Neuroprotection in early stages of Alzheimer’s disease is promoted by transthyretin angiogenic properties

While still controversial, it has been demonstrated that vascular defects can precede the onset of other AD hallmarks features, making it an important therapeutic target. Given that the protein transthyretin (TTR) has been established as neuroprotective in AD, here we investigated the influence of TTR in the vasculature. We evaluated the thickness of the basement membrane and the length of brain microvessels, by immunohistochemistry, in AβPPswe/PS1A246E (AD) transgenic mice and non-transgenic mice (NT) bearing one (TTR+/−) or two (TTR+/+) copies of the TTR gene. The angiogenic potential of TTR was evaluated in vitro using the tube formation assay, and in vivo using the chick chorioallantoic membrane (CAM) assay. AD transgenic mice with TTR genetic reduction, AD/TTR+/−, exhibited a thicker BM in brain microvessels and decreased vessel length than animals with normal TTR levels, AD/TTR+/+. Further in vivo investigation, using the CAM assay, revealed that TTR is a pro-angiogenic molecule, and the neovessels formed are functional. Also, TTR increased the expression of key angiogenic molecules such as proteins interleukins 6 and 8, angiopoietin 2, and vascular endothelial growth factor, by endothelial cells, in vitro, under tube formation conditions. We showed that while TTR reduction also leads to a thicker BM in NT mice, this effect is more pronounced in AD mice than in NT animals, strengthening the idea that TTR is a neuroprotective protein. We also studied the effect of TTR tetrameric stabilization on BM thickness, showing that AD mice treated with the TTR tetrameric stabilizer iododiflunisal (IDIF) displayed a significant reduction of BM thickness and increased vessel length, when compared to non-treated littermates. Our in vivo results demonstrate the involvement of TTR in angiogenesis, particularly as a modulator of vascular alterations occurring in AD. Since TTR is decreased early in AD, its tetrameric stabilization can represent a therapeutic avenue for the early treatment of AD through the maintenance of the vascular structure.


Introduction
Alzheimer's disease (AD) patients undergo several neurovascular changes at different levels. Brain vascular dysregulation is the earliest and strongest factor during disease progression and is followed by amyloid-β (Aβ) peptide deposition, glucose metabolism dysregulation, functional impairment, and gray matter atrophy, in this order [1]. Decreased expression of the low-density lipoprotein receptor-related protein 1 (LRP-1) and Pglycoprotein (P-gp), as well as upregulation of the receptor for advanced glycation end products (RAGE), are mechanisms reported to be changed in AD patients, leading to Aβ accumulation in the brain [2,3]. In addition to defective clearance mechanisms, increased endothelial pinocytosis, decreased number of mitochondria, decreased glucose transporter (GLUT)-1, and loss of tight/adherents junctions are features detected in AD [4]. The reduction of the capillary density is also characteristic of the AD brains [5]. This is due to an aberrant angiogenesis with premature pruning of capillary networks. This defective angiogenesis may be caused by a lack of angiogenic stimuli and unresponsive endothelium [6]. Although other authors describe increased vascular density in AD [7], the underlying angiogenic process has pathological characteristics. Some studies suggest that the promotion of angiogenesis results in concomitant blood-brain barrier (BBB) disruption and vessel leakiness [7]. Other studies defend that the vascular damage is a consequence of poor blood perfusion of the brain, leading to hypoperfusion/hypoxia causing the BBB dysfunction [8]. Other authors argue that the accumulation of Aβ in the walls of the capillaries can contribute to the reduced brain capillary density in AD via anti-angiogenic activity [9,10]. Another observed alteration in AD is the increased thickness of the vascular BM in AD [11]. Since the increase in BM thickness occurs before Aβ deposition, it is speculated that it functions as a physical barrier to the Aβ clearance across the BBB [12]. Some studies have related this BM thickening with increased collagen IV content, in AD and aging [13,14].
Transthyretin (TTR), a 55-kDa homotetrameric plasma and cerebrospinal fluid (CSF) protein, transports retinol through binding to the retinol-binding protein (RBP), which binds at the surface of TTR, and thyroxine (T4), which binds at a central hydrophobic channel formed at the dimer-dimer interface [15]. In the CSF, TTR is the main Aβ binding protein [16], providing neuroprotection by avoiding Aβ aggregation [16][17][18][19][20][21][22][23] and toxicity [17,24], and by participating in Aβ brain efflux at the BBB [25]. TTR is early decreased in AD, both in plasma [26][27][28] and in the CSF [29], probably due to its tetrameric instability [26,30], hypothesized to result in accelerated clearance and lower levels. TTR instability is also a key feature in familial amyloid polyneuropathy (FAP), a systemic amyloidosis that is usually caused by mutations in TTR. The amyloidogenic potential of the TTR variants is inversely correlated with its tetrameric stability [31], and the dissociation of the tetramer into monomers is at the basis of the events that culminate with TTR amyloid formation [32,33]. TTR stabilization, used as a therapy in FAP [34,35], can be achieved through the use of small-molecule compounds sharing molecular structural similarities with T4 and binding in the T4 central binding channel [36][37][38]. Although no TTR mutations have been found in AD patients [22], TTR stabilization has also been proposed as a therapeutic strategy to recover its ability to protect in AD [19,39], and shown beneficial in a mouse model of AD [39,40]. Iododiflunisal (IDIF), a potent TTR stabilizer, was administered to AD mice and bound plasma TTR displacing T4, resulting in decreased Aβ amyloid burden and total Aβ brain levels, and improved cognition [40]. Interestingly, TTR stabilization by IDIF improves TTR-assisted Aβ brain efflux in vitro and enhanced the expression of LRP-1 in vivo [30]. The formation of TTR-IDIF complexes enhances BBB permeability of both IDIF and TTR, in vivo [41].
TTR has also been implicated in angiogenesis and the first reports of its involvement have been described in diseases such as FAP [42]; in diabetic retinopathy (DR) [43,44], and lately, in cancer [45]. As reported, a study investigated the effect of TTR in angiogenesis by treating human umbilical vein endothelial cells (HUVECs) with wild-type (WT) TTR or a common FAP TTR mutant, V30M. The authors concluded that the TTR mutant inhibited cell migration and decreased survival relative to the WT TTR, by down-regulating several proangiogenic genes for angiopoietin-2 (Ang-2), vascular endothelial growth factor (VEGF) receptors 1 and 2, basic fibroblast growth factor (bFGF), and transforming growth factor-beta 2 (TGF-β2) [42]. In another study, to investigate how TTR affects the development of new vessels in DR, human retinal microvascular endothelial cells (hRECs) were cultured with TTR in natural and simulated DR environments (hyperglycemia and hypoxia). In the DR environment, TTR inhibited cell proliferation, migration, and tube formation, by repressing the expression of the pro-angiogenic genes Ang-2 and VEGF receptors 1 and 2 [43]. Conversely, in a low glucose environment, these angiogenesis-related features were improved by TTR. Recently, it was reported that TTR levels were increased in human serum of lung cancer patients. Additionally, TTR was shown able to promote tumor growth by enhancing several lung ECs functions as permeability, migration, and tube formation [45]. However, TTR potential in angiogenesis has never been addressed in vivo and the possible participation of TTR in brain angiogenesis and vascular alterations has never been elucidated.
Taking these evidences into account, this work aimed at investigating the angiogenic potential of TTR and at assessing its involvement in the vascular impairment that occurs in AD.
In order to study the role of TTR in collagen IV deposition or in vessel density, cohorts of littermates 7month-old female mice AD/TTR+/+ and AD/TTR+/−, cohorts of littermates 3-month-old female mice NT/ TTR+/+ and NT/TTR+/− and one cohort of 3-monthold female mice AD/TTR+/− were used. AD/TTR+/− female control mice or treated with IDIF [40] for 2 months (from 5-to 7-month-old), were used to investigate the relevance of TTR stabilization in collagen type IV levels, in AD.

Collagen IV immunohistochemistry
Free-floating 30-μm-thick coronal brain sections of mice were permeabilized with 0.25% Triton X-100 in phosphate-buffered saline (PBS) for 10 min at room temperature (RT), blocked with 5% bovine serum albumin (BSA) in PBS for 1 h at RT and incubated with primary rabbit anti-collagen IV antibody (1:100) (Abcam) in 1% BSA in PBS overnight at 4°C. Next, sections were washed with PBS and incubated with Alexa Fluor-568 goat anti-rabbit IgG antibody (1:2000) for 1 h at RT. All steps were performed with agitation. To remove tissue autofluorescence, sections were covered with Sudan black B solution (0.3% Sudan black B in 70% ethanol) applied for 5 min at RT, followed by multiple washing steps with PBS at RT with agitation. The brain sections were dried for 20 min at RT and mounted on 0.1% gelatin-coated slides with Fluoroshield TM with DAPI (Sigma-Aldrich). Sections were visualized and photographed using a Zeiss Axio Imager Z1 microscope equipped with an Axiocam MR3.0 camera and Axivision 4.9.1 software. A total of twenty-thirty randomly selected vessels in the hippocampus and cortex of each mouse was photographed at × 100 magnification, and the ratio intensity/area was measured using the ImageJ software.
To assess the vascular density of mice brains, 30-μmthick coronal brain sections were boiled at 90°C in citrate buffer for 15 min for antigenic recovery and then washed with 0.3% Triton X-100 in PBS for 10 min at RT. Tissues were blocked/permeabilized with a solution of 1% BSA and 0.5% Triton X-100 in PBS, overnight at 4°C. The coronal sections were then incubated for 72 h at 4°C with primary rabbit anti-collagen IV antibody (1: 200) (Abcam) in a solution with 1% BSA, 0.5% Triton X-100, and 2% fetal bovine serum (FBS) in PBS. After, tissues were washed with 0.3% Triton X-100 in PBS at 4°C. Next, sections were incubated with Alexa Fluor-568 goat anti-rabbit IgG antibody (1:500) overnight at 4°C, followed by washing with 0.3% Triton X-100 in PBS and then dried for 20 min at RT and mounted on silane precoated slides with Fluoroshield TM with DAPI (Sigma-Aldrich). Sections were visualized and photographed using a Zeiss Axio Imager Z1 microscope equipped with an Axiocam MR3.0 camera (Carl Zeiss) and Axiovision SE64 Rel. 4.9.1 software. A total of twenty-twenty five fields of view were randomly selected from the cortex and hippocampus of each brain section and photographed at × 20 magnification. The total length of the blood vessels per field was measured using the ImageJ software.

Production and purification of human recombinant TTR
Human recombinant WT TTR (rec TTR) was produced in a bacterial expression system using Escherichia coli BL21 [49] and purified as previously described [50]. Briefly, after growing the bacteria, the protein was isolated and purified by preparative gel electrophoresis after ion exchange chromatography. Toxin-free TTR was obtained using the Detoxi-Gel Endotoxin Removing Gel (Thermo), following the manufacturer's instructions. Analyses of the protein were performed by SDS-PAGE, western blot, and native PAGE.

Purification of human TTR from sera
Human plasma from donors who were informed of the purpose of the study and gave their written consent were collected in accordance with the approved guidelines. Samples were subjected to affinity chromatography to isolate human TTR (hTTR); for this, we used 1 mL column of NHS-activated Sepharose coupled to rabbit antihuman TTR (Dako). The column was washed with PBS and then incubated with 500 μL of human plasma for 2 h at RT. To elute TTR from the column, 5 mL of Gentle Ag/Ab elution buffer (Thermo Scientific) were applied, and 1 mL-aliquots were collected and OD 280 nm was registered.
Tube formation assay hCMEC/D3 cells, grown in 25 cm 2 flasks, at a confluence of 80-90% were incubated for 24 h with EBM-2 medium (Lonza) containing 1% FBS (Gibco) (negative control), with bFGF (35 ng/mL) (positive control), or with rec TTR at different concentrations (10, 25, 250, 500 nM, and 1 μM) or with hTTR 250 nM. Then, cells were transferred into 96-well plates, previously coated with Matrigel (Corning), and grown in the same conditions of media, bFGF or TTR for another 9 h. Then, cells were photographed using the In Cell Analyzer 2000 (GE Healthcare) (magnification × 10). The supernatants were collected, centrifuged at 14.000 rpm for 10 min and stored at − 20°C. Each condition was performed in triplicate and experiments were repeated three times.

ELISA analysis for IL-6
IL-6 was also quantified in the supernatants collected from hCMEC/D3 cells used for the tube formation, in the presence of media alone or with rec TTR at different concentrations (10, 25, 250 nM, and 1 μM), using a LE-GEND MAX™ Human IL-6 Sandwich Enzyme-Linked Immunosorbent Assay (ELISA) Kit (BioLegend) with pre-coated plates. The assay was performed according to the manufacturer's recommendations. Analysis was performed using Synergy Mx and by measuring absorbance at 450 and 570 nm. A standard curve was generated for IL-6 from 7.8 pg/mL to 500 pg/mL.

Angiogenesis chick chorioallantoic membrane (CAM) assay
Commercially available fertilized chick (Gallus gallus) eggs were horizontally incubated at 37°C, in a humidified atmosphere. On embryonic development day (EDD) 3, a square window was opened in the shell after removal of 1.5-2 mL of albumen, to allow detachment of the developing CAM. The window was sealed with a transparent adhesive tape and eggs re-incubated. On EDD10, rec TTR (1 μM), hTTR (1 μM), PBS (vehicle, negative control) and bFGF (50 ng/μL, positive control) were placed on top of the CAM, into 3 mm silicone rings, under sterile conditions (1 condition per egg). Eggs were re-sealed and returned to the incubator for an additional 72 h. On EDD13, rings were removed, the CAM was excised from embryos and photographed exovo under a stereoscope, using a × 20 magnification (Olympus, SZX16 coupled with a DP71 camera). The number of new vessels (< 20 μm diameter) growing radially towards the inoculation area was counted in a blind fashion manner.

In vivo analysis of vascular permeability
The CAM model was also used to evaluate vascular permeability or vessel leakage, as a measure of TTRinduced neo-vessels functionality. Embryos were cultured ex ovo. To prepare shell-less CAM, eggs were incubated as described above and on EDD3, the content of the egg was transferred to sterile weigh boats, covered with square Petri dishes, and returned to the incubator for additional 7 days. At EDD10, 10 μl of PBS, rec TTR (1 μM) and VEGF (4 ng/μL) were inoculated on distinct sites of the same egg, twice each, into 3 mm silicone rings under sterile conditions. Three independent experiments were performed summing a total 16 CAM sites/condition (8 eggs). After 3 days (EDD13), chicken embryos were injected intravenously with 100 ul of Evans Blue Dye (EBD, Sigma) solution (0.5% EBD, 5% BSA in PBS) and further incubated for 60 min. After incubation, embryos were perfused with saline. The tissue underlying the rings (inoculation sites) was removed, cleaned in saline, blotted dry, weight, homogenized, and incubated in 200 ul of formamide (Sigma), at 38 ºC for 48 h, to release the extravasated dye. The samples were centrifuged and 175 μl of supernatant was quantified spectrophotometrically at 620 nm. The amount of EBD in the experimental samples was calculated by interpolating to a standard curve and the concentration of EBD per g of tissue was determined.

Statistical analysis
All quantitative data were expressed as mean ± standard error of the mean (SEM). Initially, data was assessed whether it followed a Gaussian distribution. In the cases of non-Gaussian distribution comparisons between two groups were made by non-parametric Kruskal-Wallis test and comparisons between two groups were made by Student's t test with a Mann-Whitney test.
When found to follow a Gaussian distribution, differences among conditions or groups were analyzed by one-way ANOVA with the appropriate post hoc pairwise tests for multiple comparisons tests. Differences in CAM assay and IL-6 Elisa kit were analyzed using one-way ANOVA followed by Tukey's multiple comparison test. P-values lower than 0.05 were considered statistically significant. Statistical analyses were carried out using GraphPad Prism 8 software for Windows.

TTR influences vascular features in the mouse brain
For this work we used AD/TTR+/+, AD/TTR+/−, NT/ TTR+/+, and NT/TTR+/− animals. We did not analyze the respective TTR-/-animals, although we obtained them in the course of breeding, to avoid indirect effects of TTR-deficiency, including compensatory processes, that could confound our interpretations. Additionally, the TTR+/− animals, in particular, the AD/TTR+/−, are a better representation of the behavior of TTR in AD, since TTR is decreased in this pathology but not absent.

Reduction of TTR increases the thickness of the collagen IV layer in brain microvessels of AD mice
To investigate a possible relationship of TTR reduction with the thickening of the BM and with the structural vascular alterations reported in AD, we evaluated collagen IV levels in brain microvessels, in the AD/TTR mouse model. This model established in different TTR genetic backgrounds [48], bears two AD-related transgenes (APP and PSEN1) and Aβ deposition starts at around 6 months [48]. In comparison to males, females present a more severe form of AD-like disease, thus in this study we used 7-month-old female animals. AD/TTR+/− females were compared to littermates with normal TTR expression, AD/TTR+/+. Our results revealed a significantly thicker collagen IV layer in 7-months old AD/TTR+/− as compared to AD/TTR+/+ animals, both in cortex and in hippocampus microvessels (Fig. 1 A1). Altogether, our results implicate TTR in vascular processes, which are known to be early dysregulated in AD.

Reduction of TTR decreases the length of brain microvessels in AD mice
To understand if TTR affects other cerebrovascular features and if the effect observed at the level of the BM is related to angiogenesis, we measured brain vascular density in the same animals.
Both cortex and hippocampus were analyzed and our results show that, in the hippocampus, reduction of TTR resulted in decreased vessel length in AD/TTR+/− mice as compared to AD/TTR+/+ (Fig. 1 A2, right panels). In the cortex, the differences were not statistically significant ( Fig. 1 A2, left panels). These observations support the results obtained for the BM thickness and further implicate TTR in angiogenesis, especially in the hippocampus, a particularly relevant brain area in the initial stages of AD.
Reduction of TTR increases the thickness of the collagen IV layer in brain microvessels of young non-transgenic mice Although our results suggest that TTR influences the thickness of the BM, in particular the collagen IV layer, we could not determine if the effect was direct or indirect. One hypothesis is that high levels of Aβ, as it happens in AD, either due to increased production, reduced elimination or both, could be responsible for the increase in collagen IV. It is possible that AD/TTR+/-mice show increased amount of collagen IV because less TTR is available to interact with and to eliminate Aβ.
Thus, in order to unravel this question, we compared collagen IV levels in non-transgenic (NT) littermate mice with two different TTR backgrounds, NT/TTR+/+ and NT/TTR+/-allowing to understand if TTR is directly involved. Furthermore, this evaluation was performed at the age of 3 months, which in this specific AD model is prior to amyloid deposition [48]. Results presented in Fig. 1B clearly show that NT/TTR+/-mice presented significantly higher levels of collagen IV in brain microvessels of both the cortex and hippocampus, as compared to NT/TTR+/+ animals, thus suggesting that it is, in fact, a direct effect of TTR.

TTR possesses angiogenic activity TTR promotes tube formation by hCMEC/D3 cells
The tube formation assay is a powerful in vitro test encompassing EC adhesion, migration, protease activity and tube formation (capillary-like structures). Thus, and to explore the angiogenic activity of TTR, endothelial cells of human brain origin, hCMEC/D3 cells, grown under tube formation-conditions, on Matrigel, were incubated with different concentrations of rec TTR. The results are displayed in Fig. 2A, A' and Supplementary Figure 2, and reveal that TTR affects the tube formation processes in a dose dependent-manner. Concentrations of TTR equal or above 250 nM result in a significantly higher area covered by the capillary-like structures, as compared to the negative control. These TTR concentrations are well below their physiologic concentration Fig. 1 Effect of TTR reduction in vascular features in mice brain microvessels. A1 Representative images and quantification plots of collagen IV (red) levels in the BM of cortex and hippocampus vessels derived from 7-month-old AD mice with different TTR genetic backgrounds, AD/TTR/ +/+ (n = 7) and AD/TTR+/− (n = 6 for cortex; n = 5 for hippocampus), showing significantly increased levels in microvessels from AD/TTR+/− compared to AD/TTR+/+ mice. Scale bar = 5 μm. A2 Representative images and quantification plots of the length of brain vessels, as evaluated by collagen IV staining, from 7-month-old AD mice with different TTR genetic backgrounds, AD/TTR/+/+ (n = 7) and AD/TTR+/− (n = 7), showing significantly decreased vessel length in the hippocampus of AD/TTR+/− compared to AD TTR+/+. Scale bar = 50 μm. B Representative images of the cortex and hippocampus and quantification plots of collagen IV immunostaining in microvessels of NT/TTR+/+ and NT/TTR+/− 3-month-old mice, showing an increase in collagen IV content in NT/TTR+/− mice (n = 4) relative to NT/TTR+/+ littermates (n = 4). Scale bar = 5 μm. Data are expressed as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001 in plasma and are similar to their concentration in the CSF. To confirm that the angiogenic effect was indeed provided by TTR, thus excluding the possibility of contaminants associated with the protocol of recombinant protein purification from bacteria, TTR isolated from human serum (hTTR) was also evaluated, and results corroborated the angiogenic activity of TTR ( Fig. 2A,  A'). As for TTR species present in the preparations, the TTR tetramer was the only form detected, evaluated in native gel (Supplementary Figure 1).

TTR is angiogenic in vivo and the neovessels formed are functional
To further confirm the angiogenic activity of TTR, we used the CAM assay. Both rec TTR and hTTR were tested and at 1 μM induced a significantly higher angiogenic response than the negative control, as deduced by the higher number of detected new vessels (vessels with a diameter under 20 um) (Fig. 2B, B'). TTR angiogenic response was comparable to the positive control (bFGF), in particular for the rec TTR.
Also using the CAM in vivo model, we studied the permeability of the TTR-induced vessels, by quantifying the leakage of EBD. This assay indicated that the permeability of TTR-induced vessels is comparable to the negative control (PBS), and significantly different from the positive control (VEGF) (Fig. 2C). It can be inferred that TTR-induced neovessels are functional (in contrast with leakier vessels induced by VEGF). , human recombinant TTR (rec TTR, 1 μM, n = 19) or TTR isolated from human plasma (hTTR, 1 μM, n = 9). TTR, both rec TTR and hTTR, had a significantly higher angiogenic response than the negative control. Scale bar = 1 mm. C In vivo vascular permeability was measured in CAM model by quantification of leaked EBD. The permeability of the new vessels induced by TTR (n = 20) was similar to the negative control (PBS, n = 18), in contrast to the significantly higher permeability of vessels induced by VEGF (n = 18). Data are expressed as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001

TTR regulates angiogenic molecules
To further explore the molecular mechanisms underlying the angiogenic activity of TTR, supernatants of hCMEC/D3 cells grown under tube-formation conditions, in the presence of rec TTR (1 μM) or with media alone, were used to identify key targets involved in angiogenesis which could be affected by TTR.
Among the ten molecules analyzed, IL-6, IL-8, Ang-2, and VEGF were differentially overexpressed in the presence of TTR, whereas the remaining six presented concentrations below the detection limit. As shown in Fig. 3A, the expression of detected molecules was significantly increased relative to the negative control when stimulated with rec TTR (1 μM) indicating that TTR acts as a pro-angiogenic molecule by increasing the expression of those molecules. It is possible that TTR affects other angiogenic molecules, possibly even those that were undetected by the current approach.
We also confirmed the effect of IL-6 using an ELISA approach and as can be appreciated in Fig. 3B it is clear a dose-response effect as TTR concentration is increased. While at 10 nM the differences are not significant, TTR concentrations between 25 nM and 1 μM lead to significantly increased expression of IL-6, as compared to the negative control. These results are also in line with those of the tube formation assay ( Fig. 2A').

The impact of TTR reduction on BM thickening is greater in AD than in NT mice
To understand if TTR reduction impacts differently in an AD and in a non-AD environment, we analyzed the effect of the same TTR reduction on the collagen IV layer, in the AD and in the NT backgrounds (NT/ TTR+/− versus AD/TTR+/−). Figure 4 depicts the results obtained and shows that the impact of TTR reduction is greater in an AD-like environment in the Fig. 3 Quantification of angiogenesis-related proteins. Supernatants from hCMEC/D3 cells grown under conditions of tube formation were collected 9 h after incubation with media alone or with rec TTR (1 μM). A Ten of the most common angiogenesis-related proteins were quantified using bead-based LEGENDplex assay by flow cytometry. Rec TTR 1 μM revealed ability to significantly increase the levels of IL-6, IL-8, Ang-2, and VEGF. Comparisons are relative to the negative control. B IL-6 levels measured by ELISA showed that while TTR 10 nM did not affect IL-6, TTR concentrations 25 nM-1 μM increased, in a concentration-dependent manner, the levels of IL-6. Data are expressed as mean ± SEM. *p < 0.05; **p< 0.01; ***p < 0.001 hippocampus, further corroborating the neuroprotective function of TTR in AD. Given that these are 3-monthold animals and that, in this model, deposition begins at around 6 months, our results support other findings that suggest brain vascular dysregulation as the earliest factor during the disease progression.
TTR stabilization results in decreased thickness of the collagen IV layer in brain microvessels of AD mice So far, in this work, we have shown that TTR reduction worsens AD features in mice, such as BM thickening. To further investigate a possible neuroprotective effect of TTR in the vascular context, and given that TTR Fig. 4 Effect of TTR reduction on collagen IV levels, in NT and AD mice. Representative images and quantification plots of collagen IV (red) immunostaining depicting microvessels of NT/TTR+/− (n = 4) and AD/TTR+/− (n = 5) 3-month-old mice in cortex and hippocampus, showing an increase in collagen IV expression in AD/TTR+/− mice relatively to NT/TTR+/− littermates, in the hippocampus. Data are expressed as mean ± SEM. *p < 0.05. Scale bar = 5 μm stabilization is used to improve its activity, we analyzed the BM thickness in brain microvessels in AD mice treated with one TTR stabilizer, IDIF. Administration of IDIF to AD mice from the age of 5-to 7-month-old, resulted in amelioration of some AD features, such as the cognitive function and decreased Aβ brain levels [40].
In this work, we used brain slides obtained in our previous study, above mentioned [40], and performed collagen IV staining to assess BM thickness and vessels length in AD/TTR+/− animals, non-treated versus IDIF-treated. As depicted in Fig. 5, AD/TTR+/− mice treated with IDIF presented a significant reduction in the BM thickness in the cortex and hippocampus, and an increase in vessel length in the hippocampus, as compared to non-treated animals. Altogether, these results indicate that TTR stabilization might be a therapeutic target for early treatment in AD.

Discussion
TTR is a homotetrameric protein typically known as a carrier of T4 and retinol in plasma and CSF. During the last years, several functions have been attributed to this protein, in particular, as a neuroprotective protein in physiologic and in disease contexts. In ischemia models, induced by permanent middle cerebral artery occlusion, TTR has been shown to be protective, as evidenced by the significant increase in cortical infarction, cerebral edema and the microglial-leukocyte response in mice with TTR deficiency compared with normal TTR levels [51]. Also, TTR deficiency results in spatial reference memory impairment [52]. Other works showed that TTR promotes nerve regeneration and axonal growth [53,54].
In AD, TTR binds to Aβ preventing its aggregation, accumulation, and toxicity, and facilitating its efflux across the BBB [25]. This barrier is essential to maintain brain homeostasis, however, during normal aging and AD, BBB becomes dysfunctional contributing to disease progression. Molecules known to be important for Aβ brain homeostasis, such as LRP-1 and P-gp are reduced, and TTR was previously shown to increase LRP-1 Fig. 5 Effect of TTR stabilization by IDIF in the thickness of the collagen IV layer and vessel length of AD/TTR+/− mice. A Representative images and quantification plots of collagen IV (red) levels in the BM of cortex and hippocampus vessels derived from 7-month-old AD mice AD/TTR+/mice non-treated (n = 6), or treated with IDIF (IDIF treated, n = 6 for cortex; n = 9 for hippocampus), evidencing a significantly decreased collagen IV layer in treated mice. Scale bar = 5 μm. B Representative images and quantification plots of the length of brain vessels, as evaluated by collagen IV staining, from 7-month-old AD mice AD/TTR+/− non-treated (n = 7) or treated with IDIF (IDIF treated, n = 8) showing a significantly increased vessel length in the hippocampus of IDIF treated compared to non-treated mice. Scale bar = 50 μm. Data are expressed as mean ± SEM. *p < 0.05; ****p < 0.0001 expression in brain ECs and liver [25]. Thus, it is possible that TTR can regulate the neurovasculature in other ways, namely by influencing angiogenesis.
BM thickening through increased collagen IV levels is one such features observed in aging and, more severely, in AD. Previous work by González-Marrero and coworkers described concomitant reduced TTR expression and thickening of the BM at the choroid plexus (CP), in a triple transgenic mouse model of AD [55]. In addition, the authors reported increased Aβ42 in epithelial cytosol and in the stroma surrounding choroidal capillaries [55].
Here, we showed that reduction of TTR expression in an AD mouse model influenced not only the BM, resulting in a thicker collagen IV layer both in the cortex and hippocampus, but also vessel length, resulting in decreased vessel length in the hippocampus. To ascertain if the differences were due to a direct or indirect effect of TTR, collagen IV levels were evaluated also in NT 3month-old mice. NT/TTR+/− mice exhibited more collagen IV around brain microvessels than NT/TTR+/+ littermates, suggesting a direct effect of TTR. Comparing the impact of TTR reduction in the thickness of the BM and in vessel length, in the former we found differences both in the hippocampus and in the cortex, while in the later we only found differences in the hippocampus, suggesting that BM thickening precede the decrease in vessel length.
It is not yet clear what leads to increased collagen IV levels in neurovasculature, but these changes are also found in rats suffering from chronic cerebral hypoperfusion [56,57], suggesting that decreased blood flow in the brain leads to high collagen IV content around the vessels. Indeed, diminished cerebral blood flow is an early impact event during AD development [1]. The thickened and rigid vascular wall may slow down nutrient supply and waste elimination, and possibly disturb perivascular drainage. This event along with the formed barrier will potentially contribute to progressive endothelial dysfunction and to an increasing accumulation of Aβ in the brain.
We questioned if the effect of decreased TTR in the BM and vessel density could be related to TTR participation in angiogenesis, since a number of works implicated TTR in this process, showing that TTR modulates the function of ECs [43,45]. TTR is decreased in AD [26][27][28][29] but its angiogenic potential was never evaluated in vivo. Using the in vivo CAM assay, we demonstrate that WT TTR, both produced recombinantly and purified from human plasma, influences angiogenesis by promoting the formation of new functional vessels.
Previous works reported TTR angiogenic activity studied in vitro using endothelial cells from different organs (retina, lung, placenta), and here we showed that TTR promotes the formation of capillary-like structures by hCMEC/D3 cells. This suggests this activity is not limited to TTR from a specific source, although the protein used has been mainly recombinant, including mouse recombinant TTR [45]. It is noteworthy that our work, although using recombinant human TTR and TTR originated from the human liver, also shows that mice that express about half of the protein present brain vascular dysregulation. Whether this effect is produced by circulating TTR that might enter the brain or by CSF-TTR that reaches other brain areas is unknown and further studies are needed. TTR is also synthesized in the pancreas, retina, and placenta, and this locally produced TTR could explain the angiogenic involvement of TTR in other conditions, namely in neovascularization in DR.
We found that VEGF, Ang-2, IL-6, and IL-8 were significantly upregulated in the presence of the protein.
VEGF is a major driver of angiogenesis, playing a role in most of the steps of the process. Previous works have suggested a link between TTR and VEGF, and for example, elevated VEGF was found in the vitreous of patients with TTR amyloidosis [58]. Another work, proposed an interaction between the two molecules, although reporting that inhibition of VEGF in branch retinal vein occlusion (BRVO) upregulated TTR [59]. This can be interpreted as an attempt of the cells to restore VEGF levels by increasing TTR, thus corroborating our observations in hCMEC/D3 cells.
Ang-2 was previously found to be upregulated in retinal ECs after treatment with TTR [43] and plays a controversial role in angiogenesis. If, on one hand, it increases migration capacity and tube formation in brain ECs [60], on the other hand, in vivo retinal studies showed that Ang-2 promotes EC death and vessel regression if VEGF is absent. However, when in the presence of VEGF it stimulates an increase in capillary diameter, remodeling of basal lamina, proliferation, and migration of EC [61]. These studies support our findings where TTR promotes an increase of both VEGF and Ang-2, which should result in the promotion of angiogenesis.
The observed upregulation of IL-8 and IL-6 is also consistent with an increase in angiogenesis given that IL-8 enhances proliferation, survival, migration, and tube formation [62,63]; and IL-6 was shown to induce an increase in EC proliferation, migration and tube formation [64,65].
The importance of TTR in the pathogenesis of AD is also evident when 3-month-old AD mice are compared to age-matched NT animals, with the same reduction in TTR, showing a thicker BM in the hippocampus of animals with AD. Although this happens prior to amyloid-β deposition, we cannot exclude the presence of other Aβ species that can contribute to this increase, and add to the direct effect of TTR observed in the NT mice.
It is not known if the reduction of TTR in AD is a cause or effect of the disease but it is well known that TTR stability is key for its activity. Mutations in TTR, associated with amyloidosis, create tetrameric instability leading to dissociation into monomers. TTR stabilization seems also important to prevent pathological changes to the brain vasculature, and for example, heterozygous individuals with TTR T119M allele, which renders a more stable tetramer, have a reduced risk of cerebrovascular disease compared to homozygotes for WT TTR [66]. In AD, TTR stability is decreased, leading to accelerated clearance and consequently, to lower levels. We previously showed that TTR stabilization, achieved through the use of small-molecule compounds, sharing structural similarities with T4 and binding in the TTR central channel, results in improved TTR binding to Aβ [19].
One of those small-molecule stabilizers, IDIF, administrated to our AD/TTR+/− mouse model resulted in the amelioration of AD features [39,40]. In this work we showed that IDIF reverted, at least partially, the vascular alterations induced by TTR decrease.
Our results uncover angiogenesis as a mechanism in which TTR participates and importantly, it shows that TTR reduction has an impact in the vascular alterations that occur early in AD with the possibility of recovery upon TTR stabilization.

Limitations
Our work provides positive and new results but it also reveals some limitations that should be mentioned. Concerning the studies with animals, only females were used, while a final conclusion on the effect of TTR decrease regarding the vascular alterations in AD may require the use of both genders. The animal model used in this study shows gender-associated modulation of brain Aβ levels by TTR, and females present a more severe AD-like neuropathology [48], which results in a more favorable scenario to assess the involvement of TTR in AD, explaining why we carried out our experiments in females.

Conclusions
In summary, this work shows that TTR has proangiogenic properties, up-regulating molecules such as IL-6, IL-8, Ang-2, and VEGF. TTR is also involved in the early vascular alterations occurring in AD, which may be used as a target for therapeutic intervention in AD.
Additional file 1: Figure S1. Analyses of the human recombinant TTR produced in the bacterial expression system. A-SDS-PAGE analysis of two different batches (lanes 2 and 3) showing the monomer as the main form. The dimer is also detected at a much lower extent. Lane 1 refers to the protein standards used and the values refer to the respective molecular weight. B-Western blot analysis of the two different batches (lanes 1 and 2) confirms the identity of TTR. C-Native-PAGE analysis of one of the TTR batches, before (lane 1) or after detoxification (lane 2), shows that a single protein form is detected, corresponding to the tetramer Additional file 2: Figure S2. Tube formation by hCMEC/D3 cells. Representative images of tube formation by hCMEC/D3 cells. Cells were plated on Matrigel in the absence (negative control, Ctrl-) or presence (positive control, Ctrl+) of bFGF (35 ng/mL) or with 500 nM recombinant TTR, before (TTR non detoxified) or after detoxification (TTR detoxified). Scale bar = 200 μm. The quantification plot shows that treatment with TTR (either before or after detoxification) results in a significantly higher area covered by the capillary-like structures, than in the negative control. Data are expressed as mean ± SEM. * p < 0.05; ** p < 0.01 Received: 11 May 2021 Accepted: 5 August 2021