Home About us Editorial board Ahead of print Current issue Search Archives Submit article Instructions Subscribe Contacts Login 

 Table of Contents  
ORIGINAL ARTICLE
Year : 2020  |  Volume : 16  |  Issue : 1  |  Page : 7-12

Xist noncoding RNA could act as a tumor suppressor gene in patients with classical Hodgkin's disease


Institute of Electronics, Computer and Telecommunication Engineering, National Research Council of Italy, Genoa, Italy

Date of Submission23-Sep-2016
Date of Decision20-Aug-2017
Date of Acceptance24-Feb-2018
Date of Web Publication23-Jul-2018

Correspondence Address:
Stefano Parodi
Institute of Electronics, Computer and Telecommunication Engineering, National Research Council of Italy, Via De Marini, 6, 16149 Genoa
Italy
Login to access the Email id

Source of Support: None, Conflict of Interest: None


DOI: 10.4103/jcrt.JCRT_1055_16

Rights and Permissions
 > Abstract 


Background: Xist is a long noncoding RNA involved in the X chromosome inactivation in females. It may act as an onco-suppressor gene in hematologic malignancies, and its activity is strongly dependent from SATB1 gene expression. However, its potential role in Hodgkin's disease (HD) onset and progression is unknown.
Materials and Methods: Three gene expression microarray datasets were analyzed for the expression of Xist and SATB1 in patients with classical HD, namely, GDS4222 (130 patients and 54,000 gene features), GSE39134 (29 patients and 54,000 features), and E-MEXP-507 (29 patients and 27,648 probes). The first two were oligonucleotide arrays (platform: Affymetrix gene chip HG-U133-Plus2), whereas the latter was a cDNA two-channel array (platform: OncoChip. v2). Summary and time-dependent receiver operating characteristic (ROC) analysis were applied to obtain a summary measure (summary area under the ROC curve [sAUC]) of association between gene expression and unfavorable patient outcome in each probe set.
Results: Xist was overexpressed among females in each data set. A slight overexpression was associated with a good prognosis both in males (sAUC = 0.75, 95% confidence interval [CI]: 0.70–0.80) and at a lesser extent, in females (sAUC = 0.64, 95% CI: 0.59–0.69). However, this finding was limited to the analysis of the biggest database (GDS4222). No association was found between Xist and SATB1 expression.
Conclusions: A reactivation of Xist might act as an onco-suppressor gene in male patients with HD, which seems independent from SATB1 expression. The possibility that Xist could contribute to the better survival of female patients should also be investigated.

Keywords: Cancer prognosis, gene expression, Hodgkin's disease, microarray, SATB1, Xist


How to cite this article:
Parodi S. Xist noncoding RNA could act as a tumor suppressor gene in patients with classical Hodgkin's disease. J Can Res Ther 2020;16:7-12

How to cite this URL:
Parodi S. Xist noncoding RNA could act as a tumor suppressor gene in patients with classical Hodgkin's disease. J Can Res Ther [serial online] 2020 [cited 2020 Jul 12];16:7-12. Available from: http://www.cancerjournal.net/text.asp?2020/16/1/7/237380




 > Introduction Top


Xist is a noncoding long RNA involved in the initiation of X chromosome inactivation in embryonic cells of female mammals.[1] After this phase, X chromosome inactivation becomes permanent due to epigenetic mechanisms involving DNA methylation and histone deacetylation.[2]

Many studies have associated Xist expression with tumorigenesis and tumor progression in solid cancers, where it seems that Xist could have different roles as an oncogene or a tumor suppressor.[3],[4] For instance, a potential role as oncogene was observed in small cell lung cancer, hepatocellular carcinoma, testicular cancer, and glioblastoma.[3],[5],[6],[7] Conversely, a tumor-suppressive activity was found in breast and ovarian cancer.[8],[9]

With regard to hematologic malignancies, some studies have provided evidence of a potential role of Xist as an onco-suppressor gene. For instance, depletion of Xist in blood stem cells in mice can induce a highly lethal blood cancer. The cancer is female-specific and fully penetrant.[10] Furthermore, tetracycline-inducible Xist transgenes expression in thymic lymphoma was revealed to cause tumor block by X chromosome silencing in mice. In the same model, the AT-rich binding protein SATB1 was found to be essential for the Xist -mediated silencing.[11]

The potential role of Xist in the development and progression of Hodgkin's disease (HD) is unknown. A recent analysis by supervised machine learning (decision tree) of a large data set of gene expression profiles reported a poorer prognosis in patients with classical HD with a low expression of one out of the eight available Affymetrix probes.[12]

The present study is aimed at evaluating the association between Xist and the prognosis of HD patients using data from three publicly available microarray databases. The correlation between Xist and SATB1 expression was also evaluated.


 > Materials and Methods Top


Database selection

Gene expression microarray experiments including data on classical HD patients were identified by extensive research on PubMed Central database. The following keywords were employed for paper selection: (a) “Hodgkin lymphoma” or “Hodgkin disease;” (b) “gene expression;” and (c) “survival” or “prognosis” or “relapse” or “progression.” The search was restricted to articles published in English between January 2000 and June 2016.

A flow diagram illustrating the paper selection is shown in [Supplementary Figure 1]. The first group of 289 papers was selected, which included 152 articles not involving classical HD (mainly focused on non-Hodgkin's lymphoma), 29 cell line experiments, 5 micro RNA (miR) experiments and 87 researches not involving Xist gene expression. The remaining 16 papers included 11 repetitions (7 reviews, two commentaries and two reanalyzes), thus leaving five papers, corresponding to five available databases.



The largest data set (GDS4222) was the same analyzed by Parodi et al .[12] using supervised data mining techniques, and included 130 HD samples and >54,000 gene expression profiles.[13] Patient samples were withdrawn before medical treatments. First-line chemotherapy included doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) or an ABVD-like regimen. Radiotherapy was also administered when indicated. Data were downloaded from the GEO repository at the http://www.ncbi.nlm.nih.gov website.

Gene expression in GDS4222 database was assessed using the oligonucleotide Affymetrix gene chip HG-U133-Plus2.[13] Eight probes for Xist expression were available, namely, 214218_s_at , 221728_x_at , 224588_at , 224589_at , 224590_at , 227671_at , 235456_at , and 243712_at . Two SATB1 probes were also included in the microarray, namely, 203408_s_at and 241365_at .

Available data were already normalized by Robust multiarray analysis (RMA) using the RMA function in the affy package, while batch-normalization was applied using ComBat.[14]

The GSE39134 dataset was produced by the same authors of the GDS4222 using the same Affymetrix HG-U133-Plus2 platform.[15] However, contrarily to the previous study, gene expression analysis was performed on microdissected Hodgkin's and Reed–Sternberg cells (HRS). Data included 29 samples from HIV negative patients before treatment, of which 14 eventually relapsed. All patients were treated by systemic chemotherapy ABVD. Gene expression data were normalized by RMA using Bioconductor.[16] Data were downloaded from: Http://www.ncbi.nlm.nih.gov.

E-MEXP-507 is a nonnormalized database of cDNA two-channel arrays (platform: “OncoChip. v2,” Spanish National Cancer Centre, CNIO, Human Oncochip version 2.0)[17] that includes 27,648 features and 29 samples, of which 15 from patients with unfavorable outcome. Samples were drawn before treatment from patients with advanced HD (Stage II with B symptoms or bulky masses, Stage III and Stage IV), negative HIV status and ABVD treatment or similar regimens.[17] Gene expression included two clones for Xist (namely: hAD0994 in four replicates and hAC2274 in two spots), and two clones for SATB1 ( hAC9237 and hAA8112 , both duplicated).[18] Data were downloaded from the Array Express dataset: http://www.ebi.ac.uk/arrayexpress/experiments/E-MEXP-507.

With regard to the other two databases selected, they were not included in the analysis, due to either the lack of information about patient gender[19] or the too small sample size (21 patients and 6 events).[20]

For each analyzed dataset, details about patients selection, their characteristics at diagnosis, assessment of disease status, primary line treatment, and methods for gene expression analysis have been reported in the cited papers.[13],[15],[17]

Array normalization

Normalized data from GDS4222 and GSE39134 databases were available in GEO dataset as described in the previous paragraph. Accordingly, they were analyzed without applying any further preprocessing procedure.

With regard to E-MEXP-507 expression values, the log-ratio of the background subtracted channel intensity was normalized using the median value of each array.

[Supplementary Table 1] shows some patient characteristics and the proportion of invalid data in each array of E-MEXP-507. The proportion of invalid spots was very low (<5%) in all except five samples (namely, HL05, HL07, HL08, HL9, and HL14); however, it never exceeded 20%.



Statistical analysis was performed using the median of the expression value for the four spots of hAD0994 ( Xist ) and the two spots of hAC2274 ( Xist ), hAC9237 ( SATB1 ), and hAA8812 ( SATB1 ).

Statistical analysis

Descriptive statistics were reported as absolute frequencies and percentages for qualitative data, while for quantitative ones median and interquartile range were used due to the nonnormal distribution of most variables. Bivariate association between expression profiles was evaluated using the nonparametric Spearman's rho correlation coefficient.

For each patient, a favorable outcome was defined as the absence of relapse or progression at the end of follow-up in the GDS4222 and GSE31134 databases, and a favorable response to therapy (i.e., complete remission within 12 months from the first-line treatment) for E-MEXP-507 data set.

Comparison between expression values by gender and by patient outcome was performed using the Mann-Whitney U test and the area under the receiver operating characteristic (ROC) curve (AUC). In the presence of gene expressions inversely correlated with patients' outcome, ROC analysis was carried out considering as positive those expression values lower than each selected threshold, to obtain meaningful AUC values (i.e., >0.5). Accordingly, in that case, given two patients with different outcomes, AUC represents the probability of a better prognosis for the patient with the lower gene expression value.[21] The effect of time of follow-up, when available (dataset GSE39134), was taken into account applying the method of the time-dependent ROC curves.[22] A summary measure of the effect of Xist and SATB1 expression in each probe set was then obtained estimating the corresponding summary AUC (sAUC).



Where K is the number of probes, and the weights W i were obtained by the reciprocal of the AUCi variance, calculated using the method by DeLong et al .[23]

Time-dependent ROC analysis was carried out by the package timeROC, implemented in R language (version 3.2.2).[22] All the other analyses were performed using Stata for Windows statistical package (release 12.1, Stata Corporation, College Station, TX, USA). All statistical tests were two-sided and a P value lower than 0.05 was considered as statistically significant.


 > Results Top


Main patient characteristics are resumed in [Table 1]. In each data set, the majority of subjects were males. Stage II according to the Ann Arbor staging system was the most represented. No patient in Stage I was observed in the E-MEXP-507 database according to the inclusion criteria that limited the analysis to advanced HD.[17] International Prognostic Score, available in the GDS4222 database only, was low in 75% of patients. A poor outcome was observed in 29% of subjects in the GDS4222 data set, 48% in GSE39134 and 52% in E-MEXP-507.
Table 1: Demographic and clinical characteristics of patients with Hodkin's disease included in the analyses

Click here to view


Pairwise correlation between the probes of Xist and SATB1 are reported in [Supplementary Table 2], [Supplementary Table 3], [Supplementary Table 4] for the GDS4222, GSE39134 and E-MEXP-507 databases, respectively. No statistically significant association was found between Xist and SATB1 , except a small inverse correlation in the E-MEXP-507 data between hAD0994 Xist probe and the two SATB1 clones. Pairwise comparison between Xist probes was very high, but with two exceptions, namely, 235446_at and 243712_at probes in GSE39134 dataset [Xist7 and Xist8 in [Supplementary Table 3]. Correlation between the two SATB1 probes was high in both GDS4222 and GSE39134 databases [ρ = 0.77, [Supplementary Table 2] and [Supplementary Table 3], respectively] and very high in E-MEXP-507 [ρ = 0.90, [Supplementary Table 4].



[Table 2] shows the analysis of the expression of Xist and SATB1 probes by sex. Xist was clearly underexpressed in males in all comparisons with the only exception of two probes in GSE39134 database (namely, 235446_at and 243712_at) where differences were small and statistically significance not reached. No clear difference in SATB1 expression emerged by gender in any analysis.
Table 2: Comparison between the distribution of the expression profiles of the Xist and SATB1 probes by sex and data sets

Click here to view


[Table 3] shows the common estimates of association (sAUC) between Xist and SATB1 expression and the probability of a favorable outcome. Among males, a higher Xist expression was associated with a better outcome in GDS4222 database (sAUC = 0.750, 95% confidence interval [CI]: 0.703–0.797). In the other two datasets, no association was observed, but analyses were based on small sample size. Among females a small association was observed in the GDS4222 dataset (sAUC = 0.641, 95% CI: 0.588–0.694). No effect emerged from the analysis of the GSE-MEXP-507 data set, whereas the small sample size prevented to analyze data from GSE39134. With regard to SATB1 , in males a small borderline association was observed in GDS4222 dataset (sAUC = 0.603, 95% CI: 0.508–0.699). Conversely, 1-year event-free survival among GSE39134 males was inversely associated with SATB1 expression (sAUC = 0.230, 95% CI: 0.051–0.410), but results were based on only 19 patients, and 10 observed relapses. No association was found in females.
Table 3: Summary area under the receiver operating characteristic curve and related 95% confidence intervals of receiver operating characteristic curves for Xist and SATB1 profiles

Click here to view


A more detailed analysis of the association between the expression of each probe of Xist and the probability of a good outcome for the GDS4222 database is reported in [Supplementary Table 5] and in [Supplementary Figure 2]. Among males, a statistically significant association between Xist and cancer prognosis was found for six out of the eight probes, and among females in four probes.




 > Discussion Top


In the present study, a clear overexpression of Xist in female patients with HD was consistently observed in the analysis of data from all the three considered data sets. This finding is not surprising because in human adults Xist expression is normally observed only in female gender.[1] However, reactivation of Xist was revealed in cancer cells in males with nonsmall cell lung cancer or testicular cancer.[5],[6] In this latter, a potential oncogenic role of Xist was related to the inactivation of supernumerary X chromosomes.[24] In contrast to this observation, in the present study, a slightly higher expression of Xist among males was associated to a better prognosis, even if this finding was limited to probes in the biggest database (GDS4222). In the same analysis, a small protective effect among females was also observed.

It could be speculated that reactivation of Xist transcription among male HD patients could favor the killing of HRS or some specific cellular type in the tumor microenvironment by repressing the activity of the unique X chromosome. Other mechanisms of tumor suppression mediated by Xist have also been proposed that might partly explain the better prognosis reported among females with HD in many investigations.[25] For instance, Xist RNA was found to interact with miR-92b in hepatocellular carcinoma, thus inhibiting cancer progression mediated by its direct target Smad7.[3] The interaction between Xist and miR ( miR-152 transcripts) was also observed in human glioblastoma stem cells, but in this case, favoring tumor progression.[7] Furthermore, in nonsmall cell lung cancer Xist transcript was revealed to inhibit the expression of KLF2 , a gene involved in the down-regulation of cell proliferation, directly binding the catalytic subunit EZH2 of an RNA binding protein (namely, the polycomb repressive complex 2), indicating that Xist could also regulate underlying targets at transcriptional levels.[5] Finally, Xist expression might modify the response of tumor cells to some specific chemotherapeutic agents, as pointed out by an investigation on ovarian cancer that found an inverse correlation between Xist RNA levels and the response to Taxol, but not to other therapeutic drugs (carboplatin and cisplatin).[8]

The potential mechanism of reactivation of Xist in cells of HD male patients remains to be elucidated because it is not consistent with the hypothesis that at least two X chromosomes must be present to trigger the Xist -mediated silencing.[2] Furthermore, contrarily to evidence from a lymphoma model,[11] no association was observed between Xist and SATB1 expression in the three datasets analyzed. SATB1 is an AT-rich protein involved in the regulation of chromatin structure. It is expressed in embryonic stem cells and in thymic lymphoma, where it was revealed to be necessary for the Xist mediated killing of malignant cells in mice.[4],[11] However, evidence was limited to T cell lymphoma, whereas HD is a malignancy of mature B cells, originating from the germinal center or post-germinal center B lymphocytes.[26]

A previous investigation by supervised machine learning identified a Xist probe among 54,000 expression profiles in the GDS4222 database that was inversely associated with patient survival.[12] However, the statistical significance of such an effect was not assessed, and the analysis did not take into account differences between males and females. In the present study Xist effect was observed in six Xist probes in males and at a lesser extent, in four probes in females in the same data set, but not in E-MEXP-507 and GSE39134. The large variability of gene expression data from microarray experiments and the small sample size of E-MEXP-507 and GSE39134 databases could result at least in part account for the observed discrepancies. However, other major limits of the present investigation should be considered, in particular, the lack of quantitative measures of Xist and SATB1 transcripts by independent reliable techniques as quantitative real-time polymerase chain reaction. Furthermore, different methods of microarray preparation were employed (single labeled oligonucleotides vs. dual-labeled spotted arrays) and in one out of three experiments (dataset GSE39134) microdissected HRS cells were analyzed instead of tissue microarrays that generally include a variable proportion of neoplastic cells.[15]


 > Conclusion Top


In spite of the above cited limits and even if not fully consistent, results of the present study indicate that reactivation of Xist could act as an oncosuppressor gene in males with HD. This activity seems to be independent from SATB1 gene expression. The possibility that Xist could contribute to the better survival commonly observed among female patients with HD should also be investigated.

Acknowledgment

Stefano Parodi is a research fellow of the Italian MIUR Flagship project InterOmics.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.



 
 > References Top

1.
Payer B, Lee JT. X chromosome dosage compensation: How mammals keep the balance. Annu Rev Genet 2008;42:733-72.  Back to cited text no. 1
    
2.
Wutz A, Jaenisch R. A shift from reversible to irreversible X inactivation is triggered during ES cell differentiation. Mol Cell 2000;5:695-705.  Back to cited text no. 2
    
3.
Zhuang LK, Yang YT, Ma X, Han B, Wang ZS, Zhao QY, et al. MicroRNA-92b promotes hepatocellular carcinoma progression by targeting smad7 and is mediated by long non-coding RNA XIST. Cell Death Dis 2016;7:e2203.  Back to cited text no. 3
    
4.
Agrelo R. X inactivation and progenitor cancer cells. Cancers (Basel) 2011;3:2169-75.  Back to cited text no. 4
    
5.
Fang J, Sun CC, Gong C. Long noncoding RNA XIST acts as an oncogene in non-small cell lung cancer by epigenetically repressing KLF2 expression. Biochem Biophys Res Commun 2016;478:811-7.  Back to cited text no. 5
    
6.
Kawakami T, Okamoto K, Ogawa O, Okada Y. XIST unmethylated DNA fragments in male-derived plasma as a tumour marker for testicular cancer. Lancet 2004;363:40-2.  Back to cited text no. 6
    
7.
Yao Y, Ma J, Xue Y, Wang P, Li Z, Liu J, et al. Knockdown of long non-coding RNA XIST exerts tumor-suppressive functions in human glioblastoma stem cells by up-regulating miR-152. Cancer Lett 2015;359:75-86.  Back to cited text no. 7
    
8.
Huang KC, Rao PH, Lau CC, Heard E, Ng SK, Brown C, et al. Relationship of XIST expression and responses of ovarian cancer to chemotherapy. Mol Cancer Ther 2002;1:769-76.  Back to cited text no. 8
    
9.
Huang YS, Chang CC, Lee SS, Jou YS, Shih HM. Xist reduction in breast cancer upregulates AKT phosphorylation via HDAC3-mediated repression of PHLPP1 expression. Oncotarget 2016;7:43256-66.  Back to cited text no. 9
    
10.
Yildirim E, Kirby JE, Brown DE, Mercier FE, Sadreyev RI, Scadden DT, et al. Xist RNA is a potent suppressor of hematologic cancer in mice. Cell 2013;152:727-42.  Back to cited text no. 10
    
11.
Agrelo R, Souabni A, Novatchkova M, Haslinger C, Leeb M, Komnenovic V, et al. SATB1 defines the developmental context for gene silencing by xist in lymphoma and embryonic cells. Dev Cell 2009;16:507-16.  Back to cited text no. 11
    
12.
Parodi S, Manneschi C, Verda D, Ferrari E, Muselli M. Logic learning machine and standard supervised methods for Hodgkin's lymphoma prognosis using gene expression data and clinical variables. Health Informatics J 2018;24:54-65.  Back to cited text no. 12
    
13.
Steidl C, Lee T, Shah SP, Farinha P, Han G, Nayar T, et al. Tumor-associated macrophages and survival in classic Hodgkin's lymphoma. N Engl J Med 2010;362:875-85.  Back to cited text no. 13
    
14.
Johnson WE, Li C, Rabinovic A. Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics 2007;8:118-27.  Back to cited text no. 14
    
15.
Steidl C, Diepstra A, Lee T, Chan FC, Farinha P, Tan K, et al. Gene expression profiling of microdissected Hodgkin Reed-Sternberg cells correlates with treatment outcome in classical Hodgkin lymphoma. Blood 2012;120:3530-40.  Back to cited text no. 15
    
16.
Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, et al. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 2003;4:249-64.  Back to cited text no. 16
    
17.
Sánchez-Aguilera A, Montalbán C, de la Cueva P, Sánchez-Verde L, Morente MM, García-Cosío M, et al. Tumor microenvironment and mitotic checkpoint are key factors in the outcome of classic Hodgkin lymphoma. Blood 2006;108:662-8.  Back to cited text no. 17
    
18.
Tracey L, Pérez-Rosado A, Artiga MJ, Camacho FI, Rodríguez A, Martínez N, et al. Expression of the NF-kappaB targets BCL2 and BIRC5/Survivin characterizes small B-cell and aggressive B-cell lymphomas, respectively. J Pathol 2005;206:123-34.  Back to cited text no. 18
    
19.
Chetaille B, Bertucci F, Finetti P, Esterni B, Stamatoullas A, Picquenot JM, et al. Molecular profiling of classical Hodgkin lymphoma tissues uncovers variations in the tumor microenvironment and correlations with EBV infection and outcome. Blood 2009;113:2765-3775.  Back to cited text no. 19
    
20.
Devilard E, Bertucci F, Trempat P, Bouabdallah R, Loriod B, Giaconia A, et al. Gene expression profiling defines molecular subtypes of classical Hodgkin's disease. Oncogene 2002;21:3095-102.  Back to cited text no. 20
    
21.
Pepe MS. The Statistical Evaluation of Medical Tests for Classification and Prediction. Oxford: Oxford University Press; 2003.  Back to cited text no. 21
    
22.
Blanche P, Dartigues JF, Jacqmin-Gadda H. Estimating and comparing time-dependent areas under receiver operating characteristic curves for censored event times with competing risks. Stat Med 2013;32:5381-97.  Back to cited text no. 22
    
23.
DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics 1988;44:837-45.  Back to cited text no. 23
    
24.
Weakley SM, Wang H, Yao Q, Chen C. Expression and function of a large non-coding RNA gene XIST in human cancer. World J Surg 2011;35:1751-6.  Back to cited text no. 24
    
25.
Hasenclever D, Diehl V. A prognostic score for advanced Hodgkin's disease. International prognostic factors project on advanced Hodgkin's disease. N Engl J Med 1998;339:1506-14.  Back to cited text no. 25
    
26.
Weniger MA, Barth TF, Möller P. Genomic alterations in Hodgkin's lymphoma. Int J Hematol 2006;83:379-84.  Back to cited text no. 26
    



 
 
    Tables

  [Table 1], [Table 2], [Table 3]



 

Top
 
 
  Search
 
Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
Access Statistics
Email Alert *
Add to My List *
* Registration required (free)

  >Abstract>IntroductionMaterials and Me...>Results>Discussion>Conclusion>Article Tables
  In this article
>References

 Article Access Statistics
    Viewed949    
    Printed223    
    Emailed0    
    PDF Downloaded230    
    Comments [Add]    

Recommend this journal


[TAG2]
[TAG3]
[TAG4]