Transthyretin stability is critical in assisting beta amyloid clearance– Relevance of transthyretin stabilization in Alzheimer's disease

Summary

Background

The absence of transthyretin (TTR) in AD mice decreases brain Aβ clearance and reduces the low‐density lipoprotein receptor‐related protein 1 (LRP1). It is possible that neuroprotection by TTR is dependent on its tetramer structural stability, as studies using TTR mutants showed that unstable L55P TTR has low affinity for Aβ, and TTR tetrameric stabilizers such as iododiflunisal ameliorate AD features in vivo.

Methods

We firstly investigated TTR folding status in human plasma measuring the resistance to urea denaturation. The importance of TTR stability on Aβ internalization was studied in human cerebral microvascular endothelial (hCMEC/D3) and hepatoma cells (HepG2), by flow cytometry. To investigate the fate of Aβ at the blood–brain barrier, Aβ efflux from hCMEC/D3 cells seeded on transwells was measured using ELISA. Further, to assess Aβ colocalization with lysosomes, Lysotracker was used. Moreover, levels of LRP1 were assessed in the liver and plasma of mice with different TTR backgrounds or treated with iododiflunisal.

Results

We showed that TTR stability is decreased in AD and that WT TTR and drug‐stabilized L55P TTR are able to increase uptake of Aβ. Furthermore, measurement of Aβ efflux showed that stable or stabilized TTR increased Aβ efflux from the basolateral to the apical side. Moreover, HepG2 cells incubated with Aβ in the presence of WT TTR, but not L55P TTR, showed an increased number of lysosomes. Further, in the presence of WT TTR, Aβ peptide colocalized with lysosomes, indicating that only stable TTR assists Aβ internalization, leading to its degradation. Finally, we demonstrated that only stable TTR can increase LRP1 levels.

Conclusion

TTR stabilization exerts a positive effect on Aβ clearance and LRP1 levels, suggesting that TTR protective role in AD is dependent on its stability. These results provide relevant information for the design of TTR‐based therapeutic strategies for AD.

1. INTRODUCTION

One of the pathological hallmarks of Alzheimer's disease (AD), the most common form of dementia and fifth‐leading cause of death for people aged 65 and older, is the extra‐neuronal plaques consisting of beta amyloid peptide (Aβ) aggregates.1 There is growing evidence supporting the hypothesis that AD starts to develop due to an imbalance between Aβ production and clearance, which may result from a dysfunction in the clearance system and/or due to the formation of hard‐to‐clear Aβ aggregates.2 Particularly, while patients with early‐onset AD and presenilin mutations show both increased production and decreased clearance of Aβ,3 patients with late‐onset AD, comprising the majority of AD cases, only demonstrate decreased Aβ clearance.4

Aβ in the plasma is bound to a diverse number of proteins such as albumin, apoE, apoJ, transthyretin (TTR), and a soluble form of the low‐density lipoprotein receptor‐related protein 1 (sLRP1), allowing the maintenance of a constant concentration of free Aβ and inhibiting its entering back to the brain. It has been shown that in AD, besides a decrease in the expression of LRP1, the main receptor for Aβ clearance at the blood–brain barrier (BBB) and at the liver, sLRP1 levels and its affinity for Aβ are reduced.5 Moreover, in AD, serum levels of albumin–Aβ complexes,6 apolipoproteins,7 and also transthyretin 8 are shown to be decreased.

TTR, a 55‐kDa homotetrameric protein, is mainly produced and secreted by the liver and choroid plexus (CP) to the plasma and CSF, respectively. TTR physiological role as a carrier for thyroid hormones and vitamin A 9 is known for decades, as well as the involvement of its mutated forms in familial amyloid polyneuropathy (FAP).10 More recently, several studies implicated TTR in neuroprotection, including in the cases of ischemia,11 regeneration,12 memory,13 and AD.

The earliest study describing a protective role for TTR in AD dates back to 1994 when TTR was reported to be the major Aβ‐binding protein in CSF.14 Later on, several studies strengthened this hypothesis, including a study that showed amelioration of AD features in APP transgenic mice overexpressing human TTR,15 and another report that demonstrated accelerated amyloid deposition in mice with ablation of the TTR gene.16 A recent study also showed TTR attenuation in AD transgenic mouse model with CP dysfunction,17 and finally, our previous study demonstrated TTR participation in Aβ transport from the brain to the liver.18 In that study, we showed that TTR promotes Aβ internalization and efflux in a human cerebral microvascular endothelial cell line, hCMEC/D3, used as an in vitro BBB model. TTR also stimulated brain‐to‐blood Aβ permeability in these cells, suggesting that TTR interacts directly with Aβ at the BBB. We then showed that TTR crosses the monolayer of cells only in the brain‐to‐blood direction, as confirmed by in vivo studies, proposing that TTR can transport Aβ from, but not into, the brain. Moreover, TTR increased Aβ internalization by SAHep cells (human hepatoma cells) and by primary hepatocytes derived from TTR+/+ mice when compared to TTR−/− animals. Interestingly, TTR also regulates LRP1 receptor levels, as lower receptor expression was found in brains and livers of TTR−/− mice and in cells incubated without TTR.18

There are several studies reporting that TTR levels are diminished both in the CSF 19 and in the plasma 8 of patients with AD, found to occur early in the disease development.20 However, the cause of TTR reduction in AD is not known yet, but stability of its tetramer structure seems to play a significant role, leading to the stability hypothesis which postulates that an unstable TTR is cleared faster, thus explaining the lower levels. In fact, it has also been shown that plasma TTR from patients with AD presents impaired ability to carry T4 hormone. As T4 binding to TTR implies tetrameric conformation, this indicates TTR instability and supports the idea that the TTR tetramer is the main Aβ‐binding species.20 Furthermore, studies using TTR mutants showed that the lower the stability of TTR variant (and the higher its amyloidogenic potential), the weaker its affinity toward Aβ42; thus, the highly unstable TTR with a leucine substituted by a proline at position 55 (L55P TTR) was shown to bind weakly to Aβ.21

Although some screening studies in patients with AD demonstrated no significant correlation between TTR variants and AD,22 a very recent report showed a tendency of enrichment in six TTR rare variants in patients with AD in Han Chinese population.23 Also, in another recent gene‐based analysis, some TTR variants were reported to be associated with AD.24 Regardless of the possible genetic relationship between AD and TTR variants, other factors, including conformational changes resulting from aging, metal ions concentration, or interaction with other proteins, may also influence TTR stability, levels, and binding properties toward Aβ.

Most importantly, small compounds known to stabilize TTR tetrameric fold have been demonstrated to improve TTR–Aβ interaction, in vitro.25 One of such stabilizers, iododiflunisal (IDIF), when administrated to AD transgenic mice ameliorates AD features, such as cognitive function, brain Aβ deposition, and plasma Aβ levels.26 Moreover, oral administration of resveratrol, another TTR stabilizer present mainly in the grape skin, reduced significantly brain Aβ levels and amyloid deposits, increased LRP1 protein expression in both brain and liver of the AD mice, and increased TTR plasma levels.27

Therefore, this study aimed to unravel the importance of TTR stability in Aβ clearance and LRP1 expression levels at the BBB and the liver, both by comparing the effects of human recombinant wild‐type TTR (WT TTR) versus L55P TTR and by studying the impact of stabilizing L55P TTR using IDIF or resveratrol in this process.

2. MATERIALS AND METHODS

2.1. Recombinant TTR production, purification, and labeling

WT and L55P TTR were produced in a bacterial expression system using Escherichia coli BL21 and purified as previously described.28 Briefly, after growing the bacteria, the protein was isolated and purified by preparative gel electrophoresis after ion‐exchange chromatography. Protein concentration was determined by the Bradford method (Bio‐Rad Hercules, CA, USA), using bovine serum albumin as standard.

When appropriate, WT TTR and L55P TTR were labeled with Alexa Fluor 633, using the Alexa Fluor^® 633 Protein Labeling Kit (Molecular probes, alfagene, Carcavelos, Portugal), following the manufacturer's instructions.

2.2. Animals

In this work, TTR‐wild‐type mice carrying both copies of the TTR gene (TTR+/+), TTR‐heterozygous mice (TTR+/−), and TTR‐knockout mice (TTR−/−) in a SV129 background 29 were used. These mice were obtained from the littermate offspring of heterozygous breeding pairs.

The AD mouse model used in this study was generated by crossing APPswe/PS1A246E transgenic mice purchased from The Jackson Laboratory with TTR−/− mice, as previously described.16 In this project, we used cohorts of littermates APPswe/PS1A246E/TTR+/+ (AD/TTR+/+) and APPswe/PS1A246E/TTR+/− (AD/TTR+/−) and APPswe/PS1A246E/TTR−/− (AD/TTR−/−).

Animals were housed in a controlled environment (12‐hours light/dark cycle; temperature 22± 2°C; humidity 45%‐65%), with freely available food and water. All efforts were made to minimize pain and distress. All procedures involving animals were carried out in accordance with National and European Union Guidelines for the care and handling of laboratory animals and were performed in compliance with the institutional guidelines and recommendations of the Federation for Laboratory Animal Science Association (FELASA) and were approved by the National Authority for Animal Health (DGAV; Lisbon, Portugal).

2.3. Blood collection from mice and human subjects

Mice were profoundly anesthetized with an anesthetic combination of ketamine (75 mg/Kg) and medetomidine (1 mg/Kg) by intraperitoneal injection. Blood was collected from the inferior vena cava with syringes with EDTA, followed by centrifugation at 1000 rpm for 15 minutes at room temperature (RT). Supernatant was collected and frozen at −80°C until used.

Patients with AD (study group) were recruited at the Dementia Clinic, Neurology Department, Centro Hospitalar e Universitário de Coimbra, and blood was obtained by peripheral venipuncture. This study was approved by the ethics board, and all subjects or responsible caregivers, whichever appropriate, gave their informed consent. The control healthy group matched for age and gender with the study group included subjects without neurological or psychiatric history and no cognitive impairment and was recruited at LabMed Center. All subjects were informed on the purpose of the study and gave their written consent. These samples were already available in the laboratory and were used in a previously reported work.26

2.4. Evaluating tetramer stability of human plasma TTR

To evaluate TTR tetramer stability, an assay employing urea‐mediated denaturation was used as previously described in the literature 30, 31 with minor alterations. In this assay, the extent of denaturation of plasma TTR of control (n=10) and AD (n=10) human subjects was assessed by evaluating levels of folded TTR (tetramer, trimer, and dimer) versus the levels of monomer. Briefly, 3 μL of plasmas was incubated overnight with urea (6 mol L⁻¹), then cross‐linked by 25% glutaraldehyde (4 minutes) and quenched by the addition of NaBH₄ (7% in 0.1 mol L⁻¹ NaOH). SDS reducing gel loading buffer (SDS final concentration = 2%) was then added and samples were boiled for 5 minutes. Proteins were then separated using 13.5% SDS‐polyacrylamide gel electrophoresis (PAGE) and transferred to nitrocellulose membrane (Whatman™ GE Healthcare, Chalfont St. Giles, UK, Protan BA 83) using a wet system (Bio‐rad Criterion Blotter). The membrane was cut to separate the monomeric from folded TTR as the latter always comprised the larger part of the sample, and the signal was much stronger than for monomeric TTR, needing different exposure times. The membranes were then blocked for 1 hour at room temperature with 5% powdered skimmed milk in PBS containing 0.05% Tween‐20 (PBS‐T). Immunoblotting was then performed using anti‐TTR antibody (DAKO, Glostrup, Denmark, 1:2000) to detect folded and monomer TTR. The blots were developed using Clarity^™ Western ECL substrate (Bio‐Rad), and proteins were detected and visualized using a chemiluminescence detection system (ChemiDoc, Bio‐Rad). Images were analyzed and band intensities were quantified using Image Lab (Bio‐Rad, version 4.1).

2.5. Cell lines and primary hepatocytes culture

The immortalized human cerebral microvascular endothelial cell line (hCMEC/D3 cells) was kindly provided by Dr. P.O. Couraud (Institut Cochin, INSERM 1016, CNRS 8104, Université Paris Descartes, France). The hCMEC/D3 cell line, initially produced and characterized by Dr. P.O. Couraud and colleagues, was used as a model of human BBB as previously validated.32, 33 Cells were grown in 25‐cm² flasks or 12‐well plates precoated with rat tail collagen type I solution (150 μg/mL, Corning Life Sciences, Lowell, CA, USA) and cultured in EBM‐2 medium (Lonza, Walkersville, MD, USA) supplemented with 5% fetal bovine serum (FBS) (Gibco^™, Carlsbad, CA, USA.), 1% penicillin–streptomycin (Lonza), 1.4 μ mol L⁻¹ of hydrocortisone (Sigma‐Aldrich, Sintra, Portugal), 5 μg/mL of ascorbic acid (Sigma‐Aldrich), 1% chemically defined lipid concentrate (Lonza, Verviers, Belgium), 10 m mol L⁻¹ of 4‐(2‐hydroxyethyl)‐1‐piperazineethanesulfonic acid (HEPES) (Gibco^™), and 1 ng/mL of human basic fibroblast growth factor (bFGF) (Sigma‐Aldrich).

HepG2 cells, a human hepatoma line, were grown in Dulbecco's modified Eagle medium (DMEM) (Lonza) and Ham's F‐12 media (Lonza, Verviers, Belgium), supplemented with 10% FBS (Gibco^™), 100 U/mL penicillin–streptomycin (Gibco^™), 2 m mol L⁻¹ L‐glutamine (Gibco^™), and 1× nonessential amino acid solution (Lonza).

Primary hepatocytes were isolated from livers of TTR+/−, TTR−/−, or AD/TTR +/− mice with 3 months of age. To obtain these cells, two‐step collagenase perfusion of liver was performed as previously described in detail.18 Briefly, a cannula was inserted into the portal vein and perfusion medium and then collagenase solution were allowed to perfuse through the liver. The entire perfused liver was then removed to a Petri dish containing isolation medium proceeding to the next steps of filtering and centrifugations of cell suspension, counting live cells, and seeding with attachment medium for 3 hours. Then, the medium was changed to stimulation medium. After 24 hours, the medium was renewed and after 48 hours, the experiments were performed as will be described later.

2.6. Flow cytometry analysis of Aβ and TTR cellular internalization

To study Aβ and TTR uptake, hCMEC/D3 or HepG2 cells were grown to 80% confluence in 24‐well plates. Cells were incubated with 500 ng/mL (0.11 μ mol L⁻¹) of Aβ 1‐42 labeled with 5‐(and‐6)‐carboxyfluorescein (FAM‐Aβ, AnaSpec, Seraing, Belgium) in the presence of different concentrations (0.2‐2 μ mol L⁻¹) of WT or L55P TTR labeled with Alexa Fluor 633 (WT TTR‐633 or L55P TTR‐633, respectively), for 15 minutes (for hCMEC/D3 cells) or 1 hour (for HepG2 cells). To observe the effects of TTR stabilization in Aβ internalization, L55P TTR (0.2 μ mol L⁻¹) was preincubated for 1 hour at 37°C with IDIF (2 μ mol L⁻¹) or resveratrol (2 μ mol L⁻¹), before being added to the cells together with FAM‐Aβ, as described above.

We also examined the effect of IDIF and resveratrol on FAM‐Aβ uptake in primary hepatocytes derived from TTR+/− or TTR−/−. After 48 hours of seeding, cells were incubated with FAM‐Aβ (500 ng/mL) in the absence or presence of IDIF (2 μ mol L⁻¹), resveratrol (2 μ mol L⁻¹), or WT TTR (2 μ mol L⁻¹) for 2 hours.

After all incubations, cells were enzymatically detached using trypsin (Gibco^™), and then, cells resuspended in media containing FBS were centrifuged at 1 000 rpm for 5 minutes. The pellet was washed with PBS, centrifuged again, resuspended in PBS, and then fixed (final 1.5% paraformaldehyde). Finally, cells were analyzed in FACS Canto II equipment (BD Biosciences) using Blue laser excitation of 488 nm or red laser excitation of 640 nm. The flow cytometry data were analyzed using the Flowjo software, version 10.0.8.

2.7. hCMEC/D3 cells permeability to Aβ and TTR

For BBB transport experiments, hCMEC/D3 cells were seeded on type I collagen‐precoated transwells filters (polyester 12 wells, pore size 0.4 μm; Costar, Washington, DC, USA), and assays were performed 18 days after seeding. Before transport studies, and to identify filters in which the integrity of the cell monolayer was not complete, apical compartment (blood side) media were supplemented with 0.25 mg/mL fluorescein isothiocyanate (FITC)‐labeled dextran (molecular mass of 70 kDa; Sigma‐Aldrich). The concentration of FITC‐labeled dextran was determined using fluorometer (excitation: 492 nm; emission: 518 nm) in the collected media from basolateral compartment (brain side) after 1 hour. FITC‐labeled dextran concentration exceeding 125 ng/mL indicated the wells with disrupted cell layer which were excluded from Aβ transport experiments.

Transport studies were performed by supplementing the brain side with 500 pg/mL (0.11 n mol L⁻¹) Aβ1‐42 (Genscript^®, Piscataway, NJ, USA), in the absence or presence of WT TTR‐633 or L55P TTR‐633 preincubated with or without IDIF. Then, cells were incubated at 37°C, and after 24 hours, samples were collected from both brain and blood sides and Aβ 1‐42 levels in both compartments were quantified using ELISA kit (Invitrogen^™, Alfagene, Carcavelos, Portugal) following the manufacturer's recommendations. WT or L55P TTR‐633 levels in the same samples were also evaluated by fluorometer (excitation: 632/580 nm; emission: 647 nm). Furthermore, TTR and Aβ inside the same cells were also analyzed using microscopy as will be described.

2.8. Immunofluorescence for Aβ, TTR, or LRP1

hCMEC/D3, HepG2 cells, or primary hepatocytes (AD/TTR +/−) were grown on glass coverslips (Thermo Scientific, Boston, MA, USA) precoated with rat tail collagen type I solution.

To observe internalization of Aβ and TTR, HepG2 cells were incubated with FAM‐Aβ and/or WT or L55P TTR‐633 for 1 hour, washed with PBS, and fixed with acetone for 7 minutes in −20°C. After wash, coverslips were then mounted with Fluoroshield^™ with DAPI (Sigma‐Aldrich).

To assess the effect of TTR stability on LRP1 protein levels, hCMEC/D3 cells or primary hepatocytes were incubated with WT or L55P TTR for 1 hour, fixed in 4% PFA at 4°C, then permeabilized with 0.1% Triton X‐100, and blocked with 5% BSA at room temperature followed by incubation with primary antibodies against LRP1 (Rabbit anti LRP1 antibody, Abcam, Cambridge, UK 1:100). After washing, cells were incubated with Alexa Fluor 488 or 568 goat anti‐rabbit IgG antibodies (Life Technologies, Carlsbad, CA, USA, 1:1000), washed, and then mounted with DAPI.

To investigate the effect of IDIF treatment on brain LRP1 protein levels, 30‐μm‐thick coronal brain sections of mice were washed with PBS and dried on APES‐precoated slides, permeabilized with 0.25% Triton X‐100, blocked with 1% BSA, and incubated with LRP1 antibody overnight. After incubation with Alexa Fluor 488 IgG antibody, sudan black B solution (0.3%) was applied to remove tissue autofluorescence and washed well before mounting with DAPI.

All described samples were visualized and photographed by Zeiss Axio Imager Z1 equipped with an Axiocam MR3.0 camera and Axiovision 4.7 software.

To observe TTR and Aβ in cells cultured on transwell filters, after 24 hours of incubating cells as described before, filters were cut and separated from inserts carefully. Cells were then fixed using 4% PFA at 4°C, blocked with 5% BSA followed by incubation with primary antibody against Aβ (BAM‐10, Sigma, 1:200) and then Alexa Fluor 488 anti‐mouse IgG antibody (Invitrogen, 1:1000). Coverslips were then mounted and further visualized and photographed by confocal microscope (Leica SP5 AOBS SE, Wetzlar, Germany).

2.9. Study of lysosomal activity using Lysotracker deep red

To study the lysosomal activity and the fate of Aβ, HepG2 cells were seeded in 24‐well plates or coverslips. In the presence of Lysotracker deep red (Life Technologies, 1:1000) in all conditions, cells were incubated with WT TTR (2 μ mol L⁻¹) or L55P TTR (2 μ mol L⁻¹), and/or FAM‐Aβ (500 ng/mL), for 2 hours (for flow cytometry) or 18 hours (for microscopy) and fixed with 4% PFA. Then, Lysotracker inside cells was assessed using flow cytometry analysis, whereas localization of FAM‐Aβ and lysosomes was analyzed by fluorescent microscopy (Zeiss Axio Imager Z1, Oberkochen, Germany).

2.10. Protein extraction

Livers from TTR +/+ mice, AD mice with different TTR genetic backgrounds, or AD/TTR+/+ mice treated with IDIF, all with 6‐7 months of age, were subjected to protein extraction. Briefly, livers were homogenized in lysis buffer (20 m mol L⁻¹ MOPS pH 7.0; 2 m mol L⁻¹ EGTA; 5 m mol L⁻¹ EDTA; 30 m mol L⁻¹ sodium fluoride; 60 m mol L⁻¹ β ‐glycerophosphate pH 7.2; 20 m mol L⁻¹ sodium pyrophosphate; 1 m mol L⁻¹ sodium orthovanadate; 1% Triton X‐100), 1  m mol L⁻¹ phenylmethylsulfonyl fluoride (PMSF), and protease inhibitors (GE healthcare), followed by incubation on ice for 20 minutes and centrifugation at 18700 g at 4°C for 20 minutes. Resulting supernatants were used for protein analysis. Total protein concentration was quantified by the Bradford assay (Bio‐Rad), using BSA as standard.

2.11. Western blot analysis for LRP1

The expression of LRP1 protein levels was investigated by western blot using liver protein extracts (50 μg) of AD/TTR +/+ (n=4), IDIF‐treated AD/TTR+/+ (n=4), and TTR+/+ mice (n=2), aged 6‐7 months.

Briefly, proteins were separated either in 10% SDS‐PAGE or in 7% tris–acetate PAGE and then transferred to nitrocellulose membrane, and immunoblotting was performed using rabbit anti‐LRP1 (Abcam, 1:15000) or mouse anti‐β‐actin (Sigma‐Aldrich, 1:3000).

2.12. Determination of sLRP1 levels

sLRP1 levels in plasma from control and AD mice with different TTR backgrounds, untreated or treated with IDIF, were quantified using sLRP1 ELISA Kit (MyBioSource, San Diego, CA, USA), according to the manufacturer's instructions.

2.13. Statistical analysis

All quantitative data were expressed as mean ± SD and, in case of n ≥ 5, data were expressed as mean ± SEM. Initially, data were assessed whether it followed a Gaussian distribution. When found to follow a Gaussian distribution, differences among conditions or groups were analyzed by one‐way ANOVA with a Sidak's multiple comparisons test. In the cases of non‐Gaussian distribution, differences among conditions were analyzed by nonparametric Kruskal‐Wallis test and comparisons between two groups were made by Student's t test or by Mann‐Whitney test. P‐values lower than .05 were considered statistically significant (for instance: *, **, and *** in the figures denote P<.05, P<.01, and P<.001, respectively). Statistical analyses were carried out using GraphPad Prism 5 software for Windows.

3. RESULTS

3.1. TTR tetrameric structure is less stable in plasma of AD human subjects

As previously described, TTR levels are decreased in AD, not only in the CSF but also in plasma.8, 20 To assess whether the decrease in TTR concentration in plasmas from patients with AD was due to decreased tetrameric stability, we measured TTR resistance to urea denaturation, as previously described for mice treated with a TTR tetrameric stabilizer.30 The extent of denaturation of the TTR tetramer was evaluated by measuring how much tetramer or folded TTR (tetramer, trimer, and dimer) and monomer remained after urea treatment.

As it can be seen in Figure 1A, total levels of TTR are lower in AD human plasma compared to the age‐matched controls, as also previously described (207.1±8.4 vs 237.3±7.5 mg/L, respectively 20). Here, we showed that TTR stability was also decreased in AD samples compared to controls (Figure 1A, B). In other words, significantly higher levels of folded TTR and lower levels of monomer were found in controls, as compared to AD human subjects, suggesting that TTR stability is compromised in AD, followed by decreased levels of TTR in these patients.

Figure 1.

Plasma transthyretin (TTR) from Alzheimer's disease (AD) subjects is less stable than TTR from age‐matched controls. (A) Representative western blot analysis of TTR stability in plasmas from control (n=10) versus AD (n=10) human subjects; total TTR levels were decreased in AD samples, as previously reported. (B) AD human samples presented lower folded/monomer TTR ratios, compared to controls. Data are expressed as mean±SEM

3.2. Aβ internalization is increased in the presence of stable WT TTR, but not of L55P TTR

Previous work by our group suggested that the mechanism underlying TTR protection in AD involves TTR participation in Aβ efflux from the brain and uptake at the liver.18 To investigate the effects of TTR stability on Aβ internalization and clearance, we incubated both hCMEC/D3 and HepG2 cells with WT TTR‐633 or L55P TTR‐633 and/or FAM‐Aβ, and analyzed the percentage of fluorescent cells by flow cytometry. We tested different concentrations of both TTR variants (0.2‐2 μ mol L⁻¹) and observed that WT TTR always increased the internalization of FAM‐Aβ, as previously reported, contrarily to L55P TTR (data not shown).

Using 2 μ mol L⁻¹ of TTR, as shown in Figure 2A, in both hCMEC/D3 and HepG2 cells, Aβ internalization was increased in the presence of WT TTR‐633 (stable). Differently, L55P TTR‐633 decreased FAM‐Aβ internalization in hCMEC/D3 cells and had no significant effect in HepG2 cells, at this concentration. These results suggest that TTR stability is important for its participation in Aβ transport.

Figure 2.

Transthyretin (TTR) stability is critical for Aβ internalization. (A) Flow cytometry analysis of both hCMEC/D3 and HepG2 cells showed higher internalization of FAM‐Aβ in the presence of WT TTR‐633, but not L55P TTR‐633. Moreover, WT TTR‐633 internalization was also increased in the presence of FAM‐Aβ. Furthermore, uptake of L55P TTR‐633 was observed to be much higher than that of WT TTR‐633 (# and * denote significance from WT TTR and Aβ, respectively). (B) Microscopy images of HepG2 cells showing higher internalization and colocalization of FAM‐Aβ with WT TTR‐633 and not L55P TTR‐633. [TTR]=2 μ mol L⁻¹, [FAM‐Aβ] =0.1 μ mol L⁻¹, n=3 for all conditions. Data are expressed as mean ± SD

Interestingly, TTR internalization also varied, as uptake of L55P TTR‐633 was observed to be much higher than WT TTR‐633, which may be related to higher preference of cells for its degradation, probably due to its unstable conformation. Importantly, higher WT TTR‐633 internalization was observed in the presence of FAM‐Aβ, which can imply the cointernalization of these two molecules into the cells, while this was not the case for L55P TTR‐633.

Fluorescent microscopy analysis of HepG2 cells, as presented in Figure 2B, confirmed the flow cytometry results, as higher internalization of FAM‐Aβ (green) was observed in the presence of WT TTR‐633 when compared to cells incubated with FAM‐Aβ alone. Interestingly, the merged image evidenced colocalization of FAM‐Aβ with WT TTR‐633, further strengthening a role for TTR in Aβ transport. Contrarily, the presence of L55P TTR‐633 decreased FAM‐Aβ internalization, similar to the observation made by flow cytometry in hCMEC/D3 cells (Figure 2A), and this TTR variant did not colocalize with the peptide, indicating that this unstable TTR could not contribute to Aβ transport, and possibly competed with the peptide for elimination. Additionally, these results also confirmed that L55P TTR is internalized more than WT TTR.

3.3. Aβ internalization is increased when L55P TTR is stabilized with IDIF or resveratrol

As previously reported, the effect of TTR stabilizers on the TTR/Aβ interaction is best observed when using unstable TTR mutants.25 To study the importance of TTR stabilization in Aβ clearance, FAM‐Aβ internalization was assessed in the absence or presence of L55P TTR preincubated with or without IDIF or resveratrol. As depicted in Figure 3A, L55P TTR decreased FAM‐Aβ internalization in hCMEC/D3 cells, as compared to cells treated with FAM‐Aβ alone, whereas L55P TTR preincubated with IDIF or resveratrol induced higher Aβ internalization in these cells, compared to cells treated with FAM‐Aβ alone or with L55P TTR. For HepG2 cells (Figure 3B), although L55P TTR alone at this concentration (0.2 μ mol L⁻¹), unlike in hCMEC/D3 cells, increased the internalization of FAM‐Aβ, L55P TTR preincubated with IDIF further elevated the percentage of cells internalizing FAM‐Aβ, compared to cells incubated with FAM‐Aβ, alone or with L55P TTR. L55P TTR preincubated with resveratrol also increased FAM‐Aβ internalization in these cells but only when compared to FAM‐Aβ alone. Importantly, in both cell lines, the drugs alone could not increase the FAM‐Aβ internalization, showing that this effect is through stabilization of L55P TTR.

Figure 3.

L55P transthyretin (TTR) stabilization restores its ability to internalize Aβ. (A) hCMEC/D3 cells internalized less FAM‐Aβ when incubated with L55P TTR, as compared to cell incubated with FAM‐Aβ alone, while internalization of the peptide was increased upon stabilization of L55P TTR with iododiflunisal (IDIF), or resveratrol. (B) Although FAM‐Aβ internalization in HepG2 cells was increased upon incubation with L55P TTR, it was further elevated when L55P TTR was stabilized with IDIF. Cells treated with L55P TTR preincubated with resveratrol showed increased FAM‐Aβ internalization only when compared to FAM‐Aβ alone. (C) IDIF and resveratrol increased uptake of FAM‐Aβ in primary hepatocytes derived from TTR+/− mice, but not in TTR−/− hepatocytes. [TTR]=0.2 μ mol L⁻¹, [drugs]=2 μ mol L⁻¹, [FAM‐Aβ]=0.1 μ mol L⁻¹, n=3 for all conditions, and data are expressed as mean ± SD. *, #, +, and x, respectively, denote significance from Aβ, Aβ+L55P TTR, Aβ+IDIF, and Aβ+resveratrol

As shown in Figure 3C, in primary hepatocytes derived from TTR+/− mice, incubation of cells with FAM‐Aβ in the presence of IDIF or resveratrol resulted in higher levels of the peptide internalization when compared to cells incubated with FAM‐Aβ alone. To confirm that this observation was related to TTR produced and secreted by these cells, primary hepatocytes derived from TTR−/− were also used and, as expected, drugs did not produce any effect on FAM‐Aβ internalization. Importantly, addition of WT TTR increased the uptake of FAM‐Aβ in these hepatocytes. Also, as we reported before, hepatocytes from TTR+/− mice showed higher levels of FAM‐Aβ internalization compared to TTR−/− hepatocytes.

3.4. WT TTR or IDIF‐stabilized L55P TTR increases in vitro passage of Aβ through the BBB model

In order to study the importance of TTR stability in Aβ transport across the BBB, hCMEC/D3 cells were cultured in transwells inserts, as depicted in Figure 4A. In this experiment, Aβ was added to basolateral compartment (brain side), in the absence or presence of WT TTR‐633, L55P TTR‐633, IDIF, or L55P TTR‐633 preincubated with IDIF. After 24 hours, samples were collected from both chambers (brain and blood sides) and Aβ levels were measured using an ELISA kit. Results are displayed in Figure 4B, and as expected and previously shown, the permeability of the cell monolayer to Aβ was significantly increased when WT TTR‐633 was present. Differently, L55P TTR‐633 did not promote Aβ transport across the monolayer; however, upon its stabilization with IDIF, Aβ transport was significantly increased, corroborating the internalization studies.

Figure 4.

WT transthyretin (TTR), but not L55P TTR, increases hCMEC/D3 permeability to Aβ. (A) Schematic picture of the transwells system presenting the brain and blood sides; Aβ peptide was added to the brain side, in the absence or presence of WT TTR‐633, L55P TTR‐633, iododiflunisal (IDIF), or L55P TTR‐633 preincubated with IDIF. (B) Brain‐to‐blood permeability of hCMEC/D3 to Aβ, measured by Elisa, was found to be increased in the presence of WT TTR‐633 or L55P TTR‐633 preincubated with IDIF. [TTR]=0.2 μ mol L⁻¹, [Drugs]=2 μ mol L⁻¹, [Aβ]=0.1 n mol L⁻¹. (C) Assessing the intensity of Alexa Fluor 633‐labeled TTRs in both brain and blood sides, using fluorometry, showed the amount of WT TTR‐633 or L55P TTR‐633 preincubated with IDIF that remained in the brain (gray bars) was higher than that of L55P TTR‐633. (D) Confocal microscopy analysis of Aβ, WT TTR‐633, or L55P TTR‐633 levels in cells grown in transwells incubated with: (a) Aβ, (b) Aβ+WT TTR‐633, (c) Aβ+L55P TTR‐633, and (d) Aβ+L55P TTR‐633 preincubated with iododiflunisal (IDIF), (e) Fluorescent mean intensity was quantified using ImageJ software using three images from each condition. n=3 for all conditions and data are expressed as mean ± SD. *, #, +, and ^, respectively denote significance from Aβ, Aβ+L55P TTR‐633, Aβ+IDIF, and Aβ+WT TTR‐633

To inspect what happens to TTR during assisted Aβ passage, we measured the amount of labeled TTR in the samples, using fluorimetry. Figure 4C shows the fluorescent intensity of TTR‐633 (WT or L55P TTR), present in the blood side or that remained in the brain side. Interestingly, the amount of WT TTR‐633 that remained in the brain side was higher than that of L55P TTR‐633, which can imply that L55P TTR was degraded more, stayed inside cells, or passed across the monolayer faster than WT TTR‐633. According to these results, and considering the possibility of faster passage of L55P TTR‐633, its levels in the blood were expected to be higher. However, our results revealed similar amounts of both TTR variants in the compartment mimicking the blood side; therefore, both events, higher degradation (either in the brain or in blood sides) and higher passage, seem to occur at the same time.

Moreover, when L55P TTR‐633 was preincubated with IDIF, to promote its stabilization and to perform in a similar fashion to WT TTR, the remained TTR in the brain side was higher than the L55P TTR‐633 alone. Accordingly, stabilization with IDIF resulted in lower levels of degradation. Importantly, putting these results together with those of Aβ transport shown in Figure 4B, our data suggest that even if L55P TTR has passed faster through the cell monolayer, it did not assist Aβ passage, while WT TTR or L55P TTR preincubated with IDIF efficiently bound Aβ, assisting its passage and clearance across the cell monolayer.

We also assessed Aβ and TTR levels inside the same monolayer of cells by performing immunostaining and observing the fluorescence using confocal microscopy (Figure 4; the quantification is presented in Figure 4D‐e). In the presence of WT TTR‐633 (Figure 4D‐b), more Aβ entered the cells, when compared to the Aβ‐alone condition (Figure 4D‐a). Although in the presence of L55P TTR‐633 (Figure 4D‐c), levels of Aβ inside cells were similar to the ones in cells incubated with WT TTR‐633, and also considering the results presented in Figure 4B, we concluded that when stable TTR is not present, the process of Aβ transport across the monolayer is either delayed or Aβ gets arrested in the cells and cannot pass across the monolayer as effectively as in the presence of stable or stabilized TTR. The decrease in Aβ levels inside the cells observed in the presence of L55P TTR‐633 preincubated with IDIF (Figure 4D‐d) is concordant with the results of Aβ transport (Figure 4B) showing the highest percentage of passage of Aβ among the conditions tested. On the other hand, the amount of TTR inside cells confirmed what was observed before, as L55P TTR‐633 internalization was higher than WT TTR. Although the measured internalization of L55P TTR preincubated with IDIF was also high, assuming that not all L55P TTR‐633 is stabilized even after incubation with IDIF, it is possible that this nonstabilized L55P TTR‐633 stays longer in the cell (to be targeted for degradation). As a result, higher levels of stabilized L55P TTR remained in the brain side in the presence of IDIF compared to L55P TTR alone (Figure 4C).

3.5. Aβ, in the presence of WT TTR, colocalizes with lysosomes and increases lysosomal activity in HepG2 cells

In order to investigate the fate of Aβ after internalization in HepG2 cells, we used Lysotracker deep red to observe whether FAM‐Aβ enters lysosomes for degradation and whether TTR stability plays a determinant role in this process.

Cells incubated with Lysotracker deep red alone were used as control to select a gate for APC+ cells; then, we quantified the increase in the percentage of APC+ cells in those incubated with FAM‐Aβ (Figure 5A‐a), WT TTR (Figure 5A‐b), FAM‐Aβ and WT TTR (Figure 5A‐c), L55P TTR (Figure 5A‐d), or FAM‐Aβ and L55P TTR (Figure 5A‐e). The quantification, shown in Figure 5A‐f, indicates that when cells were incubated with FAM‐Aβ in the presence of WT TTR, lysosomal activity was significantly increased compared to cells incubated with FAM‐Aβ or WT TTR alone, suggesting that the goal of TTR‐assisted Aβ internalization in liver is degradation of Aβ. Cells incubated with L55P TTR alone (or together with FAM‐Aβ) also showed increased number of lysosomes, implying higher activity of the degradation system. Importantly, there was no significant difference between cells incubated with L55P TTR with or without FAM‐Aβ, indicating that the higher level of lysosomes was mainly due to the presence of L55P TTR. Thus, this increase, unlike the effect of WT TTR, aimed mainly at L55P TTR degradation.

Figure 5.

Lysosomal activity and colocalization with Aβ is increased in the presence of WT transthyretin (TTR) in HepG2 cells. (A) Flow cytometry analysis of HepG2 cells incubated with (a) FAM‐Aβ, (b) WT TTR, (c) FAM‐Aβ and WT TTR, (d) L55P TTR, or (e) FAM‐Aβ and L55P TTR. (f) Quantification of the increase in the percentage of APC+ fluorescent cells (Lysotracker deep red) using FlowJo software. n=3 for all conditions and data are expressed as mean ± SD. * and #, respectively, denote significance from Aβ and WT TTR‐633. (B) Microscopy images of HepG2 cells incubated with Lysotracker confirmed the flow cytometry results and also showed higher colocalization of FAM‐Aβ with lysosomes only when WT TTR was present. (C) Focused image of cells coincubated with Lysotracker and FAM‐Aβ in the presence of WT TTR; arrows point out yellow dots where FAM‐Aβ (green) colocalizes with lysosomes (red), indicating higher probability of FAM‐Aβ degradation, when stable WT TTR is present

This result was confirmed by fluorescence microscopy and, as depicted in Figure 5B, when Aβ was coincubated with WT TTR, the colocalization of FAM‐Aβ with lysosomes increased. Contrarily, in the presence of L55P TTR, and although the lysosomal activity was higher, colocalization of Aβ with these organelles was not increased. Figure 5C shows a focused image of cells incubated with FAM‐Aβ and WT TTR in the presence of Lysotracker and arrows point out yellow dots where Aβ (green) colocalizes with lysosomes (red), indicating higher probability of Aβ degradation, when stable WT TTR is present.

3.6. Levels of LRP1 protein expression increase in the presence of stable TTR

We had previously shown that LRP1 gene and protein expression levels are regulated by TTR.18 Therefore, to investigate the effect of TTR stability regarding LRP1 expression, we performed immunocytochemistry in hCMEC/D3 cells and mouse hepatocytes (AD/TTR+/−) in the presence of WT or L55P TTR (2 μ mol L⁻¹). As presented in Figure 6, microscopy analysis showed higher levels of LRP1 in the presence of WT TTR, but not in that of L55P TTR, for both cell types. Furthermore, LRP1 protein levels were increased in brain sections of AD mice treated with IDIF, when compared to nontreated animals.

Figure 6.

Only stable transthyretin (TTR) can up‐regulate LRP1 protein levels. hCMEC/D3 cells and mouse primary hepatocytes were incubated with cell media, WT TTR or L55P TTR. LRP1 protein levels were then assessed using immunocytochemistry, and in both cells, LRP1 protein levels were observed to be increased only in the presence of stable WT TTR. Moreover, LRP1 levels, as assessed in brain section from Alzheimer's disease mice, nontreated or treated with iododiflunisal, were found to be increased after treatment

We also carried out western blot analysis for the liver homogenates from AD/TTR+/+ and IDIF‐treated AD/TTR+/+ mice which were available in the laboratory from previous work.26 As a control, we included nontransgenic mice (TTR+/+), and as expected, these animals showed levels of LRP1 higher than the AD mice (Figure 7A). Importantly, an increase in LRP1 protein levels was observed in the liver homogenates of AD/TTR+/+ mice treated with IDIF when compared to nontreated AD/TTR+/+ mice.

Figure 7.

LRP1 and sLRP1 levels are increased, respectively, in the liver and plasma of iododiflunisal (IDIF) ‐treated animals. (A) Relative LRP1 protein expression levels were increased in the livers of Alzheimer's disease (AD) mice treated with IDIF (n=4). As expected, AD mice (n=4) showed a significant decrease in LRP1 protein levels when compared to non‐AD mice (n=2). (B) sLRP1 ELISA for plasmas of mice showed higher sLRP1 levels in TTR−/− mice when compared to transthyretin (TTR)+/+, due to sLRP1 shedding. n=4 for each group. (C) sLRP1 ELISA for plasmas of mice revealed that sLRP1 levels are increased after treatment of AD mice with IDIF. Further, plasma sLRP1 shedding, similar to nontransgenic mice, was found to be higher in AD/TTR−/− mice when compared to AD/TTR+/+. (n=5 for AD/TTR+/+ and AD/TTR+/−, n=3 for AD/TTR−/−). (D) Western blot analysis of LRP1 levels in the livers of AD mice with different TTR backgrounds, similar to nontransgenic mice, showed reduction in liver LRP1 levels in AD/TTR−/− and AD/TTR+/− compared to AD/TTR+/+. Data are expressed as mean ± SD in A and D and as mean ± SEM in B and C . *P<0.05 and **P<0.01

To further investigate the role of LRP1 in TTR‐assisted Aβ clearance, we also studied levels of sLRP1 (soluble extracellular domain of LRP1 receptor) in the plasma of mice (AD or non‐AD) with different TTR backgrounds or AD/TTR mice treated with IDIF, using an ELISA kit (Figure 7B, C). As shown in Figure 7B, unexpectedly, an increase in sLRP1 levels was observed in the plasma of TTR−/− mice compared to TTR +/+ animals. We had previously shown lower levels of LRP1 in the livers of TTR−/− mice,18 therefore suggesting that the increase in sLRP1 levels reflects an increased shedding of the LRP1 extracellular domain, a mechanism known to occur in the presence of Aβ and pathogenesis of AD,34 in mouse plasma after LPS treatment, and in human patients with rheumatoid arthritis,35 suggesting that LRP1 shedding is increased in response to inflammatory processes.

As shown in Figure 7C, IDIF treatment results in a significant increase in plasma sLRP1. However, as already discussed in Figure 7A, IDIF treatment increases also total liver LRP1 levels; thus in this case, sLRP1 levels are elevated not because of the shedding but due to the increased LRP1 expression levels. This further contributes to Aβ clearance, which is increased in IDIF‐treated mice, as previously reported,26 suggesting that Aβ binds to sLRP1 in the plasma (preventing its return to the brain) and then uses hepatic LRP1 to be internalized into the liver cells for degradation. Importantly, treatment of AD/TTR−/− mice with IDIF could not increase sLRP1 levels, which confirms that IDIF function is through the stabilization of endogenous TTR. As shown in Figure 7C, the absence of TTR in AD mice, similar to the trend observed in non‐AD mice (Figure 7B), resulted in increased sLRP1 levels. Therefore, we also performed western blot for LRP1 levels in the livers of AD mice with different TTR background, and as expected, it was observed that TTR reduction coincides with reduction in liver LRP1 levels (Figure 7D), confirming that the increase in sLRP1 levels in the absence of TTR in AD mice (Figure 7C) was due to LRP1 shedding.

Altogether, these results indicate that TTR reduction decreases liver LRP1 protein level and increases its shedding, resulting in higher plasma sLRP1 levels and, interestingly, IDIF treatment demonstrated both higher hepatic LRP1 and higher plasma sLRP1 levels leading to higher Aβ clearance, through a mechanism that should be further investigated.

4. DISCUSSION

Transthyretin, previously referred to as prealbumin, was known for decades as one of the transporters of retinol‐binding protein and thyroxine (T4) in the blood and CSF. TTR is a symmetrical tetramer composed of four identical subunits, each one made of 127 amino acids, which migrate as a complex of 55 kDa.36 This homotetrameric structure has an open channel where T4 binds,37 while retinol‐binding protein interacts with one of the dimers, at the surface.38 More recently, several studies were engaged to investigate the neuroprotective roles of TTR, including its role as an Aβ‐binding protein and its impact in AD. We have already shown that TTR participates in Aβ efflux and clearance from the brain and at the liver and that the absence of TTR in mice decreases Aβ clearance, reduces expression of the main receptor for Aβ clearance, LRP1, and increases levels of Aβ 40 and 42 in the plasmas.18

TTR has been shown to be diminished not only in aging 39 (while the highest amount was observed at age 35 in males) and inflammation (as a negative acute‐phase protein),40 but also during mild cognitive impairment (MCI) and AD.20 Decreased levels of TTR in AD have been reported not only in the CSF 19 but also in the plasma.8 However, the reason for this decrease, and whether this is a cause or effect of the disease, is not known yet. Further, there is still controversy on which TTR species binds Aβ peptide. While Jiali Du and colleagues reported that it is the monomeric specie of TTR which binds Aβ,41, 42 data obtained from the administration of TTR tetrameric stabilizers to AD transgenic mice showing improvements in AD features argue in favor of the tetramer.25 These evidences also point to TTR instability in AD, further supported by the decreased ability of plasma TTR to transport T4 in patients with AD, and by the inability of the unstable mutant L55P TTR to inhibit Aβ aggregation and toxicity.25 TTR stability has thus been proposed as a key factor in TTR/Aβ interaction and is probably also important in other contexts of CNS neuroprotection. In fact, genetic stabilization of transthyretin tetramer (carriers of TTR variant T119M) has been shown to be associated with reduced risk of cerebrovascular disease and increased life expectancy.43 Although instability of TTR may arise from factors such as aging, metal ions, or interaction with other proteins, recently it has been postulated that there may be a link between genetic TTR destabilization and AD, as some TTR rare variants were shown to be enriched in patients with AD. Furthermore, Aβ concentration in supernatants of culture media of cells overexpressing WT TTR was decreased compared to cells overexpressing TTR mutants (variants V50M and A111V), indicating that TTR variants may function differently in Aβ clearance.23

This work started by assessing the status of TTR stability in patients with AD and showed that the percentage of TTR folded species is decreased in AD plasma, as well as total TTR levels. Inappropriate folding of the TTR tetramer might result in engagement of the unfolded protein response system, resulting in accelerated clearance, explaining the lower levels.

Then, to investigate the importance of TTR stability on the recently proposed functions (ie, Aβ clearance and LRP1 expression levels), we used WT TTR and the unstable mutated L55P TTR. We showed that in both cells mimicking the BBB and the liver, L55P TTR could not perform as WT TTR and increase the internalization of the peptide; in some cases, incubation with L55P TTR even resulted in decreased Aβ uptake by the cells. Moreover, WT TTR colocalized with Aβ, whereas L55P did not, indicating that this mutant was not participating in Aβ transport. WT TTR internalization also increased in the presence of the peptide, suggesting that TTR and Aβ were cointernalized. Internalization of L55P TTR was always much higher than WT TTR but did not increase in the presence of Aβ, again suggesting that this mutant was not participating in Aβ elimination and probably was being signaled for degradation.

To further confirm the involvement of TTR stability and the potential use of TTR stabilization as a therapeutic approach in AD, we evaluated internalization of the Aβ peptide in the presence of stabilized L55P TTR, through the use of IDIF and resveratrol. IDIF has been shown to bind deep in TTR hormone‐binding channel, while providing extra hydrophobic contacts and electrostatic interactions with the protein and inducing conformational alterations, resulting in increased stability of the tetramer.44 Previous studies have shown that IDIF, administered orally to AD mice, bound TTR in plasma and improved AD features. Further, Aβ levels were shown to be reduced in the plasma, suggesting an important role for TTR stability in Aβ clearance both from the brain and from the periphery. Resveratrol, another TTR stabilizer, a natural polyphenol found in grapes, berries, peanuts, and pomegranates, has been shown to reduce neurodegeneration and cognitive decline in mice displaying AD features.45 This compound was also described to lower the levels of secreted and intracellular Aβ peptides produced from different cell lines,46 to increase TTR levels 47 and further decrease brain Aβ burden and to increase LRP1 levels in the treated AD animals.27 Moreover, resveratrol has been demonstrated to increase TTR affinity to Aβ.25 However, the effect of TTR stabilization has never been specifically studied in Aβ transport and LRP1 levels. Here, we showed that in hCMEC/D3 cells, whereas L55P TTR decreased Aβ uptake, Aβ internalization was significantly higher in the presence of L55P TTR stabilized with IDIF or resveratrol, when compared to cells incubated with Aβ with or without L55P TTR. In HepG2 cells, however, L55P TTR alone at the concentration used increased the internalization of Aβ, but L55P TTR preincubated with IDIF increased further the percentage of cells that internalized Aβ, compared to cells incubated with Aβ with or without L55P TTR. The ability of TTR stabilization to improve Aβ internalization was confirmed in primary cultures of hepatocytes, and while cultures derived from TTR+/− mice showed increased uptake of the peptide in the presence of IDIF or resveratrol, the ones derived from TTR−/− hepatocytes did not. Although we did not assess the folding status of TTR in this model, TTR tetramer dissociation and partial unfolding have been shown to precede amyloid fibril formation,48 which can be achieved by protein dilution.49 Thus, it is possible that genetic reduction of TTR also results in instability.

Permeability assays, in addition to confirming that TTR stability is important for Aβ peptide transport, have also suggested that TTR instability results in a longer retention time within the cells, probably to promote its own degradation, as also indicated by the lysosomal activity measured in the presence of both TTR variants. In fact, L55P TTR alone induced high lysosomal activity, which did not increase in the presence of Aβ, once again also indicating this variant is not binding Aβ, nor helping with its elimination. On the other hand, WT TTR led to increased lysosomal activity only in the presence of the Aβ peptide, indicating that the purpose was Aβ degradation. These results were further corroborated by microscopy analysis, which showed colocalization of Aβ and lysosomes in cells coincubated with WT TTR, but not with L55P TTR. Thus, increased lysosomal activity promoted by L55P TTR alone most likely aimed at its own degradation and not the peptide.

Aβ clearance, affected by transporter expression and activity, ligand affinity and competition, and vascular integrity,2 is compromised in AD, through a number of ways. Expression of the blood efflux transporters such as LRP1 is decreased, whereas expression of the blood influx transporter RAGE is up‐regulated.50 LRP1, which is highly expressed in the brain and liver, is involved in a number of pathways linked to AD pathogenesis. Importantly, LRP1 can regulate gene expression through its intracellular domain and can also regulate the endocytosis of several ligands including APP and Aβ bound to its extracellular domain. It mediates Aβ clearance by cellular internalization followed by lysosomal degradation and/or transcytosis of intact Aβ across the BBB to the circulation.51

We had previously shown that LRP1 gene and protein expression levels are up‐regulated by TTR,18 although the link between TTR instability and decrease in LRP1 in AD was not established. Here, we showed that stable WT TTR, but not unstable L55P TTR, can increase LRP1 protein levels, both in hCMEC/D3 cells and in primary hepatocytes. Importantly, brains of IDIF‐treated AD mice revealed higher levels of this receptor. Interestingly, it is reported that levels of hepatic LRP1 expression and function are decreased in hypothyroidism and are regulated by the thyroid hormone,52 which may demonstrate another linkage between TTR stability and LRP1 expression level as TTR stability is crucial in thyroid hormone transport.

Moreover, the soluble form of LRP1 binds and sequesters 70%‐90% of Aβ in human plasma. Oxidative changes in AD are associated with alterations in sLRP1, decreasing its affinity for Aβ.50 sLRP1 bound to Aβ 40 and 42 has been shown to be significantly reduced in MCI and AD patients,53 which can be one reason for the increased levels of free Aβ40 and Aβ42 in the plasma.54 Furthermore, inflammation, as a common feature of AD, can also influence ligand affinity by making the pH more acidic, promote hyperphosphorylation of tau, and induce conformational changes in Aβ that delay or prevent its clearance.2 On the other hand, the extracellular domain of LRP1 receptor is susceptible to shedding, which changes its endocytotic activity and, consequently, the clearance of Aβ. For instance, it has been observed that exposure of human brain endothelial cells to Aβ causes shedding of sLRP1.34 Furthermore, sLRP1 levels are increased in CSF from aged and AD subjects due to shedding.55 Thus, dysregulation of sLRP1 shedding could impair Aβ clearance, contributing to the pathogenesis of AD. Moreover, a link between inflammation and LRP1 shedding has been demonstrated through the observation of increased sLRP1 levels in mouse plasma after LPS treatment or by macrophages, in vitro, in response to inflammatory mediators, and also in human patients with rheumatoid arthritis and systemic lupus erythematosus.35 Thus, we assessed the effect of TTR and its stability in sLRP1 levels in the plasma of mice. Although sLRP1 levels were increased in plasma of TTR−/− mice, total LRP1 levels are decreased in the liver of these animals, as previously described,18 suggesting the measured sLRP1 is mainly a consequence of shedding; the same behavior was also observed in AD mice. As for the impact of IDIF administration, and in spite of the higher levels of sLRP1 in treated animals as compared to controls, total LRP1 was also increased upon treatment, indicating that, in this case, the measured sLRP1 results from an overall increment of the receptor, and not shedding.

Altogether, our results indicate that TTR stability is important for its recently described functions in assisting Aβ transport at the BBB and at the liver and also in regulating LRP1 levels and activity. In conclusion, TTR stabilization can serve as an avenue to increase both Aβ elimination and LRP1 levels, which in turn will further participate in Aβ clearance. These results provide relevant information for the design of TTR‐based therapeutic strategies for AD.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

Alemi M, Silva SC, Santana I, Cardoso I . Transthyretin stability is critical in assisting beta amyloid clearance: Relevance of transthyretin stabilization in Alzheimer's disease. CNS Neurosci Ther. 2017;23:605–619. 10.1111/cns.12707