Not all 1p/19q non-codeleted oligodendroglial tumors are astrocytic

Although 1p/19q codeletion is the genetic hallmark defining oligodendrogliomas, approximately 30-40% of oligodendroglial tumors have intact 1p/19q in the literature and they demonstrate a worse prognosis. This group of 1p/19q intact oligodendroglial tumors is frequently suggested to be astrocytic in nature with TP53 and ATRX mutations but actually remains under-investigated. In the present study, we provided evidence that not all 1p/19q intact oligodendroglial tumors are astrocytic through histologic and molecular approaches. We examined 1p/19q status by FISH in a large cohort of 337 oligodendroglial tumors and identified 39.8% lacking 1p/19q codeletion which was independently associated with poor prognosis. Among this 1p/19q intact oligodendroglial tumor cohort, 58 cases demonstrated classic oligodendroglial histology which showed older patient age, better prognosis, association with grade III histology, PDGFRA expression, TERTp mutation, as well as frequent IDH mutation. More than half of the 1p/19q intact oligodendroglial tumors showed lack of astrocytic defining markers, p53 expression and ATRX loss. TP53 mutational analysis was additionally conducted in 45 cases of the 1p/19q intact oligodendroglial tumors. Wild-type TP53 was detected in 71.1% of cases which was associated with classic oligodendroglial histology. Importantly, IDH and TERTp co-occurred in 75% of 1p/19q intact, TP53 wild-type oligodendrogliomas, highlighting the potential of the co-mutations in assisting diagnosis of oligodendrogliomas in tumors with clear cell morphology and non-codeleted 1p/19q status. In summary, our study demonstrated that not all 1p/19q intact oligodendroglial tumors are astrocytic and co-evaluation of IDH and TERTp mutation could potentially serve as an adjunct for diagnosing 1p/19q intact oligodendrogliomas.


INTRODUCTION
Oligodendroglial tumors comprise of oligodendrogliomas and oligoastrocytomas and are associated with favorable clinical behaviors compared with astrocytomas and some can be chemosensitive [1].Codeletion of 1p and 19q is a characteristic genetic signature which is observed in 39-70% of oligodendrogliomas and 21-59% of oligoastrocytomas in most series [2][3][4].Codeletion is associated with longer survival and is also a predictive marker for response to chemotherapy with procarbazine, lomustine and vincristine [1,5].
Aside from codeletion of 1p/19q, oligodendroglial tumors are strongly associated with mutations of IDH1/2, Research Paper: Pathology TERT promoter (TERTp) and they are also associated with methylation of MGMT and upregulation of PDGFRA [6][7][8].TERTp mutations are regarded as a common mechanism of upregulation of telomerase in primary glioblastoma and oligodendroglial tumors [9].1p/19q codeleted and TERTp mutated tumors are mutually exclusive with tumors exhibiting mutations in ATRX (alpha thalassemia mental retardation syndrome X linked) and TP53.The latter two are usually regarded as markers of astrocytic lineage [7,[10][11][12].
The fact remains that in most series with many coming from important international trials, 30-40% of oligodendroglial tumors are 1p/19q non-codeleted and these tumors exhibit a worse prognosis [1,[13][14][15].It has been proposed that most, if not all, of these non-codeleted tumors are actually astrocytic in nature with either TP53 or ATRX mutations [11].However, there have been few large-scale studies on molecular characterization of oligodendroglial tumors without 1p/19q codeletion.They make up a significant proportion of "oligodendroglial tumors" diagnosed in international series and daily practice and there is no standard-of-care treatment strategy.In this study, we aim to characterize the molecular and clinical features of this neglected group, interrogating them with standard biomarkers and correlating with overall survival.

DISCUSSION
In the present study, we examined a large cohort of oligodendroglial tumors (N = 337) and identified 60.2% of cases harbored 1p/19q codeletion.Since first reported by Reifenberger and colleagues in 1994 [16], combined chromosomal whole arm deletion in 1p and 19q has been referred as genetic hallmark of oligodendrogliomas [17].Mediated by an unbalanced translocation between chromosome 1 and 19, with formation of derivative chromosomes der(1;19)(q10;p10) and der(1;19)(p10;q10) and the latter one being subsequently deleted during gliomagenesis [13,18], 1p/19q codeletion was one of the most representative and remarkable molecular markers in diffuse gliomas with diagnostic, prognostic and predictive utilities in clinical practice [7,19].The marker served as a stratifier in clinical trials and was shown to predict therapeutic response to PCV chemotherapy (procarbazine, lomustine and vincristine) and radiotherapy in oligodendroglial tumors [1,5].1p/19q codeletion will be one of the molecular markers being incorporated into the integrated morphological and molecular diagnosis of CNS tumors as proposed by the International Society of Neuropathology-Haarlem Consensus Guidelines [20].
Genomic landscape of grades II and III diffuse gliomas was comprehensively interrogated in the landmark study reported by the TCGA Research Network using multi-omic platforms [39].The genomewide analyses identified concordant classification of three prognostically relevant molecular subgroups of lower grade gliomas based on unsupervised clustering of multi-omic data, which can be robustly identified by IDH mutation and 1p/19q codeletion.Importantly, IDH mutated, 1p/19q codeleted gliomas were characterized by mutations in CIC, FUBP1, NOTCH1, and TERT promoter, suggesting these mutations could serve as oligodendroglial lineage markers.IDH mutated gliomas lacking 1p/19q codeletion showed high frequency of TP53 mutation (94%) and ATRX inactivation (86%), compatible with their role as astrocytic markers.These substantial genomic data provided the basis of our choice of astrocytic and oligodendroglial markers in our current study (Supplementary Table S3).It was worthwhile to note that in the TCGA report, OncoSign analysis based on recurrent copy-number variations, mutations, and gene fusions, was performed and the group identified four dominant OncoSign classes, namely OSC1 to OSC4, which were correlated with the IDH-1p/19q defined subgroups.1p/19q non-codeleted oligodendrogliomas were enriched in OSC3 and were associated with IDH mutation and TERTp mutation, validating the observations in our study.Another remarkable study from Weller et al. interrogated grades II and III diffuse gliomas using genome-and transcriptomewide profiling and reported molecular subgroups harboring distinct genomic aberrations and expression patterns which provided prognostically significant information on top of IDH mutation and 1p/19q codeletion [26].Genomic profiling revealed five molecular subgroups with Group I tightly correlating with IDH mutation and 1p/19q codeletion and exhibiting the best clinical outcome, and Group V harboring glioblastoma-like copy number changes including 7q gain and 10q loss, and showing the poorest survival.Importantly, the genomic profiling provided clinically relevant information with biological implication for prognostication of diffuse gliomas beyond histologic classification and grading.
1p/19q non-codeleted oligodendrogliomas in our study were associated with TERTp mutation (p = 0.007) and demonstrated frequent IDH mutation (67.4%).Comutations of IDH and TERTp occurred in 75% of 1p/19q intact, TP53 wild-type oligodendrogliomas in our cohort.Notably, IDH mutation was also frequently detected in the 1p/19q intact oligodendroglial tumors and conferred chemo-radiation sensitivity in the absence of 1p/19q codeletion as shown by Cairncross and colleagues [21].In terms of classification and diagnostic utility, presence of IDH mutation, even without 1p/19q codeletion, has been suggested to indicate the diagnosis of oligodendrogliomas over other tumors with similar histo-morphological features [40].Multiple studies have shown the association between TERTp mutation and 1p/19q codeletion as well as classic oligodendroglial morphology [6,9,10,41,42].Our current study validated previous observations and additionally demonstrated the association between TERTp mutation and classic oligodendroglial morphology in a molecular background of intact 1p/19q.Notably, in the study by Killela and colleagues examining more than 470 diffuse gliomas, co-mutations of IDH and TERTp were identified in 79.3% (69/87) of oligodendrogliomas while 1p/19q codeletion was only observed in 54% (47/87) of the cases [41].Our study and previous observations suggested the potential of IDH mutation plus TERTp mutation in assisting the diagnosis of oligodendroglioma for a tumor with clear cell morphology but lacking 1p/19q codeletion.
It was crucial to avoid misclassification of true 1p/19q codeleted oligodendrogliomas as non-deleted tumors (ie false negativity) in our study.We employed FISH and used commercial probe targeting the "minimally deleted regions" [43][44][45] in order to maximize the sensitivity in detecting 1p/19q codeletion [2].The technique had been well standardized and adopted in various studies [28,[31][32][33]38] and could be applied robustly on formalin-fixed paraffin-embedded tissues in routine diagnostic neuropathology practice.Importantly, FISH probes targeting at the same loci were also adopted by the EORTC and RTOG clinical trials [1,5].Since 1p/19q loss classically involved deletion of the entire 1p and 19q arms, tumors with retained minimally deleted regions in our study were extremely unlikely to be false negative cases.
We further examined astrocytic markers, p53 expression and ATRX loss, in the 1p/19q non-codeleted oligodendroglial tumors.Co-evaluation of the two astrocytic immuno-markers revealed more than 50% of the 1p/19q non-codeleted oligodendroglial tumors did not show p53 positivity and ATRX loss, i.e. they were not molecularly astrocytic.Among 38 cases of mixed oligoastrocytomas with analyzable data of ATRX expression and TERTp mutation, the two biomarkers showed mutual exclusivity in all but one case [Supplementary Figure S1e].TERTp and ATRX mutations represented independent genetic mechanisms in maintaining telomere lengths in tumorigenesis of diffuse gliomas [42].In the study by Sahm and colleagues investigating 43 mixed oligoastrocytomas, 1p/19q codeletion and ATRX loss were mutually exclusive in all but one case, suggesting parting with the diagnosis of oligoastrocytoma.According to these authors, all noncodeleted oligodendroglial tumors are in fact astrocytic [11].With the morphological and molecular evidence of several cases of "true" oligoastrocytoma being reported [46][47][48], complete farewell to oligoastrocytoma can be argued.Notably, in our series, we also did not find any true cases of "oligoastrocytoma".
Clinical trials by Cairncross et al. and van den Bent et al. have set the standard therapeutic protocol for patients with 1p/19q codeleted anaplastic oligodendroglial tumors [1,5].The former research group has additionally demonstrated that the survival benefit of PCV (procarbazine, lomustine, and vincristine) in 1p/19q non-codeleted anaplastic oligodendroglial tumors was associated with IDH mutation [21].Clinical benefits of radiation plus PCV have been extended to treatment of low grade diffuse gliomas [49].The chemoradiotherapy was associated with prolonged overall survival and progression-free survival in subsets of oligodendroglioma, oligoastrocytoma, and IDH mutated low grade gliomas.Our group has previously reported the potential clinical value of TERTp mutation in predicting differential responses to genotoxic therapies in grades II and III diffuse gliomas, on top on IDH status [50].Our current study identified a major portion of oligodendrogliomas lacking 1p/19q codeletion was characterized by IDH and TERTp co-mutation, with a background of ATRX nuclear positivity wild-type TP53 (Figure 5), indicating that these patients could potentially benefit from chemoradiotherapy.
In summary, this study examined a large cohort of oligodendroglial tumors and demonstrated that astrocytic tumors could not account for all 1p/19q non-codeleted oligodendroglial tumors and oligodendroglioma lacking 1p/19q codeletion may form a distinct subgroup of diffuse glioma.Co-evaluation of IDH and TERTp mutation had diagnostic value for oligodendroglioma and could serve as an adjunct in oligodendroglial tumors after evaluation of the 1p/19q status.

Patients and tissue samples
Oligodendroglial tumors diagnosed from our service archives at the Prince of Wales Hospital (The Chinese University of Hong Kong) and Huashan Hospital (Fudan University, Shanghai) were included in the present study.Totally 337 oligodendroglial tumors were included in our study and 71 of them were from Hong Kong and 266 from Shanghai.Only cases which were reviewed by two neuropathologists (HK N and H C) as oligodendroglial were included based on The 2007 WHO classification of tumors of the central nervous system [51].The present study used only formalin-fixed, paraffin-embedded material.Clinical and survival data of the patients were retrieved from the respective institutional medical record systems.This study was approved by the Ethics Committee of Shanghai Huashan Hospital and the New Territories East Cluster-Chinese University of Hong Kong Ethics Committee.The cohort in this study partially overlapped with previous studies [6,15,52].

Fluorescence In Situ hybridization for chromosome 1p and 19q
Fluorescent in situ hybridization (FISH) was performed on formalin-fixed, paraffin-embedded tumor tissue to detect deletion of chromosome 1p and 19q [15].4-μm-thick sections were deparaffinized, treated with sodium thiocyanate and followed by digestion with pepsin solution at 37°C.Dual-color-probe hybridization was then performed using Vysis 1p36/1q25 and 19q13/19p13 FISH Probe Kit (Abbott Molecular) and the spectrum-green-probe was labeled on chromosome 1q and 19p, respectively.Both probes and target tumor DNA were denatured in an 80°C oven for 30 minutes and followed by an overnight incubation at 37°C.Nuclei were counterstained with Vectashield mounting medium containing 4', 6-diamidino-2-phenylindole (Vector Laboratories, Burlingame, CA, USA) and the number of FISH signals was assessed under a Zeiss Axioplan fluorescence microscope (Carl Zeiss Microscopy LLC, NY, USA) equipped with a triple-pass filter (DAPI/ Green/Orange).Hybridizing signals of at least 100 nonoverlapping nuclei were enumerated and a sample will be considered as 1p or 19q deleted when more than 25% of counted nuclei presented one target (orange) signal and two reference (green) signals [2].

Mutation analysis of IDH1/2, TERT promoter and TP53
Mutation status of IDH1/2, TERT promoter and TP53 were studied with direct sequencing [6,52,53].Tissues from representative areas with tumor content > 70% were collected from 5 to 7 dewaxed formalinfixed, paraffin-embedded sections and re-suspended in 10mM Tris-HCl buffer (pH 8.5) which also contained proteinase K with a final concentration of 2 mg/ml.After that, the mixture was incubated at 55°C for 2-18 h and then at 98°C for 10 min.After the above incubation, the buffer comprising the cell lysate was used for polymerase chain reaction (PCR) analysis of IDH1, IDH2, TP53 and TERTp.We performed direct sequencing for IDH1, IDH2, TERT promoter region and exon 4-9 of TP53 as per previous protocols [6,15,[52][53][54].All mutational status was validated by sequencing of a newly amplified DNA fragment.Sequences of primers used for PCR analysis of IDH1/2, TERT promoter and TP53 were listed in Supplementary Table S1.For sequencing of IDH1/2 and TP53, target DNA fragments were amplified in a total 10μl reaction volume which containing 1μl of cell lysate, 1μl of MgCl2 (25mM), 1μl of 10*PCR buffer II, 0.2μl of each deoxyribonucleoside triphosphate (10mM), 0.3μl of each forward and reverse primers of target DNA fragment (10nmol), 0.075μl of AmpliTaq Gold DNA polymerase (Life Technologies Corporation, Hong Kong, China).PCR for TERT promoter region was performed in 10μl reaction mixture containing 1μl of cell lysate, 0.3μl of each forward and reverse primers of target DNA fragment (10nmol), and 5μl of KAPA HiFi HotStart ReadyMix DNA Polymerase (Kapa Biosystems Wilmington, DE, USA).PCR reaction started with a denature procedure of 95°C for 10 min, then followed by 45 cycles of 95°C for 20 sec, annealing temperature (Supplementary Table 1) for 20 sec and 72°C for 30 sec, and a final extension step of 72°C for 3 min.Products were then treated with Exonuclease I (TakaRa Biotechnology Limited, Dalian, China) of 2μl (0.25U/ μl) per 5μl PCR product at 37°C for 15min followed by 80°C for 15min.Sequencing of target DNA fragment was performed using BigDye Terminator Cycle Sequencing kit v.1.1 (Life Technologies).The Genetic Analyzer 3130xl was used for the following sequencing and the results were analyzed by Sequencing Analysis Software.Hotspot mutations of IDH1/2 and TERTp, all missense mutations of TP53 could be detected.All detected missense mutations of TP53 were listed in Supplementary Table S2.

Immunohistochemistry of ATRX, p53 and PDGFRA
Immunohistochemistry was performed for detecting the expression of p53, PDGFRA, and ATRX.4μm thick formalin-fixed, paraffin-embedded (FFPE) sections of each sample were de-waxed by xylene and rehydrated in graded alcohols.Sections were then treated with citrate buffer (pH 6.0) by heating for antigen retrieval.After antigen retrieval, they were subjected to immuno-staining by BenchMark XT automated tissue staining systems (Ventana Medical Systems, Tucson, AZ, USA) using validated protocols for detecting the expression of p53 and ATRX.Sections were incubated with antibodies of anti-p53 (Dako DO-7, 1:100) and anti-ATRX (SIGMA CAT: HPA001906, 1:400) at 37°C for 32 min and followed by an incubation with UltraView HRP-conjugated multimer antibody reagent (Ventana Medical Systems).Subsequent antigen detection was performed using UltraView diamino benzidine chromogen step (Ventana Medical Systems).Expression of PDGFRA was detected using the automated Bond-max system (Leica Bond-Max) with anti-PDGFRA antibody (Santa Cruz, C-20, 1:200) and subsequent antigen detection was performed using Ultra View diamino benzidine chromogen step (Bond-Max).At last, slides were counterstained with Mayer's hematoxylin.A tumor was scored p53 positive if > 10% of tumor cells showed strong nuclear staining [53].Samples with more than 10% of tumor cells showing positive nuclear staining of ATRX were scored as positive.Endothelial cells, cortical neurons and infiltrating inflammatory cells were generally positive and served as internal positive controls [12].Cytoplasmic and membrane staining were considered for evaluation of PDGFRA expression.Both the distribution and intensity of staining were semi-quantitatively scored as previously reported [55].

Statistical analysis
Statistical analysis was performed using IBM SPSS Statistics Version 20 (IBM Corporation, NY, USA).Univariate analysis was conducted using Chisquare or Fisher's test to compare categorical variables.Independent-Samples T Test was used to compare mean age between 2 populations.Kaplan-Meier estimator and log rank test were performed for univariate survival analysis.Whereas a Cox proportional hazards model was employed for multivariate survival analysis.Statistical significance was considered when p < 0.05 (two-side).

Figure 5 :
Figure 5: Overview of the sequential work-flow based on The 2016 WHO classification of CNS tumors and Haarlem scheme [20, 56].mixed, mixed oligoastrocytic histology; CFO, classic for oligodendroglial morphology.

Table 3 : Multivariate analysis of clinical and molecular variables in oligodendroglial tumors
mut, mutant; wt, wild-type; n, number of cases with data available.