GDF11 Rejuvenates Cerebrovascular Structure and Function in an Animal Model of Alzheimer’s Disease

Abstract

Cerebral amyloid angiopathy (CAA) is present in up to 90% of patients with Alzheimer's disease (AD), and may interact with classical neuropathology to exacerbate cognitive decline. Since growth differentiation factor 11 (GDF11) can activate vascular remodeling, we tested its effects on cognitive function and neuroinflammatory changes of AD model mice. We intravenously administered GDF11 or vehicle daily to 12-month-old transgenic mice overexpressing the amyloid-β protein precursor (AβPP)/PS1). Cognitive function was monitored using the Morris water maze, and after conclusion of the treatment, we assessed the morphology and presence of inflammatory markers in the cerebral vasculature. Subchronic treatment of adult AβPP/PS1 mice with GDF11 rescued cognitive function and ameliorated cerebrovascular function. In particular, the de novo genesis of small blood vessels and the expression of vascular-related proteins were significantly higher than in the vehicle-treated AβPP/PS1 mice, whereas the expressions of the inflammatory markers Iba-1 and GFAP significantly decreased in proportion to the lower ratio of two forms of amyloid-β (Aβ_40/42). Daily intravenous treatment with GDF11-injection can rejuvenate respects of cognition and cerebrovascular changes in AD mice.

Keywords

Alzheimer’s disease cerebral amyloid angiopathy cerebral blood flow growth differentiation factor 11 neuroinflammation

INTRODUCTION

Cerebrovascular pathology in Alzheimer’s disease (AD) may interact with parenchymal neurodegenerative changes in a manner aggravating disease progression and cognitive decline [1 –3]. Clinical data indicate significant cerebral amyloid angiopathy (CAA) in postmortem brain from AD patients, which is characterized by pathological amyloid-β (Aβ) deposition on leptomeningeal and large penetrating cortical vessels [4]. CAA is a key cerebrovascular comorbidity, being present in up to 90% of AD patients [5 –8].

CAA results from the failure to eliminate Aβ from the cerebral vasculature [9], ultimately contributing independently to the pervasive cerebrovascular dysfunction in AD. This association of AD and CAA is characterized by impaired neurovascular and metabolic regulation of cerebral blood flow (CBF), in conjunction with aberrations in vascular morphology and density [10, 11]. Furthermore, neuropathological findings in AD patients and AD mouse models strongly implicate inflammation as a significant contributor to disease pathogenesis [12]. Postmortem and in vivo molecular imaging studies concur in showing that inflammatory processes mediated by microglial activation can contribute importantly to the development and progression of cognitive deficits in CAA and AD [13, 14] through a common mechanism arising from impaired Aβ clearance and resultant inflammation [9 , 16].

A recent study suggests that circulating protein GDF11 found in blood of young mammals can induce vascular remodeling and regeneration, while ameliorating age-related impairments in cognitive function and synaptic plasticity in aged mice [17]. Cognitive benefits of infusions of blood from young animals are mediated by GDF11 as well as other cytokines and C-C motif Chemokine 11 (CCL11) [17, 18]. Other recent research has shown that GDF11 can restore immune competency in brain [19]. Hence, these results imply that angiogenesis can play a crucial role in vascular recovery, and raise the promise that restorative treatment with GDF11 and other factors may present a new avenue for intervention against neurodegenerative diseases such as AD, which have a neurovascular component.

Laser speckle flowmetry (LSF) is a non-invasive imaging modality for measuring CBF with high temporal and spatial resolution [20, 21], which furthermore affords repeated longitudinal measurements in the same animal. AD and CAA can independently lead to pronounced cerebrovascular dysfunction, characterized by impaired neurovascular and metabolic regulation of CBF, along with aberrations in vascular morphology and density. As such, LSF can be used to monitor effects of GDF11 treatment on vascular pathology in animal models in preclinical studies aiming to establish possible disease modifying treatments for AD. At present, the relationship between vascular regenerative and pro-cognitive effects of GDF11 has not been clearly established. Therefore, we tested the efficacy of GDF11 treatment as an intervention against cerebrovascular and associated cognitive deficits in adult AβPP/PS1 mice. We hypothesized that the treatment would ameliorate cognition and cerebrovascular function by promoting angiogenesis and relieving the vessel wall burden of Aβ, thus preserving CBF in the target brain region of prefrontal cortex (PFC), while reducing the extent of neuroinflammatory changes in the brain parenchyma.

MATERIALS AND METHODS

Experimental animals

AβPP/PS1 double transgenic mice, B6C3-Tg (APPswe/PS1dE9) 85Dbo/J, were obtained from the Jackson Laboratory (USA). These mice express the Swedish AβPP mutant AβPP as well as PS1, and are characterized by a substantial intracerebral accumulation of Aβ. Mice were raised in individual cages in a 12-h light-dark cycle at constant temperature with free access to food and water. We chose to study (n = 40) 12-month-old Tg mice, based on previous reports of the temporal course of CAA in this model [22 –26], and used (n = 20) age-matched 20 wild-type (Wt) littermates as controls in our study. All experimental animals were be housed in identical conditions so that unexpected side effects on behavioral tests can be avoided. The study was approved by the Shanghai Ethics Committee, and all experiments were performed in accordance with the guidelines from the Chinese Animal Welfare Agency.

GDF11 injection

The AβPP/PS1 Tg mail mice were randomly assigned to one of two groups; the GDF11 group (n = 20) received injections of recombinant GDF11 (rGDF11; Abnova, Taiwan) at a daily dose totaling 0.1 mg/kg in a volume of 0.1 ml, and the control group (n = 20) received daily injections of PBS. Under anesthesia, we injected GDF11 by mouse tail vein injection fixator. The Wt mouse control group did not receive any treatment except to give anesthesia, because the Wt group was normal aging mice, it did not do any traumatic treatment. For injections, mice were anesthetized with 0.25% pentobarbital (Sigma USA; 40 mg/kg, i.p.) for twice daily administration of GDF11or PBS via a tail vein. Each infusion lasted approximately 3 min, and the twice-daily treatment for 28 days.

Behavior assessment

At four weeks after the completion of treatment or start of treatment, we carried out behavioral studies of spatial learning and memory using the Morris water maze (MWM) according to the standard procedures as adapted to mice [27]. The complete MWM apparatus used in our study was purchased from Shanghai Mobile Datum Information Technology Co, Ltd. Swimming paths were recorded by a computerized video imaging analysis system. MWM training comprised two procedures: the place navigation test and the spatial probe test. The place navigation test assessed the learning ability of the mice in the water maze during six successive days to record the escape latency while searching for the platform, to a maximum of 60 s. The spatial probe test to evaluate memory retention was performed on the afternoon of the seventh day. Here, the frequency of crossing over the position of the platform place and the time spent in each of target quadrants were recorded.

Laser speckle contrast imaging procedures and analysis

At baseline and again after four weeks of GDF11 injection, we undertook thinned skull in vivo LSF to determine the CBF on leptomeningeal or superficial cortical vessels according to standard procedures adapted to mice [28]. LSF is a noninvasive imaging technique, which has been widely used to measure the CBF with high temporal and spatial resolution, and lends itself to longitudinal monitoring of CBF in individual animals [20, 21]. We chose the PFC, a brain region critical for cognitive processes, as the region of interest (ROI) [29]. The mice were anesthetized with pentobarbital as above and fixed on a stereotaxic apparatus (David Kopf Instruments, Tujunga, CA, USA). With reference to the atlas Mouse Brain in Stereotaxic Coordinates, CBF recordings were made over the left prefrontal cortex for an ROI measuring 3×4 mm.

Immunohistochemistry/blood vessel density detection

After the LSF test, with animals under deep anesthesia, the brains of mice from each group were removed, fixed, and serially sectioned by a sliding microtome. Fluorescent immunohistochemistry was performed by using specific antibodies to identify Aβ (mouse monoclonal, 1:200; Abcam), CD31, VGEF/VGEFR, (Rabbit polyclonal, 1:300; Abcam), Iba-1, and GFAP (glial-fibrillary acidic protein) (rabbit monoclonal, 1:200; Abcam) as described previously [30, 31]. To analyze the relationship between Aβ deposition and angiogenesis, sections were double-labeled for Aβ and VGEF/VGEFR immunohistochemistry. Other sections were double labeled for Aβ and Iba-1, CD31 or GFAP for identifying the extent of on inflammatory changes in the vascular wall. The nuclei were counter stained with Hoechst 33,342 (Sigma-Aldrich). Images representing either a single confocal Z slice or Z stacks were acquired with a Zeiss 880 confocal system were counted separately by two blind observers. Ten serial sections of 40μm thick with a 10-mm interval of each two adjacent sections from each animal in each group (n = 4) were analyzed in this study. Immunohistochemistry of paraffin-embedded tissue blocks was performed using specific antibodies to identify CD31 (rabbit polyclonal, 1:200; Sigma). The number of positive cells was counted in the prefrontal cortical region under the light microscope (400×). For each section, ten visual fields in PFC were chosen at random for CD31-positive cell counting, which was expressed as mean number of CD31 positive cells per mm².

Reverse transcription and semi-quantitative real-time RT-PCR

Total RNA was isolated from the collected frozen PFC samples using Trizol according to the manufacturer’s instructions (Takara Bio Inc). In brief, 2-3 mg template RNA was used to synthesize the first strand of complementary DNA using a reverse transcription kit purchased from Takara. Real-time PCR of complementary DNA was performed (ABI PRISM 7500 Sequence Detection System; Applied Biosystems, USA) using the forward and reverse primer sequences were as follows: GFAP forward: 5’-CGGAGACGCATCACCTCTG-3’, reverse: 5’-AGGGAGTGGAGGAGTCA-3’ [32]; Iba-1 forward: 5’-TGGGAGTTAGCAAGGGAATG-3’, reverse: 5’-AGACGCTGGTTGTCTTAGGC-3’ [33]; and GAPDH forward: 5’-GGCATGGACTGTGGTCATGA-3’, reverse: 5’-TTCACCACCATGGAGAAGGC-3’ [34].

Data were analyzed using a comparative critical threshold (Ct) method, where the amount of target signal normalized to the endogenous control (GAPDH) and relative to the control sample.

Western blotting

Western blot and the relative protein quantity analysis were performed as previously described [30]. Tissue samples came from the PFC. In brief, cortical tissues were homogenized in RIPA buffer (Sigma–Aldrich) in the presence of protease inhibitors. Total protein concentration was determined using the bicinchoninic acid assay (BCA). Equal amounts of the protein samples were loaded for each lane. The primary antibodies were rabbit polyclonal antibody against VGEF/VGEFR, Iba-1, GFAP, Aβ₄₂, Basement membrane layer IV collagen antigen (Collagen IV), laminin (1:500-1000; Abcam, USA), and β-actin/GAPDH (1:10,000, Sigma). Western blots were quantified with horseradish peroxidase-conjugated immunoglobulin G, with densitometric analysis of the bands using a KS400 image analysis system (version 3.0, Karl Zeiss, Germany).

Enzyme-linked immunosorbent assay (ELISA)

The levels of soluble and insoluble Aβ (i.e., the ratio of Aβ_40/42) in the brain of AD mice were quantified according to procedures described previously [14]. In brief, the supernatant fraction was collected analysis of the soluble Aβ ELISA. Tissue samples came from the PFC. The remaining sodium dodecyl sulfate-insoluble pellet was sonicated, dissolved in 70% formic acid, and re-centrifuged. The supernatant was separated for the insoluble Aβ ELISA. The total concentrations of Aβ determined in each sample using commercially available quantitative sandwich ELISAs (Immuno-Biological Laboratories, Co., Ltd., Gunma, Japan) according to the manufacturer’s instructions, with absorbance at 450 nm measured using an ELISA reader (Multiskan EX; Labsystem, Helsinki, Finland). The analyses were always performed in duplicate and in a coded manner to ensure that the operator was blind to the characteristics of the animals.

Ultrastructure

Anesthesia, brain tissue of PFC collection, and electron microscopy were performed as previously described [35]. A portion of the left prefrontal cortex measuring approximately 1×1×1 mm was trimmed under a microscope and processed for ultrastructural analysis using transmission electron microscopy (Philips CM 120; The Netherlands). The integrity of the vascular structures was noted, and the extent of Aβ plaque deposition in vessel walls was noted after photography at magnifications of 4,000–46,000 X.

Statistical analysis

All data were normally distributed and expressed as mean±standard deviation (S.D.). For multiple mean comparisons, the data were analyzed by one-way ANOVA and LSD post-test, followed by Bonferroni multiple comparison correction using standard statistics software (SPSS, 14.0). p < 0.05 was considered as a statistically significant difference.

RESULTS

GDF11 injection rescued the cognitive deficits in AD mice

Previous studies have shown an increase in Aβ plaque deposition and impairment in learning and memory abilities with aging in AβPP/PS1 mice [36]. We first used the MWM test to test if GDF11 treatment improved or rescued spatial learning and memory deficits in our AβPP/PS1 mice. As expected, after four weeks of GDF11 treatment, the spatial learning and memory scores in the Tg- GDF11 mouse group were significantly better than for the Tg-PBS mice. However, the time spent in the original platform quadrant and the number of times crossing the original platform remained lower for Tg-GDF11 mice than corresponding scores for untreated Wt mice (n = 20, p < 0.05, Fig. 1). Taken together, these results indicate that GDF11 treatment can rescue impaired spatial learning and memory deficits commonly seen in adult AβPP/PS1 mice.

Fig.1

GDF11 treatment rescued cognitive deficits in AD mice. Four weeks after GDF11treatment, the spatial learning and memory were measured by the Morris water maze test. A) In the place navigation test, GDF11-treated AβPP/PS1 (Tg-GDF11) mice exhibited significantly shorter escape latency with days 3–6 of training than did Tg-PBS mice (n = 10, p < 0.001). These latencies were similar to those of wild-type (Wt) mice (n = 10, p < 0.05). B) In the spatial probe test, Tg-GDF11 mice spent a significantly longer time in the target quadrant than Tg-PBS mice (n = 10, p < 0.01), and similar to Wt mice (n = 10, p > 0.05). C) Tg- GDF11 mice crossed the platform significantly more often than Tg-PBS mice (n = 10, p < 0.01), matching the crossing rate for Wt mice (n = 10, p > 0.05). Significance of difference: p < 0.001, ^#p < 0.01 versus Wt.

GDF11 treatment rescued CBF in cerebral cortex

At four weeks after GDF11 injection, we saw a conspicuous rescue of cerebrovascular function, with CBF approximately 87% of that seen in at baseline in the Wt mice (Fig. 2). Compared with Tg-PBS mice, the GDF11-treatment group showed nearly 1.6-fold increase of CBF in the cortical region of interest relative to baseline (Fig. 2), thus attaining CBF 87% of that in Wt mice. In contrast, the CBF in Wt mice did not change to four weeks follow-up.

Fig.2

GDF11 treatment restored CBF in the prefrontal cortex of AD model mice 4 weeks after the injection. Quantification of the relative CBF in ROI (% of baseline for each animal). n = 6 for each group. ^*p < 0.05, Tg-GDF11 versus Wt mice; and ^#p < 0.01, Tg-GDF11 versus Tg-AD mice.

GDF11 injection into AD mice reduced cerebral vascular Aβ burden in AD mice

Aβ peptide plays a key pathogenic role in AD. We saw elevated Aβ deposition in vessel walls and narrowing of the vascular lumen in PFC regions of the untreated Tg-mice (Fig. 3), whereas GDF11 treatment reduced the vascular Aβ burden in cerebral of adult AD mice. We found that GDF11 treatment did not change the total level of Aβ₄₂ to western blot analysis and ELISA (n = 4, p > 0.05), but significantly reduced the ratio of Aβ₄₀/Aβ₄₂ in brain tissue to ELISA (n = 4, p < 0.01). Meanwhile, GDF11-treatment stimulated VGEF expression in vessel walls, and reduced plaque and activated the expression of Iba+ cells (Fig. 3). Taken together, these data indicated that GDF11 injection can reduced the amount of soluble Aβ and reduces the deposition of Aβ plaques in the vascular wall, and further improved the vascular function in AD mice.

Fig.3

GDF11 treatment of AD model mice reduced cerebral vascular Aβ. ELISAs for soluble and insoluble levels of Aβ₄₀ and Aβ₄₂ in the brain and plasma reveal significantly lower Aβ₄₀/Aβ₄₂ ratio in brain tissue (A) and plasma (B) of GDF11-treated mice (n = 4, p < 0.01). Aβ₄₀/Aβ₄₂ ratio changes in plasma (B) were similar to those in brain, although the total insoluble Aβ₄₂ level AD mice was unaltered to western blot of prefrontal cortex tissue (n = 4, p > 0.05) (C, D). E-H) Abundant plaque deposition was evident surrounding cortical blood vessels in AD mice, in association with significant narrowing of the lumen (I). J, K) GDF11-treatment stimulated VGEF expression in vessel walls, and reduced plaque (Red) and the density of Iba+ microglia (L).

Reduce vascular wall surrounding inflammatory factor, promote new blood vessels to formation and increase vascular density

We undertook various immunostaining to observe the responses of markers for brain endothelial cells (CD31), microglia (Iba1) and astrocytes (GFAP) and Aβ deposition in the PFC region to treatment of AD mice with GDF11. We saw conspicuously lower expression of Iba1 and GFAP in the GDF11-treated mice compared to the Tg-PBS group (Fig. 4). This figure also illustrates that Aβ load to immunofluorescence was similar in the GDF11 and PBS treated Tg mice, as was likewise the total Aβ protein to western blot (n = 4, p > 0.05). Nonetheless, glial activation markers were still higher than in the Wt mice (Fig. 4). There was no evidence of activated glia or Aβ plaques in brain of the Wt mice. Relatively lower gene expression and protein levels of Iba-1 and GFAP were observed in the prefrontal cortical region of GDF11 mice compared with Tg-PBS mice (n = 4, p < 0.01, p < 0.001), but these markers were nonetheless high than in the Wt mice (n = 4, p < 0.05).

Fig.4

GDF11 treatment reduced vascular wall and perivascular inflammatory markers, promoted new blood vessel formation, and increase vascular density. In the GDF11-treated group, abundant blood vessels (CD31+endothelial cells) were found in the prefrontal cortex (A), in contrast to the sparse distribution seen in untreated AD mice (Tg-PBS), but similar to density in Wt mice. B) Quantification of the CD31+ cells showed that there was significantly greater microvessels density in the GDF11 treated mice. n = 6 for each group. ^**p < 0.01, versus the AD mice; there was no difference between Tg-GDF11 mice and Wt mice.

We also quantified angiopoietins levels by western blotting. We found that GDF11 treatment enhanced the expression of VGEF/VGEFR antigens relative to finding in to Tg-PBS mice (n = 4, p < 0.01), and there was no significant difference between the protein levels in GDF11-treated Tg and Wt animals (n = 4, p > 0.05) (Figs. 5 and 6).

Fig.5

GDF11 treatment decreased mRNA and protein levels of GFAP and Iba-1 in AD mice. A, B) Quantification of the total mRNA levels of GFAP and Iba-1 in the prefrontal cortex (PFC) regions by real-time PCR analysis (n = 4 per group). C) Representative western blot images of relevant proteins in the PFC at 4 weeks after GDF11 treatment. D) Quantitation of the expression of GFAP and Iba-1 protein expression in PFC regions by western blotting. The data, expressed as the ratio GFAP/Iba-1/β-actin, represent the mean±standard deviation of three separate experiments (n = 4 per group). GDF11-treated Tg mice versus Wt group ^*p < 0.05, ^#p < 0.01 versus the group of Tg-GDF11 or Wt mice. F-H, O-Q) Representative co-expression of VGEF and Iba-1 in Tg-GDF11 mice, and in Wt mice (I-K) and representative co-expression of VGEF and Iba-1 in AD mice (L-N). R, S) Double immunofluorescence analysis shows more activation and localization of GFAP and CD31 in vessel walls in PFC from AD model mice. Expression of Iba1 and GFAP were conspicuously decreased in GDF11-treated mice (S) compared to Tg-PBS mice (R).

Fig.6

GDF11-treatment enhanced the expression of angiopoietin-related proteins. A) At 4 weeks post-treatment, there was increased expression of VGEF (green). Aβ positive regions are also indicated (red). Untreated AD mice had more Aβ plaque deposition in the vascular wall and reduced expression of VEGF were found, whereas the treatment VGEF expression (green) in vessel walls, stimulating some new small blood vessels without Aβ (red) was observed (A). In Wt mice, Aβ plaques were absent, and there was high expression of VGEF in vessel walls. Scale bars: 30μm in Tg-GDF11 mice, 25μm in Tg-PBS mice, 80μm in Wt mice. B) Quantified angiopoietins levels by western blot. VGEF/VGEFR antigens at four weeks after GDF11 treatment compared to Tg-PBS mice (n = 4, ^**p < 0.01). There was no significant difference in protein levels between Tg-GDF11 and Wt animals (n = 4, p > 0.05), but there were conspicuous increases in collagen IV and laminin antigenicity in PFC of AD mice with GDF11 treatment compared to Tg-PBS mice (n = 4, & p < 0.01), and similar to Wt mice (n = 4, p > 0.05).

We had hypothesized that the expression of basement membrane related proteins would be increased by GDF11-treatment. As predicted, we found conspicuously higher collagen IV and laminin antigen levels in the GDF11-treated group compared to the PBS-treated AD mice (n = 4, p < 0.01), which were lower than in the Wt mice (n = 4, p < 0.05) (Fig. 6).

Ultrastructure showed more normal vessels in the cerebral regions in AD mice

We found that the number of new normal vessels was significantly higher in the PFC of GDF11 mice compared to Tg-PBS mice (Fig. 7). That figure also shows the abundance of morphologically normal vessels in the Tg-GDF11 mice, with integral vasculature, wide lumen, and closely connected endothelial cells. In contrast, the integrity of vessels was comprised in the Tg-PBS mice, which showed uneven thickening of the wall of small vessels. Ground-glass opacity amyloid and evidence of Aβ aggregation were observed in tunica media, meanwhile, coarse and damaged in intima were also found, these pathologic changes suggested involvement of the inter membrane and tunica media, but in the absence of luminal stenosis. The untreated Tg mice showed organelles damaged in endothelial cells, such as mitochondria swollen, mitochondrial crest, and evidence of glial cell proliferation in the surrounding wall (Fig. 7). Wt mice showed closely connected vascular endothelial cells, integral basement membrane, and uniform wall thickness (Fig. 7).

Fig.7

Electron microscopy showed more normal microvascular structure in the prefrontal cortex of AD model mice four weeks after GDF11 treatment. After GDF11 injection, abundance of morphologically normal vessels was observed in the Tg-GDF11 mice, with integral vasculature and closely connected endothelial cells (A-C), normal mitochondrion were found in Tg-GDF11 mice (red arrow), and Aβ plaques also were observed in the vascular wall (A, C, black arrow). The Tg-PBS mice had uneven thickening of small blood vessel walls, mainly involving the inter membrane (E, red thin arrow), and meanwhile Aβ accumulation also was found in the vascular wall of Tg-PBS mice (F, black thin arrow). Wt mice likewise had connected vascular endothelial cells, integrated basement membrane, and uniform walls (G). Scale bars: 2μm (A, D, F), 2.0μm (C), 4.0μm (B, F), and 1μm (G).

DISCUSSION

Our study demonstrated that intravenous administration of rGDF11 enhances angiogenesis in Tg mice, as evidenced by increased expression of VGEF/VGEFR, CD31, Collagen IV, and laminin; these mice also showed decreased glial activation accompanied by a decrease in the ratio of Aβ_40/42 in brain parenchyma and lower expression of inflammation markers in the vasculature. This rescue of vascular function was associated with improved cognition in tests of spatial memory relative to untreated AD mice.

CAA is a major factor leading to impaired CBF in AD, which can be attributed to a defect in the clearance of Aβ from vascular tissue. Indeed, autopsy examinations show that some 80% of AD patients also manifest CAA [9]. Studies in transgenic animals show widespread Aβ plaques, notably impinging on brain microvasculature [5, 6]. The factors mediating these association have been unclear, but recent work has indicated that the circulating protein GDF11 is a rejuvenating factor for skeletal muscle, and can also improve cerebral vascular function and enhance neurogenesis in aging mice [18]. Based on these findings, we predict that GDF11 holds promise as a disease modifying treatment for AD via rejuvenation of cerebrovascular function.

The integrities of cerebrovascular architecture, capillary density, and CBF have all been reported to decline with aging [10, 11]. As such, we explored the potential GDF11 treatment as a moderator of CAA mechanisms in an established mouse model of AD. Neuronal loss is pronounced in PFC of AD patients [37], and functional/anatomic cerebrovascular impairments in PFC correlated with cognitive deficits and behavioral abnormalities in AD patients [38]. Given this background, we chose the PFC as the region of interest for analyzing the effects of GDF11-treatment on CAA-related pathologies in AD mice.

Consistent with our hypothesis, we found that with GDF11 treatment, CBF was conspicuously rejuvenated in PFC of adult AβPP/PS1 mice at four weeks after GDF11 injection. The expression of the vascular neurogenesis markers VGEF/VGEFR was increased in GDF11-treated mice relative to the PBS vehicle group. Another report has shown that the cytokine exercise brain protection factor increased the expression of VEGF, VEGF receptors, and angiopoietin receptor, while improving perfusion of brain tissue [39]. Hence, we consider that GDF11 activates the VEGFR signaling pathway, leading to endothelial cell growth and improved perfusion.

We found that the expression of the endothelial cell marker CD31 was increased significantly at weeks after GDF11 injection. Angiogenesis is a process involving proliferation and sprouting of endothelial cells and their subsequent formation of new vessels, which plays a critical role in functional recovery from brain insults such as stroke and traumatic brain injury [40, 41]. Teng et al. demonstrated that endothelial cells from ischemic brain tissue stimulate the proliferation and differentiation of neural stem cells in vitro, apparently via signals from the pro-angiogenesis factor VEGF and the chemokine stromal derived factor 1α secreted from endothelial cells [42]. Our present study demonstrates that GDF11 treatment stimulated the proliferation of endothelial cells and improved the integrity of the vascular basement membrane in brain of AD mice. We suppose that these structural changes underlie the rescue of CBF in these mice.

The brain microvascular basement membrane plays an important role in the maintenance of cerebral microvascular integrity. The basement membrane constituents’ laminin and collagen IV are well-established markers of cerebrovascular structures [43, 44]. Therefore, we hypothesized that GDF11 treatment would increase the expression of these markers in AD mouse brain. As predicted, we saw conspicuously higher collagen IV and laminin antigen levels in the group of AD mice with GDF11 treatment, along with an increase in microvessel density. This observation is consistent with new blood vessel formation, rescuing the CBF deficit in AD mice arising from Aβ plaque accumulation.

In an earlier report, enhanced angiogenesis not only increased CBF in the PFC region, but also stimulated neurogenesis, both of which contributed to functional recovery in AD mice [42]. The rectification of the CBF defect is indicative for cerebrovascular functional recovery, which may underlie the rescue of cognitive function in AD mice. We suppose that angiogenesis due to the GDF11-induced elevation in VEGF and VEGFR is a factor in the pro-cognitive effects seen in AβPP/PS1 mice.

The total Aβ level, as well as the ratio of β- and γ-secretase products Aβ₄₀ and Aβ₄₂ (Aβ₄₀/Aβ₄₂), are factors determining both onset and the severity of CAA [4 , 45–47]. AβPPDutch animals displaying a very high Aβ₄₀/Aβ₄₂ ratio develop severe CAA, indicating that the majority of vascular amyloid deposits are composed of Aβ₄₀ [4]. While Aβ₄₂ might be essential for the initiation of vascular amyloid deposition, an increase of total Aβ and higher Aβ₄₀/Aβ₄₂ ratio favors the progression of vascular Aβ load [48]. We find that GDF11 treatment relieves the Aβ burden in microvessel walls, and attenuates the progression of the present AD model. These benefits were accompanied with a decreased Aβ_40/42 ratio in brain tissues and blood plasma, without a reduction in the total Aβ level.

We now report that GDF11 treatment attenuated the activation and proliferation of glial cells, in association with reduced gene and protein expression of the glial markers Iba-1 and GFAP. A growing body of evidence suggests that there is a close association between neuroinflammatory reactions and amyloidogenesis in AD and in animal models [49]. For example, AβPP/PS1 transgenic mice have an excess of activated glia in proximity to Aβ plaques [34]. We suppose that the present GDF11 treatment had primary effects on amyloid clearance across cerebral microvessels, which propagated to attenuated cerebral plaque deposition, and reduced local inflammation.

In our hands, GDF11-treatment of AD mice restored CBF to the range for Wt mice, indicating that the vascular remodeling manifests in improved cerebral hemodynamics. Meanwhile, we saw increased density of microvessels in the PFC region. We suppose that the changes in laminin and Collagen IV expression, as well as basement membrane changes, are relevant to the observed improvements in cerebral perfusion, all arising from attenuated intravascular Aβ accumulation. In particular, we find that GDF11-treatment has primary benefits for tunica media of vascular lesions, with subsequent effects on the intima, ultimately bringing about sparing from luminal stenosis.

We saw evidence for small blood vessel wall thickening in untreated Tg mice, this in association with perivascular gliosis. However, GDF11 treatment rescued these changes. In general, new vessels can form either by sprouting from existing capillaries or form de novo from circulating endothelial progenitors. Blood-born factors also influence angiogenesis: Serum from young animals stimulated endothelial cell proliferation in vivo, but serum from older animals was less effective [10]. Findings of that study confirmed that GDF11 has a direct biological effect on endothelial cells through the p-SMAD pathway, ultimately increasing CBF and angiogenesis in aged mice. Attribution of causality is difficult, but we suppose that the pro-cognitive effect of GDF11 treatment seen in the present study of AD mice may be secondary to improved perfusion, resulting in better substrate delivery to brain. More importantly, although we still found Aβ aggregation in tunica media after GDF11 treatment, but intima presented normal structure, however, the untreated Tg mice showed coarse and damaged in intima and organelles damaged in endothelial cells, such as mitochondria swollen, mitochondrial crest. These data indicate that GDF11 injection can play an important role in supplementing functional mitochondrial and further alleviation of the course of CAA. The mitochondria are key organelles responsible for producing adenosine triphosphate for normal function and also serve a role in mediating specific cell death pathways [50]. We propose that injected GDF11 play an important role in rescuing functional mitochondrial of vascular endothelial cell land further alleviation of the course of AD to some extent. Since proliferation of endothelial cells itself promotes vascular remodeling [51, 52]. Thus, we contend that the DGF11 treatment initiates new angiogenesis and vascular remodeling, thus supporting cognitive improvement. This concept warrants further study, since increasing the angiogenesis with GDF11 would be a potential disease modifying intervention strategy for AD.

Conclusion

In our study, we show that GDF11 treatment mice significantly rejuvenated cerebrovascular function in AβPP/PS1. Our other findings suggest GDF11 stimulated angiogenesis by upregulation of VGEF and its receptor, and alleviated Aβ burden in vessel walls in association with decreased inflammatory markers in brain. These results are promising for the translational testing of GDF11 as a restorative treatment of CAA occurring in association with AD.

Footnotes

ACKNOWLEDGMENTS

This research was supported by The Shanghai Municipal Science and Technology Commission medical guide project No.134119b2100, 16411969200; Shanghai Municipal Commission of Health and Family Planning project No. 201640035).

Authors’ disclosures available online ().

References

Hofman

, Ott

, Breteler

, Bots

, Slooter

, van Harskamp

, van Duijn

, Van Broeckhoven

, Grobbee

(1997) Atherosclerosis, apolipoprotein E, and prevalence of dementia and Alzheimer’s disease in the Rotterdam Study, Lancet 18, 151–154.

Luchsinger

, Reitz

, Honig

, Tang

, Shea

, Mayeux

(2005) Aggregation of vascular risk factors and risk of incident Alzheimer disease, Neurology 23, 545–551.

Hayden

, Zandi

, Lyketsos

, Khachaturian

, Bastian

, Charoonruk

, Tschanz

, Norton

, Pieper

, Munger

, Breitner

, Welsh-Bohmer

; Cache County Investigators (2006) Vascular risk factors for incident Alzheimer disease and vascular dementia: The Cache County study, Alzheimer Dis Assoc Disord 20, 93–100.

Herzig

, Winkler

, Burgermeister

, Pfeifer

, Kohler

, Schmidt

, Danner

, Abramowski

, Stürchler-Pierrat

, Bürki

, van Duinen

, Maat-Schieman

, Staufenbiel

, Mathews

, Jucker

(2004) Abeta is targeted to the vasculature in a mouse model of hereditary cerebral hemorrhage with amyloidosis, Nat Neurosci 7, 954–960.

Vinters

(1987) Cerebral amyloid angiopathy. A critical review, Stroke 18, 311–324.

Jellinger

(2002) Alzheimer disease and cerebrovascular pathology: An update, J Neural Transm 109, 813–836.

Breteler

(2000) Vascular risk factors for Alzheimer’s disease: An epidemiologic perspective, Neurobiol Aging 21, 153–160.

Breteler

(2000) Vascular involvement in cognitive decline and dementia. Epidemiologic evidence from the Rotterdam Study and the Rotterdam Scan Study, Ann N Y Acad Sci 903, 457–465.

Weller

, Boche

, Nicoll

(2009) Microvasculature changes and cerebral amyloid angiopathy in Alzheimer’s disease and their potential impact on therapy, Acta Neuropathol 118, 87–102.

10.

Snowdon

(2003) Nun Study. Healthy aging and dementia: Findings from the Nun Study, Ann Intern Med 2, 450–454.

11.

Cordonnier

, van der Flier

(2011) Brain microbleeds and Alzheimer’s disease: Innocent observation or key player? Brain 134, 335–344.

12.

Choo

, Alukaidey

, White

, Grubman

(2013) Neuroinflammation and copper in Alzheimer’s disease, Int J Alzheimers Dis 2013, 145345.

13.

Wirths

, Breyhan

, Marcello

, Cotel

, Brück

, Bayer

(2010) Inflammatory changes are tightly associated with neurodegeneration in the brain and spinal cord of the APP/PS1KI mouse model of Alzheimer’s disease, Neurobiol Aging 31, 747–757.

14.

Zhang

, Stafford

, Wang

, Arnold

, Wolk

, Detre

(2012) Microvascular perfusion based on arterial spin labeled perfusion MRI as a measure of vascular risk inAlzheimer’s disease, J Alzheimers Dis 32, 677–687.

15.

Weller

, Subash

, Preston

, Mazanti

, Carare

(2008) Perivascular drainage of amyloid-beta peptides from the brain and its failure in cerebral amyloid angiopathy and Alzheimer’s disease, Brain Pathol 18, 253–266.

16.

Carare

, Hawkes

, Weller

(2014) Afferent and efferent immunological pathways of the brain. Anatomy, function and failure, Brain Behav Immun 36, 9–14.

17.

Villeda

, Plambeck

, Middeldorp

, Castellano

, Mosher

, Luo

, Smith

, Bieri

, Lin

, Berdnik

, Wabl

, Udeochu

, Wheatley

, Zou

, Simmons

, Xie

, Longo

, Wyss-Coray

(2014) Young blood reverses age-related impairments in cognitive function and synaptic plasticity in mice, Nat Med 20, 659–663.

18.

Katsimpardi

, Litterman

, Schein

, Miller

, Loffredo

, Wojtkiewicz

, Chen

, Lee

, Wagers

, Rubin

(2014) Vascular and neurogenic rejuvenation of the aging mouse brain by young systemic factors, Science 9, 630–634.

19.

Wang

, Shen

, Li

, Chen

, Wang

, Qin

, Zhou

, Peng

, Feng

, Li

, Liang

(2015) Splenocytes derived from young WT mice prevent AD progression in APPswe/PSENldE9 transgenic mice, Oncotarget 28, 20851–20862.

20.

Strong

, Bezzina

, Anderson

, Boutelle

, Hopwood

, Dunn

(2006) Evaluation of laser speckle flowmetry for imaging cortical perfusion in experimental stroke studies: Quantitation of perfusion and detection of peri-infarct depolarisations, J Cereb Blood Flow Metab 26, 645–653.

21.

Boas

, Dunn

(2010) Laser speckle contrast imaging in biomedical optics, J Biomed Opt 15, 011109.

22.

Jankowsky

, Slunt

, Ratovitski

, Jenkins

, Copeland

, Borchelt

(2001) Co-expression of multiple transgenes in mouse CNS: A comparison of strategies, Biomol Eng 17, 157–165.

23.

Savonenko

, Xu

, Melnikova

, Morton

, Gonzales

, Wong

, Price

, Tang

, Markowska

, Borchelt

(2005) Episodic-like memory deficits in the APPswe/PS1dE9 mouse model of Alzheimer’s disease: Relationships to beta-amyloid deposition and neurotransmitter abnormalities, Neurobiol Dis 18, 602–617.

24.

Garcia-Alloza

, Robbins

, Zhang-Nunes

, Purcell

, Betensky

, Raju

, Prada

, Greenberg

, Bacskai

, Frosch

(2006) Characterization of amyloid deposition in the APPswe/PS1dE9 mouse model of Alzheimer disease, Neurobiol Dis 24, 516–524.

25.

O’Leary

, Brown

(2009) Visuo-spatial learning and memory deficits on the Barnes maze in the 16-month-old APPswe/PS1dE9 mouse model of Alzheimer’s disease, Behav Brain Res 19, 120–127.

26.

Stover

, Brown

(2012) Age-related changes in visual acuity, learning and memory in the APPswe/PS1dE9 mouse model of Alzheimer’s disease, Behav Brain Res 16, 75–85.

27.

Billings

, Oddo

, Green

, McGaugh

, LaFerla

(2005) Intraneuronal Abeta causes the onset of early Alzheimer’s disease-related cognitive deficits in transgenic mice, Neuron 3, 675–688.

28.

Zhang

, Yu

, Zhou

, Zhang

, Wu

, Zhang

, Bai

, Jia

, Zhang

, Tian

, Wu

, Hu

(2013) Early exercise improves cerebral blood flow through increased angiogenesis in experimental stroke rat model, J Neuroeng Rehabil 26, 43.

29.

Arnsten

, Li

(2005) Neurobiology of executive functions: Catecholamine influences on prefrontal cortical functions, Biol Psychiatry 57, 1377–1384.

30.

Zhang

, Wu

, Wang

, Gu

, Zhang

, Xia

(2018) Neural stem cell transplantation decreases neuroinflammation in a transgenic mouse model of Alzheimer’s disease, J Neurochem 136, 815–825.

31.

Zhang

, Gu

, Shen

, Zhang

, Wang

(2015) Neural stem cell transplantation enhances mitochondrial biogenesis in a transgenic mouse model of Alzheimer’s disease-like pathology, Neurobiol Aging 36, 1282–1292.

32.

Shinjyo

, Di Marzo

(2013) The effect of cannabichromene on adult neural stem/progenitor cells, Neurochem Int 63, 432–437.

33.

Luo

, Wang

, Li

, Wei

, Zhang

, Liu

, Wang

, Niu

, Wang

(2014) Maternal bisphenol a diet induces anxiety-like behavior in female juvenile with neuroimmune activation, Toxicol Sci 1, 364–373.

34.

Zhang

, Wu

, Wang

, Gu

, Zhang

, Xia

(2016) Neural stem cell trans plantation decreases neuroinflammation in a transgenic mouse model of Alzheimer’s disease, J Neurochem 136, 815–825.

35.

Zhang

, Wang

, Sha

, Ni

, Li

, Gu

(2014) Neural stem cell transplants improve cognitive function without altering amyloid pathology in an APP/PS1 double transgenic model of Alzheimer’s disease, Mol Neurobiol 50, 423–437.

36.

Marjanska

, Curran

, Wengenack

, Henry

, Bliss

, Poduslo

, Jack

Jr , Ugurbil

, Garwood

(2005) Monitoring disease progression in transgenic mouse models of Alzheimer’s disease with proton magnetic resonance spectroscopy, Proc Natl Acad Sci U S A 16, 11906–11910.

37.

Salat

, Kaye

, Janowsky

(2001) Selective preservation and degeneration within the prefrontal cortex in aging and Alzheimer disease, Arch Neurol 58, 1403–1408.

38.

Arnsten

, Li

(2005) Neurobiology of executive functions: Catecholamine influences on prefrontal cortical functions, Biol Psychiatry 1, 1377–1384.

39.

Cheng

, Kuzuya

, Kim

, Song

, Hu

, Inoue

, Nakamura

, Di

, Sasaki

, Tsuzuki

, Shi

, Okumura

, Murohara

(2010) Exercise training stimulates ischemia-induced neovascularization via phosphatidylinositol 3-kinase/Akt-dependent hypoxia-induced factor-1 alpha reactivation in mice of advanced age, Circulation 17, 707–716.

40.

Guo

, Liu

, Zhang

, Bergeron

, Cui

, Dong

, Zhang

(2009) Correlation of CD34+ cells with tissue angiogenesis after traumatic brain injury in a rat model, J Neurotrauma 26, 1337–1344.

41.

Morgan

, Kreipke

, Roberts

, Bagchi

, Rafols

(2009) Neovascularization following traumatic brain injury: Possible evidence for both angiogenesis and vasculogenesis, Neurol Res 29, 375–381.

42.

Teng

, Zhang

, Wang

, Zhang

, Morris

, Gregg

, Wu

, Jiang

, Lu

, Zlokovic

, Chopp

(2008) Coupling of angiogenesis and neurogenesis in cultured endothelial cells and neural progenitor cells after stroke, J Cereb Blood Flow Metab 28, 764–771.

43.

Moser

, Schmidt-Kastner

, Hinterhuber

, Humpel

(2003) Brain capillaries and cholinergic neurons persist in organotypic brain slices in the absence of blood flow, Eur J Neurosci 18, 85–94.

44.

Moser

, Reindl

, Blasig

, Humpel

(2004) Brain capillary endothelial cells proliferate in response to NGF, express NGF receptors and secrete NGF after inflammation, Brain Res 13, 53–60.

45.

Herzig

, Paganetti

, Staufenbiel

, Jucker

(2007) BACE1 and mutated presenilin-1 differently modulate Abeta40 and Abeta42 levels and cerebral amyloidosis in APPDutch transgenic mice, Neurodegener Dis 4, 127–135.

46.

Herzig

MC1

, Van Nostrand

, Jucker

(2006) Mechanism of cerebral beta-amyloid angiopathy: Murine and cellular models, Brain Pathol 16, 40–54.

47.

Kumar-Singh

(2009) Hereditary and sporadic forms of Abeta-cerebrovascular amyloidosis and relevant transgenic mouse models, Int J Mol Sci 23, 1872–1895.

48.

McGowan

, Pickford

, Kim

, Onstead

, Eriksen

, Yu

, Skipper

, Murphy

, Beard

, Das

, Jansen

, Delucia

, Lin

, Dolios

, Wang

, Eckman

, Dickson

, Hutton

, Hardy

, Golde

(2005) Abeta42 is essential for parenchymal and vascular amyloid deposition in mice, Neuron 21, 191–199.

49.

Giannoni

, Arango-Lievano

, Neves

, Rousset

, Baranger

, Rivera

, Jeanneteau

, Claeysen

, Marchi

(2016) Cerebrovascular pathology during the progression of experimental Alzheimer’s disease, Neurobiol Dis 88, 107–117.

50.

Onyango

, Lu

, Rodova

, Lezi

, Crafter

, Swerdlow

(2010) Regulation of neuron mitochondrial biogenesis and relevance to brain health, Biochim Biophys Acta 1802, 228–234.

51.

Hutter-Schmid

, Kniewallner

, Humpel

(2015) Organotypic brain slice cultures as a model to study angiogenesis of brain vessels, Front Cell Dev Biol 2, 52.

52.

Morin-Brureau

, Lebrun

, Rousset

, Fagni

, Bockaert

, de Bock

, Lerner-Natoli

(2011) Epileptiform activity induces vascular remodeling and zonula occludens 1 downregulation in organotypic hippocampal cultures: Role of VEGF signaling pathways, J Neurosci 20, 10677–10688.