Age-dependent dysfunction of the cerebrovascular system in the zebrafish telencephalon

The brain is an essential organ that controls various biological activities via the nervous system. The cerebral blood vessels supply oxygen and nutrients to neuronal cells and carry away waste products, which is essential in maintaining brain functions. Aging affects cerebral vascular function and decreases brain function. However, the physiological process of age-dependent cerebral vascular dysfunction is not fully understood. In this study, we examined aging effects on cerebral vascular patterning, vascular function, and learning ability in adult zebrafish. We found that the tortuosity of the blood vessels was increased, and the blood flow rate was reduced with aging in the zebrafish dorsal telencephalon. Moreover, we found cerebral blood flow positively correlated with learning ability in middle-old-aged zebrafish, as in aged humans. In addition, we also found that the elastin fiber decreased in the middle-old-aged fish brain vessel, suggesting a possible molecular mechanism underlying vessel dysfunction. Therefore, adult zebrafish may serve as a useful model for studying the aging-dependent decline in vascular function and human diseases such as vascular dementia.


Introduction
The brain is an organ that plays a central role in the nervous system and has important functions such as memory, learning, and sensations, and regulating breathing and the heartbeat. The cerebral blood vessels that line the brain supply oxygen and nutrients to neurons and glial cells and carry away waste products, thus playing an important role in maintaining brain functions. Therefore, a decline in cerebrovascular function significantly impacts brain function (Wardlaw et al., 2021).
Aging is closely related to cerebrovascular function. Resting hippocampal blood flow decreases with aging (Heo et al., 2010), and resting hippocampal blood flow is positively correlated with spatial memory performance in aged people (Heo et al., 2010;Yang et al., 2017). In addition, aging is one of the risk factors for developing cerebrovascular dysfunction, and this might lead to vascular dysfunction-dependent diseases such as vascular dementia and lacunar infarction (Fulop et al., 2019;Rundek et al., 2022;Torres-Simón et al., 2022;Yang et al., 2017). However, the physiological process of age-dependent cerebral vascular dysfunction is not fully understood. Zebrafish are useful model animals for analyzing basic biological phenomena and human diseases at the molecular level (Dooley and Zon, 2000). In a previous study, we showed that learning ability declined in aged fish (Yang et al., 2018). This result suggests that aging may affect the cerebral vascular system and decrease brain function in zebrafish. Several transgenic lines were developed to facilitate imaging of the blood vessel network and blood flow (Beis et al., 2005;Kitaguchi et al., 2009). Furthermore, combining these lines with transparent lines (White et al., 2008) makes it possible to observe cerebral blood vessels in adult zebrafish without surgical intervention. Therefore, zebrafish might help analyze aging-dependent changes in the cerebral vascular system. However, how cerebral blood flow changes with aging and the relationship between blood flow change and learning ability in adult zebrafish are unclear. Furthermore, recent studies suggest that agingdependent expression changes in genes regulating the vessel wall extracellular matrix (ECM) affect vascular vessel morphogenesis and function in mammals (Rammos et al., 2014;Ungvari et al., 2018). However, it is unclear whether the changes in aging-dependent gene expression in the vascular vessels are conserved across vertebrates.
In this study, we focused on the dorsal telencephalon, which is critical for learning ability in adult zebrafish (Aoki et al., 2013;Mueller, 2012). We observed changes in the blood vascular network and cerebral blood flow in the zebrafish dorsal telencephalon with aging and examined the relationship between cerebral blood flow status and learning ability. In addition, we also examined the changes in gene expression, which can modulate vascular morphogenesis and function.

Live imaging
Anesthesia was performed by diluting ethylene glycol monophenyl ether (Wako) with breeding water. After anaesthetizing the fish in a 10 cm deep petri dish for 30 s with 0.05 % ethylene glycol monophenyl ether, the fish was fixed by sandwiching it in the notch of the sponge. A peristaltic pump was used to supply 0.04 % anesthetic to the mouth of the fish at a flow rate of 5 ml/min. Imaging was performed by confocal microscopy (Leica SP8) with a 10× water immersion lens (Fig. 1). A focal plane (optical section size = 17.9 μm) was taken at 2.29 flame/s, and approximately 100-150 sequential images were taken by sifting the z-axis with 6 μm steps. After imaging, the breeding water was refluxed with a peristaltic pump, and after awakening from anesthesia, the fish were returned to the breeding system. Z-projections of the images and tracing of the vascular vessels were performed by Fiji (Schindelin et al., 2012) and Illustrator (Adobe).

Measurement of tortuosity
Z-projection images of the cerebral vascular network in Tg(kdrl: EGFP); Tg(gata1:mRFP) taken under the described conditions in 2.2 were generated. The merged view of the GFP channel and RFP channel was binarized and skeletonized with Fiji (Schindelin et al., 2012). Then, all detected blood vessels were divided into segmented vessels, and the curvilinear length and straight length of each segmented vessel were measured with MATLAB (Fig. S1). The tortuosity was calculated by dividing the curvilinear length by the straight length of all segmented vessels. The ratio of the number of high tortuosity (over 1.2) to the total number of vessels was defined as a high tortuosity rate. We tracked the tortuosity of the same animals as much as possible, but due to technical limitations such as spontaneous death during keeping, it was difficult to track tortuosity of same animals during the whole period. In this study, we analyzed the tortuosity of segmented vessels of 57 animals in total and observed the same animals of segmented vessels over 3.9 months on average.

Measurement of blood flow
Z-projection images of Tg(kdrl:EGFP); Tg(gata1:mRFP) taken under the described conditions in 2.2 were generated with Fiji. In our confocal microscope imaging conditions, it was difficult to analyze precisely the presence or absence of blood flow in blood vessels with a diameter of 10 μm or less. In our z-projection imaging condition, when the blood cells are flowing, the red fluorescence makes a line-shaped trajectory. On the other hand, when blood cells stopped, the red fluorescence makes dotlike spots. Therefore, we defined green blood vessels, whose diameter of 10 μm or greater, with red blood cell fluorescence line-shaped trajectory as vessels with blood flow. Green blood vessels, whose diameter of 10 μm or greater, with red blood cell fluorescence dot-like spots or no red blood cell fluorescence signal was defined as vessels without blood flow. We manually measured each length of the vessel with Fiji and calculated the blood flow rate by dividing the length of a blood vessel with blood flow by the total blood vessel length.

Measurement of blood vessel numbers and bleeding rate
By comparison with the image taken one month ago, the number of newly formed blood vessels and disappeared blood vessels was measured in the dorsal telencephalon of the zebrafish. We calculated the rate of individuals with apparent bleeding in dorsal telencephalon from z-projection images of Tg(kdrl:EGFP); Tg(gata1:mRFP).

Active avoidance test
Learning ability was measured with an active avoidance test by discrete spaced training, as previously reported (Yang et al., 2018). The active avoidance rate was calculated as the escape success rate in 50 trials. Fig. 1. Illustration of the live imaging method. The fish were fixed by the sponge. A peristaltic pump was used to supply 0.04 % anesthetic to the mouth of the fish at a flow rate of 3-5 ml/min. Imaging was performed by confocal microscopy. The illustration was created with BioRender.com.

qRT-PCR
Four-to five-mpf and 14-to 15-mpf fish were anesthetized and dissected under a stereomicroscope, and the whole brain was removed on ice. Then, the whole brain was homogenized by a plastic pestle, and total RNA was isolated using Sepasol®-RNAISuper G (Nakarai Tesque). Reverse transcription was performed by ReverTra Ace (TOYOBO) with random primers. qRT-PCR was performed with a LightCycler® 96 System (Roche) using THUNDERBIRD® qPCR Mix (TOYOBO). The primers for qRT-PCR are shown in the supplemental file (Table S1). The relative expression was calculated by the ΔΔCt method.

Elastic fiber staining
Adult zebrafish were anesthetized and dissected under a stereomicroscope, and the whole brain was removed. Then, the whole brain was fixed with 4 % PFA at 4 • C overnight. Telencephalon was separated from the whole brain, dehydrated by cold acetone, and embedded in Technovit 8100 (Kulzer). The polymerized plastic block was sectioned into 7 μm sections with a microtome. Sectioned samples were washed PBS 5 min and treated with 25 % ethanol/PBS for 1 min, 50 % ethanol/PBS for 1 min, 70 % ethanol/PBS for 1 min, and 70 % ethanol/1 % hydrochloric acid/PBS for 1 min, sequentially. Then, samples were stained with Weigert's resorcin-fuchsin solution (Muto Pure Chemicals Co., Ltd). The sections were washed with 70 % ethanol/1 % hydrochloric acid/PBS for 5minn, then washed with running water for 10 min. After washing, samples were stained with van Gieson solution (A 10:1 mixture of saturated picric acid solution and 1 % acid fuchsin solution) and washed with distilled water for 1 s. After staining, samples were dehydrated and mounted with Bioleit (Okenshoji Co., Ltd). Images were captured using an inverted microscope (Axio Observer Z1, Zeiss) and digital camera (DS-2500, Sato Shouji Inc).

Statistical analysis
Statistical analysis was performed by GraphPad Prism (GraphPad software) or the R program. The detailed method is described in the figure legends.

The high tortuosity rate increased in an age-dependent manner in the telencephalon
To evaluate the effect of aging on blood vessel patterning, we examined the blood vessel network every month starting at 2 months post fertilization (mpf). Using Tg(kdrl:EGFP); Tg(gata1:mRFP), Casper; Tg (kdrl:EGFP), and Tg(gata1:mRFP); nacre zebrafish lines (Beis et al., 2005;Kitaguchi et al., 2009;White et al., 2008), we imaged the blood vasculature with GFP and followed red blood cells with RFP in the dorsal telencephalon, which is responsible for learning in zebrafish (Aoki et al., 2013;Mueller, 2012) (Figs. 1, 2A). The lifespan of Tg(kdrl:EGFP); Tg (gata1:mRFP), Casper was 18 mpf in our rearing environment (Fig. S2). Therefore, we conducted continuous imaging analysis until 16 mpf. Then, we compared the brain vascular patterning of the same fish at the young stage (3 mpf), young-middle-aged stage (8 mpf), and middle-oldaged stage (12 mpf). The major brain vascular patterning of the dorsal telencephalon was not drastically changed with aging (Figs. 2A, B, S1). However, there was an expansion of the vascular network between 3 mpf and 8 mpf (Figs. 2A,B,S3). The expansion of the vascular network is due to the increase in brain size accompanying growth, because, in zebrafish, significant body size growth occurs between 3 mpf and 8 mpf (Yang et al., 2019). In addition, the formation of new vascular vessels and the disappearance of some vascular vessels were observed (Figs. 2A, B, S3, yellow and magenta arrowheads).
Tortuosity, a score of the vascular curve, is one of biomarker which indicates the aging-dependent phenotype of vessels (Ciuricȃ et al., 2019). Therefore, we measured the tortuosity of vessels in the zebrafish dorsal telencephalon. Z-stack images of blood vessels were projected, binarized and skeletonized sequentially. Then, all detected blood vessels were divided into segmented vessels, and the curvilinear length and straight length of each segmented vessel were measured (Figs. 2C and S1). The tortuosity was calculated by dividing the curvilinear length by the straight length. We found the average tortuosity significantly increased age-dependently, although it was highly variable among each vessel (Fig. 2D). In the human brain, it has been reported that tortuosity over 1.2, which is not seen in younger age groups, appears in small arterial vessels after the age of 50 (Thore et al., 2007). Thus, we calculated the percentage of tortuosity above 1.2, termed the high tortuosity rate (HTR) and found HTR was significantly higher in the middle-old stage (13, 15, 16 mpf) than in the young stage (2 mpf) (Fig. 2E). These data show that tortuosity increases with aging in zebrafish. Therefore, tortuosity of vessels and HTR might be useful indicators for detecting changes in the vascular network with aging in adult zebrafish.

The blood flow rate was decreased in the telencephalon in an agedependent manner
We then next evaluated the effect of aging on cerebral vascular function in zebrafish. We imaged the red blood cells flow and calculated the blood flow rate (BFR) in the zebrafish dorsal telencephalon. Using Tg (kdrl:GFP); Tg(gata1:RFP), Casper; Tg(kdrl:GFP) and Tg(gata1:RFP); nacre, we performed live imaging of the blood vessels and blood flow in the dorsal telencephalon. BFR was calculated by dividing the length of a blood vessel with blood flow by the total blood vessel length (Figs. 3A and S4, please see the detailed method in the Methods section). In fish from 5 to 8 mpf, BFR was over 90 %. After 9 mpf, BFR was slightly decreased compared to younger fish. After 14 mpf, BFR further declined with aging (Fig. 3B). These data showed that aging affected cerebral vascular function in zebrafish.

Spontaneous cerebral bleeding tends to increase with aging
We found nontraumatic spontaneous bleeding in some cases (Fig. 4A). Then, we examined the relationship between spontaneous bleeding and aging. Spontaneous bleeding in the telencephalon was observed at most of the ages examined (Fig. 4A, B). However, the bleeding rate (the percentage of individuals with bleeding among all observed individuals) tended to increase with aging ( Fig. 4B).
In many cases, young fish could survive until the next month even if they had cerebral hemorrhage, probably due to the regeneration capacity of the zebrafish brain (Kizil et al., 2012) (Fig. 4B, blue bars). On the other hand, after 13 mpf, fish that showed cerebral hemorrhage tended to die during that month (Fig. 4B, magenta bars).

The remodeling of the blood vessel network was observed even in aged fish
The cerebral vessel network changed with aging ( Fig. 2A). To evaluate changes in the vascular network with aging, we measured the number of newly formed blood vessels and the number of disappeared blood vessels in the dorsal telencephalon region every month after 3 mpf. Angiogenesis and degeneration of blood vessels were observed in the telencephalon at each age (Fig. 5). However, there was a difference in their age-dependent occurrence. Angiogenesis was more frequently observed in young fish (3-4 mpf) than in young-middle-aged fish (5-11 mpf). In addition, it increased in the middle-old-aged fish again (12-15 mpf) (Fig. 5, light grey bars). Blood vessel disappearance was observed in all stages but was slightly higher in aged fish (Fig. 5, dark grey bars).

BFR correlates with memory and learning ability in aged fish
Previously, we showed that aged zebrafish had lower learning ability than young zebrafish (Yang et al., 2018). Therefore, we examined the relationship between BFR and learning performance. We measured the active avoidance learning abilities of individual fish and examined the correlation between BFR and learning ability in individual fish.
In young (5-6 mpf) and young-middle-aged (9 mpf) fish, the avoidance rate, which indicates learning ability, was high, and a positive correlation between BFR and learning ability was not observed (Fig. 6, correlation coefficient = 0.25 and − 0.57, respectively). On the other hand, middle-old-aged fish (14 mpf) showed a positive correlation between BFR and learning ability (Fig. 6, correlation coefficient = 0.64). BFR in young and young-middle fish tended to be higher than that in middle-old-aged fish (Figs. 3 and 6). However, some young and youngmiddle-aged fish showed low learning ability, even though they showed high BFR. Therefore, factors other than BFR might also play roles in learning in the younger stage.

Gene expression changes in the zebrafish telencephalon during aging
To investigate the aging effect on gene expression in brain vessels, we performed qRT-PCR analysis. First, we examined the expression of marker genes for differentiated neurons (elav3), neural stem cells (her4 and her15.1), vascular endothelial cells (pecam and cxcl12b), glial cells (gfap and fabp7a), microglia (aif1 l and ccl34.1), and pericytes (pdgfrb) (Fig. S3) (Raj et al., 2018). The glial marker gfap expression was significantly increased in the middle-old-aged brain, whereas there was no difference in the expression of other marker genes between young and middle-old-aged brains (Fig. S5).
Next, we focused on the genes col4, elastin, and emilin1, which encode ECM components and modulate vascular vessel morphogenesis and function (Duca et al., 2016;Heinz, 2021;Steffensen et al., 2021). We found that elastina (elna), emilin1a, and emilin1b expression was significantly increased in the middle-old-aged fish brain (Fig. 7A). However, the expression of col4a1 and col4a2, which encode collagen IV ⍺1 chains and one ⍺2 chain, was not changed (Fig. 7A). Then, we performed elastic fiber staining. Contrary to the qRT-PCR analysis results, the elastic fiber of the vascular vessel was reduced age-dependently in the zebrafish  Fig. S2. Right panel: vessels with blood flow (red) and those without blood flow (blue) were identified, and the blood flow rate (the ratio of the length of a blood vessel with blood flow to the total length of the blood vessel) was calculated. See method section for details on how to take blood vessel images. Scale bar, 500 μm.

Cerebral vascular patterning was changed by aging in the adult zebrafish brain
In this study, we showed that aging increased the tortuosity of the blood vessels in the dorsal telencephalon. An increase in the tortuosity of the cerebral vasculature has also been observed in the human brain (Thore et al., 2007), suggesting a common mechanism in vertebrates for changes in blood vessels with aging. Elastic fibers are the main components of the vessel wall ECM (Duca et al., 2016). Previous research reported that aging induces the fragmentation and calcification of elastic fibers. Thus, aging reduces the elasticity of vascular vessels, affecting vessel morphology in mammals (Duca et al., 2016;Heinz, 2021;Ungvari et al., 2018). In this study, we showed that aging reduces the elastic fiber of vascular vessels in zebrafish telencephalons. Therefore, agingdependent reduction of elastic fiber might be a common phenomenon among vertebrates.
However, we showed an increase in the expression of the elastic fiber component genes elna, emlin1a, and emilin1b in the middle-old-aged zebrafish brain compared to young zebrafish. Elna codes tropoelastin, a precursor of elastin, the main component of the vessel wall ECM (Duca et al., 2016;Heinz, 2021). Emilin1 is mainly associated with elastic fibers and microfibrils in blood vessels, and has been implicated in elastogenesis and maintaining blood vascular cell morphology (Colombatti et al., 2000(Colombatti et al., , 1985. Therefore, aging-dependent increases in elna, emlin1a, and emilin1b might be a compensating effect for a decrease in elastic fiber. Further detailed temporal observation of elastic fiber dynamics in adult zebrafish brain vascular vessels is needed to clarify the relationship among aging, elastic fiber dynamics, and vascular vessel morphogenesis. Collagen is another significant component of the vessel wall ECM. Collagen IV is a type of collagen related to several artery diseases (Klarin et al., 2019;Schunkert et al., 2011;Steffensen et al., 2021). However, in our observations, the expression of col4a1 and col4a2 was not changed at the transcriptional level. The increase in the ratio of collagen to elastin affects vascular vessel morphogenesis in mammals (Duca et al., 2016). Therefore, spatiotemporal analysis of collagen matrix dynamics might be required to understand the effect of aging on vascular morphogenesis and function in zebrafish.
We found that the cerebral vasculature maintains its remodeling ability at least until 15 mpf in zebrafish. Young fish showed higher levels of angiogenesis, which may be related to the growth of the brain. Middle-old-aged fish also showed similar level of angiogenesis. In addition to the biosynthesis and modification of the vessel wall ECM, several studies in rodent models have shown that age-dependent vascular remodeling is attributed to Ras signaling, MCP-1/CCR2 signaling, and a decline in circulating levels of growth hormone, IGF-1, and oestrogens (Spinetti et al., 2004;Ungvari et al., 2018). Thus, these factors may also play an essential role in aging-dependent vascular remodeling in zebrafish.
Although we cannot rule out the possibility that the stress of restraint and repeated imaging deteriorate cerebral hemorrhage, spontaneous bleeding tended to increase in aged fish. In a mouse model, aging promotes cerebral microhemorrhage by exacerbating hypertension-induced oxidative stress and metalloproteinase activation, which cause vascular vessel dysfunction (Toth et al., 2015;Wakisaka et al., 2010). Therefore, increased spontaneous bleeding in aged fish may also be due to agingdependent vascular dysfunction, such as a decline in elasticity due to decreased elastic fibers.

Aging-dependent dysfunction of the neurovascular units might affect the patterning of the cerebral vascular network
Pericytes form neurovascular units with vascular endothelial cells, glial cells, and neurons (Brown et al., 2019;Sweeney et al., 2016). Aging decreases pericyte function and might impair the interaction between pericytes and endothelial cells (Bennett and Kim, 2021). Furthermore, pericyte loss has been suggested to occur during normal aging (Bennett and Kim, 2021). However, in this research, the expression of the pericyte marker gene pdgfrb was not changed in middle-old-aged fish. On the other hand, the glial marker gene gfap was increased in middle-old-aged fish brains. Aging-dependent upregulation of gfap is common among mammals due to the increase in the volume of astrocytes and increased transcription (Middeldorp and Hol, 2011). Since aging-dependent increases in the number of astrocytes are much more modest, this increase in gfap transcription could arise from increased oxidative stress (Middeldorp and Hol, 2011). Therefore, the functions of the neurovascular units might decline with aging, even in the middle-old-aged stage. Aging-dependent dysfunction of the neurovascular units might affect the patterning of the cerebral vascular network and increase HTR. To explore these possibilities, additional studies are needed to evaluate the relationship between neurovascular units and aging in zebrafish adult brains.

Age-dependent decrease in BFR might be involved in memory impairment in zebrafish
Blood vessels are essential for supplying oxygen and nutrients to neurons. Thus, impaired blood vessel function might decrease learning activity in aged fish. We found that in aged fish, memory learning ability positively correlated with BFR. In humans, resting hippocampal blood flow was positively correlated with spatial memory performance (Heo et al., 2010). Therefore, a common molecular mechanism in humans and zebrafish may underlie the deterioration of learning ability in the aged brain caused by decreased blood flow.
In young people, there was no positive correlation between hippocampal blood flow and spatial memory performance (Heo et al., 2010). Similarly, in young-middle-aged zebrafish, a positive correlation between BFR and learning ability was not observed in this study. Many young and young-middle-aged zebrafish showed a relatively higher BFR than middle-old-aged zebrafish. Although a strikingly low BFR may affect the learning ability in younger fish, BFR above a threshold value might not affect learning ability. Additional studies are required to clarify this issue.
In human, blood pressure is the driving force for cerebral blood flow and the dysfunction of cardiovascular system affects cerebral blood flow (Tarumi and Zhang, 2018). It is expected that changes in blood pressure by aging also affect BFR in zebrafish. It is necessary to develop a method for measuring blood pressure in adult zebrafish in the future. A limitation of this study is BFR measurement under anesthesia using ethylene glycol monophenyl ether, which reduces the heartbeats in fish (Lambooij et al., 2009). Future studies on the BFR measurement of a live zebrafish brain without anesthesia would help overcome the limitation.
In conclusion, this study showed that aging affects the brain vascular structure and decreases the BFR in the zebrafish dorsal telencephalon. BFR correlates with learning ability in aged zebrafish, similar to aged people. We found the aging-dependent decline in elastic fiber of cerebral vascular vessel in zebrafish. We also found upregulation of gfap suggesting dysfunction of the neurovascular unit. These are common phenomenon among vertebrates. Therefore, the protection of elastic fibers from breakage and maintenance of the neurovascular unit in a younger state may prevent the decline of cerebrovascular function due to aging. Elastic fiber and neurovascular units are potential therapeutic targets for age-dependent cerebrovascular dysfunction and vascular dysfunctiondependent diseases such as vascular dementia and cerebral infarction. Therefore, zebrafish might be a helpful model for analyzing the physiological process of the effect of aging on vascular function and developing therapeutic agents and treatments for age-dependent cerebrovascular dysfunction.

Submission declaration and verification
All authors have reviewed the final version of the manuscript and approve it for publication. Authors declare that the work described has not been published previously and is not under consideration for publication elsewhere.

Declaration of competing interest
The authors declare that they have no competing interests.

Data availability
No data was used for the research described in the article.