Detection of ATRX and IDH1-R132H immunohistochemistry in the progression of 211 paired gliomas

Recurrence and progression to higher grade lesions are key biological events and characteristic behaviors in the evolution process of glioma. A small residual population of cells always escapes surgery and chemoradiation, resulting in a typically fatal tumor recurrence or progression. IDH mutation (isocitrate dehydrogenase) and ATRX (alpha-thalassemia/mental retardation, X-linked) loss/mutation occur in association and may represent early genetic alterations in the development of gliomas. However, their prognostic value in the evolution of gliomas still needs further investigation. Two hundreds and eleven serial sampling of gliomas were included in our study. We used immunohistochemistry (IHC) to detect IDH1-R132H mutation and ATRX status and showed that the IDH1-R132H and (or) ATRX status could be necessary to provide the basic molecular information for the “integrated diagnosis” of gliomas. We illustrated an evaluation formula for the evolution of gliomas by IDH1-R132H combined with ATRX immunohistochemistry and identified the association of IDH1-R132H/ATRX loss accompanied by longer progression time interval of patients with gliomas. Furthermore, we observed that most recurrences had a consistent IDH1 and ATRX status with their matched primary tumors and demonstrated the progressive pattern of grade II astrocytoma/oligodendroglial tumors and anaplastic oligoastrocytoma with or without IDH1-R132H. Identification of IDH1-R132H and ATRX loss status in the primary-recurrent gliomas may aid in treatment strategy selection, therapeutic trial design, and clinical prognosis evaluation.


INTRODUCTION
Diffuse gliomas are classified into astrocytomas, oligoastrocytomas and oligodendrogliomas of grade II, grade III and glioblastoma based on the 2007 World Health Organization (WHO) Classification of Tumors of the Central Nervous System [1]. Although patients with low grade gliomas (LGGs, grade II) [2] have a more favorable prognosis than patients with high grade gliomas (grade III and IV), in 50-75% of patients with

Frequency and diagnostic value of IDH1-R132H and ATRX loss in gliomas
To test the impact of IDH1-R132H and ATRX on routine diagnostic neuropathology and their evaluation for progression of gliomas, we analyzed a series of 211 serial sampling of glioma tissues, including 103 astrocytomas (A, AA), 25 oligodendrogliomas (O, AO), 123 oligoastrocytomas (OA, AOA) and 181 glioblastomas (pGBM, sGBM, rGBM) by immunohistochemistry for IDH1-R132H and ATRX evaluation.
Then we performed receiver operating characteristic (ROC) analysis to delineate the sensitivity and specificity of IDH1-R132H and (or) ATRX loss for WHO histology classification. As shown in Figure 3, the area under roc curve (AUC) of using IDH1-R132H as a diagnostic biomarker for discriminating between pGBM and grade II/III glioma, sGBM was 0.7414 (sensitivity 63.19%, specificity 85.09%, Figure 3A). The AUC of using ATRX loss as a diagnostic biomarker for discriminating between pGBM, oligodendroglioma (WHO grade II/III) and sGBM, astrocytoma (WHO grade II/III) was 0.8241 (sensitivity 76.67%, specificity 88.15%, Figure 3B). When combining IDH1-R132H with ATRX loss as diagnostic biomarkers for discriminating between pGBM, oligodendroglioma (WHO grade II/III) and sGBM, astrocytoma (WHO grade II/III), the AUC was 0.7407 (sensitivity 53.33%, specificity 94.815%, Figure 3C). We also performed ROC analysis for distinguishing primary and secondary GBMs using IDH1-R132H and (or) ATRX. When we just used IDH1-R132H status for ROC analysis, the AUC is 0.7644 (the sensitivity is 67.8%, the specificity is 85.09%, Figure 3D). When we just used ATRX loss status for ROC analysis, the AUC is 0.82 (the sensitivity is 87.72%, the specificity is 76.27%, Figure 3E). When we combined IDH1-R132H and ATRX loss status for ROC analysis, the AUC is 0.7832 (the sensitivity is 61.02%, the specificity is 95.61%, Figure 3F). These results indicated that IDH1-R132H and (or) ATRX loss status might have a good predictive ability for WHO histology diagnosis. www.impactjournals.com/oncotarget Prognostic value of IDH1-R132H and ATRX loss in the evolution of gliomas We analyzed the cohort of patients with primary and recurrent gliomas to observe the value of IDH1-R132H or ATRX as biomarkers evaluating the progressionfree survival of patients with gliomas. In our cohort, we estimated the progression-free survival (PFS) in WHO grade II-IV patients and observed a remarkable separation (Median PFS of low grade gliomas, LGGs = 830 days, Median PFS of anaplastic gliomas, AGs = 495 days, Median PFS of GBMs = 394 days, p < 0.0001, log-rank test, Table 2).
PFS was significantly longer in patients with grade II gliomas harboring IDH1-R132H than in patients with IDH1 wild-type tumors (Median PFS of IDH1-R132H = 893 days, Median PFS of IDH1-wt = 580 days, p = 0.0009, log-rank test; Figure 4A). Although PFS was longer in patients with AG or pGBM harboring IDH1-mutated tumors than in patients with IDH1 wild-type tumors, this difference was not significant (p = 0.0702 and p = 0.0914, respectively, log-rank test; Figure 4B, 4C). Although there was no significant difference between ATRX loss and ATRX expression in the cohort of patients with grade II or grade III tumors ( Figure 5A, 5B), ATRX loss was associated with longer progression-free survival in patients with grade II/III oligoastrocytomas or grade II/III gliomas ( Figure 5C, 5D). Patients with astrocytic tumors losing ATRX had a longer progression-free survival than patients with tumors expressing ATRX ( Figure 5E). In additions, there was a trend that patients with IDH1-R132H and ATRX loss grade II/III gliomas had longer PFS than patients with IDH1-R132H and ATRX positive grade II/III gliomas and PFS of patients with GBM was similar to that of patients with IDH wild type ( Figure 5F). We have demonstrated the basic clinical characteristics among the IDH-R132H/ATRXloss, IDH-R132H/ATRXexpr, IDH-WT and GBM groups in the Supplementary Table1. After incorporation of age, WHO grade, IDH1 and ATRX status into the multivariate model, the prognostic value of IDH1-R132H was still significant (Table 3).

IDH1-R132H and ATRX loss in matched primary and recurrent gliomas
First, we excluded these pairs that the primary gliomas were high grade tumor while the recurrence was lower grade tumor and then analyzed the pattern of glioma progression. As shown in Table 4, the recurrent tumors of primary LGGs included LGGs, AGs and GBMs of   which rates corresponded to 29% (29/100), 35% (35/100) and 36% (36/100), respectively. The time interval of the three progressive patterns was analyzed. There was no difference in the progression-free survival when LGG malignantly recurred to LGG, AG or GBM (Median PFS = 802 days, 830 days and 835 days, respectively, Figure 6A). However, the time interval of AG to GBM was significantly shorter than that of AG to AG (Median PFS = 365 days and 611 days, respectively, p = 0.0067, log-rank test). The progressive time interval of GBM to GBM was similar to that of AG to GBM (Median PFS = 395 days, p = 0.8446, Figure 6B). IDH1-R132H occurred stably in the primary tumors and matched recurrent tumors, although not all the recurrent tumors showed consistent IDH1-R132H status as the corresponding primary tumors (

Progressive pattern of primary glioma with or without IDH1-R132H
Generally, diffuse LGGs malignantly transformed into AGs, followed by secondary GBM, but occasionally, there might be no evidence of the former one. Meanwhile, some recurrences had different morphology from initial tumors (astrocytic to oligodendroglial and vice versa). In our cohort, we discriminated the IDH1-R132H initial tumors from the IDH1-WT tumors and analyzed the evolution pattern of 125 primary-recurrent paired tumors, 49 recurrences of which were originated from grade II A ( Figure 7A), 47 were from grade II O ( Figure 7B) and 29 were grade III AOA ( Figure 7C). We observed that Figure 3: ROC curves of using IDH1-R132H and (or) ATRX status in discriminating the classification of glioma histology. The AUC of using IDH1-R132H as diagnostic biomarkers for discriminating between primary GBM and grade II/III glioma, secondary GBM is 0.7414 (the sensitivity is 63.19%, the specificity is 85.09%, (A). The AUC of using ATRX loss as diagnostic biomarkers for discriminating between primary GBM, oligodendroglioma (WHO grade II/III) and secondary GBM, astrocytoma (WHO grade II/III) is 0.8241 (the sensitivity is 76.67%, the specificity is 88.15%, (B). When IDH1-R132H combined with ATRX loss serves as diagnostic biomarkers for discriminating between astrocytoma (WHO grade II/III), secondary GBM, and primary GBM, oligodendroglioma (WHO grade II/III), the AUC is 0.7407 (the sensitivity is 53.33%, the specificity is 94.815%, (C). www.impactjournals.com/oncotarget regardless of with or without IDH1-R132H, the histology of recurrences from initial grade II A could be sGBM, AOA and AA. Although with the similar evolution pattern, patients with IDH1-R132H grade II A had a longer progressive time interval compared with patients with IDH1-WT grade II A (Median PFS = 895 days, Median PFS = 574 days, respectively, p = 0.0096, Figure 7D). Recurrences from IDH1-R132H grade II O mainly showed sGBM, OA morphology. In contrast, there were no main histology characteristics in the recurrences from the initial IDH1-WT grade II O. Patients with IDH1-R132H grade II O had a longer progression-free survival than patients with IDH1-WT grade II O (Median PFS = 858.5 days, Median PFS = 605 days, respectively, p = 0.0325, Figure 7E). GBM and AOA were the main recurrent histology from primary AOA. However, most of relapses from IDH1-R132H AOA were recurrent AOA; most of relapses from IDH1-WT AOA were sGBM. Patients with IDH1-R132H AOA also had a longer progression-free survival than patients with IDH1-WT ones (Median PFS = 629 days, Median PFS = 328 days, respectively, p = 0.0037, Figure 7F).

DISCUSSION
It is accepted that IDH mutation and ATRX mutation were used as diagnostic and prognostic biomarkers for molecular classification of gliomas [12,13]. However, their prognostic value for glioma progression is not clear, especially ATRX loss.
In our study, we used immunohistochemistry to detect IDH1-R132H and ATRX status instead of sequencing, aiming to provide a mature and simple tool for clinical practice. In the light of the "ISN-Haarlem" consensus, the routine approach to all diffuse astrocytic and oligodendroglial gliomas begins with performing IHC for ATRX and IDH1-R132H expression. Our results showed that the IDH1-R132H and (or) ATRX loss status could be necessary to provide the basic molecular information for the "integrated diagnosis" of gliomas.
IDH mutation is thought to be an early if not the initial event in the development of low-grade astrocytomas and oligodendrogliomas. In support of this hypothesis, IDH mutation is found in secondary glioblastomas but  LGG, low grade glioma; pA, primary astrocytoma; pO, primary oligodendroglioma; pOA, primary oligoastrocytoma; AG, anaplastic glioma; pAA, primary anaplastic astrocytoma; pAO, primary anaplastic oligodendroglioma; pAOA, primary anaplastic oligoastrocytoma; pGBM, primary glioblastoma.
Oncotarget 16390 www.impactjournals.com/oncotarget rare in primary glioblastomas [6]. IDH1 mutations have been shown to alter the enzymatic activity of the encoded protein, leading to up-regulation of hypoxia inducible factor-1α (HIF-1α) [14,15]. HIF-1α plays an important role in the process of angiogenesis while also supporting tumor cell survival and proliferation [14,16]. Although the properties of the IDH1 mutation would promote tumor growth, IDH mutation commonly indicates a favorable prognosis independent of WHO grades [7,[17][18][19].
Mutations and loss of expression of alphathalassemia/mental retardation syndrome X-linked (ATRX) was first reported in pancreatic neuroendocrine tumors [20]. ATRX protein plays a variety of key roles at tandem repeat sequences within the genome, including prevention of replication fork stalling, the deposition of a histone variant, and the suppression of a homologous recombination-based pathway of telomere maintenance [9]. Recently, mutation/loss of ATRX was identified as a potent biomarker in grade II-III gliomas and was associated with recurrent ones [21]. ATRX gene mutations were significantly correlated with ALT positivity [22]. Loss of ATRX in ALT frees macroH2A1.1 to bind and sequester tankyrase 1, thereby preventing resolution of sister telomere cohesion. Forced resolution of sister telomere cohesion induces excessive recombination between non-homologs, genomic instability, and impaired cell growth, indicating the ATRX-macroH2A1. 1-tankyrase axis as a potential therapeutic target in ALT tumors [23]. Inhibition of the protein kinase ATR, a critical regulator of recombination recruited by RPA, disrupts ALT and triggers chromosome fragmentation and apoptosis in ALT cells. The cell death induced by ATR inhibitors is highly selective for cancer cells that rely on ALT, suggesting that such inhibitors may be useful for treatment of ALT-positive cancers [24]. In patients with 1p/19q noncodeleted tumors with IDH mutations, those who were ATRX positive might have benefitted more than those who were negative from pre-RT PCV [25]. Whether these noncodeleted ATRX-positive tumors are biologically identical to codeleted tumors, save for absence of visible loss of 1p/19q, or represent laboratory errors in codeletion testing is unknown.
We previously reported that reduced expression of ATRX inhibited the migration, promoted apoptosis and reduced proliferation of glioma cells, suggesting the important role of ATRX in gliomas. Similarly, a report depicted how loss of ATRX could affect neuroprogenitor cell apoptosis [26].
IDH1-R132H and ATRX expression status are essential to the diagnosis for gliomas. IDH1/2 mutations dominate in WHO grade II/III gliomas (also called as lower grade gliomas) and secondary GBM. According to several studies, IDH1-R132H (G395A) is the most common mutation (90%). The NOA-04 trial also showed that ATRX status and IDH1-R132H might be initially detected for the "integrated decision" in the further [11]. In addition, oligodendrogliomas have the 1p/19q codeletion and TERT promoter mutations phenotype, while TP53 and ATRX

Figure 5: Kaplan-Meier analysis of progression-free survival in patients with ATRX protein loss and expression.
In the cohort of patients with grade II or grade III tumors (A) and (B), patients with ATRX loss had a trend for longer PFS compared with those with ATRX expression. ATRX loss was associated with longer progression-free survival in patients with grade II/III oligoastrocytomas or grade II/III gliomas (C) and (D). Patients with astrocytic tumors losing ATRX had a longer progression-free survival than patients with tumors expressing ATRX (E). In additions, there was a trend that patients with IDH1-R132H and ATRX loss grade II/III gliomas had longer PFS than patients with IDH1-R132H and ATRX positive grade II/III gliomas and PFS of patients with GBM was similar to that of patients with IDH wild type (F).
Oncotarget 16391 www.impactjournals.com/oncotarget LGG, low grade glioma; pA, primary astrocytoma; pO, primary oligodendroglioma; pOA, primary oligoastrocytoma; AG, anaplastic glioma; pAA, primary anaplastic astrocytoma; pAO, primary anaplastic oligodendroglioma; pAOA, primary anaplastic oligoastrocytoma; pGBM, primary glioblastoma; IDH, isocitrate dehydrogenase, ATRX, alpha-thalassemia/mental retardation, X-linked. Abbreviations: PFS, progression-free survival; LGG, low grade glioma; AG, anaplastic glioma; GBM, glioblastoma; IDH, isocitrate dehydrogenase, ATRX, alpha-thalassemia/mental retardation, X-linked.   Oncotarget 16393 www.impactjournals.com/oncotarget mutations predominantly occur in grade II/III astrocytoma and secondary GBM. Maybe, oligoastrocytomas (WHO grade II/III) may be separated into two groups, genetically matching oligodendroglioma on one and astrocytoma on the other side based on the molecular information, for example, IDH1/2, 1p/19q and ATRX and so on [27,28]. The frequency of IDH mutations is rare in primary GBMs that have TERT promoter mutations, EGFR alteration and PTEN loss. Yao et al. analyzed IDH1 and IDH2 status at codon 132 of IDH1 and codon 172 of IDH2 by direct sequencing and anti-IDH1-R132H immunohistochemistry in 53 paired samples and their recurrences [29]. We also found that IDH1-R132H and ATRX expression status of the recurrent tumor was most consistent with that of their initial tumor and that IDH1-R132H positive gliomas had a significantly longer progression-free survival (PFS). In addition, our cohort contained a larger sample size, thus, we assessed the diagnosis value IDH1-R132H and (or) ATRX status for human gliomas and observed that the progressive pattern of gliomas based on the precious cohort [2,11,[30][31][32]. The "Haarlem Consensus Guidelines for Nervous System Tumor Classification" [10] suggest that some entities will require molecular information to provide an "integrated" diagnosis, which is based on several layers comprising (i) the integrated diagnosis as top layer, followed by (ii) histological classification, (iii) WHO grade, and (iv) molecular information. In the present study, for the first time to our knowledge, we illustrated an evaluation formula for the evolution of gliomas by IDH1-R132H combined with ATRX immunohistochemistry and identified the association between IDH1-R132H/ATRX loss and longer progression time interval of gliomas. In addition, we also observed that most recurrences had a consistent IDH1 and ATRX status with their matched primary tumors and demonstrated the progressive pattern of grade II astrocytoma/oligodendroglial tumors and anaplastic oligoastrocytoma with or without IDH1-R132H.  [1]. The distinction between primary and secondary GBM was based on the presence of prior specimens showing lower grade gliomas. All patients provided written informed consent, and the study was approved by the ethics committees of the participating hospitals. Patients' progression-free survival data were recorded when the relapse occurred. Patients who underwent biopsy alone or lost follow-up were excluded from the survival analysis.

Immunohistochemistry for IDH1-R132H and ATRX
Immunostaining was performed according to the manufacturer's protocol. In brief, formalin-fixed, paraffinembedded tissue sections cut to four micrometer were dried at 80°C for 15 min and dewaxed in xylene, rinsed in graded ethanol, and rehydrated in double-distilled water. The sections were then treated with 3% H 2 O 2 for 5 min at room temperature (RT) to block endogenous peroxidase activity. For antigen retrieval, slides were pretreated by steaming in sodium citrate buffer (10 mM sodium citrate, pH 6.0) for 15 min at 100°C. After washing with phosphate-buffered saline for 3 min, the sections were immunostained with an anti-human IDH1-R132H antibody (at 1:60 dilution, H09, Dianova, Hamburg, Germany) or an anti-human ATRX antibody (at 1:800 dilution, ab97508, Abcam), and incubated at 4°C over night. After washed by 3 changes of PBS buffer, the tissues were covered by anti-mouse/rabbit polymer HRP-label for 30 min at RT. Staining reaction was performed through covering tissue by prepared DAB chromogen solution, and incubating approximately for 10 min to allow for proper brown color development. Standard of IDH1-R312H staining: (1) a strong cytoplasmic immunoreaction product was scored positive; (2) a weak diffuse staining and staining of macrophages were not scored positive [33]. Standard of ATRX staining according to German Cancer Research Center (DKFZ) [11]: nuclear ATRX loss was scored as specific if tumor cell nuclei were unstained while nuclei of non-neoplastic cells such as endothelia, microglia, lymphocytes and reactive astrocytes were strongly positive.

Statistical analysis
Receiver operating characteristic (ROC) curves were constructed to determine the discriminatory capacity of IDH1-R132H and (or) ATRX loss for diagnosis. Kaplan-Meier analysis was performed to estimate the survival time of different subgroups and a log-rank test was used to test prognostic differences. Comparisons of binary and categorical patient characteristics between subgroups were performed by the use of the Fisher's exact test All statistical computations were performed with the statistical software environment R version 3.2.0 (http://www.r-project.org/), GraphPad Prism Version 6.01 or Microsoft Excel 2013. P value < 0.05 was considered statistically significant.