Brain ACE2 Activation Following Brain Aminopeptidase A 1 Blockade by Firibastat in Salt-dependent Hypertension 2

In the brain, aminopeptidase A (APA), a membrane-bound zinc metalloprotease, generates 30 angiotensin III from angiotensin II. Brain angiotensin III exerts a tonic stimulatory effect on 31 the control of blood pressure in hypertensive rats and increases vasopressin release. Blocking 32 brain angiotensin III formation by the APA inhibitor prodrug RB150/firibastat normalizes 33 arterial blood pressure in hypertensive DOCA-salt rats without inducing angiotensin II 34 accumulation. We therefore hypothesized that another metabolic pathway of brain angiotensin 35 II, such as the conversion of angiotensin II into angiotensin 1-7 by angiotensin-converting 36 enzyme 2 (ACE2) might be activated following brain APA inhibition. We found that the 37 intracerebroventricular administration of RB150/firibastat in conscious DOCA-salt rats both 38 inhibited brain APA activity and induced an increase in brain ACE2 activity. Then, we 39 showed that the decreases in blood pressure and arginine-vasopressin release resulting from 40 brain APA inhibition with RB150/firibastat were reduced if ACE2 was concomitantly 41 inhibited by MLN4760, a potent ACE2 inhibitor, or if the Mas receptor (MasR) was blocked 42 by A779, a MasR antagonist. Our findings suggest that in the brain, the increase in ACE2 43 activity resulting from APA inhibition by RB150/firibastat treatment, subsequently increasing 44 angiotensin 1-7 and activating the MasR while blocking angiotensin III formation, contributes 45 to the antihypertensive effect and the decrease in arginine-vasopressin release induced by 46 RB150/firibastat. RB150/firibastat treatment constitutes an interesting therapeutic approach to 47 improve BP control in hypertensive patients by inducing in the brain renin-angiotensin 48 system, hyperactivity of the beneficial ACE2/Ang 1-7/MasR axis while decreasing that of the deleterious APA/Ang II/Ang III/ATI


52
All the components of the systemic renin-angiotensin system (RAS), including the precursor 53 and enzymes required for the production and metabolism of angiotensin peptides and type 1 54 (AT1) and type 2 (AT2) angiotensin II (Ang II) receptors, have been identified in the brain 55 (1,2). Among the bioactive peptides of the brain RAS, Ang II and angiotensin III (Ang III) 56 have similar affinities for AT1 and AT2 receptors (3). Studies with selective aminopeptidase 57 A (APA; EC 3.4.11.7) and aminopeptidase N (APN; EC 3.4.11.2) inhibitors (3S)-3-amino-4-58 sulfanyl-butane-1-sulfonic acid (EC33) and (2S)-2-amino-4-methylsulfanyl butane thiol 59 (PC18), respectively (4,5) have shown that APA generates Ang III from Ang II in the brain, 60 whereas APN metabolizes brain Ang III into angiotensin IV (6). By using these tools, we 61 subsequently demonstrated that Ang III, and not Ang II, as established in the periphery, is one 62 decrease results from a decrease in AVP release, a decrease in sympathetic tone and an 77 improvement in baroreflex function (13,14). These observations allowed the selection of 78 RB150/firibastat as the first drug-candidate for clinical development. Clinical phase IIa and 79 IIb trials provide pharmacological proof of principle for the efficacy of brain APA inhibition 80 for decreasing BP in hypertensive patients, especially in Black patients, who are often salt-81 sensitive and where monotherapy with angiotensin I-converting enzyme (ACE) inhibitors or 82 AT1 receptor antagonists is less effective (15,16). 83 The blockade of brain Ang III formation by RB150/firibastat might be expected to lead to 84 higher brain Ang II levels, thereby increasing the stimulation of brain AT1 receptors, 85 offsetting the decrease in such stimulation achieved by lowering brain Ang III levels ( Figure  86 1). However, this does not seem to occur since BP decreased, and we can therefore assume 87 that Ang II is rapidly metabolized by another metabolic pathway to generate a peptide with no 88 affinity for AT1 receptors. The enzyme responsible may be angiotensin-converting enzyme 2 89 (ACE2, EC 3.4.17.23), a membrane-bound zinc metalloprotease that converts Ang II to 90 angiotensin 1-7 (Ang 1-7),which binds to the G-protein-coupled Mas receptor (MasR) with 91 high affinity (17,18) ( Figure 1). We hypothesized that the intracerebroventricular (icv) 92 infusion of an APA inhibitor in hypertensive DOCA-salt rats may, by blocking the conversion 93 of Ang II into Ang III, enhance the conversion of Ang II into Ang 1-7 by ACE2. In agreement 94 with this hypothesis, the chronic central infusion of Ang 1-7 activates the brain MasR, 95 lowering mean arterial BP (MABP) in  and contributing to the NO-96 mediated inhibitory tone of AVP release during hemorrhage (20). Moreover, ACE2 activators 97 such as XNT (1-[(2-dimethylamino) ethylamino]-4-(hydroxymethyl)-7-[(4-methylphenyl) 98 sulfonyloxy]-9H-xanthene-9-one) and diminazene (DIZE) have been shown to decrease BP in 99 different experimental models of hypertension (21)(22)(23). However, the fact that activation of 100 ACE2 by these compounds was responsible for their hypotensive effect was challenged by 101 Haber et al. (24) who showed that the BP-lowering effect of XNT seen in wild-type animals 102 was also observed in ACE2 knockout mice. 103 Our work consisted therefore to investigate whether the blockade of brain APA activity by icv 104 RB150/firibastat injection activated brain ACE2 activity and whether the subsequent 105 RB150/firibastat-induced decrease in BP and AVP release involved the ACE2/Ang 1-7/MasR 106 axis. 107 . 108 Downloaded from http://portlandpress.com/clinsci/article-pdf/doi/10.1042/CS20201385/905546/cs-2020-1385.pdf by guest on 13 March 2021 7 DOCA implantation and we used DOCA-salt rats four weeks after DOCA implantation. On 134 arrival in our animal facility, the animals were housed, three to five animals per cage, under 135 standardized conditions with a 12 h light/12 h dark cycle, at 22°C and 50% humidity. Animals 136 had ad libitum access to water with 0.9% NaCl and 0.2% KCl for DOCA-salt rats or regular 137 water for WKY rats and standard maintenance chow. 138 All of the animal experiments were carried out in accordance with the National Institutes of 139 Health Guide for the Care and Use of Laboratory Animals. WKY rats were assigned 140 randomly to four groups: saline, RB150/firibastat, MLN4760 and A779. DOCA-salt rats were 141 assigned randomly to six groups: saline, RB150/firibastat, MLN4760, RB150/firibastat + 142

Surgical Methods 189
At least 48h before the experiment, rats were anesthetized by the intraperitoneal injection of 190 pentobarbital (50-55 mg/kg). Rats were placed in a stereotaxic apparatus, and a hole was 191 drilled in the right side of the skull (stereotaxic coordinates: 0.8 mm posterior to bregma and 192 1.5 mm lateral to the midline) (30) for the introduction of a guide cannula into the lateral 193 cerebral ventricle. This cannula was inserted to a depth of 3.8 mm below the surface of the 194 skull, as described by Reaux et al. (9). Inhibitors were injected by inserting a 33-gauge 195 stainless steel internal cannula into the guide cannula so that it extended 1 mm beyond the tip 196 of the guide into the lateral ventricle. This injector was connected to a 10-µL Hamilton 197 syringe via polyethylene (PE20) tubing. A left femoral artery catheter (PE50) filled with 198 heparinized saline (250 units/mL) was also inserted for the recording of arterial BP. The 199 catheters were tunneled subcutaneously, to exit from the neck. Then, the animals received for 200 pain relief subcutaneous injection of 1 mg/kg meloxicam (Metacam,Boehringer Ingelheim,201 Germany) and were maintained at a temperature of 37°C throughout their recovery from 202 anesthesia and surgery. The experiments were initiated two days after surgery, to allow the 203 rats time to recover. 204

BP recording 206
Baseline MABP was recorded before drug administration. One hour after the start of BP 207 recording, RB150/firibastat (100 µg, 272 nmol), MLN4760 (5 µg, 11.6 nmol), A779 (10 µg, 208 11.5 nmol) or saline was administered to the conscious unrestrained rats by the icv route. We 209 chose to administer RB150/firibastat via the icv route while this orally administered 210 compound enters the brain because MLN4760 and A779 are not able to enter the brain 211 following systemic administration. After acute icv treatment with RB150/firibastat, 212 MLN4760, A779 or saline, BP was monitored for 4 hours. BP was recorded in conscious 213 unrestrained animals with a PowerLab / We used mice injected by the icv route with exogenous Ang II to increase AVP release from 222 the posterior pituitary in the blood circulation as previously described (6,7) and evaluated the 223 effects of the i.c.v. injections of the APA and ACE2 inhibitors or the MasR antagonist on 224 Ang II -induced AVP release. For this purpose, mice received, by the icv route, a single dose 225 of saline or the following drugs: Ang II (5 ng, 5 pmol), RB150/firibastat (50 µg), MLN4760 226 (5 µg), A779 (500 ng), or Ang II (5 ng), or Ang II (5 ng) + RB150/firibastat (50 µg) + 227 MLN4760 (5 µg), or Ang II (5 ng) + RB150/firibastat (50 µg) + A779 (500 ng), according to 228 the method of Haley and McCormick (31). Mice were killed 1 min after the injection, and 229 trunk blood (0.5-1 mL) was collected in chilled tubes containing 50 µL of 0.3 M EDTA pH 230 7.4. Samples were stored on ice for a maximum of 15 min before centrifugation (2,600 x g for 231 15 min) at 4°C. Plasma (200 µL) was collected and transferred to polypropylene tubes 232 containing 50 µL of 3 M HCl. It was stored at -80°C until AVP radioimmunoassay (RIA) as 233 previously described (6). 234 Two hundred fifty µL trifluoroacetic acid (TFA, 1% final in water) were added to the samples 235 before they were loaded onto the C18 Sep-Pak cartridges first equilibrated with 2 mL 100% 236 acetonitrile followed by 5 mL TFA 1%. The columns were washed with 3 mL TFA1% and 237 AVP was eluted with 1.5 mL 100% acetonitrile. The eluates were dried then redissolved in 238 Results. 255 We first evaluated the effects of the icv administration of RB150/firibastat and the ACE2 256 inhibitor, MLN4760 (32), on brain ACE2 and APA activities in conscious mice and DOCA-257 salt hypertensive rats. We then investigated whether the icv administration of MLN4760 or 258 the MasR antagonist, A779 (33)

MNL4760 on Brain APA and ACE2 Activities in Alert Mice and DOCA-Salt Rats 273
In mice: We assessed the ability of intracerebroventricular (icv) RB150/firibastat to block 274 brain APA activity in alert control mice. The inhibition of brain APA activity ( Figure 3A) by 275 RB150/firibastat (50 µg) peaked at 90% after 5 minutes, as previously described (10), 276 subsequently decreasing to 40% after 30 min (56.8  4.2 vs. 95.0  4.1 nmol of GluNA 277 hydrolyzed per mg of protein per h, P < 0.001). In conscious mice, the icv injection of 278 MLN4760 (5 µg) had no significant effect on brain APA activity, as shown by comparison 279 with mice receiving icv injections of saline (93.4  3.2 nmol of GluNA hydrolyzed per mg of 280 protein per h). We then investigated the effects of MLN4760 on brain ACE2 activity ( Figure  281 3B). Half an hour after the icv administration of 5 µg MLN4760, brain ACE2 activity was 282 90% lower than that in mice receiving an icv injection of saline, (12568 ± 1877 vs 123641 ± 283 7335 RFU per mg protein per h). By contrast, the icv administration of RB150/firibastat (50 284 µg) had no significant effect on brain ACE2 activity (126356 ± 9768 RFU per mg protein per 285 h) after 30 min. 286 In DOCA-salt rats: Brain APA activity in hypertensive DOCA-salt rats was 75% higher than 287 that in normotensive rats (79.6  3.6 vs 45.6  1.0 nmol of GluNA hydrolyzed per mg of 288 protein per h, P < 0.001) ( Figure 3C). The icv injection of RB150/firibastat (100 µg) in 289 DOCA-salt rats significantly decreased brain APA activity, to levels 41% lower than those in 290 DOCA-salt rats receiving icv saline injections, 30 minutes after injection (46.8  2.6 vs. 79.6 291  3.6 nmol of GluNA hydrolyzed per mg of protein per h, P < 0.001). This treatment resulted 292 in values similar to those obtained for the brains of normotensive rats (45.6  1.0 nmol of 293 GluNA hydrolyzed per mg of protein per h). By contrast, the icv injection of MLN4760 (5 294 µg) in DOCA-salt rats had no significant effect on brain APA activity (76.6  3.0 nmol of 295 GluNA hydrolyzed per mg of protein per h), which was similar to that in DOCA-salt rats 296 receiving icv saline injections, 30 minutes after injection. Brain ACE2 activity in 297 hypertensive DOCA-salt rats was 50% lower than that in normotensive rats (119899 ± 17121 298 vs. 239363 ± 37728 RFU per mg protein per h, P<0.001) ( Figure 3D). In DOCA-salt rats, 299 RB150/firibastat (100 µg) administered by the icv route significantly increased brain ACE2 300 activity, by 45% (174149 ± 13462 RFU per mg protein per h, P < 0.05) relative to DOCA-salt 301 rats receiving icv saline injections, resulting in values only slightly lower (-27%) than those 302 obtained for normotensive rats (239363 ± 37728 RFU per mg protein per h). The brain ACE2 303 activity measured 30 minutes after the icv administration of MLN4760 (5 µg) in DOCA-salt 304 rats was 40% lower (72255 ± 8904 RFU per mg protein per h, P < 0.05) than that in DOCA-305 salt rats receiving icv saline injections ( Figure 3D). It would have been interesting to correlate 306 the changes in brain APA and ACE2 activities induced by the icv administration of 307 RB150/firibastat and MLN4760 with the concomitant changes in the release of brain AngII, 308 AngIII and Ang1-7 levels in the synaptic cleft. One of the main difficulties is the low quantity 309 of the neuropeptide released as compared to its intraneuronal steady-state level. Microdialysis 310 sampling technique for in vivo measurement of neuropeptide release was applied with success 311 for peptides such as Met-enkephalin, highly concentrated in some brain structures (34) but 312 more difficult for angiotensins exhibiting very low brain concentration (35,36). 313 314

Effects of MLN4760 or A779 Intracerebroventricularly Injected Alone or in 315
Combination with RB150/firibastat on Arterial Blood Pressure in Conscious WKY or 316

390
The overactivity of the brain RAS including the brain APA/Ang II/Ang III/AT1R and 391 ACE2/Ang 1-7/MasR axes plays a major role in the development and maintenance of 392 hypertension (37,38). We previously showed that the inhibition of brain APA by the icv 393 injection of the APA inhibitor EC33 in hypertensive rats blocks the formation of brain Ang III 394 (6), inducing a decrease in arterial BP, whereas the icv injection of the APN inhibitor EC27 or 395 PC18 induces brain Ang III accumulation (6,7), which in turn causes BP increase (9,10). This 396 effect is mediated by AT1 but not by AT2 receptors since the PC18-induced BP increase was 397 blocked by the co-administration of losartan, an AT1 receptor antagonist but not by that of 398 PD 123319, an AT2 receptor antagonist (9). Furthermore, the PC18-induced BP increase was 399 blocked by pretreatment with EC33, showing the existence in the brain of the endogenous 400 cascade Angiotensin II→Angiotensin III→Angiotensin IV under the action of APA and APN 401 respectively. Altogether these previous data indicate the predominant role for brain Ang III in 402 the control of BP. This role for AngIII has also been demonstrated in the control of AVP 403 release from the posterior pituitary in the bloodstream (6-8). As the blockade of brain AngIII 404 formation by EC33 decreases BP, this suggests that no brain Ang II accumulation occurred 405 because if this had been the case, we should have observed an increase in BP due to AngII-406 induced AT1 receptors activation. Consequently, these data suggest that another metabolic 407 pathway of Ang II must be activated, leading to the conversion of brain Ang II into another 408 Ang fragment with no affinity for AT1 receptors. We hypothesized that this metabolic 409 pathway might correspond to ACE2. With this aim, we first checked in vitro that 410 RB150/firibastat specifically inhibited recombinant mouse APA and that MLN4760 411 specifically inhibited purified mouse ACE2. We also checked that RB150/firibastat and MLN 412 4760 had no inhibitory effects on ACE2 and APA, respectively. We then showed in vivo that 413 the icv injection of RB150/firibastat in conscious control mice inhibited brain APA activity, 414 with no effect on brain ACE2 activity. Conversely, after the icv administration of MLN4760, 415 brain ACE2 activity was 90% lower than that in mice receiving icv injections of saline, with 416 no effect on brain APA activity. These data demonstrate the specificity and selectivity of 417 action of these inhibitors for their respective targets in vivo. These data are in agreement with 418 previous studies showing in vitro, the lack of affinity of RB150 for other zinc 419 metalloproteases involved in the production or metabolism of vasoactive peptides, such as 420 APN, ACE, endothelin-converting enzyme 1 and neutral endopeptidase 24.11, and the 421 absence of binding of RB150 to AT1 and AT2 or endothelin A and B receptors (13) as well as 422 by the lack of affinity of MLN4760 for ACE or carboxypeptidase A (32). 423 We then explored the central effects of these inhibitors in conscious hypertensive DOCA-salt 424 rats. We found that hypertensive DOCA-salt rats exhibited significant higher levels of brain 425 APA activity and significant lower levels of brain ACE2 activity than normotensive rats. 426 These changes are consistent with those reported in previous studies on DOCA-salt rats, 427 SHRs and genetic mouse models of hypertension (13,39-41). Both the higher level of brain 428 APA activity and the lower level of brain ACE2 activity, normally expected to lead to an 429 increase in brain Ang III levels and a reduction in brain Ang 1-7 levels, increase BP in 430 DOCA-salt rats relative to normotensive rats. Indeed both the hyperactivity of the APA/Ang 431 II/Ang III/AT1R axis and the decreased activity of the ACE2/Ang 1-7/MasR axis contribute 432 to the brain RAS hyperactivity and high BP observed in this strain ( Figure 7A) (14,42). The 433 icv injection of RB150/firibastat in alert DOCA-salt rats significantly decreased brain APA 434 hyperactivity relative to DOCA-salt rats receiving icv saline injections, resulting in a 435 normalization of brain APA activity in RB150/firibastat-treated DOCA-salt rats to levels 436 found in normotensive rats. By contrast, under the same experimental conditions, treatment 437 by RB150/firibastat induced a significant increase in brain ACE2 activity relative to DOCA-438 salt rats receiving saline injections, resulting in values slightly lower than those measured in 439 Downloaded from http://portlandpress.com/clinsci/article-pdf/doi/10.1042/CS20201385/905546/cs-2020-1385.pdf by guest on 13 March 2021 the brains of normotensive rats. The decrease in brain ACE2 activity in DOCA-salt 440 hypertensive rats has been shown to be due to ADAM17 upregulation leading to ACE2 441 shedding and decrease in ACE2 activity through activation of AT1 receptors (43-45). 442 Moreover, in a neuronal cell line, AngII was shown to decrease ACE2 mRNA and protein 443 expression through AT1 receptors (46). Considering that 1) AngII and AngIII have similar 444 affinities for AT1 receptors(47-49) and 93% of the angiotensin material released in the PVN, 445 after simulation by veratridine or water deprivation, was in the form of AngIII (50), we may 446 expect in hypertensive rats, that the increase in brain AngII/AngIII-immunoreactivy (51) and 447 brain APA activity (13,39) leading to increased released AngIII by activating AT1 receptors 448 should also increase ADAM17 activity, increasing the shedding of ACE2 and producing a 449 decrease in ACE2 activity. Finally, firibastat treatment by blocking the formation of brain 450 AngIII (10) which cannot anymore activate AT1 receptors, should result in a decrease in 451 ADAM17 activity, reducing the shedding of ACE2 and producing an increase in brain ACE2 452 activity. This remains to be explored. 453 Our results suggest that RB150/firibastat acts through a dual mechanism. First, it normalizes 454 brain APA activity to control levels in normotensive rats and blocks brain Ang III formation, 455 thereby preventing activation of the brain AT1R by Ang III and decreasing BP (6,7,10). 456 Second, it increases brain ACE2 activity to levels close to those found in the control group, 457 thereby increasing the rate of conversion of brain Ang II into Ang 1-7, favoring MasR 458 activation and decreasing BP (43). Both these mechanisms ( Figure 7B) could potentially be 459 involved in the BP decrease induced by RB150/firibastat. 460 Since we previously demonstrated the first mechanism of action (6,7,10), in this study, we 461 sought to demonstrate if the second mechanism might play a role in the antihypertensive 462 effect of RB150/firibastat. For this purpose, we used a pharmacological approach, by means 463 of APA and ACE2 inhibitors (RB150/firibastat and MLN4760, respectively) and a MasR 464 antagonist, A779, alone or in combination with RB150/firibastat, to explore the role of the 465 brain ACE2/Ang 1-7/MasR axis in the RB150/firibastat-induced BP decrease in conscious 466 hypertensive DOCA-salt rats. 467 The inhibition of brain APA activity and the blockade of brain Ang III formation (10)  Altogether, our data suggest that in hypertensive DOCA-salt rats, the RB150/firibastat-480 induced BP decrease, partly blocked by the ACE2 inhibitor or the MasR antagonist is due to: 481 i) blocking the formation of brain Ang III by APA, preventing AT1R stimulation and 482 decreasing BP and ii) increasing ACE2 activity, thereby enhancing the conversion of Ang II 483 into Ang 1-7, which, prevents Ang II-induced AT1R activation and favors Ang1-7-induced 484 MasR activation, leading to BP decrease ( Figure 7B). In agreement with this conclusion, 485 chronic icv Ang 1-7 perfusion was shown to attenuate hypertension in DOCA-salt rats and 486 removing brain ACE2 from cell membrane by ADAM17 contributes to the development of 487 neurogenic hypertension (19,43). 488 Nevertheless, the partial blockade of the firibastat-induced BP decrease by MLN4760 may be 489 due to the activation of another metabolic pathway of brain Ang II, different from ACE2. The 490 enzymes responsible may be a prolylcarboxypeptidase (53) or a prolyloligopeptidase (54)  491 known to convert Ang II in 55). Morover, Ang II may be converted in 492 angiotensin A, itself metabolized in alamandine, a peptide known to decrease BP and AVP 493 release via activation of the Mas-related G protein-coupled receptor D (42,56,57). 494 Alamandine may also be produced directly from Ang 1-7 by decarboxylation of the Asp 495 residue (42). These possibilities remain to be investigated. 496 We further explored the involvement of the brain ACE2/Ang 1-7/MasR axis in the decrease in 497 AVP release triggered by icv RB150/firibastat administration. For this purpose, we have used 498 mice icv injected with exogenous Ang II, a model that we previously developed when we 499 evaluated the central effects of APA and APN inhibitors, EC33 and EC27 respectively, on 500 AVP release (6). This study allowed to show that the stimulatory action of Ang II on AVP 501 release depends upon the prior conversion of Ang II to Ang III. 502 In the current study, we showed that the icv injection of exogenous Ang II converted into Ang 503 III by APA doubled plasma AVP levels, whereas the concomitant icv injection of 504 RB150/firibastat with Ang II, which blocks the formation of Ang III from Ang II (10), 505 induced a 81% decrease in Ang II-induced AVP release, as previously shown after EC33 506 treatment (6). Conversely, the icv injection of MLN4760 did not significantly affect basal 507 AVP levels or Ang II-induced AVP release. However, the concomitant icv injection of 508 MLN4760 with Ang II and RB150/firibastat reduced by 76% the inhibitory effect of RB150 509 on Ang II-induced AVP release. Thus, when brain APA activity was blocked by 510 RB150/firibastat, exogenous Ang II could not be converted into Ang III, but in the presence 511 of MLN4760, which blocks brain ACE2 activity, Ang II could not also be converted into Ang 512 1-7, preventing MasR stimulation by Ang 1-7 and thus Ang II accumulates and activates the 513 AT1 receptors. This results in an increase in AVP release in the bloodstream. This explains 514 why the decrease in Ang II-induced AVP release induced by RB150/firibastat was lower in 515 the presence of brain ACE2 inhibition. This conclusion is strenghtened by the observation 516 that the blockade of brain MasR by icv injection of A779 reduced by 75% the inhibitory 517 effect of RB150/firibastat on Ang II-induced AVP release, whereas A779 alone had no 518 significant effect on AVP release. Given the short in vivo half-life of Ang 1-7, it would be 519 interesting to evaluate, in experimental models of hypertension, 520 the effects on arterial BP and AVP release of metabolically-stable Ang 1-7 analogs, built 521 according to the strategy we used to develop active metabolically stable apelin-17 analogs, 522 consisting to the original addition of a fluorocarbon chain to the N terminus of the peptide 523 (58). Alternatively, a recently developed pharmaceutical formulation including angiotensin 524 (Ang)-(1-7) in hydroxypropyl β-cyclodextrin could be also used (59). 525 Otherwise, ACE2 is the entry point for the SARS-CoV-2 strain underlying the COVID-19 526 epidemic (60). The entry of SARS-CoV-2 into the cells through membrane fusion markedly 527 down-regulates ACE2 and decreases ACE2 activity at the cell surface. ACE2 down-528 regulation was proposed to be particularly detrimental in subjects with pre-existing ACE2 529 deficiency associated with hypertension, diabetes and cardiovascular diseases (61,62). The 530 additional ACE2 deficiency after viral invasion might amplify the dysregulation between the 531 'adverse' APA→AngII→AngIII→AT1 receptor axis and the 'protective' ACE2→Ang1-532 7→Mas receptor axis (63). Because of the presence of ACE2 in the CNS, SARS-CoV-2 533 invasion causes brain injury and neurological symptoms (64,65). In this context, firibastat 534 treatment inhibiting brain APA activity and inducing higher brain ACE2 activity may 535 constitute an interesting therapeutic approach in patients with SARS-CoV-2 infection. However, 536 until assessed, it cannot be ruled out that the increase in membrane-bound ACE2 activity does 537 not increase the entry of SARS-CoV2 into cells. 538 In conclusion, these data show that brain APA inhibition following RB150/firibastat 539 treatment, in DOCA-salt rats blocking the conversion of brain Ang II into Ang III, leads to 540 the activation of another metabolic pathway of brain Ang II involving brain ACE2, 541 converting brain Ang II into Ang 1-7, which, by activating the MasR, partially contributes to 542 the antihypertensive effect of RB150/firibastat and its inhibitory effect on AVP release. These 543 findings strengthen further evaluation of RB150/firibastat in hypertensive patients. If the 544 efficacy of firibastat is confirmed in the current pivotal multicenter placebo-controlled Phase 545 III study called FRESH (Firibastat in treatment-RESistant Hypertension) (ClinicalTrials.gov 546 Identifier: NCT04277884), which aims to assess the efficacy of firibastat in 500 hypertensive 547 subjects in whom hypertension remains uncontrolled despite treatment with at least two 548 classes of antihypertensive drugs, firibastat could constitute the first of a new class of 549 centrally acting antihypertensive agents and be beneficial to improve BP control in patients 550 with difficult-to-treat or resistant hypertension. 551 552 Clinical perspectives 553 • We conducted the study to clarify the research question if the blockade of the brain 554 deleterious APA/Ang II/Ang III/ATI R axis by RB150/firibastat increases the activity of the 555 beneficial brain ACE2/Ang 1-7/MasR axis. 556 • We found that the icv administration of RB150/firibastat in hypertensive DOCA-salt rats, 557 both normalized brain APA hyperactivity and induced an increase in brain ACE2 activity. 558 Then, we showed that the decreases in blood pressure and arginine-vasopressin release 559 resulting from brain APA inhibition with RB150/firibastat were reduced if ACE2 was 560 concomitantly inhibited by MLN4760, a potent ACE2 inhibitor, or if the Mas receptor 561 (MasR) was blocked by A779, a MasR antagonist. 562 • Our findings suggest that brain APA inhibition following RB150/firibastat treatment in 563 hypertensive rats by blocking the conversion of brain Ang II into Ang III, leads to the 564 activation of another metabolic pathway of brain Ang II involving brain ACE2, converting 565 brain Ang II into which  expressed in pg/mL plasma. The mean value for the control group was 181 ± 13 pg/mL of 868 plasma. The data shown are the mean values ± SEM. Groups were compared by one-way 869 ANOVA followed by Sidak's multiple comparison tests. *** P < 0.0001 versus control; ** P 870 < 0.001 versus Ang II and * P < 0.05 versus Ang II + RB150/firibastat, ns, not significant. 871 Effects of the Mas receptor antagonist A779 or RB150/firibastat injected alone or in 872 combination, via the icv route, on Ang II-induced AVP release in conscious mice. Mice 873 received 10 µl saline (control, n=32) or Ang II (5 ng, n=21) via the icv route, in the absence 874 or presence of A779 (500 ng, n=17), RB150/firibastat (50 µg, n= 16) or the combination of 875 RB150/firibastat (50 µg) and A779 (500 ng, n=19). (B) Plasma AVP levels were determined 876 1 min after injection, by RIA, and are expressed in pg/mL of plasma. The mean value for the 877 control group was 152 ± 17 pg/mL of plasma. The data shown are mean values ± SEM. The 878 groups were compared by one-way ANOVA followed by Sidak's multiple comparison tests. 879 *** P < 0.0001 versus control; ** P < 0.001 versus Ang II and * P < 0.05 versus Ang II + 880 RB150/firibastat, ns, not significant.