LAM Cells as Potential Drivers of Senescence in Lymphangioleiomyomatosis Microenvironment

Senescence is a stress-response process characterized by the irreversible inhibition of cell proliferation, associated to the acquisition of a senescence-associated secretory phenotype (SASP), that may drive pathological conditions. Lymphangioleiomyomatosis (LAM) is a rare disease in which LAM cells, featuring the hyperactivation of the mammalian Target of Rapamycin Complex 1 (mTORC1) for the absence of tuberin expression, cause the disruption of the lung parenchyma. Considering that LAM cells secrete SASP factors and that mTOR is also a driver of senescence, we deepened the contribution of senescence in LAM cell phenotype. We firstly demonstrated that human primary tuberin-deficient LAM cells (LAM/TSC cells) have senescent features depending on mTOR hyperactivation, since their high positivity to SA-β galactosidase and to phospho-histone H2A.X are reduced by inducing tuberin expression and by inhibiting mTOR with rapamycin. Then, we demonstrated the capability of LAM/TSC cells to induce senescence. Indeed, primary lung fibroblasts (PLFs) grown in LAM/TSC conditioned medium increased the positivity to SA-β galactosidase and to phospho-histone H2A.X, as well as p21WAF1/CIP1 expression, and enhanced the mRNA expression and the secretion of the SASP component IL-8. Taken together, these data make senescence a novel field of study to understand LAM development and progression.


Introduction
Cellular senescence is a physiological stress response in which cells are in a permanent status of cycle arrest. This process is highly dynamic, and it represents an important mechanism through which tissue homeostasis is maintained, as it regulates tissue remodelling and repair during development and in adulthood [1]. Nevertheless, an irreversible proliferation arrest might also contribute to the loss of tissue homeostasis by limiting its regenerative capacity, inducing a pro-inflammatory status that might sustain the onset of pathological condition and age-related diseases, i.e., atherosclerosis and pulmonary idiopathic fibrosis (IPF) [2]. Besides as the inhibition of cell cycle, senescence should be considered also as the activation of protein secretion and cell flattening. Senescent cells acquire the capability to alter the functions of neighbouring cells through the release of senescence-associated secretory phenotype (SASP) factors, composed of pro-inflammatory cytokines, chemokines, growth factors, and matrix metalloproteinases. These molecules promote inflammation, invasion, angiogenesis, and cell proliferation, which might drive the development of several pathologies, including cancer [3]. Indeed, senescent fibroblasts The demonstration of senescent features in LAM cells and of their capability to propagate senescence phenotype in fibroblasts introduces an interesting approach to investigate novel mechanisms to modulate LAM progression.

Evaluation of Senescence in LAM/TSC Cells and PLFs
Senescence is a heterogenous process that needs to be analysed by using multiple approaches since no single universal marker has been identified. The expression of SA-βgalactosidase is a good biomarker to detect senescent cells, both in culture and in tissues, with a histochemical assay [19]. Moreover, the phosphorylation of the histone H2A.X on Ser139 occurs in senescent cells to promote the assembly of DNA repair and checkpoint factors, thus initiating the senescence program in response to DNA damage [20]. To investigate if senescence could contribute to the LAM development and progression, we studied the senescent features of a LAM cell line, the human primary LAM/TSC cells isolated from chylous thorax of a patient affected by LAM associated to TSC [16], and of PLFs, obtained from a lung biopsy, which express tuberin and do not display mTOR hyperactivation ( Figure 1a). Comparing these LAM/TSC cells to PLFs, we observed a significantly higher level of SA-β-galactosidase staining ( Figure 1b) and a higher percentage of phospho-histone H2A.X-positive cells compared with PLFs ( Figure 1c). These data sustain a LAM/TSC cell senescent behaviour, and make these cells a good model to study senescence in LAM.

Effect of the Induction of Tuberin Expression and mTOR Inhibition on LAM/TSC Senescent Features
The lack of tuberin leads to the hyperactivation of mTOR, which exerts many functions in controlling cell growth and metabolism, and that has a role in sustaining senescence, being considered a positive regulator of SASP [21]. LAM/TSC cells do not express tuberin because of a germline TSC2 mutation and the hypermethylation of the TSC2 promoter, making these cells an interesting model in which restore tuberin expression by the exposure to chromatin remodelling agents such as 5-azacitydine [16]. To analyse if the LAM/TSC cell senescent features might depend on the mTOR hyperactivation, senescence was firstly evaluated following the induction of tuberin expression. As expected, the phosphorylation of S6, the substrate of S6 kinase (S6K), a functional marker of mTOR activity [22], was reduced by 1 µM 5-azacytidine treatment for 96 h, which caused tuberin expression ( Figure 2a). Interestingly, also the levels of the phosphorylated form of ZRF1, which is a substrate of S6K that drives senescence reprogramming in fibroblasts [23], were impaired, while p21 WAF1/CIP1 expression was not significantly affected (Figure 2b), likely suggesting that tuberin expression only influences the pathways downstream mTOR but it does not have effects on the cell cycle arrest in senescence. As further confirmation of this hypothesis, a significant decrease of the positivity to the SA-β-galactosidase staining was observed in tuberin-expressing LAM/TSC cells following 5-azacytidine treatment (Figure 2c), while the induction of tuberin only slightly reduced the cells positive to phospho-histone H2A.X (5%) compared to tuberin null LAM/TSC cells (7.8%) analysed by flow cytometry (data not shown). Of note, the treatment with 5-azacytidine did not significantly affect LAM/TSC cell viability, as demonstrated by trypan blue exclusion assay, in accordance with our previous study [16]. Moreover, the incubation with 5-azacytidine for 96 h did not cause a relevant apoptotic process as demonstrated by the absence of cleaved fragments of the apoptotic markers Caspase 3 and PARP by Western Blot and by the low positivity to annexin V analysed by flow cytometry (Supplementary Figure S1). To confirm that the expression of tuberin can decrease the senescent features of LAM/TSC cells through the modulation of mTOR activity, LAM/TSC cells were grown for 48 h with the mTOR inhibitor rapamycin at the concentration of 0.5 ng/mL and 1 ng/mL. Similar to the data shown above, rapamycin caused the reduction of S6 and ZRF1 phosphorylation, and inhibited the p21 WAF1/CIP1 expression even if only at the higher drug concentration (Figure 3a,b). Consistently, the percentage of SA-β-galactosidase positive LAM/TSC cells was reduced by rapamycin treatment (Figure 3c) as well as the percentage of cells expressing the phosphorylated histone H2A.X analysed by flow cytometry (3.04% and 1.95% for LAM/TSC cells treated with 0.5 ng/mL and 1 ng/mL of rapamycin, respectively, and 7.8% in control LAM/TSC cells) (data not shown). LAM/TSC cell viability, evaluated by trypan blue exclusion assay, was not significantly affected by the 48-h incubation with 0.5 ng/mL or 1 ng/mL of rapamycin (Supplementary Figure S2a), consistently with our previous demonstration that, at this time point, the proliferation rate of LAM/TSC cells in the presence of rapamycin concentrations up to 5 ng/mL is similar to controls [16]. As shown for 5-azacytidine treatment, the 48-h treatment with 0.5 ng/mL or 1 ng/mL of rapamycin did not induce apoptosis in LAM/TSC cells (Supplementary Figure S2b,c). β-actin was used as loading control. Densitometric analysis of p-ZRF1/β-actin and p21/β-actin are presented. Results are shown as mean ± SEM (n = 3). Student's t-Test, ** p < 0.01 (c) Quantification of β galactosidase-positive LAM/TSC cells, considered as the percentage of stained cells on the total number of cells in each field, was performed on 3 independent experiments in standard condition (C) or after the treatment with 5-Azacytidine for 96 h (5-Aza). 8 images for each experiment were evaluated. Results are shown as mean ± SEM (n = 24). Student's t-Test, **** p < 0.0001. ANOVA with Tukey Test, ** p < 0.01 (b) Senescent markers p21 WAF1/CIP1 (p21) and phospho-ZRF1 (p-ZRF1) were analysed by Western Blot analysis in LAM/TSC cells (C) and in LAM/TSC cells treated for 48 h with 0.5 ng/mL (R0.5) or 1 ng/mL (R1) of rapamycin. β-actin was used as loading control. Densitometric analysis of p-ZRF1/β-actin and p21/β-actin are shown. Results are shown as mean ± SEM (n = 3). ANOVA with Tukey, ns (c) Quantification of SA-β galactosidase positive LAM/TSC cells was performed in standard condition (C) or after the treatment with 0.5 ng/mL (R0.5) or 1 ng/mL (R1) of rapamycin for 48 h on 3 independent experiments evaluating 10 images each experiment. Results are shown as mean ± SEM (n = 30). ANOVA with Tukey Test, **** p < 0.0001.

Evaluation of the Effect of Conditioned Medium (CM) of LAM/TSC Cells on PLFs Proliferation and Apoptosis
Considering that LAM cells represent only a small population in the lung LAM microenvironment but cause a diffuse parenchyma disruption [7], we developed an in vitro model to investigate the capability of LAM/TSC cells to influence PLFs, by likely inducing a LAM-supporting microenvironment. PLFs were grown in the CM of LAM/TSC cells to mimic a paracrine communication between these two cellular populations. As control, PLFs were grown on their own CM, in order to avoid artefacts due to the culture in an exhausted medium. Firstly, LAM/TSC cell CM reduced the PLFs proliferation, with a significant decrease in PLFs number starting from 48 h (Figure 4a,b). The delayed PLFs growth in LAM/TSC cell CM was not due to an increased PLFs death, as demonstrated by the unaltered levels of full-length apoptotic markers Caspase 3 and PARP, by any significant difference in their cleaved fragments ( Figure 4c) and the low positivity to annexin V, which is not significantly different between PLFs grown on their own CM and in LAM/TSC cell CM ( Figure 4d).

Effect of LAM/TSC Cell CM on Senescent Features of PLFs
To investigate if the delay of PLFs proliferation might be due to the induction of a senescent phenotype caused by LAM/TSC cell CM, senescent markers were studied. The positivity to SA-β-galactosidase and to phospho-histone H2A.X of PLFs grown for 48 h in LAM/TSC cell CM was significantly higher compared to controls (Figure 5a,b). To confirm the capability of LAM/TSC cell CM to induce senescence, IMR90 cells, a fibroblast cell line usually used as a senescence model [5], were grown for 48 h in the CM from LAM/TSC cells. Similar to what was observed for PLFs, the IMR90 cells positive to SA-β-galactosidase and to phospho-histone H2A.X were significantly increased (Figure 5c,d). It should be considered that the SA-β-galactosidase activity can be influenced by high levels of glucose (e.g., 22 mM) and in general by medium components [24]. Since IMR90 cells grow in DMEM and LAM/TSC cell CM was composed of Type II medium, senescent features of IMR90 grown in their own Type II CM and in their own DMEM CM were compared to evaluate the influence of the medium composition. We did not observe any significant differences between the two groups, indicating that the different DMEM and Type II medium formulations, at our experimental conditions, did not influence the modulation of senescence as evaluated by the used hallmarks (Figure 5c,d: columns 1 and 2). Likely, the difference between the glucose concentrations in the two media is not high enough to influence senescence. DMEM contains 1 g/L (5 mM) of glucose, lower than Type II medium that is composed by 1:1 mixture of DMEM and Ham's F12, whose glucose concentration is 1.802 g/L (10 mM). Moreover, the different concentration of the serum, which has the same source and is 10% in DMEM and 15% in Type II medium, appears to not be relevant in our experimental conditions to modulate senescence.  To better study the effect of LAM/TSC cell CM on senescence in this model, p21 WAF1/CIP1 and p16 INK4A expression, biomarkers involved in signalling pathway specific for cellcycle arrest, which inhibit cyclin-dependent kinase, were evaluated by Western Blot. The induction of senescence might be driven by p21 WAF1/CIP1 , whose expression is increased upon PLFs cultured in LAM/TSC cell CM, while, surprisingly, p16 INK4A was unaltered (Figure 6a,b). mTOR activity in PLFs seems to be unaffected by 48 h of growing in LAM/TSC CM, since S6 phosphorylation was not altered (Figure 6c).

Evaluation of IL-8 Expression and Secretion in PLFs, LAM/TSC Cells, and in PLFs Grown in CM
IL-8 is a chemokine overexpressed in the most senescent cells as a SASP component and its high levels in lung parenchyma have been related to the pathogenesis and progression of lung diseases, such as IPF [25]. IL-8 is expressed in various cell types, including fibroblasts, and is highly secreted by LAM/TSC cells [16,26]. The mRNA relative expression and the secretion of IL-8 were significantly higher in LAM/TSC cells compared to PLFs suggesting a SASP IL-8 role in sustaining LAM/TSC cell senescent phenotype (Figure 7a,c). When PLFs were grown in LAM/TSC cell CM, the IL-8 mRNA relative expression significantly increased (Figure 7b). In a similar way, PLFs grown in LAM/TSC cell CM secreted higher IL-8 levels than PLFs grown in their CM (Figure 7c

Discussion
In this study, we demonstrated that human primary tuberin-deficient LAM/TSC cells have a senescent phenotype, as shown by the activation of the SA-β-galactosidase enzyme and by the phosphorylation of the histone H2A.X, which are both significantly higher in LAM/TSC cells compared to human PLFs, which do not have hyperactivation of mTOR. The inhibitory modulation of the senescent features by rapamycin and by the induction of tuberin expression through 5-azacytdine indicated that, in LAM/TSC cells, senescence is dependent on mTOR hyperactivation. Finally, we demonstrated that LAM/TSC cell CM can induce senescence in PLFs, with a key role of the SASP factor IL-8, which is highly expressed and secreted in LAM/TSC cells and PLFs grown in LAM/TSC cell CM.
Among the physiological and pathological processes involved in the LAM onset, we focused on senescence, hypothesizing that it might promote the loss of the parenchymal structure in the lungs, causing a severe impairment of the lung function, as already demonstrated for other respiratory diseases, such as IPF [2]. Senescence could be a novel mechanism to explain the highly variable manifestation and progression of LAM. At the state of the art, this hypothesis was only proposed but not demonstrated so far [27]. Indeed, it is already known that LAM cells exploit several mechanisms to induce the remodelling of the lung parenchyma, including the secretion of molecules such as matrix metalloproteinases [11], kathepsin K [10], VEGF-D [7], IL-6 [28], and IL-8 [16], which are also SASP components released by senescent cells to modulate the behaviour of neighbouring cells.
Moreover, as further confirmation of this hypothesis, the evidence of an early senescent phenotype in mouse Tsc2-null embryo fibroblasts (MEF lines), which bear a constitutive activation of mTORC1 due to loss-of-function mutations in TSC1, TSC2, or PTEN, was reported [29]. The absence of tuberin in LAM/TSC cells leads to mTOR hyperactivation likely causing senescence. Indeed, senescent cells might overcome the cell cycle arrest through the hyperactivation of the growth-promoting mTOR pathway [30], leading to lysosomal hyperfunction that can be detected following the activity of the lysosomal protein SA-β-galactosidase, used as senescence biomarker in senescent fibroblasts and keratinocytes, as well as in aged skin in vivo [31]. In LAM/TSC cells, the hyperactivation of mTOR appears to play a central role in senescence, as demonstrated by the decreased SA-β galactosidase positivity after the downregulation of mTOR by inducing tuberin expression or by using its inhibitor rapamycin. This last result confirms the senolytic property of the mTOR inhibitors [32]. At the state of the art, rapamycin is the only drug approved for the treatment of LAM that has a cytostatic and not cytotoxic action, and discontinuation of the treatment leads to tumour regrowth and decline in pulmonary function [33][34][35]. Considering a possible senescent behaviour of LAM cells, the relapse after rapamycin treatment suspension might be due to the reactivation of the so-called geroconversion, an irreversible state of senescence characterized by cellular growth sustained by mTOR. Indeed, we observed that the inhibition of mTOR by rapamycin reduces the phosphorylation of ZRF1, which drives the mTOR-dependent senescent programme [23], without changes in p21 WAF1/CIP1 expression, indicating that LAM/TSC cells are senescent in spite of a state of cell cycle arrest. As further confirmation of the role of mTOR hyperactivation in sustaining the senescent phenotype of LAM/TSC cells, we obtained similar results on phospho-ZRF1 and p21 WAF1/CIP1 expression when LAM/TSC cells were treated with 5-azacytidine, which causes tuberin expression. Since 5-azacytidine is a methyltransferase inhibitor, this drug is expected to induce senescence in a p53 dependent manner as reported for oral squamous cell carcinoma, hepatocellular carcinoma cells, and other solid tumour cell lines [36]. Conversely, our results indicate that the treatment with 5-azacytidine, inducing tuberin expression, reduced the senescent features of LAM/TSC cells, which decrease the positivity to SA-β galactosidase. These data are consistent with the recent study in which 24 h of incubation with 5-azacytidine inhibited senescence in insulinoma cell lines established from transgenic mice [37]. In this case, the drug treatment caused cytotoxicity, which was not observed in our experimental conditions. We previously demonstrated that 5-azacytidine treatment also reduced migratory capability and MMP2 and MMP7 secretion of LAM/TSC cells, likely because of the induction of tuberin expression [11]. Given the broad inhibitory action of 5-azacytidine on hypermethylation that leads to chromatin remodelling, as often occurs in several types of cancers, we have to consider that this drug might cause other demethylating effects besides the induction of tuberin expression, making this point a limitation of our study that could be overcome by the retroviral transduction of TSC2. However, all the effects of 5-azacytidine that we know to occur in our model are not supposed to influence what we observed as reduction of senescence features and, considering together the rapamycin and 5-azacytidine action, we may reasonably affirm that LAM/TSC cell senescent features are caused by mTOR hyperactivation due to the absence of tuberin. In the controversial discussion of the role of senescence in cancer, considering the neoplastic characteristics of LAM cells, these data indicate that cancer cells are in fact pro-senescent and capable to promote senescence, as shown in cells with loss of PTEN [38].
LAM/TSC cells survive in an anchorage-independent manner, degrade the extracellular matrix through the secretion of matrix metalloproteinases, suggesting a migratory ability, and secrete high levels of IL-6 and IL-8 [16]. All this evidence also demonstrated the capability of these cells to create a supporting microenvironment for their survival in the lung. To model the possible communication between LAM/TSC cells and PLFs (not by means of a direct contact), PLFs were grown in the CM of LAM/TSC cells. PLFs proliferation was significantly reduced probably because of the activation of p21 WAF1/CIP1 protein that induces cell cycle arrest without increasing cellular apoptosis, as shown by the annexin V results and expression of Caspase 3 and PARP. In this condition, also a significantly higher activation of SA-β-galactosidase and higher phosphorylation of histone H2A.X were observed, confirming a senescent behaviour of PLFs induced by LAM/TSC cell CM but not by their own CM. The LAM/TSC cell CM also caused senescent features in IMR90 cell line sustaining the capability of LAM/TSC cells to induce senescence in fibroblasts. Very recently, the capability of LAM cells to modulate the LAM-Associated fibroblasts (LAFs) phenotype was demonstrated in an in vitro model of LAM microtissues, where non-expressing-tuberin LAM cells cause the increased expression of the Fibroblasts Growth Factor 7 protein, which acts as a chemotactic and mitogen factor for epithelial cells [18]. This evidence sustains the role of the LAM microenvironment in controlling LAM development.
IL-8 is a potent neutrophil chemotactic factor that plays important roles under several pathological and physiological conditions including senescence, which mediates the spreading of the SASP response in the microenvironment [6]. We demonstrated that LAM/TSC cells not only express and secrete more IL-8 than PLFs, but they are also able to cause a significant increase of IL-8 mRNA levels and IL-8 secretion in PLFs grown in LAM/TSC cell CM, while PLF CM does not cause any alteration in IL-8 expression and secretion. This evidence suggests that, in LAM, senescent LAM cells might reduce the proliferation of lung resident cells and, at the same time, might promote a pro-inflammatory microenvironment that ultimately has detrimental effects on the lung tissue. Indeed, even if IL-8 has a clear role as a SASP molecule, its secretion by fibroblasts, alveolar macrophages, and alveolar cells and its involvement in neutrophil migration and infiltration in lungs suggest a more complex physiological implication, making the deregulation of IL-8 secretion responsible for pathological processes, i.e., in IPF [25], acute respiratory distress syndrome [26], and chronic obstructive pulmonary disease [39]. Besides the crucial role of IL-8 in lung inflammation, little is known concerning its involvement in LAM or TSC. We observed that high IL-8 secretion occurs in LAM/TSC cells, and it is decreased after the treatment with 5-azacytidine indicating that its production in these cells might be dependent on tuberinsensitive pathways [16]. However, the fact that IL-8 secretion is increased in PLFs when grown in LAM/TSC cell CM and the evidence that in this condition PLFs become senescent suggest that IL-8 secretion in PLFs occurs in a senescence-dependent manner. In LAM or TSC models, the SASP cytokine IL-6 has been better studied than IL-8, demonstrating an increase in TSC2-deficient cells in an mTORC1-dependent manner and an up-regulation in plasma of LAM patients compared to healthy subjects [28]. In the in vitro model, IL-6 is involved in migration and proliferation as they are inhibited by the treatment with anti-IL-8 antibody. Similar to IL-6, the IL-8 high secretion of LAM/TSC cells might be responsible of other processes than senescence in LAM microenvironment.
Concluding, the demonstration that LAM cells are senescent and have the capability to induce senescence in neighbouring cells introduces an interesting approach to investigate novel targets to control LAM progression. Indeed, to understand the involvement of mTOR and of its specific age-related targets in LAM senescent cells might help to better evaluate the effect of rapamycin in treating LAM. Finally, interfering with IL-8 expression and secretion may help to conceive a novel approach to delay the progression of the lung parenchyma disruption.

Cell Culture and Treatments
LAM/TSC cells were previously obtained from the chylous thorax of an LAM/TSC2 patient who had given written informed consent according to the Declaration of Helsinki [16]. The study was approved by the Institutional Review Board of Milan's San Paolo Hospital. As previously reported, LAM/TSC cells can be grown as a stabilized cell line in Type II medium, composed by a 1:1 mixture of stable Glutamine-containing DMEM/Ham's F12 (Aurogene s.r.l, Rome, Italy) supplemented with 200 nmol/L hydrocortisone (Sigma-Aldrich, St. Louis, MO, USA), 10 ng/mL epidermal growth factor (EGF; Sigma-Aldrich), 100 µg/mL penicillin/streptomycin (Sigma-Aldrich), and 15% foetal bovine serum (Sigma-Aldrich). LAM/TSC cells were routinely checked for morphological, biochemical, and genetic features.
Human Primary Lung Fibroblasts (PLFs) were obtained from a lung biopsy. The tissue was washed in PBS, cut into small pieces, and placed in cell culture dishes with complete Type II medium for 3-5 days. When the fibroblasts reached the confluency, they were sub-cultured as a stabilized cell line.
In order to avoid non-specific effects due to aging cells that could be confused with senescence, all the experiments with the primary cells were performed within 20 days from cell thawing, meaning between passage 3 to 5.
IMR90 cells, a human foetal lung fibroblast cell line (ATCC, Rockville, MD, USA), were maintained in stable Glutamine-containing DMEM supplemented with 10% FBS and 100 µg/mL penicillin/streptomycin. To prevent the onset of a senescent phenotype due to cell aging, IMR90 cells were used within a population doubling level (PDL) 20 to 30. In this PDL range, IMR90 cells duplicate at a constant rate, indicating that senescence process is not still begun, while at PDL levels over 40, a growing delay starts (data not shown). PDL was calculated with the following formula: PDL = [PDL 0 + 3.22 × (Log 10 CN Y − Log 10 CN P )] where PDL 0 is the initial PDL, CN Y is the number of cells yielded, and CN P is the number of cells plated.
The CM was obtained from LAM/TSC cells and PLFs grown in standard condition for 72 h and pre-cleared from suspended cells and debris by centrifugation at 400× g for 5 min at room temperature. PLFs or IMR90 cells were grown for 48 h in LAM/TSC cell CM or in their own CM as control.
Proliferation assay was performed by plating 5 × 10 4 PLFs in CM of LAM/TSC cells or in their own CM, as control. The cells were harvested and counted after 24, 48, 72, and 96 h using a Neubauer chamber. Cells viability was evaluated with trypan blue exclusion assay as previously described [16].

SA-βGalactosidase Assay
In total, 7 × 10 4 LAM/TSC cells, PLFs or IMR90 cells were seeded, grown as indicated for each experiment, and stained using the Senescence β-Galactosidase Staining Kit (Cell Signalling Technology), according to manufacturer's instruction. Images were acquired with a bright-field microscope, and senescent cells were evaluated as the percentage of the stained blue senescent cells over the total number of cells by counting at least 10 fields for each sample in three independent experiments.

Flow Cytometric Analysis
The histone H2A.X Phosphorylation assay Kit (Sigma Aldrich, Cat. No. 17-344) was used according to manufacturer's instruction. Briefly, LAM/TSC cells, PLFs, and IMR90 cells, grown as reported for each experiment, were collected, washed in PBS, and fixed in 2% paraformaldehyde for 20 min on ice. Direct labelling was performed by incubating the cells with the antibody for histone H2A.X FITC-conjugated or normal mouse IgG isotype control FITC-conjugated (0.3 µg for each sample) in an appropriate volume of Permeabilization Solution for 20 min on ice. Finally, samples were washed with Wash Buffer.
For apoptosis evaluation, PLFs grown in LAM/TSC cell CM, in their own CM, or treated with 500 µM H 2 O 2 for 2 h were detached with Accutase (Sigma-Aldrich), washed in PBS, and incubated for 15 min at room temperature with APC Annexin V (BD Biosciences, Franklin Lakes, NJ, USA, Cat. No. 550475) in an appropriate volume of Annexin V Binding Buffer (BD) and washed in PBS.

ELISA
The total amount of IL-8 in cell culture media was measured by Enzyme-Linked Immunosorbent Assay (ELISA) kit (Cat. No. IL-8: SEA080Hu-96, Cloud-Clone Corp, Katy, TX, USA) according to the manufacturer's instruction.

qPCR
Total RNA was extracted from LAM/TSC cells or PLFs, grown in standard condition or in the CM, using NucleoZOL (Macherey-Nagel, Düren, Germany) and phenol-clorophorm method. cDNA was synthesized from 1 µg of total RNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Thermo Fisher Scientific) and random primers. qPCR was performed on a Roche LightCycler 480 qPCR machine using precision TB Green Premix Ex TaqII (Tli RNaseH Plus, Takara Bio, Saint-Germain-en-Laye, France) qPCR Master Mix with SYBR Green (Takara Bio, Saint-Germain-en-Laye, France). First-strand DNA synthesis reactions without reverse transcriptase were used as controls. Triplicate samples are run in all reactions. The quantification cycle (∆Cq) value and the ∆∆Cq were calculated relative to control samples using Cq threshold values that are normalized to the housekeeping gene GAPDH, RPLP0, RPL13A.

Statistical Analysis
Statistical analysis was performed with the GraphPad Prism 7.02 software. (GraphPad Software, San Diego, CA, USA). Differences between the mean measurement of each sample were analysed using Student's t-test for single comparison or one-way Analysis of Variance (ANOVA) followed by the Tukey's test for multiple comparisons, when appropriate. Results are presented as mean ± standard error of the mean (SEM). p values were considered significant as reported in each figure.