Abstract
Aggregation of tau protein into intracellular deposits is a pathognomonic feature of tauopathies such as Alzheimer’s disease (AD) and lowering tau is a prominent therapeutic strategy under development. However, the physiological function of tau protein is not well known, particularly in the periphery. Lowering tau protein risks disrupting its physiological role leading to unwanted effects. In this study, the presence of tau protein in cardiac tissue is confirmed and the functional role in the cardiovascular system and the consequences of its loss were explored. Isolated right and left atria and small mesenteric arteries from wild type and tau deficient (KO) mice of two age groups (13 and 23 months old) were used to assess cardiovascular phenotypes. Tau KO mice showed an increased systolic blood pressure and cardiac hypertrophy at 13 months, which was accompanied by a significantly lower right atrial rate and a subtle decrease in the maximum contractility to calcium, isoprenaline, and electrical sympathetic nerve stimulation. Aging tau KO mice to 23 months resulted in cardiac hypertrophy with significantly attenuated left atrial contractility, increased blood pressure, and sensitivity of isolated mesenteric arteries to angiotensin II contraction and isoprenaline relaxation compared to their younger counterparts. This study supports a functional role of tau in the heart and loss of this protein leads to a deterioration in cardiovascular performance which worsens with age. Taken together, these results provide insight into the peripheral function of tau protein, and give caution to the therapeutic strategy of lowering tau protein.
INTRODUCTION
Insoluble, intraneuronal deposits of the micro-tubule-associated protein tau in neurofibrillary tangles are the defining pathology of a class of diseases termed tauopathies. Tauopathies include Alzheimer’s disease (AD), progressive supranuclear palsy, and frontotemporal dementia, where aggregated tau species are thought to confer neurotoxicity. Tau has also been shown to mediate the toxicity of amyloid-β (Aβ), since a decrease in tau in both cell culture and mice protects against Aβ-related toxicity [1–4], including cellular toxicity [1], cognitive deficits [2, 4], and axonal transport failure [3]. Therefore, therapeutic agents that lower tau may possess the dual benefit of directly decreasing tau toxicity, as well as attenuating Aβ toxicity; for these reasons, lowering tau is a prominent therapeutic strategy.
However, tau is an abundant protein, and therapeutically lowering tau may cause unintended actions. A number of studies have shown the efficacy of tau immunotherapy in transgenic mouse models [5, 6],but as these mice have substantially higher tau levels, it is difficult to evaluate if these drugs may also cause off-target consequences due to tau attenuation. In contrast, vaccination with tau protein of wild type (WT) mice induced a pathogenic immune response resulting in tauopathy-like phenotypes [7]. Tau antibody therapy administered to the J20 AβPP mutant mouse (without tau overexpression) also caused unexpected and unexplained sudden death after chronic treatment [8], which cautions against lowering tau.
We previously reported that tau reduction in mice causes an age-dependent, iron-mediated Parkinsonian neurodegeneration [9–12]. The age-dependent motor impairment phenotype of tau knockout (KO) mice has been supported by other groups [13, 14]. We showed that tau protein facilitates the trafficking of amyloid-β protein precursor (AβPP) to the neuronal plasma membrane, which promotes ferroportin-mediated cellular iron export [9]. Loss of tau or AβPP causes iron elevation and motor disability, which could be rescued by iron chelation[9, 15].
Although it is not often a focus of the field, tau and AβPP are also present in peripheral organs [16, 17]. Here, we explore the presence of tau protein in the heart, and its potential function in the cardiovascular system. We report that tau is present in the heart and loss of tau in the heart causes elevated blood pressure and altered cardiac performance in aged mice.
MATERIALS AND METHODS
Mice
All mice were housed in a specific pathogen-free facility according to standard animal care protocols and fed standard laboratory chow (Code 102108, Barastoc, Ridley AgriProducts) and tap water ad libitum. All mouse procedures were approved by the Florey Institute animal ethics committee (10-017, 14-041) and were performed in accordance with the Australian code for the care and use of animals for scientific purposes (8th edition, 2013, National Health and Medical Research Council, Canberra). Mice on a background of sv129 and C57Bl/6 [18] were originally obtained from Dr. M. Vitek (Duke University) and maintained homozygously, with mutants backcrossed to the parental inbred strain every three generations as previously described [4, 12].
Heart dissection
Tau KO mice and background control mice (C57BL6/SV129) aged approximately 1 and 2 years were deeply anesthetized by i.p. injection of sodium pentobarbitone (80 mg/kg; Lethabarb, New South Wales, Australia), and the chest was opened after checking for the loss of the pinch reflex. The heart was exposed and a needle inserted into the left ventricle to perfuse the body with ice-cold saline; the right atrium was cut to allow the escape of returning blood. The heart was then isolated and either fixed in 10% neutral buffered formalin for immunohistochemistry or stored at –80°C until required for western blot.
Sample preparation and western blot
Whole heart tissue was sonicated in phosphate-buffered saline (PBS, pH 7.4) with EDTA-free protease inhibitor cocktail (1:50, Roche, Indianapolis, IN, USA) and phosphatase inhibitor I and II (1:1000) and 0.1% SDS, and centrifuged at 40,000 g for 30 min as described by Lei et al. [9]. Total protein concentration of the supernatant was determined using a BCA protein assay (Pierce, Rockford, IL, USA) and equal protein concentrations (25 μg) were loaded to 4–12% bis-Tris gels with NuPAGE MES running buffer (Invitrogen, Carlsbad, CA, USA). After completion of the electrophoresis, the gels were transferred to nitrocellulose membranes by using an iBlot dry blotting system (Invitrogen); membranes were blocked with milk (5% v/v) and probed with primary human tau antibody (1:2000 dilution, Dako, Glostrup, Denmark, catalogue no. A0024) overnight and a secondary mouse anti-human IgG-HRP conjugated antibody (1:5000 dilution, Invitrogen). An enhanced chemiluminescence detection system (GE Healthcare, Uppsala, Sweden) was used for developing the gel; Fujifilm LAS-3000 was used for visualization.
Immunohistochemistry
Thin tissue sections (7 μm) from the heart were cut from cooled and hardened paraffin embedded heart with a cryostat. Sections were treated with 5% aqueous hydrogen peroxide and rinsed with tap water. After a brief immersion in Tris buffer bath for 5 min, the sections were blocked in 0.2% casein and incubated with polyclonal rabbit anti Tau primary antibody (1:400 dilution, Dako, catalogue no. A0024) for 1 h. The sections were then incubated further with a secondary biotinylated anti mouse / rabbit immunoglobulin (1:5000 dilution, Invitrogen) for 10 min followed by peroxide conjugated streptavidin. Finally, activated 3’3’-diaminobenzidine solution (1% in PBS + 1% COCl2, 1% NiSO4; 1:3000) was used and color change was monitored microscopically for 1–3 min. Tau immunostained slides were stained with a brief 10–15 s treatment of hematoxylin (Dako) and bluing solution (Dako), and the sections were imaged using a DMLB Leica microscope.
In vitro pharmacological experiments
Separate groups of 13- and 23-month-old mice were deeply anesthetized by spontaneous inhalation of 5% isoflurane (Baxter Healthcare Pty. Ltd., NSW, Australia) in 95% O2 and killed by decapitation. The thoracic cavity was opened and the heart rapidly removed and placed in an ice-cold Krebs-Henseleit physiological salt solution (PSS) of the following composition (mmol/L): NaCl 119; KCl 4.69; MgSO4.7H2O 1.17; KH2PO4 1.18; glucose 11; NaHCO3 25; CaCl2.6H2O 2.5; EDTA 0.026 saturated with carbogen (O2 95%; CO2 5%) at pH 7.4. The abdomen was then opened and approximately 10 cm of jejunum and its attached vascular fan removed and pinned out on a Silastic-bottomed petri dish filled with ice-cold PSS solution that had half the glucose concentration to the above. In the 13-month-old group, female mice were used while the 23-month-old group included both males and females.
In vitro experimental protocols
Isolated right and left atria
For cardiac tissue experiments, the method described by Wright and Angus [19] was used with modifications. Briefly, the atria were dissected free from the ventricles in oxygenated PSS; the right and left atria separated and mounted in 15 ml organ baths maintained at 36.5±0.5°C and bubbled with 95% O2:5% CO2 at pH 7.4. Two stainless steel hooks were passed through the atrium and vertically suspended on support legs attached to an isometric force transducer (FTO3C, Grass Instruments, Warwick, RI, USA). Right atria were stretched twice to 0.5 g force in 10 min intervals. While the right atria were allowed to spontaneously beat, the left atria were paced with square wave pulses (1 Hz, 0.2 ms, 1.5 times threshold voltage) via two punctate platinum electrodes. A force-length curve was constructed for each left atrium by passivelystretching the tissues using an adjustable micrometer and setting the force to the stretch that gave a plateau active force. Tissue contractions were recorded using an isometric force transducer connected to a PowerLab 8SP (ADInstruments, Bella Vista, NSW, Australia). Changes in spontaneous right atrial rate (beats/min) and left atrial force (g) were recorded using LabChart 7 software (ADInstruments). Once a stable baseline was established following equilibration (30–60 min), response curves to various drugs were performed. Two concentration-response curves, first for isoprenaline (0.01–100 nM) followed by methacholine (0.01–100 μM), were obtained per single right atrium. The second experiment with methacholine was only attempted if the atrium returned to a stable baseline rate after washing out the isoprenaline.
For the left atria, 2-3 concentration (or field stimulation)-response curves were obtained per single tissue in the following order: isoprenaline (0.1–300 nM), calcium (9 mM), and electrical stimulation (1–32 field pulses). The tissue was washed repeatedly with PSS and enough time given for recovery between each protocol. For assessing the role of neuronal stimulation on inotropic responses of the left atria the method of Serone and Angus [20] was used with some modifications. Briefly, left atria were first incubated with benextramine (1 μM) for 10 min to irreversibly block presynaptic α2-adrenoceptor-mediated inhibition of norepinephrine release [21]. Then the tissue was washed with PSS to prevent non-selective effects of benextramine and incubated with atropine (1 μM) and desipramine (0.1 μM) for 10 min to block any parasympathetic contribution and reuptake of norepinephrine, respectively, to allow a maximal left atrial response from sympathetic nerve stimulation. Electrical field stimulation for 1, 2, 4, 8, 16, and 32 consecutive field pulses (1 Hz, 5 ms duration, 120 V) was then applied in the refractory period of each atrial beat following the driving punctate stimulus. The atria were given enough time to recover from each block of stimulation before the next ensued and the subsequent changes in atrial force (g) were measured.
Isolated mesenteric arteries
Second or third order mesenteric arteries (250–350 μm internal diameter) were dissected as described in Angus and Wright [22]. Two 40 μm diameter wires were threaded through the lumen of 2 mm length segments of arteries mounted in separate myography baths (Model 610M; Danish Myo Technology, Aarhus, Denmark) for isometric force measurement. Responses were then measured by a PowerLab 4/30 A/D converter (ADInstruments) and recorded on LabChart 7 data acquisition software (ADInstruments). After an equilibration period in PSS at 37°C, the arteries were passively stretched according to a normalization protocol [23] and adjusted to a diameter setting of 90% of that determined for an equivalent transmural pressure of 100 mmHg (D100). After 30 min, the arteries were exposed to a potassium depolarizing solution (124 mM K+ replacing Na+ in PSS; termed KPSS) and norepinephrine (10 μM) for 2 min. A second exposure to KPSS solution (only) was used to provide a reference contraction.
Contraction responses were assessed by performing cumulative response curves to methoxamine (0.01–100 μM) and angiotensin II (0.1–100 nM). For relaxation experiments, an initial stable precontraction was obtained using U46619, a thromboxane mimetic, before completing response curves to sodium nitroprusside (0.001–30 μM), acetylcholine (0.001–3 μM), and isoprenaline (0.001–30 μM). The concentration of U46619 (10–100 nM) necessary for the initial precontraction was determined by gradual increments to cause a submaximal response of about 80% KPSS.
In vivo blood pressure measurement
In the 13-month-old group, systolic blood pressure was measured in conscious, restrained mice using a CODATM computerized non-invasive blood pressure system (Kent Scientific, Torrington, CT, USA) that uses a volume pressure recording (VPR) sensor technology. Each mouse was placed in a warmed restraining tube sitting on a surface board maintained at 37°C. The tail of each mouse was passed through a cuff to occlude blood flow and another distal cuff with the VPR sensor to measure blood pressure. The VPR sensor was connected to a controller linked to a computer that acquired data by CODATM software version 4.1. Mice were acclimatized to the environment and apparatus by placing them in the restraining tubes for 10 min for 2-3 consecutive days and on experiment days prior to acquiring data. An average of at least 3 measurements was taken on separate days. The room temperature was maintained around 26°C and tails were heated to more than 30°C (as confirmed by an infrared thermometer) by using a small heater fan adjusted toward the tail; core temperature was not affected. Mice spent a maximum of 20 min in the restrainer and were carefully inspected for signs of stress before and during BPmeasurement.
The older 23-month-old mice were anesthetized by i.p. injection of ketamine/xylazine (100 mg/kg and 10 mg/kg) and placed on a heat pad to maintain core temperature. The neck area was shaved and a local anesthetic, lidocaine (1%), was injected. The nose was placed inside in a cone that delivers room air supplemented with oxygen. A small (∼1 cm long) incision was made in the neck to access the left carotid artery and the artery was isolated with 6/0 silk and polyvinylchloride fluid-filled catheter (0.5 mm o.d.) was carefully inserted in to the carotid artery and secured. The catheter was connected to a pressure transducer (Cobe, Argon Medical, Tx, USA) for the continuous measurement of blood pressure. Phasic blood pressure was recorded via a PowerLab amplifier (ADInstruments) and LabChart 7 data acquisition software. Systolic, diastolic and mean arterial pressure were derived from the phasic blood pressure every 3 min for 20 min and an averagewas taken.
Drugs
The drugs used and suppliers were: acetylcholine bromide (Sigma, St. Louis, MO, USA), angiotensin II amide (Auspep, Tullamarine, Victoria, Australia), atropine sulphate (Sigma), benextramine tetrachloride (Sigma), calcium chloride solution (Scharlab, Sentmenat, Spain), desipramine hydrochloride (Sigma), isoprenaline hydrochloride (Sigma), acetyl-β-methylcholine chloride (Sigma), methoxamine (Sigma), [–]-norepinephrine bitartarate (Sigma), sodium nitroprusside (Enzo, Farmingdale, NY, USA), U46619 (9,11-dideoxy-9a,11a-methanoepoxy prostaglandin F2α; Tocris Biosciences, Bristol, UK). All drugs were prepared in ‘Type-1, ultrapure’ water (Milli-Q, Merck Millipore) except U46619, which was made to 10 mM in DMSO and then diluted in Milli-Q water. Stock solutions were stored at –20°C.
Statistical analyses
All data are expressed as mean±SEM from n experiments each from separate mice. The response of isolated mesenteric arteries is expressed as a change in effective active pressure (ΔP) which is the change in tension (ΔT) per internal circumference. This is important as it takes the diameter of vessels in to account. The Laplace relationship, ΔP = (2πΔT)/L, where L is the length of the internal circumference of the vessel, was used to calculate the effective transmural pressure from the change in force (ΔF) in mN. The data in isolated mesenteric arteries are expressed as change in effective active pressure in mmHg (conversion factor 1 mN/mm2 = 7.502 mmHg). For the relaxation experiments, only the arteries that gave near complete relaxation (>80%) at the highest agonist concentration from the precontractile tone were included in the data and analysis. Each sigmoidal concentration-response curve was fitted using Prism 7 (GraphPad Software, La Jolla, CA, USA). The pEC50±SEM values (the negative log10M of agonist concentration that caused 50% of the maximum response) and Emax (maximum response) were determined for each group.
Unless otherwise stated, two-way analysis of variance (ANOVA) with Sidak’s test for multiple comparison of either rows or columns was used to analyse the effect of the two independent parameters: age, a row factor, and genotype, a column factor. In all cases, the multiplicity adjusted p value for individual comparisons between the different groups or the adjusted p value for the main effect of either genotype or age is reported. p < 0.05 is considered statistically significant.
RESULTS
Mice and isolated tissue characteristics
The presence of tau in the heart of mice (on C57BL6/SV129 background) was confirmed using western blot (Fig. 1A) and immunohistochemistry (Fig. 1) of heart tissue sections, in accordance with a previous report [16]. Tau KO and background control mice used in this study had a similar body weight at 13 months (WT: 41.7±1.8 g and KO: 44.5±2.4 g) and 23 months (WT: 42.3±2.4 g and KO: 42.2±3.0 g, p > 0.05; Fig. 2A). Tau KO mice had an elevated heart-to-body weight ratio at 13 months compared to WT mice (p = 0.029; Fig. 2B). This difference was no longer evident at 23 months as the heart-to-body weight ratio of 23-month-old WT mice was increased to a level similar to that of the 13- and 23-month-old tau KO mice. The heart-to-body weight ratios of the separate heart chambers were also assessed from a subset of these mice to evaluate where the hypertrophy was occurring. In both WT and tau KO mice, there was a marked increase in the left atrial weight ratio at 23 months (Fig. 2C). Further, the left atria of tau KOs had a significantly increased weight ratio at 23 months compared to the WTs (p = 0.0002; Fig. 2C). At 23 months, the right atria were significantly enlarged compared to their respective 13-month-old counterparts in both WT (p = 0.033) and KO mice (p = 0.004; Fig. 2D). Within the tau KO or WT groups, a similar effect of age was seen in the ventricles where the 23-month tissues were heavier than respective 13-month tissues (p = 0.009 and 0.023 in WT and KO, respectively; Fig. 2E). However, there was no difference in the normalized weight of either the right atria or ventricles between WT and tau KO mice at either 13 or 23 months (Fig. 2D, E).
In vivo and in vitro cardiac phenotype
Blood pressure was measured in conscious restrained (13 months) and anesthetized (23 months) mice. Tau KO mice had greater systolic blood pressure than age-matched WT mice at both 13 months (WT: 97±4 mmHg, KO: 112±5 mmHg, p = 0.032) and 23 months (WT: 87±4 mmHg, KO: 110±11 mmHg, p = 0.038; Fig. 2F). There was a non-significant trend for a corresponding increase in diastolic (Fig. 2G) and mean blood pressure (data not shown) in tau KO mice.
In functional assays of isolated right and left atria, the resting (baseline) left atrial force of tissues from 23-month-old KO mice was significantly lower than in 23-month-old WT atria (p = 0.029; Fig. 3A). Although a trend for decreased force was seen in the left atria from younger tau KO mice, this was not statistically significant. Similarly, after exposure to a high calcium concentration (9 mM), the peak inotropic response was significantly depressed in the older tau KO mice (WT: 0.22±0.02 g/mg, KO: 0.15±0.02 g/mg, p = 0.002; Fig. 3B). The left atria from both groups of 23-month-old mice had decreased inotropic responses to calcium compared to their respective 13-month-old counterparts (p < 0.05; Fig. 3B). In response to another positive inotropic agent isoprenaline, tau KO left atria showed a similar phenotype as with calcium: a significantly decreased maximum response (Emax) in 23-month tissues compared to WT tissues (WT: 0.23±0.02, KO: 0.16±0.02, p = 0.008; Fig. 3C) and in tissues from older mice of both genotypes, the Emax was less than in 13-month atria (p < 0.01). The range of response to isoprenaline (i.e., from baseline to Emax) was also significantly less (only 47-48% of the range in 13-month atria) in 23-month tau KO (p = 0.002) and WT (p = 0.0003) left atria compared to respective 13-month tissues. To assess if this cardiac phenotype also affected the innervation of the left atria, electrical field stimulation of sympathetic nerves was tested. Similar to the response of tau KO left atria to calcium and isoprenaline, there was a decreased contractile response of left atria from both 13-month-old tau KO (WT: 0.28±0.01 g/mg, KO: 0.22±0.01 g/mg after 32 field pulses, p = 0.017) and 23-month-old tau KO mice (WT: 0.20±0.02 g/mg, KO: 0.13±0.02 g/mg after 32 field pulses, p = 0.003; Fig. 3D). Aging also significantly decreased stimulation-induced inotropy in both 23-month WT and KO atria compared with their respective 13-month group (p = 0.0004 each). However, there were no differences in the sensitivity of the tau KO and WT left atria to contraction by either isoprenaline or electrical fieldstimulation.
The isolated right atria of 13-month-old tau KO mice had significantly decreased basal intrinsic rate compared to the 13-month WT atria (WT: 368±18 beats/min, KO: 277±15 beats/min, p = 0.004; Fig. 4A). By 23 months of age, there was an elevation of the right atrial basal rate in both WT and tau KO tissues and there was no longer any difference between genotypes. When positive chronotropic responses to isoprenaline were assessed, there were no differences between any of the groups in either sensitivity (pEC50) or maximum response (Emax) (p > 0.05, Fig. 4B). The negative chronotropic responses of the right atria to methacholine were not different in either sensitivity or Emax between ages or genotypes (Fig. 4C), despite the lower basal rate in 13-month tau KO tissues.
Vascular phenotype
Isolated small mesenteric arteries were examined to investigate whether loss of tau affected vascular reactivity. After normalization to an isometric stretch equivalent to a transmural pressure of 100 mmHg, the small resistance arteries had a comparable internal diameter in both WT (267±9 μm in 13-month-old and 257±8 μm in 23-month-old) and tau KO groups (283±13 μm in 13-month-old and 256±7 μm in 23-month-old). Three different types of vasodilator agents with differing mechanisms of action— sodium nitroprusside (Fig. 5A), acetylcholine (Fig. 5B), and isoprenaline (Fig. 5C)— all relaxed mesenteric arteries that were precontracted with U46619, a thromboxane mimetic. There was a similar sensitivity (pEC50) to either sodium nitroprusside or acetylcholine in arteries isolated from WT or tau KO mice, and there was no effect of aging with comparable pEC50 values at all ages (Table 1). Aged tau KO (23-month-old) arteries showed a small 6.3-fold increase in sensitivity to isoprenaline relaxation compared to their younger counterparts (p = 0.013;Table 1).
Methoxamine and angiotensin II caused concentration-dependent contractions of mesenteric arteries (Fig. 6). The vessels from the aged 23-month-old WT mice were 3-fold less sensitive to methoxamine than the 13-month-old WT group pEC50 (p = 0.02; Table 1); the maximum contractile response was also decreased (p = 0.008; Fig. 6A, Table 1). In contrast, there was no change with age in the responses to methoxamine in the tau KO group. However, the 23-month tau KO arteries had an increased sensitivity to angiotensin II compared to 23-month WT arteries (3.1 fold, p = 0.015, Table 1), as well as the 13-month tau KO arteries (5.2 fold, p = 0.002, Fig. 6B; Table 1). The Emax to angiotensin II was only decreased in the arteries from 23-month-old WT mice (Table 1; Fig. 6B).
DISCUSSION
The microtubule-associated protein tau has an important role in stabilizing microtubules, axonal transport, and modifying organization of cytoskeletal network in neurons [24]. However, the functions of tau protein, especially in the periphery, are not well understood, but indeed important to characterize if tau-attenuating compounds are to be investigated as novel therapies for tauopathies. In this study, we characterized the cardiovascular phenotype of tau KO mice. The findings of this study confirm the presence of tau in the cardiac tissue of mice (previously identified in cardiac tissue of rats [16] and humans [17]) and show that tau KOs express a detrimental cardiovascular phenotype that is aggravated by age.
At 13 months of age, tau KOs exhibited increased systolic blood pressure with cardiac hypertrophy, lower intrinsic right atrial rate, and a subtle decrease in left atrial contractile function. At 23 months, a similar increase in systolic blood pressure with increased heart mass and a further decline in left atrial contractility was observed in tau KOs compared to WTs. Increased load in the heart due to chronically increased blood pressure is known to initially lead to cardiac hypertrophy as a compensatory response, which may then progress to a decompensated stage [25]. Consistent with this, 13-month-old tau KO mice showed increased blood pressure, increased heart-to-body weight ratio, and a trend toward decreased resting as well as peak calcium responses; a phenotype of hypertrophy without decompensation. However, in 23-month-old mice, there was a decompensation in the left atria of tau KOs as evidenced by the significantly decreased contractile response to calcium (resting, peak and delta (peak – resting)), isoprenaline, and electrical nerve stimulation.
Aged WT mice had differences in both cardiac and vascular function. There was an increased heart-to-body weight ratio of the different heart chambers, decreased inotropic responses to calcium, isoprenaline, and electric field stimulation, and decreased sensitivity of mesenteric arteries to contraction by methoxamine. These changes are similar to those seen with human and mouse cardiac aging [26, 27]. The 13-month-old tau KO mice showed comparable changes to the 23-month-old WT mice in many of the assays, perhaps representing an early aging phenotype.
We used two different techniques to measure the blood pressure: the tail cuff system was used for the conscious restrained 13-month-old mice and intra-arterial fluid-filled catheter for the anesthetized 23-month-old mice. Previous studies by both our group (unpublished) and others [28] show that the two measurements are highly correlated in measuring blood pressure. Both measurements have certain limitations. Although tail-cuff is the simplest method, stress associated with balloon inflation and restraining will influence the results [28]. The use of anesthesia in the catheterization technique is associated with decreased blood pressure and cardiac output [29]. The fact that a similar trend was seen using the two measurements gives us confidence that the results are robust. However, the use of anesthesia in the 23-month-old mice may have resulted in an underestimate of (conscious) blood pressure values.
Tau may have an important physiological function in the cytoskeletal assembly, autonomic innervation, or the conduction system of the heart and vasculature. A study by Kim and colleagues showed that tau is one of the key microtubule-stabilizing proteins involved in endothelial cell lumen formation and vascular tube morphogenesis [30]. Suppression of tau by siRNA inhibited endothelial cell lumen formation and decreased the post-translational acetylation and detyrosination of tubulin [30]. Tau proteolysis by calpain is also reported to be responsible for collapse of microtubules and failure of VEGF-driven angiogenesis [31]. The endothelial cell-cardiac myocyte interaction is known to play a key role in regulating survival, cytoskeletal organization, and function of the myocardium [32], thus it is plausible that tau loss may have caused the observed cardiac phenotype in our study by affecting this interaction or the process of angiogenesis.
Microtubules also play a significant role in determining the cardiomyocyte response in cardiac hypertrophy. In this study, deficiency of tau showed a hypertrophic phenotype with significantly depressed left atrial contractility to the β-adrenoceptor agonist isoprenaline (Fig. 3C). Consistent with our result, a previous study by Palmer and colleagues [33] in isolated cardiomyocytes showed that both microtubule assembly and β-adrenoceptor responsiveness were attenuated in a rat model of cardiac hypertrophy, which was partially reversed by a taxol-induced increase in microtubule assembly. Numerous gene mutations of cytoskeletal proteins, which are potentially affected by tau, have been linked with human dilated and hypertrophic cardiomyopathies [34], with significant differential expression of tau mRNA and protein levels [35, 36]. Tau may also be linked with cardiac development and function through activity-dependent neuroprotective protein (ADNP). ADNP binds chromatin to regulate the expression of several genes and its loss leads to impaired cardiac development during embryogenesis [37] and causes tau hyperphosphorylation [38]. It is possible that tau loss may also cause iron accumulation in the cardiomyocytes similar to substantia nigra neurons of tau KO mice [9] which is linked with different heart diseases [39]. The specific mechanism by which tau loss causes this cardiovascular phenotype needs further investigation, but regardless, these results provide insight into the peripheral function of tau protein, and give caution to the therapeutic strategy of lowering tau protein.
Footnotes
ACKNOWLEDGMENTS
We thank Mr. Mark Ross-Smith and Dr. Makhala Khammy for assistance with the myography experiments, Ms. Linda Cornthwaite-Duncan for assistance with the in vivo experiments, and the Florey Histology Unit (Ms. Mirjana Bogeski) for assistance with the immunostaining of tau in the heart. The study was supported by funds from the Australian Research Council, the National Health & MedicalResearch Council of Australia, the Cooperative Research Centre for Mental Health, Alzheimer’s Australia Dementia Research Foundation and the National Natural Science Foundation of China (81571236).
