Prognostic value of increase in transcript levels of Tp73 ΔEx2-3 isoforms in low-grade glioma patients

Glial tumours are a devastating, poorly understood condition carrying a gloomy prognosis for which clinicians sorely lack reliable predictive parameters facilitating a sound treatment strategy. Tp73, a p53 family member, expresses two main classes of isoforms – transactivatory activity (TA)p73 and ΔTAp73 – exhibiting tumour suppressor gene and oncogene properties, respectively. The authors examined their expression status in high- and low-grade adult gliomas. Isoform-specific real-time reverse transcription-polymerase chain reaction was used for the analysis of Tp73 isoform transcript expression in a series of 51 adult patients harbouring glial tumours, in order to compare tumour grades with each other, and with non-tumoural samples obtained from epileptic patients as well. Our data demonstrate increase of TAp73 and ΔTAp73 transcript levels at onset and early stage of the disease. We also show that ΔEx2–3 isoform expression in low-grade tumours anticipates clinical and imaging progression to higher grades, and correlates to the patients’ survival. Expression levels of P1 promoter generated Tp73 isoforms – and particularly ΔEx2–3 – indeed allow for prediction of the clinical progression of low-grade gliomas in adults. Our data are the first such molecular biology report regarding low-grade tumours and as such should be of help for sound decision-making.

Glioblastomas are divided into primary or de novo, and secondary tumours originating from low-grade (World Health Organization -WHO -grade II) progressing to high-grades, that is anaplastic (WHO grade III) gliomas and eventually to highly malignant glioblastomas (WHO grade IV). Although our knowledge of the molecular mechanisms involved in tumorigenesis and progression has significantly improved over fairly recent years, this has yet to result in an improved therapeutic strategy. To date, no reliable parameter has allowed for anticipation of clinical progression from lower to higher grades. Such criteria would be of paramount importance in therapeutic management of these patients.
The Tp73 gene, a close relative of the tumour suppressor Tp53 gene, contains two independent promoters P1 and P2 (Melino et al, 2002). P1 promoter controls the TAp73 transcripts containing exons 1 -3 that encode N-terminal sequences with transactivatory activity (TA), and DEx2, DEx2 -3 and DN' transcripts collectively designate DTAp73 that lack a fully competent TA domain (Stiewe et al, 2002a, b). The TAp73 isoform is capable of inducing cell cycle growth arrest and apoptosis and is also involved in the response of p53 to death stimuli (Flores et al, 2002). These features are consistent with a tumour suppressor function. In contrast, the DTA transcripts do not induce cell growth arrest or cell death. Significantly, their overexpression in NIH3T3 cells induces malignant transformation and tumour growth in nude mice (Stiewe et al, 2002a, b), which is consistent with an oncogene status. The P2 promoter located downstream of exon 3 0 controls DNp73 transcripts lacking the TA domain that likewise exhibit antiapoptotic activity and are therefore functionally similar to DTAp73 transcripts (Grob et al, 2001;Nakagawa et al, 2002). DNp73 has been shown to facilitate the immortalisation of MEFs and to cooperate with oncogenic Ras in transforming primary fibroblasts in vitro and in inducing MEF-derived fibrosarcomas in vivo in nude mice (Petrenko et al, 2003). Because the tetramerisation domain and the DNA-binding domain of the TAp73 protein are conserved in DNp73 and DTAp73 isoforms, both have been shown to exert a dominant-negative activity on TAp73 and p53 through oligomerisation with the two proteins and/or by competition for p53/TAp73 target genes (Stiewe et al, 2002a, b). Consistent with these data, upregulation of the DNp73 isoform has been described in neuroblastomas with poor prognosis and thereby defines this parameter as a strong and independent predictor of such poor prognosis (Casciano et al, 2002). Moreover Bergamaschi et al (2003) and Irwin et al (2003) have shown that TAp73 is induced by a wide variety of chemotherapeutic agents and that chemosensitivity is related to TAp73 function.
In view of the potential relevance of Tp73 to tumorigenesis, the expression status of the gene has been addressed in various tumour types. Significant results have been achieved only in more recent works taking into account the recently described DTAp73 transcripts by using appropriate couples of primers, designed to specifically assess contribution of the DTAp73 isoforms among the different P1-generated transcripts collectively referred to as P1 transcripts (Ng et al, 2000;Cui et al, 2005). The most convincing evidence to date comes from a work on a series of 100 ovarian tumours of all the known histological types (Concin et al, 2004). On the whole, these data point to the activation of the P1 promoter and the deregulation of P1-controlled DTA p73 isoforms in the tumorigenic process. Furthermore, because the DN'p73 isoform is upregulated in ovarian tumours but not in other cancers, the data argue in favour of a selective deregulation among the P1 transcripts in accordance with types of cancer.
We have undertaken a prospective study in order to determine the involvement of Tp73 isoforms in gliomagenesis and their prognostic relevance. In the first part we performed a global analysis of P1-and P2-generated transcripts on high-and lowgrade gliomas, whereas in the second part we focused on low-grade gliomas and performed a discriminant analysis of P1-generated transcripts.

Patients
Tissues from 51 adult patients harbouring glial tumours were successively collected during surgery at the Department of Neurological Surgery, University of Poitiers, France, with signed informed consent of all patients and the approval of the ethics committee of the Poitou-Charentes area. These patients were free from any past medical history, especially with regards to brain surgery, brain radiation therapy or chemotherapy. Our series also included three tumour-free patients, operated for refractory epilepsy, obtained from Neurobiotec s (Lyon, France) in compliance with the ethics committee. For each patient a sample was collected at the presumed higher-grade region and named the 'central' sample. Whenever deemed reasonable -that is without significant additional functional risk -a second sample was harvested in the vicinity of the tumour and called 'peripheral'.
All patients were treated between January 2002 and May 2003. Distribution of patients in groups according to the WHO pathology classification appears in Table 1. Twenty-nine patients underwent tumour resection, whereas 22 benefited from stereotactic biopsy. The first series, dedicated to Tp73 involvement, included 28 high-grade and 11 low-grade patients (Table 1A), and 12 were added to this group in a second series focusing on Tp73 isoforms (Table 1B). All low-grade patients had non-enhancing tumours on MRI imaging, and pathology-proven low-grade tumours. In all cases progression to a higher grade was pathology-proven. Tumour diagnosis and grading were established according to the WHO criteria (Kleihues and Cavenee, 2000) and were systematically revised by two expert neuropathologists. Results are expressed in DC t which is the threshold cycle differences (CtGAPDH-CtGene) and in 2 ÀDDCt which is the comparative threshold between non-tumoural and tumoural tissue where DDC t ¼ DC tTumoural ÀDC tNonTumoural and represents fold change value between tumoural and non-tumoural tissue. P1 -p73, TAp73, DEx2p73 and DEx2 -3p73 show significantly increased transcript levels as compared to non-tumoural epileptic tissues (Po10 -3, Po10 -4, P ¼ 0.01 and P ¼ 0.03, respectively). There is no difference of expression between low-and high-grade gliomas. (*) Non-tumoural tissue value was based on the analysis of three epileptic patients. The mean of DC t values was considered as normal expression. (ND: not done).

Analysis of Tp73 isoform transcripts
sample was divided into two parts: one was dedicated to smears, the second was immediately frozen in liquid nitrogen in the operating room, and stored at -801C until usage.
RNA isolation and complementary DNA preparation Total RNA was extracted from tumour tissues using the RNAeasy s Mini Kit (Qiagen, Courtaboeuf, Paris, France) according to the manufacturer's instructions with minor modifications, for exclusion of contaminated genomic DNA. The spin-column membranes were treated with DNase (Qiagen) for 15 min at room temperature before elution. Ten microliters of DNAse-treated total RNA was transcribed into cDNA using Superscriptt II RnaseH -and random hexamers (Invitrogene s , Courtaboeuf, Paris, France).
Real-time reverse transcription-polymerase chain reaction and relative quantification We assessed levels of Tp73 isoform mRNA transcripts by real-time quantitative PCR in the ABI PRISM 7000 sequence detection system (Applera, Courtaboeuf, Paris, France). Primer sequences are listed in Figure 1. All primers were queried against the no redundant Human Genome Database. Probe sequence is 5 0 -CAGTTCAATCTGCTGAGCA-3 0 . All primer pairs detected a unique specific cDNA. The forward primers for DEx2p73 as well as for DEx2 -3p73 were designed to specifically recognise on the exon -exon boundaries (exon1/exon3 for DEx2p73 and exon1/exon4 for DEx2 -3p73). Indeed, these primers could not hybridise on TAp73 transcripts as the selected boundaries exist only in the variants forms. The forward primer  for the global analysis of promoter 1 generated transcripts (P1 -p73) and promoter 2 generated transcript (DNp73) and those of isoform-specific amplification of individual NH2 terminus Tp73 isoforms. The forward primers for DEx2p73 as well as for DEx2 -3p73 were designed on the exon-exon boundary (exon1/exon3 for DEx2p73 and exon1/exon4 for DEx2 -3p73). These primers could not hybridise on TAp73 transcripts as the boundary exists only in the variants forms. The forward primer used for specific TAp73 amplification is localised in exon 2 and the reverse on the exon3/exon4 boundary and could only amplifiy TAp73. Specificity for DN'p73 was achieved by the unique combination of the upstream (on exon3) and downstream (on exon 3 0 ) primers.
used for specific TAp73 amplification is localised in exon 2 and the reverse on the exon3/exon4 boundary and could only amplify TAp73. Specificity for DN'p73 was achieved by the unique combination of the upstream (on exon3) and downstream (on exon 3 0 ) primers. However, the specific amplification of TAp73, DEx2p73, DEx2 -3p73 and DN'p73 transcripts was checked on gel and gives as expected a single band of 168, 115, 178 and 90 pb. This specific amplification was confirmed by sequencing the products.
The reactions were carried out first by using the Taqman s chemistry for the global screening of P1-generated p73 transcripts (P1 -p73) and P2-generated transcripts (DNp73). We then used SyberGreen chemistry assessing of discriminant Tp73 isoform expression. Briefly, the PCR reactions were performed in 25 ml reaction volumes consisting of 1 Â Taqman Universal PCR Master-Mix or SyberGreen Universal PCR Master-Mix (Applera), one out of 20 of the reverse transcription reaction, 300 nm of each primer and 200 nm of probe when Taqman chemistry was used. The samples were submitted to amplification as follows: heating at 501C for 2 min, 951C for 10 min followed by 40 cycles at 951C Â 15 s, 601C Â 1 min. Each RNA sample was tested in duplicate and a negative control was included in every plate. The computed tomography value was defined as the cycle number (C t ) at which the fluorescence crossed the threshold. Range of threshold cycles was C t : 25 -37.6 for P1 -TAp73, C t : 26.9 -39.8 for DNp73 C t : 22.89 -39.43 for TAp73, C t : 26.73 -37.33 for DEx2p73, C t : 23.24 -38.75 for DEx2 -3p73 and C t :30.32 -36.36 for DN'p73. Results were normalised using the endogenous reference GAPDH. The threshold cycle differences is given in Table 1. The amplification efficiency of the reference gene was similar to that of the target genes. We employed the relative quantification method described in Applied Biosystems User Bulletin No. 2 (Applied Biosystem User Bulletin, 1997) and by Livak and Schmittgen (2001), in which the amount of target, normalised to the endogenous reference GAPDH and relative to the non-tumoural epileptic tissue, is indicated by the 2 ÀDDCt formula where DDC t ¼ DC tTumoural ÀDC tNonTumoural .

Statistical analysis
The data distribution of expression levels being non-Gaussian, non-parametric tests were used for data analyses. Increase in transcript levels of each Tp73 isoform was analysed by signed tests, whereas differences between tumour tissues were tested by the Wilcoxon rank sum test. Each isoform was dichotomised according to its median value. Survival was estimated according to the Kaplan -Meier method and compared between groups by means of the log-rank test. For correlations among the various isoforms and patient age, the threshold cycle differences were used and the Spearman correlation test was applied. Multivariate analyses were performed with the Cox proportional hazard model. Models were fitted for the threshold cycle differences and for the dichotomised variable. All tests were two-sided and the type I error was set at 5%. Statistical analyses were performed with use of SAS Statistical Software (Release 8.02).

RESULTS
Global analysis of P1 -(P1 -p73) and P2 -(DNp73) generated transcript expression levels conducted in 11 patients shows that P1 -p73 transcript levels are increased in low-grade gliomas (Po10 À3 ) when compared to epileptic tissue samples taken as non-tumoural controls (Table 1C). In contrast, tumour samples displayed no increase of DNp73 transcript levels. The same observation was carried out in 28 high-grade gliomas (Po10 À3 ). Because no significant difference was found in P1 -p73 transcript expression between high-and low-grade tumours, deregulation of the P1 promoter that generates the TAp73 and DTAp73 isoforms was likely to occur early in the tumorigenic process. This upregulation of the P1 -p73 transcripts was observed in the centre of the tumours whatever their grade. In contrast, although the high-grade tumour periphery still displayed high levels of the transcript, this expression was significantly (P ¼ 0.031) lower on the periphery of low-grade tumours (Table 2).
On the whole, these data suggest that low-and high-grade tumours can be distinguished on the basis of their P1 -p73 transcript expression status in their central and peripheral areas. In addition this selective expression of P1 -p73 between these areas in low-grade tumours argues in favour of an evolving process.
Analysis of P1-p73 transcripts (TAp73, DEx2p73 and DEx2 -3p73 and DN'p73) shows that TAp73, DEx2p73 and DEx2-3p73 transcripts levels were increased (Po0.05) as compared to the expression spectrum displayed by non-tumoural, control tissue (Table 1C and Figure 2). On the contrary, expression of the DN' isoform in tumour samples did not appear to differ from that of controls. In addition, expressions of these different isoforms were correlated to each other as significant associations were observed between TAp73 and DEx2p73 levels (Po10 À2 ) and TAp73 and DEx2 -3p73 levels (Po10 À3 ). These data led to the conclusion that, with the exception of the DN' isoform, all P1 transcripts were upregulated in the tumours. Again, these findings argued in favour of the early deregulation of the P1 promoter in tumours.
Differences in the expression levels were observed within tumoural and non-tumoural tissue samples. This observation was particularly noticeable for TAp73 and DEx2 -3p73 transcript levels with a variation of 430 000-fold and 410 000-fold within this groups (lowest versus highest expression). A similar observation was reported by Concin et al, 2004 concerning ovarian carcinomas where DEx2 -3p73 show in normal tissues a variation of 47000-fold from the lowest to highest measured. One possible explanation for this variation range is that this variation may reflects perceptibly the heterogeneity of the tumours, patient-topatient variability or both. Results are expressed in DC t which is the threshold cycle differences (CtGAPDH-CtGene) and in 2 ÀDDCt which is the comparative threshold between non-tumoural and tumoural tissue where DDC t ¼ DC tTumoural ÀDC tNonTumoural and represents fold change value between tumoural and non-tumoural tissue. There is a statistically significant difference for P1 -p73 expression between the periphery of the tumour and the centre of the tumour for low-grade gliomas (P ¼ 0.031 in Wilcoxon test).
Tp73 DEx2-3 isoforms in low-grade glioma patients M Wager et al In order to determine the clinical relevance of these findings with regards to low-grade glioma outcome, survival analysis was performed as regards overall survival rates. Taking as references the median values of mRNA levels of different transcripts -above and below the 50th percentile of the range of the latter -two groups were defined. This approach was dictated by the impossibility of estimating the absolute copy numbers of transcripts. Group 1 included patients in whom the expression levels for each transcript species were higher than their respective median values. Group 2 included patients whose expression levels were lower or equal to the median values. Among the isoforms, expression levels of the DEx2 -3p73 transcript proved to be a strong prognostic marker for overall survival in our low-grade glioma cohort (P ¼ 0.007, Figure 3A). Although 10 out of 12 (83%) DEx2 -3p73 low expressor patients (group 2) were still alive as of the last follow-up, nine out of eleven (82%) of the high expressor (group 1) patients were deceased. The median overall survival time for high expressor patients was 13 months (IC95% 4 -20) whereas median survival time for low expressor patients was not reached. The high expresssor patients (group 1) had a relative amount of DEx2-3p73 mRNA 422-fold compared with non-tumoural epileptic tissue whereas patients from group 2 had a relative amount of DEx2 -3p73 mRNA o ¼ 22-fold as compared to non-tumoural epileptic tissue.
In multivariate analysis including age and DEx2 -3p7 expression, age is a significant prognostic factor (P ¼ 0.017) and a trend is observed for DEx2 -3p73 (P ¼ 0.070). However, when DEx2 -3 expression is classified in low and high groups, the only factor that remains significant, even when age is taken into account, is DEx2 -3p73 (P ¼ 0.025).
A similar observation was also carried out, but to a lesser extent, for DEx2p73 as high expressor patients had a worse outcome than did low expressor patients (P ¼ 0.043 and Figure 3B). In contrast, TAp73 and DN'p73 transcripts did not appear to be involved as the survival rates in the two groups were similar (P ¼ 0.144 and P ¼ 0.906, respectively Figure 3C and D).
As we know that the neutralising role of these transdominant forms take place through oligomerisation with TAp73 and p53 proteins it is crucial to confirm that RNA overexpression of the pronostic splice variants as far as proteins are concerned. Unfortunately no specific antibodies for any of the DTAp73 forms currently exist. We cannot therefore propose the existence of a dominant negative mechanism with DEx2 -3p73 and DEx2p73 variants in low-grade gliomas. However, as mRNA expression of DEx2 -3p73 and to a lesser extent DEx2p73 appeared to be correlated to low-grade gliomas patients survival and to be a strong prognostic marker in these tumours, our data, although suggestive on the nefastious role of these transcripts, do not allow to definitely conclude on this point.

Tp73 involvement in glial tumorigenesis
The role of the Tp73 gene in oncogenic process has been underlined by several reports dealing with different tumour types. Results including ours have clearly shown that the expression of the gene rises frequently in a number of tumours although considerable variations in the levels of transcripts from patient to patient can be noted (Concin et al, 2004). A previous work dealing with gliomas and using a semiquantitative RT-PCR assay has shown that a similar situation also exists in this type of tumour as overexpression of TAp73 mRNA occurred in high-grade gliomas whereas only a few tumours displayed DNp73 expression (Ugur et al, 2004). The present study confirms these data and extends them by reporting increase in transcript levels of the P1-generated Tp73 transcripts, namely TAp73, DEx2p73 and DEx2-3p73, (but not DN'p73) in a cohort of low-grade adult gliomas. Interestingly DNp73 transcript does not appear to differ from that of control samples. Thus, deregulation of the P1 promoter and emergence of particular RNA populations are likely to be an early event as this deregulation is likewise observed in high-grade gliomas and could possibly influence cell evolution toward malignancy. The abundance of Tp73 mRNA isoforms in specific cell type is complex and is likely to result from differential expression, RNA stability and splicing. In our study, increase in transcript levels of P1-generated isoforms could be explained by coupled mechanisms: activation of the P1 promoter itself and selective splicing process leading to an increase of certain mRNA species (TAp73, DEx2p73 and DEx2 -3p73 but not DN'p73). The expression of some transcripts could be restricted to particular areas of the tumour suggesting differential splicing in this area. Addressing this question would require to perform RNA status analysis of at the individual tumour cells level.
The selective patterns of expression of the transdominantTp73 gene isoforms observed in a variety of tumours (Ng et al, 2000;Cui et al, 2005), should reflect a particular distribution of the factors that contribute to splicing such as HnRNP proteins. Incidentally, several reports have shown the existence of particular populations of HnRNP proteins in a variety of tumours suggesting that their expression is implicated in selective activation of gene isoforms expression (Carpenter et al, 2006).
Tumour-specific variants have been reported to affect transcript stability and thus, accumulation of the variant transcripts could result from an increased stability. Regarding this point, impaired mRNA turnover and stability are shown to play a critical role in the activation of specific genes during the cellular response to mitogens, immunological triggers stressful stimuli and differentiation agents (Bashirullah et al, 2001;Fan et al, 2002). Several parameters control the mRNA turnover: RNA-binding proteins (RBPs) that bind to specific RNA sequences and either increase or decrease the transcript half-life (Brennan and Steitz, 2001;); regions located in the 3 0 -untranslated regions (3 0 -UTRs) of mRNA. In the latter, differences in 3 0 -UTRs of splice variants could be recognised by proteins expressed specifically in some normal tissues or tumoural tissues or in a particular areas of the tumour (Jensen and Whitehead, 2004). Finally, the RBP HuR, upon binding to RNA HuR, has been shown to stabilise it and alter its translation (Brennan and Steitz, 2001;Mazan-Mamczarz et al, 2003).
Another aspect is the differential expression of Tp73 transcripts in the peripheral and central areas of low-grade tumours as compared to high-grade gliomas that overexpress the gene in both areas. Beside the possibility of discriminating between high-and low-grade tumours, this result suggests that malignant cells from the periphery of low-grade gliomas differ from those located in the tumour centre. By using the two dimensional profiling strategy Li et al (2006) found a large number of genes that showed coordinated changes in transcript abundance and splicing, indicating that many distinct steps in gene expression from transcription, stability control and splicing may be coupled in different cell types (Li et al, 2006). In view of this observation it is tempting to speculate that the peripheral cells in low-grade tumours that differ from their central area congeners may evolve toward a more malignant stage or disappear to be replaced by tumour cells coming from the centre. We are considering the possibility of addressing these hypothesis by carrying out transcriptome analysis of tumour areas isolated by microdissection.

Tp73 isoform prognostic value
A survival analysis aimed at determining the clinical relevance of the increase in DTAp73 isoform transcript levels in low-grade adult gliomas showed that patients with higher DEx2 -3p73 contents in their tumours presented shorter survival than those with lower amounts of this transcript. Strikingly, increase in the TAp73 isoform transcript levels was also found in all patients, thereby arguing in favour of a dominant negative role for the DEx2 -3p73 isoform in survival. This situation mimics that of ovarian tumours where a better overall survival was noted for patients exhibiting low expression of the DN'p73 isoform than for those with high expression, whatever the levels of TAp73 in the tumours (Concin et al, 2004). Moreover, DEx2 -3p73 is a prognostic factor in low-grade gliomas even when age is considered.
Up until now, the status of the Tp73 gene involving its various isoforms has been examined in ovarian tumours, hepatocarcinomas, vulval cancers, oesophagus and gliomas (this report) (Ng et al, 2000;Cui et al, 2005). Although these studies point to increase in transcript levels of transdominant DTAp73 isoforms, it may be noted that, depending on the tumour pathology, a selective expression of one of the transdominant isoforms appears to be detected in certain tumour types, supporting the notion that among the various P1-generated isoforms, differential regulation exists. Given the necessarily limited number of patients included in this study, our data strongly suggest that the DEx2 -3p73 isoform is selectively expressed in gliomas. It will be of interest to extend these observations to other tumour types in order to determine whether a more specific expression spectrum exists. Obviously, such selective patterns should reflect an abnormal distribution in tumour cells of the factors that contribute to splicing. In any case, whatever the identity of the expressed transdominant DTAp73 isoform, the biological significance directly relies on the neutralising role of this transdominant form to TAp73 and/or p53.
Although the gross natural history of adult low-grade glioma is known, a given individual outcome remains unpredictable. This lack of individual predictive factors results in daunting difficulties once therapeutic decisions are involved. When a low-grade glioma is newly diagnosed, clinicians have to decide when to proceed to treatment: not too early because therapeutic means -extensive surgical removal, radiation therapy, chemotherapy -are contestable as pertains to a benign tumour, but not too late either, that is before progression to higher grades of these infiltrating tumours. Up until now the indicators of such a change are clinical symptoms and imaging changes (e.g. contrast enhancement) but they only accompany, and fail to anticipate clinical tumour progression. New criteria allowing for earlier detection of tumour grade progression and correspondingly more expeditious therapeutic adaptation would be of heightened value as regards optimal treatment of these patients. Nevertheless, the use of DEx2 -3p73 as such an indicator has one limitation, which is inherent to our methodology: as there exists no absolute quantification of this increase in transcript levels, it has got to be compared with a non-tumoural control group.
In conclusion, to the best of our knowledge this is the first report of a single molecular prognostic marker criterion for lowgrade glioma survival in adults. This could lead to early treatment of the newly diagnosed adult patients harbouring lowgrade gliomas and exhibiting increase in transcript levels of DEx 2 -3p73.