Tone in the absence of L-NAME amounted to ∼1.7 mN, present after the normalization procedure. To demonstrate that this increase in tone involves ROK, ring preparations from BAs were treated after the normalization procedure (20–25 min equilibration in PSS see “Methods”) with 3 μmol/L Y27632 for 10 min, followed by addition of 100 μmol/L of the pan-NOS inhibitor L-NAME for another 20 min in the presence of Y27632. Pretreatment with Y27632 did not alter stretch-induced tone (1.69 ± 0.1 mN in the presence of Y27632 versus 1.73 ± 0.28 in non-treated time matched controls), but prevented the increase in L-NAME induced tone, which was 1.66 ± 0.1 mN in the presence of Y27632 and 2.7 ± 0.5 mN in the absence of Y27632 (Figures 1A, C).

We then tested whether ROK inhibition also attenuated the agonist-induced tone by the thromboxaneA₂ analogue, U46619. In the presence of Y27632, the cumulative concentration-response relation was shifted to the right; the pEC₅₀ value was significantly lower with than without Y27632 in time-matched controls (6.3 ± 0.1 logarithmic units vs. 7.1 ± 0.18; p < 0.01; n = 5). Maximal force was also lower (F_max 3.68 ± 0.37 mN with vs. 5.89 ± 1.3 mN without Y27632, p < 0.05; t-test; n = 5; Figure 1B). However, Y27632 at a concentration of 3 μmol/L (concentration able to nearly completely dephosphorylate ROK-site of targeting subunit of MLCP, MYPT1-T853; Figures 5A, D; Figures 8C, D) attenuated U46619 contractile response by only ∼40% compared to complete inhibition of L-NAME induced tone, suggesting that a part of the U46619-induced tone involves one or more additional mechanisms, distinct from ROK activation.

The inhibitory effect of the ROK inhibitor on L-NAME and U46619 induced tone may partially result from reduced phosphorylation of MYPT1-T696 (Puetz et al., 2009), a site that is phosphorylated by several protein kinases including ROK. Phosphorylation of this site inhibits MLCP activity in vitro (Feng et al., 1999). To test the involvement of phosphorylation this site in L-NAME and U46619 induced contractile activity, we used BAs from 2 months old wild type and heterozygous MYPT1-T696A/+ mice, in which the phosphorylation propensity of MYPT1 at T696 was genetically lowered by mutating the threonine-residue into a non-phosphorylatable alanine (MYPT1-T696A/+).

Ring preparations from MYPT1-T696A/+ and WT BA were stimulated after mounting and equilibration in PSS by cumulatively increasing concentrations of U46619. U46619 induced a concentration dependent rise in tone with similar F_max and pEC₅₀ values in mutant and WT littermates (p > 0.05; n = 7–6; Figures 1D–E). After wash-out of the agonist with PSS, the vessels were pre-incubated for 20 min in L-NAME (100 μmol/L), followed by a second U46619 concentration-response relation (Figures 1D, E). In time-matched controls, the second concentration-response relation was obtained in the absence of L-NAME (n = 4–5; Supplementary Figure S2). L-NAME shifted the U46619-concentration-response curve to the left in both WT and mutant BA and increased F_max, whereby neither F_max nor the pEC₅₀ values differed between MYPT1-T696A/+ and WT littermates (p > 0.05; n = 7–6; Figures 1D, E). No significant difference between the first and second U46619 concentration response curve was observed in time matched controls (n = 5–4; Supplementary Figure S2). L-NAME-induced tone was significantly less in mutant BA than in WT (mutant: 1.95 ± 0.1 mN vs. WT 2.82 ± 0.2 mN, n = 7–6; Figures 1D, F). Thus, the effect of the mutation of MYPT1 and Y27632 on L-NAME induced tone were of similar magnitude. All contractile parameters of BAs from MYPT1-T696A/+ and WT mice are given in Table 1.

Next, we tested in separate experiments whether the reduction in tone induced by L-NAME in y-BAs from MYPT1-T696A/+ mice was related to lower phosphorylation levels of the regulatory light chain of myosin (MLC₂₀) under basal conditions. For technical reasons, the vessels were mounted on two wires (ø 25 µm) but not stretched as for mechanical experiments. Basal phosphorylation levels of MLC₂₀-S19 in WT and MYPT1-T696A/+ were similar (Figures 2A, B). In line with the attenuated force, incubation with L-NAME was associated with increased MLC₂₀-S19 phosphorylation in BA of WT, but not in BA of MYPT1-T696A/+.

We further tested, whether observed differences in MLC₂₀-S19 phosphorylation were caused by differences in the phosphorylation of MYPT1 at T696 and T853 and thus in activity of MLCP. Under control conditions, phosphorylation of MYPT1 at T696 (PSS, no treatment) was higher in BA from WT-mice than in those from MYPT1-T696A/+-mice (p < 0.05; unpaired t-test; n = 4; Figures 2A, C). Nominal maximal pMYPT1-T696 phosphorylation was assessed by stimulation with 0.1 μmol/L calyculin, a type1 phosphatase inhibitor (Chen et al., 2015). As expected, MYPT1-T696 phosphorylation in calyculin treated mutant BA amounted to only ∼50% of that in WT BA (Supplementary Figure S3). We did not observe a significant change in phosphorylation of pMYPT1-T853 between WT and mutant mice, under control conditions (PSS, no treatment) or after incubation with L-NAME (Figures 2A, D). This result suggests that this site, which is phosphorylated only by ROK, is not affected by silencing phosphorylation in MYPT1-T696.

To test whether altered PKG activity may contribute to the lower L-NAME induced tone, we determined phosphorylation of MYPT1-S695, which is specifically phosphorylated by PKG. L-NAME decreased phosphorylation of this site, indicated that NO-PKG signaling is constitutively active in this preparation. Phosphorylation of pMYPT1-S695 was similar in WT and mutant BA (Figures 2A, E). These results also indicated that the alanine mutation of T696 did not affect the phosphorylation of the neighboring S695, and further, that the NO-PKG pathway was not affected by the mutation.

The level of phosphorylation of MYPT1-T696 was proposed to define the intrinsic Ca²⁺-sensitivity of the contractile machinery of urinary bladder smooth muscle (Khromov et al., 2009). To test whether this applies to basilar arteries, we obtained Ca²⁺-force relations in α-toxin permeabilized preparations. Permeabilization with α-toxin generates pores into the plasma membrane which allow diffusion of molecules <1 kDa into the intracellular space. This permits intracellular [Ca²⁺] to be controlled by EGTA so that the Ca²⁺-responsiveness of the contractile machinery can be measured without confounding changes in cytosolic [Ca²⁺]. After permeabilization, the preparations were equilibrated for 20 min in pCa >8 relaxing solution and then challenged with cumulatively increasing concentrations of [Ca²⁺] (Figure 3). We found that force in MYPT1-T696A/+ BA was lower than in WT BA at intermediate, but not at maximal [Ca²⁺] (Figure 3; Table 2). The pEC₅₀, the neg. logarithm of concentration of Ca²⁺ for half maximal contraction, and the Hill coefficient of the Ca²⁺-force relation were lower than in WT (Table 2). Thus, under Ca²⁺-clamped conditions, the Ca²⁺-sensitivity was lower in mutant compared to WT vessels, consistent with a reduced inhibitory phosphorylation of MLCP. This result indicates that a lower Ca²⁺-sensitivity contributes to the lower L-NAME-induced tone in mutant y-BA.

In contrast to y-BA, in senescent BA (s-BA, >24 months old) a slow spontaneous rise in force was observed (Figure 4A) after stretching them to IC90 and equilibration in PSS for 20 min during the normalization procedure (see Methods). This tone amounted to 3.5 ± 0.9 mN in s-BAs from WT mice, which was not reduced in s-BA MYPT1-T696A/+ mice (3.1 ± 1.3 mN; Figures 4A, C; n = 8). The preparations from both genotypes were then incubated with 100 μmol/L L-NAME for 20 min, followed by stimulation with increasing concentrations of U46619 (0.001–3 μmol/L, Figure 4A). As in the experiments with y-BA, treatment of s-BA with 100 μM L-NAME increased tone in both mouse lines to a similar extent (4.5 ± 0.5 mN in WT vs. 4.2 ± 0.5 mN in mutant; n = 8). However, unlike the case with y-BA, the mutation did not blunt L-NAME induced tone (Figures 4A, C). U46619 concentration-response curves were similar in s-BAs from both mouse lines (Figures 4A, C, D; Table 3).

In y-BAs, inhibition of ROK blunted completely L-NAME-induced tone. We therefore investigated the effect of Y27632 in WT and MYPT1-T696A/+ s-BAs. U46619 concentration relationships and F_max after Y27632 treatment were similar between s-BAs from both groups (Figures 4B–D).

In s-BAs from MYPT1-T696A/+ mice, the basal pMLC₂₀-S19 and pMYPT1-T696 were lower than in WT arteries (Figures 5A–C). Incubation with 3 μmol/L of Y27632 reduced pMLC₂₀-S19 only in WT s-BAs (Figures 5A, B). pMYPT1-T853 in WT and mutant s-BAs did not differ under basal conditions or L-NAME-, or Y27631-treatment (Figures 5A, D).

Stretch-induced tone and L-NAME induced tone development under resting conditions, i.e., in PSS, is much slower than agonist induced contraction. Such slow force development has been observed in neonatal mice, in which smooth muscle myosin II was genetically ablated and force development was ascribed to non-muscle myosin (NM-II) (Morano et al., 2000). Blebbistatin has been used as a small molecule inhibitor of NM-II (Limouze et al., 2004; Rhee et al., 2006), but its specificity has recently been questioned as it was reported to also inhibit smooth muscle myosin II (Eddinger et al., 2007). The active (−) form of blebbistatin reduced stretch-induced tone and L-NAME induced tone in s-BAs to a similar degree as Y27632 (Figures 6A–C; Table 3). Under (+) blebbistatin treatment stretch-induced tone was 2.0 ± 0.3 mN vs. 2.2 ± 0.3 mN in heterozygous s-BAs from MYPT1-T696A/+ animals and amounted to 3.1 ± 0.3 mN and 3.0 ± 0.2 mN after 20 min treatment with L-NAME (n = 5; Figure 6A; and Table 3). In 2 of 5 experiments application of 30 µmol/L (+) Blebbistatin had no additional effect on tone. The concentration responsiveness and F_max of U46619 in the presence of blebbistatin (−) or (+) were similar to those measured previously in Y27632 treated group (Figures 6A, C, D; Table 3).

Next we tested whether blebbistatin reduced basal MLC₂₀-S19 phosphorylation or inhibitory phosphorylation of MYPT1. Phosphorylation of pMLC₂₀-S19, pMYPT1-T696, and pMYPT1-T853 in the presence of blebbistatin (−) was determined in s-BAs from WT and from MYPT1-T696A/+ mice using phosphospecific antibodies. Neither pMLC₂₀-S19, nor pMYPT1-T696/T853, were reduced by blebbistatin (−) in BAs from both mouse lines (Figures 7A–D). These experiments support the notion that the hyper-contractile phenotype of senescent BAs involves increase in ROK activity and likely non-muscle myosin-II cross-bridge cycling.

Contractile activity requires that NM myosin-II proteins assemble into filaments. A predictor of filament competent NM myosin is the level of phosphorylation of its heavy chains (MHC) at S1943 (Zhang and Gunst, 2017). Here, we tested 1) whether NM-II S1943 phosphorylation was higher in senescent than in young arteries, 2) whether Y27632 decreased phosphorylation of this site and 3) whether the expression pattern of global or IIa and IIb isoforms of NM-II are altered by senescence. Phosphorylation of NM-MHC-S1943 was quantified by Western blot analysis using phosphospecific antibodies. In s-BAs, basal S1943 phosphorylation was significantly higher under control conditions (time matched controls to the Y-27632 treated preparations) than in y-BAs (Figures 8A, B). Incubation with 3 μmol/L Y27632 reduced the immunoreactivity against pNM-II-S1943 in s-BAs, but not in y-BAs (Figures 8A, B). No difference in the expression levels of NM-II-total between y- and s-BAs was observed (Supplementary Figures S5A, B). Interestingly, using the same scan intensity, the intensity of the immunoreactive signal obtained with NM-IIa antibodies was larger than with NM-IIb antibodies, supporting the view that NM-IIa is the predominant isoform in BAs of both ages (Supplementary Figures S5A, C, D).

In these experiments, phosphorylation of MYPT1-T853 as a marker for ROK-activity was monitored in parallel. Phosphorylation of T853 was significantly higher in s-BAs than in y-BAs. In both age groups, Y27632 strongly reduced the immunoreactivity of pMYPT1-T853 (Figures 8 C, D). These experiments indicate that ROK regulates NM-II S1943 phosphorylation.

PAK1 ist another target of ROK, which phosphorylates PAK1 at threonine 432 and thereby activates the enzyme (Zhang et al., 2018). We assessed the phosphorylation level of PAK1 using phosphospecific antibodies against pPAK1-T432 in Western Blots prepared with the same samples we used for analysis of NM-II-S1943- and MYPT1-T853-phosphorylation. The immunoreactive signal of PAK1-T423 in s-BAs was higher than in y-BAs. In both groups, incubation with Y27632 reduced the PAK1-T423 phospho-signal (Figures 8E, F).

PAK1 has several downstream targets involved in regulation of tone. One of them, caldesmon was reported to elicit a contraction without an increase in MLC₂₀ phosphorylation when phosphorylated by PAK1 (Van Eyk et al., 1998). Thus, phosphorylation of caldesmon by PAK1 and reversal of its inhibitory effect on cross-bridge cycling is a potential mechanism that could account for the stretch-induced tone in s-BAs. Unfortunately, there are no commercial phosphospecific antibodies available to test whether caldesmon was phosphorylated by PAK1. Therefore, we investigated the involvement of caldesmon in the regulation of stretch-induced tone in arteries from heterozygous mice, carrying one allele in which Cald1 gene was deleted by homologous recombination of the Cald1 gene (Putz et al., 2021). As homozygosity is prenatally lethal, only heterozygous mice (Cald1^+/−) were employed (Putz et al., 2021). In heterozygous old basilar arteries (20–21 months old; o-BAs) expression of caldesmon was ∼50% lower than in WT mice (n = 4, Figures 9A, B). Interestingly, stretch-induced tone was observed in o-BAs obtained from heterozygotes Cald1^+/− mice at an earlier age of 20–21 months, compared to the BAs from controls MYPT1-T696A/+-mice in which such a tone appeared at senescence (age >24 months). At this age, o-BAs from WT mice did not exhibit stretch-induced tone yet (Figures 9C, D). The magnitude of stretch-induced tone of o-BAs from Cald1^+/− mice was similar to that in WT s-BAs (3.5 ± 0.9 mN in s-BAs from WTs vs. 3.5 ± 0.6 mN in o-BAs from Cald1^+/−; n = 8). There was no difference between both genotypes considering L-NAME- and U46619-induced tone (Figures 9C, E). As in previous experiments, application of 3 μmol/L Y27632 reduced stretch-induced tone in both groups (Figures 9C–F). We did not observe significant differences in stretch-induced or L-NAME-induced tone, nor in maximal -induced tone between y-BAs from WT and Cald1^+/− mice (12–14 weeks old) (Supplementary Figure S4).

Inhibition of cross-bridge cycling by caldesmon is exerted by the C-terminal, actin-binding domain (Pfitzer et al., 1993b). Binding to myosin was proposed to position caldesmon in such a way that the actin-binding domain can inhibit cross-bridge cycling (Lee et al., 2000). We tested this in a second mouse line, CaD-ΔEx2^−/−, in which expression of the strong myosin binding domain, which is encoded by Exon 2 of the Cald1 gene, was ablated (Figure 10) (Pfitzer et al., 2005). Expression of the truncated CaD did not differ from WT (Figure 10A). Whereas WT CaD bound to myosin in solution, binding was nearly abolished in truncated caldesmon (CaD-ΔEx2^−/−) (Figure 10B). Binding to actin was not affected (Figure 10B). o-BAs from homozygous mice expressing the truncated caldesmon were subjected to the same protocols used in previous experiments. No significant change of vascular tone in o-BAs from CaD-ΔEx2^−/− mice compared to their time matched WT controls was observed (Figures 10C–F).

The G/F-actin decreased in a ROK-dependent manner by agonist stimulation of airway smooth muscle cells (Zhang and Gunst, 2017) or under basal conditions in old BAs (Lubomirov et al., 2017). To validate that targeting of MYPT1-T696 does not influence F-actin polymerization, we next measured the G/F-actin ratio in s-BAs from WT and MYPT1-T696A/+ animals. No change of G-actin immunoreactivity in supernatants, used as read out for the G-actin fractions, was detected in BAs from WT compared to MYPT1-T696A/+ animals (Figures 11A, B).

In the next series of experiments, we investigated whether maximal stimulation with calyculin resulted in a redistribution of F-actin in favor of subcortical actin. Calyculin treatment induced a maximal increase in force, starting 7–10 min after application and reached its maximum after 30 min (not shown). In the previous experiments, we showed that in BAs, calyculin treatment maximally activates ROK as seen by phosphorylation of MYPT1-T853 (Figure 7A). Alexa Fluor™ 555 phalloidin staining of F-actin filaments showed changes in localization in endothelial and smooth muscle cells. In s-BA endothelial cells F-actin localized to the cortex, while F-actin, stained by Alexa Fluor™ 555 phalloidin, in endothelium of y-BAs was diffuse and could not be attributed to a subcellular niche (Figures 11C, D). Interestingly, in smooth muscle cells from y-BAs F-actin localized predominantly to the cell-poles (Figure 11C).

There is ample evidence that stretch-induced cerebrovascular tone in young arteries relates to two important components, namely, 1) the plasma membrane depolarization, leading to a rise in intracellular Ca²⁺-concentration and activation of the specific Ca²⁺/calmodulin-dependent protein kinase, MLCK, and 2) Ca²⁺-sensitization via inhibition of MLCP (Walsh and Cole, 2013), both inducing an increase in phosphorylation of MLC₂₀. Here, we show that there is a marked difference in the mechanical behavior of y-BAs and s-BAs and its regulation in the absence of a contractile agonist. In both y-BAs and s-BAs, there is a rapid rise in tone followed by a decline to a lower level upon stepwise stretching the vessels to IC90 during the normalization procedure as expected from the length tension relation. The tone at IC90 is stable in y-BAs during the whole experiment, i.e., there is no slow increase in tone, and basal tone recovers after washout of the contractile agonist U46619. In contrast, in s-BAs tone at IC90 is not stable, but rather slowly increases. This rise in tone we denote as stretch-induced tone. In y-BAs, such a stretch-induced tone was observed only after inhibition of NO-signaling by L-NAME.

At both ages, the developed stretch-induced tone and L-NAME induced tone were inhibited by the ROK inhibitor Y27632, putting ROK center stage in regulating tone in basilar arteries, noteworthy, even in the absence of a contractile agonist. Our experiments indicate that not only NO-PKG signaling (Lubomirov et al., 2017; Lubomirov et al., 2018), but also ROK signaling is constitutively active in BAs. In y-BAs, NO-PKG signaling, known to inhibit RhoA-ROK-MLCP signaling at different levels along the pathway, prevents the spontaneous rise in force, while in s-BAs, the balance between NO-PKG and ROK signaling is shifted in favor of the latter. The mechanisms that underlie the constitutive activity of ROK unmasked by inhibition of NO-release are currently unknown. It is possible that stretching the vessels, either by opening of stretch sensitive cation channels (Mederos y Schnitzler et al., 2011) and/or by activating G-protein receptors (Momotani et al., 2011; Takefuji et al., 2013) leads to a Ca²⁺-influx, which then via tyrosine kinase Pyk2 might activate Rho-ROK signaling [(Mills et al., 2015); reviewed in (Brozovich et al., 2016)]. It is of interest that in aged renal vessels phosphorylation and activity of 90 kDa ribosomal S6 kinase (RSK2) are higher than in young renal vessels (Lubomirov et al., in press). RSK2 is a non-canonical MLC₂₀-kinase (Suizu et al., 2000; Artamonov et al., 2018) and also activates ROK (Shi et al., 2018), and hence may be responsible for stretch-induced tone in s-BAs. We are currently exploring this possibility.

The downstream targets of ROK appear to differ between y-BA and s-BA. In y-BA, ROK appears to act through the canonical pathway, i.e., ROK-induced inhibitory phosphorylation of MYPT1 at T696 and T853, resulted in an increase in phosphorylation of MLC₂₀ [reviewed in Puetz et al. (2009)]. We do not exclude the possibility that a rise in intracellular Ca²⁺-concentration also leads to MLC₂₀ phosphorylation and increase in tone. The ROK inhibitor, Y27632, lowers phosphorylation of both sites, thus increasing the activity of MLCP, which in turn lowers MLC₂₀ phosphorylation. Y27632 prevents the increase in L-NAME-induced force. However, which one of the two residues or perhaps both are responsible for MLCP inhibition is under debate (Feng et al., 1999; Khromov et al., 2009). To address this, we used a mouse model (MYPT1-T696A/+ mice), in which the phosphorylation propensity of T696 was genetically lowered. In y-BAs, the mutation prevented L-NAME induced tension and phosphorylation of MLC₂₀, as did Y27632.

Surprisingly, stretch-induced tone in s-BAs, although ROK dependent, appears not to be mediated by MLC₂₀ phosphorylation. This notion is based on the observation that basal MLC₂₀ phosphorylation but not stretch-induced tone was reduced in s-BAs carrying the MYPT1-T696A/+ mutation. In other words, stretch-induced tone develops in s-BAs despite low levels of MLC₂₀ phosphorylation. As the vessels for phosphorylation determination were not stretched because of technical reasons, albeit pharmacological treatment was identical to the mechanical experiment, we cannot exclude the possibility that stretching the vessel would have increased MLC₂₀-phosphorylation. These data are in contrast to our previous observation, where the T696A mutation reduced both, phosphorylation of MLC₂₀ and tone in young and old cerebral arteries (Lubomirov et al., 2017; Lubomirov et al., 2018). It appears from our study that regulation of cerebrovascular tone in young and old age occurs predominantly through MLC₂₀ phosphorylation, the canonical pathway, whereas in senescence these pathways play a minor role.

Several non-canonical pathways, in addition to the canonical one, have been shown to regulate tone in tracheal (Zhang and Gunst, 2019), and vascular smooth muscle (Kim et al., 2010; Moreno-Dominguez et al., 2013) by acting through both actin and NM-II. A series of recent publications demonstrated that ROK regulates polymerization of subcortical actin at cell adhesomes in response to contractile stimuli (Zhang et al., 2016; Zhang et al., 2018). Actin polymerization and assembly are multistep processes involving a large number of other regulatory proteins.

Furthermore, according to Gunst and co-workers (Zhang et al., 2016; Zhang et al., 2018), key events are RhoA mediated NM-II filament assembly, as well as ROK mediated phosphorylation and activation of PAK1. Our study suggests that these events are also involved in regulating stretch-induced tone and L-NAME induced contractions in senescence, notably in the absence of a contractile agonist. Phosphorylation of both NM-II and PAK1 was much more prominent in s-BAs than in y-BAs and was reduced by Y27632 in s-BAs along with the attenuation of stretch-induced tone and L-NAME induced tone. We note that global inhibition of phosphatase activity using calyculin resulted in increased subcortical F-actin (Figure 11). Our findings are in keeping with a previous report that myogenic tone was associated with a small increase in filamentous actin (Moreno-Dominguez et al., 2013). Different from airway smooth muscle, where these phosphorylation events play an important role in agonist induced contraction at young age [reviewed in Zhang et al. (2015)], they appear to be silenced in cerebral arteries in young age, but they are markers of cerebrovascular senescence and intrinsically active, i.e., in the absence of a contractile agonist.

In addition to activation of signaling molecules that eventually lead to actin polymerization, PAK1 also works through the actin binding protein caldesmon. Phosphorylation of caldesmon reverses its inhibitory effect on actomyosin MgATPase activity thereby increasing force in skinned smooth muscle at low levels of MLC₂₀ phosphorylation (Van Eyk et al., 1998). Technically, it was not possible to monitor PAK1 mediated phosphorylation of caldesmon due to lack of commercial antibodies. Nor was it possible to inhibit caldesmon-phosphorylation by PAK1. Therefore, we used BAs from heterozygous mice, in which the Cald1 gene was ablated, thus reducing the caldesmon content to ∼50% of WT mice (Figure 9). The decrease in caldesmon level was expected to reduce its inhibitory activity. Old BAs from Cald1^+/− mice exhibited stretch-induced tone at an age in which stretch-induced tone was not yet present in WT. Similar to our findings, downregulation of caldesmon in carotid arteries ex vivo by siRNA resulted in cross-bridge cycling in unstimulated tissue (Smolock et al., 2009). It is noteworthy that caldesmon immunoreactivity dramatically decreased as a long-term consequence of subarachnoid hemorrhage (Oka et al., 1996; Doi et al., 1997) a status, which is typically associated with a hyper-contractile response of the vasculature. Our finding corroborates previous findings (Smolock et al., 2009; Putz et al., 2021) showing that caldesmon acts as a molecular brake on tone in aged cerebral vessels (Figure 9). Inhibition of caldesmon by phosphorylation, e.g., by upregulation of ROK-PAK1 signaling cascade in senescence, might therefore be a mechanism that leads to a hyper-contractile state in the absence of a contractile stimulus in senescence. Interestingly, myogenic constriction in young rat cerebral arteries was not associated with altered caldesmon phosphorylation, and was proposed not to be involved in tone regulation (Moreno-Dominguez et al., 2014).

Caldesmon, which is expressed in smooth muscle cells of the walls of visceral organs as well as in the vasculature including the medial layer of human cerebral vessels (Kacem et al., 2006), has two important functional domains (Morgan and Gangopadhyay, 2001). Inhibition of actomyosin interaction is exerted by the C-terminal actin-binding domain (Katsuyama et al., 1992; Pfitzer et al., 1993a), while the N-terminal domain binds to myosin and was proposed to be important for positioning caldesmon (Hemric and Chalovich, 1990; Lee et al., 2000). Blocking of myosin binding by inhibitory peptides resulted in an increase in resting tone in permeabilized arteries (Pfitzer et al., 2005). Ablation of exon2 of the Cald1 gene in mice (CaD-ΔEx2^−/−), which encodes the strong myosin binding domain, resulted in expression of a truncated caldesmon but the expression level did not differ from WT mice. In vitro binding of the truncated caldesmon, isolated from these mice, had lost myosin but not actin binding (Figure 10). However, vascular tone in o-BAs from homozygous CaD-ΔEx2^−/− mice was not different from that in WT mice. This indicates that the C-terminal domain is sufficient for inhibition of force.

Taken together, our work corroborated the notion of the central role ROK has as a signaling node that orchestrates several downstream processes, essential for regulating vascular tone both in y- and s-BAs. Our observations expand previous reports that inhibition of ROK reduced vascular tone of pulmonary, cerebral, coronary, and mesenteric arteries from different species including humans, typically performed in young arteries (Fu et al., 1998; Matsumura et al., 2001; Kitazono et al., 2002; Grisk et al., 2012). However, our work suggests that the downstream ROK activated processes differ especially in senescent cerebral arteries from those in young and old ones. We propose that in y-and o-BAs the canonical pathways, i.e., inhibition of MLCP by ROK and increase in phosphorylation of MLC_20, predominate tone regulation, while in senescence regulation of tone involves ROK related thin filament linked mechanisms.

Several in vivo studies in animal models demonstrated that ROK inhibition prevented coronary and cerebral artery vasospasm and neointimal formation after balloon injury of carotid arteries (Kandabashi et al., 2000; Sato et al., 2000; Sawada et al., 2000). However, ROK inhibition may have a number of unwanted side effects. This is because ROK serves as a master regulator in non-muscle cells, regulating such diverse functions as cell-motility, -migration, and -adhesion, thrombocyte aggregation, tumor cell activity as well as the barrier function of endothelial cells (Aslan and McCarty, 2013; Constantin, 2016; Kataoka and Ogawa, 2016; Shimokawa et al., 2016).

1. Therefore, one goal of the present work was to delineate the ROK substrates NM-II and PAK1-phosphorylation as regulators of spontaneous tone generation in the senescent cerebrovascular system. Based on our study, it is reasonable to conduct further studies, investigating whether targeting of these phosphoproteins would be an approach to treat cerebrovascular dysfunction in the elderly.