Exosomal Long Non-Coding RNAs in Lung Diseases

Within the non-coding genome landscape, long non-coding RNAs (lncRNAs) and their secretion within exosomes are a window that could further explain the regulation, the sustaining, and the spread of lung diseases. We present here a compilation of the current knowledge on lncRNAs commonly found in Chronic Obstructive Pulmonary Disease (COPD), asthma, Idiopathic Pulmonary Fibrosis (IPF), or lung cancers. We built interaction networks describing the mechanisms of action for COPD, asthma, and IPF, as well as private networks for H19, MALAT1, MEG3, FENDRR, CDKN2B-AS1, TUG1, HOTAIR, and GAS5 lncRNAs in lung cancers. We identified five signaling pathways targeted by these eight lncRNAs over the lung diseases mentioned above. These lncRNAs were involved in ten treatment resistances in lung cancers, with HOTAIR being itself described in seven resistances. Besides, five of them were previously described as promising biomarkers for the diagnosis and prognosis of asthma, COPD, and lung cancers. Additionally, we describe the exosomal-based studies on H19, MALAT1, HOTAIR, GAS5, UCA1, lnc-MMP2-2, GAPLINC, TBILA, AGAP2-AS1, and SOX2-OT. This review concludes on the need for additional studies describing the lncRNA mechanisms of action and confirming their potential as biomarkers, as well as their involvement in resistance to treatment, especially in non-cancerous lung diseases.


Introduction
The purpose of this review is to summarize the current knowledge in the field of long non-coding RNAs (lncRNAs) and exosomal-lncRNAs involved in lung diseases such as Idiopathic Pulmonary Fibrosis (IPF), Chronic Obstructive Pulmonary Disease (COPD), asthma, and lung cancer. To this aim, we collected information from RNAseq and microarray data when available. Furthermore, we standardized each gene and microRNA (miRNA) name to the official gene symbols and miRNA nomenclatures.
The following three significant steps describe the overall methodology used to process the literature. First, using the PubMed database, we seek for publications related to lncRNAs in each disease. Secondly, we standardized the gene names using the official gene symbols and the Ensembl identifier. Finally, we overlapped the standardized gene lists from each disease and identified ten lncRNAs associated with at least two of the diseases mentioned above. These ten lncRNAs are H19, MALAT1, MEG3, FENDRR, CDKN2B-AS1, TUG1, HOTAIR, GAS5, LINC00861, and CCDC18-AS1.  (30-150 nm), which originate from the endosomal pathway by the formation of the early endosomes, late endosomes, and, ultimately, multivesicular bodies (MVBs). These vesicles are released to the extracellular microenvironment through MVB fusion to the plasma membrane and exocytosis. Exosomes contain many components of donor cells, including cell-surface proteins, lipids, metabolites, and genetic material, which confer them functional properties. Exosomes can transfer information to the target cell by (1) interacting with the cell surface, via a receptor-mediated mechanism, or by delivering its content to the target cell through (2) endocytosis, macropinocytosis or (3) through a direct fusion of exosomal membrane with the plasma membrane.

Exosomes Are Playing an Essential Role in Lung Diseases
In a physiological context, various cell types from the lung microenvironment participate in the regulation of lung homeostasis via the secretion of exosomes presenting anti-inflammatory/anti-fibrotic properties [20,[37][38][39]. Noxious stimuli exposure (e.g., CS, allergens, infections, air pollutants) can impact nucleic acid cargo (miRNAs, lncRNAs) of lung-derived exosomes and alter their protective properties. Indeed, CS exposure induces the alteration of the composition of bronchial epithelial-derived exosomes composition, with an upregulation of hsa-miR-21 and hsa-miR-210 [22,23], which in turn dysregulates several cellular processes associated with the progression of the COPD. In the asthma context, several studies have reported an alteration of inflammatory-related exosomal miRNAs from airway biofluids (BALF, and sputum supernatants) [21,24]. In conclusion, lung exosomes released in the pathological context present an alteration of their composition, which in turn may impact the progression of lung diseases.

Long Non-Coding Rnas
Thanks to an international effort through the FANTOM (fantom.gsc.riken.jp) and the ENCODE (encodeproject.org) projects, we know that the non-coding sequences cover 98% of the human genome and that the transcribed part alone covers 90%. Within these RNA sequences that lack protein-coding capacities, lncRNAs are any expressed RNAs of more than 200 nt in length [40,41]. The current classification of lncRNAs gathers five categories, according to their original genomic location in regards to their corresponding protein-coding gene. These five categories are • intergenic, • intronic, • sense, • antisense, and • bidirectional [42]. According to the NONCODE (noncode.org) database file "NONCODEv5_human_hg38_lncRNA.gtf", 172,216 lncRNA transcripts can theoretically be found in humans [43]. From our literature screening in January 2020, we found hundreds of them potentially associated with COPD, asthma, IPF, or lung cancer. This review focuses on the lncRNAs commonly found between these four lung diseases. While we will summarize their mechanisms of action regarding these diseases, additional studies may also help to understand the full picture of lncRNAs [42,[44][45][46][47]].

Transcript-Regulating LncRNAs
Under the mRNA transcript degradation process, mature miRNAs bind to the Argonaute RISC Catalytic Components 1 to 4 (AGO1-4). This complex will then target an mRNA 3'UTR leading to the degradation of the mRNA. However, transcript-regulating lncRNAs (treg-lncRNAs) can prevent such mRNA degradation. The lncRNA can act as an RNA decoy for a miRNA, leading to a • miRNA sequestration, • miRNA degradation, or a • translational repression of the mRNA [48]. The miRNA sequestration and the miRNA degradation are part of the competing endogenous RNA (ceRNA) network, which aims at circumventing miRNAs from their original targets [49]. The sequestration controls the miRNA abundance in the cell and inhibits its activity [41]. This process also called "miRNA sponge", happens just before a miRNA could regulate its target mRNAs through a physical binding between the lncRNA and the miRNA. More precisely, the miRNA binding site, located in the 3'UTR of the lncRNA, does not allow for the degradation of the lncRNA, as it would occur for protein-coding transcripts. Instead, mismatched nucleotides in the lncRNA binding site lead to sequestration of the miRNA by the lncRNA. Subsequently, a sufficient amount of lncRNAs would act as a "sponge," disabling the mRNA regulation by the sequestrated miRNA [41]. Furthermore, one lncRNA can have multiple miRNA targets, and circular lncRNAs can also regulate the miRNA activity [47]. Interestingly, lncRNAs may be involved in positive feedback loops while targeting miRNAs. For example, a recent study by Qu et al. describes the upregulation of lncRNA ZEB1-AS1 observed in NSCLC cells. ZEB1-AS1 can sequester the miRNA hsa-miR-409-3p, which leads to an increase in the mRNA and protein levels of ZEB1. In return, ZEB1 binds the promoter region of ZEB1-AS1 and activates its expression [50].
NAT-lncRNAs Specific Regulation: Few mechanisms may be specific to Natural Antisense transcripts (NATs). NATs are RNA sequences that are complementary to and overlap with either protein-coding or non-coding transcripts [42,51]. Cis-NATs are transcribed from the same genomic locus and have a perfect complementarity with their target mRNA transcript. Trans-NATs are transcribed from a different genomic locus and have an imperfect complementarity with their target mRNA transcript [42]. The NATs may act on the transcription through transcriptional interference, RNA masking, and RNA "A to I" editing [42]. NATs may also regulate the abundance of mRNAs by • suppressing the translation through polysome displacing, • promoting the mRNA decay through 3'UTR binding, or • modulating the mRNA stability and increasing its expression level through the formation of sense or antisense pairs [44]. While the majority of the studies described their actions in cis, no study invalidates a possible operation in trans.

Epigenetics-Regulating LncRNAs
Epigenetics-regulating lncRNAs (epi-lncRNAs) are lncRNAs that may guide the polycomb chromatin domains until polynucleosome compaction [42,52]. The Polycomb group of proteins was identified as a transcriptional-repressive complex, named the Polycomb Repressive Complex (PRC). Essential members of the PRC are the PRC1 and PRC2. The PRC1 contains the CBX7 protein, and the PRC2 contains the EZH1, EZH2, EED, and the SUZ12 subunits. Importantly, the PRC2 is highly conserved between plants and animals and can create polycomb chromatin domains with the PRC1 to help polynucleosome compaction [53]. Indeed, PRC1's CBX7 and PRC2's EZH2-SUZ12 can tether epi-lncRNAs, which in return will guide the polycomb chromatin domain through its repressive action. Then, PRC2's EZH2 and EZH1 trimethylates the histone 3 at lysine 27 (H3K27me3), which will become an anchor site for the PRC1. Once the PRC1 is attached, it mono-ubiquitinates the H2A on K119 (H2AK119Ub), eventually leading to the repression of the targeted gene [42,53]. However, the PRC2 may tether many types of RNA without a precise binding site. In consequence, assigning only the lncRNAs to the recruitment of the PRC2 may be premature [54]. Moreover, epi-lncRNAs may regulate gene expressions independently [47]. For example, HOTAIR can repress a limited number of genes through H3K27me3 without the involvement of the PRC2 complex [55]. Hence, these recent observations point up the need for a better understanding of the epigenetics modulation triggered by epi-lncRNAs.
Besides, the lncRNAs are generally expressed at low levels when compared to protein-coding genes [56,57]. This low expression underlines a fine regulation of their target mRNA, which could dramatically impact the behavior of the receiving cell during intercellular communication. However, the mechanisms are not clear enough to understand how an epi-lncRNAs, which acts in the nucleus, can reach its targets in another cell through vesicle transportation.

LncRNAs and Their Exosomes in Lung Diseases
From a PubMed screening, done in January 2020, we found associations only between lung cancers and both lncRNAs and exosomes. Therefore, our strategy was first to build a list of the most published lncRNAs found in at least two lung diseases within asthma, COPD, IPF, and lung cancers. Importantly, this information comes from either the main text, the figures, or the gene expression datasets that were available. After overlapping the official gene symbols, we found the ten following lncRNAs in at least two diseases: H19, MALAT1, MEG3, FENDRR, CDKN2B-AS1, TUG1, HOTAIR, GAS5, LINC00861, and CCDC18-AS1. Interestingly, we did not find publications reporting LINC00861 and CCDC18-AS1 in lung cancers.
Hereafter we will describe the ten lncRNAs in their associated disease. We will then summarize the eight lncRNAs found in lung cancers into interaction networks and tables listing their promising clinical interests. Figures 2-4 summarize these lncRNAs and their known actions in the COPD, IPF, and asthma, respectively. Dedicated networks will then cover each of these lncRNAs in lung cancers. Tables 1 and 2 respectively provide the downstream targets of the eight lncRNAs and their possible use as biomarkers in the lung diseases mentioned above. Table 3 provides lung cancer treatment resistances associated with the eight lncRNAs. Moreover, from this list, only H19, MALAT1, HOTAIR, and GAS5 were associated with exosomes and lung cancers in our PubMed search. Therefore, we will also shortly describe results on exosomes for six additional lncRNAs that were related to only lung cancers. These lncRNAs are UCA1, lnc-MMP2-2, GAPLINC, TBILA, AGAP2-AS1, and SOX2-OT.

H19 and Asthma
Austin et al. found H19 downregulated in airway smooth muscle (ASM) cells from non-severe asthma when compared to healthy patients. Please refer to their microarray experiment in supplemental Table 10 [127]. However, the authors did not focus on H19 in their study, which requires further efforts to assess its clinical impact and mechanism of action.

H19 and COPD
H19 is upregulated in the quadriceps of the low fat-free mass index (FFMI) COPD patients when compared to normal FFMI. H19 hosts the hsa-miR-675, a miRNA also upregulated in low FFMI COPD patients. The increase of H19 expression may be the consequence of an altered methylation of its region. Besides, H19 expression is associated with the downregulation of MYOD1 and hsa-miR-519a (hsa-miR-519a-1 or hsa-miR-519a-2) in male patients with severe COPD [68]. Subsequently, we can suspect that H19 sequester hsa-miR-519a following demethylation, thus contributing to increase the susceptibility to a low FFMI for the COPD patients.

Regulators of H19:
In NSCLC tumor tissues, FOXF2 can bind the promoter of H19 and can increase its expression, causing a PTEN downregulation [135]. Additionally, H19 may be a direct transcriptional target of and is induced by MYC in NSCLC tumor tissues. Indeed, MYC binds to H19 promoter's E-boxes to facilitate histone acetylation and transcriptional initiation. Furthermore, MYC can downregulate the expression of IGF2 independently [136,137]. Besides, Shahdoust et al. found H19 among the seven most differentially expressed lncRNAs in the human airway epithelium of cigarette smokers when compared to non-smokers [138]. Similar results, obtained in cdk-4/hTERT-immortalized human bronchial epithelial cells (HBEC), described the H19 upregulation following prolonged CSC exposure. Interestingly, the same authors also found a general diminution of H4K16ac and H4K20me3 and an overall increase of H3K27me3 levels [74]. Nevertheless, H19 methylation status remains low, as RIOX2 can remove methyl groups from H3K9me3 on the H19 promoter, leading to de-repress H19 transcription [139]. Importantly, Liu et al. suggested that H19 demethylation may precede the methylations that silence tumor suppressor genes such as p16-CDKN2A, MGMT, DAPK, E-cadherin (CDH1), and CDH13 [74]. Moreover, the histone alterations coincided with a decreased DNMT1 and an increased DNMT3B expressions, as well as the activation of the WNT/β-catenin signaling pathway during prolonged CSC exposure [74]. Indeed, the authors found that WNT ligands, such as WNT2, WNT5A, WNT6, and WNT10A, and the Wnt signaling targets FOXN1 and TCF7, were up-regulated [74].
H19 as ceRNA: Among its known functions, H19 was reported as a heavy miRNA regulator. Indeed, H19 can sequester: • hsa-miR-107, to release the expression of NF1 in NSCLC tumor tissues [128] • hsa-miR-200a, to release the expression of ZEB1 and ZEB2 in NSCLC tumor tissues [130] • hsa-miR-29b-3p and hsa-miR-17, to release the expression of STAT3 in NSCLC tumor tissues [131,132] • hsa-miR-196b to release the expression of LIN28B and induce cell growth in NSCLC tumor tissues [142] • hsa-miR-138 (MIR138-1, MI0000476 or MIR138-2, MI0000455) to release the expression of PDK1 in NSCLC tumor tissues [129] • hsa-miR-484 to release the expression of ROCK2 and increase the levels of phosphorylated JUN as well as the mesenchymal markers N-cadherin (CDH2), vimentin (VIM), ZEB1 and SNAI1 while decreasing the level of the epithelial marker CDH1 in NSCLC tumor tissues and cell lines [133,134,143]. A similar observation can be made with the downregulation of hsa-miRNA-203 (hsa-miRNA-203a or hsa-miRNA-203b), which was associated with VIM and SNAI1 upregulation and CDH1 downregulation in NSCLC tumor tissues [143].
H19 at the clinical level: H19 upregulation in NSCLC tumor tissues was associated with advanced tumor-node-metastasis (TNM) stages and negatively correlated with patient Overall Survival (OS) [96,97]. H19 expression was also higher in stage III and IV NSCLC, while hsa-miR-21 expression was higher in stage I and II NSCLC when compared to non-tumor lung tissues [97]. Besides, plasma levels of H19 were significantly increased in NSCLC patients when compared to patients with benign lung disease [89]. Additionally, a nucleotide polymorphism, the H19-rs217727 C>T, was found significantly associated with an increased risk of lung cancer [84]. Consequently, with further efforts to confirm these results in large independent cohorts, H19 would make a great biomarker to diagnose or to assess a genetic predisposition to lung cancers. H19 behavior against treatment: Wang et al. found a negative correlation between the H19 upregulation in NSCLC tumor tissues and the Cisplatin (DDP) response [96].
Exosomal H19: Tumor-released lncRNA H19 (exo-H19) can promote gefitinib resistance via packaging into exosomes in NSCLCs. While the experiments were on gefitinib-resistant NSCLC cell lines, the authors assessed the expression of H19 in both gefitinib-resistant and parental sensitive cells. H19 expression was increased in gefitinib-resistant NSCLC cells and was described as secreted through the incorporation into exosomes, which was mediated by HNRNPA2B1. Moreover, exosome-mediated transfer of H19 conferred gefitinib resistance to the recipient NSCLC cells [148].

MEG3
Maternally Expressed 3 (MEG3) is a maternally expressed imprinted RNA gene localized on the cytogenetic band 14q32.2, in the DLK1-DIO3 imprinted region. MEG3 gene has 50 transcripts, all identified as lncRNAs. MEG3 is expressed in many tissues under normal conditions and interacts with TP53 [149].

MEG3 and Asthma
In CD4 + T-cells of asthmatic patients, MEG3 is upregulated when compared to healthy patients. Moreover, it displays pro-inflammatory properties linked to the increase of Th17 associated cytokines IL17A and IL22. It acts through hsa-miR-17 sequestration, which in turn releases the expression of RORγt (RORC), leading to an increase of Treg/TH17 [59].

MEG3 and Lung Cancer
Downregulated MEG3 can regulate cell proliferation, EMT, and apoptosis in NSCLC tumor tissues from patients with an advanced pathological stage, as well as in cell lines. Importantly, MEG3 was more downregulated in stages III+IV when compared to stages I+II, and it also increases cell viability and proliferation, while reducing the expression of autophagy [120]. Conversely, MEG3 was significantly less methylated in the tumor of smoker patients with clinical early-stage NSCLC, as compared to non-cancerous tissue. This MEG3 demethylated region (DMR) was associated with a hypermethylated DIO3 and two hypomethylated (DLK1 and RTL1) [150]. Subsequently, the low MEG3 expression observed in NSCLC patients with an advanced pathological stage may be due to a deletion of the MEG3-DMR locus or could be due to the deletion of a transcription factor binding MEG3 promoter [150][151][152]. MEG3 uses the following mechanisms, as summarized in Figure 6.
DDP-resistance Regulators of MEG3: MEG3 promoter methylation was reported in 96% of NSCLC tumor tissues, which mainly contributes to its downregulation [153,154]. Alternatively, phosphorylated RB1 has been described to activate DNMT1, which in turn, will also decrease the methylation of the MEG3 locus [155].
MEG3 regulated genes: Two studies realized in NSCLC tumor tissues described MEG3 as an activator of TP53 by decreasing the levels of MDM2 [149,153]. Additionally, MEG3 can decrease BCL2 expression by promoting BAX, and also decrease MAP1LC3A expression [156,157]. Besides, Su et al. described MEG3 as inversely correlated with PCNA [153].
MEG3 at the clinical level: MEG3 low-expression in NSCLC tumor tissues was associated with short-term survival in two independent public datasets [106]. Besides, MEG3 genotype rs4081134 SNP (AA) was associated with a lung cancer risk in Chinese patients [85]. While being promising, these findings should, however, be confirmed in additional large independent cohorts to classify MEG3 as a reliable biomarker.
Exosomal MEG3: From the PubMed screening, we found no studies on MEG3 associated with lung diseases and exosomes. However, three recent studies could describe exosomal MEG3 in high-grade serous carcinoma [163], cervical cancer [164], and Hunner-type interstitial cystitis [165]. These studies underline that MEG3 may be involved in intercellular communication, especially in cancers, and therefore further research on this topic is needed to assess its relevance in lung diseases.

MALAT1
Metastasis Associated Lung Adenocarcinoma Transcript 1 (MALAT1) is an RNA gene localized on the cytogenetic band 11q13.1. MALAT1 gene has 17 transcripts, all identified as lncRNAs. MALAT1 may also act as a transcriptional regulator for numerous genes and is involved in the cell cycle regulation and pre-mRNA splicing [166].

MALAT1 and Asthma
In Zhu et al. RNA sequencing data, MALAT1 was observed as upregulated in the blood of highly-expressed IgE eosinophilic asthmatic (EA) patients when compared to healthy patients [61]. The authors denoted the T cell receptor signaling as one of the main pathways impaired in EA patients. Moreover, Qiu et al. heatmap (Figure 2A) displays MALAT1 as upregulated in CD4 + cells from asthmatic patients when compared to healthy patients [59].

MALAT1 and COPD
MALAT1 was found the most abundant lncRNA in whole blood cells from former and current smokers with COPD [167].
MALAT1 at the clinical level: Several studies described a high MALAT1 level associated with a poor prognosis and short OS in NSCLC tumor tissues [76,79,[101][102][103]105]. Besides, MALAT1 level was low in the serum of patients with NSCLC; however, it still lacks some specificity [90][91][92][93]. Furthermore, both MALAT1 and SOX9 expressions were associated with age, tumor size, and TNM stage, making these two genes potential candidates for prognosis tools [76,79,104].
MALAT1 behavior against treatment: High MALAT1 expression was associated with a DDP chemo-resistance in NSCLC tumor tissues [76,103,184,188] and cell lines [182]. Interestingly, a feedback loop between MALAT1 and SOX9 may amplify this resistance [76] Exosomal MALAT1: Exosomal MALAT1 (exo-MALAT1) was described as upregulated in the serum of 77 NSCLC patients, and its expression was related to tumor stage and lymphatic metastasis [174]. While being a potential new biomarker for tumor stage diagnosis, further studies should be conducted on larger cohorts to confirm the predictive power of exo-MALAT1.

FENDRR
FOXF1 adjacent non-coding developmental regulatory RNA (FENDRR) is an RNA gene localized on the cytogenetic band 16q24.1. FENDRR gene has 14 transcripts, all identified as lncRNAs. This gene is transcribed bidirectionally with FOXF1 on the opposite strand. FENDRR may bind the polycomb repressive complex 2 (PRC2) to promote the methylation of its targeted genes.

FENDRR and IPF
FENDRR levels were low in fibrotic human lung cells and mouse primary lung fibroblasts. Interestingly, Huang et al. hypothesized that the TGFB1/SMAD3 signaling pathway might cause these low levels. Besides, FENDRR may inhibit fibroblast activation and reduces pulmonary fibrosis by capturing ACO1, thus reducing the iron levels, and by sequestrating the profibrotic hsa-miR-214 [64].

Regulators of FENDRR:
Two studies described FENDRR as hypermethylated through a presumable involvement of EZH2. This hypermethylation could explain its downregulation in NSCLC tumor tissues and cell lines (Calu-1 and H1975) [193,194].
FENDRR regulated genes: In NSCLC tumor tissues, FENDRR can specifically bind to the 3'UTR of ABCB1, thus blocking HuR (ELAVL1) binding to ABCB1 3'UTR, and therefore resulting in the decrease of ABCB1 expression [195]. Besides, Xu et al. found a negative correlation between FENDRR and ABCC10 expressions in NSCLC tumor tissues and cell lines (A549) [116].
FENDRR at the clinical level: FENDRR low-expression in tumor tissues is strongly associated with TNM 1 stage in LUAD patients. Furthermore, when associated with LINC00312, FENDRR showed a diagnostic value in detecting these LUAD patients [88]. Nevertheless, to confirm the diagnostic power of FENDRR on the TNM 1 stage of LUAD patients, additional studies should be conducted on a broader spectrum of lung cancers, including different cancer subtypes and TNM stages.

FENDRR behavior against treatment:
The low expression of FENDRR observed in NSCLC tumor tissues was correlated with chemo-resistance to DDP [116]. However, additional studies are required to confirm this result.
Exosomal FENDRR: From the PubMed search, we found no studies associating FENDRR to exosomes. However, FENDRR is an important lncRNA that controls the occurrence of metastasis. Indeed, low FENDRR expression was associated with distant metastasis and allowed the downregulation of the metalloproteinase inhibitor TIMP2 by the lack of competition with hsa-miR-761. Therefore, the metalloproteinase can degrade the extracellular matrix (ECM) and facilitates tumor metastasis [196]. It would, therefore, be interesting to seek for FENDRR in the extracellular vesicles of lung cancer associated with distant metastasis.

TUG1
Taurine Up-Regulated 1 (TUG1) is an RNA gene localized on cytogenetic band 22q12.2. TUG1 gene has 20 transcripts, all identified as lncRNAs. TUG1 interacts with the PRCs to regulate the transcription and may also act as a ceRNA targeting miRNAs.

TUG1 and Asthma
In ASM of Sprague Dawley rats, rat TUG1 was described upregulated and release rat Fgf1 through rno-miR-590-5p sequestration, thus leading the increase of cell proliferation and migration [62].

TUG1 and COPD
Gu et al. found TUG1 as upregulated, in sputum and lung tissues from COPD patients with or without a smoking history. They also demonstrated that TUG1 releases the expression of DUSP6 through the sequestration of hsa-miR-145-5p, thus contributing to the inhibition of inflammation and airway remodeling [71]. Besides, in TGFB1 treated BEAS-2B and HFL1 cells, TUG1 could block cell proliferation through the inhibition of αSMA (ACTA2) and fibronectin 1 (FN1) expressions [72].

TUG1 and Lung Cancer
TUG1 was the only lncRNA described as downregulated in NSCLC tumor tissues and upregulated in Small Cell lung cancer (SCLC) tissues and the serum of LUAD patients. Interestingly, TUG1 downregulation was significant in the tumor tissues of male donors only and was associated with Squamous Cell Carcinoma (SCC) and LUAD tumor subtypes. This decrease was, however, strongly and significantly correlated to GAS5 decrease in Female donors and combined tumors when compared to adjacent non-cancerous tissues (ANCTs) [95]. Furthermore, few studies described TUG1 associated with cell proliferation, migration, invasion, apoptosis, and autophagy [198][199][200]. It is worth underlining that Ghaforui-Fard et al. summarized the studies of TUG1 to a broader range of cancers in Table 1 of their review [198]. TUG1 uses the following mechanisms, as summarized in Figure 9.
TUG1 at the clinical level: On the one hand, TUG1 was upregulated in the serum of LUAD patients when compared to healthy serums [86,198]. On the other hand, TUG1 low expression in NSCLC tumor tissues was associated with a high TNM stage and a poor patient outcome [105,107,108,198].
TUG1 behavior against treatment: The low expression of TUG1 observed in NSCLC cells was associated with chemo-resistance to DDP. When overexpressed, TUG1 promoted the sensitivity of NSCLC cells to DDP, leading to apoptosis, in vitro, and in vivo [83].
Exosomal TUG1: From the PubMed search, we found no studies associating TUG1 with lung diseases and exosomes. However, two recent studies mentioned exosomal TUG1, in MCF-7 cells, the levels of TUG1 were moderately elevated in exosomes when compared to cells [201], and TUG1 was up-regulated in the serum exosomes of colorectal cancer (CRC) patients [202]. Besides, Guo et al. described TUG1 as ceRNA to hsa-miR-221, which enables PTEN expression [83]. Together, these results suggest that TUG1 may be involved in intercellular communication to synchronize cellular proliferation. Therefore, further research on this topic is needed to assess its relevance in lung cancers.

CDKN2B-AS1 and IPF
In the peripheral blood of IPF patients, CDKN2B-AS1 is downregulated when compared to healthy controls. The adjacent gene, CDKN2A, is transcribed simultaneously with CDKN2B-AS1 and is also downregulated in IPF patients. Importantly, Du et al. described that both CDKN2B-AS1 and CDKN2A might regulate the TP53 signaling pathway [63]. Indeed, the CDKN2A protein is known to stabilize TP53 in NSCLC [206], and the authors found the p53-signaling pathway as the top target gene-associated pathway in IPF patients.

CDKN2B-AS1 and Asthma
In their study on bronchial asthma at exacerbation (BA-E) and bronchial asthma at remission (BA-R), Ye et al. described CDKN2B-AS1 as upregulated in the plasma of patients with both types of bronchial asthma when compared to healthy subjects. Moreover, CDKN2B-AS1 is upregulated in BA-E patients when compared to BA-R patients. These authors also found a negative correlation between CDKN2B-AS1 and hsa-miR-125a in all patients, suggesting miRNA sequestration. In Tables 2 and 3 from the same study, the authors found CDKN2B-AS1 negatively correlated with FEV_1/FVC in BA-E patients, and positively correlated with pro-inflammatory cytokines, such as TNF in BA-E patients and IL17A in both BA-E and BA-R patients [58].

CDKN2B-AS1 and COPD
Ge et al. described CDKN2B-AS1 as downregulated in the plasma from acute exacerbations of COPD (AECOPD) when compared to stable COPD or healthy patients. These authors also found a negative correlation between the expression of CDKN2B-AS1 and inflammatory cytokines such as TNF, IL1B, IL17A, and Leukotriene B4 (LTB4) in both AECOPD and stable COPD patients. Furthermore, they found another negative correlation between CDKN2B and IL8 (CXCL8) only in AECOPD patients [67].

CDKN2B-AS1 and Lung Cancer
CDKN2B-AS1 was described as upregulated in NSCLC tumor tissues and cell lines and was among the top three lncRNAs with high connectivity with differentially expressed protein-coding genes [87,88,207]. CDKN2B-AS1 high expression is also known to promote cell proliferation, cell migration, and to be involved in apoptosis [208,209]. CDKN2B-AS1 uses the following mechanisms, as summarized in Figure 10.
CDKN2B-AS1 regulated genes: In NSCLC tumor tissues, Alsibai et al. found a strong positive correlation between the expressions of CDKN2B-AS1 and the tumor suppressors p15-CDKN2B and p14-CDKN2A, but not p16-CDKN2A. Interestingly, expressed CDKN2B-AS1 can stabilize the PRC complexes to repress the expression of p15, p14, and p16, leading to activate the cell cycle [207]. Few additional studies described the ability of CDKN2B-AS1 to decrease the expression of PARP1, cleaved-PARP1, and cleaved-CASP3, and to increase BCL2 and CASP3 expressions [87,121].
CDKN2B-AS1 at the clinical level: Lin et al. found a positive correlation between CDKN2B-AS1 high-expression and the differentiation grade and TNM stages in LUAD [87]. CDKN2B-AS1 high-expression in NSCLC tumor tissues was also associated with poor patient OS [94]. Intriguing results from Du et al. suggested that a low CDKN2B-AS1 expression in the peripheral blood of IPF patients may promote the occurrence of lung cancers by regulating the P53 signaling pathway [63]. However, further investigations should be conducted on IPF patients that developed lung cancer to confirm this hypothesis. Interestingly, CDKN2B-AS1 SNPs were strongly associated with the risk of developing a LUAD [210][211][212][213][214].

CDKN2B-AS1 behavior against treatment: CDKN2B-AS1 high expression increases PTX resistance of A549 cells [87]. Interestingly, Zhang et al. showed that cancer-associated fibroblasts (CAFs) contribute to the high level of MDK in the tumor micro-environment of Oral Squamous Cell Carcinoma (OSCC) tissues, thus promoting a DDP resistance via a high expression of CDKN2B-AS1 [121].
Exosomal CDKN2B-AS1: From the PubMed search, we did not find studies associating CDKN2B-AS1 with lung diseases and exosomes. Moreover, only one recent study mentioned exosomal CDKN2B-AS1 as significantly higher in the urine of BC patients when compared to healthy subjects [215]. However, CDKN2B-AS1 is an important lncRNA that can decrease the expression levels of PARP1, which plays a crucial role in DNA repair [87,121]. Since the alteration of the DNA repair mechanism is part of the hallmark of cancers, it would be interesting to seek for CDKN2B-AS1 in the extracellular vesicles of early-stage lung tumors.

HOTAIR
HOX Transcript Antisense RNA (HOTAIR) is an RNA gene localized on the cytogenetic band 12q13.13. HOTAIR gene has five transcripts, all identified as lncRNAs.

HOTAIR and COPD
In Male BALB/c mice exposed for four days with CS as well as human bronchial epithelial (HBE) cells treated with CSE, STAT3 activation led to the upregulation of HOTAIR and EZH2. Additionally, the levels of inflammatory factors, IL6 and CXCL8, as well as the EMT markers, CDH2, VIM, and ACTA2, increased, while CDH1 levels decreased. Nevertheless, these results shall be confirmed in COPD patients [69].

HOTAIR and Lung Cancer
In Nakagawa et al.'s study on NSCLC tumor tissues, 22.1% of the patients showed at least a two-fold increased expression of HOTAIR. This increase was more frequent in patients with an advanced stage of the tumor than in patients with other stages [98]. Furthermore, HOTAIR expression could upregulate cell migration and anchorage-independent cell growth [87,98,146]. Under hypoxic conditions, HOTAIR also enhances cell proliferation, migration, invasion, EMT, the formation of cancer stem cells (CSCs), and inhibits G0/G1 cell-cycle arrest and cell apoptosis [73,112,[216][217][218]. More generally, the HOX cluster-embedded lncRNAs (HOX-lncRNAs) plays a significant role in the regulation of their adjacent coding genes and several HOX-lncRNAs, including HOTTIP, HOXA11-AS, HOTAIRM1, HOXA-AS3, HOXA10-AS, HOTAIR, and HAGLR, which are dysregulated in lung cancer [218]. HOTAIR uses the following mechanisms, as summarized in Figure 11.
Regulators of HOTAIR: In the NSCLC cell line A549, HOTAIR was described as upregulated by hypoxia and CSE [216,217], and is a direct target of HIF-1α (HIF1A), which acts through interaction with putative hypoxia-responsive elements (HREs) in the upstream region of HOTAIR [216]. Besides, pro-inflammatory IL6 can activate STAT3 in an autocrine path, and STAT3 will then increase HOTAIR expression by interacting with its promoter [217]. In NSCLC tumor tissues and A549 cell line, Caveolin 1 (CAV1) was described to upregulate HOTAIR [219]. Interestingly, a specific negative regulation loop involves HOTAIR and TP53 in NSCLC tumor tissues. Indeed, two TP53 binding sites were found on HOTAIR's promoter and can suppress HOTAIR transcription after TP53 binding. HOTAIR can, in turn, modify the promoter of TP53 by increasing H3K27me3 leading to TP53 repression [220].
HOTAIR recruits the PRC2: Fang et al. recently described a negative feedback regulator loop involving HOTAIR in SCLC cell lines (NCI-H69 and NCI-H446). Indeed, HOTAIR may upregulate EZH2 and H3K27me3 levels, which in turn can repress HOTAIR, leading to change HOXA1 methylation [225]. Interestingly, both HOTAIR 5' and 3' ends may be involved in the cell cycle dysregulation. Indeed, in NSCLC cell lines (95C, 95D, and YTMLC-90), Liu et al. demonstrated that RB1 and E2F1 are both regulated by HOTAIR5' via the PRC2 (EZH2, SUZ12, and EED) complex and by HOTAIR3' via the LSD1/ CoREST/ REST complex. Both complexes may, therefore, act on the WNT/β-catenin signaling pathway and promote EMT when coupled with histone H3 lysine 27 methylation and lysine 4 demethylation [80].
HOTAIR at the clinical level: HOTAIR high expression in NSCLC tumor tissues coincides with greater tumor size, advanced TNM stage, lymph node metastasis or lymph-vascular invasion, and short disease-free interval [98]. Its expression was also related to a reduced OS in NSCLC tumor tissues [99]. Furthermore, the expressions of HOTAIR in patients with stage I and II were lower than those with stage III and IV NSCLC tumors [100]. Besides, both H19 and HOTAIR were identified as non-invasive diagnostic biomarkers in the sputum of lung cancer patients [109]. The diagnosis of head-and-neck squamous cell carcinoma (HNSCC) can also be improved by combining the high expression of HOTAIR to the high expression of CASC9 [231]. Altogether, these encouraging results underline the potential of HOTAIR as a diagnostic biomarker.
HOTAIR behavior against treatment: HOTAIR high-expression contributes to DDP resistance via CDKN1A downregulation in LUAD tumor tissues, and experimental downregulated HOTAIR in A549 cells promoted DDP sensitivity [112,113]. It also contributes to Atractylenolide 1 and Erlotinib resistances by activating PDK1 and EZH2, in LUAD cells, in vitro and in vivo [114]. Besides, the downregulation of HOTAIR can increase the SCLC cell lines' sensitivity to DDP, Adriamycin, and Etoposide, through decreasing DNMT1 and DNMT3B expressions, leading to the reduction of HOXA1 methylation [115]. Furthermore, the downregulation of HOTAIR can increase NSCLC cell lines sensitivity (A549, H460, H1299, NCI-H460, HCC-827) to Crizotinib through the inhibition of ULK1-phosphorylation. This sensitivity leads to the suppression of tumor growth and triggers the cell cycle arrest and the apoptosis signaling pathway [81].
Exosomal HOTAIR: Exosomal HOTAIR (exo-HOTAIR) appeared in the exosomes from bronchoalveolar lavage (BAL) of smokers, NSCLC, and healthy patients, but without significant differences between the three conditions [232]. However, in a recent letter to editors, Zhang et al. found exo-HOTAIR more expressed in the blood samples from LCC patients when compared to LUAD or SCC patients. Moreover, these authors described that A549 and H1299 cells treated with exo-HOTAIR increased the level of cellular HOTAIR. The authors concluded that exo-HOTAIR promotes proliferation, migration, and invasion of the cells through the sequestration of hsa-miR-203 (hsa-mir-203a or hsa-mir-203b) [233]. Interestingly, this interaction between hsa-miR-203 and HOTAIR was also reported with similar effects in renal cell carcinoma cells. In this study, Dasgupta et al. described that the sequestration of hsa-miR-203 decreases CDH1, PTEN, CDKN1A, and CDKN1B levels, while it increases the expression of VIM [234].

GAS5
Growth Arrest Specific 5 (GAS5) is an RNA gene localized on the cytogenetic band 1q25.1. GAS5 gene has 31 transcripts, 20 identified as retained introns, and 11 identified as lncRNAs. GAS5 can bind the DNA binding domain of the glucocorticoid receptor, which disable it from regulating the transcription of its target genes.

GAS5 and Asthma
In Qiu et al. study, the heatmap Figure 2A displays an upregulated GAS5 in CD4 + T-cells from patients with severe asthma vs. healthy patients, with a fold change greater than 2 [59]. These results add to Keenan et al. previous observations in bronchial epithelial cells (BEAS-2B) and primary human airway smooth muscle (ASM) cultures. Indeed, pro-inflammatory mediators, TNF, and IL1α (IL1A) were observed to promote GAS5 upregulation in ASM and BEAS-2B cells, which in turn can modulate glucocorticoid activity and thus may mediate glucocorticoid insensitivity [60].

GAS5 and Lung Cancer
GAS5 expression levels are low in lung cancers. Interestingly, this decrease in NSCLC is significant in male donors only [95]. GAS5 is involved in cellular proliferation, metastasis, and autophagy. GAS5 uses the following mechanisms, as summarized in Figure 12.
GAS5 regulated genes: In NSCLC tumor tissues, GAS5 decrease correlates strongly and significantly to the decrease of FAS-AS1 and THRIL in male donors and combined tumors, as well as the increase of NEAT1 in male donors and combined tumors. GAS5 decrease also correlates with TUG1 increase in female donors and combined tumors, and with PVT1 increase in female donors when compared to ANCTs. GAS5 expression is also associated with the upregulation of IGF2BP2 and the downregulation of FXR1 [95]. Moreover, GAS5 downregulates the expression of TNRC6A, ZC3H7B, and UPF1, while it can upregulate the expression of EIF4A3, TIA1, TIAL1, and HNRNPC [95]. Besides, in NSCLC tumor tissues and cell lines (A549, H1299, H1975, HCC827), GAS5 can deregulate the expression of phospho-EGFR, phospho-MAPK1, phospho-AKT1, and IGF1R. Interestingly, GAS5 overexpression inversely correlates with the activation of the EGFR pathway [110].
Exosomal GAS5: In Cheng et al. study on urethane-induced lung cancer mouse model, lung cancer-derived exosomal GAS5 (exo-GAS5) affects the proliferation, apoptosis, and tube formation of human umbilical vein endothelial cells (HUVECs). The overexpression of GAS5 leads to an increase of exo-GAS5, upregulates PTEN expression, and inhibits the phosphorylation of PI3K/AKT, through hsa-miR-29b-3p sequestration [238]. Furthermore, exo-GAS5 expression was lower in the serum of 64 NSCLC patients when compared to healthy controls. This low expression of exo-GAS5 was associated with larger tumor size and advanced TNM [239]. While being a potential new biomarker for the diagnosis of Stage I NSCLCs, further studies should be conducted on larger cohorts to confirm the predictive power of exo-GAS5.

LINC00861
Long Intergenic Non-Protein Coding RNA 861 (LINC00861) is an RNA gene localized on the cytogenetic band 8q24.13. LINC00861 gene has 9 transcripts, all identified as lncRNAs.

LINC00861 in Asthma
In RNAseq available data, LINC00861 transcript 201 (LINC00861-201) was found upregulated in eosinophilic asthma and eosinophilic asthma with high Igg patients vs. healthy patients [61].

LINC00861 in COPD
In COPD smokers, LINC00861 expression is lower than in non-smoking COPD or healthy patients. LINC00861 and the uncharacterized LOC101928100 RNA gene were co-expressed with RORA, while the authors observed an upregulation of hsa-miR-218-5p. The authors hypothesized that the LINC00861/LOC101928100 upregulation releases the expression of RORA through the sequestration of both hsa-miR-218-5p and hsa-miR-15a [240].

CCDC18-AS1 in Asthma
In asthma and severe asthma when compared to healthy controls, CCDC18-AS1 was upregulated, as shown by Figure 2B of Chen et al. study [241].

Exo-UCA1
Urothelial Cancer-Associated 1 (UCA1) is an RNA gene localized on the cytogenetic band 19p13.12. UCA1 has 45 known transcripts, all of them described as lncRNAs. UCA1 can promote cell proliferation and resistance to gefitinib-induced cell apoptosis. It can also sequester miR-143 to release FOSL2 expression, leading to gefitinib resistance of epidermal growth factor receptor-positive (EGFR+) NSCLCs [243].
exo-UCA1 behavior against in lung cancer treatment Kwok et al. demonstrated that the transfer of extracellular Vesicle-Associated-RNAs could induce drug resistance in ALK-Translocated lung Adenocarcinoma [244]. Moreover, exo-UCA1 levels are high in both gefitinib-resistant NSCLC cells and their secreted exosomes [243].

Exo-GAPLINC
Gastric Adenocarcinoma Associated, Positive CD44 Regulator, Long Intergenic Non-Coding RNA (GAPLINC) is an RNA gene localized on the cytogenetic band 18p11.31. GAPLINC has four transcripts, all described as lncRNAs. Exosomal GAPLINC (exo-GAPLINC), was described to promote erlotinib resistance in NSCLCs [246]. Erlotinib is a tyrosine kinase inhibitor that is effective in patients with or without EGFR mutations but appears to be more effective in patients with EGFR mutations.

Exo-TBILA and Exo-AGAP2-AS1
TGF-Beta Induced lncRNA (TBILA) is an RNA gene localized on the cytogenetic band 3q13.2. AGAP2 Antisense RNA 1 (AGAP2-AS1) is an RNA gene localized on the cytogenetic band 12q14.1. Both TBILA and AGAP2-AS1 have a unique transcript identified as a lncRNA. Both exosomal TBILA (exo-TBILA) and exosomal AGAP2-AS1 (exo-AGAP2-AS1) expression are higher in the serum of NSCLC patients than in exosomes-depleted serum (EDS) from NSCLC patients or serum of healthy patients. Importantly, these two lncRNAs are very stable in the bloodstream, which makes them promising biomarkers. Moreover, they have only one known transcript each, which reduces potential screening errors. TBILA can discriminate all NSCLC patients, while AGAP2-AS1 is better as distinguishing SCC patients from healthy controls. Additionally, the authors mentioned that the combination of TBILA/AGAP2-AS1 with Cyfra21 (KRT19), a protein widely used in clinical practices, could distinguish all NSCLC patients from healthy controls in their study [247].

Exo-SOX2-OT
SOX2 Overlapping Transcript (SOX2-OT) is an RNA gene localized on the cytogenetic band 3q26.33. SOX2-OT has 104 known transcripts, all identified as lncRNAs Exosomal SOX2-OT (exo-SOX2-OT) was significantly upregulated in Lung Squamous Cell Carcinoma (LSCC) patients when compared to non-LSCC patients. Exo-SOX2-OT levels in plasma correlated with tumor size and TNM stages, and might also reflect SOX2-OT expression in tumors [248]. Nevertheless, with 104 known transcripts, all identified as lncRNAs, further studies should identify which transcripts are useful for diagnosis.

Concluding Remarks and Future Perspectives
LncRNAs are promising molecules for the better understanding of protein-gene regulations and subsequent pathways that define lung diseases. In the current review, we attempted to cover well-described lncRNAs associated with at least two lung diseases within asthma, IPF, COPD, and lung cancers. First, we searched into the PubMed database for publications related to lncRNAs in each of the four lung diseases. Secondly, we built and described the networks of molecular interactions of the lncRNAs H19, MALAT1, MEG3, FENDRR, CDKN2B-AS1, TUG1, HOTAIR, and GAS5, within each disease. Thirdly, we reported the clinical relevance of each of these lncRNAs, focusing on the biomarker and the treatment response aspects. Finally, we covered ten additional lncRNAs that were described only in lung cancers under their exosomal form.
Based on the literature covered here, it is evident that the interaction networks are far more complex than those presented here. Indeed, these networks are an extension of the following signaling pathways found in common for different diseases or lncRNAs: • The WNT/β-catenin signaling pathway, has been associated with H19, MALAT1, and HOTAIR in NSCLC [74][75][76]80] • FENDRR and H19 are upstream regulators of the fibrosis and associated with the TGFB/SMAD3 signaling pathway in IPF [64,65] • MEG3 and HOTAIR target the apoptosis pathway in COPD and NSCLC, respectively [70,81]. • GAS5, TUG1, H19, and MALAT1 are upstream regulators of the PTEN/PI3K/AKT signaling pathway in NSCLC [77,78,82,83,111,135,235] • ANRIL (CDKN2B-AS1) is an upstream regulator of the P53 signaling pathway in both IPF and NSCLC [63]. Hence, all these studies on lncRNAs increased the list of upstream regulators of crucial cancer signaling pathways.
These lncRNAs have, therefore, the potential to sustain the tumor state of a cell. They were otherwise previously described as oncogenes and tumor-suppressors in lung cancers. In this context, H19, HOTAIR, MALAT1, and CDKN2B-AS1 are oncogenic lncRNAs [87,176,[249][250][251], while MEG3, FENDRR, and GAS5 are tumor suppressor lncRNAs [116,154,[252][253][254][255]. Intriguingly, TUG1 can be one or the other regarding lung cancer subtype. TUG1 would, therefore, act as an oncogene in SCLCs and many human cancers, but would act as a tumor-suppressor in NSCLCs [256]. Therefore, such versatile behavior suggests either a tumor-specific mechanism of action or the presence of an upstream regulator with a tumor-dependent expression.
Moreover, we can link these interaction networks to the hallmarks of cancer. Due to their implication in multiple steps of the cancer progression, we can group the eight lncRNAs according to Hanahan and Weinberg's classification [257] under the following hallmarks: Genome instability and mutation: MALAT1 and CDKN2B-AS1 may be key players of the "Genome instability and mutation" hallmark since they can decrease the expression levels of PARP1 [87,121,171]. H19 may also contribute to the "Genome instability and mutation" as well as the "Evading growth suppressors" hallmarks. CSC exposure induces an overall increase of H3K27me3 levels, which would repress many genes [74]. However, RIOX2 may demethylate H19 before the DNA-repair gene MGMT and the cyclin-dependent kinase inhibitor p16-CDKN2A are methylated [74,139]. Subsequently, a decreased level of MGMT would lead to chromosomal alterations, while a decrease in p16-CDKN2A would lead to inhibit the cell cycle arrest in G1 and G2 phases.
Activating invasion and metastasis: FENDRR may be a key player of the "Activating invasion and metastasis" hallmark. The low FENDRR expression observed in NSCLC tumor tissues allows the ECM degradation by the metalloproteinases and thus facilitates the metastasis. Indeed, under normal conditions, FENDRR is supposed to sequester hsa-miR-761, which will permit the increase of the metalloproteinase inhibitor TIMP2, leading to the degradation of the extracellular matrix [196].
The eight lncRNAs may provide an interesting angle to circumvent the treatment resistances observed in lung cancers. Throughout the review, we listed treatment resistance to Cisplatin, Vincristine, Paclitaxel, Erlotinib, Atractylenolide 1, Adriamycin, Etoposide, Crizotinib, Gefitinib, and ionizing radiations. Surprisingly, HOTAIR alone was involved in the resistance to Crizotinib, Cisplatin, Erlotinib, and Atractylenolide 1 in NSCLC cells [81,[112][113][114], and to Cisplatin, Adriamycin, and Etoposide in SCLC cells [115]. Although these results require confirmation, HOTAIR could be a promising target for future treatments. These results also underline the importance of studying lncRNAs in lung cancer.
The lncRNAs described here may be promising non-invasive biomarkers for the diagnosis or prognosis of asthma, COPD, and lung cancers. Indeed, GAS5, MEG3, MALAT1, or CDKN2B-AS1 may be used to diagnose asthma [58, 59,61]. TUG1 could be used to diagnose the COPD, and in association with CDKN2B-AS1, it could also help predict acute exacerbations in COPD patients [67,71]. In IPF patients, CDKN2B-AS1 is again a new biomarker that could predict the occurrence of lung cancers [63]. In lung cancers, MALAT1 and H19 are interesting for the diagnosis of NSCLCs [89][90][91][92][93]. TUG1 is also a promising biomarker for diagnosis of LUAD [86], whereas H19 or HOTAIR can help discriminate SCC from LUAD [109]. Nevertheless, while being very encouraging results, most of these studies lack some confirmation from additional studies in large independent cohorts. Furthermore, the majority of these lncRNAs studied as individual biomarkers may help define better the classification of lung diseases. For example, using machine learning algorithms on previously established biomarkers combined with these lncRNAs, new disease subtypes may be revealed.
Accumulative pieces of evidence display lncRNAs as communication entities to regulate their micro-environment, thus being potentially involved in a local disease spread. While non-cancerous diseases have received some attention, such as osteoarthritis or chronic inflammatory diseases including rheumatoid arthritis, systemic lupus erythematosus, and psoriasis [258], we mentioned the lack of studies covering exosomal-lncRNAs in asthma, COPD, and IPF. Hence, further efforts are needed to identify exosomal-lncRNAs in these three lung diseases. Conversely, the studies on exosomal-lncRNAs in lung focused on cancers. Although we did not find any shreds of evidence that exosomes transport MEG3, FENDRR, TUG1, and CDKN2B-AS1 in lung cancers, the mechanisms of action of these lncRNAs strongly suggest their involvement in cancer invasion. First, these lncRNAs may trigger important hallmarks of cancers, such as "Genome instability and mutation" for CDKN2B-AS1 [87,121], "Activating invasion and metastasis" for FENDRR [196], "Resisting cell death" for MEG3 and CDKN2B-AS1 [118,154,157,160,207], and "Sustaining and proliferative signaling" for TUG1 [83]. Secondly, previous studies in other cancers found an exosomal expression of MEG3, TUG1, and CDKN2B-AS1 [163][164][165]201,202,215]. These findings suggest that MEG3, TUG1, and CDKN2B-AS1 may also be involved in lung cancer intercellular communication. Surprisingly, FENDRR remains unassociated with exosomes in any diseases. Nevertheless, its implication in invasion and metastasis [196] suggests a late expression during the tumor progression, which could explain the lack of literature.
Altogether, throughout these numerous studies, we first underlined the importance of describing the lncRNAs mechanisms of action. By continuously seeking their putative regulators and downstream targets, new diagnostic tools and further treatments may arise. This review also compiled knowledge on exosomal-lncRNAs in lung diseases, and emphasize the essential position of some lncRNAs in the transcription regulation. Hence, we would be thrilled to read more investigations, in lung cancer, on exosomal MEG3, FENDRR, TUG1, and CDKN2B-AS1 or treatment resistance involving HOTAIR. We would be even more thrilled to read investigations on exosomal-lncRNAs in asthma, COPD, and IPF. Here, we highlighted eight well-described members of the lncRNAs' ocean. While these lncRNAs may trigger new research interests, it is of prime importance to identify all of them and to describe their mechanisms of action to better understand lung diseases.

Conflicts of Interest:
The authors declare no conflict of interest.

Abbreviations
The following abbreviations are used in this manuscript: