**4. Role of non-coding RNAs in CF**

### **4.1. What is the impact of mutations in microRNA target sites on the** *CFTR* **gene?**

### *4.1.1. Mutations in the 3'UTR of the CFTR gene*

Assessing the putative impact of single nucleotide polymorphisms (SNPs) in the 3'UTR of *CFTR* is essential to define the pathological importance of these cis-regulatory motifs in noncoding regions, especially after the development of Next-Generation Sequencing (NGS) technologies. Recent work identified the SNP c.\*1043A>C on the 3'UTR of the *CFTR* gene in one patient with a CFTR-related disorder and congenital bilateral absence of vas deferens [61]. This SNP is located in a region predicted to interact with miR-433 and miR-509-3p. Expression analysis demonstrated that the c.\*1043A>C mutation increases the affinity for miR-509-3p and slightly decreases that for miR-433. *In vitro*, these two miRNAs reduces CFTR protein expres‐ sion. The authors suggested that the very low expression of miR-509-3p in normal human bronchial epithelial (NHBE) cells could explain the mild phenotype caused by this mutation. Thus, the c.\*1043A>C mutation, by acting as a mild *CFTR* mutation that enhances the affinity for an inhibitory miRNA, could represent a novel pathogenic mechanism in CF.

### *4.1.2. Methods to investigate the impact of CFTR gene mutations*

Bioinformatics tools, such as miRNA binding site prediction programs (listed in Table 2), could be used to predict whether at a mutated position there is a cis-regulatory motif and whether it corresponds to an miRNA binding site. Moreover, tools like RNAhybrid could allow predicting the binding energy of the mutated motif. *In vitro* luciferase reporter gene assays together with mutagenesis approaches to create degenerate 3'UTR-*CFTR* sequences can be used to study the different variants (Figure 5Ad).

### **4.2. Deregulation of non-coding RNAs in CF**

*3.2.2. Functional approaches*

290 Cystic Fibrosis in the Light of New Research

can be assessed using different techniques.

**4. Role of non-coding RNAs in CF**

*4.1.1. Mutations in the 3'UTR of the CFTR gene*

To evaluate *in vitro* the impact of miRNAs on *CFTR* gene regulation, several approaches can be used (Figure 5 A). a) First of all, *in silico* analysis by using bioinformatics tools can be performed (see 3.2.1). b) The role of specific miRNAs could be assessed *in vitro* by using airway cell lines and luciferase gene reporter assays, in which the luciferase coding sequence is under the control of the 3' UTR of *CFTR,* as previously described [11,3]. c) By transfecting mimics or precursors, miRNAs can be overexpressed and the relative level of luciferase expression reflects the effect of a given miRNA on the 3'UTR-*CFTR*. Inhibitors may also be used to confirm the specific effect of the miRNA under study. d) To validate the direct effect of the identified miRNAs (for instance, miRNA-494 and miRNA-101), the cis element for miRNAs binding in the 3'UTR-*CFTR* can be mutated by site-directed mutagenesis (created mutation). Target-site blockers may also be used to inhibit binding to the tested motif [3]. e) To validate these effects, the endogenous level of *CFTR* mRNA following miRNA overexpression or down-regulation

**4.1. What is the impact of mutations in microRNA target sites on the** *CFTR* **gene?**

for an inhibitory miRNA, could represent a novel pathogenic mechanism in CF.

Bioinformatics tools, such as miRNA binding site prediction programs (listed in Table 2), could be used to predict whether at a mutated position there is a cis-regulatory motif and whether it corresponds to an miRNA binding site. Moreover, tools like RNAhybrid could allow predicting the binding energy of the mutated motif. *In vitro* luciferase reporter gene assays together with mutagenesis approaches to create degenerate 3'UTR-*CFTR* sequences can be

*4.1.2. Methods to investigate the impact of CFTR gene mutations*

used to study the different variants (Figure 5Ad).

Assessing the putative impact of single nucleotide polymorphisms (SNPs) in the 3'UTR of *CFTR* is essential to define the pathological importance of these cis-regulatory motifs in noncoding regions, especially after the development of Next-Generation Sequencing (NGS) technologies. Recent work identified the SNP c.\*1043A>C on the 3'UTR of the *CFTR* gene in one patient with a CFTR-related disorder and congenital bilateral absence of vas deferens [61]. This SNP is located in a region predicted to interact with miR-433 and miR-509-3p. Expression analysis demonstrated that the c.\*1043A>C mutation increases the affinity for miR-509-3p and slightly decreases that for miR-433. *In vitro*, these two miRNAs reduces CFTR protein expres‐ sion. The authors suggested that the very low expression of miR-509-3p in normal human bronchial epithelial (NHBE) cells could explain the mild phenotype caused by this mutation. Thus, the c.\*1043A>C mutation, by acting as a mild *CFTR* mutation that enhances the affinity Clinical manifestations of CF are various including chronic pulmonary inflammation and infection that strongly contribute to the morbidity and mortality of these patients [62]. Overinflammation precedes chronic infection that is then amplified by pathogens. Notably, the protease–antiprotease balance, which is responsible for lung remodelling, is disrupted in CF airways early in life and then this imbalance is chronically maintained [63]. Pulmonary tissues in patients with CF are usually infected by antibiotic-resistant *Pseudomonas aeruginosa*. Interleukin-8 (IL-8), IL-6 and Tumor Necrosis Factor alpha (TNF-alpha) are pro-inflammatory cytokines that are highly expressed in CF lung epithelial cells and allow the recruitment of neutrophils [62]. Lipopolysaccharide (LPS) and IL-1beta, which bind to Toll-like receptor 4 (TLR4) and IL-1R, respectively, are also involved. The identification of the molecular events involved in lung epithelium injury and repair is essential for understanding CF physiopathol‐ ogy. Expression levels of miRNAs physiologically vary greatly among tissues. Recent advan‐ ces explored their effects on influencing signaling pathways in CF. Identification of deregulated miRNAs may offer possible future directions for clinical applications.

### *4.2.1. Deregulated lncRNAs in CF*

To date, only one publication has reported aberrant expression of specific lncRNAs in CF bronchial epithelium *in vivo* [64]. To establish the lncRNA profiles of CF and non-CF bronchial epithelium, 10 CF and 12 non-CF (controls) bronchial brushing samples were analyzed using a human lncRNA Array v3.0 (ArrayStar). In this way, more than 30586 lncRNAs and 26109 protein coding transcripts were evaluated. Overall, 1063 lncRNAs, most of which were intergenic, were differentially expressed in non-CF and CF bronchial brushing samples. RTqPCR analysis of the differential expression of well-known ncRNAs (XIST, MALAT1, HO‐ TAIR, and TLR8 antisense) did not confirm the down-regulation of MALAT1 and HOTAIR in CF samples compared to controls. Interestingly, MALAT1 (lnc-SCYL1-1\*) and HOTAIR (lnc-SMUG1-7\*) have been described as oncogenic lncRNAs in lung cancer (\*asterisks meaning the existence of several isoforms).

### *4.2.2. Deregulated miRNAs in CF*

miRNA profiling studies identified various miRNAs with altered expression in CF (summar‐ ized in Figure 6). For instance, *in vivo* the expression of miR-145, miR-223 and miR-494 is increased in CF bronchial brushing samples (individuals with at least one p.Phe508del CFTR allele) compared to non-CF controls and correlates with decreased p.Phe508del *CFTR* expression. Moreover, these three miRNAs inhibit *CFTR* mRNA expression [65]. These authors also highlighted a relationship between their regulation and CFTR chloride channel activity. Specifically, they showed that treatment with inh-172, a specific inhibitor of the CFTR chloride channel, significantly increases the level of miR-145, miR-223 and miR-494 in 16 HBE14orespiratory epithelial cells. These data are in agreement with the hypothesis by Wenming Xu *et al*. [66] that CFTR chloride channel alteration affects the miRNA profile. Another study reported that miR-509-3p and miR-494 are increased in well-differentiated primary cultures of human CF but not of non-CF airway epithelia, [67]. Other miRNAs have been found to be

Interface.

regulation lead to high IL-8 expression [73].

regulator of CTSS expression via IRF1 in CF epithelial cells [74].

4.2.3. Methods to study deregulated non-coding RNA

deregulated in CF, such as miR-155, miR-126 and miR-31 [68]. miR-155 is overexpressed in IB3-1 CF cells compared to IB3-1 control cells and in *ex vivo* CF cells (bronchial brushing samples versus normal human bronchial epithelial cells and CF neutrophils CD66+ versus control cells). Moreover, in IB3 cells, miR-155 up-regulation is related to defective CFTR chloride channel activity (following exposure to inh-172). This leads to repression of SHIP1, a well-known effector in the regulation of inflammation, and consequently to activation of the PI3K/AKT pathway that stabilizes IL-8 mRNA through MAPK. Thus, miR-155 up-regulation contributes to the maintenance of a pro-inflammatory phenotype. Furthermore, miR-155 antagomir leads to IL-8 down-regulation in IB3-1CF cells and could represent a candidate treatment for CF. The biogenesis of miR-155 has been in part elucidated and involves the inflammatory RNA binding proteins KSRP and TTP [69]. KSRP promotes miR-155 production, while TTP down-regulates, via miR-1, miR-155 mature expression in CF lung epithelial cells. misfolded p.Phe508del CFTR proteins. A complete analysis of the pathways triggered in CF cells seems essential to identify/develop novel treatments for CF.

Figure 6: Deregulated miRNAs in different CF models. The schematic lists miRNAs that have been reported to be upregulated or down-regulated in CF (tissues, cell lines, primary cultures) versus (vs) non-CF samples. \*ALI: Air–Liquid **Figure 6.** Deregulated miRNAs in different CF models. The schematic lists miRNAs that have been reported to be upregulated or down-regulated in CF (tissues, cell lines, primary cultures) versus (vs) non-CF samples. \*ALI: Air–Liquid Interface.

Another study assessed the miRNA profile, by using Agilent microarray, of CF IB3-1 cells infected or not with Pseudomonas aeruginosa, a model mimicking the inflammatory response observed in pulmonary tissues of CF patients [73]. Two miRNAs (miR-93 and miR-494) were found to be strongly deregulated in infected cells. These two miRNAs are predicted to interact with the 3'UTR of IL-8 mRNA. Down-regulation of miR-93 was confirmed in two other bronchial epithelial cell lines (CF CuFi-1 and non-CF Nuli-1 cells). In non-infected CF cells, miR-93 is strongly expressed and is involved in IL-8 down-regulation combined with low NF-KB recruitment to the IL-8 promoter. In CF cells infected with Conversely, miR-126 down-regulation has an anti-inflammatory role to compensate the immunity response. Oglesby *et al.* [70] reported that miR-126 is consistently decreased in CF compared to non-CF airway epithelial cells and this reduction correlates with up-regulation of TOM1, a negative regulator of TLR2, TLR4, IL-1, IL-1β and TNF-alpha [71,72]. TOM1, which also negatively regulates NF-κB, may play an anti-inflammatory role in CF lung. The authors

Pseudomonas aeruginosa, the combined effects of high NF-KB recruitment to the IL-8 gene promoter and miR-93 down-

Finally, miR-31 is down-regulated in CF bronchial brushing cells compared to non-CF cells [74]. In CF epithelial cells, miR-31 negatively modulates the expression of IRF1, a transcription factor that regulates the level of cathepsin S (CTSS). CTSS is overexpressed in CF airways cell lines, such as bronchial (CFBE), tracheal (CFTE) and CF primary bronchial epithelial cells (CF-PBECs) and has been detected in CF lung secretions. CTSS activates the epithelial sodium channel and cleaves and inactivates antimicrobial proteins such as surfactant A, lactoferrin and members of the Bdefensin family, thus contributing to lung inflammation in patients with CF [63-66]. Moreover, in a cohort of paediatric patients with CF, it was found that CTSS level correlates with the decline of lung function. Thus, miR-31 is a potential

Methods to quantify miRNAs in CF and non-CF samples are depicted in Figure 5B. Approaches to identify dysregulated

lncRNAs in CF samples are detailed in Figure 7 and databases for lncRNAs are listed in Table 4.

hypothesized that the observed reduction in miR-126 expression in CF cells may be due to ER stress induced by accumulation of misfolded p.Phe508del CFTR proteins. A complete analysis of the pathways triggered in CF cells seems essential to identify/develop novel treatments for CF.

Another study assessed the miRNA profile, by using Agilent microarray, of CF IB3-1 cells infected or not with *Pseudomonas aeruginosa*, a model mimicking the inflammatory response observed in pulmonary tissues of CF patients [73]. Two miRNAs (miR-93 and miR-494) were found to be strongly deregulated in infected cells. These two miRNAs are predicted to interact with the 3'UTR of *IL-8* mRNA. Down-regulation of miR-93 was confirmed in two other bronchial epithelial cell lines (CF CuFi-1 and non-CF Nuli-1 cells). In non-infected CF cells, miR-93 is strongly expressed and is involved in IL-8 down-regulation combined with low NFκB recruitment to the IL-8 promoter. In CF cells infected with *Pseudomonas aeruginosa*, the combined effects of high NF-κB recruitment to the *IL-8* gene promoter and miR-93 downregulation lead to high IL-8 expression [73].

Finally, miR-31 is down-regulated in CF bronchial brushing cells compared to non-CF cells [74]. In CF epithelial cells, miR-31 negatively modulates the expression of IRF1, a transcription factor that regulates the level of cathepsin S (CTSS). CTSS is overexpressed in CF airways cell lines, such as bronchial (CFBE), tracheal (CFTE) and CF primary bronchial epithelial cells (CF-PBECs) and has been detected in CF lung secretions. CTSS activates the epithelial sodium channel and cleaves and inactivates antimicrobial proteins such as surfactant A, lactoferrin and members of the β-defensin family, thus contributing to lung inflammation in patients with CF [63-66]. Moreover, in a cohort of paediatric patients with CF, it was found that CTSS level correlates with the decline of lung function. Thus, miR-31 is a potential regulator of CTSS expression via IRF1 in CF epithelial cells [74].

### *4.2.3. Methods to study deregulated non-coding RNA*

deregulated in CF, such as miR-155, miR-126 and miR-31 [68]. miR-155 is overexpressed in IB3-1 CF cells compared to IB3-1 control cells and in *ex vivo* CF cells (bronchial brushing samples versus normal human bronchial epithelial cells and CF neutrophils CD66+ versus control cells). Moreover, in IB3 cells, miR-155 up-regulation is related to defective CFTR chloride channel activity (following exposure to inh-172). This leads to repression of SHIP1, a well-known effector in the regulation of inflammation, and consequently to activation of the PI3K/AKT pathway that stabilizes IL-8 mRNA through MAPK. Thus, miR-155 up-regulation contributes to the maintenance of a pro-inflammatory phenotype. Furthermore, miR-155 antagomir leads to IL-8 down-regulation in IB3-1CF cells and could represent a candidate treatment for CF. The biogenesis of miR-155 has been in part elucidated and involves the inflammatory RNA binding proteins KSRP and TTP [69]. KSRP promotes miR-155 production, while TTP down-regulates, via miR-1, miR-155 mature expression in CF lung epithelial cells.

misfolded p.Phe508del CFTR proteins. A complete analysis of the pathways triggered in CF cells seems essential to

miRNA dysregulation in CF airways

 miR-101 miR-138

CF vs non CF bronchial brushing

CFBE41o- vs 16HBE14o-

miR-126

 miR-145 miR-223 miR-494

 miR-145 miR-223 miR-494 miR-126

Figure 6: Deregulated miRNAs in different CF models. The schematic lists miRNAs that have been reported to be upregulated or down-regulated in CF (tissues, cell lines, primary cultures) versus (vs) non-CF samples. \*ALI: Air–Liquid

**Figure 6.** Deregulated miRNAs in different CF models. The schematic lists miRNAs that have been reported to be upregulated or down-regulated in CF (tissues, cell lines, primary cultures) versus (vs) non-CF samples. \*ALI: Air–Liquid

Conversely, miR-126 down-regulation has an anti-inflammatory role to compensate the immunity response. Oglesby *et al.* [70] reported that miR-126 is consistently decreased in CF compared to non-CF airway epithelial cells and this reduction correlates with up-regulation of TOM1, a negative regulator of TLR2, TLR4, IL-1, IL-1β and TNF-alpha [71,72]. TOM1, which also negatively regulates NF-κB, may play an anti-inflammatory role in CF lung. The authors

ALI\* culture of CF vs non CF airway epithelia (trachea and primary bronchi)

 miR-494 miR-509-3p miR-138

miR-146-a CF vs non CF

Another study assessed the miRNA profile, by using Agilent microarray, of CF IB3-1 cells infected or not with Pseudomonas aeruginosa, a model mimicking the inflammatory response observed in pulmonary tissues of CF patients [73]. Two miRNAs (miR-93 and miR-494) were found to be strongly deregulated in infected cells. These two miRNAs are predicted to interact with the 3'UTR of IL-8 mRNA. Down-regulation of miR-93 was confirmed in two other bronchial epithelial cell lines (CF CuFi-1 and non-CF Nuli-1 cells). In non-infected CF cells, miR-93 is strongly expressed and is involved in IL-8 down-regulation combined with low NF-KB recruitment to the IL-8 promoter. In CF cells infected with Pseudomonas aeruginosa, the combined effects of high NF-KB recruitment to the IL-8 gene promoter and miR-93 down-

Finally, miR-31 is down-regulated in CF bronchial brushing cells compared to non-CF cells [74]. In CF epithelial cells, miR-31 negatively modulates the expression of IRF1, a transcription factor that regulates the level of cathepsin S (CTSS). CTSS is overexpressed in CF airways cell lines, such as bronchial (CFBE), tracheal (CFTE) and CF primary bronchial epithelial cells (CF-PBECs) and has been detected in CF lung secretions. CTSS activates the epithelial sodium channel and cleaves and inactivates antimicrobial proteins such as surfactant A, lactoferrin and members of the Bdefensin family, thus contributing to lung inflammation in patients with CF [63-66]. Moreover, in a cohort of paediatric patients with CF, it was found that CTSS level correlates with the decline of lung function. Thus, miR-31 is a potential

Methods to quantify miRNAs in CF and non-CF samples are depicted in Figure 5B. Approaches to identify dysregulated

lncRNAs in CF samples are detailed in Figure 7 and databases for lncRNAs are listed in Table 4.

identify/develop novel treatments for CF.

292 Cystic Fibrosis in the Light of New Research

Nasal cavity brushing

IB3-1 CF vs

IB3/S9 miR-155

 miR-145 miR-494

miR-126

Interface.

Interface.

regulation lead to high IL-8 expression [73].

regulator of CTSS expression via IRF1 in CF epithelial cells [74].

4.2.3. Methods to study deregulated non-coding RNA

Methods to quantify miRNAs in CF and non-CF samples are depicted in Figure 5B. Approaches to identify dysregulated lncRNAs in CF samples are detailed in Figure 7 and databases for lncRNAs are listed in Table 4.




**Table 4.** Free databases for lncRNAs

**Name Website Characteristics References**

blast search of putative lncRNAs.

localization, functional evidence.

entries are based on literature data.

structure, genomic context.

lncRNAs.

data.

genes

and cell types.

Gene expression repository for human and mouse

Includes both microarrays and in situ hybridization

Human lncRNAs catalog from manually annotated

Includes evolutionary conservation, secondary

Data available via the UCSC Genome Browser

Expression data from RNAseq accross 24 tissues

Database containing a large collection on noncoding transcripts including annotated and nonannotated sequences from the H-inv database,

Human reference catalog for lincRNAs

lincRNA features (sequence, structure, transcriptional and orthology features).

NONCODE and RNAdb databases.

mRNA of genomic RNAs.

reaction.

Database integrating the diverse body of

experimental knowledge on functional interactions between ncRNAs (except tRNAs and rRNAs) and protein-related biomacromolecules such as proteins,

Functional interactions (both physical interactions and other forms of interactions) eliciting a cellular

Component of ScienceWiki for community curation

Database for functional eukaryotic lncRNAs. Allow

Gene expression data, evolutionary conservation, structural information, genomic context, subcellular

Integrated knowledge database dedicated to ncRNAs. NONCODE specific ID for each ncRNA with a conversion tool to RefSeq and Ensembl. Includes all ncRNAs, except transfer RNAs and ribosomal RNAs, ncRNA sequences and relative information (expression, cellular location, chromosomal information...). More than 80% of

of human lncRNAs. [109]

[110,111]

[112,113]

[114]

[119]

[115]

[116]

[117]

lncRNAWiki http://lncrna.big.ac.cn/

294 Cystic Fibrosis in the Light of New Research

lncRNAdb http://www.lncrnadb.org/

NONCODE http://www.noncode.org/

GENCODE http://

fRNAdb http://www.ncrna.org/

NPInter http://www.bioinfo.org/

NRED : ncRNA expression database

Human Body Map lincRNAs index.php/Main\_Page

http:// nred.matticklab.com/cgibin/ncrnadb.pl

www.gencodegenes.org/

http:// www.broadinstitute.org/ genome\_bio/ human\_lincrnas/

frnadb

NPInter/index.php

**Figure 7.** Strategies to study lncRNAs in CF and non-CF samples. Few techniques allow the quantitative analysis of lncRNAs in biological samples. RNAseq: global sequence screening is currently an easy option thanks to NGS and small RNA sequencing. This method allows detecting and quantitating in a biological sample all non-coding tran‐ scripts. However, as lncRNAs are very weakly expressed, this strategy might not be fully appropriate. To obtain a sig‐ nificant profile of ncRNAs, more than 180 million reads are required, whereas for protein-coding transcripts, less than 30 million are needed. Therefore, microarray analysis remains a powerful tool for global profiling of lncRNAs. In total, more than 30,586 lncRNAs have been analyzed based on the last lncRNA databases. AONs: antisense oligonucleotides.

### **4.3. Impact of mutations in lncRNA or miRNA genes**

To date, no ncRNA mutation has been described in CF. However, alterations in RNA sequence and/or structure can affect the synthesis, maturation and turnover of ncRNAs. Changes in RNA molecules can be introduced in different ways. For instance, SNPs may affect miRNA biogenesis. miRNA tailing can modify pre-miRNAs and mature miRNAs. RNA editing can modify nucleotide sequences of RNA transcripts. NGS technologies, including exome se‐ quencing and complete re-sequencing of the *CFTR* gene, could reveal mutations in lncRNA sequences that may affect CF severity and outcome.
