Transthyretin (TTR)-related amyloidoses are diseases characterized by extracellular deposition of amyloid fibrils and aggregates in tissues composed of insoluble misfolded TTR that becomes toxic. Previous studies have demonstrated the ability of small compounds in preventing and reversing TTR V30M deposition in transgenic mice gastrointestinal (GI) tract as well as lowering biomarkers associated with cellular stress and apoptotic mechanisms. In the present study we aimed to study TTR V30M aggregates effect in autophagy, a cellular mechanism crucial for cell survival that has been implicated in the development of several neurodegenerative diseases. We were able to demonstrate in cell culture that TTR V30M aggregates cause a partial impairment of the autophagic machinery as shown by p62 accumulation, whereas early steps of the autophagic flux remain unaffected as shown by autophagosome number evaluation and LC3 turnover assay. Our studies performed in TTR V30M transgenic animals demonstrated that tauroursodeoxycholic acid (TUDCA) and curcumin effectively reverse p62 accumulation in the GI tract pointing to the ability of both compounds to modulate autophagy additionally to mitigate apoptosis. Overall, our in vitro and in vivo studies establish an association between TTR V30M aggregates and autophagy impairment and suggest the use of autophagy modulators as an additional and alternative therapeutic approach for the treatment of TTR V30M-related amyloidosis.
Familial Amyloid Polineuropathy is a life threatening disorder caused by mutant TTR V30M for which a cure is still lacking. The molecular mechanisms of the disease are not yet completely understood. This study was performed in order to understand the relationship between TTR V30M and autophagy, a cellular mechanism recently implicated in the development of other neurodegenerative diseases.
Our in vitro data shows evidence of autophagy impairment in the presence of TTR V30M aggregates and the ability of TUDCA and curcumin in re-establishing the autophagic flux in the gastrointestinal tract of a transgenic TTR mouse model, a common site for TTR deposition.
These results propose the use of autophagy modulators as an alternative therapeutic strategy or to be tested in combination with current treatments for TTR V30M related amyloidosis.
Familial amyloid polyneuropathy (FAP) is an autosomal hereditary disease characterized by the extracellular deposition of amyloid fibrils essentially composed by mutant transthyretin (TTR) . The most common mutation is characterized by the substitution of a methionine residue for valine at position 30 (V30M) . Portugal is the main focus of the TTR V30M mutation, with approximately 600 kindreds affected, making it a major national health problem .
It is accepted that amyloid fibril formation may be initiated by tetramer dissociation into non-native monomers that partially misfold and have high predisposition to aggregate [4,5] contributing to the cellular toxicity observed in FAP .
Several studies support the involvement of different cellular mechanisms and up- or down-regulation of specific markers either as cause or consequence of the disease. But until now only a few of them have addressed autophagy in FAP and even those studies are superficial [7–10]. More clarification regarding FAP in this basic and yet complex mechanism that is autophagy is warranted.
Macroautophagy (hereinafter referred to as autophagy), is a finely tuned mechanism that involves delivery of large protein aggregates, defective organelles and other cellular debris to lysosomes for degradation and possible recycling. It was only in the last decade that this process has gained relevance and became a focus of interest in the scientific community when increasing evidence suggested that it may be related to the development of several diseases. The concept of autophagy has dramatically changed over the years; at first it was looked at as a cell-death pathway, now, as more knowledge has gathered, it is clear that it is a crucial mechanism for cell survival under critically stressful conditions .
The identification of (some of) the key players of this complex machinery greatly facilitated the detection and evaluation of the autophagic status as well as the design of drugs or compounds capable of modulating the process.
Two of the most broadly used autophagic markers are light chain 3 (LC3)-II and p62. LC3 is cleaved by Atg4 to form LC3-I. The latter is then conjugated to phosphatidylethanolamine to form LC3-II which in turn is targeted to the forming phagophore. When membrane closure is imminent, LC3-II molecules on the cytosolic side of the autophagosome detach and can be recycled, whereas luminal LC3-II remains attached to the membrane until fusion with lysosomes occurs where degradation takes place .
p62/sequestosome 1 (SQSTM-1), on the other hand, is responsible for cargo sequestration into autophagosomes where it provides the linkage between the referred cargo and the LC3-II associated with the membrane. This bond is maintained until delivery and degradation in lysosomes. Therefore this marker provides a reliable readout for autophagic degradation .
In the present study we show results on these markers both in cell culture and in animal studies.
MATERIALS AND METHODS
EGFP–LC3 plasmid was a gift from Professor Luís Almeida from the Center for Neuroscience and Cell Biology–CNC, Coimbra, Portugal.
Human embryonic kidney (HEK)-293 cells (European Collection of Cell Cultures) stably expressing TTR wild-type (WT) and TTR V30M with a tetracysteine tag to allow TTR fluorescent detection  were cultured in EMEM (Eagle's minimum essential medium, Lonza), supplemented with 10% (v/v) FBS, 2 mM L-glutamine, 100 unit/ml penicillin/streptomycin (all from Gibco) and 1% (v/v) non-essential amino acids (Sigma–Aldrich). SH-SY5Y cell line (human neuroblastoma, European Collection of Cell Cultures) was cultured in EMEM/Ham's F12 (1:1, Lonza), supplemented with 10% FBS, 2 mM L-glutamine, 100 unit/ml penicillin/streptomycin and 1% non-essential amino acids. All cell lines were maintained at 37°C, in a 5% CO2 humidified atmosphere.
As described by Batista et al. , the modified HEK-293 cell line used in this study had been obtained as follows: the human TTR cDNA was cloned into NheI and XhoI sites of the CSCW2-IRES-mCherry lentiviral vector . Expression of the vector is driven by the cytomegalovirus (CMV) promoter. The TTR V30M variant was generated with the QuikChange site-directed mutagenesis kit (Agilent) according to the manufacturer's instructions and using the WT TTR vector as a template. Lentiviral vector stocks were produced as previously described . Titres were determined on HEK-293T cells as transducing units (tu) using serial dilutions of vector stocks and infection in the presence of 8 μg/ml Polybrene (Millipore).
HEK-293 cells were then infected with the lentivirus vectors at a multiplicity of infection (M.O.I.)=100 tu/cell in the presence of 8 μg/ml polybrene. At 48 h after infection, cells were assayed for expression of the transgenes by measurement of mCherry expression, yielding on average 90% of infected cells with no apparent toxicity. Cells were then propagated.
TTR V30M aggregation detection
Conditioned medium from HEK-293 cells expressing TTR V30M was subjected to a slot-blot filter assay, using conditioned medium having TTR WT as control. Briefly, medium containing 600 ng of either TTR WT or TTR V30M was applied to a 0.2 μm pore cellulose acetate membrane filter (Whatman) using a manifold system and subjected to vacuum. The membrane was then immunoblotted using rabbit polyclonal anti-TTR (Dako, A0002) as a primary antibody. Filter-retained aggregates were detected by the ECL method (GE Healthcare Bioscience).
Conditioned medium preparation and ELISA
HEK-293 cells stably expressing TTR WT or TTR V30M were seeded in 100 mm diameter Petri dishes. At 24 h after plating, medium was replaced with Dulbecco's modified Eagle's medium (DMEM) without Phenol Red supplemented with 5% FBS, 2 mM L-glutamine and 100 units/ml penicillin/streptomycin and cells were allowed to produce and secrete TTR for ~24 h. Afterwards, medium was collected, centrifuged for 10 min to discard cells and other cellular debris, filtered through a 0.45 μm filter and TTR concentration was measured by ELISA. For ELISA quantification, 96-well plates (Maxisorp, Nunc) were coated overnight at 4°C with anti-TTR (Dako, A0002) in carbonate/bicarbonate coating buffer, pH 9.6, and blocked, also overnight with 5% (w/v) skim milk in PBS; conditioned media were loaded into the wells for 1 h at room temperature; a second anti-TTR antibody (Abcam, ab9015) was used followed by phosphatase alkaline-conjugated anti-sheep antibody (Sigma–Aldrich, A5187); development was performed with alkaline phosphatase yellow liquid substrate (Sigma–Aldrich, P7998). Absorbance was measured at 405 nm.
This GFP–LC3 molecule has a particularity that, when delivered to the lysosomes, the LC3 part is sensitive to degradation whereas the GFP protein is relatively resistant, therefore it is possible to make assumptions about flux status upon observation of a higher or lower expression of the free GFP band by immunoblot.
SH-SY5Y cells were plated in six-well multiplates containing a 10 mm glass coverslip in each well and transiently transfected with EGFP–LC3 plasmid for 24 h. Medium was then replaced with normal or conditioned medium containing 2 μg of TTR V30M for 18–20 h. Bafilomycin A1 (100 nM, Sigma–Aldrich, B1793) or vehicle (DMSO, Sigma–Aldrich) were added in two different regimens: (i) either for the entire time of the experiment (20 h) or (ii) 2 h prior to experiment termination (at 18 h).
After 18–20 h, coverslips were collected and cells were fixed in 4% paraformaldehyde and mounted with Vectashield mounting medium with DAPI (Vector Laboratories). The number of punctate dots per cell in GFP–LC3-positive cells were counted using a TCS SP5 II laser-scanning confocal microscope (Leica Microsystems). At least 15 cells were counted for each condition using ImageJ software (NIH). The remaining cells in each well were lysed and immunoblotted.
Cells were lysed using radioimmune precipitation assay (RIPA) buffer (50 mM Tris/HCl, 150 mM NaCl, 0.5% sodium deoxycholate, 1% NP-40, 0.1% SDS, 1 mM PMSF, 1 mM Na3VO4, 50 mM NaF and 1× protease inhibitors mixture from GE Healthcare Bioscience). The supernatant from the centrifuged lysates was collected and the concentration of protein was determined using the Bio-Rad assay kit (Bio-Rad Laboratories, 500-0006). Protein (20–30 μg) was then mixed with 4× Laemmli buffer containing 240 mM Tris/HCl, 8% SDS, 40% glycerol, 0.04% Bromophenol Blue and 10% 2-mercaptoethanol and incubated for 5 min at 95°C. Proteins were separated by SDS/PAGE and transferred onto nitrocellulose membranes (Pall Life Sciences). After blocking with 5% non-fat dry milk in TBS plus 0.1% Tween 20, the membranes were incubated with the indicated primary antibodies overnight at 4°C. Primary antibodies used were: anti-SQSTM1/p62 (Abcam, ab56416), anti-LC3A/B (Cell Signaling Technology, 12741), anti-GFP (Santa Cruz Biotechnology, sc-9996) and anti-β-actin (Sigma–Aldrich, A5441). Detection was achieved using horseradish peroxidase-conjugated secondary antibodies anti-mouse (Thermo, 31432) and anti-rabbit (The Binding Site, AP311), and visualized with Clarity Western ECL (Bio-Rad Loboratories). Immunoblot quantitative analysis was performed using ImageJ software (NIH).
Proteasome activity assay
SH-SY5Y cells were incubated with normal medium or conditioned medium containing either TTR WT or TTR V30M for 24 h or 48 h. Cells were then harvested in lysis buffer (50 mM HEPES (pH 7.5), 5 mM EDTA, 150 mM NaCl, 1% Triton X-100 and 2 mM ATP) and centrifuged for 15 min at 21500 g at 4°C. The activity of the 20S proteasome, the catalytic core of the proteasome complex, was then measured according to the manufacturer's instructions (Millipore, APT280). Briefly, whole-cell lysate protein extract (35 μg) was incubated in the provided buffer with fluorophore-linked peptide substrate LLVY-7-amino-4-methylcoumarin (LLVY-AMC) for 90 min at 37°C. Proteasome activity was measured by quantification of relative fluorescent units from the release of the fluorescent cleaved product LLVY-AMC by using a fluorescence plate reader (Synergy 2, Biotek) at 360/460 nm and subtracting background levels.
The biological samples were obtained from two previous studies, one with tauroursodeoxycholic acid (TUDCA)  and another with curcumin . The experiment involving TUDCA was conducted in a transgenic mice model bearing the human TTR V30M mutation, in a TTR-null background . In the present study 9-month-old animals were used and received treatment with TUDCA (Calbiochem), at a concentration of 500 mg/kg per day in the drinking water for 3 months. An age-matched control group of animals was maintained under the same conditions, drinking regular tap water. After treatment and at 12 months of age, animals were killed after anaesthesia and organs were harvested.
The study performed with curcumin (Sigma–Aldrich) was carried out using a transgenic mice model where human TTR V30M is expressed in a heat-shock transcription factor 1 (Hsf1) heterozygous background . The study was performed using 14-month-old mice from this same experiment, fed with 2% (w/w) curcumin mixed in standard chow for 6 weeks, whereas an age-matched control group was given plain standard chow.
Sections (3 μm thick) from the gastrointestinal (GI) tract were analysed. Immunohistochemistry was performed using Mouse on Mouse (M.O.M.) Immunodetection Kit (Vector Laboratories) according to the manufacturer's instructions. Antigen unmasking was performed with a sodium citrate solution and a permeabilization step with 0.2% Triton X-100 in PBS was included. The primary antibody used was anti-SQSTM1/p62 (Abcam, ab56416). For colour development we used 3,3′-diaminobenzidine (Sigma–Aldrich, D5637) as substrate. Tissue sections were examined with an Olympus BX50 light microscope equipped with an Olympus DP71 digital camera for image acquisition. Each slide was analysed in at least five different representative areas. Semi-quantitative immunohistochemistry analysis was performed using Image-Pro Plus software (Media Cybernetics).
For fluorescent detection, colon sections from 12-month-old mice were double labelled with anti-p62 and anti-TTR, overnight at 4°C; anti-mouse antibody conjugated to Alexa Fluor 568 and anti-rabbit antibody conjugated to Alexa Fluor 488 (Invitrogen), respectively, were then added for 1 h at room temperature. Slides were then mounted with mounting medium containing DAPI (Vector Laboratories) and visualized under a TCS SP5 II laser-scanning confocal microscope(Leica Microsystems).
Data are presented as the means ± S.E.M. for at least three independent experiments. Multiple comparison one-way ANOVA was used in autophagosome analyses experiments. Ratio-paired Student's t test was used to compare blots. Mann–Whitney statistics were applied in immunohistochemistry semi-quantitative analyses. Whenever P values were less than 0.05 the differences were considered statistically significant and properly marked with asterisks as explained in the figure legends.
TTR V30M leads to late-stage autophagy impairment
We started by evaluating the effect of TTR WT and TTR V30M in the autophagic status of SH-SY5Y cell line. For that we used conditioned media from HEK-293 cells stably producing and secreting TTR V30M. This TTR variant has higher susceptibility to form aggregates in the conditioned media than TTR WT when evaluated by an immune dot-blot filter assay that retains only aggregated material (results not shown). These media, freshly collected and quantified for each assay, was then added to SH-SY5Y cells and then autophagic markers were assessed. First, LC3-II and p62 levels were quantified in the presence or absence of bafilomycin A1, a lysosomal inhibitor affecting vacuolar (V-)ATPases and preventing acidification of endosomes and lysosomes . Experiments were performed at different time points.
p62 recognizes toxic cellular waste and is selectively incorporated into autophagosomes through direct binding to LC3-II which is then degraded by autophagy in the autolysosomes. Although LC3 changes are normally observed at earlier time points, clearance of autophagy substrates such as p62 may require more time as described by Klionsky et al. . Accumulation of p62 correlates with autophagy impairment; we did in fact observe higher levels of p62 in SH-SY5Y cells incubated with TTR V30M for 48 h, in a statistically significant fashion. TTR WT, on the other hand, did not have an impact on p62 accumulation (Figure 1A). Because proteasome impairment may also contribute to p62 accumulation, we proceeded to evaluate the impact of the different conditioned media in proteasome activity. Our results show that it was not affected by either TTR WT or TTR V30M in both time points tested (24 and 48 h) (Figure 1B). Taken together, these results directed the main focus of our investigation on TTR V30M in the experiments that followed.
Impairment of later steps of autophagy upon incubation with TTR V30M
Regarding LC3-II levels, upon incubation with TTR V30M a modest decrease–in the absence of bafilomycin A1–is observed when compared with control media (complete media with vehicle). This would suggest an increased delivery of substrates to lysosomes or decreased autophagosome synthesis (Figure 1C). It could also be attributable to lysosome-dependent degradation similar to that which Tanida et al.  observed in HeLa cells under starvation. The inclusion of bafilomycin A1 also led to no significant differences between tested conditions and this tendency was maintained for both shorter and longer incubations (results not shown). As suggested by Gamerdinger et al. , subtracting LC3-II levels in the absence of bafilomycin to LC3-II levels in the presence of this compound roughly translates to the amount of LC3-II that has been degraded (Figure 1D). However, together with the previous data obtained from p62 protein levels, this hinted to the hypothesis of TTR acting as a blocker of autophagy at a later stage rather than enhancing the synthesis and degradation of autophagosomes.
TTR V30M has no effect on autophagosome number
To visually detect possible differences in autophagosome formation/number we transiently transfected SH-SY5Y cells with EGFP–LC3. This chimaera displays a green fluorescence that changes from a diffuse to a punctate staining when autophagy is induced (or blocked before the degradation step). Regarding autophagosome number, there were no differences between controls and TTR V30M incubated cells either with or without bafilomycin A1 (Figures 2A and 2B) which supports the above results from immunoblotting using cell lysates; thus the hypothesis of impaired autophagy initiation is definitely discarded. This assay also offers the possibility of measuring autophagosome size. No significant differences in autophagosome size were registered between each pair of samples.
TTR V30M does not affect autophagosome formation but causes a delay in the degradation step
TTR V30M causes a delay in free GFP degradation
We assessed free GFP protein levels which, depending on the cell type, may only be detectable under non-saturating doses of lysosomal inhibitors or under conditions that attenuate lysosome acidity; otherwise the degradative mechanism is too efficient to allow GFP accumulation .
In our cell culture set-up, a GFP band was detected with a stronger signal in cells incubated with TTR V30M (Figure 2C). This event could have two possible explanations: (i) either the flux is more active and more GFP–LC3 is being cleaved producing more free GFP fragments, or (ii) the degradative mechanism is partially impaired (as observed when using non-saturating doses of inhibitors) meaning that the autophagic machinery is active enough to cleave GFP–LC3-II but less effective as noticed by the delay in GFP degradation. Taking into consideration that in the presence of TTR V30M there is no augmented production of autophagosomes as seen in the above fluorescent assay, this higher level of free GFP cannot be explained by a more active flux but rather by partial impairment at a later stage of the degradative mechanism.
In V30M transgenic mice p62 accumulation is associated with higher V30M TTR levels in intestinal murine cells
As in cell culture TTR V30M displays a direct effect in p62 accumulation, we used confocal microscopy to search whether a correlation also in terms of localization existed in tissues from V30M transgenic mice. Our studies were conducted in a transgenic mice model bearing the human TTR V30M mutation, in a TTR-null background. The model displays non-fibrillar TTR deposition in the GI tract at as early as 3 months of age. We proceeded by immunolabelling colon samples with anti-TTR and anti-p62. Besides TTR extracellular deposition, confocal images clearly showed intracellular colocalization between both markers in the intestinal villi (Figure 3); so cells that have a higher TTR burden also have massive amounts of accumulated p62 which corroborates the previous results obtained in our cell culture assays regarding the link between TTR and late-stage autophagy impairment.
TTR V30M co-localizes with p62 in colon of 12-month-old TTR V30M transgenic mice
TUDCA effectively reverses p62 accumulation in the V30M transgenic GI tract
Next we proceeded to compare p62 immunostaining patern in the same animal model (V30M mice) where a group was subjected to a TUDCA treatment for 3 months having a non-treated control group for comparison. TUDCA is an hydrophilic bile acid proven to have anti-apoptotic and antioxidant properties. In fact, previous studies from our laboratory focusing on the GI tract of the same strain of mice showed that this compound has the capability of reducing apoptotic and oxidative stress biomarkers such as immunoglobulin heavy-chain-binding protein (BiP), eukaryotic initiation factor 2α (eIF2α), Fas and 3-nitrotyrosine, usually associated with TTR deposition. Of higher relevance, it was able to significantly reduce levels of TTR deposition . We were interested in investigating whether the improvements previously observed in TUDCA-treated V30M mice also involve any effect on autophagy. As seen in Figure 4, p62 levels dramatically decreased to almost undetectable values in colon samples of the treated group. Other organs from the GI tract were analysed, namely oesophagus, stomach and duodenum, and although the same tendency was observed, the differences between the treatment groups and controls were not statistically significant (results not shown). These results indicate that, besides other cellular mechanisms, TUDCA is able to modulate autophagy; ultimately it enhances/re-establishes the autophagic flux which in turn may be direct or indirectly involved in the pathological cascade of FAP, thus contributing to the clearance of TTR V30M aggregates.
p62 accumulation is decreased in colon samples of animals treated with TUDCA
Curcumin effectively reverses p62 accumulation in the V30M transgenic GI tract
We used another V30M transgenic mouse strain with earlier and aggravated intestinal non-fibrillar deposition treated with curcumin, a well-known drug that promotes autophagy , to investigate the effect of treatment on p62 accumulation. In earlier studies, 7-month-old mice were fed with 2% (w/w) curcumin mixed in standard chow for 6 weeks . This treatment led to a significant reduction in TTR deposition in the GI tract as did curcumin treatment starting at 14 months of age and terminating at 15.5 months . We analysed the latter tissues in terms of expression of the autophagic marker p62. The stomach and duodenum samples of the curcumin-treated group hardly presented any p62 staining in comparison with the controls (Figure 5). This indicates that, similarly to what happened in the TUDCA study, curcumin is able to directly or indirectly modulate autophagy in the transgenic TTR mouse model.
p62 accumulation is dramatically decreased in stomach and duodenum samples of animals treated with curcumin
Defects in the autophagic machinery have been linked to neurodegenerative conditions such as Huntington's, Parkinson's and Alzheimer's diseases that are characterized by an aberrant accumulation of endogenous proteins. Defects in autophagy in FAP have hardly been described.
In the present study, we showed a causal effect of TTR V30M aggregates in autophagy in a cell model and the modulating ability of this process in vivo by two different compounds (TUDCA and curcumin).
Our experiments were focused on two of the most well characterized autophagic markers, p62 and LC3. Special caution was taken when interpreting LC3-II level changes as similar results can be consequence of two opposite steps, i.e. formation and degradation of autophagosomes. Very recently, Rubinzstein and co-workers designed a flowchart in order to assist autophagy researchers in understanding the observed changes in LC3-II protein levels in Western blots upon use of a compound or inhibitor  which we followed in our approach.
We observed a modest decrease in LC3-II when cells were exposed to TTR V30M aggregates for 20 h. This outcome could be due to an initial blockage in autophagosome formation (hence the lower LC3-II levels) or, alternatively, TTR V30M could be promoting higher in autophagosome degradation. However, upon inclusion of lysosomal inhibitors that generally helps in distinguishing these two scenarios, no significant differences were observed in the presence/absence of inhibitor.
We also followed the proposal of Gamerdinger et al.  of an alternative approach to infer modifications in autophagic flux. Overall we were unable to obtain statistically significant differences that allow us to make conclusions. Furthermore, autophagosome number and accumulation evaluated using visual detection of EGFP–LC3 revealed that the number of counted autophagosomes either in control condition or in the presence of TTR V30M were remarkably similar independently of the presence of lysosomal inhibitors. This translates into the assumption that the number of autophagosomes being generated and ready for delivery in the lysosome is quite similar, i.e. TTR V30M is not affecting the autophagosome generation rate in any noticeable way.
The GFP–LC3 data demonstrate that the presence of TTR V30M leads to higher amounts of free GFP indicating that the degradative machinery becomes partially impaired and less effective in dismantling GFP when compared with control conditions. Overall, with these latter in vitro studies we demonstrate that TTR aggregates, even at small amounts, are able to disrupt the autophagic machinery. Autophagosome–lysosome fusion does not seem to be defective because generally, when blocked, larger autophagosomes are generated possibly due to autophagosome–autophagosome fusion which was not observed in our experiments.
In contrast with LC3-II, interpretation of changes in p62 detection levels are usually more straightforward. The amount of p62 generally increases when autophagy is inhibited and decreases when autophagy is induced . Such an increase was observed in our cell culture upon incubation with TTR V30M, supporting a decrease in autophagosomal clearance with statistical significance. The impact of TTR WT in p62 was also evaluated but no significant changes were observed when compared with the control group. Accumulation of p62 can also occur as result of proteasome impairment. For this reason we evaluated the impact of TTR WT and TTR V30M on proteasome activity, using complete media as control. Our data demonstrated that the proteasome activity remained unaffected by both TTR WT and TTR V30M conditioned media, for as long as 48 h of incubation. This discounted the possible involvement of the proteasome in p62 accumulation. Using the human TTR V30M transgenic mouse model we proposed to investigate the status of the autophagy substrate p62 in colon sections. The double staining co-localization observed in murine colon cells was remarkably specific. p62 appeared accumulated within cells that also have TTR. Although we did not proceed to identifying these cells, previous results from Gonçalves et al.  point out that enteric glial cells are able to internalize TTR in vivo in a FAP mouse model, suggesting that oligomer uptake could impair the autophagic machinery leading to p62 accumulation ultimately to cell death.
The majority of FAP patients have early and severe autonomic nervous system dysfunction, involving the GI tract with infiltration of aggregated material in the interspace of two adjacent ganglia, leading to neuronal loss. A study conducted by Wixner et al. , however, showed no significant difference in the number of enteric neurons in myenteric ganglia of gastric tissue between controls and TTR V30M patients. The most remarkable feature observed in this research was the significantly lower abundance of gastric interstitial cells of Cajal (ICC) in patients when compared with non-amyloidosis controls. The mechanisms behind the depletion of ICC are still not clear.
So far, there is no curative treatment for FAP patients. Liver transplantation performed on early-onset FAP patients is the only procedure able to stabilize this neuropathy in the long term, although it presents some limitations and disadvantages [25,26]. Despite removing the main source of mutated TTR, liver is not the only structure where this protein is being produced.
Cytotoxicity described in FAP is related to TTR V30M aggregation and tissue deposition. This is closely related to increased levels of oxidative metabolites that trigger oxidative stress, inflammation, ER stress and activation of apoptosis; impairment of the autophagic flux certainly would contribute to cell toxicity.
TUDCA, a naturally synthesized biliary acid affecting, among others, endoplasmic reticulum stress and oxidative stress, demonstrated promising results in human TTR V30M mice not only by reducing apoptotic and oxidative stress biomarkers that show up-regulation in animal models for FAP, but also by lowering TTR deposits in the GI tract . Although already being used in clinical trials for different diseases , the mechanism of action of this compound is not completely understood. With this work, we showed that TUDCA is able to modulate autophagy in human TTR V30M mice by reversing the p62 accumulation observed in the control group.
Similar results were observed upon curcumin treatment. Curcumin has been broadly used by researchers and demonstrated to have a protective effect in many diseases involving antimicrobial, antitubercular  and anticancer  mechanisms with the additional ability of modulating innate immunity . It has also been reported as an autophagy inducer . Work from our laboratory by Ferreira et al.  demonstrated the antioxidative and anti-inflammatory as well as anti-apoptotic properties of curcumin in human TTR V30M mice while lowering TTR deposition. In this previous work, however, autophagic pathway was not addressed. Results obtained from our research clearly shows that autophagy is also a target of the multifunctional activity of curcumin contributing massively to the reduction of p62 accumulation.
The effects of both compounds are very promising regarding FAP therapeutics research. However, the collected data still raises some questions in terms of discrimination of cause and effect in autophagy. It is not completely clear whether: (i) the autophagic flux is restored because these compounds are directly involved in the removal of TTR V30M aggregates; (ii) these compounds act directly in other pathways that play a major role in aggregate removal, with an effect in signalling cascades involving oxidative stress and inflammation that in turn contribute to re-establishment of autophagy; or (iii) these treatments directly increase autophagic activity and subsequently the re-establishment of the autophagic flux helping, even if only partially, in removing the aggregates from the cells and contributing to the reduction of oxidative stress and inflammation.
Other therapeutic strategies for the treatment of FAP are currently being explored, with some of them already on clinical trials, but a curative therapy is yet to be developed; they comprise molecule stabilizers (that prevent TTR tetramer dissociation necessary for fibril formation), namely diflunisal  and tafamidis [33,34]; gene therapy using antisense oligonucleotides  and siRNAs; combination of compounds such as doxycy-cline/TUDCA is underway [36,37] with preliminary promising results.
As stated before, the relationship between autophagy and FAP had not been established, although defects in autophagy have been linked to many neurodegenerative diseases such as Huntington's, Parkinson's and Alzheimer's diseases that are characterized by an aberrant accumulation of endogenous proteins. Improvement of autophagy has been successful in promoting clearance of toxic proteins thus reducing the symptoms of neurodegeneration in cell and animal models [38–42]. The data obtained in our study in FAP models point in the same direction and encourage the search for new therapeutic alternatives for the treatment of FAP either through the design of new/more specific drugs or through the combination of current treatments with specific autophagic inducers.
Cristina Teixeira was responsible for the acquisition, analysis and interpretation of data and manuscript writing. Maria João Saraiva wrote and reviewed the manuscript for important intellectual content. Maria do Rosário Almeida provided material from treated animals and revised the manuscript.
We thank Dr Luis Almeida from Coimbra University for helpful suggestions, Dr Paula Sampaio from I3S for help with confocal microscopy and Paula Gonçalves for tissue processing.
This work was supported by Fondo Europeo de Desarrollo Regional (FEDER) funds through the Operational Competitiveness Programme–COMPETE [grant number FCOMP-01-0124-FEDER-022718 (PEST-c/SAU/LA0002/2011) FCT-FEDER for unit 4293 in partnership with PT2020]; and the Portuguese Foundation for Science and Technology (FCT) [grant number PTDC/BIM-AMEC/0282/2012].