Cholesteryl hemiazelate causes lysosome dysfunction impacting vascular smooth muscle cell homeostasis

Vascular smooth muscle cells (VSMCs) play a crucial role in the development of cardiovascular diseases (CVDs). VSMCs have been implicated in all stages of human atherosclerotic plaque development. In early stages of the disease, VSMCs are responsible for the intimal thickening due to increased proliferation, migration and production of extracellular matrix (Newby and Zaltsman, 1999). This aberrant VSMC migration and proliferation culminates with the formation of a fibrous cap that is regarded as an adaptive response to atheroma formation and beneficial rather than detrimental to the pathology (Allahverdian et al., 2018; Greig et al., 2012). Nevertheless, the role of VSMCs in atherogenesis is not entirely beneficial. VSMCs within atherosclerotic lesions can also acquire a macrophage-like phenotype that gives rise to foam cells, and undergo apoptosis and senescence (Feil et al., 2014; Shankman et al., 2015). Indeed, studies suggest that ∼50% and 70% of foam cells, which are characterized by an aberrant lysosomal and cytosolic lipid accumulation, in human and murine lesions, respectively, are VSMC derived (Allahverdian et al., 2014; Vengrenyuk et al., 2015; Wang et al., 2019). The transformation of VSMCs into foam cells was first described in the 1970s, by de Duve, in the aorta of cholesterol-fed rabbits (de Duve, 1974). The sequestration of lipids, such as free cholesterol (FC) and cholesteryl esters, within lysosomes hinders their mobilization by cells, leading to lysosome dysfunction and enlargement (Jerome and Lewis, 1990; Jerome et al., 1991; Minor et al., 1991). In VSMCs, lysosome dysfunction may promote atherosclerosis by reducing their ability to clear dying cells and necrotic debris, exacerbating inflammation and promoting further cell death (Marques et al., 2021). Furthermore, lysosomes are the endpoint of the autophagic pathway, a critical process for VSMC function, phenotype, and survival (Xu et al., 2010; Ding et al., 2013).

In atherosclerotic lesions, the vast amount of oxidized low-density lipoproteins (oxLDL) generated in the arterial intima represents one of the sources of undigested lipids that accumulate within the lysosomes of foam cells (Yancey and Jerome, 1998). In VSMCs, oxLDL can be internalized by scavenger receptors and through micropinocytosis (Chellan et al., 2016). However, it remains unclear which components of oxLDL elicit the lysosomal dysfunction and subsequent loss of VSMC homeostasis observed in atherosclerotic plaques. OxLDL represents a complex mixture of lipid oxidation products (Greig et al., 2012), making it difficult to associate a specific biological response with an individual oxLDL component and so, knowledge of the extent to which each oxidation product contributes to the pathology is required to further state any correlation between cause and effect.

Our group has identified and quantified cholesteryl hemiesters (ChEs), stable oxidation end-products of cholesteryl-polyunsaturated fatty acid esters in LDL, in plasma of cardiovascular disease (CVD) patients and the most prevalent ChE is cholesteryl hemiazelate (ChA) (Estronca et al., 2012; Domingues et al., 2021 preprint; Matthiesen et al., 2021). In addition, as proof-of-concept, our group demonstrated that cholesteryl hemisuccinate (ChS), a commercially available but biologically irrelevant ChE, can induce irreversible lysosomal lipid accumulation and inflammation in macrophages, mimicking what has been described to occur in atherosclerotic lesions (Domingues et al., 2017; Estronca et al., 2012).

Considering the critical role of lysosomes in atherogenesis, it is imperative to extend our understanding of the molecular and cellular mechanisms underlying lysosomal dysfunction. For this reason, we decided to assess whether ChA can induce lysosomal dysfunction and loss of cell homeostasis in a murine model of VSMCs. Our data indicate that ChA is able to induce lysosome dysfunction in VSMCs leading to an exuberant accumulation of neutral lipids. The dysfunctional enlarged lysosomes are localized mainly in the perinuclear region of the cells and these outcomes can be the consequence of lysosomal luminal pH changes. Interestingly, the microphthalmia-transcription factor E (MiT or TFE) family of proteins (hereafter denoted MiT/TFE family), involved in responding and adapting to lysosomal stress, and autophagy are not activated. VSMCs with these features also proliferate and migrate less and become stiffer than control cells. Altogether, we present evidence that ChA is pathogenic towards VSMCs with potential impact in atherosclerosis progression and in their protective role in plaque stability.

ChA is an amphiphile cholesterol derivative that acquires a negative charge at neutral pH (Fig. 1A). Given its poor solubility in cell culture medium, ChA was delivered to the VSMCs via ChA:POPC (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine) liposomes (65:35, molar ratio). POPC liposomes were always used as control. Based on our previous publications (Domingues et al., 2017, 2021), we know that, in macrophages, ChEs accumulate in late endosomal compartments (late endosomes/lysosomes, hereafter referred to as lysosomes). Similar to FC, ChEs can be stained by filipin, a polyene macrolide antibiotic (Domingues et al., 2017, 2021 preprint). Thus, we firstly analysed the intracellular distribution of ChA in VSMCs exposed to ChA:POPC liposomes. After 72 h incubation, we observed a stronger and more distinct filipin staining in cells treated with ChA:POPC compared to POPC-exposed cells (Fig. 1B). In ChA-treated cells, filipin stained large round vesicular structures that were surrounded with lysosomal-associated membrane protein 1 (LAMP1), a marker of lysosomes (Fig. 1B, insets). However, we cannot exclude that part of the filipin staining could be attributed to labelling of accumulated intralysosomal FC. Similar results were obtained when lysosomes were stained with anti-LAMP2 antibodies. These data suggest that ChA causes the enlargement of lysosomes in VSMCs and that ChA might be accumulated inside these organelles. To better assess the impact of ChA in the lysosomal structure and the nature of the storage material, we performed a transmission electron microscopy analysis of VSMCs exposed to this lipid. Unlike what was seen in control cells, VSMCs incubated with ChA liposomes revealed the presence of enlarged vesicular structures with electron-lucent material compatible with lipid accumulation visualized by light microscopy (Fig. 1C, red arrows). In accordance with previous observations (Domingues et al., 2017; Estronca et al., 2012), quantification of the area of lysosomes in VSMCs incubated with ChA for 72 h showed a clear shift towards larger lysosomes compared to those in cells incubated with POPC liposomes (Fig. 1D). To confirm whether the observed enlargement was exclusive to lysosomes or impacted other organelles of the endocytic pathway, we immunostained VSMCs exposed to ChA for early-endosome antigen 1 (EEA-1), a marker of early endosomes, and LAMP2. As shown in Fig. 1E, we did not see changes in the area of EEA-1-stained structures or colocalization of EEA-1 with LAMP2 in POPC or ChA-exposed VSMCs. This indicates that the lipid-induced enlargement occurs exclusively in lysosomes and not in other organelles of the endocytic compartment.

Since it has been reported that cholesterol crystals can damage lysosome membranes with loss of its integrity, we decided to assess whether ChA had a similar effect. For this purpose, we immunostained the ChA-treated VSMCs for galectin-3 (Gal3, also known as LGALS3), a protein that binds to sugars on the inner leaflet of lysosome membranes, which are exposed when lysosomes become permeable (Maejima et al., 2013). Thus, the presence of Gal3 on lysosomes indicates loss of lysosome membrane integrity. As shown in Fig. 1F, ChA-treated VSMCs lysosomes were not decorated with Gal3 puncta, indicating that the membrane of these organelles was kept intact. However, when ChA-treated VSMCs were challenged with an agent that permeabilizes lysosomal membranes acutely, the lysosomotropic di-peptide L-leucyl-L-leucine methyl ester (LLOMe), their lysosomal membranes were more sensitive to damage than lysosomal membranes of control VSMCs, as judged by the presence of Gal3 puncta in lysosomes (Fig. 1F).

Next, we sought to clarify the consequences of the detected enlargement and ChA accumulation for lysosomal homeostasis and function. First, we investigated the storage of other lipids besides ChA and FC in the ChA-exposed lysosomes. Using the fluorescent neutral lipid dye BODIPY 493/503, we found that the LAMP2-positive enlarged lysosomes in VSMCs incubated with ChA were enriched in neutral lipids (Fig. 2A, see insets). In contrast, in POPC-exposed cells, there were no detectable BODIPY-positive lysosomes and just a limited number of small lipid droplets (Fig. 2A, see insets), indicating that neutral lipids are not being stored under these conditions. ChA, as stated above, is a polar lipid and cannot be visualized with BODIPY. However, we can speculate that once ChA starts to accumulate in lysosomes the degradative capacity of these organelles is compromised, leading to the accumulation of internalized neutral lipids that exist in the cell culture medium, as well as other cargo, contributing to their enlargement. To assess whether limited lysosomal lipid exit capacity was also contributing to the enlargement in ChA-treated cells, we measured the egress towards the Golgi of the glycosphingolipid lactosylceramide (LacCer) conjugated with BODIPY. LacCer can be taken up by endocytosis and upon exiting the lysosomes it is routed to the Golgi and not to the plasma membrane (Choudhury et al., 2002; Sharma et al., 2003). We evaluated the exit of LacCer from lysosomes in live cells by analysing its colocalization with the probe Lysotracker, which stains acidic organelles. ChA-treated VSMCs presented higher levels of LacCer colocalization with Lysotracker-positive vesicles (Fig. 2B, thick arrows) when compared with POPC-treated cells that already show some LacCer in the Golgi (Fig. 2B, thin arrows), as quantified in Fig. 2C. These data suggest an impairment of lysosomal cargo exit due to ChA treatment.

Next, we decided to evaluate the impact of ChA in lysosomal luminal pH. Breakdown of incoming endocytic and autophagic substrates within lysosomes, is pH dependent and, changes in pH could explain the outcomes described above (Fig. 2A–C). The lysosomal pH gradient is assured by the activity of a proton-pumping V-type ATPase (Mindell, 2012). We probed lysosomal pH in VSMCs treated with ChA employing a combination of two dextrans, which readily reach lysosomes via the endocytic pathway – a pH insensitive dextran conjugated to Alexa Fluor 647 (dextran–647) and a pH-sensitive dextran conjugated to fluorescein isothiocyanate (FITC; dextran–FITC). VSMCs exposed to ChA for 72 h presented an approximately twofold higher dextran–FITC:dextran–647 ratio, when compared to POPC-treated cells (Fig. 2D,E), indicating a significant increase in intralysosomal pH. Bafilomycin treatment was used as a positive control (Fig. 2E). Thus, lysosome luminal pH changes may be one of the mechanisms by which ChA causes the accumulation of undigested material within lysosomes.

Next, we determined the activity of several lysosomal hydrolases in ChA-treated VSMCs using artificial fluorescent substrates (in ex vivo assays). The activities of the glycosidases β-hexosaminidase and β-galactosidase were similar between the lysates of POPC- and ChA-treated cells (72 h incubation) (Fig. 2F). In contrast, the activity of the lysosomal acid lipase (LAL, encoded by the Lipa gene) as well as cathepsin B (CTSB) activity were ∼30% decreased in ChA-treated VSMCs when compared to POPC-treated cells (Fig. 2F). Like cathepsins, LAL is synthesized as a zymogen, requiring proteolytic processing at low pH to become active (Ameis et al., 1994; Zschenker et al., 2004). On the other hand, cathepsin D (CTSD) activity was increased in ChA-exposed cells, which may be the consequence of the reported self-activation of pro-CTSD in the acidic assay buffer conditions. In general, these data indicate that in ChA-treated VSMCs the activity of lysosomal hydrolases is either unchanged (β-hexosaminidase and β-galactosidase) or decreased (CTSB and LAL).

To clarify whether ChA was directly affecting the CTSB and LAL activity, we performed in vitro inhibition assays employing human recombinant enzymes in the presence of POPC or ChA liposomes. While increasing ChA concentrations did not significantly alter the activity of recombinant CTSD (data not shown), CTSL (data not shown) or LAL (Fig. 2G) towards their respective artificial substrates, CTSB activity towards the Z-Arg-Arg-AMC substrate was progressively inhibited (Fig. 2H). In this assay, CTSB activity was reduced by up to 20% of control levels in the presence of ChA liposomes (Fig. 2H), suggesting that this lipid may be able to directly inhibit the protease. To confirm whether CTSB inhibition by ChA plays a role in the cellular phenotypes observed, we decided to expose VSMCs simultaneously to ChA and a specific CTSB inhibitor, ZRLR (Reich et al., 2009; Wieczerzak et al., 2007). Indeed, pharmacological abrogation of CTSB activity led to a significant worsening of the lysosomal phenotype (Fig. 2I), as evaluated by the percentage of cells presenting enlarged lysosomes at 48 h (Fig. 2J). Suggesting that CTSB inhibition may contribute to the development of the lysosomal phenotype observed. This led us to speculate whether boosting the activity of CTSB and LAL could prevent the accumulation of neutral lipids and consequent lysosomal enlargement. LAL is responsible for the breakdown of LDL-derived cholesteryl esters and triglycerides within lysosomes and the recombinant enzyme is used as a treatment for patients suffering from LAL deficiency. Following the same rationale, we decided to perform a rescue experiment in which VSMCs were treated with ChA liposomes for 48 h followed by a 24 h exposure to recombinant human (rh)CTSB and/or LAL in the culture medium. Treatment with rhLAL was able to fully revert the lysosomal hypertrophy (Fig. 2K,L) and neutral lipid accumulation (Fig. 2K) caused by ChA. The results with rhCTSB did not reveal a significant improvement under these experiment conditions, which may relate to the relatively long maturation time (>24 h) of rhCTSB taken up by endocytosis (di Spiezio et al., 2021).

As described for other experimental settings (Martina et al., 2016), lysosomal stress could trigger the activation of the MiT/TFE family of transcription factors (TFs) with subsequent increase in lysosome biogenesis and autophagy (Napolitano and Ballabio, 2016). Therefore, next we decided to assess whether ChA was promoting the nuclear translocation and activation of these TFs.

The cellular localization and activity of MiT/TFE TFs is mainly controlled by their phosphorylation status. They are phosphorylated by the mechanistic target of rapamycin complex 1 (mTORC1) or other kinases, and localized in the cytosol in their inactive form (Settembre et al., 2013; Medina et al., 2015). Upon starvation or conditions of lysosomal dysfunction, these TFs are quickly de-phosphorylated and translocated to the nucleus, where they activate the transcription of the CLEAR gene network, leading to an upregulation of genes involved in lysosome functioning and autophagy (Martina et al., 2016; Napolitano and Ballabio, 2016). Taking this into account, we studied the translocation into the nucleus of TFEB, TFE3 and MiTF at different time-points in VSMCs treated with ChA or POPC. Torin-1 and Torin-2, inhibitors of mTOR activity, were used as positive controls. We found a significant increase in the nuclear fraction of MiTF after 24 and 48 h of ChA treatment, compared to POPC-treated cells (Fig. 3A,B) and an increase in nuclear TFE3 48 h after treatment (Fig. 3C,D). Curiously, we did not observe nuclear translocation of TFEB in VSMCs stably expressing human TFEB–GFP (hTFEB–GFP, Fig. 3E,F). To assess whether this could have occurred earlier, we performed the same analysis at earlier timepoints (Fig. S1A–F). At these timepoints, there was also no significant increase in the nuclear fraction of hTFEB (Fig. S1E–F). Importantly, at 72 h of incubation, when changes at the lysosomal level were striking, ChA-treated cells showed decreased levels of total hTFEB–GFP protein (Fig. 3G,H). Similar results were obtained for endogenous murine TFEB (Fig. 3G,I). Furthermore, increased levels of phosphorylated TFEB, its inactive form, were also observed (Fig. 3G,J).

Since mTOR is responsible for TFEB phosphorylation, we decided to address its activation status in ChA-treated VSMCs (Fig. 3K–M). The ratio of phosphorylated-to-total mTOR levels tended to be higher in ChA-treated than in control VSMCs (Fig. 3K,L). In addition, phospho-mTOR beautifully decorated the enlarged lysosomes in ChA-treated VSMCs (Fig. 3M, insets). This leads us to speculate that, at 72 h, ChA may be sufficient to hyperactivate mTOR, culminating with the phosphorylation of its effectors, namely TFEB.

Besides their nuclear translocation, the expression levels of these TFs have also been shown to increase upon exposure to lysosomal stressors (Lu et al., 2017; Pan et al., 2019). MiT/TFE family members are encoded by four distinct genes: Mitf, Tfeb, Tfe3 and Tfec (Hemesath et al., 1994). To assess whether ChA exposure might impact their transcription in VSMCs, we evaluated the expression level of these genes by quantitative real-time RT-PCR (qRT-PCR). VSMC starvation and treatment with Torin-1 were used as positive controls. No major alterations in the mRNA levels of the transcription factors Mitf, Tfe3 and Tfeb were observed in ChA-treated when compared to POPC-treated VSMCs (Fig. 4A–C). The only exceptions were a slight elevation in Mitf transcript levels after 24 h incubation with ChA (Fig. 4A) and a decrease in Tfe3 at 48 and 72 h (Fig. 4B) and Tfeb expression at 24 h (Fig. 4C).

Once in the nucleus, MiTF and TFE3 can promote cell adaptation to lysosome stress by upregulating transcription of numerous lysosomal and autophagic genes (Martina et al., 2016; Sergin et al., 2017). Thus, we assessed next the transcription levels of lysosomal genes (Fig. 4D–J) and autophagic genes (Fig. 4K–L) at different time-points. To our surprise, for all evaluated lysosomal and autophagic genes, we did not observe a significant transcription increase in ChA-treated VSMCs (Fig. 4D–L). In fact, the majority of the genes showed, at the later timepoints, a reduction of their transcript levels when compared to POPC-treated cells. We found that gene expression reduction was more pronounced and significant for Lipa, Mcln1, Lamp2b and Ctsb (Fig. 4D,F,G,J) than for the other genes analysed. Interestingly, the decreased expression of the lysosomal genes coincided with the decline of lysosomal degradation capacity in ChA-treated cells (Fig. 2F).

Next, we analysed the protein levels of autophagy-related genes, whose transcription is decreased in ChA-treated cells (Fig. 4K-L). Microtubule-associated proteins 1A/1B light chain 3B (LC3B, also known as MAP1LC3B; hereafter denoted LC3) is an essential protein for autophagy, as it is crucial for substrate selection and autophagosome biogenesis. In particular, the levels of membrane-associated (lipidated) LC3 (denoted LC3-II) are commonly used as a marker for autophagosomes (Nakatogawa, 2020). At basal conditions, VSMCs treated with ChA presented a tendency towards higher levels of the lipidated LC3-II after 48 h treatment with ChA liposomes, but this tendency was absent at 72 h incubation (Fig. S2A,B). This suggests that the number of autophagosomes was not increased upon ChA treatment and that autophagy was not stimulated. Another protein essential for autophagy is SQSTM1, an autophagy receptor that targets proteins for selective autophagy, which did not exhibit any change in our experimental settings (Fig. S2A,C). Since autophagy does not appear to be stimulated, this may indicate an impairment in autophagic cargo degradation due to lysosome dysfunction.

To gauge the autophagic flux in ChA-treated VSMCs, we inhibited the latter with the weak base chloroquine (CQ). CQ blocks the autophagic flux by increasing lysosomal pH, as well as decreasing autophagosome-lysosome fusion (Mauthe et al., 2018). As shown in Fig. S2B, the LC3-II:LC3-I (LC3-I is non-lipidated LC3) ratio upon treatment with CQ in ChA-exposed and in control cells was similar. Overall, these data imply that, despite the translocation of MITF and TFE3, autophagy in murine VSMCs is not being stimulated by exposure to ChA. This outcome could be attributed, at least partially, to the mTOR activation and subsequent TFEB and possibly unc-51-like kinase (ULK) phosphorylation, both of which are autophagy inhibitors (Kim et al., 2011). Furthermore, the observation that lysosome membrane remains intact in ChA-treated cells (Fig. 1F) makes the occurrence of lysophagy unlikely.

As we observed an inhibition of the transcription of lysosomal genes and lysosomes of VSMCs are dysfunctional following exposure to ChA, we set out to investigate whether ChA exposure results in any change of structural and luminal lysosomal proteins. The protein levels of the lysosomal transmembrane protein LAMP1 were significantly increased in cells treated with ChA, in comparison to POPC-treated cells (Fig. 5A,B). Next, we evaluated the protein levels of luminal lysosomal cathepsins, some of the most abundant soluble lysosomal hydrolases. Analysis of CTSD protein levels revealed a significant increase in pro-CTSD levels (reflected in the total protein levels), which was not accompanied by an increase in mature (lysosomal) CTSD (Fig. 5C–E), suggesting that pro-CTSD is not being properly processed in the lysosomes to the mature (active) form of the enzyme (Marques et al., 2020). When CTSB protein levels were analysed, an increase in its pro-form was also observed (Fig. 5F,G). Immunostaining of CTSD revealed an increase in protein levels within the enlarged lysosomes of ChA-exposed VSMCs (Fig. 5H–I). To understand whether the observed increase in LAMP1 and pro-cathepsins levels was complemented by a similar elevation in lysosome number, we counted the number of lysosomes in POPC- and ChA-treated VSMCs (72 h incubation). Surprisingly, the number of lysosomes was significantly reduced in VSMCs exposed to ChA liposomes (Fig. 5J). Together, our observations indicate that even though lysosomal mass is increased, that did not translate into an increased number of lysosomes, but rather in the enlargement of these organelles. To further understand the increase in lysosomal mass we decided to evaluate whether ChA was affecting lysosomal protein degradation. For that, we treated VSMCs with the translational inhibitor cycloheximide, and after 6 and 12 h post-chasing the total-CTSD levels were analysed. The relative levels of the protease (compared to t=0) were always higher in ChA-treated VSMCs than in control cells (Fig. 5K,L).

Next, we decided to evaluate the impact of lysosome dysfunction in VSMCs viability, an event that can contribute to plaque instability (Grootaert et al., 2015, 2018). We evaluated necrosis and apoptosis by propidium iodide (PI)/annexin V staining. Necrosis was also evaluated by lactate dehydrogenase (LDH) release at 72 h incubation time. We found that ChA-treated cells did not present any increase in the staining by annexin V up to 72 h, when compared with control cells (Fig. 6A). Moreover, when PI staining and LDH activity in the medium was measured, no differences between control and ChA-treated cells were observed (Fig. 6B). These results indicate that ChA treatment did not induce apoptosis or necrosis. However, observations under the microscope suggested that in the presence of ChA the number of VSMC was lower in comparison with control cells. To confirm this observation, we counted the number of cells at 48 and 72 h and, as shown in Fig. 6C, a significant decrease on cell number was observed at both time points in the presence of ChA (Fig. 6C). This can be achieved through delaying the proliferation rate and/or by arresting the cell cycle. To address these two hypotheses, we first evaluated Ki-67 mRNA levels. The results showed a significant decrease in Ki-67 (also known as Mki67) mRNA levels at 72 h, under ChA conditions (Fig. 6D). Given these results, we further evaluated the proliferation using the vital dye carboxyfluorescein succinimidyl ester (CFSE), whose fluorescence intensity decreases with every cell division. As expected, CFSE fluorescence intensities in ChA-treated VSMCs at 48 and 72 h were higher than for the controls (Fig. 6E). These results suggest that there is a decline in VSMC proliferation. Taking these results into account, together with the increase in lysosomal mass, we decided to search for cell senescence markers in VSMCs exposed to ChA.

Cellular senescence can be established by the activation of two different tumour suppressor pathways: CDKN2A (p16)-RB (retinoblastoma) and TP53 (p53)–cyclin-dependent kinase inhibitor 1A (CDKN1A, p21). Thus, we assessed whether these two senescence pathways were activated in VSMCs treated with ChA. As shown in Fig. 7A,B, at 48 h p16 levels were higher in ChA-treated than in control cells. This is in line with the lower number of cells at 48 h (Fig. 6C). Nevertheless, p16 levels in ChA-treated cells at 72 h returned to POPC expression levels. Yet, when we evaluated p21 protein levels, we could observe an increase in ChA-treated cells compared to controls at 72 h (Fig. 7C,D). The expression of p21 is orchestrated by the transcription factor p53, which promptly translocates to the nucleus in response to different stress signals, driving p21 expression (Macleod et al., 1995). After 72 h exposure to ChA, VSMCs presented a significantly increased ratio of nuclear versus cytoplasmic p53, providing evidence for its nuclear translocation (Fig. 7E,F). This is in line with the increased p21 expression levels at this time point. However, this outcome did not result in translocation of γH2AX, a histone involved in DNA repair (Collins et al., 2020), into the nucleus (Fig. 7G), suggesting that our results are more compatible with a state of reversible cell cycle arrest (quiescence) than with cell senescence. Similarly, ChA-treated VSMCs do not exhibit a senescence-associated secretory phenotype (SASP), as shown in Fig. 7H–K.

VSMC migration to the intima is one of the first steps of atherosclerosis, being critical to the formation of the protective fibrous cap over the atheroma. Lysosomal dysfunction may influence VSMC migration, as traffic of some migration machinery goes through the late endosome-lysosome-plasma membrane pathway, namely integrins and matrix metalloproteinases (Dozynkiewicz et al., 2012; Monteiro et al., 2013). Thus, we investigated whether ChA is able to alter the migration ability of VSMCs. To this end, wound healing experiments were performed in the presence of mitomycin C, to synchronize the cell cycle and ensure that the experiment was performed with a similar number of cells. The results revealed that VSMC migration was markedly reduced even at 4 h post ChA addition (Fig. 8A). Taking this into account, the decrease in VSMC migration observed upon ChA exposure may be associated with alterations in lysosome function, as some cells already featured enlarged lysosomes at 8 h of ChA exposure (Fig. 8B, see insets). Afterwards, we evaluated the transmigration capacity of VSMCs after 72 h of ChA exposure. The results indicated that the migration capacity of ChA-treated cells was decreased when compared to controls (Fig. 8C,D). Furthermore, taking the previous results into account, we hypothesized that the described lysosomal dysfunction of ChA-treated cells may have a role in VSMC migration and trans-migration. To investigate this, we treated cells with CQ for 4 h and measured their transmigratory pattern. The results revealed that, upon CQ treatment, the migration capacity was strikingly reduced when compared with ChA treated and control cells (Fig. 8C,D). Overall, these results indicate that ChA treatment leads to a decrease in the migration capacity of VSMCs upon acute (8 h) and chronic (72 h) exposure, and that lysosomes may play an important role in this process.

ChA may also affect the biomechanical characteristics of cells, with consequences in migration and transmigration. To evaluate whether ChA alters the biomechanical properties of VSMCs, atomic force microscopy (AFM) measurements were performed. The AFM tip penetration depth and the Young's modulus were the quantitative parameters extracted from the AFM quantitative images, combining imaging with force spectroscopy. The results revealed that ChA-treated cells present lower penetration depth, indicating that they are less capable of deforming than control cells (Fig. 8E, F). Moreover, ChA-treated VSMCs showed a higher Young's modulus, when compared to the controls (Fig. 8E,G), also indicating that these cells are stiffer than their controls. Owing to the force applied on the cells, the penetration depth values exceeded the thickness of cortical actin. Thus, it is expected that lysosomes would also contribute to the observed increase in cell stiffness. From Fig. 8E (height images), it can also be noticed that VSMCs treated with ChA change their morphology, becoming less elongated and organized in a monolayer, when compared with the control cells. ChA treatment led to local changes on cell deformation (lower penetration depth), apparently at the cytoplasm level, when compared with the penetration depth QI image of the control cells.

Altogether, our data suggest that lysosome dysfunction, cell cycle arrest and changes of the biomechanical properties in ChA-treated VSMCs can lead to pro-atherogenic-like phenotypes.

One of the early characteristics of atherogenesis, observed in lesions of different origin, is the sequestration of lipids by macrophages and VSMCs in their lysosomes. It is accepted that lysosome dysfunction is at the centre of the atherosclerosis pathogenic hub (Fowler, 1980; Fowler et al., 1980; Geer et al., 1961; Jerome and Lewis, 1985; Shio et al., 1979; Wolinsky et al., 1974; Yancey and Jerome, 1998). The decline in lysosomal function is postulated to cause and/or facilitate atherosclerosis. In contrast with the vast amount of literature on lysosome dysfunction in macrophages, very little is known about lysosome dysfunction in VSMCs. Here we found that ChA, a stable end oxidation product of cholesteryl linoleate and arachidonate found increased in plasma of CVD patients (Domingues et al., 2021preprint; Matthiesen et al., 2021) accumulates in VSMCs and causes lysosome malfunction, recapitulating many of the effects seen with oxLDL, including induction of foam cell formation. We also demonstrate that, as consequence of these alterations, these cells become stiffer, migrating and proliferating less. Regarding the ChA concentrations employed, one should bear in mind that liposomes are not very efficient vehicles for ChA delivery (less than 3% of ChA is internalized) (Domingues et al., 2021 preprint), and so we expect the ‘real’ ChA concentrations to be on the micromolar range. This corresponds to the range of concentrations found in the plasma of CVD patients (Domingues et al., 2021 preprint; Matthiesen et al., 2021).

ChA accumulates in lysosomes inhibiting cargo degradation and egress, culminating with lysosomal neutral lipid accumulation and acquisition of a foam cell-like phenotype. The dysfunctional lysosomes are enlarged and perinuclear. Furthermore, ChA-treated VSMCs exhibit an increase in lysosomal mass but have less lipid-loaded lysosomes than control VSMCs. These outcomes can be explained by the fact that ChA is an amphiphilic molecule and partitions via passive diffusion and transmembrane translocation into all cell membranes. Thus, like cholesterol, it is expected that ChA affects membrane lateral packing density, causing changes in the v-ATPase activity, increased luminal pH and decreased acidic hydrolase activity (Maxson and Grinstein, 2014). It is also conceivable that lysosome dysfunction is not only caused by changes in the biophysical properties of cell membranes and lysosome pH. For example, it is possible that ChA, like oxLDL, partly inactivates some lysosomal enzymes. Indeed, here we demonstrate that ChA can directly inhibit the activity of lysosomal CTSB. ChA is not a CTSB substrate; however, we can envision that some hydrophobic interactions may occur between the lipid and the enzyme resulting in the inhibition of the latter. In line with this hypothesis, complete pharmacological inhibition of CTSB activity led to a worsening of the lysosomal phenotype in VSMCs exposed to ChA. In contrast, ChA does not appear to affect LAL activity in vitro. Nonetheless, the addition of recombinant LAL to ChA-treated VSMCs is sufficient to overcome lysosome enlargement and neutral lipid accumulation. As reported in the literature, these results can be explained by the low levels of LAL in VSMCs (Dubland et al., 2021). In our experimental setup, the cells were exposed to ChA for 48 h followed by a 24 h treatment with rhLAL (in the absence of ChA). At the 48 h timepoint, the changes in lysosomal pH are likely not as dramatic as observed at 72 h, and in this way the processing of the exogenous LAL is not hindered by the alkanization of the lysosomal pH, as described by Dubland et al. (2021). Thus, the addition of this recombinant enzyme will facilitate the hydrolysis of neutral lipids, with the loss of the foam cell phenotype.

An increase in lysosome area can then be explained by the accumulation of undigested material inside lysosomes or by the increase in lysosome fusion and decrease in fission or even by an increase in de novo lysosome biogenesis (de Araujo et al., 2020; Saffi and Botelho, 2019). However, in ChA-treated VSMCs, the increase in lysosome area cannot be attributed to lysosomal biogenesis. As such, MiT/TFE TFs, key players in cellular adaptation to lysosome dysfunction, are not activated in ChA-treated VSMCs. TFEB nuclear translocation is even inhibited. The increase in lysosomal area and mass can also be driven by a delay in intracellular degradation of lysosomal protein components, similar to what was described for LAMP2 in the absence of protective protein (also known as cathepsin A) (Cuervo et al., 2003). Another possibility is that mTOR mediates a TFEB-independent boost in lysosomal protein translation, similar to that shown to take place in macrophages (Hipolito et al., 2019). In order for this to occur, both S6K and 4E-BP effectors must be phosphorylated by mTORC1 (Hipolito et al., 2019). However, in our experimental settings, we could only observe an increase in the phosphorylation of S6K but not 4E-BP (data not shown), making this hypothesis less likely.

ChA-treated VSMCs are also stiffer than control cells (see working model shown in Fig. S3). The increase in cell stiffness may explain, in part, the effects on cell quiescence and migration that are crucial to form the fibrous cap that stabilizes atherosclerotic lesions. Finally, the increase in cell stiffness together with lysosome dysfunction could abrogate phagocytosis of lipids, dead cells and cell debris (Santarino et al., 2017; Viegas et al., 2012) contributing to the aetiology and development of the atherosclerotic lesions.

Cholesteryl hemiazelate [ChA; cholesteryl O-(8-carboxyoctanoyl)] was prepared following a general procedure described in the literature for the synthesis of cholesteryl hemisuccinate (Klein et al., 1974). However, in order to optimize reaction conditions, molar equivalents of the reactants, reaction time and purification conditions were adapted. The synthetic strategy involves the reaction of commercially available cholesterol with azelaic anhydride, which was prepared by the reaction of azelaic acid with acetyl chloride, as described elsewhere (Hill and Carothers, 1933). Cholesterol was reacted with 2.6 molar equivalents of azelaic anhydride in dry pyridine under reflux for 7 h. Purification by flash chromatography with chloroform/methanol/ammonia (50:5:0.25), followed by recrystallization from ethanol, gave the target cholesteryl hemiazelate as a white solid in 57% yield. After the reaction, ChA is purified by flash chromatography with chloroform/methanol/ammonia (to remove unreacted cholesterol) and recrystallized from ethanol. ChA purity is assessed by nuclear magnetic resonance (NMR; ¹H and ¹³C) and melting point to ensure that the product was not contaminated with unreacted reagents, pyridine or any solvent.

Lipid aqueous suspensions were prepared by mixing POPC (Avanti Polar Lipids, Alabaster, AL, USA) and ChA at 35:65 molar ratio in an azeotropic mixture of chloroform and methanol, and then incubated for 30 min. The solvent was evaporated using a rotary evaporator and dried during 30 min in a 65°C water bath. The lipid film was hydrated with a buffer solution 20 mM HEPES, 0.11 M NaCl, 1 mM EDTA, pH 7.4 in a water bath at 65°C for at least 1 h. Samples were submitted to mild sonication for 10 min and extruded through a polycarbonate filters (Nucleopore, Whatman, Little Chalfont, UK) with a pore diameter of 0.4 μm. This process was repeated at least six times. During extrusion, the water-jacketed extruder (Lipex Biomembranes, Vancouver, British Columbia, Canada) was maintained at 65°C. ChA concentration was determined using the Liebermann–Burchard protocol (Huang et al., 1961).

Murine aorta smooth muscle cells (MOVAS, CRL-2797™, ATCC, Manassas, VA, USA) were grown in Dulbecco's modified Eagle's medium (DMEM GlutaMAX, Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% heat inactivated fetal bovine serum (FBS, Gibco), 1 mM pyruvate (Gibco) and 0.2 mg/ml G-418 (Sigma), and passaged two to three times before their use in the different assays. Cells were grown in a humidified incubator at 37°C under 5% CO₂. After 24 h of seeding, VSMCs were incubated with ChA:POPC (65:35, molar ratio, ChA) or POPC liposomes, as control, for up to 72 h, as indicated in the figure legends. The concentrations used in most of the experiments was 1500 µM ChA and 807.7 µM POPC (the equivalent to the POPC concentration present in ChA liposomes). The MOVAS cell line was recently authenticated and tested negative for mycoplasma contamination.

Primary antibodies used were anti-LAMP1 and anti-LAMP2 (1:500; Hybridoma Bank, 1D4B and ABL-93), anti-EEA-1 (1:50; Santa Cruz Biotechnology, SC-6415), anti-CTSD (1:100; SICGEN, AB0043-200), anti-MITF (1:50; Tebu-bio, B0512), anti-TFE3 (1:100; Sigma, HPA023881), anti-p53 (1:50; SICGEN, AB0154-200), anti-SQSTM1 (1:50; Abgent, AP2183B), anti-phospho-histone H2A.X (1:500; Cell Signaling Technology, 20E3 #9718), and anti-phospho-mTOR (Ser2448) (1:50; Cell Signaling, #5536). As a positive control for H2A.X staining, the cells were treated with 100 µM etoposide (Sigma, E1383) 16 h prior to fixation.

For immunofluorescence (IF) staining, VSMCs were grown in the presence of lipids on glass coverslips. After incubation, cells were fixed with 4% paraformaldehyde (PFA) for at least 30 min, followed by quenching of the aldehyde groups with ammonium chloride and permeabilization with saponin (0.1% in PBS). After permeabilization, cells were blocked with 1% BSA in PBS or for lipid staining 2% fish gelatin in PBS (Bodipy and filipin). Cells were then incubated with the primary antibodies in blocking solution for 1 h at room temperature (RT) (LAMP1 and LAMP2, p53, SQSTM1, EEA-1, CTSD and H2A.X) or overnight at 4°C (MITF, TFE3 and TFEB). Then they were washed, and finally incubated for 1 h with the secondary antibodies conjugated with a fluorophore. Antibody dilution was 1:500 for secondary antibodies conjugated with Cy3 and 1:250 for those conjugated with Cy5. Secondary antibodies used were from Jackson ImmunoResearch Laboratories (West Grove, PA, USA). Neutral lipids were stained with BODIPY 493/503 (diluted 1:500 from a saturated ethanolic solution of BODIPY) for 1 h. Free cholesterol and ChA were stained with filipin (25 μg/ml in PBS) for 30 min at room temperature. DAPI was used to visualize nuclei (1:800; Fluka). Coverslips were mounted with mowiol/DABCO and images were obtained using an AxioVision microscope (Axio Observer Z2), a Zeiss LSM710 confocal microscope or a Zeiss LSM980 confocal microscope with a 63× oil-immersion objective, using the appropriate filter sets (1.4 NA).

Images were randomly acquired, and ten different fields were imaged for each experimental condition. For a given staining, images were acquired with the same settings. Images were obtained from three independent experiments and in total more than 15 cells were analysed per condition. Quantification of the number and size of lysosomes was performed using ImageJ software. Background was subtracted from the total fluorescence intensity of each region of interest (ROI). For nucleus translocation, the ratio of the nucleus and cytoplasm fluorescence intensities was also calculated using ImageJ.

For transmission electron microscopy, cells were seeded on glass coverslips and, after POPC and ChA treatment, fixed in 2% PFA with 2% glutaraldehyde (Electron Microscopy Sciences, Hatfield, PA, USA) in 0.1 M phosphate buffer. Cells were then osmicated, treated with tannic acid, dehydrated, infiltrated with Epon and mounted on Epon stubs, all as previously described (Tomas et al., 2004). After polymerization overnight at 60°C, coverslips were removed from Epon stubs with liquid nitrogen. 70 nm sections were cut en face and stained with lead citrate, before examination on a Jeol 1010 transmission electron microscope and images acquired with a Gatan OriusSC100B charged coupled device camera.

Cells were seeded in a Nunc™ Lab-Tek™ Chamber Slide (Thermo Fisher Scientific) and treated for 3 days with ChA or POPC. At day 2, cells were incubated overnight simultaneously with dextran conjugated with FITC-dextran (200 μg/ml; Sigma) and dextran conjugated to Alexa Fluor 647 (50 μg/ml; Thermo Fisher Scientific). To ensure that lysosomes were labelled by the overnight pulse, cells were washed with fresh culture medium and analysed by fluorescence microscopy after 3 h in dextran-free medium. Image acquisition was performed in CO₂-independent medium, and the settings were not changed between different samples. To measure the fluorescence intensity ratio between FITC–dextran and Alexa Fluor 647 within the lysosomes, the organelles were delimited and the fluorescence intensities of the two channels were quantified with ImageJ Fiji software (Coloc2 plugin).

Cells were seeded in a Nunc™ Lab-Tek™ Chamber Slide (Thermo Fisher Scientific) and treated for 3 days with ChA or POPC. At day 3, cells were incubated with BODIPY^® FL C5-Lactosylceramide – LacCer (2.5 μM; Invitrogen, Carlsbad, CA, USA) in DMEM with 1% FBS for 45 min and LysoTracker™ Red DND-99 (200 nM; Invitrogen) to stain the acidic organelles, for 15 min at 37°C. Cell surface fluorescence of LacCer was removed by washing the cells three times with 10% serum. Cells were followed under a confocal microscope for 3 h at 37°C in DMEM with 10 mM HEPES and 10% FBS. Images were obtained for both dyes. Colocalization was calculated using the Manders (M1) coefficient from ImageJ JACoP plugin.

CTSD activity was assayed as previously described (Marques et al., 2020). CTSB activity was determined in a similar manner using 20 μM of cathepsin B substrate Z-RR-AMC (Enzo Life Sciences, Farmingdale, NY, USA) (di Spiezio et al., 2021). β-hexosaminidase activity was measured with 1.97 mM 4-methylumbelliferyl-N-acety-β-D-glucosaminide (Sigma) in 150 mM citrate-Na₂HPO₄ (pH 4.0) buffer. β-galactosidase activity was measured with 0.64 mM 4-methylumbelliferyl-β-D-galactoside (Sigma) in 150 mM citrate-Na₂HPO₄ (pH 4.0) buffer with 0.2 M NaCl. The 4-methylumbelliferyl (4-MU) substrates for these two gycosidases were kindly gifted by Prof. Johannes Aerts (University of Leiden, The Netherlands). For these two activities, after incubation at 37°C for 30 min, the reaction was quenched with 0.3 M glycine adjusted with NaOH to pH 10.6. LAL activity was assayed based on the method described by Hamilton et al. (2012), with slight alterations. Briefly, cells were lysed in reaction buffer (100 mM sodium acetate pH 4.0 with 1% Triton X-100). The lysates were cleared by centrifugation (17,000 g for 10 min) and 5 µg of lysate were incubated for 90 min at 37°C with a final concentration of 0.345 mM 4-MU-palmitate (Santa Cruz Biotechnology, Dallas, TX, USA) and 0.03% cardiolipin (w/v; Sigma) in 100 µl reaction buffer. Samples were incubated in the presence and absence of 30 µM Lalistat2 (Sigma), to inhibit LAL activity, and the reaction was quenched with 150 mM EDTA pH 11 (190 µl). For all 4-MU substrates, fluorescence was measured by using an Tecan Infinite F200 PRO microplate reader (Tecan, Männedorf, Switzerland), with λ_exc=366 nm and λ_em=445 nm.

Recombinant human CTSB (rhCTSB) was kindly provided by Prof. Paul Saftig and Dr Alessandro Di Spiezio (University of Kiel, Germany). rhCTSB (2 µg) was pre-incubated for 30 min at room temperature in 100 µl of 50 mM sodium acetate, pH 5.5, 0.1 M NaCl, 1 mM EDTA, and 0.2% Triton X-100, for auto-activation of the proCTSB into the mature (active) form of the enzyme. After this period, 20 μM of cathepsin B substrate Z-RR-AMC was added. The POPC and ChA:POPC liposomes were added to this mixture in the indicated concentrations (50, 100, 250, 500, 1000, 1500, 2000 and 2500 µM) and incubated for 30 min on ice. Leupeptin (25 µM, Sigma, #L2884) was used as a negative control. The samples were incubated 3 h at 37°C for hydrolysis of the substrate to occur. Afterwards, the reaction was quenched with 190 µl of 0.3 M glycine adjusted with NaOH to pH 10.6. Recombinant human LAL (rhLAL, Kanuma^®, Sebelipase alfa) was kindly provided by Dr Eugen Mengel (SphinCS GmbH, Institute of Clinical Science for LSD, Hochheim, Germany). The rhLAL assay was similar to that described for rhCTSB. Briefly, 100 ng of rhLAL were pre-incubated in assay buffer (100 mM sodium acetate pH 4.0 with 1% Triton X-100) for 30 min at room temperature. Then, the substrate 4MU-palmitate (0.345 mM) was added together with the liposomes and incubated 30 min on ice. Lalistat2 was used as a negative control. Samples were incubated 1 h at 37°C and 190 µl of 150 mM EDTA pH 11 was used to stop the reaction. The fluorescence was read in a microplate reader MolecularDevices SpectraMax i3x (exc: 360 nm; em: 440 nm).

The cells were incubated for 48 h with the liposomes in coverslips. Afterwards, fresh culture medium supplemented with 1 µg rhLAL or 5 µg rhCTSB was added, and cells were incubated for a further 24 h. For the CTSB inhibition experiment, the cells were treated for 48 h with liposomes in the presence of 100 µM ZRLR, a specific CTSB inhibitor kindly provided by Dr Ewa Wieczerzak (Department of Biomedical Chemistry, Faculty of Chemistry, University of Gdañsk, Gdañsk, Poland; Reich et al., 2009; Wieczerzak et al., 2007). After the mentioned incubations, the cells were fixed with 4% PFA and processed for immunohistochemical analysis.

Cells were cultured in coverslips for 72 h in the presence of liposomes, after which 2 mM L-leucyl-L-leucine methyl ester hydrochloride (LLOMe; Bertin, #16008) dissolved in DMSO was added to the culture medium, and cells were incubated for a further 2 h. DMSO was used as control. The cells were fixed with 4% PFA and processed for immunohistochemical analysis.

The plasmid pN1-CMV-TFEB-GFP from was obtained from Addgene (#38119 deposited by Dr Shawn Ferguson's lab). The TFEB–GFP sequence was amplified using primers (forward, 5′-GGGACAAGTTTGTACAAAAAAGCAGGCTAAATGGCGTCACGCATAGGGTTGCGCAT-3′; reverse, 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTATTACTTGTACAGCTCGTCCATGCCGAG-3′) containing the attB1/attB2 Gateway^® sites. The amplified sequences were cloned into pDONRTM (Thermo Fisher Scientific) using the Gateway™ BP Clonase™ II enzyme (Thermo Fisher Scientific), followed by subcloning into pLenti6 (Thermo Fisher Scientific) with the Gateway™ LR Clonase™ II enzyme according to the manufacturer's instructions. All lentiviral particles were produced by co-transfection of pLenti6 and pMD 2.G (VSV-G protein) and psPAX2 (Rev and Pol proteins) into a producer cell line 293STAR RDPro (ATCC). Recombinant viral particles were harvested 48 days later, cleared for cell debris by centrifugation at 3200 g for 10 min and used.

Cells were seeded on six-well format plates (100,000 cells per well), corresponding to approximately confluency after 24 h. Supernatant of the producer cell line containing the viral particles was added to the cells (500 µl per well). Antibiotic selection was initiated 72 h after transduction by replacing the medium with culture medium supplemented with 10 µg/ml blasticidine (Sigma, #15205). After selection, the mixed clone was cultured for two passages, after which the population was sorted in a FACS Aria III Cell Sorter. The population with the 2% highest TFEG–GFP signal was selected and cultured in the presence of antibiotic (gentamicin; Sigma, 1405-41-0) and antifungal (amphotericin B; Thermo Fisher Scientific, 15290018) until confluency. After the first passage, the cells were cultured in medium with blasticidine, which was removed during the experiments.

The cells were cultured in glass coverslips and incubated with liposomes for the indicated period. As positive control, the cells were treated with 1 µM Torin-1 (ApexBio, A8312) or 0.5 µM Torin-2 (LC Labs, T-8448) in DMSO and added to the cells in culture medium without serum.

Cells were seeded and exposed to ChA or POPC for 48 or 72 h. To detect cell death, annexin V-PI double staining was undertaken with the FITC Annexin V Apoptosis Detection Kit I (BD Biosciences, Sparks, MD, USA) and used according to the standard protocol provided by the supplier. Briefly, after lipids exposure, cells were washed 1× with PBS and incubated with annexin V-PI solution for 30 min, protected from light, in an agitator. Then, images were taken with a Zeiss Axiovert 40 microscope, and PI- and annexin V-positive cells were counted. As positive control, cells were exposed to 800 µM of H₂O₂ for 2 h prior to staining.

The activity of the cytoplasmic enzyme LDH in the supernatant, which is used to assess plasma membrane integrity, was determined in VSMCs. After treatment with POPC or ChA for 72 h, the supernatant was collected, and cells were lysed. LDH activity of the supernatant, cell culture medium and cell lysates (intracellular content) was determined with the Pierce LDH Cytotoxicity Assay Kit (Thermo Fisher Scientific), according to the standard protocol provided by the supplier.

Cell proliferation was assessed by two different methods. Cell number was assessed by counting the viable cells at each time point, with a Bürker chamber, in the presence of Trypan Blue, a dye exclusion. Cell proliferation was also evaluated by measuring the fluorescence of carboxyfluorescein succinimidyl ester (CFSE) at 48 and 72 h incubation time, in the presence of the lipids. CFSE covalently labels long-lived intracellular molecules with the fluorescent dye. When a CFSE-labelled cell divides, its progeny is endowed with half the number of carboxyfluorescein-tagged molecules and, thus, each cell division can be assessed by measuring the corresponding decrease in cell fluorescence via flow cytometry. In detail, VSMCs were seeded and allowed to grow for 24 h. Then, we used the CellTrace™ CFSE Cell Proliferation Kit (3 µM; Thermo Fisher Scientific), following the protocol provided by the supplier. After 48 or 72 h with POPC and ChA, cells were tripsinized and fixed with 4% PFA. CFSE levels were acquired in a BD FACScantoTMII flow cytometer (CFSE fluorescence in the 488 nm laser with filter 530/30 BP). The software for acquisition was BD FACSDIVA Diva, and data from at least 5000 cells were analysed with FlowJo. Cellular proliferation is an essential feature of the adaptive immune response. The introduction of the division tracking dye CFSE has made it possible to monitor the number of cell divisions during proliferation and to examine the relationship between proliferation and differentiation.

Primary antibodies used were anti-LAMP1 (1:1000; Hybridoma Bank, 1D4B), anti-LC-3 (1:1000; Sigma, L8918), anti-SQSTM1 (1:500; Abnova, H00008878-M0), anti-calnexin (1:1000; SICGEN, AB0037), anti-cathepsin D (1:500; Abcam, ab75852), anti-cathepsin B (1:500; R&D Systems, AF965), anti-p21 (1:1000; Santa Cruz Biotechnology, sc-397), anti-GAPDH (1:500; SICGEN, AB0049), anti-mTFEB (1:500; Assay Biotech, C10428), anti-hTFEB (1:1000; Cell Signalling #4240S), anti-phospho-TFEB (1:1000; Millipore, ABE1971-I), anti-p16-INK4A (1:500; Proteintech, 10883-1-AP), anti-mTOR (7C10) (1:500; Cell Signaling, #2983), and anti-phospho-mTOR (Ser2448) (1:500; Cell Signaling, #5536).

Cell lysates were prepared in lysis buffer (50 mM Tris-HCl, pH 8.0, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 2 mM EDTA, 150 mM NaCl) in the presence of protease inhibitors (Sigma) and phosphatase inhibitor (Calbiochem). Lysates were cleared by centrifugation at 4°C for 30 min at 17,000 g and protein concentrations were determined using the Pierce^™ BCA Protein Assay Kit (Thermo Fisher Scientific). Samples were then mixed with 4× Laemmli buffer (4% SDS, 20% glycerol, 10% 2-mercaptoethanol, 0.004% Bromophenol Blue and 0.125 M Tris-HCl), heated for 5 min at 95°C and loaded (20–40 μg) on 12–15% SDS polyacrylamide gels. After electrophoresis, proteins were transferred into activated PVDF membranes in 1× transfer buffer (25 mM Tris, 192 mM glycine and 20% methanol) for 75 min, at 300 mA, with gentle agitation. Membranes were then blocked with blocking buffer (5% non-fat dry milk and 0.1% Tween-20 in TBS) for 1 h at room temperature, followed by an overnight at 4°C with the primary antibody diluted in blocking buffer.

After incubation with primary antibodies, membranes were washed with 1× TBS tween and incubated for 1 h at room temperature with the corresponding horseradish peroxidase-conjugated secondary antibody (Bio-Rad, Hercules, CA, USA) diluted in blocking buffer. Blots were visualized using ECL Prime Western Blot Detection reagent (GE Healthcare) and a Chemidoc Touch Imaging System (Bio-Rad). Bands were quantified using Image Lab 6.0 software (Bio-Rad).

Cells were cultured for 48 h in the presence of ChA liposomes. Afterwards, the medium was removed and replaced by fresh medium with 10 µg/ml cycloheximide (Sigma, C1988) in DMSO. DMSO was used as control. After the mentioned incubation periods (6 or 12 h) the cells were washed and proposed for western blot analysis.

As positive control for the induction of the expression of lysosomal genes, cells were starved for 4 h by incubation with medium without serum or treated with 1 µM Torin-1 for 4 h also in medium without serum. Total RNA was extracted with the NZY total RNA isolation kit (NZYTech, Lisbon, Portugal) and reverse transcription was performed using NZY first-strand cDNA synthesis kit (NZYTech), following the protocol provided by the supplier. Quantitative PCR was performed in a 96-well plate using the SYBR green master mix (NZYtech) using an AB7300 real-time PCR thermal cycler with Step One software (v2.2.2; Applied Biosystems, Waltham, MA, USA) and QuantStudio™ 5 Real-Time PCR System (Thermo Fisher). Gapdh and Pgk1 were used as housekeeping genes to normalize the expression. Target gene expression was determined by relative quantification (ΔΔCt method) to the housekeeping reference gene and the control sample. Primer sequences are indicated in Table S1.

Supernatant from VSMCs were collected after 72 h of treatment with POPC and ChA. Cytokines levels in the medium were quantified with a Mouse IL-6 Uncoated ELISA kit (Invitrogen), according to the manufacturer's instructions. Since the number of cells in POPC- and ChA-treated cells were different, cytokines levels were normalized to the total cell protein levels, which was quantified as described for the western blots.

Cell migration was assessed by means of a wound-healing assay. VSMCs were grown to confluence and 20 mg/ml mitomycin C (Sigma) was added to block the cell cycle 2 h prior to the wound. Cells were wounded by a scratch injury line made with a sterile cell scraper. Cells were then treated with POPC and ChA, and images were taken every 2 h up to 8 h incubation.

For transmigration assays, we used Transwell cell culture chambers (Costar, Cambridge, MA, USA) containing filters with 8 μm pore size. VSMCs were treated with POPC and ChA for 72 h and then trypsin-harvested. Chloroquine (Sigma, C6628) was used as positive control for migration inhibition. Cells were suspended in serum-deprived medium and added (10⁵ cells/well) in the upper chamber of a 24-well plate. Then, complete medium was added to the lower chamber and cells were allowed to transmigrate for 4 h. After incubation, cells were fixed with 4% PFA and stained with 0.1% (w/v) Crystal Violet in 20% methanol, for 15 min. Cells remaining in the upper chamber of the Transwells were removed with a cotton swab and cells that had migrated onto the lower-surface membranes were counted. Images were captured using an Axiovert 40C inverted microscope (Carl Zeiss) equipped with a Powershot A640 digital camera (Canon).

Cell elasticity was measured using a NanoWizard II atomic force microscope (JPK Instruments, Berlin, Germany) mounted on top of an Axiovert 200 inverted microscope (Carl Zeiss, Jena, Germany). Nanoindentation experiments were carried out on live cells, at 25°C, in DMEM. Quantitative imaging (QI) mode was used to scan the cells. For these measurements, non-functionalized qp-BioAC CB2 AFM cantilevers (Nanosensors, Neuchâtel, Switzerland) with partial Au coating and quartz-like tips (nominal force constant of 100 pN/nm) were used. Differential interference contrast (DIC) microscopy was used to position the tip on top of the ChA-treated and control cells (only with POPC). QI images of 100 µm×100 µm in a z-range of 2.4 µm were acquired. Images of 256×256 pixels with a pixel time of 30 ms were performed. For every contact between cell and tip, the distance between the cantilever and the cell was adjusted to maintain a maximum applied force of 650 pN before retraction. QI height, adhesion and elasticity images were acquired. Images were analysed to obtain the cells Young's modulus (E), using JPK Image Processing software v. 6.0.55, by the application of the Hertzian model. AFM tip penetration depth onto cells was also evaluated. This parameter was analysed by the position of the maximal movement of the piezo sensor in the z-axis, which corresponds to the z-axis coordinate when the sensor reaches an indentation force of 650 pN, subtracting the z-axis position of the sensor when the tip begins the contact with the cell surface. Values above 100 kPa were not considered for the analysis of the stiffness of the cells. A cut-off of values above 1 µm of cell indentation height was also performed for the cell penetration depth images.

Data are representative of at least three independent experiments and values depicted on graphs are expressed as mean±s.d., unless stated otherwise. Statistical analysis [two-tailed unpaired t-test, one-way ANOVA followed by a Tukey post-test (Fig. 8D) or two-way ANOVA followed by Sidak's post-test] was performed using the GraphPad Prism software v. 8.0.2. P<0.05 (*), P<0.01 (**), P<0.001 (***) and P<0.0001 (****) were considered to be statistically significant.

We acknowledge the technical support of the CEDOC Microscopy and Flow cytometry facilities. We would like to thank Dr Eugen Mengel for providing us with the recombinant LAL and Prof. Paul Saftig and Dr Alessandro Di Spiezio for the recombinant CTSB. We also thank Dr Ewa Wieczerzak for the gift of the ZRLR inhibitor, Dr Miguel Cavadas for the Torin-1 and Dr Elena Baena for the Torin-2. Some of the text and Figs 4–8 and some panels of Figs 1–3 and 5 in this paper formed part of Liliana Teresa da Silva Alves’ PhD thesis in the Faculdade de Ciências Médicas (NOVA Medical School) at Universidade NOVA de Lisboa and Universidade do Algarve in 2020.

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.254631.