High dose everolimus promotes a pro-inammatory phenotype and facilitates the epithelial to mesenchymal transition of primary bronchial epithelial cells isolated from cystic brosis patients.

Background: Lung transplantation is still the best therapeutic option for cystic brosis (CF) patients, but, unfortunately, immunosuppressive therapies, often employed at high dosages to avoid acute rejection, may induce severe complications. In particular, patients treated with high dose of mammalian target of rapamycin inhibitors (mTOR-Is) may experience lung brosis (including bronchiolitis obliterans-organizing pneumonia). Although epithelial to mesenchymal transition (EMT) of airway cells has a central role in this process, the complete biological machinery is not completely claried. Methods: In order to improve our knowledge on this process, primary bronchial epithelial cells carrying F508del mutation were treated with 5 and 100 nM everolimus (EVE) for 24 hours. Subsequently, RNA was hybridized to the Human HT-12 v3 Expression BeadChip (Illumina). Real-Time PCR was, then, used to validate microarray results and to measure major EMT biomarkers. Trans-epithelial resistance was measured by Millicell-ERS ohmmeter. Results: High dosage EVE induced a signicant up-regulation of 42 genes and a down-regulation of 12 genes. After pathway analysis by Gene Set Enrichment Analysis and Ingenuity Pathway Analysis, we found that most of them were implicated in the pro-inammatory pathway. Real-Time PCR validated these results and revealed that, in addition to pro-inammatory genes (IL-1α, IL-8, Pim-1 Oncogene), EVE at high dosage was able to up-regulate major EMT biomarkers (such as: alpha-smooth muscle actin, connective tissue growth factor and metalloproteinase 12). In lung, EMT is the convergence point between inammation and the progression of brotic damage. Additionally, EVE at this dosage reduced the trans-epithelial resistance (altering tight junction strength). In contrast, lower EVE did not trigger similar effects. Conclusions: We demonstrated that high dose EVE may trigger a pro-inammatory/brotic biological machinery in bronchial epithelial cells from CF patients. Our results, although obtained in vitro, suggest that the Five hundred ng total RNA from each sample was reverse transcribed into cDNA using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems), following the manufacturer’s instructions. Real-time PCR amplication reactions were performed in duplicate via SYBR Green chemistry on CFX-connect (Bio-Rad) and SsoAdvanced™ Universal SYBR® Green Supermix (Bio-Rad). Primers for alpha-smooth muscle actin (α-SMA), Pim1 oncogene (PIM1), CTGF, matrix metalloproteinase (MMP)-12, IL-8, IL-1α and β-actin were obtained from Qiagen (QuantiTect Primer Assay). The comparative Ct method (ΔΔCt) was used to quantify gene expression and the relative quantication was calculated as 2 − ΔΔCt . The presence of non-specic amplication products was excluded by melting curve analysis. the biological/molecular effects of the mTOR-inhibitors on bronchial epithelial cells from the whole transcriptomic prole of these cells after EVE 24 h of CF epithelial signicant deregulation of FDR 5%) Gene set enrichment showed that most of the up-regulated genes were involved in the inammatory response pathway a well-dened biological network including several immune-inammatory mediators and cytokines.


Background
Cystic brosis (CF) represents one of the most common genetic disease in Caucasian population, with an autosomal recessive mode of transmission. It affects more than 70,000 individuals worldwide [1] with median survival age expected to be 43.6 years [2].
This pathological condition is caused by mutations in the gene encoding cystic brosis transmembrane conductance regulator (CFTR), an anion channel localized at the apical membrane of several polarized epithelia [3] that mediates chloride and bicarbonate secretion [4] as well as epithelial sodium channelmediated sodium absorption [5]. The most common CFTR mutation is the deletion of the phenylalanine in position 508 of the polypeptide chain, known as Phe508del or F508del [6].
CF is a systemic disease, involving in ammation of low airways and pancreatic ducts due to the secretion of an abnormally dense mucous. In the lungs, defects in anion transport lead to reduced hydration, pH and thickened of the airway surface liquid, leading to thick sticky mucus build up and impaired mucociliary clearance and host defenses [7]. This results in a chronic lung infection that causes respiratory tissue injury leading to respiratory insu ciency [8,9], the main cause of morbidity and mortality in CF patients. Therefore, lung transplantation is an established therapeutic option for end-stage lung disease in patients with CF improving survival and quality of life [10][11][12]. Even if the advances in immunosuppressive therapy and its management have dramatically reduced the acute rejection rate, the long-term life expectancy is still around 45% at 10 years [13]. The main limitation to better long-term survival after lung transplantation remains bronchiolitis obliterans syndrome (BOS), which is the most common form of chronic lung allograft dysfunction [14,15].
Its exact pathogenesis is still unclear, even if an immune-mediated process, probably driven by speci c epithelial and endothelial antigens of the donor, is suspected. BOS is thought to be the end-result of persistent damage to the bronchial epithelium leading to an excessive in ammatory response, local myo broblast recruitment, brosis and ultimately complete obliteration of the airway lumen by brotic matrix [16][17][18].
It is undoubtedly that in this context immunosuppressive therapy could have a role. Everolimus (EVE) is a mammalian target of rapamycin inhibitors (mTOR-I) widely used as immunosuppressant in solid organ transplant recipients that acts by halting T-cell progression from the G1 to the S phase of cell cycle, leading to inhibition of IL-2-induced protein synthesis and cellular proliferation [19,20]. Its administration can also lead to several side effects, which may impose the reduction or withdrawal of the drug. Lung toxicity is a relative common consequence of mTOR-I, showing different clinical forms such as interstitial lymphocytic pneumonia, lymphocytic alveolitis, bronchiolitis obliterans, pulmonary brosis or a combination thereof [21][22][23][24]. Such damage seems to be dose-dependent since the dose reduction or drug discontinuation led in most of the cases to the complete resolution of symptoms.
signi cant for the uploaded data set. Fischer's exact test with false discovery rate (FDR) option was used to calculate the signi cance of the canonical pathway.
Validation of microarray results and measurement of epithelial-mesenchymal transition (EMT) biomarkers by quantitative Real-Time PCR Five hundred ng total RNA from each sample was reverse transcribed into cDNA using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems), following the manufacturer's instructions. Real-time PCR ampli cation reactions were performed in duplicate via SYBR Green chemistry on CFX-connect (Bio-Rad) and SsoAdvanced™ Universal SYBR® Green Supermix (Bio-Rad). Primers for alpha-smooth muscle actin (α-SMA), Pim1 oncogene (PIM1), CTGF, matrix metalloproteinase (MMP)-12, IL-8, IL-1α and β-actin were obtained from Qiagen (QuantiTect Primer Assay).
The comparative Ct method (ΔΔCt) was used to quantify gene expression and the relative quanti cation was calculated as 2 − ΔΔCt . The presence of non-speci c ampli cation products was excluded by melting curve analysis.
Evaluation of transepithelial resistance (TER) in primary bronchial epithelial cell of CF patients treated with different doses of everolimus TER was measured daily by a Millicell-ERS ohmmeter with electrodes (Millipore) that were inserted into both sides of the lter. Alternating current applied between the electrodes was within ± 20 µA at a frequency of 12.5 Hz. The resistance of the monolayer was multiplied by the effective surface area to obtain the electrical resistance of the monolayer (Ω cm2). The background TER of the blank Transwell lter was subtracted from the TER of the cell monolayer. Once stable resistances were obtained, different culture media (control, EVE 5 nM, EVE 100 nM, TGF-β 20 ng/ml) were tested. After the addition of test solutions, measurements were taken at 24 h.

Statistical analysis
Statistical analysis for transcriptomic data was performed by Genespring GX 11.0 software (Agilent Tech Inc., Santa Clara, CA, USA). Identi cation of genes differentially expressed between treated cells and relative controls was carried out with false discovery rate (FDR) method of Benjamini-Hochberg and gene probe sets were ltered based on the FDR (adjusted-P value with multiple testing on 1000 permutations) and foldchange. Fold change lter was set to 1.5-fold in each comparison. Only genes that were signi cantly (adjusted-p value < 0.05 and fold-change > 1.5) modulated were considered for further analysis.
To understand the biological role of the hyper-expressed genes after treatment with 100 nM EVE, we used the Gene Set Enrichment Analysis (GSEA) software. GSEA is a computational method that determines Page 6/31 whether a set of genes, de ned a priori, shows statistical signi cance by comparing the differences in expression between two biological states (phenotypes).
For Real-Time PCR data, the differences between control and treated cell were compared using Two-tailed Student's t-test. A p-value < 0.05 was set as the level of signi cance for all tests.

Results
High dose everolimus (EVE) deregulated pro-in ammatory genes in primary bronchial epithelial cells isolated from cystic brosis (CF) patients.
To investigate the biological/molecular effects of the mTOR-inhibitors on bronchial epithelial cells from CF patients, we evaluated the whole transcriptomic pro le of these cells before and after EVE treatment.
Bioinformatic analysis revealed that 24 h treatment of primary CF bronchial epithelial cells with high dose EVE (100 nM) was able to induce a signi cant deregulation of 54 genes (42 resulted hyper-and 12 downexpressed, p < 0.0001, FC > 1.5, FDR < 5%) ( Table 1). Gene set enrichment analysis showed that most of the up-regulated genes were involved in the in ammatory response pathway (Table 2), a well-de ned biological network including several immune-in ammatory mediators and cytokines.  Differently, as shown in Table 3, the treatment with low dose EVE (5 nM) did not induce any change of the expression level of genes encoding for in ammatory mediators. Instead, this low dose caused a hyperexpression of genes involved in key regulators of the response to stress and cell cycle (Table 4).  Principal component analysis (PCA) and volcano plot graphically showed the degree of separation of untreated versus treated cells with both EVE doses (Fig. 1).
Additionally, Ingenuity Pathway Analysis (IPA), a web-based bioinformatics application that allows researchers to functional analyze microarray data, con rmed enrichment analysis. Also, with this analytic tool high dose of EVE was able to up-regulate biological factors involved in the in ammatory pathway ( Fig. 2A).
Low dose of EVE, instead, deregulated biological factors indirectly associated with the PI3K-AKT pathway, typically regulated by mTOR-I ( Fig. 2A).
Real Time-PCR validated microarray-selected in ammatory genes.
As showed in Fig. 3, the mRNA levels of the Interleukin (IL)-1 alpha, IL-8 and Proto-oncogene serine/threonine-protein kinase 1 (PIM1) resulted signi cantly up-regulated after treatment of primary bronchial epithelial cells from CF patients with high concentration EVE (100 nM) compared to untreated cells (CTR).
Everolimus was able to hyper-express pro-brotic biological elements.
Since in ammation plays a key role in pulmonary brosis in CF patients, we decided to assess whether EVE was able to induce, in our in vitro model, a concomitant hyper-activation of epithelial to mesenchymal transition (EMT), the convergence point between in ammation and the progression of brotic diseases.
As expected, high dose EVE (100 nM) was able to increase the expression of EMT biomarkers, such as alpha-SMA, CTGF, and MMP-12.
On the contrary, lower levels of EVE (5 nM) did not produce any pro-EMT effect (Fig. 4A-C).

High dose EVE reduced transepithelial resistance (TER)
To assess whether high concentration of EVE (100 nM) was able also to alter tight junction (a direct effect of EMT), we measured in primary bronchial epithelial cells from CF patients, the TER with Millicell_ERS ohmmeter before and after treatment.
Concordantly with gene expression results, EVE-treated cells presented a signi cant dysfunction of tight intercellular junctions similar to that induced by TGF-β (positive control) (Fig. 4D).

Discussion
Mammalian target of rapamycin (mTOR) inhibitors are effective therapeutic option in the immunosuppression after organ transplantation. However, currently they are prescribed to 5-10% of transplant recipients [31]. This is mainly due to the high incidence of side effects requiring dose reduction or, in many cases, drug withdrawal [32]. Pulmonary toxicity is relatively frequent with several clinical manifestations: lymphocytic interstitial pneumonitis, lymphocytic alveolitis, bronchiolitis obliterans with organizing pneumonia, focal pulmonary brosis, diffuse alveolar hemorrhage or a combination thereof [21][22][23][24].
The pathogenesis of this condition is complex and is thought to involve a number of processes that lead to an altered alveolar environment and to an abnormal repair process causing epithelial-to-mesenchymal transition (EMT) and pulmonary brosis [33][34][35][36][37]. In particular, it has been suggested that a cell-mediated autoimmune response could play a role when cryptic alveolar antigens are exposed after mTOR-I induced pulmonary toxicity [38]. In addition, a T-cell mediated, delayed-type hypersensitivity may be an alternative pathogenic mechanism [39].
Recently our research group showed that high doses of EVE were able to induce EMT in renal proximal tubular epithelial cells, hepatic stellate cells, type II pneumocytes and immortalized bronchial epithelial cells [24][25][26]. In addition, in wild-type primary bronchial epithelial cells we found that high dosage EVE (contrarily to low dosage) induced the upregulation of genes involved in collagen synthesis/metabolism with CTGF and MMP-12 having a key role [29].
Therefore, we decided to further investigate the effect of EVE in primary bronchial epithelial cells derived from CF patients. This could mimic the in vivo effects of this agent on bronchial cells of patients receiving a lung transplantation, an established therapeutic option for end-stage lung disease in patients with CF that improves survival and quality of life [10][11][12].
Through the transcriptome analysis, we showed that high dose EVE (100 nM) altered the expression of 54 genes (42 up-and 12 down-regulated) most of which were related to in ammatory response such as IL-8, IL-1α, and PIM1.
IL-8 is produced by in ammatory cells such as monocytes/macrophages [40] and by a variety of resident cells such as airway epithelial cells [41] and endothelial cells [42]. In addition to serving as a chemoattractant for immune cells, IL-8 promotes cell proliferation, motility, EMT and has proangiogenic functions [43][44][45]. Several studies have described a role for this cytokine in pulmonary brosis [46][47][48][49]. Its level is high in lung, serum and bronchoalveolar lavage uid of idiopathic pulmonary brosis patients and correlated signi cantly with impairment of lung function parameters [47,48]. This cytokine is also released by mesenchymal progenitor cells (cell of origin for broblasts) and by an autocrine manner, promotes the proliferation and motility of these cells [50].
IL-1α is constitutively expressed in the cytoplasm and nuclei of non-haematopoietic cells such as epithelial cells in tissues including the lung, liver and kidney [51]. Several observations have proposed a role for this cytokine in brotic lung diseases. Its level is increased in bronchoalveolar lavage (BAL) samples from lung transplant recipients with bronchiolitis obliterans syndrome compared to patient-matched BAL samples acquired in other times following transplant or in BAL samples from stable transplant patients [52]. IL-1α is upregulated in the lungs of patients with chronic obstructive pulmonary disease [53] and autoantibodies against IL-1α were detected in the sera of rapidly progressive idiopathic pulmonary brosis patients compared to patients on the rst hospital day [54]. Moreover, it has been demonstrated that damaged epithelial cells release IL-1α that triggers in ammatory response in broblasts [55].
PIM1 encodes a serine/threonine kinase that mediates crosstalk between signalling pathways, including independent Smad proteins and c-Myc, which target downstream transcription factors (ZEB1, ZEB2, Snail1, Snail2 and Twist) to trigger EMT [56]. Although there are no previous studies focusing on the role of PIM1 in brotic lung disease, we can suppose that its hyper-expression could drive in part the EVE-induced pulmonary pro-brotic effects.
In our model several markers of brosis were over-expressed in cells under high dose EVE providing further evidence of an intimate relationship between in ammation and brosis.
Low dose of EVE (5 nM) determined a signi cant deregulation of 89 genes (87 hyper-expressed). In this case, pathways analysis showed the involvement of processes related to stress response.
Furthermore, the reduction of trans-epithelial resistance following treatment with high dose of EVE is an indicator of the progressive loss of integrity of the tight junctions between the epithelial cells.
Finally, our results demonstrates for the rst time that EVE at high dose can induce a pro-in ammatory state in bronchopulmonary cells of patients affected by CF, which in turn can determine brogenesis through different biological pathways, contributing to subsequent loss of the graft.
In the next future, the development of drugs modulating these pathways may represent an excellent therapeutic choice to minimize such complications. Furthermore, the present study underlines the importance of using mTOR-I at the lowest possible therapeutic dosage, alone or in combination with calcineurin or cytostatic inhibitors. In vivo studies might be useful to con rm our hypothesis. Availability of data and materials

Abbreviations
The datasets generated and analysed during the current study are available at: http://www.ebi.ac.uk/arrayexpress/experiments/E-MTAB-8123 Competing interests