SLC6A14 Modulates Lung Disease Severity in Cystic Fibrosis and Affects mTOR Phosphorylation and Bronchial Epithelial Repair

Cystic brosis (CF), due to variants in CFTR gene, is associated with chronic infection/inammation responsible for airway epithelium alteration and lung function decline. Modier genes induce phenotype variability between people with CF (pwCF) carrying the same CFTR variants. Among these, the gene encoding for the amino acid transporter SLC6A14 has been associated with lung disease severity and age of primary airway infection by the bacteria Pseudomonas aeruginosa. In this study, we investigated whether the single nucleotide polymorphism (SNP) rs3788766, located within SLC6A14 promoter, is associated with lung disease severity in a large French cohort of pwCF. We also studied the consequences of this SNP on SLC6A14 promoter activity and the role of SLC6A14 in mammalian target of rapamycin (mTOR) signaling pathway and airway epithelial repair. We conrm that SLC6A14 SNP rs3788766 is associated with lung disease severity in pwCF (p=0.020; n=3,257, pancreatic insucient, aged 6 to 40 years old), with the minor allele G being deleterious. In bronchial epithelial cell lines decient for CFTR, SLC6A14 promoter activity is reduced in the presence of the rs3788766 G allele. SLC6A14 inhibition with a specic pharmacological blocker reduced 3 H-arginine transport, mTOR phosphorylation and bronchial epithelial repair rates in wound healing assays. In conclusion, SLC6A14 rs3788766 G allele is associated with lower lung function in pwCF. SLC6A14, whose transcriptional promoter activity varies according to rs3788766 genotype, is involved in mTOR signaling and bronchial epithelial repair. This study suggests that SLC6A14 might inuence CF lung phenotype via mTOR and epithelial repair mechanisms modulation.


Introduction
Cystic brosis (CF), the most common lethal autosomal recessive genetic disease in the Caucasian population, results from pathogenic variants in the CF transmembrane conductance regulator (CFTR) gene [1][2][3]. Manifestations of CF occur in several organs including the pancreas, liver and the intestine [4][5][6], but lung disease is the main cause of morbidity and mortality in people with CF (pwCF). CF lung disease is characterized by chronic airway colonization with microorganisms, including the most common CF life-threatening pathogen Pseudomonas aeruginosa, exacerbated in ammation and lung tissue damage due to abnormal repair of the airway epithelium [7].
Variability in the clinical phenotype of pwCF carrying identical CFTR variants and living in the same environment involves genetic modi ers, which are expected to contribute to almost 50% of CF lung phenotype [8]. In a large cohort of pwCF (n = 6,365), we previously identi ed ve CF lung disease modi er loci by genome wide association studies (GWAS) [9], including the solute carrier family 6 member 14 gene (SLC6A14, also known as ATB 0,+ ). SLC6A14 is located on chromosome X and encodes for the neutral and cationic amino acid transporter SLC6A14 that concentrates amino acids into cells (with the exception of proline, glutamate and aspartate) by using a sodium and chloride electrochemical gradient [10][11][12]. This transporter has been shown to be upregulated in several cancers [13] and to be involved in cell growth, proliferation and the mechanistic target of rapamycin (mTOR) pathway [14].
The SLC6A14 gene has pleiotropic effects in pwCF [13], with several SLC6A14 single nucleotide polymorphisms (SNPs) being associated with different phenotypes such as meconium ileus (MI) occurrence, a severe neonatal intestinal obstruction [15], lung disease severity [9,16,17], and age at rst P. aeruginosa infection [16,17]. In particular, the SNP rs3788766, located within the promoter region of SLC6A14, has been previously associated with both MI and lung function variability in pwCF [17,18,16]. However, the functional consequences of this SNP on SLC6A14 expression and function are still unknown.
Here, we rst analyzed the association between SLC6A14 SNP rs3788766 and lung function by genotyping a large French cohort of pwCF. Then, we studied how rs3788766 regulate transcriptional activity of the SLC6A14 promoter. Finally, we investigated the possible consequences of SLC6A14 activity modulation in CF bronchial epithelial cells.

Material And Methods
Genotype-phenotype association study Participants As of January 31st, 2021, 4,975 pwCF had been included in the French CF modi er gene study [9]. The study was approved by the French Human Ethics Committee (CPP n°2004/15), and information collected by the Commission Nationale de L'informatique et des Libertés (n°04.404). Informed written consent was obtained from each participant and/or parents or guardians. For the analysis, we excluded pancreaticsu cient participants since they had milder disease (n = 820), non-genotyped participants for the SLC6A14 SNP rs3788766 (n = 436), and participants without forced expiratory volume in 1 s (FEV 1 ) measurements (n = 177). We also excluded participants under 6 years of age, since their spirometry data was less reliable, as well as those over 40 years old, in order to limit selective survival bias (n = 285).

Lung function and genotyping
Measurements of FEV 1 were quarterly collected according to international CF care recommendations [19] and expressed as percent-predicted values (ppFEV 1 ) using Global Lung Function Initiative (GLI) equations [20]. To assess the lung disease severity, FEV 1 were transformed to the Survival Adjusted Kulich Normalized (SaKnorm Z-value) CF-speci c lung phenotype [21]. Lung function and lung disease severity were analyzed over the last 3 years; except for post-lung transplant patients and patients under CFTR modulator therapy (ivacaftor and lumacaftor-ivacaftor) for whom FEV 1 measurements were analyzed over the 3 years prior to the event. SLC6A14 SNP rs3788766 was genotyped using Kompetitive Allele Speci c PCR (KASP) chemistry (LGC, Teddington, UK).

Immunohistochemistry
Human lung biopsies were obtained from a healthy 51-year-old male smoker and from a 29 years old male with CF (homozygous for the F508del variant), respectively from Hôpital Foch, Suresnes 92150, France. Biopsies were collected and processed in compliance with the current French public health legislation (articles L.1235-2 and L.1245-2, code de la santé publique, www.legifrance.gouv.fr). The institution informed the participant and made sure that they were not opposed to the use of surgical samples for research purposes. Staining was performed using 5-µm thick para n sections from formalin-xed para n-embedded lung biopsies. Immunolabeling for SLC6A14 was performed on a Bond-III® automat (Leica, Leica Biosystems, Nussloch, Germany) using anti-SLC6A14 antibody (PA5-51855, Invitrogen, Carlsbad, CA, USA; 1:100). Reagents SLC6A14 inhibitor α-methyltryptophan (α-MT; M8377 was from Sigma-Aldrich, Saint-Quentin Fallavier, France) was solubilized in 100% methanol (MeOH) to achieve 1, 2.5 or 5 mM (used concentrations are speci ed in gures legends). Equivalent volumes of MeOH alone were used for control conditions, thus reaching 0.87%, 2.18% and 4.36% respectively.
Non-CF and CF primary human bronchial epithelial cells (HBEC) (Epithelix, Plan-les-Ouates, Switzerland) (characteristics in Table 3) were grown on plastic supports (1x10 5 cells/well) until con uent in hAEC culture medium supplemented with antibiotics (Epithelix). Gaussia luciferase (GLuc) reporters driven by the SLC6A14 promoter harboring either the A or G allele of rs3788766 (GeneCopoeia, Rockville, MD, USA) were used. Calu-3-CFTR-KD cells were seeded in 24-well plates and transfected at 60% con uence with 1 µg/ml of the GLuc SLC6A14 promoter reporter using Lipofectamine 3000 (Invitrogen). After 24 h of transfection, culture media were collected, centrifuged 10 min at 10,000x g and supernatants were stored at -20°C. SLC6A14 promoter activity was quanti ed by measuring Gaussia luciferase and secreted embryonic alkaline phosphatase (SEAP) used as endogenous reporter, was measured using Secrete-Pair™ Dual Luminescence Assay kit (GeneCopoeia). SLC6A14 promoter activity is represented as the ratio of GLuc normalized by SEAP.
L-Arginine uptake quanti cation SLC6A14 amino acid transport was studied as described by others [23]. Brie y, cells cultured in 12-well plates were washed and then incubated in HEPES buffer ( Wound-healing assay Cell monolayers grown on plastic supports were injured mechanically (3 wounds per well) as previously described [24,25]. Afterwards, cells were washed with their culture medium to remove detached cells and treated with either α-MT (1, 2.5 or 5 mM) or MeOH which acted as vehicle control (respective equivalent volumes of MeOH alone). Photographs of the wounds were taken at two different positions on each wound using an inverted microscope with a X4 objective at t = 0 h and t = 6 h post-wounding. Images were analyzed with Image J software (https://imagej.nih.gov/ij/index.html) to measure areas of the wounds at t = 0 h and t = 6 h and mean wound closure (% of the area at t = 0 h) were calculated.

Statistical analysis
Genotype-phenotype analysis Descriptive statistics are reported as mean ± standard deviation (SD) or percentages. Association between lung disease severity and SLC6A14 rs3788766 genotypes was evaluated by linear regression. We applied additive SNP coding, and the reference allele (i.e., the allele with the highest frequency in the European population) was taken from annotations of the human genome (http://www.ensembl.org).
Fisher's exact test was used to test conformance of the allele frequencies with the Hardy-Weinberg equilibrium. A P-value of less than 5% was interpreted as evidence of a statistically signi cant association. Analyses were carried out using R software (version 3.6.3, http://www.R-project.org/).

In vitro data
All data are presented as mean ± SEM and the number of repeated experiments is indicated in the gure legends. GraphPad Prism version 7.05 (GraphPad Software, San Diego, CA, USA) was used to analyze all data. Paired or unpaired t-tests were used to compare two groups. One-way ANOVA were used for comparison of more than two groups and followed by appropriate post-hoc tests as indicated in the gure's legends. Values of p<0.05 were considered to be signi cant. In gures, statistical differences are indicated as p < 0.05 (*), p < 0.01 (**) and p < 0.001 (***) or non-signi cant (NS).

Results
SLC6A14 SNP rs3788766 is associated with lung function in people with CF Among the 4,975 pwCF included in the French CF modi er gene study and after application of the exclusion criteria, 3,257 pwCF were analyzed in the genotype/phenotype association study. Demographic characteristics of the participants are summarized in Table 1. The minor allele frequency (MAF) of SLC6A14 SNP rs3788766 in our cohort was 0.38, similar to that reported in Europeans (0.36). As SLC6A14 is located on the X chromosome, Hardy-Weinberg equilibrium p-value was computed by Fisher's exact test among females. Results showed that our cohort does not signi cantly diverge from Hardy-Weinberg equilibrium (p-value = 0.459). We also found that SLC6A14 SNP rs3788766 was associated with lung function with the G allele being deleterious. Linear regression models estimated that pwCF carriers of the minor allele G had a signi cant increase in lung disease severity, which was measured by an average loss in SaKnorm Z-value of 0.038 ± 0.016 for each G allele (p = 0.020) ( Table 2). Overall, an average decrease of ~ 1.5% of ppFEV 1 was observed in patients carrying at least one rs3788766 G allele (  The SLC6A14 SNP rs3788766, located within SLC6A14 promoter (Fig. 1a), is likely to affect SLC6A14 mRNA expression. Quantitaive-trait Loci (QTL) expression data extracted from GTEx (https://www.gtexportal.org/home/) show that rs3788766 G is associated with a decrease of SLC6A14 transcript expression in different tissues (Fig. 1b), including the pituitary (p-value = 3.0x10 − 8 ) and the minor salivary gland (p-value = 1.8x10 − 5 ). In the lung, a diminished but not statistically signi cant expression is observed (AA vs GG genotypes). It is worth mentioning that, for the GTEx project, the prefered location for the lung tissue collection is in the inferior segment of the left upper lobe, 1 cm below the pleural surface, avoiding any large arteries, veins, and bronchi. We observed by immunohistochemistry that SLC6A14 is predominantly expressed in the bronchial epithelium of both healthy participants and those with CF (Fig. 1c). Therefore, SLC6A14 expression using QTL analysis might not inform about SNP consequences on SLC6A14 expression in the lung. Thus, to determine whether rs3788766 affects SLC6A14 promoter activity in bronchial epithelial cells, we used SLC6A14 promoter reporters carrying either the A or the G allele of this SNP and assessed the reporters expression. We observed that Calu-3-CFTR-KD cells transfected with SLC6A14 promoter reporter plasmid carrying the G allele had a lower luciferase activity (12.1% reduction) compared to cells transfected with the A allele (Fig. 1d). This result indicates that the G allele of rs3788766, i.e. the minor allele, is associated with a decreased SLC6A14 promoter activity.

SLC6A14 inhibition in human bronchial epithelial cells regulate mTOR phosphorylation and epithelial repair
Since the SLC6A14 rs3788766 G allele is likely to reduce SLC6A14 mRNA expression level and consequently its activity as an amino acid transporter, we sought to investigate the cellular consequences of a decreased activity of SLC6A14 in CF bronchial epithelial cells. Thus, we inhibited its activity in bronchial epithelial cells using α-MT, a speci c pharmacological blocker of SLC6A14 [26]. Firstly, we showed that α-MT induced a 54%-decrease of 3 H-Arginine transport in Calu-3-CFTR-KD cells (Fig. 2a). A similar effect was observed in Calu-3-CFTR-WT (Fig. S2a). We then measured LDH release to ensure that α-MT was not toxic at the doses used (Fig. S1). We showed that cytotoxicity levels in Calu-3-CFTR-KD treated with 1mM, 2.5mM and 5mM of α-MT were similar to those of cells treated with vehicle (Fig. S1a).
As SLC6A14 is involved in colon cancer cell proliferation, migration and invasion [27], we wondered whether it could play a role in bronchial epithelial repair, a process which involves both cell proliferation and migration mechanisms. Thus, we performed scratch assay experiments on Calu-3-CFTR-KD monolayers, treated or without αMT at t = 0 h and for the following 6 h of repair (Fig. 2b and c). Quantitative analysis highlighted a dose-dependent inhibition of wound closure with decreases of 8%, 27% and 39%, at 1mM, 2.5mM and 5mM of α-MT, respectively (Fig. 2b). In contrast, no signi cant wound closure inhibition was observed at 1mM and 2.5mM of α-MT in Calu-3-CFTR-WT, but a 31% decrease of wound closure was observed at 5mM of α-MT (Fig. S2b). To determine that this effect was not restricted to the Calu-3 cell lines, we performed similar experiments in primary HBECs isolated from patient with CF homozygous for the F508del CFTR variant (Fig. 3) or from healthy subjects (Fig. S3). Signi cant decreases of 25% and 36% of arginine transport were observed in CF HBECs treated with 2.5mM and 5mM of α-MT, respectively (Fig. 3a). Similar to the Calu-3 cells, no increase in cytotoxicity has been observed in non-CF HBECs treated or not with α-MT (Fig. S1b). A decrease of 32% and 56% of wound closure after 6 h of repair was observed in CF HBECs with 2.5mM and 5mM of α-MT (Fig. 3b), respectively, while 26% and 79% decreases were found in non-CF HBECs with 2.5mM and 5mM of α-MT ( Fig. S3a), respectively. Finally, we wondered if SLC6A14 amino acid transport inhibition could have an impact on mTOR activity in primary bronchial epithelial cells, as it was previously shown in pancreatic and colonic cells [14,27]. Therefore, we evaluated mTOR activation by western blot in primary HBECs treated or not with 2.5mM of α-MT. From results generated, we observed that SLC6A14 activity inhibition induces a signi cant decrease of mTOR phosphorylation in primary CF (Fig. 3c)

cells. A similar effect is
observed in non-CF (Fig. S3b) HBECs, however without reaching signi cance.

Discussion
Lung disease severity is highly variable among pwCF, with CFTR, the environment and modi er genes all contributing to this variability. Among the modi er genes, SLC6A14 is of particular interest because it has been associated with both lung and digestive phenotypes in pwCF [9,[16][17][18]. This suggests a major pleiotropic role of SLC6A14 in the overall pathophysiology of the disease [13]. Here, we con rmed the association between SLC6A14 SNP rs3788766 and the lung function of pwCF, and further demonstrated that carrying the minor allele of rs3788766 induces a decreased SLC6A14 promoter activity. We nally demonstrated that a reduced SLC6A14 amino acid transport activity alters wound repair mechanisms and modulates the mTOR pathway in human CF bronchial epithelial cells.
Our study rst showed, using a large French CF cohort (n = 3,257), that pwCF carrying at least one minor allele G of the SLC6A14 SNP rs3788766 exhibit reduced lung function compared to those carrying two major allele A, con rming prior studies [16,17]. Li et al. performed a sub-analysis from the original 1,661 Canadian CF Gene Modi er Study participants of a previous study [18] and showed that rs3788766 is associated with both pediatric lung disease severity and earlier age at rst acquisition of P. aeruginosa [16]. Other SLC6A14 SNPs, also associated with lung function of pwCF, have been identi ed either by GWAS [9] or genotyping [13]. Beside lung phenotype, Sun et al. showed that the SNP rs3788766 is associated with digestive manifestations of CF, such as increased MI susceptibility in a cohort of 3,763 pwCF [15]. Those results were further con rmed in a genome-wide association investigation performed by the International CF Gene Modi er Consortium with 6,770 pwCF [28]. Altogether, these results emphasize the major involvement of this particular SNP of SLC6A14 in CF clinical variability.
To understand how this SNP can contribute to CF pathophysiology and because of its location within SLC6A14 regulatory region, we evaluated its impact on SLC6A14 promoter activity. We found that the minor allele G, previously identi ed as the deleterious allele regarding CF patient's lung function, is associated with a decrease of SLC6A14 promoter activity in bronchial epithelial cells. This is the rst report showing that a SLC6A14 SNP might in uence SLC6A14 transcription in the context of CF. Indeed, SLC6A14 expression and function have been mostly investigated in cancers so far [29]. Recently, some studies have explored its role in CF pathophysiology and began to explain the reasons for its identi cation as a modi er gene of CF lung and intestinal diseases. Firstly, Di Paola et al. demonstrated that the inhibition of SLC6A14 amino acid transport increased P. aeruginosa attachment to human bronchial epithelial cells by enhancing L-arginine levels in the airway surface liquid [23]. Arginine transport through SLC6A14 also seems to increase F508del-CFTR protein by enhancing nitric oxide (NO) production and activating cGMP or PKG pathways [30]. NO production increase has also been suggested to contribute to anti-infectious response because it is well-known to have bactericidal effects on P. aeruginosa [31]. Thus, SLC6A14 seems to modulate CFTR activity and could participate in the infectious process of CF airways by P. aeruginosa. Concerning the role of SLC6A14 in the intestine, it has been suggested that SLC6A14 involvement in MI susceptibility could be related to intestinal uid secretion defect in CF, which was worsened in Slc6a14-KO CF mice carrying the major mutation F508del [32].
Here, we report, for the rst time, that SLC6A14 is involved in bronchial epithelial repair. In healthy epithelia, repair processes involving cell proliferation, migration and differentiation, facilitate epithelial integrity restoration and function. In CF, repair mechanisms are altered and chronic infections with various pathogens and exacerbated in ammation induce progressive epithelial damage [33]. Our results demonstrated that, in Calu-3 cells and primary HBEC, pharmacological inhibition of SLC6A14 activity resulted in a delayed epithelial repair. SLC6A14 involvement in cell migration and proliferation has been previously described. Indeed, Sikder et al. rst showed that SLC6A14 function favors cell proliferation and invasion in colon cancer LS174T cell line [27]. In addition, Mao et al. showed that SLC6A14 overexpression or knockdown respectively promotes or inhibits migration and proliferation of colorectal cancer cells (HCT-116 and Caco-2 cells) in vitro [34]. They also found that the pharmacological inhibitor α-MT inhibited cell proliferation as well as the fact that SLC6A14 promoted colorectal cancer cell proliferation and migration via the JAK2/STAT3 pathway. SLC6A14 involvement in cell proliferation has also been shown in other cancer cells such as pancreatic cancer cells [14].
In addition to its role in epithelial repair, we highlighted that it may be involved in the mTOR pathway. The mTOR pathway balances anabolism and catabolism in order to control key cellular processes such as cell growth or proliferation. It is very sensitive to amino acids starvation especially leucine and arginine [35]. SLC6A14 implication in mTOR pathway was previously described in pancreatic cancer cell lines [14] for which α-MT-mediated SLC6A14 blockade induces the decreased phosphorylation of proteins involved in mTOR pathway including 4E-BP1, eIF-2α and S6kinase. This interplay between SLC6A14 and mTOR was also recently con rmed in colon cancer LS174T cell line treated with α-MT [27]. In intestinal epithelium, mTOR is involved in wound healing and the re-establishment of barrier function following injury [36]. Consistent with the literature, we con rmed the relation between SLC6A14 activity and mTOR activation. However, how airway epithelial repair, mTOR and SLC6A14 are related remains unknown.
Recently, SLC6A14 was shown to be a target for Wnt-signaling [27], which is known to be one of the key pathways involved in lung repair and regeneration in response to injury [37]. Further work is thus necessary to fully understand the consequences of SLC6A14 blockade on the molecules of the mTOR pathway speci cally related to this wound repair process.
To conclude, we con rmed that SLC6A14 rs3788766 genotype in uences the lung disease severity of pwCF. This study also suggests that SLC6A14 might in uence CF lung phenotype via mTOR signaling pathway and epithelial repair processes modulation.

Declarations
Funding JM received a doctoral fellowship from the French Ministry of Higher Education, Research and Innovation. MR received a post-doctoral fellowship from the French cystic-brosis non-pro t organization Vaincre la mucoviscidose (RF20180502243, RF20190502451).

Con icts of interest statement
There are no con icts of interest to declare.

Consent to participate
Written informed consent was obtained from adults, and for patients<18 years old there was consent from parents or guardians for participation in the study. The study was approved by the French ethical committee (CPP n°2004/15) and the information collection was approved by CNIL (n°04.404).

Consent to participate
Not applicable