HB-EGF depolarizes hippocampal arterioles to restore myogenic tone in a genetic model of small vessel disease

Abstract

Vascular cognitive impairment, the second most common cause of dementia, profoundly affects hippocampal-dependent functions. However, while the growing literature covers complex neuronal interactions, little is known about the sustaining hippocampal microcirculation. Here we examined vasoconstriction to physiological pressures of hippocampal arterioles, a fundamental feature of small arteries, in a genetic mouse model of CADASIL, an archetypal cerebral small vessel disease. Using diameter and membrane potential recordings on isolated arterioles, we observed both blunted pressure-induced vasoconstriction and smooth muscle cell depolarization in CADASIL. This impairment was abolished in the presence of voltage-gated potassium (K_V1) channel blocker 4-aminopyridine, or by application of heparin-binding EGF-like growth factor (HB-EGF), which promotes K_V1 channel down-regulations. Interestingly, we observed that HB-EGF induced a depolarization of the myocyte plasma membrane within the arteriolar wall in CADASIL, but not wild-type, arterioles. Collectively, our results indicate that hippocampal arterioles in CADASIL mice display a blunted contractile response to luminal pressure, similar to the defect we previously reported in cortical arterioles and pial arteries, that is rescued by HB-EGF. Hippocampal vascular dysfunction in CADASIL could then contribute to the decreased vascular reserve associated with decreased cognitive performance, and its correction may provide a therapeutic option for treating vascular cognitive impairment.

Introduction:

CADASIL (Cerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and Leukoencephalopathy) is an archetypal monogenetic form of cerebral small vessel disease (SVD) that emerges as the most common heritable cause of stroke and vascular cognitive impairment in adults (Chabriat et al. 2009). It offers a lens to investigate SVD pathogenesis as there is no specific treatment, despite the profound impact on human health worldwide (Pantoni 2010; Iadecola et al. 2019; Wardlaw, Smith and Dichgans 2019). CADASIL disease is caused by dominant mutations in the NOTCH3 receptor that lead to the extracellular deposition of NOTCH3 ectodomain (Notch3^ECD) and other proteins–including one called tissue inhibitor of metalloproteinases 3 (TIMP3) (Monet-Leprêtre et al. 2013).

The well-characterized CADASIL mouse model TgNotch3^R169C, overexpressing a naturally-occurring mutation in patients, has been an invaluable tool in our comprehension of how TIMP3 accumulation impairs cerebral blood flow (CBF) hemodynamics at an early stage of the disease. Specifically, we have found that elevated extracellular TIMP3 acts through inhibition of a disintegrin and metalloprotease 17 (ADAM17) to inhibit ectodomain shedding of the epidermal growth factor receptor (EGFR) ligand, the heparin-binding EGF-like growth factor (HB-EGF) (Capone et al. 2016a; 2016b). In healthy brain vascular smooth muscle cells (VSMCs), EGFR activation promotes subtype 1 voltage-dependent K⁺ (K_V1) channel down-regulation. The suppression of this pathway in CADASIL results in an elevated K_V1 current density (Dabertrand et al. 2015; Capone et al. 2016b), essentially from K_V1.5 channels, that abnormally hyperpolarizes VSMC membrane potential and acts as a brake on contraction (Nelson and Quayle 1995; Faraci and Heistad 1998).

Small arteries and arterioles within the brain, and other organs, exist in a partially constricted state, termed myogenic tone, in response to the luminal pressure exerted on the vessel walls by the blood (Bayliss 1902). The primary mechanism underlying myogenic tone involves VSMC membrane depolarization and subsequent Ca²⁺ entry through voltage-dependent calcium channels, which activates the contractile machinery (Harder 1984; Knot and Nelson 1998). As blood pressure increases, the arteriole constricts and vice-versa, which allows a constant perfusion to be maintained over a range of physiological blood pressure (Lassen 1959); myogenic tone also protects downstream capillaries against potentially disruptive high blood pressure and, in the brain, it creates the vasodilatory reserve necessary to locally increase CBF to match neuronal activation (Nishimura et al. 2007; Fernández-Klett et al. 2010; Shih et al. 2013). In the CADASIL mice, myogenic tone is blunted by K_V1 channel up-regulation in surface (pial) arteries and cortical parenchymal arterioles as K⁺ channels oppose the plasma membrane depolarization (Joutel et al. 2010; Dabertrand et al. 2015; Capone et al. 2016b; Koide et al. 2018). We found that decreasing K_V1 current density to levels measured in control animals through EGFR activation by exogenous HB-EGF is sufficient to restore myogenic tone in the aforementioned arterioles (Dabertrand et al. 2015; Capone et al. 2016b).

CADASIL patients develop progressive cognitive decline and brain atrophy over the course of the disease, with the latter being a strong predictor of the former (Chabriat et al. 2009). In particular, patients exhibit reduced cerebral blood flow (Fujiwara et al. 2012) and hypometabolism in the limbic system (Su et al. 2019; Schoemaker et al. 2019), and hippocampal atrophy is a strong predictor of cognitive decline in CADASIL (O’Sullivan et al. 2009). Recent studies have highlighted the specificity of the hippocampal microcirculation in regard of the blood brain barrier integrity during vascular cognitive impairment (Montagne et al. 2020) or sepsis (Zhang et al. 2019), but have also pointed to potential differences in the regulation of myogenic tone (Johnson and Cipolla 2016; Johnson, Miller and Cipolla 2020). Here we sought to investigate the impact of the CADASIL-causing mutation on membrane potential and the myogenic response from hippocampal parenchymal arterioles (HiPAs) in CADASIL mice.

Materials and Methods:

Animal Model.

All experimental protocols used in this study were in accord with the institutional guidelines approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Colorado, Anschutz Medical Campus. TgNotch3^WT (WT) and TgNotch3^R169C (CADASIL) mouse lines have been previously described (Joutel et al. 2010). Six-month-old male mice were euthanized by i.p. injection of sodium pentobarbital (100mg/kg) followed by rapid decapitation. We focus on mice at this age because this is at an early stage of the disease progression and for the sake of comparison with our relevant previous studies (Joutel et al. 2010; Dabertrand et al. 2015; Capone et al. 2016b). TgNotch3^WT and TgNotch3^R169C mice (on an FVB/N background) overexpress rat wild-type NOTCH3 and the CADASIL-causing NOTCH3(R169C) mutant protein, respectively, to a similar degree (approximately fourfold) compared with the levels of endogenous NOTCH3 in Non-Tg mice. Expression of CADASIL-causing mutations at normal endogenous levels produce a less severe CADASIL-like phenotype and with a later age of onset, likely because the slowly developing mutant phenotype is unable to manifest during the short lifespan of a mouse. Overexpression of the mutant protein overcomes this constraint and is thus a key feature of this model (Ayata 2010; Joutel et al. 2010).

Diameter Measurements.

Hippocampal parenchymal arterioles (HiPAs) were obtained as previously described (Pires, Dabertrand and Earley 2016; Rosehart, Johnson and Dabertrand 2019). Briefly, after euthanasia, the brain was removed and placed into chilled (4°C) MOPS-buffered saline (composition:135 mM NaCl, 5 mM KCl, 1 mM KH₂PO₄, 1 mM MgSO₄, 2.5 mM CaCl₂, 5 mM glucose, 3 mM MOPS, 0.02 mM EDTA, 2 mM pyruvate, 10 mg/mL bovine serum albumin, pH 7.3 at 4°C). Arterioles that emanate from the anterior, middle, and posterior hippocampal arteries were dissected free of the surrounding tissue, cannulated on borosilicate glass micropipettes (with one end occluded) in an organ chamber (University of Vermont Instrumentation and Model Facility), and pressurized using an arteriography system (Living Systems Instrumentation, Inc., St. Albans, VT, USA). Vessel internal diameter was continuously monitored using a CCD camera and edge-detection software (IonOptix, Westwood, MA, USA). HiPAs were superfused (4 mL/min) with prewarmed (36.5°C ± 1°C), gassed (5% CO₂, 20% O₂, 75% N₂) artificial cerebrospinal fluid (aCSF) for at least 30 min. The composition of aCSF was 125 mM NaCl, 3 mM KCl, 26 mM NaHCO₃, 1.25 mM NaH₂PO₄, 1 mM MgCl₂, 4 mM glucose, 2 mM CaCl₂, pH 7.3 at room temperature with gas aeration. Passive diameter was obtained in nominally Ca²⁺-free aCSF (0 mM [Ca²⁺]_o with 5 mM EGTA). Arteriolar tone was calculated with the following equation: [(passive diameter – active diameter)/(passive diameter)] × 100.

Arteriole Biomechanical Properties Calculation.

Considering HiPAs as thick-walled cylindrical tubes, we used the following equations over the pressure-diameter range of 20 to 80 mmHg, as previously described (Coulson et al. 2002). Circumferential stress θσ(r) was calculated using the following equation:

$$\theta \sigma (r)=\frac{r_i{}^2{p}_i}{r_e{}^2-r_i{}^2}\left(1+\frac{r_i{}^2}{r^2}\right)$$

where r is the arteriolar radius, r_i is the internal radius, r_e is the external radius, and p_i is the luminal pressure. Circumferential strain ɛθ(r) was calculated using the following equation:

$$\epsilon \theta (r)=\frac{r-r_0}{r_0}$$

where r₀is the radius at 20 mmHg. Overall passive stiffness β was calculated using the following equation:

$$\text{ln}\left(\frac{P}{P_s}\right)=\beta \left(\frac{d}{d_s}-1\right)$$

where P is the internal pressure, P_s is the reference pressure chosen in the physiological pressure range (60 mmHg), d is the external diameter, and d_s is the external diameter of the vessel at the reference pressure.

Drug Experiments.

Only viable HiPAs, defined as those that developed pressure-induced myogenic tone greater than 15% at 40 mmHg, were used in the experiments. Drug compounds were mixed to their respective molarities and infused with aliquots of aCSF before they were added to the chamber bath. Effects of drug exposure on vessel diameter were recorded using measurements described above. Endothelial function was tested by assessing the vasodilator response to NS309 (1 μM), an activator of endothelial small- and intermediate-conductance, calcium-sensitive K⁺ channels (SK and IK, respectively). HiPAs not responding to NS309, typically with a near max vasodilation as displayed in Fig. 3A and 3B, were discarded from the study. NS309 dilation was calculated as % of maximum dilation using the following equation [(NS309 diameter - baseline diameter)/(0 Ca²⁺ diameter - baseline diameter)]. Changes in arterial diameter were calculated as the percentage change from baseline using the following equation [(change in diameter)/(initial diameter)] × 100.

Fig. 3. HiPAs from CADASIL mice are hypersensitive to voltage-gated potassium (KV) channel blocker 4-Aminopyridine (4-AP).

(A) Representative traces of pressurized (40 mmHg) HiPAs from WT and CADASIL mice showing NS309-induced dilation (1 μM) and 4-AP-induced constriction (1 mM). (B, C, and D) Summary data from 7 WT and 7 CADASIL mice (**P < 0.01; n.s. not significant; unpaired t-test).

Arteriolar Smooth Muscle Membrane Potential Recordings.

HiPAs were pressurized as described above. Membrane potential was recorded in myocytes, using glass microelectrodes filled with 0.5 M KCl (tip resistance, 100–150 MΩ). A WPI Intra 767 amplifier (Sarasota, FL, USA) was used for recording membrane potential. Analog output from the amplifier was obtained using AxoScope (Molecular Devices, Sunnyvale, CA, USA) software (sample frequency, ~400 Hz). Criteria for accepting recordings were 1) an abrupt negative deflection of potential as the microelectrode was advanced into the cell, 2) stable membrane potential for at least 1 minute, and 3) an abrupt change in potential to approximately 0 mV after retracting the electrode from the cell. For each PA, membrane potential was recorded in at least 2 different cells and averaged.

Reagents.

NS309 and paxilline were purchased from Tocris Bioscience (USA); All other chemicals and reagents were obtained from Sigma-Aldrich (USA). The vehicle for HB-EGF solutions was 0.2-μm–filtered PBS containing 0.1% BSA.

Statistical Analysis.

Data in figures and text are presented as means ± standard error of the mean (SEM). Statistical testing was performed using GraphPad Prism 8 software. All data passed the Kolmogorov–Smirnov test for normality. Statistical significance was determined using either unpaired two-tailed Student’s t-test or two-way repeated measures ANOVA followed by Bonferroni post-hoc test. A maximum of two HiPAs per mouse were used throughout the study.

Results:

To extend our previous studies on pial arteries and cortical parenchymal arterioles to the hippocampal microcirculation, we first examined diameter changes of isolated and pressurized hippocampal parenchymal arterioles (HiPAs) from 6-month-old mice overexpressing wild-type Notch3 (WT) or mutated Notch3^R169C (CADASIL) transgenes in response to increases in luminal pressure (Joutel et al. 2010; Dabertrand et al. 2015; Capone et al. 2016b). As observed in cortical arterioles, HiPAs displayed similar values of myogenic tone, around 10%, when subjected to a luminal pressure of 20 mmHg (Fig. 1). Increasing the pressure to 40 mmHg, the blood pressure experienced in vivo by arterioles of this size (Baumbach, Sigmund and Faraci 2003), caused constriction in both groups. However, HiPAs from CADASIL mice constricted significantly less than the WT HiPAs, showing myogenic tone percentage values of 20.1 ± 1.1 vs 33.6 ± 4.2, respectively, a significant difference which persisted at 60 and 80 mmHg (Fig. 1). Hence these data suggest that the cerebrovascular dysfunction observed in pial and cortical arteries/arterioles also targets vessels within deeper structures of the brain.

Fig. 1. Myogenic responses are decreased in hippocampal parenchymal arterioles (HiPAs) from CADASIL mice.

(A) Typical recordings of HiPA diameter during 20 mmHg luminal pressure increments from 20 to 80 mmHg. Arterioles are first let to equilibrate at 20 mmHg for 20 min before increasing the pressure (blue and red active traces). The same pressure ramp is then conducted in the absence of Ca²⁺ (black passive traces). (B) Summary myogenic tone data from 7 WT and 7 CADASIL mice (*P < 0.05, **P < 0.01; two-way repeated measures ANOVA followed by Bonferroni post-hoc test).

We previously reported that the posterior cerebral artery, which perfuses the hippocampus (Xiong et al. 2017), is smaller in this CADASIL mouse model than in control animals (Joutel et al. 2010; Baron-Menguy et al. 2017). Thus, we then compared arteriolar diameter in active (after development of myogenic tone) and passive (in the absence of extracellular Ca²⁺) conditions (Fig. 2A and 2B). While arterioles from the CADASIL mice appeared to have a smaller diameter at the physiological pressure of 40 mmHg, the difference was not significant in active and passive conditions. We further compared the circumferential stress and strain values for the HiPAs in passive condition over the entire range of pressures and we did not find significant differences. Plotting the circumferential stress-strain relationships showed the typical “J”-shaped curve with a good exponential fit for both WT and CADASIL HiPAs, respectively R² = 0.9921 and R² = 0.9990 (Fig. 2C). The overall stiffness parameter β, calculated in passive conditions, also showed no difference between WT (4.35 ± 0.33) and CADASIL (4.09 ± 0.40) HiPAs (Fig. 2D). Taken together, these data suggest that biomechanical properties are not significantly altered by the CADASIL-causing mutation, at least at this early stage.

Fig. 2. Comparison of HiPAs biomechanical properties between WT and CADASIL mice.

(A) Active diameter, (B) passive diameter, (C) passive circumferential stress-strain relationship from 20 to 80 mmHg, and (D) stiffness (reference value 60 mmHg) of the HiPAs from 7 WT and 7 CADASIL mice used in Fig. 1. Parameters are not significantly different between WT and CADASIL conditions (two-way repeated measures ANOVA followed by Bonferroni post-hoc test).

Our previous work identified that a cerebrospecific up-regulation of voltage-gated potassium channel K_V1.5 in CADASIL VSMCs acts as a brake on pressure-induced depolarization and then limits cortical arteriolar constriction in response to physiological pressures (Nelson and Quayle 1995; Dabertrand et al. 2015). A hint that a higher Kv channel density was blunting the myogenic response was the hypersensitivity of the CADASIL arterioles to the Kv1 blocker 4-aminopyridine (4-AP). We then subjected HiPAs from WT and CADASIL mice to a bath solution containing 1 mM 4-AP (Fig. 3A). HiPAs from CADASIL mice displayed a more pronounced constriction in the presence of 4-AP compared to HiPAs from WT mice (Fig 3C). Interestingly, the percentage of tone values in the presence of 1 mM 4-AP were 45.2 ± 7.8 and 41.4 ± 2.8 for HiPAs from WT and CADASIL mice, respectively, and were not statistically different (P = 0.64). This last result illustrates how inhibiting K_V channels corrects the difference between WT and CADASIL HiPAs by bringing them to a similar tone baseline.

In contrast, 1 μM paxilline, which blocks large-conductance Ca²⁺-sensitive K⁺ (BK) channels, did not induce a significant constriction in HiPAs from WT or CADASIL animals (Fig 3A and 3D). This is consistent with our previous observations that despite expression of BK channels by VSMCs of intraparenchymal arterioles, their activity remains minimal under physiological conditions (Dabertrand, Nelson and Brayden 2012; 2013; Dabertrand et al. 2015). Taken together these results suggest that the CADASIL-causing mutation is also impairing hippocampal microcirculation reactivity via increased K_V channel activity in VSMCs.

Our previous studies demonstrated that activation of the epidermal growth factor receptor (EGFR) down-regulates plasma membrane K_V1 channels, which can be used to rescue cerebrovascular reactivity in the CADASIL mouse model (Dabertrand et al. 2015; Capone et al. 2016b). Exogenous addition of the EGFR agonist heparin-binding EGF-like growth factor (HB-EGF) decreases the elevated number of Kv1.5 channels in VSMCs so that, at 30 ng/mL, it lowers the K_V current density to WT level and therefore restores myogenic tone. We first observed that bath application of 30 ng/mL HB-EGF did not increase myogenic responses of HiPAs from WT mice (Fig 4A and 4B). On the opposite, the presence of HB-EGF nearly doubled the constriction induced by the increase in pressure from 20 to 40 mmHg and, over the range of pressure tested, HiPAs from CADASIL mice reached myogenic tone levels comparable to the ones observed in WT (Fig. 4C and 4D).

Fig. 4. Heparin-binding EGF-like growth factor (HB-EGF) restores myogenic responses in HiPAs from CADASIL mice.

(A) Typical recordings of the internal diameter of the same pressurized WT HiPA subjected to 20 mmHg luminal pressure increments from 20 to 80 mmHg in absence (blue trace) or presence (green trace) of HB-EGF. The solid black line represents the passive arterial diameter measured in the absence of Ca²⁺ at each luminal pressure. (B) Summary myogenic tone data for HiPAs from 5 WT mice during myogenic constriction in response to increasing luminal pressure (n.s. not significant; two-way repeated measures ANOVA followed by Bonferroni post-hoc test). (C) Typical recordings of the internal diameter of the same pressurized CADASIL HiPA in absence (red trace) or presence (purple trace) of HB-EGF. The solid black line represents the passive arterial diameter measured in the absence of Ca²⁺. (D) Summary myogenic tone data for HiPAs from 7 CADASIL mice during myogenic constriction in response to increasing luminal pressure (**P < 0.01, ****P < 0.0001; two-way repeated measures ANOVA followed by Bonferroni post-hoc test).

Compared to WT condition, VSMCs in the walls of cortical arterioles from CADASIL mice display a membrane potential more hyperpolarized by 9.4 mV, at 40 mmHg. The relation between membrane potential and arteriole diameter is very steep and this difference is sufficient to account for the altered myogenic tone (Dabertrand et al. 2015; Koide et al. 2018). However, whether restoration of myogenic tone by exogenous HB-EGF also corrects the membrane potential abnormality has not been investigated so far. Here, a comparison between WT and CADASIL HiPAs pressurized at 40 mmHg revealed a smaller, yet significant, 4.7 mV difference in membrane potential (Fig. 5). Consistent with a K_V1.5 channel down-regulation, application of HB-EGF brought the average membrane potential of HiPAs towards a less negative value, from −36.7 mV to −34.6 mV (Fig. 5A and 5C). Remarkably, this effect presented a larger amplitude in HiPAs from CADASIL mice, depolarizing the membrane potential from −41.4 mV to −34.7 mV (Fig. 5B and 5C). Consistent with the lack of effect of HB-EGF on WT arteriolar tone, statistical analysis revealed that HB-EGF had no effect on membrane potential in WT arterioles whereas it did have an effect in CADASIL HiPAs. In the presence of HB-EGF, CADASIL membrane potential was not significantly different from WT membrane potential, plus or minus HB-EGF (Fig. 5C). Collectively, these results indicate that HB-EGF rescues myogenic tone in CADASIL by shifting ionic conductances to a less depolarized equilibrium.

Fig. 5. HB-EGF hyperpolarizes membrane potential of HiPAs from CADASIL mice more than HiPAs from WT mice.

(A and B) Membrane potential (mV) of VSMCs in pressurized HiPAs at 40 mmHg in absence or presence of HB-EGF from WT (A) or CADASIL (B) mice. (C) Membrane potential summary data expressed as means ± SEM from 6 WT and 5 CADASIL mice (*P < 0.05, ***P < 0.001; two-way repeated measures ANOVA followed by Bonferroni post-hoc test).

Discussion:

In the present study, we investigated the impact of a CADASIL-causing mutation on the myogenic responses of hippocampal arterioles ex vivo. We found that arterioles from CADASIL mouse hippocampi display a diminished pressure-induced constriction compared to WT. This difference can be reverted by bath application of the EGFR agonist, HB-EGF, that acts by correcting the abnormal hyperpolarization of the VSMCs within the arteriolar wall.

Previously, we measured a 60% increase in K_V1 channel current density in cortical arteriolar smooth muscle cells from the same CADASIL mouse model (Dabertrand et al. 2015; Capone et al. 2016b). This channelopathy-like defect constitutes the molecular basis of the blunted myogenic response in this model. As the luminal pressure increases, the VSMCs’ membrane potential depolarizes, which in turn increases the open probability of K_V channels, particularly in the physiological membrane potential range. The resulting K⁺ outward current acts as a negative feedback loop on membrane depolarization, and therefore, as a brake on pressure-induced vasoconstriction (Nelson and Quayle 1995). This explains why 4-AP, by blocking K_V1 channels and then releasing the figurative brake on depolarization, promotes VSMCs contraction. The increased current density in CADASIL causes 4-AP to induce a larger constriction which, in fine, eliminates the difference in myogenic tone, supporting the concept that K_V1 channels are also up-regulated in hippocampal arterioles. This interpretation is consistent with the fact that the difference in the myogenic response between WT and CADASIL HiPAs only emerges when luminal pressure reaches physiological pressures (40 mmHg and above), i.e. when membrane potential depolarizes towards less negative values and activates K_V conductances (Nelson and Quayle 1995). Finally, while an effect of EGFR activation on a depolarizing influence, like TRPM4, could not be ruled out, the depolarization of hippocampal VSMCs induced by HB-EGF in CADASIL HiPAs, strongly support the concept of a pathogenic K_V1 channel up-regulation, as cerebral VSMCs display an EGFR-mediated pathway that promotes K_V1 channel down-regulation (Ishiguro et al. 2006; Koide et al. 2007; Dabertrand et al. 2015).

Our precedent studies established that TIMP3 aggregation in the vascular extracellular matrix is a key contributor to cerebrovascular dysfunction in CADASIL (Monet-Leprêtre et al. 2013; Capone et al. 2016a). As TIMP3 accumulates, it inactivates ADAM17, a metalloproteinase controlling the ectodomain shedding of proHB-EGF, which then leads to less soluble HB-EGF being released (Blobel 2005). The subsequent depressed EGFR activity results in decreased endocytosis, and therefore an increased surface expression, of K_V channels (Dabertrand et al. 2015; Capone et al. 2016b). The restoration of myogenic tone by exogenous HB-EGF in CADASIL HiPAs to WT levels supports the concept that the ADAM17/HB-EGF/EGFR signaling module exists and is also depressed by TIMP3 in the hippocampal arterioles. This observation is of importance as the regulation of myogenic tone varies between hippocampal and cortical arterioles. In two recent and interesting studies on rat HiPAs by Johnson et al., inhibition of endothelial nitric oxide synthase (eNOS) failed to induce vasoconstriction, an observation in stark contrast to cortical arterioles where eNOS exerts a tonic vasodilatory influence (Johnson and Cipolla 2016; Johnson, Miller and Cipolla 2020). Here we found that HiPAs displayed a 15% lower tone, supported by a less depolarized membrane potential, at 40 mmHg compared to our previous studies conducted on cortical arterioles (Dabertrand, Nelson and Brayden 2012; Dabertrand et al. 2013; 2015; De Silva et al. 2018). This decreasing trend does not line up with a silenced eNOS activity but suggests potential differences in myogenic tone regulation, nonetheless. Additional studies should help to clarify the specificity of the hippocampal microcirculation.

In older adults, resting hippocampal blood flow is positively correlated to performance in spatial memory (Heo et al. 2010). Remarkably, a mixed hippocampal vascular supply from both the posterior cerebral artery and the anterior choroidal artery, defined as a larger vascular reserve, is tied to a higher hippocampal volume (Perosa et al. 2020). This appears to be an advantage for hippocampal-centered cognitive tasks and, importantly, SVD patients with mixed vascular supply, and then a larger vascular reserve, score better in verbal memory tests than the ones presenting a single supply pattern (Perosa et al. 2020). Moreover, hippocampal atrophy is a strong predictor of cognitive decline in cerebrovascular diseases (Fein et al. 2000; Zuloaga et al. 2015; van der Flier et al. 2018). Consistent with these observations, in their 2009 study conducted on 144 CADASIL patients, O’Sullivan and colleagues concluded that hippocampal atrophy correlates with progression of cognitive dysfunction independently of global brain volume (O’Sullivan et al. 2009). Our data provide the first direct measurement of a vascular dysfunction in isolated hippocampal microcirculation from a CADASIL mouse model at an early stage of the disease. We found a pronounced trend of HiPAs having a smaller diameter in CADASIL compared to WT, although it did not reach statistical significance. This differs from our previous measurements made in the posterior cerebral artery (Joutel et al. 2010; Baron-Menguy et al. 2017). One explanation could be that the posterior cerebral artery diameter can be measured consistently in the same region in different animals, while our approach, by dissecting arterioles out of the hippocampus, does not allow this reproducibility. The arteriolar diameter decreases along the microvascular arborescence, and this introduces a variability that could mask the difference in diameter. A decrease in the myogenic tone of hippocampal arterioles in CADASIL mice also diminishes the extent to which arterioles can dilate and therefore decreases the vasodilatory reserve. This impairment would then represent a contributing mechanism to the hippocampal dysfunction by greatly limiting local dilation in response to neuronal activity during neurovascular coupling and hippocampal blood flow autoregulation.

Deficits in cognitive domains such as orientation and short-term memory are very common in vascular dementia and line up with functions typically involving the hippocampus. In this regard, as a major cognitive brain area, the hippocampus needs a constant local blood perfusion matching its high metabolic rate, which makes it particularly vulnerable to SVDs and, likely, a starting point for vascular cognitive impairment (van der Flier et al. 2018). Early vascular cognitive impairment pathomechanisms are not clearly identified but appear to include endothelial dysfunction, impairment of autoregulation, and neurovascular uncoupling leading to reduced blood flow and increased blood brain barrier permeability (Sweeney et al. 2019). In CADASIL however, other mechanisms may also be contributing as hippocampal neurogenesis, a process necessary to learning and memory, appears impaired in the TgNotch3^R169C mouse model used in the present study (Ehret et al. 2015; Klein et al. 2017). Moreover, the effect of Notch3^ECD and TIMP3 aggregates at the pericyte level may further impact HiPAs reactivity during capillary-to-arteriole electrical signaling in neurovascular coupling (Joutel et al. 2010; Longden et al. 2017).

In conclusion, our study supports the concept that arterioles in the hippocampus of CADASIL mice display a reduced myogenic tone comparable to what is measurable in pial and cortical arterioles (Joutel et al. 2010; Dabertrand et al. 2015; Capone et al. 2016b). The hypersensitivity towards 4-AP and the abnormal VSMCs hyperpolarization suggest that K_V1 channel up-regulation underlies this defect as well. The rescue of myogenic tone by HB-EGF also supports the concept that the regulation of K_V1 channel trafficking by EGFR is inhibited. Here we further show that HB-EGF restores myogenic tone by causing a depolarization of a larger amplitude in CADASIL than in WT VSMCs. Our findings provide potential therapeutic targets to increase hippocampal blood flow.

Highlights.

• Functional study of isolated and pressurized mouse hippocampal arterioles

• Hippocampal arterioles from CADASIL model mouse display lower myogenic responses

• Biomechanical properties of the CADASIL hippocampal arterioles are not modified at an early stage of the disease

• CADASIL-specific K_V1 channel up-regulation abnormally hyperpolarizes myocytes

• Epidermal growth factor receptor agonist, HB-EGF, depolarizes hippocampal myocytes to restore myogenic tone in CADASIL