Contents lists available atScienceDirect
Experimental and Molecular Pathology
journal homepage:www.elsevier.com/locate/yexmp
Expression analysis of NF- κ B interacting long noncoding RNAs in breast cancer
Sepideh Dashti
a, Soudeh Ghafouri-Fard
a,⁎, Farbod Esfandi
b, Vahid Kholghi Oskooei
c,d, Shahram Arsang-Jang
e, Mohammad Taheri
f,⁎aDepartment of Medical Genetics, Shahid Beheshti University of Medical Sciences, Tehran, Iran
bGenIran Lab, Tashkhis Gene Pajohesh, Tehran, Iran
cDepartment of Laboratory Sciences, School of Paramedical Sciences, Torbat Heydariyeh University of Medical Sciences, Torbat Heydariyeh, Iran
dNeuroscience Research Center, Torbat Heydariyeh University of Medical Sciences, Torbat Heydariyeh, Iran
eClinical Research Development Center (CRDU), Qom University of Medical Sciences, Qom, Iran
fUrogenital Stem Cell Research Center, Shahid Beheshti University of Medical Sciences, Tehran, Iran
A R T I C L E I N F O Keywords:
Breast cancer lncRNA CEBPA CHAST ADINR DICER1-AS1 HNF1A-AS1 NKILA ATG5
A B S T R A C T
The nuclear factor-κB (NF-κB) has a prominent role in development of breast cancer and response of patients to conventional therapies. Several factors regulate the activity of this transcription factor.
In the current investigation, we compared expression levels of five long non-coding RNAs (lncRNAs) with putative interactions with NF-κB namelyCHAST, ADINR, DICER1-AS1, HNF1A-AS1 andNKILAbetween 78 breast cancer tissues and their paired adjacent non-cancerous tissues (ANCTs). We also assessed expression levels ofATG5andCEBPAmRNA coding genes that are functionally linked with NF-κB signaling in these two sets of samples.
All assessed genes except forNKILAwere significantly down-regulated in tumoral tissues compared with ANCTs. Expression ofNKILAwas not significantly different between tumoral tissues and ANCTs. Expression levels ofCEBPAandHNF1A-ASwere significantly associated with cancer stage (Pvalues of 0.03 and 0.02 respectively). Expression levels ofATG5tended to be associated with mitotic rate (P= .05). The association between expression levels ofATG5and tumor size was also significant (P= .02). Expression ofCHASTwas significantly associated with PR status (P= .04) and tended to be associated with ER status (P= .05). Finally, expression ofNKILAwas significantly associated with first pregnancy age (P= .01). No other significant as- sociation was detected between expression levels of assessed genes and clinical parameters. Expression levels of mentioned genes were significantly correlated with each other. The most significant correlations were found betweenCHASTandADINR(correlation coefficients of 0.78 and 0.69 in tumoral tissues and ANCTs respec- tively). Based on the area under curve (AUC) values,DICER1-ASandCEBPAhad the best performance in dif- ferentiation of tumoral tissues from ANCTs (AUC values of 0.92 and 0.90 respectively. Combination of transcript quantities of six genes could differentiate these two sets of samples with 92.3% sensitivity, 91% specificity and diagnostic power of 95%.
The current project highlights dysregulation of NF-κB-associated genes in breast cancer tissues and suggests them as potential diagnostic markers in breast cancer.
1. Introduction
Breast cancer as the most frequent female cancer (Ghoncheh et al., 2016) has been the focus of several investigations. Among the most important factors in the development of breast cancer is the nuclear
factor-κB (NF-κB). NF-κB participates in the evolution and progression of human breast cancer (Wang et al., 2015b). Besides, induction of NF- κB expression results in the evolution of invasive breast cancers which are not sensitive to conventional therapies. Suppression of NF-κB has sensitized cancer cells to the therapeutic effects of these modalities and
https://doi.org/10.1016/j.yexmp.2019.104359
Received 5 November 2019; Received in revised form 6 December 2019; Accepted 11 December 2019
Abbreviation:NF-κB, nuclear factor-κB; lncRNAs, long non-coding RNAs; AUC, area under curve; CEBP, CCAAT/enhancer-binding protein; ADINR, adipogenic differentiation induced noncoding RNA; CHAST, cardiac hypertrophy-associated transcript; ROC, receiver operating characteristics
⁎Corresponding authors.
E-mail addresses:[email protected](S. Ghafouri-Fard),[email protected](M. Taheri).
Available online 14 December 2019
0014-4800/ © 2019 Elsevier Inc. All rights reserved.
T
subsequently has improved survival of patients with breast cancer (Wang et al., 2015b). Consequently, understanding the mechanisms of regulation of NF-κB expression has practical significance. Long non- coding RNAs (lncRNAs) as important regulators of gene expression have been shown to interact with NF-κB. NF-KappaB Interacting LncRNA (NKILA), whose expression is induced by NF-κB, attaches to NF-κB/IκB, and obscures phosphorylation motifs of IκB, thus suppressing IKK- mediated IκB phosphorylation and NF-κB activation. Notably,NKILA has an indispensable role in suppression of NF-κB over-activation (Liu et al., 2015). It has been noted for more than two decades that NF-κB has interactions with another transcription factor namely CCAAT/en- hancer-binding protein (CEBP) (Xia et al., 1997). More recently, it has been demonstrated that NF-κB p50 modulates CEBPα expression (Wang et al., 2009). Moreover, C/EBPα has a role in relocation of histone deacetylases from NF-κB p50 homodimers and stimulation of NF-κB target genes (Paz-Priel et al., 2011). An lncRNA namely adipogenic differentiation induced noncoding RNA(ADINR), is transcribed from up- stream of the CEBPA gene and its expression is correlated with ex- pression of CEBPA. ADINRhas also been shown to regulate CEBPA expression in cis through recruitment of histone methyl-transferase complexes (Xiao et al., 2015).
The lncRNAcardiac hypertrophy-associated transcript(CHAST) is an lncRNA with remarkable over-expression in hypertrophic heart tissue and in embryonic stem cell–originated cardiomyocytes after hyper- trophic impulses. Animal studies have shown thatChastsilencing af- fects Wnt signaling (Viereck et al., 2016). Based on the close relation- ship between Wnt/β-Catenin and NF-κB signaling pathways (Ma and Hottiger, 2016), this lncRNA might also be involved in regulation of NF- κB signaling.
The lncRNA DICER1-AS1 modulates autophagy of osteosarcoma cells through miR-30b/ATG5 axis (Gu et al., 2018). Autophagy parti- cipates in numerous cellular activities controlled by NF-κB such as cell survival, differentiation and senescence. Besides, autophagy and NF-κB have shared upstream controlling signals. Both positive and negative feedback loops have been recognized between these two pathways (Trocoli and Djavaheri-Mergny, 2011).
NF-kB signaling has also have interaction with another transcription factor namely the hepatocyte nuclear factor-1A (HNF-1A) (Bao et al., 2015). An antisense RNA which is transcribed from the opposite di- rection ofHNF-1Ahas been shown to participate in the pathogenesis of esophageal cancer (Yang et al., 2014).
Based on the above-mentioned evidences, we compared expression of CHAST, ADINR, DICER1-AS1, HNF1A-AS1 andNKILA lncRNAs as well as two mRNA coding genes which are linked with two of these lncRNAs namelyATG5andCEBPAin breast cancer samples and their pared adjacent non-cancerous tissues (ANCTs).
2. Material and methods 2.1. Patients
In the present investigation, we recruited 78 female breast cancer patients. Patients aged between 29 and 84 ((mean ± (SD):
52.82 ± 13.71). Both tumoral tissues and their matching ANCTs were acquired during surgery, quickly transferred in liquid Nitrogen to the laboratory and stored in −80 °C. The diagnosis was confirmed by a pathologist. Malignant samples included one papillary carcinoma, seven invasive lobular carcinoma, one ductal carcinoma in situ and 69 invasive ductal carcinoma samples. Samples were gathered from pa- tients hospitalized in Farmanieh and Sina Hospitals during 2016–2018.
The study protocol was approved by Ethical Committee of Shahid Beheshti University of Medical Sciences and all methods were per- formed in accordance with the relevant guidelines and regulations.
Informed written consent forms were signed by all study participants.
2.2. Expression analysis
RNA was isolated from tissue specimens using Hybrid-R™ kit (GeneAll®, Seoul, South Korea). RNA samples were treated with DNAse I (Thermo Fisher Scientific, Lithuania) to get rid of DNA remnants.
Next, the RNA quantity and quality was assessed and cDNA was made from extracted RNA using FIREScript RT cDNA Synthesis Kit (Solis BioDyne, Estonia). Relative transcription levels of genes were counted in all specimens using RealQ Plus Master Mix Green (AMPLICON, Denmark). Transcript levels of genes were normalized toB2Mexpres- sion levels. Experiments were performed in the rotor gene 6000 ma- chine (Corbett, Australia) in duplicate.Table 1summarizes the primer features. Primers for amplification ofCHASThave been taken from a previous study (Viereck et al., 2016).
2.3. Statistical methods
Expression of genes in tumoral specimens and ANCTs were mea- sured considering the efficiency values. The association between clin- ical variables and expression quantities was gauged using Chi-square test or Fisher's exact-test depending on the situation. The level of sig- nificance was fixed at P < .05. The suitability of mRNA levels of mentioned genes in breast cancer diagnosis was assessed using receiver operating characteristics (ROC) curves. The significance of difference in mean values of genes expressions between two paired groups, the Kruschke's was used. A t student prior distribution was supposed for parameters with 4000 iteration and 2000 burn-outs. The highest den- sity interval (HDI) was calculated based on Bayesian approach Table 1
The sequences of primers used for expression assays.
Name Type Sequence Primer Length PCR Product Length
NKILA-F lncRNA AACCAAACCTACCCACAACG 20 108
NKILA-R ACCACTAAGTCAATCCCAGGTG 20
CEBPA-DT (ADINR)-F lncRNA AGGGTGGATGTGCTGTGATGAAGA 24 98
CEBPA-DT (ADINR)-R AGTCCATAACACCTCCGCAGACAA 24
CEBPA-F mRNA ATTGCCTAGGAACACGAAGCACGA 24 161
CEBPA-R TTTAGCAGAGACGCGCACATTCAC 24
DICER1-AS1-F lncRNA CGAAGAAATGGAATAACTTCCAAC 24 125
DICER1-AS1-R TTGGTCCAAACACAGAAGATC 21
ATG5-F mRNA TTCGAGATGTGTGGTTTGGAC 21 134
ATG5-R CACTTTGTCAGTTACCAACGTCA 23
HNF1A-AS1-F lncRNA TCAAGAAATGGTGGCTAT 18 148
HNF1A-AS1-R GCTCTGAGACTGGCTGAA 18
CHAST-F lncRNA GCAGAGGGTGCCAACTTGTA 20 109
CHAST-R TCTCAGGGAAATCAGATTGCGG 22
B2M-F mRNA AGATGAGTATGCCTGCCGTG 20 105
B2M-R GCGGCATCTTCAAACCTCCA 20
(Kruschke, 2013). In addition, Bayesian regression model was used to examine the effects of independent variables on gene expressions. TheP values were estimated from frequentist methods including quantile regression and mixed effects models. The spearman correlation was used to examine the size and direction of association among continues variables. The ‘quantreg’, ‘ggplot2’, and ‘BEST’ packages were used in R 3.5.2 environment.
3. Results
3.1. General data of patients
Table 2summarizes the general characteristics of study participants.
3.2. Expression levels genes in tumoral tissues versus ANCTs
All assessed genes except forNKILAwere significantly down-regu- lated in tumoral tissues compared with ANCTs. Expression of NKILA was not significantly different between tumoral tissues and ANCTs.
Table 3andFig. 1show expression levels of assessed genes in these two sample groups.
3.3. Associations between expression of genes and clinical data
Expression levels ofCEBPAandHNF1A-ASwere significantly asso- ciated with cancer stage (P values of 0.03 and 0.02 respectively).
Expression levels of ATG5tended to be associated with mitotic rate (P = .05). The association between expression levels ofATG5 and tumor size was also significant (P= .02). Expression ofCHAST was
significantly associated with PR status (P= .04) and tended to be as- sociated with ER status (P= .05). Finally, expression ofNKILAwas significantly associated with first pregnancy age (P= .01). Table 4 shows the results of association analysis between expression of genes and clinical and demographic parameters. Moreover, the quantile re- gression analysis showed association between expression ofHNF1A-AS and history of abortion (Beta = 0.21, standard error = 0.09,t= 2.32, P= .023, 95% confidence interval = [0.03, 0.41]).
3.4. ROC curve analysis
In order to assess the diagnostic power of the mentioned genes, ROC curves were depicted for 6 genes whose expressions were different between tumoral tissues and ANCTs (Fig. 2). Based on the area under curve (AUC) values,DICER1-ASandCEBPAhad the best performance in differentiation of tumoral tissues from ANCTs (AUC values of 0.92 and 0.90 respectively. Combination of transcript quantities of six genes could differentiate these two sets of samples with 92.3% sensitivity, 91% specificity and diagnostic power of 95% (Table 5).
ATG5.DICER1-AShad the best diagnostic power among all assessed genes.
3.5. Correlations between expression levels of genes
Expression levels of mentioned genes were significantly correlated with each other (Fig. 3). The most significant correlations were found betweenDICER1-ASandNKILAas well as betweenCHASTandADINR.
4. Discussion
The NF-κB family of transcription factors participates in regulation of several cellular activities such as proliferation, differentiation and immune responses (Pires et al., 2017). Such vast domain of functions indicates their role in the pathogenesis of breast cancer. This patho- genic role is exerted through different mechanisms such as epithelial-to- mesenchymal transition (Pires et al., 2017), up-regulation of certain cell surface glycoproteins (Smith et al., 2014) and expression of the proin- flammatory cytokine IL-6 (Gordon et al., 2005). Although the role of lncRNAs in the regulation of many cancer-related pathway has been assessed previously, associations between NF-κB signaling and lncRNAs have been the focus of few studies. For instance, the interactions be- tween NF-κB and the lncRNAsHOTAIRandPCAT1have been evaluated in the context of prostate cancer (Xiang et al., 2018;Shang et al., 2019).
In breast cancer, the cross-talk between NKILA and NF-κB inhibits cancer metastasis (Liu et al., 2015). Based on the scarcity of data re- garding the associations between lncRNAs and NF-κB signaling espe- cially in the context of breast cancer, we performed the current study to address this issue. We reported significant down-regulation ofCHAST, ADINR, DICER1-AS1, HNF1A-AS1lncRNAs as well asATG5andCEBPA mRNAs in tumoral tissues compared with ANCTs. However, NKILA expression was not significantly different between two sets of speci- mens. Liu et al. have reported that NKILA expression is induced by inflammatory cytokines through NF-κB signaling. Meanwhile, NKILA negatively regulates NF-κB signaling and suppresses NF-κB-mediated breast cancer metastasis (Liu et al., 2015). The reason for the observed similar levels ofNKILAbetween tumoral tissues and ANCTs in our study might be the influence of tumor microenvironment including in- flammatory cytokines on the expression profile of ANCTs. Such spec- ulation also resolves the discrepancy between our results and the results of Liu et al. study since they compared expression level of this lncRNA between tumoral tissues and benign breast samples but not ANCTs (Liu et al., 2015).
We also demonstrated lower levels ofCHASTlncRNA in tumoral tissues versus ANCTs. The primers used for amplification of this lncRNA have been taken from Vierek et al. study (Viereck et al., 2016). Further in silico assays showed that these primers also amplify PLEKHM1P Table 2
The general characteristics of study participants (SD: standard deviation).
Parameters Values
Age (mean ± SD (range)) 52.82 ± 13.71 (29–84)
Menarche age (mean ± SD (range)) 13.12 ± 1.51 (10–18) Menopause age (mean ± SD (range)) 49.46 ± 4.84 (38–60) First pregnancy age (mean ± SD (range)) 21.35 ± 4.97 (14–37) Breast feeding duration (months) (mean ± SD
(range)) 46.12 ± 46.08 (0–240)
Abortion history
Yes 18.9%
No 81.1%
Cancer stage (%)
I 27.4
II 31.6
III 34.2
IV 6.8
Overall grade (%)
I 18.8
II 52.2
III 29
Mitotic rate (%)
I 42.9
II 42.9
III 14.3
Tumor size (%)
< 2 cm 28.2
≥2 cm, < 5 cm 69
≥5 cm 2.8
Estrogen receptor (%)
Positive 19.4
Negative 80.6
Progesterone receptor (%)
Positive 24.3
Negative 75.7
Her2/neu expression (%)
Positive 18.3
Negative 81.7
Ki67 expression (%)
Positive 100
Negative 0
pseudogene. Animal studies have shown thatChastnegatively regulates expression ofPlekhm1p(Viereck et al., 2016). Future studies are needed to assess whetherCHASTcould be a negative regulator ofPLEKHM1Pin humans and address the issue of being a pseudogene. We also reported significant association betweenCHASTexpression and PR status and a tendency toward association between its expression and ER status.
Animal studies have demonstrated that Chast is induced by nuclear factor of activated T cells (NFAT) signaling (Viereck et al., 2016).
Ehring et al. have reported that progesterone can block certain K+
channels, leading to depolarization of the membrane potential and subsequent inhibition of NFAT-driven gene expression (Ehring et al., 1998). The mentioned evidences might explain the observed associa- tions betweenCHASTexpression and hormone receptors status.
Besides, we detected decreased expressions ofADINRandCEBPAin tumoral tissues versus ANCTs and significant correlations between ex- pression levels of this pair of lncRNA-mRNA which is in accordance with the proposed role forADINRin regulation ofCEBPA(Xiao et al., 2015). Our results are also in line with a previous study which de- monstrated prominently decreased or unnoticeable expression of CEBPA protein in a subset of breast cancer samples (Gery et al., 2005).
Although Gery et al. reported association between ER/ PR status and CEBPA expression (Gery et al., 2005), we could not detect such asso- ciations. However, we demonstrated associations between CEBPA
expression and cancer stage which might be explained by the role of CEBPA in growth inhibition, cell cycle arrest and decreased anchorage- independent cell growth (Gery et al., 2005).
The observed down-regulation ofDICER1-AS1in breast cancer tis- sues in our study is in obvious contrast with its expression pattern in osteosarcoma (Gu et al., 2018). However, the positive correlation be- tween its expression and expression ofATG5is in line with Gu et al.
study which has shown thatDICER1-AS1 silencing decreases protein level of ATG5 (Gu et al., 2018). Future in vitro experiments are needed to assess the effects ofDICER1-AS1down-regulation in the pathogenesis of breast cancer. The observed down-regulation of ATG5 in breast cancer cells supports the previous study by Han et al. which showed the association between downregulation of ATG5-dependent macro- autophagy and breast cancer cell metastasis (Han et al., 2017). We also detected associations betweenATG5levels and tumor size. Our results are in line with Wang et al. study which reported expression levels of ATG5 as biomarkers for favorable disease outcome in breast cancer patients (Wang et al., 2015a).
Contrary to the bulk of evidences regarding the oncogenic role of HNF1A-AS1in many cancers (Yang et al., 2014;Cai et al., 2017;Fang et al., 2017), we reported significant down-regulation of this lncRNA in breast cancer tissues versus ANCTs which might reflect a distinct role for this lncRNA in breast cancer. In vitro studies are needed to verify Table 3
The results of Bayesiant-test for comparison of the relative expression of genes between two paired groups (IQR: [Quartile 1, Quartile3]).
Gene Median of ∆CT [IQR] Relative Expression difference Standard deviation Effect Size P-value 95% HDI Tumoral tissue Non-tumoral tissue
CEBPA 7.44 [5.6, 10.09] 2.4 [0.52, 3.97] −3.83 2.89 −1.34 < 0.0001 [−4.49, −3.14]
CHAST 7.53 [5.96, 9.2] 4.29 [2.69, 6.04] −2.151 2.499 −0.868 < 0.0001 [−2.72, −1.57]
ADINR 6.68 [5.62, 9.33] 5.12 [3.28, 6.59] −1.418 2.515 −0.568 < 0.0001 [−2, −0.84]
DICER1-AS 5.91 [3.44, 8.88] −1.1 [−3.46, 0.97] −4.77 3.32 −1.45 < 0.0001 [−5.53, −4.01]
HNF1A-AS 9.21 [6.98, 11.53] 4.49 [2.86, 5.82] −3.41 3.24 −1.06 < 0.0001 [−4.15, −2.64]
NKILA 3.29 [0.27, 6.23] 2.33 [−0.08, 4.65] −0.444 3.611 −0.124 0.299 [−1.27, 0.39]
ATG5 7.59 [6.46, 9.53] 5.41 [4.56, 6.39] −1.73 2.18 −0.8 < 0.0001 [−2.25, −1.23]
Fig. 1.Relative expression of genes in tumoral tissues and ANCTs. Expression of genes between paired samples were compared using pairedt-test. Significance is indicated by *.
Table4 Theresultsofassociationanalysisbetweenexpressionofgenesandclinicalanddemographicparameters. ParametersCEBPAup- regulationCEBPA down- regulation
PvalueCHASTup- regulationCHASTDown- regulationPvalueADINRup- regulationADINRdown- regulationPvalueDICER1-AS up-regulation Age0.110.380.59 <55years6(14.3%)36(85.7%)10(23.8%)32(76.2%)14(33.3%)28(66.7%)4(9.5%) ≥55years1(2.8%)35(97.2%)5(13.9%)31(86.1%)10(27.8%)26(72.2%)2(5.6%) Stage0.030.610.51 14(20%)16(80%)3(15%)17(85%)5(20%)15(75%)3(15%) 21(4.3%)22(95.7%)4(17.4%)19(82.6%)5(21.7%)18(78.3%)0(0%) 30(0%)25(100%)5(20%)20(80%)10(40%)15(60%)1(4%) 41(20%)4(80%)2(40%)3(60%)2(40%)3(60%)1(20%) HistologicalGrade10.20.82 11(7.7%)12(92%)0(0%)13(100%)4(30.8%)9(69.2%)0(0%) 23(8.3%)33(91.7%)8(22.2%)28(77.8%)9(25%)27(75%)3(8.3%) 31(5%)19(95%)3(15%)17(85%)6(30%)14(70%)1(20%) MitoticRate0.510.86 11(3.7%)26(96.3%)5(18.5%)22(81.5%)7(25.9%)20(74.1%)1(3.7%) 23(11.1%)24(88.9%)5(18.5%)22(81.5%)7(25.9%)20(74.1%)3(11.1%) 30(0%)9(100%)1(11.1%)8(88.9%)3(33.3%)6(66.7%)0(0%) Tumorsize0.670.060.07 <22(10%)18(90%)3(15%)17(85%)7(35%)13(65%)1(5%) 2–53(6.1%)46(93.9%)9(18.4%)40(81.6%)12(24.5%)37(75.5%)2(4.1%) >50(0%)2(100%)2(100%)0(0%)2(100%)0(0%)0(0%) ERStatus10.050.2 Negative1(7.1%)13(92.9%)5(35.77%)9(64.3%)6(42.9%)8(57.1%)1(7.1%) Positive4(6.9%)54(93.1%)8(13.8%)50(86.2%)15(25.9%)43(74.1%)2(3.4%) PRStatus10.040.18 Negative1(5.9%)16(94.1%)6(35.3%)11(64.7%)7(41.2%)10(58.8%)1(5.9%) Positive4(7.5%)49(92.5%)7(13.2%)46(86.8%)13(24.5%)40(75%)2(3.8%) Her2Status10.691 Negative4(6.9%)54(93.1%)10(17.2%)48(82.8%)17(29.3%)41(70.7%)2(3.4%) Positive1(7.7%)12(92.3%)3(23.1%)10(76.9%)4(30.8%)9(69.2%)1(7.7%) Firstpregnancyage10.691 15to20years3(9.7%)28(90.3%)6(19.4%)25(80.6%)11(35.5%)20(64.5%)1(3.2%) 21to25years2(9.1%)20(90.9%)6(27.3%)16(72.7%)7(31.8%)15(68.2%)2(9.1%) ≥26years0(0%)9(100%)1(11.1%)8(88.9%)3(33.3%)6(88.9%)0(0%) ParametersDICER1-AS down- regulation
PvalueHNF1A-AS up-regulationHNF1A-AS down- regulation
PvalueNKILAup- regulationNKILAdown- regulationPvalueATG5up- regulationATG5down- regulationPvalue Age0.680.20.430.21 <55years38(90.5%)8(19%)34(81%)20(47.6%)22(52.4%)
12 (28.6%)
30(71.4%) ≥55years34(94.4%)3(8.3%)33(91.7%)14(38.9%)22(61.1%)6(16.7%)30(83.3%) (continuedonnextpage)
Table4(continued) ParametersDICER1-AS down- regulation
PvalueHNF1A-AS up-regulationHNF1A-AS down- regulation
PvalueNKILAup- regulationNKILAdown- regulationPvalueATG5up- regulationATG5down- regulationPvalue Stage0.080.020.850.35 117(85%)4(20%)16(80%)8(40%)12(60%)5(25%)15(75%) 223(100%)0(0%)23(100%)10(43.5%)13(56.5%)3(13%)20(87%) 324(96%)3(12%)22(88%)13(52%)12(48%)8(32%)17(68%) 44(80%)2(40%)3(60%)2(40%)3(60%)2(40%)3(60%) Histological Grade0.810.620.320.69 113(100%)1(7.7%)12(92.3%)8(61.5%)5(36.5%)3(23.1%)10(76.9%) 233(91.7%)4(11.1%)32(88.9%)15(41.7%)21(58.3%)7(19.4%)29(80.6%) 319(80%)4(20%)16(80%)7(35%)13(65%)6(30%)14(70%) MitoticRate0.50.420.940.05 126(96.3%)2(7.4%)25(92.6%)12(44.4%)15(55.6%)2(7.4%)25(92.6%) 224(88.9%)5(18.5%)22(81.5%)11(40.7%)16(59.3%)9(33.3%)18(66.7%) 39(100%)2(22.2%)7(77.8%)3(33.3%)6(66.7%)2(22.2%)7(77.8%) Tumorsize10.070.320.02 <219(95%)4(20%)16(80%)8(40%)12(60%)6(30%)14(70%) 2–547(95.9%)4(8.2%)45(91.8%)20(40.8%)29(59.2%)8(16.3%)41(83.7%) >52(100%)1(50%)1(50%)2(100%)0(0%)2(100%)0(0%) ERStatus0.480.060.820.17 Negative13(92.9%)4(28.6%)10(71.4%)6(42.9%)8(57.1%)5(35.7%)9(64.3%) Positive56(96.6%)5(8.6%)53(91.4%)23(39.7%)35(60.3%)11(19%)47(81%) PRStatus10.090.980.16 Negative16(94.1%)4(23.5%)13(76.5%)7(41.2%)10(58.8%)6(35.3%)11(64.7%) Positive51(96.2%)4(7.5%)49(92.5%)22(41.5%)31(58.5%)
10 (18.9%)
43(81.1%) Her2Status0.460.350.751 Negative56(96.6%)6(10.3%)52(89.7%)23(39.7%)35(60.3%)
13 (22.4%)
45(77.6%) Positive12(92.3%)3(23.1%)10(76.9%)6(46.2%)7(53.8%)3(23.1%)10(76.9%) First
pregnancy age
0.730.750.010.27 15to20years30(96.8%)4(12.9%)27(87.1%)12(38.7%)19(61.3%)8(25.8%)23(74.2%) 21to25years20(90.9%)4(18.2%)18(81.8%)12(54.5%)10(45.5%)4(18.2%)18(81.8%) ≥26years9(100%)2(22.2%)7(77.8%)0(0%)9(100%)0(0%)9(100%)
this speculation. We also detected associations between expression of HNF1A-AS1and cancer stage which further imply its role in the car- cinogenesis process. Moreover, the observed association between ex- pression ofHNF1A-ASand history of abortion suggests that this lncRNA might be involved in the “abortion-breast cancer link”.
The significant pairwise correlation between expression levels of mentioned genes implies the presence of single regulatory mechanism for their expression which should be assessed in future. We also de- monstrated high diagnostic power of six out of seven assessed genes in differentiation of malignant status of breast cells. Notably, combination of transcript levels of mentioned genes had an excellent diagnostic power in this regard.
Our study has some limitations including lack of assessment of status of NF-kB activation or expression of NF-kB downstream genes and lack of evaluation of expression of autophagy markers.
5. Conclusion
Our study reveals dysregulation of NF-κB-associated genes in breast cancer and suggests a panel of these genes which can distinguish cancer tissues from ANCTs. The diagnostic power of this panel should be evaluated in larger sample sizes.
Ethics approval and consent to participate
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/
or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Ethical ap- proval was obtained from the ethical committee of Shahid Beheshti University of Medical Sciences (IR.SBMU.MSP.REC.1397.1022).
Informed consent was obtained from the patients.
Fig. 2.ROC curves for prediction of breast cancer based on expression levels of CEBPA, CHAST, ADINR, DICER1-AS, HNF1A-ASand.
Table 5
The results of ROC curve analysis (a: Youden index, b Significance level P (Area = 0.5), Estimate criterion: optimal cut-off point for transcript quantifi- cation).
Estimate
criterion AUC Ja Sensitivity Specificity P-valueb
CEBPA > 5.51 0.90 0.7 78.2 92.3 < 0.0001
CHAST > 5.83 0.79 0.49 75.6 73.1 < 0.0001
ADINR > 5.61 0.7 0.36 75.6 60.3 < 0.0001
DICER1-AS > 2.08 0.92 0.72 83.3 88.5 < 0.0001
HNF1A-AS > 5.57 0.86 0.6 89.7 70.5 < 0.0001
ATG5 > 6.21 0.81 0.54 80.8 73.1 < 0.0001
Combination of
all genes > 0.51 0.95 0.83 92.3 91 < 0.0001
Fig. 3.The scatterplot matrix depicting the correlations between expression levels of genes. Histograms of the variables are shown in the diagonal, and correlation coefficients are in the upper part of the matrix.
Consent of publication Not applicable.
Availability of data and materials
The analysed data sets generated during the study are available from the corresponding author on reasonable request.
Funding
The present article is financially supported by “Research Department of the School of Medicine Shahid Beheshti University of Medical Sciences”.
Authors' contributions
MT and SD conducted the experiments. VKO, SA and FE performed the data analyses. SGF designed and supervised the study and wrote the manuscript. All authors read the manuscript drafts, contributed edits, and approved the final manuscript.
Declaration of Competing Interest
The authors declare they have no conflict of interest.
Acknowledgement
The current study was supported by a grant from Shahid Beheshti University of Medical Sciences.
References
Bao, C., Li, Y., Huan, L., Zhang, Y., Zhao, F., Wang, Q., Liang, L., Ding, J., Liu, L., Chen, T., Li, J., Yao, M., Huang, S., He, X., 2015. NF-kappaB signaling relieves negative reg- ulation by miR-194 in hepatocellular carcinoma by suppressing the transcription factor HNF-1alpha. Sci Signal 8, ra75.
Cai, L., Lv, J., Zhang, Y., Li, J., Wang, Y., Yang, H., 2017. The lncRNA HNF1A-AS1 is a negative prognostic factor and promotes tumorigenesis in osteosarcoma. J. Cell. Mol.
Med. 21, 2654–2662.
Ehring, G.R., Kerschbaum, H.H., Eder, C., Neben, A.L., Fanger, C.M., Khoury, R.M., Negulescu, P.A., Cahalan, M.D., 1998. A nongenomic mechanism for progesterone- mediated immunosuppression: inhibition of K+ channels, Ca2+ signaling, and gene expression in T lymphocytes. J. Exp. Med. 188, 1593–1602.
Fang, C., Qiu, S., Sun, F., Li, W., Wang, Z., Yue, B., Wu, X., Yan, D., 2017. Long non-coding RNA HNF1A-AS1 mediated repression of miR-34a/SIRT1/p53 feedback loop pro- motes the metastatic progression of colon cancer by functioning as a competing en- dogenous RNA. Cancer Lett. 410, 50–62.
Gery, S., Tanosaki, S., Bose, S., Bose, N., Vadgama, J., Koeffler, H.P., 2005. Down-reg- ulation and growth inhibitory role of C/EBPalpha in breast cancer. Clin. Cancer Res.
11, 3184–3190.
Ghoncheh, M., Pournamdar, Z., Salehiniya, H., 2016. Incidence and mortality and
epidemiology of breast Cancer in the world. Asian Pac. J. Cancer Prev. 17, 43–46.
Gordon, A.H., O’keefe, R.J., Schwarz, E.M., Rosier, R.N., Puzas, J.E., 2005. Nuclear factor- kappaB-dependent mechanisms in breast cancer cells regulate tumor burden and osteolysis in bone. Cancer Res. 65, 3209–3217.
Gu, Z., Hou, Z., Zheng, L., Wang, X., Wu, L., Zhang, C., 2018. LncRNA DICER1-AS1 promotes the proliferation, invasion and autophagy of osteosarcoma cells via miR- 30b/ATG5. Biomed. Pharmacother. 104, 110–118.
Han, Q., Deng, Y., Chen, S., Chen, R., Yang, M., Zhang, Z., Sun, X., Wang, W., He, Y., Wang, F., Pan, X., Li, P., Lai, W., Luo, H., Huang, P., Guan, X., Deng, Y., Yan, J., Xu, X., Wen, Y., Chen, A., Hu, C., Li, X., Li, S., 2017. Downregulation of ATG5-dependent macroautophagy by chaperone-mediated autophagy promotes breast cancer cell metastasis. Sci. Rep. 7, 4759.
Kruschke, J.K., 2013. Bayesian estimation supersedes the t test. J Exp Psychol Gen 142, 573–603.
Liu, B., Sun, L., Liu, Q., Gong, C., Yao, Y., Lv, X., Lin, L., Yao, H., Su, F., Li, D., Zeng, M., Song, E., 2015. A cytoplasmic NF-kappaB interacting long noncoding RNA blocks IkappaB phosphorylation and suppresses breast cancer metastasis. Cancer Cell 27, 370–381.
Ma, B., Hottiger, M.O., 2016. Crosstalk between Wnt/β-catenin and NF-κB signaling pathway during inflammation. Front. Immunol. 7, 378.
Paz-Priel, I., Houng, S., Dooher, J., Friedman, A.D., 2011. C/EBPalpha and C/EBPalpha oncoproteins regulate nfkb1 and displace histone deacetylases from NF-kappaB p50 homodimers to induce NF-kappaB target genes. Blood 117, 4085–4094.
Pires, B.R., Mencalha, A.L., Ferreira, G.M., de Souza, W.F., Morgado-Díaz, J.A., Maia, A.M., Corrêa, S., Abdelhay, E.S., 2017. NF-kappaB is involved in the regulation of EMT genes in breast cancer cells. PLoS One 12 (1) e0169622.
Shang, Z., Yu, J., Sun, L., Tian, J., Zhu, S., Zhang, B., Dong, Q., Jiang, N., Flores-Morales, A., Chang, C., Niu, Y., 2019. LncRNA PCAT1 activates AKT and NF-kappaB signaling in castration-resistant prostate cancer by regulating the PHLPP/FKBP51/IKKalpha complex. Nucleic Acids Res.
Smith, S.M., Lyu, Y.L., Cai, L., 2014. NF-kappaB affects proliferation and invasiveness of breast cancer cells by regulating CD44 expression. PLoS One 9, e106966.
Trocoli, A., Djavaheri-Mergny, M., 2011. The complex interplay between autophagy and NF-kappaB signaling pathways in cancer cells. Am. J. Cancer Res. 1, 629–649.
Viereck, J., Kumarswamy, R., Foinquinos, A., Xiao, K., Avramopoulos, P., Kunz, M., Dittrich, M., Maetzig, T., Zimmer, K., Remke, J., Just, A., Fendrich, J., Scherf, K., Bolesani, E., Schambach, A., Weidemann, F., Zweigerdt, R., De Windt, L.J., Engelhardt, S., Dandekar, T., Batkai, S., Thum, T., 2016. Long noncoding RNA Chast promotes cardiac remodeling. Sci. Transl. Med. 8.
Wang, D., Paz-Priel, I., Friedman, A.D., 2009. NF-κB p50 regulates C/EBPα expression and inflammatory cytokine-induced neutrophil production. J. Immunol. 182, 5757–5762.
Wang, L., Yao, L., Zheng, Y.Z., Xu, Q., Liu, X.P., Hu, X., Wang, P., Shao, Z.M., 2015a.
Expression of autophagy-related proteins ATG5 and FIP200 predicts favorable dis- ease-free survival in patients with breast cancer. Biochem. Biophys. Res. Commun.
458, 816–822.
Wang, W., Nag, S.A., Zhang, R., 2015b. Targeting the NFkappaB signaling pathways for breast cancer prevention and therapy. Curr. Med. Chem. 22, 264–289.
Xia, C., Cheshire, J.K., Patel, H., Woo, P., 1997. Cross-talk between transcription factors NF-kappa B and C/EBP in the transcriptional regulation of genes. Int. J. Biochem. Cell Biol. 29, 1525–1539.
Xiang, S., Zou, P., Wu, J., Zheng, F., Tang, Q., Zhou, J., Hann, S.S., 2018. Crosstalk of NF- kappaB/P65 and LncRNA HOTAIR-mediated repression of MUC1 expression con- tribute to synergistic inhibition of castration-resistant prostate Cancer by Polyphyllin 1-enzalutamide combination treatment. Cell. Physiol. Biochem. 47, 759–773.
Xiao, T., Liu, L., Li, H., Sun, Y., Luo, H., Li, T., Wang, S., Dalton, S., Zhao, R.C., Chen, R., 2015. Long noncoding RNA ADINR regulates Adipogenesis by transcriptionally ac- tivating C/EBPalpha. Stem Cell Reports 5, 856–865.
Yang, X., Song, J.H., Cheng, Y., Wu, W., Bhagat, T., Yu, Y., Abraham, J.M., Ibrahim, S., Ravich, W., Roland, B.C., Khashab, M., Singh, V.K., Shin, E.J., Yang, X., Verma, A.K., Meltzer, S.J., Mori, Y., 2014. Long non-coding RNA HNF1A-AS1 regulates pro- liferation and migration in oesophageal adenocarcinoma cells. Gut 63, 881–890.
