WP, van den Bent MJ, et al. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 2005;352:987-96. [5] Hamperl H. Beiträge zur normalen und pathologischen Histologie menschlieher Speicheldriisen. Z Mikrosk Anat Forsch 1931;27:1-55. [6] Roth SI, Olen E, Hansen LS. The eosinophilic cells of the parathyroid (oxyphil cells), salivary (oncocytes), and thyroid (Hürthle cells) glands. Light and electron microscopic observations. Lab Invest 1962; 11:933-41. [7] Tremblay G, Pearse AG. Histochemistry of oxidative enzyme systems in the human thyroid, with special reference to Askanazy cells. J Pathol Bacteriol 1960;80:353-8. [8] Ghadially FN. Diagnostic Electron Microscopy of Tumors. 2nd ed. London: Butterworth & Company; 1985. [9] Ragazzi M, de Biase D, Betts CM, et al. Oncocytic carcinoma of the breast: frequency, morphology and follow-up. HUM PATHOL 2011;42: 166-75. [10] Gasparre G, Romeo G, Rugolo M, Porcelli AM. Learning from oncocytic tumors: why choose inefficient mitochondria? Biochim Biophys Acta 1807;2011:633-42. [11] Kovacs K, Horvath E. Pituitary “chromophobe” adenoma composed of oncocytes. Arch Pathos 1973;95:235-9. [12] Roncaroli F, Riccioni L, Cerati M, et al. Oncocytic meningioma. Am J Surg Pathol 1997;21:375-82. [13] Sav A, Scheithauer BW, Mazzola CA, Ketterling SR, Thompson SJ, Reilly MH. Oncocytic choroid plexus carcinoma: case report. Clin Neuropathol 2010;29:14-20. [14] Vajtai I, von Gunten M, Fung C, Brekenfeld C, Kappeler A, Reinert MM. Oncocytic ependymoma: a new morphological variant of highgrade ependymal neoplasm composed of mitochondrion-rich epithelioid cells. Pathol Res Pract 2011;207:49-54. [15] Utsuki S, Oka H, Kijima C, et al. Subependymal giant cell astrocytoma with oncocytic change. Brain Tumor Pathol 2011;28:53-7. [16] Wallace DC. A mitochondrial paradigm of metabolic and degenerative diseases, aging, and cancer: a dawn for evolutionary medicine. Annu Rev Genet 2005;39:359-407. [17] DiMauro S, Schon EA. Mitochondrial respiratory-chain diseases. N Engl J Med 2003;348:2656-68. [18] Clay Montier LL, Deng JJ, Bay Y. Number matters control of mammalian mitochondrial DNA copy number. J Genet Genomics 2009;36:125-31. [19] Chatterjee A, Mambo E, Sidransky D. Mitochondrial DNA mutations in human cancer. Oncogene 2006;25:4663-74. [20] Tseng LM, Yin PH, Chi CW, et al. Mitochondrial DNA mutations and mitochondrial DNA depletion in breast cancer. Genes Chromosomes Cancer 2006;45:629-38. [21] Wang Y, Liu VW, Xue WC, Cheung AN, Ngan HY. Association of decreased mitochondrial DNA content with ovarian cancer progression. Br J Cancer 2006;95:1087-91. [22] Gasparre G, Iommarini L, Porcelli AM, et al. An inherited mitochondrial DNA disruptive mutation shifts to homoplasmy in oncocytic tumor cells. Hum Mutat 2009;30:391-6. [23] Lee HC, Wei YH. Mitochondrial role in life and death of the cell. J Biomed Sci 2000;7:2-15.

1876 [24] Singh KK, Russell J, Sigala B, Zhang Y, Williams J, Keshav KF. Mitochondrial DNA determines the cellular response to cancer therapeutic agents. Oncogene 1999;18:6641-6. [25] Oliva CR, Moellering DR, Gillespie GY, Griguer CE. Acquisition of chemoresistance in gliomas is associated with increased mitochondrial coupling and decreased ROS production. PLoS One 2011;6:e24665. [26] Oliva CR, Nozell SE, Diers A, et al. Acquisition of temozolomide chemoresistance in gliomas leads to remodeling of mitochondrial electron transport chain. J Biol Chem 2010;285:39759-67. [27] Liang BC. Evidence for association of mitochondrial DNA sequence amplification and nuclear localization in human low-grade gliomas. Mutat Res 1996;354:27-33. [28] Morandi L, Franceschi E, de Biase D, et al. Promoter methylation analysis of O6-methylguanine-DNA methyltransferase in glioblastoma: detection by locked nucleic acid based quantitative PCR using an imprinted gene (SNURF) as a reference. BMC Cancer 2010;10:48. [29] Morandi L, de Biase D, Visani M, et al. Allele specific locked nucleic acid quantitative PCR (ASLNAqPCR): an accurate and cost-effective

G. Marucci et al.

[30]

[31]

[32]

[33]

[34] [35]

assay to diagnose and quantify KRAS and BRAF mutation. PLoS One 2012;7(4):e36084. Cossarizza A, Riva A, Pinti M, et al. Increased mitochondrial DNA content in peripheral blood lymphocytes from HIV-infected patients with lipodystrophy. Antivir Ther 2003;8:315-21. Costa-Guda J, Tokura T, Roth SI, Arnold A. Mitochondrial DNA mutations in oxyphilic and chief cell parathyroid adenomas. BMC Endocr Disord 2007;7:8. Gasparre G, Porcelli AM, Bonora E, et al. Disruptive mitochondrial DNA mutations in complex I subunits are markers of oncocytic phenotype in thyroid tumors. Proc Natl Acad Sci 2007;104:9001-6. Iommarini L, Calvaruso MA, Kurelac I, Gasparre G, Porcelli AM. Complex I impairment in mitochondrial diseases and cancer: parallel roads leading to different outcomes. Int J Biochem Cell Biol 2013;45:47-63. Tallini G. Oncocytic tumours. Virchows Arch 1998;433:5-12. Gasparre G, Kurelac I, Capristo M, et al. A mutation threshold distinguishes the antitumorigenic effects of the mitochondrial gene MTND1, an oncojanus function. Cancer Res 2011;71:6220-9.