bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1 2 HIF-2α is essential for development and function 3 4 5 6 David Macías1, Andrew S. Cowburn1,2, Hortensia Torres-Torrelo3, Patricia 7 Ortega-Sáenz3, José López-Barneo3, and Randall S. Johnson1,4,5. 8 9 1Department of Physiology, Development and Neuroscience. 2Department of 10 Medicine, University of Cambridge, Cambridge, UK; 3Instituto de Biomedicina 11 de Sevilla (IBiS), Seville, Spain. 4Department of Cell and Molecular Biology, 12 Karolinska Institute, Stockholm, Sweden. 13 14 15 16 17 18 19 20 21 22 23 Keywords: , carotid body, sensing, HIF-2α, HVR, physiological 24 responses 25 26 27 28 5For correspondence: Professor Randall S. Johnson, Department of 29 Physiology, Development and Neuroscience, University of Cambridge, 30 Downing Street, Cambridge. CB2 3EG. United Kingdom. E-mail: 31 [email protected] 32

33

1 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

34 Summary

35 Mammalian adaptation to oxygen flux occurs at many levels, from shifts in

36 cellular metabolism to physiological adaptations facilitated by the sympathetic

37 nervous system and carotid body (CB). Interactions between differing forms of

38 adaptive response to hypoxia, including transcriptional responses

39 orchestrated by the Hypoxia Inducible transcription Factors (HIFs), are

40 complex and clearly synergistic. We show here that there is an absolute

41 developmental requirement for HIF-2a, one of the HIF isoforms, for growth

42 and survival of oxygen sensitive glomus cells of the carotid body. The loss of

43 these cells renders mice incapable of ventilatory responses to hypoxia, and

44 this has striking effects on processes as diverse as arterial pressure

45 regulation, exercise performance, and glucose homeostasis. We show that

46 the expansion of the glomus cells is correlated with mTORC1 activation, and

47 is functionally inhibited by rapamycin treatment. These findings demonstrate

48 the central role played by HIF-2a in carotid body development, growth and

49 function.

50

51

52

53

54

55

56

57

58

2 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

59 Introduction

60 Detecting and responding to shifts in oxygen availability is essential for animal

61 survival. Responding to changes in oxygenation occurs in cells, tissues, and

62 at the level of the whole organism. Although many of responses to oxygen flux

63 are regulated at the transcriptional level by the hypoxia inducible transcription

64 factor (HIF), at tissue and organismal levels there are even more dimensions

65 to the adaptive process. In vertebrates, the actors in adaptation are found in

66 different peripheral chemoreceptors, including the neuroepithelial cells of the

67 gills in teleosts, and in the carotid body (CB) of mammals. The CB regulates

68 many responses to changes in oxygenation, including alterations in ventilation

69 and rates.

70

71 The carotid body is located at the carotid artery bifurcation, and contains

72 glomeruli that contain O2-sensitive, neuron-like tyrosine hydroxylase (TH)-

73 positive glomus cells (type I cells). These glomus cells are responsible for a

74 rapid compensatory hypoxic ventilatory response (HVR) when a drop in

75 arterial O2 tension (PO2) occurs (Lopez-Barneo, Ortega-Saenz, et al., 2016;

76 Teppema & Dahan, 2010). In addition to this acute reflex, the CB plays an

77 important role in acclimatization to sustained hypoxia, and is commonly

78 enlarged in people living at high altitude (Arias-Stella & Valcarcel, 1976; Wang

79 & Bisgard, 2002) and in patients with chronic respiratory diseases (Heath &

80 Edwards, 1971; Heath, Smith, & Jago, 1982). This enlargement is induced

81 through a proliferative response of CB stem cells (type II cells)(Pardal,

82 Ortega-Saenz, Duran, & Lopez-Barneo, 2007; Platero-Luengo et al., 2014).

3 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

83 During embryogenesis, the initiating event in CB organogenesis occurs when

84 multipotent neural crest cells migrate toward the dorsal aorta, to give rise to

85 the sympathoadrenal system (Le Douarin, 1986). Segregation of sympathetic

86 progenitors from the superior cervical ganglion (SCG) then creates the CB

87 parenchyma (Kameda, 2005; Kameda, Ito, Nishimaki, & Gotoh, 2008). This

88 structure is comprised of O2-sensitive glomus cells (type-I cells) (Pearse et al.,

89 1973), -like sustentacular cells (type-II cells) (Pardal et al., 2007) and

90 endothelial cells (Annese, Navarro-Guerrero, Rodriguez-Prieto, & Pardal,

91 2017).

92

93 The cellular response to low oxygen is in part modulated by two isoforms of

94 the HIF-α transcription factor, each with differing roles in cellular and

95 organismal response to oxygen flux: HIF-1α and HIF-2α. At the organismal

96 level, mouse models hemizygous for HIFs-α (Hif-1α+/- and Hif-2α+/-) or where

97 HIF inactivation was induced in adults showed contrasting effects on CB

98 function, without in either case affecting CB morphogenesis (Hodson et al.,

99 2016; Kline, Peng, Manalo, Semenza, & Prabhakar, 2002; Y. J. Peng et al.,

100 2011).

101

102 Transcriptome analysis has confirmed that Hif-2α is the second most

103 abundant transcript in neonatal CB glomus cells (Tian, Hammer, Matsumoto,

104 Russell, & McKnight, 1998; Zhou, Chien, Kaleem, & Matsunami, 2016), and

105 that it is expressed at far higher levels than are found in cells of similar

106 developmental origins, including superior cervical ganglion (SCG) sympathetic

107 neurons (Gao et al., 2017). It has also been recently shown that Hif-2α, but

4 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

108 not Hif-1α, overexpression in sympathoadrenal cells leads to enlargement of

109 the CB (Macias, Fernandez-Aguera, Bonilla-Henao, & Lopez-Barneo, 2014).

110 Here, we report that HIF-2α is required for the development of CB O2-

111 sensitive glomus cells, and that mutant animals lacking CB function have

112 impaired adaptive physiological responses.

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

5 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

133 Results

134 Sympathoadrenal Hif-2α loss blocks carotid body glomus cell

135 development

136 To elucidate the role of HIFα isoforms in CB development and function, we

137 generated mouse strains carrying Hif-1α or Hif-2α embryonic deletions (TH-

138 HIF-1αKO and TH-HIF-2αKO) restricted to catecholaminergic tissues by

139 crossing them with a mouse strain expressing cre recombinase under the

140 control of the endogenous tyrosine hydroxylase (TH) promoter (Macias et al.,

141 2014).

142

143 Histological analysis of the carotid artery bifurcation dissected from adult TH-

144 HIF-2αKO mice revealed virtually no CB TH+ glomus cells (Figure 1A, bottom

145 panels) compared to HIF-2αWT littermate controls (Figure 1A, top panels).

146 Conversely, TH-HIF-1αKO mice showed a typical glomus cell organization,

147 characterized by scattered clusters of TH+ cells throughout the CB

148 parenchyma (Figure 1B, bottom panels) similar to those found in HIF-1αWT

149 littermate controls (Figure 1B, top panels). These observations were further

150 corroborated by quantification of the total CB parenchyma volume (Figure 1D-

151 1F).

152

153 Other catecholaminergic organs whose embryological origins are similar to

154 those of the CB, e.g., the superior cervical ganglion (SCG), do not show

155 significant differences in structure or volume between TH-HIF-2αKO mutants

156 and control mice (Figure 1A-C). This suggests a specific role of HIF-2α in the

157 development of the CB glomus cells, and argues against a global role for the

6 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

158 gene in development of catecholaminergic tissues. Further evidence for this

159 comes from phenotypic characterization of adrenal medulla (AM) TH+

160 chromaffin cells. No major histological alterations were observed in adrenal

161 glands removed from TH-HIF-2αKO and TH-HIF-1αKO mutant mice compared

162 to HIF-2αWT and HIF-1αWT littermate controls (Figure 1G and H). Consistent

163 with this, the amount of catecholamine (adrenaline and noradrenaline) present

164 in the urine of TH-HIF-2αKO and TH-HIF-1αKO deficient mice was similar to

165 that found in their respective littermate controls (Figure 1I).

166

167 To determine deletion frequencies in these tissues, we crossed a loxP-flanked

168 Td-Tomato reporter strain (Madisen et al., 2010) with TH-IRES-Cre mouse

169 mice. Td-Tomato+ signal was only detected within the CB, SCG and AM of

170 mice expressing cre recombinase under the control of the Th promoter

171 (Figure 1-figure supplement 1A). Additionally, SCG and AM from HIF-2αWT,

172 HIF-1αWT, TH-HIF-2αKO and TH-HIF-1αKO were quantified for Hif-2α and Hif-

173 1α deletion efficiency using genomic DNA. As expected, there is a significant

174 level of deletion of Hif-2α and Hif-1α genes detected in TH-HIF-2αKO and TH-

175 HIF-1αKO mutant mice compared to HIF-2αWT and HIF-1αWT littermates

176 controls (Figure 1-figure supplement 1B and C).

177

178 To determine whether absence of CB glomus cells in adult TH-HIF-2αKO mice

179 is a result of impaired glomus cell differentiation or cell survival, we examined

180 carotid artery bifurcations dissected from TH-HIF-2αKO mice at embryonic

181 stage E18.5 (i.e., 1-2 days before birth), and postnatally (P0) (Figure 2A).

182 Differentiated TH+ glomus cells were found in the carotid bifurcation of TH-

7 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

183 HIF-2αKO mutant mice at both stages (Figure 2A). However, there is a

184 significant reduction in the total CB parenchyma volume and in the number of

185 differentiated TH+ glomus cells in TH-HIF-2αKO mice at E18.5, and this

186 reduction is even more evident in newborn mice (Figure 2B and C). Since CB

187 volume and cells number of mutant mice decreased in parallel, no significant

188 changes were seen in TH+ cell density at these two time points, however

189 (Figure 2D).

190

191 Consistent with the results described above (Figure 1C), the SCG volume was

192 not different in mutant newborns relative to wild type controls (Figure 2E). We

193 next assessed cell death in the developing CB, and found a progressive

194 increase in the number of TUNEL+ cells within the CB parenchyma of TH-HIF-

195 2αKO mice compared to HIF-2αWT littermates (Figure 2F and G). These data

196 demonstrate that HIF-2α, but not HIF-1α, is essential for the differentiation of

197 CB glomus cells.

198

199 Impaired ventilatory response and whole-body metabolic activity in TH-

200 HIF-2αKO mice exposed to hypoxia

201 Although TH-HIF-2αKO deficient mice did not show an apparent phenotype in

202 the AM or SCG, it was important to determine whether their secretory or

203 electrical properties were altered (Figure 3-figure supplement 1). The

204 secretory activity of chromaffin cells under basal conditions and in response to

205 hypoxia, and high (20mM) extracellular K+ was not changed in

206 TH-HIF-2αKO mice (8-12 weeks old) relative to control littermates (Figure 3-

207 figure supplement 1A-D). Current-voltage (I-V) relationships for Ca2+ and K+

8 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

208 currents of dispersed chromaffin cells (Figure 3- figure supplement 1E) and

209 SCG neurons (Figure 3-figure supplement 1F) were also similar in HIF-2αWT

210 and TH-HIF-2αKO animals. These data further support the notion that HIF-2α

211 loss in the sympathoadrenal lineages predominantly affects CB glomus cells.

212

213 To study the impact of the TH-HIF-2αKO mutation on respiratory function, we

214 examined ventilation in normoxia (21% O2) and in response to environmental

215 normobaric hypoxia (10% O2) via whole-body plethysmography. Figure 3A

216 illustrates the changes of respiratory rate during a prolonged hypoxic time-

217 course in HIF-2αWT and TH-HIF-2αKO littermates. HIF-2αWT mice breathing

218 10% O2 had a biphasic response, characterized by an initial rapid increase in

219 respiratory rate followed by a slow decline. In contrast, TH-HIF-2αKO mice,

220 which have normal ventilation rates in normoxia (Figure 3A-C), respond

221 abnormally throughout the hypoxic period. Averaged respiratory rate and

222 minute volume during the initial 5 minutes of hypoxia show a substantial

223 reduction in ventilation rate in TH-HIF-2αKO mutant mice relative to HIF-2αWT

224 controls (Figure 3B and C).

225

226 Additional respiratory parameters altered by the lack of CB in response to

227 hypoxia are shown in Figure 3-supplement 2. Consistent with these data,

228 hemoglobin saturation levels dropped to 50-55% in TH-HIF-2αKO mice, as

229 compared to 73% saturation in control animals after exposure to 10% O2 for 5

230 minutes (Figure 3D). HIF-1αWT and TH-HIF-1αKO mice have similar respiratory

231 responses to 21% O2 and 10% O2, which indicates that HIF-1α is not required

9 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

232 for the CB-mediated ventilatory response to hypoxia (Figure 3E-G; Figure 3-

233 supplemental 2F-J).

234 We next determined metabolic activity in HIF-2αWT and TH-HIF-2αKO

235 littermates in a 21% O2 or 10% O2 environment. Circadian metabolic profiles

KO 236 (VO2 and VCO2) were similar in WT and TH-HIF-2α mutant mice (Figure 3H

237 and I). After exposure to hypoxia, however, TH-HIF-2αKO mice showed lower

238 VO2 and VCO2 peaks and had a significantly hampered metabolic adaptation

239 to low oxygen (Figure 3J and K).

240

241 Deficient acclimatization to chronic hypoxia in mice with CBs loss

242 The CB grows and expands during sustained hypoxia, due to the proliferative

243 response of glia-like (GFAP+) neural crest-derived stem cells within the CB

244 parenchyma (Arias-Stella & Valcarcel, 1976; Pardal et al., 2007). Despite the

245 absence of adult TH+ oxygen sensing cells in TH-HIF-2αKO mice, we identified

246 a normal number of CB GFAP+ stem cells in the carotid artery bifurcation of

247 these animals (Figure 4-figure supplement 1A). We then determined the rate

248 at which CB progenitors from TH-HIF-2αKO mice were able to give rise to

249 newly formed CB TH+ glomus cells after chronic hypoxia exposition in vivo.

250 Typical CB enlargement and emergence of double BrdU+ TH+ glomus cells

WT 251 was seen in HIF-2α mice after 14 days breathing 10% O2 (Figure 4-figure

252 supplement 1B-D), as previously described (Pardal et al., 2007). However, no

253 newly formed double BrdU+ TH+ glomus cells were observed in TH-HIF-2αKO

254 mice after 14 days in hypoxia (Figure 4-figure supplement 1B-D). These

255 observations are consistent with the documented role of CB glomus cells as

10 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

256 activators of the CB progenitors proliferation in response to hypoxia (Platero-

257 Luengo et al., 2014).

258

259 HIF-2αWT, HIF-1αWT and TH-HIF-1αKO mice adapted well and survived up to

260 21 days in a 10% O2 atmosphere normally. However, long-term survival of the

KO 261 TH-HIF-2α deficient mice kept in a 10% O2 environment was severely

262 compromised, and only 40% reached the experimental endpoint (Figure 4A).

263 Mouse necropsies revealed internal general haemorrhages, which are most

264 likely the cause of the death. TH-HIF-2αKO mice also showed increased

265 haematocrit and haemoglobin levels relative to HIF-2αWT littermates when

266 maintained for 21 days in a 10% O2 atmosphere (Figure 4B and C).

267 Consistent with this, after 21 days in hypoxia, heart and spleen removed from

268 TH-HIF-2αKO mice were enlarged relative to HIF-2αWT littermates, likely due to

269 increased blood viscosity and extramedullary hematopoiesis, respectively

270 (Figure 4D and E). TH-HIF-1αKO mice did not show significant differences

WT 271 compared to HIF-1α littermates maintained either in a 21% O2 or 10% O2

272 environment (Figure 4-figure supplement 2A-D).

273

274 Chronic exposure to hypoxia induces pulmonary arterial hypertension (PAH)

275 in mice (Cowburn et al., 2016; Stenmark, Meyrick, Galie, Mooi, & McMurtry,

276 2009). The right was found to be significantly enlarged in TH-HIF-

KO 277 2α mice relative to control mice after 21 days of 10% O2 exposure,

278 indicating an aggravation of pulmonary hypertension (Figure 4F). This result

279 was further confirmed by direct measurement of right ventricular systolic

280 pressure (RVSP) on anaesthetised mice (Figure 4G). histology of mice

11 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

281 chronically exposed to 10% O2 (Figure 4H) shows that the percentage of fully

282 muscularized lung peripheral vessels is significantly increased in TH-HIF-2αKO

283 mice relative to controls. The tunica media of lung peripheral vessels was also

284 significantly thicker in TH-HIF-2αKO mice compared to HIF-2αKO littermates

285 after 21 days of hypoxic exposure (Figure 4I). Taken together, these data

286 demonstrate that pulmonary adaptation to chronic hypoxia is highly reliant on

287 a normally functioning CB.

288

289 Cardiovascular homeostasis is altered in mice lacking CBs

290 An important element of the CB chemoreflex is an increase in sympathetic

291 outflow (Lopez-Barneo, Ortega-Saenz, et al., 2016). To determine whether

292 this had been affected in TH-HIF-2αKO mutants, cardiovascular parameters

293 were recorded on unrestrained mice by radiotelemetry. Circadian blood

294 pressure monitoring showed a constitutive hypotensive condition that was

295 more pronounced during the diurnal period (Figure 5A and B) and did not

296 correlate with changes in , subcutaneous temperature or activity

297 (Figure 5C; Figure 5-figure supplement 1A and B). We next subjected

WT KO 298 radiotelemetry-implanted HIF-2α and TH-HIF-2α littermates to a 10% O2

299 environment preceded and followed by a 24 hour period of normoxia (Figure

300 5D-F; Figure 5-figure supplement 1C-E). Normal control HIF-2αWT mice have

301 a triphasic response to this level of hypoxia, characterised by a short (10

302 minutes) initial tachycardia and hypertension, followed by a marked drop both

303 in heart rate and ; a partial to complete recovery is then seen

304 after 24-36 hours (Figure 5D and E; Figure 5-figure supplement 1C and

305 D)(Cowburn, Macias, Summers, Chilvers, & Johnson, 2017). However, while

12 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

306 TH-HIF-2αKO mice experienced similar, but less deep, hypotension 3 hours

307 post hypoxic challenge, they failed to recover, and showed a persistent

308 hypotensive state throughout the hypoxic period (Figure 5D; Figure 5-figure

309 supplement 1C and D). Interestingly, the bradycardic effect of hypoxia on

310 heart rate was abolished in TH-HIF-2αKO animals (Figure 5E). Subcutaneous

311 temperature, an indirect indicator of in the skin, showed

312 an analogous trend to that seen in the blood pressure of TH-HIF-2αKO

313 mutants (Figure 5F). Activity of TH-HIF-2αKO mice during hypoxia was slightly

314 reduced relative to littermate controls (Figure 5-figure supplement 1E).

315

316 The carotid body has been proposed as a therapeutic target for neurogenic

317 hypertension (McBryde et al., 2013; Pijacka et al., 2016). To determine the

318 relationship between CB and hypertension in these mutants, we performed

319 blood pressure recordings in an II (Ang II)-induced experimental

320 hypertension model (Figure 5G). As shown, TH-HIF-2αKO animals showed

321 lower blood pressure relative to HIF-2αWT littermate controls after 7 days of

322 Ang II infusion (Figure 5H; Figure 5-figure supplement 1F and G). Heart rate,

323 subcutaneous temperature and physical activity were unchanged throughout

324 the experiment (Figure 5K-M; Figure 5-figure supplement 1). Collectively,

325 these data highlight the CB as a systemic cardiovascular regulator which,

326 contributes to baseline blood pressure modulation, is critical for

327 cardiovascular adaptations to hypoxia and delays the development of Ang II-

328 induced experimental hypertension.

329

13 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

330 Adaptive responses to exercise and high glucose are affected in mice

331 with CB dysfunction

332 We next questioned whether this mutant, which lacks virtually all CB function,

333 can inform questions about adaptation to other physiological stressors. In the

334 first instance, we assessed metabolic activity during a maximal exertion

335 exercise test. Oxygen consumption and production (VO2 and

WT KO 336 VCO2, respectively) were comparable in HIF-2α and TH-HIF-2α mice at

KO 337 the beginning of the test. However, TH-HIF-2α mutants reached VO2max

338 much earlier than wild type controls, with significantly decreased heat

339 generation (Figure 6A, B and D). There was no shift in energetic substrate

340 usage, as evidenced by the lack of difference in respiratory exchange ratio

341 (RER) (Figure 6C). The time to exhaustion trended lower in mutants, and

342 there was a significant reduction in both aerobic capacity and power output

343 relative to littermate controls (Figure 6E-G). Interestingly, although no

344 significant changes were found in glucose blood content across the

345 experimental groups and conditions (Figure 6H), lactate levels in TH-HIF-2αKO

346 mice were significantly lower than those observed in HIF-2αWT mice after

347 running (Figure 6I).

348

349 We next examined the response of TH-HIF-2αKO mice to hyperglycemic and

350 hypoglycemic challenge. Baseline glucose levels of HIF-2αWT and TH-HIF-

351 2αKO animals were not statistically different under fed and fasting conditions

352 (Figure 6J and K), and metabolic activity during fasting was unaffected (Figure

353 6-figure supplement 1A-D). However, TH-HIF-2αKO mutant mice showed

354 significant changes in glucose tolerance, with no change in insulin tolerance

14 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

355 when compared to HIF-2αWT littermates (Figure 6L and M). TH-HIF-2αKO mice

356 also showed an initially hampered lactate clearance relative to control mice

357 (Figure 6N). Together, these data demonstrate that the mutants demonstrate

358 intriguing roles for the CB in metabolic adaptations triggered by both exercise

359 and hyperglycemia.

360

361 CB development and proliferation in response to hypoxia require

362 mTORC1 activation

363 Hypoxia is an mTORC1 activator that can induce cell proliferation and growth

364 in a number of organs, e.g., the lung parenchyma (Brusselmans et al., 2003;

365 Cowburn et al., 2016; Elorza et al., 2012; Hubbi & Semenza, 2015; Torres-

366 Capelli et al., 2016). Physiological proliferation and differentiation of the CB

367 induced by hypoxia closely resembles CB developmental expansion (Annese

368 et al., 2017; Pardal et al., 2007; Platero-Luengo et al., 2014); therefore a

369 potential explanation for the developmental defects in CB expansion observed

370 in above could be defects in mTORC1 activation.

371

372 To test this hypothesis, we first determined mTORC1 activation levels in the

373 CB of mice maintained at 21% O2 or 10% O2 for 14 days. This was done via

374 immunofluorescent detection of the phosphorylated form of the ribosomal

375 protein S6 (p-S6) (Elorza et al., 2012; Ma & Blenis, 2009). In normoxia, there

376 was a faint p-S6 expression within the TH+ glomus cells of the CB in wild type

377 mice; this contrasts with the strong signal detected in SCG sympathetic

378 neurons (Figure 7A, left panels; Figure 7-figure supplement 1A, left panel).

379 However, after 14 days in a 10% O2 environment, we saw a significantly

15 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

380 increased p-S6 signal within the CB TH+ cells while the SCG sympathetic

381 neurons remained at levels similar to those seen in normoxia (Figure 7A and

382 B; Figure 7-figure supplement 1A and B). TH expression, which is induced by

383 hypoxia (Schnell et al., 2003), was also significantly elevated in CB TH+ cells,

384 and trended higher in SCG sympathetic neurons (Figure 7B; Figure 7-figure

385 supplement 1B).

386

387 We next assessed the role of HIF-2α in this hypoxia-induced mTORC1

388 activation. We began with evaluating expression of p-S6 in late stage

389 embryos. In E18.5 embryos there was a striking accumulation of p-S6; this is

390 significantly reduced in HIF-2α deficient CB TH+ cells in embryos from this

391 stage of development (Figure 7C and D). p-S6 levels in SCG TH+ sympathetic

392 neurons of HIF-2αWT and TH-HIF-2αKO E18.5 embryos were unaffected,

393 however (Figure 7-figure supplement 1C and D).

394

395 TH expression trended lower in both CB and SCG TH+ cells from TH-HIF-

396 2αKO E18.5 embryos relative to HIF-2αWT littermate controls (Figure 7D;

397 Figure 7-figure supplement 1D). These data indicate that HIF-2α regulates

398 mTORC1 activity in the CB glomus cells during development. To determine

399 whether this pathway could be pharmacologically manipulated, we then

400 studied the effect of mTORC1 inhibition by rapamycin on CB proliferation and

401 growth following exposure to environmental hypoxia. mTORC1 activity, as

402 determined by p-S6 staining on SCG and CB TH+ cells after 14 days of

403 hypoxic treatment, was inhibited after in vivo rapamycin administration (Figure

404 7-figure supplement 1E). Interestingly, histological analysis of rapamycin

16 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

405 injected wild type mice maintained for 14 days at 10% O2 showed decreased

406 CB parenchyma volume, TH+ cells number, and showed a dramatic reduction

407 in the number of proliferating double BrdU+ TH+ cells when compared to

408 vehicle-injected mice (Figure 7E-H).

409

410 In addition, we assessed the ventilatory acclimatization to hypoxia (VAH), as a

411 progressive increase in ventilation rates is associated with CB growth during

412 sustained exposure to hypoxia (Hodson et al., 2016). Consistent with the

413 histological analysis above, rapamycin treated mice had a decreased VAH

414 compared to vehicle-injected control mice following 7 days at 10% O2 (Figure

415 7K). Haematocrit levels were comparable between vehicle control and

416 rapamycing treated wild type mice after 7 days in hypoxia (Figure 7-figure

417 supplement 1F).

418

419 Collectively, these data indicate that mTORC1 activation is required for CB

420 expansion in a HIF-2α-dependent manner, and this activation regulates CB

421 growth and subsequent VAH under prolonged low pO2 conditions.

422

17 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

423 Discussion

424 Metabolic homeostasis depends on a complex system of O2-sensing cells and

425 organs. The carotid body is central to this, via its regulation of metabolic and

426 ventilatory responses to oxygen (Lopez-Barneo, Macias, Platero-Luengo,

427 Ortega-Saenz, & Pardal, 2016). At the cellular level, the HIF transcription

428 factor is equally central for responses to oxygen flux. However, how HIF

429 contributes to the development and function of CB O2-sensitivity is still not

430 completely understood. Initial studies performed on global Hif-1α-/- or Hif-2α-/-

431 mutant mice revealed that HIFs are critical for development (Compernolle et

432 al., 2002; Iyer et al., 1998; J. Peng, Zhang, Drysdale, & Fong, 2000; Ryan, Lo,

433 & Johnson, 1998; Scortegagna et al., 2003; Tian et al., 1998), but only HIF-

434 2α-/- embryos showed defects in catecholamine homeostasis, which indicated

435 a potential role in sympathoadrenal development (Tian et al., 1998).

436

437 Investigations using hemizygous Hif-1α+/- and Hif-2α+/- mice showed no

438 changes in the CB histology or overall morphology (Kline et al., 2002; Y. J.

439 Peng et al., 2011). However, our findings demonstrate that embryological

440 deletion of Hif-2α, but not Hif-1α, specifically in sympathoadrenal tissues

+ 441 (TH ) prevents O2-sensitive glomus cell development, without affecting SCG

442 sympathetic neurons or AM chromaffin cells of the same embryological origin.

443 It was previously reported that HIF-2α, but not HIF-1α overactivation in

444 catecholaminergic tissues led to marked enlargement of the CB (Macias et al.,

445 2014). Mouse models of Chuvash polycythemia, which have a global increase

446 in HIF-2α levels, also show CB hypertrophy and enhanced acute hypoxic

447 ventilatory response (Slingo, Turner, Christian, Buckler, & Robbins, 2014).

18 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

448 Thus either over-expression or deficiency of HIF-2α clearly perturbs CB

449 development and, by extension, function.

450

451 Sustained hypoxia is generally thought to attenuate cell proliferation in a HIF-

452 1α-dependent fashion (Carmeliet et al., 1998; Goda et al., 2003; Koshiji et al.,

453 2004). However, environmental hypoxia can also trigger growth of a number

454 of specialized tissues, including the CB (Arias-Stella & Valcarcel, 1976; Pardal

455 et al., 2007) and the pulmonary arteries (Madden, Vadula, & Kurup, 1992). In

456 this context, the HIF-2α isoform seems to play the more prominent role, and

457 has been associated with pulmonary vascular remodeling as well as with

458 bronchial epithelial proliferation (Brusselmans et al., 2003; Bryant et al., 2016;

459 Cowburn et al., 2016; Kapitsinou et al., 2016; Tan et al., 2013; Torres-Capelli

460 et al., 2016).

461

462 CB enlargement induced by hypoxia to a large extent recapitulates the cellular

463 events that give rise to mature O2-sensitive glomus cells in development. In a

464 recent study, global HIF-2α, but not HIF-1α, inactivation in adult mice was

465 shown to impair CB growth and associated VAH caused by exposure to

466 chronic hypoxia (Hodson et al., 2016). Therefore, we used this physiological

467 proliferative response as a model to understand the developmental

468 mechanisms underpinning of HIF-2α deletion in O2-sensitive glomus cells.

469

470 Contrasting effects of HIF-α isoforms on the mammalian target of rapamycin

471 complex 1 (mTORC1) activity can be found in the literature that mirror the

472 differential regulation of cellular proliferation induced by hypoxia in different

19 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

473 tissues (see above) (Brugarolas et al., 2004; Elorza et al., 2012; Liu et al.,

474 2006). Our findings reveal that mTORC1, a central regulator of cellular

475 metabolism, proliferation and survival (Laplante & Sabatini, 2012), is

476 necessary for CB O2-sensitive glomus cell proliferation; and further, that

477 impact in the subsequent enhancement of the hypoxic ventilatory response

478 induced by chronic hypoxia. Importantly, we found that mTORC1 activity is

479 reduced in HIF-2α-deficient embryonic glomus cells, but not in SCG

480 sympathetic neurons, which likely explains the absence of a functional adult

481 CB in these mutants.

482

483 As HIF-2α regulates mTORC1 activity to promote proliferation, this

484 connection could prove relevant in pathological situations; e.g., both somatic

485 and germline Hif-2α gain-of-function mutations have been found in patients

486 with , a catecholamine-secreting tumor of neural crest origin

487 (Lorenzo et al., 2013; Zhuang et al., 2012). Indeed, CB tumors are 10 times

488 more frequent in people living at high altitude (Saldana, Salem, & Travezan,

489 1973). As further evidence of this link, it has been shown that the mTOR

490 pathway is commonly activated in both paraganglioma and

491 pheochromocytoma tumors (Du et al., 2016; Favier, Amar, & Gimenez-

492 Roqueplo, 2015).

493

494 CB chemoreception plays a crucial role in homeostasis, particularly in

495 regulating cardiovascular and respiratory responses to hypoxia. Our data

KO 496 show that mice lacking CB O2-sensitive cells (TH-HIF-2α mice) have

497 impaired respiratory, cardiovascular and metabolic adaptive responses.

20 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

498 Although previous studies have linked HIF-2α and CB function (Hodson et al.,

499 2016; Y. J. Peng et al., 2011), none of them exhibited a complete CB loss of

500 function. This might explain the differing effects observed in hemizygous Hif-

501 2α+/- mice (Hodson et al., 2016; Y. J. Peng et al., 2011) compared to our TH-

502 HIF-2αKO mice.

503 Mice with a specific HIF-1α ablation in CB O2-sensitive glomus cells (TH-HIF-

504 1αKO mice) did not show altered hypoxic responses in our experimental

505 conditions. Consistent with this, hemizygous Hif-1α+/- mice exhibited a normal

506 acute HVR (Hodson et al., 2016; Kline et al., 2002) albeit some otherwise

507 aberrant CB activities in response to hypoxia (Kline et al., 2002; Ortega-

508 Saenz, Pascual, Piruat, & Lopez-Barneo, 2007).

509

510 Alterations in CB development and/or function have recently acquired

511 increasing potential clinical significance and have been associated with

512 diseases such as heart failure, obstructive sleep apnea, chronic obstructive

513 pulmonary disease, sudden death infant syndrome and hypertension

514 (Iturriaga, Del Rio, Idiaquez, & Somers, 2016; Lopez-Barneo, Ortega-Saenz,

515 et al., 2016).

516

517 We describe here the first mouse model with specific loss of carotid body O2-

518 sensitive cells: a model that is viable in normoxia but shows impaired

519 physiological adaptations to both acute and chronic hypoxia. In addition, we

520 describe new and likely important roles for the CB in systemic blood pressure

521 regulation, exercise performance and glucose management. Further

21 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

522 investigation of this mouse model could aid in revealing roles of the carotid

523 body in both physiology and disease.

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

22 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

547 Material and Methods

548 Mice husbandry

549 This research has been regulated under the Animals (Scientific Procedures)

550 Act 1986 Amendment Regulations 2012 following ethical review by the

551 University of Cambridge Animal Welfare and Ethical Review Body (AWERB).

552 Targeted deletion of Hif-1α, Hif-2α and Td-Tomato in sympathoadrenal

553 tissues was achieved by crossing homozygous mice carrying loxP-flanked

554 alleles Hif1atm3Rsjo (Ryan et al., 1998), Epas1tm1Mcs (Gruber et al., 2007) or

555 Gt(ROSA)26Sortm9(CAG-tdTomato)Hze (Madisen et al., 2010) into a mouse strain of

556 cre recombinase expression driven by the endogenous tyrosine hydroxylase

557 promoter (Thtm1(cre)Te)(Lindeberg et al., 2004), which is specific of

558 catecholaminergic cells. LoxP-flanked littermate mice without cre

559 recombinase (HIF-1αflox/flox, HIF-2αflox/flox or Td-Tomatoflox/flox) were used as

560 control group (HIF-1αWT, HIF-2αWT and Td-Tomato, respectively) for all the

561 experiments. C57BL/6J wild type mice were purchased from The Jackson

562 laboratory (Bar Harbor, Maine, US). All the experiments were performed with

563 age and sex matched control and experimental mice.

564

565 Tissue preparation, immunohistochemistry and morphometry

566 Newborn (P0) and adult mice (8-12 weeks old) were killed by an overdose of

567 sodium pentobarbital injected intraperitoneally. E18.5 embryos were collected

568 and submerged into a cold formalin solution (Sigma-Aldrich, Dorset, UK). The

569 carotid bifurcation and adrenal gland, were dissected, fixed for 2 hours with

570 4% paraformaldehyde (Santa Cruz Biotechnology) and cryopreserved (30%

571 sucrose in PBS) for cryosectioning (10 µm thick, Bright Instruments, Luton,

23 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

572 UK). Tyrosine hydroxylase (TH), glial fibrillary acidic protein (GFAP),

573 bromodeoxyuridine (BrdU) and phospho-S6 ribosomal protein (p-S6) were

574 detected using the following polyclonal antibodies: rabbit anti-TH (1:5000,

575 Novus Biologicals, Abindong, UK; NB300-109) for single immunofluorescence

576 or chicken anti-TH (1:1000, abcam, Cambridge, UK; ab76442) for double

577 immunofluorescence, rabbit anti-GFAP (1:500, Dako, Cambridge, UK;

578 Z0334), rat anti-BrdU (1:100, abcam, ab6326), and rabbit anti-phospho-S6

579 ribosomal protein (1:200, Cell Signaling, 2211), respectively. Fluorescence-

580 based detection was visualized using, Alexa-Fluor 488- and 568-conjugated

581 anti-rabbit IgG (1:500, ThermoFisher, Waltham, MA US), Alexa-Fluor 488-

582 conjugated anti-chicken IgG (1:500, ThermoFisher) and Alexa-Fluor 488-

583 conjudaged anti-rat IgG (1:500, ThermoFisher) secondary antibodies.

584 Chromogen-based detection was performed with Dako REAL Detection

585 Systems, Peroxidase/DAB+, Rabbit/Mouse (Dako, K5001) and counterstained

586 with Carrazzi hematoxylin. For the Td-Tomato detection, carotid bifurcation

587 and adrenal gland were dissected, embedded with OCT, flash-frozen into

588 liquid N2 and sectioned (10 µm thick, Bright Instruments). Direct Td-Tomato

589 fluorescence was imaged after nuclear counterstain with 4',6-diamidino-2-

590 phenylindole (DAPI).

591 CB glomus cell counting, area measurements and pixel mean intensity were

592 calculated on micrographs (Leica DM-RB, Wetzlar, Gernany) taken from

593 sections spaced 40 µm apart across the entire CB using Fiji software

594 (Schindelin et al., 2012) as previously reported (Macias et al., 2014). CB and

595 SCG volume was estimated on the same micrographs according to Cavalieri’s

596 principle. Lung tissue was removed in the distended state by infusion into the

24 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

597 trachea of 0.8% agarose and then fixed in 4% paraformaldehyde for paraffin

598 embedding. Lung sections (6 µm thick) were stained with H&E and EVG stain

599 to assess morphology (Merck, Darmstadt, Germany). Vessel medial thickness

600 was measured on micrographs using Fiji software. To determine the degree of

601 muscularization of small pulmonary arteries, serial lung tissue sections were

602 stained with anti-α smooth muscle actin (1:200, Sigma-Aldrich, A2547).

603 Tissue sections were independently coded and quantified by a blinded

604 investigator.

605

606 Gene deletion

607 Genomic DNA was isolated from dissected SCG and AM using DNeasy Blood

608 & Tissue Kit (Qiagen, Hilden, Germany). Relative gene deletion was assayed

609 by qPCR (One-Step Plus Real-Time PCR System; ThermoFisher Scientific)

610 using TaqMan™ Universal PCR Master Mix (ThermoFisher Scientific) and

611 PrimeTime Std qPCR Assay (IDT, Coralville, IA US).

612 For Hif-1α,

613 Hif1_probe: 5’-/56-FAM/CCTGTTGGTTGCGCAGCAAGCATT/3BHQ_1/-3’

614 Hif1_F: 5’-GGTGCTGGTGTCCAAAATGTAG-3’

615 Hif1_R: 5’-ATGGGTCTAGAGAGATAGCTCCACA-3’

616 For Hif-2α,

617 Hif2_probe: 5’-/56-FAM/CCACCTGGACAAAGCCTCCATCAT/3IABkFQ/-3’

618 Hif2_F: 5’-TCTATGAGTTGGCTCATGAGTTG-3’

619 Hif2_R: 5’-GTCCGAAGGAAGCTGATGG-3’

620 Relative gene deletion levels were calculated using the 2ΔΔCT method.

621

25 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

622 Tissue preparation for amperometry and patch clamp experiments

623 Adrenal glands and SCG were quickly removed and transferred to ice-cooled

624 PBS. Dissected AM was enzymatically dispersed in two steps. First, tissue

625 was placed in 3 ml of extraction solution (in mM: 154 NaCl, 5.6 KCl, 10

626 Hepes, 11 glucose; pH 7.4) containing 425 U/ml collagenase type IA and 4-5

627 mg of bovine serum albumin (all from Sigma-Aldrich), and incubated for 30

628 min at 37°C. Tissue was mechanically dissociated every 10 minutes by

629 pipetting. Then, cell suspension was centrifuged (165 g and 15-20ºC for 4

630 min), and the pellet suspended in 3 ml of PBS containing 9550 U/ml of trypsin

631 and 425 U/ml of collagenase type IA (all from Sigma-Aldrich). Cell suspension

632 was further incubated at 37°C for 5 min and mechanically dissociated by

633 pipetting. Enzymatic reaction was stopped by adding ≈10 ml of cold complete

634 culture medium (DMEM/F-12 (21331-020 GIBCO, ThermoFisher Scientific)

635 supplemented with penicillin (100 U/ml)/streptomycin (10 µg/ml), 2 mM L-

636 glutamine and 10% fetal bovine serum. The suspension was centrifuged (165

637 g and 15-20ºC for 4 min), the pellet gently suspended in 200 µl of complete

638 culture medium. Next, cells were plated on small coverslips coated with 1

639 mg/ml poly-L-lysine and kept at 37ºC in a 5% CO2 incubator for settlement.

640 Afterwards, the petri dish was carefully filled with 2 ml of culture medium and

641 returned to the incubator for 2 hours. Under these conditions, cells could be

642 maintained up to 12-14 hours to perform experiments.

643 To obtain SCG dispersed cells, the tissue was placed in a 1.5 ml tube

644 containing enzymatic solution (2.5 mg/ml collagenase type IA in PBS) and

645 incubated for 15 min in a thermoblock at 37°C and 300 rpm. The suspension

646 was mechanically dissociated by pipetting and centrifuged for 3 min at 200 g.

26 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

647 The pellet was suspended in the second enzymatic dispersion (1mg/ml of

648 trypsin in PBS), incubated at 37°C, 300 rpm for 25 min and dissociated again

649 with a pipette. Enzymatic reaction was stopped adding a similar solution than

650 used for chromaffin cells (changing fetal bovine serum for newborn calf

651 serum). Finally, dispersed SCG cells were plated onto poly-L-lysine coated

652 coverslips and kept at the 37ºC incubator. Experiments with SCG cells were

653 done between 24-48h after finishing the culture.

654

655 To prepare AM slices, the capsule and adrenal gland cortex was removed and

656 the AM was embedded into low melting point agarose at when reached 47 ºC.

657 The agarose block is sectioned (200 µm thick) in an O2-saturated Tyrode's

658 solution using a vibratome (VT1000S, Leica). AM slices were washed twice

659 with the recording solution (in mM: 117 NaCl, 4.5 KCl, 23 NaHCO3, 1 MgCl2, 1

660 CaCl2, 5 glucose and 5 sucrose) and bubbled with carbogen (5% CO2 and

661 95% O2) at 37 ºC for 30 min in a water bath. Fresh slices are used for the

662 experiments during 4-5h after preparation.

663

664 Patch-clamp recordings

665 The macroscopic currents were recorded from dispersed mouse AM and SCG

666 cells using the whole cell configurations of the patch clamp technique.

667 Micropipettes (≈3 MΩ) were pulled from capillary glass tubes (Kimax, Kimble

668 Products, Rockwood, TN US) with a horizontal pipette puller (Sutter

669 instruments model P-1000, Novato, CA US) and fire polished with a

670 microforge MF-830 (Narishige, Amityville, NY US). Voltage-clamp recordings

671 were obtained with an EPC-7 amplifier (HEKA Electronik, Lambrecht/Pfalz,

27 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

672 Germany). The signal was filtered (3 kHz), subsequently digitized with an

673 analog/digital converter (ITC-16 Instrutech Corporation, HEKA Electronik) and

674 finally sent to the computer. Data acquisition and storage was done using the

675 /pulsefit software (Heka Electronik) at a sampling interval of 10 µsec.

676 Data analysis is performed with the Igor Pro Carbon (Wavemetrics, Portland,

677 OR US) and Pulse/pulsefit (HEKA Electronik) programs. All experiments were

678 conducted at 30-33°C. Macroscopic Ca2+, Na+, and K+ currents were recorded

679 in dialyzed cells. The solutions used for the recording of Na+ and Ca2+

680 currents contained (in mM): external: 140 NaCl, 10 BaCl2, 2.5 KCl, 10

681 HEPES, and 10 glucose; pH 7.4 and osmolality 300 mOsm/kg; and internal:

682 110 CsCl, 30 CsF, 10 EGTA, 10 HEPES, and 4 ATP-Mg; pH 7.2 and

683 osmolality 285 mOsm/kg. The external solution used for the recording of

+ 684 whole cell K currents was (in mM): 117 NaCl, 4.5 KCl, 23 NaHCO3, 1 MgCl2,

+ 685 2.5 CaCl2, 5 glucose and 5 sucrose. The internal solution to record K

686 currents was (in mM): 80 potassium glutamate, 50 KCl, 1 MgCl2, 10 HEPES,

687 4 MgATP, and 5 EGTA, pH 7.2. Depolarizing steps of variable duration

688 (normally 10 or 50 msec) from a holding potential of -70 mV to different

689 voltages ranging from -60 up to +50 mV (with voltage increments of 10 mV)

690 were used to record macroscopic ionic currents in AM and SCG cells.

691

692 Amperometric recording

693 To perform the experiments, single slices were transferred to the chamber

694 placed on the stage of an upright microscope (Axioscope, Zeiss, Oberkochen,

695 Germany) and continuously perfused by gravity (flow 1–2 ml min-1) with a

696 solution containing (in mM): 117 NaCl, 4.5 KCl, 23 NaHCO3, 1 MgCl2, 2.5

28 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

697 CaCl2, 5 glucose and 5 sucrose. The ‘normoxic’ solution was bubbled with 5%

698 CO2, 20% O2 and 75 % N2 (PO2 ≈ 150 mmHg). The ‘hypoxic’ solution was

699 bubbled with 5% CO2 and 95% N2 (PO2 ≈ 10–20 mmHg). The ‘hypercapnic’

700 solution was bubbled with 20% CO2, 20% O2 and 60 % N2 (PO2 ≈ 150

701 mmHg). The osmolality of solutions was ≈ 300 mosmol kg−1 and pH = 7.4. In

702 the 20 mM K+ solution, NaCl was replaced by KCl equimolarly (20 mM). All

703 experiments were carried out at a chamber temperature of 36°C.

704 Amperometric currents were recorded with an EPC-8 patch-clamp amplifier

705 (HEKA Electronics), filtered at 100 Hz and digitized at 250 Hz for storage in a

706 computer. Data acquisition and analysis were carried out with an ITC-16

707 interface and PULSE/PULSEFIT software (HEKA Electronics). The secretion

708 rate (fC min−1) was calculated as the amount of charge transferred to the

709 recording electrode during a given period of time.

710

711 Catecholamine measurements

712 Catecholamine levels were determined by ELISA (Abnova, Taipei, Taiwan)

713 following manufacture’s instructions. Spontaneous urination was collected

714 from control and experimental mice (8-12 weeks old) at similar daytime

715 (9:00am). HIF-2αWT and TH-HIF-2αKO or HIF-1αWT, TH-HIF-1αKO samples

716 were collected and assayed in different days.

717

718 Tunel assay

719 Cell death was determined on CB sections using Click-iT™ TUNEL Alexa

720 Fluor™ 488 Imaging Assay (ThermoFisher Scientific), following manufacture’s

29 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

721 indications. TUNEL+ cells across the whole CB parenchyma were counted

722 and normalized by area.

723

724 Plethysmography

725 Respiratory parameters were measured by unrestrained whole-body

726 plethysmography (Data Sciences International, St. Paul, MN US). Mice were

727 maintained in a hermetic chamber with controlled normoxic airflow (1.1 l/min)

728 of 21% O2 (N2 balanced, 5000ppm CO2, BOC Healthcare, Manchester, UK)

729 until they were calm (30-40 minutes). Mice were then exposed to pre-mixed

730 hypoxic air (1.1 l/min) of 10% O2 (N2 balanced, 5000ppm CO2, BOC

731 Healthcare) for 30 min (time-course experiments) or 10 min (VAH

732 experiments). Real-time data were recorded and analyzed using Ponemah

733 software (Data Sciences International). Each mouse was monitored

734 throughout the experiment and periods of movements and/or grooming were

735 noted and subsequently removed from the analysis. The transition periods

736 between normoxia and hypoxia (2-3 min) were also removed from the

737 analysis.

738

739 Oxygen saturations

740 Oxygen saturations were measured using MouseOx Pulse Oximeter (Starr

741 Life Sciences Corp, Oakmont, PA US) on anaesthetised (isofluorane) mice.

742 Anaesthetic was vaporized with 2 l/min of 100% O2 gas flow (BOC

743 Healthcare) and then shift to 2 l/min of pre-mixed 10% O2 flow (N2 balanced,

744 5000ppm CO2, BOC Healthcare) for 5 min. Subsequently, gas flow was

745 switched back to 100% O2.

30 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

746

747 Whole-body metabolic assessments

748 Energy expenditure was measured and recorded using the Columbus

749 Instruments Oxymax system (Columbus, OH US) according to the

750 manufacturer’s instructions. Mice were randomly allocated to the chambers

751 and they had free access to food and water throughout the experiment with

752 the exception of fasting experiments. An initial 18-24h acclimation period was

753 disregarded for all the experiments. Once mice were acclimated to the

754 chamber, the composition of the influx gas was switched from 21% O2 to 10%

755 O2 using a PEGAS mixer (Columbus Instruments). For maximum exertion

756 testing, mice were allowed to acclimatize to the enclosed treadmill

757 environment for 15 min before running. The treadmill was initiated at 5

758 meters/min and the mice warmed up for 5 minutes. The speed was increased

759 up to 27 meters/min or exhaustion at 2.5 min intervals on a 15° incline.

760 Exhaustion was defined as the third time the mouse refused to run for more

761 than 3-5 seconds or when was unable to stay on the treadmill. Blood glucose

762 and lactate levels were measured before and after exertion using a StatStrip

763 Xpress (nova biomedical, Street-Waltham, MA US) glucose and lactate

764 meters.

765

766 Chronic hypoxia exposure and in vivo CB proliferation

767 Mice were exposed to a normobaric 10% O2 atmosphere in an enclosed

768 chamber with controlled O2 and CO2 levels for up to 21 days. Humidity and

769 CO2 levels were quenched throughout the experiments by using silica gel

770 (Sigma-Aldrich) and spherasorb soda lime (Intersurgical, Wokingham, UK),

31 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

771 respectively. Haematocrit was measured in a bench-top haematocrit

772 centrifuge. Haemoglobin levels were analyzed with a Vet abc analyzer

773 (Horiba, Kyoto, Japan) according to the manufacture’s instructions. Heart,

774 spleen tissues were removed and their wet weights measured. Right

775 ventricular systolic pressures were assessed using a pressure-volume loop

776 system (Millar, Houston, TX US) as reported before (Cowburn et al., 2016).

777 Mice were anaesthetized (isoflurane) and right-sided heart catheterization

778 was performed through the right jugular vein. were processed for

779 histology as describe above. Analogous procedures were followed for mice

780 maintained in normoxia.

781 For hypoxia-induced in vivo CB proliferation studies, mice were injected

782 intraperitoneally once a week with 50 mg/kg of BrdU (Sigma-Aldrich) and

783 constantly supplied in their drinking water (1mg/ml) as previously shown

784 (Pardal et al., 2007). In vivo mTORC inhibition was achieved by

785 intraperitoneal injection of rapamycin (6 mg kg-1 day-1). Tissues were collected

786 for histology after 14 days at 10% O2 as described above. VAH was recorded

787 in vehicle control and rapamycin-injected C57BL/6J mice after 7 days in

788 chronic hypoxia by plethysmography as explained above.

789

790 Blood pressure measurement by radio-telemetry

791 All radio-telemetry hardware and software was purchased from Data Science

792 International. Surgical implantation of radio-telemetry device was performed

793 following University of Cambridge Animal Welfare and Ethical Review Body’s

794 (AWERB) recommendations and according to manufacturer’s instructions.

795 After surgery, mice were allowed to recover for at least 10 days. All baseline

32 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

796 telemetry data were collected over a 72h period in a designated quiet room to

797 ensure accurate and repeatable results. Blood pressure monitoring during

798 hypoxia challenge was conducted in combination with Columbus Instruments

799 Oxymax system and PEGAS mixer. Mice were placed in metabolic chambers

800 and allowed to acclimate for 24h before the oxygen content of the flow gas

801 was reduced to 10%. Mice were kept for 48h at 10% O2 before being returned

802 to normal atmospheric oxygen for a further 24h. For hypertension studies,

803 mice were subsequently implanted with a micro osmotic pump (Alzet,

804 Cupertino, CA; 1002 model) for Ang II (Sigma-Aldrich) infusion (490 ng min-1

805 kg-1) for 2 weeks. Blood pressure, heart rate, temperature and activity were

806 monitored for a 48h period after 7, 14 and 21 days after osmotic pump

807 implantation.

808

809 Glucose, insulin and lactate tolerant tests

810 All the blood glucose and lactate measurements were carried out with a

811 StatStrip Xpress (nova biomedical) glucose and lactate meters. Blood was

812 sampled from the tail vein. For glucose tolerance test (GTT), mice were

813 placed in clean cages and fasted for 14-15 hours (overnight) with free access

814 to water. Baseline glucose level (0 time point) was measured and then a

815 glucose bolus (Sigma-Aldrich) of 2 g/kg in PBS was intraperitoneally injected.

816 Glucose was monitored after 15, 30, 60, 90, 120 and 150 minutes. For insulin

817 tolerance test, mice were fasted for 4 hours and baseline glucose was

818 measured (0 time point). Next, insulin (Roche, Penzberg, Germany) was

819 intraperitoneally injected (0.75 U/kg) and blood glucose content recorded after

820 15, 30, 60, 90 and 120 minutes. For lactate tolerance test, mice were injected

33 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

821 with a bolus of 2 g/kg of sodium lactate (Sigma-Aldrich) in PBS and blood

822 lactate measured after 15, 30, 60, 90 and 120 minutes.

823

824 Statistical analysis

825 Particular statistic tests and sample size can be found in the figures and figure

826 legends. Data are presented as the mean ± standard error of the mean (SEM).

827 Shapiro-Wilk normality test was used to determine Gaussian data distribution.

828 Statistical significance between two groups was assessed by un-paired t-test in

829 cases of normal distribution with a Welch’s correction if standard deviations

830 were different. In cases of non-normal distribution (or low sample size for

831 normality test) the non-parametric Mann-Whitney U test was used. For multiple

832 comparisons we used two-way ANOVA followed by Fisher’s LSD test. Non-

833 linear regression (one-phase association) curve fitting, using a least squares

834 method, was used to model the time course data in Figs. 3, 5 and

835 Supplemental Fig. S6. Extra sum-of-squares F test was used to compare

836 fittings, and determine whether the data were best represented by a single

837 curve, or by separate ones. Individual area under the curve (AUC) was

838 calculated for the circadian time-courses in Figs. 3, 5; Supplemental Figs. S6,

839 S7, followed by un-paired t-test or non-parametric Mann-Whitney U test as

840 explained above. Statistical significance was considered if p ≤ 0.05. Analysis

841 was carried out using Prism6.0f (GraphPad software, La Jolla, CA US) for Mac

842 OS X.

843

844 Ackowledgements

845 This work is funded by the Wellcome Trust (grant WT092738MA).

34 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

846

847 Competing interests:

848 No competing interest declared.

849

850

851 Author Contributions

852 Conceptualization, D.M. and R.S.J.; Investigation, D.M., A.C., H.T.T., and

853 P.O.S.; Writing-Original Draft, D.M. and R.S.J.; Writing-Review & Editing,

854 D.M., A.C., H.T.T., P.O.S., J.L.B., and R.S.J.; Supervision, D.M., J.L.B. and

855 R.S.J.; Funding Acquisition, R.S.J.

856

857

35 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

858 References 859 Annese, V., Navarro-Guerrero, E., Rodriguez-Prieto, I., & Pardal, R. (2017). 860 Physiological Plasticity of Neural-Crest-Derived Stem Cells in the Adult 861 Mammalian Carotid Body. Cell Rep, 19(3), 471-478. doi: 862 10.1016/j.celrep.2017.03.065

863 Arias-Stella, J., & Valcarcel, J. (1976). Chief cell hyperplasia in the human 864 carotid body at high altitudes; physiologic and pathologic significance. 865 Hum Pathol, 7(4), 361-373.

866 Brugarolas, J., Lei, K., Hurley, R. L., Manning, B. D., Reiling, J. H., Hafen, E., 867 . . . Kaelin, W. G., Jr. (2004). Regulation of mTOR function in response 868 to hypoxia by REDD1 and the TSC1/TSC2 tumor suppressor complex. 869 Genes Dev, 18(23), 2893-2904. doi: 10.1101/gad.1256804

870 Brusselmans, K., Compernolle, V., Tjwa, M., Wiesener, M. S., Maxwell, P. H., 871 Collen, D., & Carmeliet, P. (2003). Heterozygous deficiency of hypoxia- 872 inducible factor-2alpha protects mice against pulmonary hypertension 873 and right ventricular dysfunction during prolonged hypoxia. J Clin 874 Invest, 111(10), 1519-1527. doi: 10.1172/JCI15496

875 Bryant, A. J., Carrick, R. P., McConaha, M. E., Jones, B. R., Shay, S. D., 876 Moore, C. S., . . . Lawson, W. E. (2016). Endothelial HIF signaling 877 regulates pulmonary fibrosis-associated pulmonary hypertension. Am J 878 Physiol Lung Cell Mol Physiol, 310(3), L249-262. doi: 879 10.1152/ajplung.00258.2015

880 Carmeliet, P., Dor, Y., Herbert, J. M., Fukumura, D., Brusselmans, K., 881 Dewerchin, M., . . . Keshert, E. (1998). Role of HIF-1alpha in hypoxia- 882 mediated apoptosis, cell proliferation and tumour angiogenesis. Nature, 883 394(6692), 485-490. doi: 10.1038/28867

884 Compernolle, V., Brusselmans, K., Acker, T., Hoet, P., Tjwa, M., Beck, H., . . . 885 Carmeliet, P. (2002). Loss of HIF-2alpha and inhibition of VEGF impair 886 fetal lung maturation, whereas treatment with VEGF prevents fatal 887 respiratory distress in premature mice. Nat Med, 8(7), 702-710. doi: 888 10.1038/nm721

889 Cowburn, A. S., Crosby, A., Macias, D., Branco, C., Colaco, R. D., 890 Southwood, M., . . . Johnson, R. S. (2016). HIF2alpha-arginase axis is 891 essential for the development of pulmonary hypertension. Proc Natl 892 Acad Sci U S A, 113(31), 8801-8806. doi: 10.1073/pnas.1602978113

893 Cowburn, A. S., Macias, D., Summers, C., Chilvers, E. R., & Johnson, R. S. 894 (2017). Cardiovascular adaptation to hypoxia and the role of peripheral 895 resistance. Elife, 6. doi: 10.7554/eLife.28755

36 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

896 Du, J., Tong, A., Wang, F., Cui, Y., Li, C., Zhang, Y., & Yan, Z. (2016). The 897 Roles of PI3K/AKT/mTOR and MAPK/ERK Signaling Pathways in 898 Human Pheochromocytomas. Int J Endocrinol, 2016, 5286972. doi: 899 10.1155/2016/5286972

900 Elorza, A., Soro-Arnaiz, I., Melendez-Rodriguez, F., Rodriguez-Vaello, V., 901 Marsboom, G., de Carcer, G., . . . Aragones, J. (2012). HIF2alpha acts 902 as an mTORC1 activator through the amino acid carrier SLC7A5. Mol 903 Cell, 48(5), 681-691. doi: 10.1016/j.molcel.2012.09.017

904 Favier, J., Amar, L., & Gimenez-Roqueplo, A. P. (2015). Paraganglioma and 905 phaeochromocytoma: from genetics to personalized medicine. Nat Rev 906 Endocrinol, 11(2), 101-111. doi: 10.1038/nrendo.2014.188

907 Gao, L., Bonilla-Henao, V., Garcia-Flores, P., Arias-Mayenco, I., Ortega- 908 Saenz, P., & Lopez-Barneo, J. (2017). Gene expression analyses 909 reveal metabolic specifications in acute O2 -sensing chemoreceptor 910 cells. J Physiol, 595(18), 6091-6120. doi: 10.1113/JP274684

911 Goda, N., Ryan, H. E., Khadivi, B., McNulty, W., Rickert, R. C., & Johnson, R. 912 S. (2003). Hypoxia-inducible factor 1alpha is essential for cell cycle 913 arrest during hypoxia. Mol Cell Biol, 23(1), 359-369.

914 Gruber, M., Hu, C. J., Johnson, R. S., Brown, E. J., Keith, B., & Simon, M. C. 915 (2007). Acute postnatal ablation of Hif-2alpha results in anemia. Proc 916 Natl Acad Sci U S A, 104(7), 2301-2306. doi: 917 10.1073/pnas.0608382104

918 Heath, D., & Edwards, C. (1971). The carotid body in cardiopulmonary 919 disease. Geriatrics, 26(12), 110-111 passim.

920 Heath, D., Smith, P., & Jago, R. (1982). Hyperplasia of the carotid body. J 921 Pathol, 138(2), 115-127. doi: 10.1002/path.1711380203

922 Hodson, E. J., Nicholls, L. G., Turner, P. J., Llyr, R., Fielding, J. W., Douglas, 923 G., . . . Bishop, T. (2016). Regulation of ventilatory sensitivity and 924 carotid body proliferation in hypoxia by the PHD2/HIF-2 pathway. J 925 Physiol, 594(5), 1179-1195. doi: 10.1113/JP271050

926 Hubbi, M. E., & Semenza, G. L. (2015). Regulation of cell proliferation by 927 hypoxia-inducible factors. Am J Physiol Cell Physiol, 309(12), C775- 928 782. doi: 10.1152/ajpcell.00279.2015

929 Iturriaga, R., Del Rio, R., Idiaquez, J., & Somers, V. K. (2016). Carotid body 930 chemoreceptors, sympathetic neural activation, and cardiometabolic 931 disease. Biol Res, 49, 13. doi: 10.1186/s40659-016-0073-8

37 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

932 Iyer, N. V., Kotch, L. E., Agani, F., Leung, S. W., Laughner, E., Wenger, R. H., 933 . . . Semenza, G. L. (1998). Cellular and developmental control of O2 934 homeostasis by hypoxia-inducible factor 1 alpha. Genes Dev, 12(2), 935 149-162.

936 Kameda, Y. (2005). Mash1 is required for glomus cell formation in the mouse 937 carotid body. Dev Biol, 283(1), 128-139. doi: 938 10.1016/j.ydbio.2005.04.004

939 Kameda, Y., Ito, M., Nishimaki, T., & Gotoh, N. (2008). FRS2 alpha 2F/2F 940 mice lack carotid body and exhibit abnormalities of the superior cervical 941 sympathetic ganglion and carotid sinus nerve. Dev Biol, 314(1), 236- 942 247. doi: 10.1016/j.ydbio.2007.12.002

943 Kapitsinou, P. P., Rajendran, G., Astleford, L., Michael, M., Schonfeld, M. P., 944 Fields, T., . . . Haase, V. H. (2016). The Endothelial Prolyl-4- 945 Hydroxylase Domain 2/Hypoxia-Inducible Factor 2 Axis Regulates 946 Pressure in Mice. Mol Cell Biol, 36(10), 1584-1594. 947 doi: 10.1128/MCB.01055-15

948 Kline, D. D., Peng, Y. J., Manalo, D. J., Semenza, G. L., & Prabhakar, N. R. 949 (2002). Defective carotid body function and impaired ventilatory 950 responses to chronic hypoxia in mice partially deficient for hypoxia- 951 inducible factor 1 alpha. Proc Natl Acad Sci U S A, 99(2), 821-826. doi: 952 10.1073/pnas.022634199

953 Koshiji, M., Kageyama, Y., Pete, E. A., Horikawa, I., Barrett, J. C., & Huang, 954 L. E. (2004). HIF-1alpha induces cell cycle arrest by functionally 955 counteracting Myc. EMBO J, 23(9), 1949-1956. doi: 956 10.1038/sj.emboj.7600196

957 Laplante, M., & Sabatini, D. M. (2012). mTOR signaling in growth control and 958 disease. Cell, 149(2), 274-293. doi: 10.1016/j.cell.2012.03.017

959 Le Douarin, N. M. (1986). Cell line segregation during peripheral nervous 960 system ontogeny. Science, 231(4745), 1515-1522.

961 Lindeberg, J., Usoskin, D., Bengtsson, H., Gustafsson, A., Kylberg, A., 962 Soderstrom, S., & Ebendal, T. (2004). Transgenic expression of Cre 963 recombinase from the tyrosine hydroxylase locus. Genesis, 40(2), 67- 964 73. doi: 10.1002/gene.20065

965 Liu, L., Cash, T. P., Jones, R. G., Keith, B., Thompson, C. B., & Simon, M. C. 966 (2006). Hypoxia-induced energy stress regulates mRNA translation 967 and cell growth. Mol Cell, 21(4), 521-531. doi: 968 10.1016/j.molcel.2006.01.010

38 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

969 Lopez-Barneo, J., Macias, D., Platero-Luengo, A., Ortega-Saenz, P., & 970 Pardal, R. (2016). Carotid body oxygen sensing and adaptation to 971 hypoxia. Pflugers Arch, 468(1), 59-70. doi: 10.1007/s00424-015-1734- 972 0

973 Lopez-Barneo, J., Ortega-Saenz, P., Gonzalez-Rodriguez, P., Fernandez- 974 Aguera, M. C., Macias, D., Pardal, R., & Gao, L. (2016). Oxygen- 975 sensing by arterial chemoreceptors: Mechanisms and medical 976 translation. Mol Aspects Med, 47-48, 90-108. doi: 977 10.1016/j.mam.2015.12.002

978 Lorenzo, F. R., Yang, C., Ng Tang Fui, M., Vankayalapati, H., Zhuang, Z., 979 Huynh, T., . . . Prchal, J. T. (2013). A novel EPAS1/HIF2A germline 980 mutation in a congenital polycythemia with paraganglioma. J Mol Med 981 (Berl), 91(4), 507-512. doi: 10.1007/s00109-012-0967-z

982 Ma, X. M., & Blenis, J. (2009). Molecular mechanisms of mTOR-mediated 983 translational control. Nat Rev Mol Cell Biol, 10(5), 307-318. doi: 984 10.1038/nrm2672

985 Macias, D., Fernandez-Aguera, M. C., Bonilla-Henao, V., & Lopez-Barneo, J. 986 (2014). Deletion of the von Hippel-Lindau gene causes 987 sympathoadrenal cell death and impairs chemoreceptor-mediated 988 adaptation to hypoxia. EMBO Mol Med, 6(12), 1577-1592. doi: 989 10.15252/emmm.201404153

990 Madden, J. A., Vadula, M. S., & Kurup, V. P. (1992). Effects of hypoxia and 991 other vasoactive agents on pulmonary and cerebral artery smooth 992 muscle cells. Am J Physiol, 263(3 Pt 1), L384-393.

993 Madisen, L., Zwingman, T. A., Sunkin, S. M., Oh, S. W., Zariwala, H. A., Gu, 994 H., . . . Zeng, H. (2010). A robust and high-throughput Cre reporting 995 and characterization system for the whole mouse brain. Nat Neurosci, 996 13(1), 133-140. doi: 10.1038/nn.2467

997 McBryde, F. D., Abdala, A. P., Hendy, E. B., Pijacka, W., Marvar, P., Moraes, 998 D. J., . . . Paton, J. F. (2013). The carotid body as a putative 999 therapeutic target for the treatment of neurogenic hypertension. Nat 1000 Commun, 4, 2395. doi: 10.1038/ncomms3395

1001 Ortega-Saenz, P., Pascual, A., Piruat, J. I., & Lopez-Barneo, J. (2007). 1002 Mechanisms of acute oxygen sensing by the carotid body: lessons 1003 from genetically modified animals. Respir Physiol Neurobiol, 157(1), 1004 140-147. doi: 10.1016/j.resp.2007.02.009

39 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1005 Pardal, R., Ortega-Saenz, P., Duran, R., & Lopez-Barneo, J. (2007). Glia-like 1006 stem cells sustain physiologic neurogenesis in the adult mammalian 1007 carotid body. Cell, 131(2), 364-377. doi: 10.1016/j.cell.2007.07.043

1008 Pearse, A. G., Polak, J. M., Rost, F. W., Fontaine, J., Le Lievre, C., & Le 1009 Douarin, N. (1973). Demonstration of the neural crest origin of type I 1010 (APUD) cells in the avian carotid body, using a cytochemical marker 1011 system. Histochemie, 34(3), 191-203.

1012 Peng, J., Zhang, L., Drysdale, L., & Fong, G. H. (2000). The transcription 1013 factor EPAS-1/hypoxia-inducible factor 2alpha plays an important role 1014 in vascular remodeling. Proc Natl Acad Sci U S A, 97(15), 8386-8391. 1015 doi: 10.1073/pnas.140087397

1016 Peng, Y. J., Nanduri, J., Khan, S. A., Yuan, G., Wang, N., Kinsman, B., . . . 1017 Prabhakar, N. R. (2011). Hypoxia-inducible factor 2alpha (HIF-2alpha) 1018 heterozygous-null mice exhibit exaggerated carotid body sensitivity to 1019 hypoxia, breathing instability, and hypertension. Proc Natl Acad Sci U 1020 S A, 108(7), 3065-3070. doi: 10.1073/pnas.1100064108

1021 Peng, Y. J., Yuan, G., Ramakrishnan, D., Sharma, S. D., Bosch-Marce, M., 1022 Kumar, G. K., . . . Prabhakar, N. R. (2006). Heterozygous HIF-1alpha 1023 deficiency impairs carotid body-mediated systemic responses and 1024 reactive oxygen species generation in mice exposed to intermittent 1025 hypoxia. J Physiol, 577(Pt 2), 705-716. doi: 1026 10.1113/jphysiol.2006.114033

1027 Pijacka, W., Moraes, D. J., Ratcliffe, L. E., Nightingale, A. K., Hart, E. C., da 1028 Silva, M. P., . . . Paton, J. F. (2016). Purinergic receptors in the carotid 1029 body as a new drug target for controlling hypertension. Nat Med, 1030 22(10), 1151-1159. doi: 10.1038/nm.4173

1031 Platero-Luengo, A., Gonzalez-Granero, S., Duran, R., Diaz-Castro, B., Piruat, 1032 J. I., Garcia-Verdugo, J. M., . . . Lopez-Barneo, J. (2014). An O2- 1033 sensitive glomus cell-stem cell synapse induces carotid body growth in 1034 chronic hypoxia. Cell, 156(1-2), 291-303. doi: 1035 10.1016/j.cell.2013.12.013

1036 Ryan, H. E., Lo, J., & Johnson, R. S. (1998). HIF-1 alpha is required for solid 1037 tumor formation and embryonic vascularization. EMBO J, 17(11), 3005- 1038 3015. doi: 10.1093/emboj/17.11.3005

1039 Saldana, M. J., Salem, L. E., & Travezan, R. (1973). High altitude hypoxia and 1040 chemodectomas. Hum Pathol, 4(2), 251-263.

1041 Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., 1042 Pietzsch, T., . . . Cardona, A. (2012). Fiji: an open-source platform for

40 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1043 biological-image analysis. Nat Methods, 9(7), 676-682. doi: 1044 10.1038/nmeth.2019

1045 Schnell, P. O., Ignacak, M. L., Bauer, A. L., Striet, J. B., Paulding, W. R., & 1046 Czyzyk-Krzeska, M. F. (2003). Regulation of tyrosine hydroxylase 1047 promoter activity by the von Hippel-Lindau tumor suppressor protein 1048 and hypoxia-inducible transcription factors. J Neurochem, 85(2), 483- 1049 491.

1050 Scortegagna, M., Ding, K., Oktay, Y., Gaur, A., Thurmond, F., Yan, L. J., . . . 1051 Garcia, J. A. (2003). Multiple organ pathology, metabolic abnormalities 1052 and impaired homeostasis of reactive oxygen species in Epas1-/- mice. 1053 Nat Genet, 35(4), 331-340. doi: 10.1038/ng1266

1054 Slingo, M. E., Turner, P. J., Christian, H. C., Buckler, K. J., & Robbins, P. A. 1055 (2014). The von Hippel-Lindau Chuvash mutation in mice causes 1056 carotid-body hyperplasia and enhanced ventilatory sensitivity to 1057 hypoxia. J Appl Physiol (1985), 116(7), 885-892. doi: 1058 10.1152/japplphysiol.00530.2013

1059 Stenmark, K. R., Meyrick, B., Galie, N., Mooi, W. J., & McMurtry, I. F. (2009). 1060 Animal models of pulmonary arterial hypertension: the hope for 1061 etiological discovery and pharmacological cure. Am J Physiol Lung Cell 1062 Mol Physiol, 297(6), L1013-1032. doi: 10.1152/ajplung.00217.2009

1063 Tan, Q., Kerestes, H., Percy, M. J., Pietrofesa, R., Chen, L., Khurana, T. S., . . 1064 . Lee, F. S. (2013). Erythrocytosis and pulmonary hypertension in a 1065 mouse model of human HIF2A gain of function mutation. J Biol Chem, 1066 288(24), 17134-17144. doi: 10.1074/jbc.M112.444059

1067 Teppema, L. J., & Dahan, A. (2010). The ventilatory response to hypoxia in 1068 mammals: mechanisms, measurement, and analysis. Physiol Rev, 1069 90(2), 675-754. doi: 10.1152/physrev.00012.2009

1070 Tian, H., Hammer, R. E., Matsumoto, A. M., Russell, D. W., & McKnight, S. L. 1071 (1998). The hypoxia-responsive transcription factor EPAS1 is essential 1072 for catecholamine homeostasis and protection against heart failure 1073 during embryonic development. Genes Dev, 12(21), 3320-3324.

1074 Torres-Capelli, M., Marsboom, G., Li, Q. O., Tello, D., Rodriguez, F. M., 1075 Alonso, T., . . . Aragones, J. (2016). Role Of Hif2alpha Oxygen Sensing 1076 Pathway In Bronchial Epithelial Club Cell Proliferation. Sci Rep, 6, 1077 25357. doi: 10.1038/srep25357

1078 Wang, Z. Y., & Bisgard, G. E. (2002). Chronic hypoxia-induced morphological 1079 and neurochemical changes in the carotid body. Microsc Res Tech, 1080 59(3), 168-177. doi: 10.1002/jemt.10191

41 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1081 Zhou, T., Chien, M. S., Kaleem, S., & Matsunami, H. (2016). Single cell 1082 transcriptome analysis of mouse carotid body glomus cells. J Physiol, 1083 594(15), 4225-4251. doi: 10.1113/JP271936

1084 Zhuang, Z., Yang, C., Lorenzo, F., Merino, M., Fojo, T., Kebebew, E., . . . 1085 Pacak, K. (2012). Somatic HIF2A gain-of-function mutations in 1086 paraganglioma with polycythemia. N Engl J Med, 367(10), 922-930. 1087 doi: 10.1056/NEJMoa1205119 1088

1089

1090

1091

1092

1093

1094

1095

1096

1097

1098

1099

1100

1101

1102

1103

1104

1105

1106

1107

1108

42 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1109 Figures

1110 Macias_Figure 1

1111

1112

43 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1113 Macias_Fig1-figure supplement 1

1114

44 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1115 Macias_Figure 2

1116

1117

1118

1119

1120

1121

1122

1123

1124

45 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1125 Macias_Figure 3

1126

1127

1128

1129

1130

1131

1132

1133

1134

1135

1136

1137

1138

46 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1139 Macias_Figure 3-figure supplement 1

1140

1141

1142

1143

1144

1145

1146

1147

1148

1149

1150

1151

47 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1152 Macias_Figure 3-figure supplement 2

1153

1154

1155

1156

48 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1157 Macias_Figure 4

1158

1159

1160

1161

1162

1163

1164

1165

1166

1167

1168

49 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1169 Macias_Figure 4-figure supplement 1

1170

50 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1171 Macias_Figure 4-figure supplement 2

1172

1173

1174

1175

1176

1177

1178

1179

1180

1181

1182

1183

1184

1185

1186

51 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1187 Macias_Figure 5

1188

1189

1190

1191

52 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1192 Macias_Figure 5-figure supplement 1

1193

1194

1195

53 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1196 Macias_Figure 6

1197

1198

1199

1200

1201

1202

1203

1204

1205

1206

1207

1208

1209

1210

54 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1211 Macias_Figure 6-figure supplement 1

1212

1213

1214

1215

1216

1217

1218

1219

1220

1221

1222

1223

1224

1225

55 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1226 Macias_Figure 7

1227

1228

1229

1230

1231

56 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1232 Macias_Figure 7-figure supplement 1

1233

1234

1235

1236

1237

1238

1239

1240

1241

1242

57 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1243 Figure Legends

1244 Figure 1. Selective loss of carotid body glomus cells in sympathoadrenal-

1245 specific Hif-2α, but not Hif-1α, deficient mice. (A and B) Tyrosine hydroxylase

1246 (TH) immunostaining on carotid bifurcation sections from HIF-2αWT (A, top

1247 panels), TH-HIF-2αKO (A, bottom panels), HIF-1αWT (B, top panels) and TH-

1248 HIF-1αKO (B, bottom panels) mice. Micrographs showed in A (bottom) were

1249 selected to illustrate the rare presence of TH+ glomus cells in TH-HIF-2αKO

1250 mice (white arrowheads). Dashed rectangles (left panels) are shown at higher

1251 magnification on the right panels. SCG, superior cervical ganglion; CB, carotid

1252 body; ICA, internal carotid artery; ECA, external carotid artery. Scale bars:

1253 200 µm (left panels), 50 µm (right panels). (C-F) Quantification of total SCG

1254 volume (C), total CB volume (D), TH+ cell number (E) and TH+ cell density (F)

1255 on micrograph from HIF-2αWT (black dots, n = 4 for SCG and n = 6 for CB),

1256 TH-HIF-2αKO (red squares, n = 4), HIF-1αWT (black triangles, n = 4) and TH-

1257 HIF-1αKO (blue triangles, n = 3) mice. Data are expressed as mean ± SEM.

1258 Mann-Whitney test, **p < 0.001; ns, non-significant. (G and H) TH-

1259 immunostained adrenal gland sections from HIF-2αWT (G, top panel), TH-HIF-

1260 2αKO (G, bottom panel), HIF-1αWT (H, top panel) and TH-HIF-1αKO (H, bottom

1261 panel) littermates. AM, adrenal medulla. Scale bars: 200 µm. (I) Normalized

1262 adrenaline (left) and noradrenaline (right) urine content measured by ELISA.

1263 HIF-2αWT (black dots, n = 8), TH-HIF-2αKO (red squares, n = 7), HIF-1αWT

1264 (black triangles, n = 7) and TH-HIF-1αKO (blue triangles, n = 4). Mann-Whitney

1265 test, ns, non-significant.

1266

58 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1267 Figure 1-figure supplement 1. Sympathoadrenal gene deletion by TH-IRES-

1268 Cre mouse strain. (A) TH-Cre-mediated recombination in carotid bifurcation

1269 (left panels) and adrenal gland (right panels) sections dissected from the TH-

1270 Td-tomato reporter mice (Cre+) compared to Td-Tomato (Cre-) mice. Notice

1271 the presence of Td-tomato fluorescence into the CB glomus cells, SCG

1272 sympathetic neurons and adrenal medulla due to TH-Cre activity (lower

1273 panels). Dashed lines delineate the location of SCG and AM within Td-tomato

1274 (Cre-) sections. SCG, superior cervical ganglion; CB, carotid body; AM,

1275 adrenal medulla. Scale bars: 100 µm for carotid bifurcation and 200 µm for

1276 adrenal gland sections. (B-C) Relative Hif-2α (B) and Hif-1α (C) gene deletion

1277 of dissected SCG (left graphs) and AM (right graphs) from TH-HIF-2αKO (red,

1278 n = 4) and TH-HIF-1αKO (blue, n = 4) compared to their respective littermate

1279 controls (HIF-2αWT, n = 4; HIF-1αWT, n = 2). Data are expressed as mean

1280 ±SEM. Unpaired t-test, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

1281

1282 Figure 2. Progressive CB glomus cells death during development of TH-

1283 HIF2αKO mouse. (A) Representative micrographs illustrating CB TH+ glomus

1284 cells appearance at embryo stage E18.5 (left) and postnatal stage P0 (right)

1285 in HIF-2αWT (top panels) and TH-HIF-2αKO (bottom panels) mice. Dashed

1286 rectangles (left panels) are shown at higher magnification on the right panels.

1287 SCG, superior cervical ganglion; CB, carotid body; ICA, internal carotid artery;

1288 ECA, external carotid artery. Scale bars: 100 µm. (B-E) Quantitative

1289 histological analysis of CB volume (B), TH+ cells number (C), TH+ cell density

1290 (D) and SCG volume (E) from TH-HIF-2αKO compared to HIF-2αWT mice. HIF-

1291 2αWT (E18.5, n = 8; P0, n = 4; SCG, n = 5), TH-HIF-2αKO (E18.5, n = 7; P0, n

59 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1292 = 5; SCG, n = 5). SCG, superior cervical ganglion. Data are expressed as

1293 mean ±SEM. Two-way ANOVA, *p < 0.05, **p < 0.01, ***p < 0.001, ****p <

1294 0.0001, ns, non significant. (F) Representative pictures of double TH+ and

1295 TUNEL+ (white arrows) stained carotid bifurcation sections from HIF-2αWT (top

1296 panels) and TH-HIF-2αKO (bottom panels) mice at E18.5 (left) and P0 (right)

1297 stages. (G) Number of TUNEL+ cells within the CB (TH+ area) of HIF-2αWT

1298 (E18.5, n = 9; P0, n = 5) and TH-HIF-2αKO (E18.5, n = 5; P0, n = 4) mice at

1299 E18.5 and P0 stages. Data are expressed as mean ±SEM. Two-way ANOVA,

1300 *p < 0.05, **p < 0.01, ****p < 0.0001. Scale bars: 50 µm

1301

1302 Figure 3. Effects of HIF-2α and HIF-1α sympathoadrenal loss on the HVR

1303 and whole-body metabolic activity in response to hypoxia. (A and E) HVR in

1304 TH-HIF-2αKO (A, red line, n = 4) and TH-HIF-1αKO (E, blue line, n = 3)

1305 deficient mice compared to control HIF-2αWT (A, black line, n = 4) and HIF-

WT 1306 1α (E, black line, n = 3) mice. Shift and duration of 10% O2 stimulus (30

1307 minutes) are indicated. Shaded areas represent the period of time analysed in

1308 B and C or F and G, respectively. Data are presented as mean breaths/min

1309 every minute ±SEM. Two-way ANOVA, **p < 0.01, ns, non-significant. (B and

1310 C) Averaged respiratory rate (B) and minute volume (C) of HIF-2αWT (black, n

1311 = 4) and TH-HIF-2αKO (red, n = 4) mice before and during the first 5 minutes

1312 of 10% O2 exposure (shaded rectangle in A). Data are expressed as mean

1313 ±SEM. Two-way ANOVA, *p < 0.05, ***p < 0.001, ****p < 0.0001, ns, non-

1314 significant. (D) Peripheral O2 saturation before, during and after hypoxia in

1315 HIF-2αWT (black, n = 8) and TH-HIF-2αKO (red, n = 9) mice. Data are

1316 expressed as mean percentage every 10 seconds ±SEM. Two-way ANOVA,

60 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1317 ****p < 0.0001. (F and G) Averaged respiratory rate (F) and minute volume (F)

1318 of HIF-1αWT (black, n = 3) and TH-HIF-1αKO (blue, n = 3) mice before and

1319 during the first 5 minutes of 10% O2 exposure (shaded rectangle in E). Data

1320 are expressed as mean ±SEM. Two-way ANOVA, *p < 0.05, **p < 0.01. (H

1321 and I) Baseline metabolic activity of HIF-2αWT (black, n = 3) and TH-HIF-2αKO

1322 (red, n = 3) mice. Data are presented as mean oxygen consumption (VO2, H)

1323 and carbon dioxide generation (VCO2, I) every hour ± SEM. White and black

1324 boxes depict diurnal and nocturnal periods, respectively. Area under the curve

1325 followed by Mann-Whitney test, ns, non-significant. (J and K) Whole-body

1326 metabolic analysis of HIF-2αWT (black, n = 4) and TH-HIF-2αKO (red, n = 6)

1327 mice in response to 10% O2. Shift and duration of 10% O2 stimulus are

1328 indicated. Data are presented as mean oxygen consumption (VO2, J) and

1329 carbon dioxide generation (VCO2, K) every hour ± SEM. White and black

1330 boxes depict diurnal and nocturnal periods, respectively. Recovery across the

1331 hypoxia period was analysed by one-phase association curve fitting. ****p <

1332 0.0001.

1333

1334 Figure 3-figure supplement 1.Secretory and electrophysiological properties

1335 in AM and SCG from TH-HIF-2αKO mice. (A and B) Representative

1336 amperometric recordings of catecholamine release by chromaffin cells from

WT KO 1337 HIF-2α and TH-HIF-2α mice (8-12 weeks old) in response to low PO2

1338 (PO2 ≈ 10-15 mmHg), hypercapnia (20% CO2) and 20 mM KCl. (C)

1339 Percentage of AM slices with response to hypoxia, 20% CO2 and 20 mM KCl,

1340 respectively. HIF-2αWT n = 8; TH-HIF-2αKO, n = 7. (D) Averaged secretion rate

1341 of chromaffin cells from HIF-2αWT and TH-HIF-2αKO mice exposed to hypoxia

61 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1342 (PO2 ≈ 10-15 mmHg), hypercapnia (20% CO2) and 20 mM KCl. Each data

1343 point represents the mean of 2-3 recordings per mouse. HIF-2αWT, n = 8; TH-

1344 HIF-2αKO, n = 7. Data are expressed as mean ±SEM. One-way ANOVA, non-

1345 significant. (E and F) Ca2+ and K+ current/voltage (I/V) curves of AM

1346 chromaffin cells (E) and SCG sympathetic neurons (F) dissociated from HIF-

1347 2αWT (n = 9 cells from 3 mice) and TH-HIF-2αKO (n = 10 cells from 3 mice)

1348 mice. Data are expressed as mean ±SEM. Two-way ANOVA, ns, non-

1349 significant.

1350

1351 Figure 3-figure supplement 2. Respiratory parameters of TH-HIF-2αKO and

1352 TH-HIF-1αKO mice exposed to acute hypoxia. (A-J) Averaged tidal volume (A

1353 and F), inspiration time (B and G), expiration time (C and H), peak inspiratory

1354 flow (D and I) and peak expiratory flow (E and J) of TH-HIF-2αKO (n = 4) and

1355 TH-HIF-1αKO (n = 3) respect to control mice (HIF-2αWT, n = 4; HIF-1αWT, n =

1356 3) before and during the first 5 minutes breathing 10% O2. Data are

1357 expressed as mean ±SEM. Two-way ANOVA, *p < 0.05, **p < 0.01, ***p <

1358 0.001.

1359

1360 Figure 4. Impaired acclimatisation to chronic hypoxia and severe pulmonary

1361 hypertension in mice lacking CBs. (A) Kaplan-Meier survival curve of the

WT 1362 indicated mice housed at 10% O2 up to 21 days. HIF-2α (n = 12), TH-HIF-

1363 2αKO (n = 11), HIF-1αWT (n = 8) and TH-HIF-1αKO (n = 8). ***p < 0.001. (B and

1364 C) Haematocrit (B) and haemoglobin (C) levels of HIF-2αWT (black) and TH-

KO 1365 HIF-2α (red) littermates kept in normoxia (21% O2) and after 21 days at

WT KO 1366 10% O2. In B, HIF-2α (21% O2, n = 6; 10% O2, n = 21), TH-HIF-2α (21%

62 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

WT 1367 O2, n = 7; 10% O2, n = 8). In C, HIF-2α (21% O2, n = 5; 10% O2, n = 8), TH-

KO 1368 HIF-2α (21% O2, n = 5; 10% O2, n = 7). Data are expressed as mean

1369 ±SEM. Two-way ANOVA, *p < 0.05, ***p < 0.001, ****p < 0.0001, ns, non-

1370 significant. (D and E) Normalized heart (D) and spleen (E) weight of HIF-2αWT

KO 1371 (black) and TH-HIF-2α (red) before and after 21 days in hypoxia (10% O2).

WT KO 1372 In D, HIF-2α (21% O2, n = 16; 10% O2, n = 23), TH-HIF-2α (21% O2, n =

WT 1373 15; 10% O2, n = 13). In E, HIF-2α (21% O2, n = 5; 10% O2, n = 23), TH-HIF-

KO 1374 2α (21% O2, n = 5; 10% O2, n = 13). Data are expressed as mean ±SEM.

1375 Two-way ANOVA, ****p < 0.0001, ns, non-significant. (F) Fulton index on

WT 1376 dissected from HIF-2α (black, 21% O2, n = 17; 10% O2, n = 23) and

KO 1377 TH-HIF-2α (red, 21% O2, n = 15; 10% O2, n = 17) mice before and after 21

1378 days in hypoxia (10% O2). Data are expressed as mean ±SEM. Two-way

1379 ANOVA, ****p < 0.0001, ns, non-significant. RV, right ventricle. LV, left

1380 ventricle. (G) Right ventricular systolic pressure (RVSP) recorded on HIF-

WT KO 1381 2α (black, 21% O2, n = 6; 10% O2, n = 7) and TH-HIF-2α (red, 21% O2, n

1382 = 5; 10% O2, n = 5) mice maintained in normoxia (21% O2) or hypoxia (10%

1383 O2) for 21 days. Data are expressed as mean ±SEM. Two-way ANOVA, **p <

1384 0.01, ***p < 0.001, ****p < 0.0001, ns, non-significant. (H and I) Lung

1385 peripheral vessels muscularization (H) and arterial medial thickness (I) in HIF-

WT KO 1386 2α and TH-HIF-2α mice exposed to 10% O2 for 14 and/or 21 days. In H,

WT KO 1387 HIF-2α (10% O2 14d, n = 4; 10% O2 21d, n = 3), TH-HIF-2α (10% O2

WT 1388 14d, n = 6; 10% O2 21d, n = 3). In I, HIF-2α (21% O2, n = 5; 10% O2 21d, n

KO 1389 = 8), TH-HIF-2α (21% O2, n = 5; 10% O2 21d, n = 8). Data are expressed

63 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1390 as mean ±SEM. Two-way ANOVA, *p < 0.05, **p < 0.01, ****p < 0.0001, ns,

1391 non-significant.

1392

1393 Figure 4-figure supplement 1. Impaired hypoxia-induced CB proliferation in

1394 TH-HIF-2αKO mice. (A) Glial fibrillary acidic protein (GFAP) immunostaining to

1395 illustrate the presence of CB stem cells in carotid bifurcation of both HIF-2αWT

1396 and TH-HIF-2αKO mice. Scale bars: 50 µm. (B) Double TH+ BrdU+

1397 immunofluorescence performed on carotid bifurcation sections from HIF-2αWT

KO 1398 (top) and TH-HIF-2α (bottom) mice exposed to 10% O2 for 14 days. Dashed

1399 rectangles are shown at higher magnification on the right panels. Newly

1400 formed TH+ cells are indicated with white arrows. Scale bars: 200 µm (left

1401 pictures), 50 µm (right pictures). (C and D) Quantification of total CB volume

1402 (C, n = 4 per genotype and condition) and double BrdU+ TH+ cells number (D,

1403 n = 3 per genotype and condition) in HIF-2αWT (black) and TH-HIF-2αKO (red)

1404 mice maintained for 14 days in hypoxia (10% O2). Data are presented as

1405 mean ±SEM. Unpaired t-test, **p < 0.01, ***p < 0.001.

1406

1407 Figure 4-figure supplement 2. Adaptation to hypoxia in TH-HIF-1αKO

WT 1408 animals. (A) Haematocrit levels of HIF-1α (black, 21% O2, n = 4; 10% O2, n

KO 1409 = 8) and TH-HIF-1α (blue, 21% O2, n = 4; 10% O2, n = 8) mice housed at

1410 21% O2 or after 21 days at 10% O2. Data are expressed as mean ±SEM.

1411 Two-way ANOVA, ****p < 0.001, ns, non-significant. (B and C) Normalized

1412 heart (B) and spleen (C) weight of HIF-1αWT (black) and TH-HIF-1αKO (blue)

WT 1413 mice before and after 21 days of hypoxia (10% O2) exposure. In B, HIF-1α ,

KO 1414 21% O2, n = 9; 10% O2, n = 8; TH-HIF-1α , n = 9, per condition. In C, 21%

64 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1415 O2, n = 4 per genotype; 10% O2, n = 5 per genotype. Data are expressed as

1416 mean ±SEM. Two-way ANOVA, *p < 0.05, ns, non-significant. (D) Fulton

1417 index on hearts dissected from HIF-1αWT (black) and TH-HIF-1αKO (blue) mice

WT 1418 before and after 21 days in hypoxia (10% O2). HIF-1α , 21% O2, n = 7; 10%

KO 1419 O2, n = 8; TH-HIF-1α , n = 8, per condition. Data are expressed as mean

1420 ±SEM. Two-way ANOVA, ****p < 0.0001, ns, non-significant. RV, right

1421 ventricle. LV, left ventricle.

1422

1423 Figure 5. Effects of CB dysfunction on blood pressure regulation. (A-C)

1424 Circadian variations in baseline systolic blood pressure (A, continued line),

1425 diastolic blood pressure (A, broken line), mean blood pressure (B) and heart

1426 rate (C) of HIF-2αWT (black, n = 10) and TH-HIF-2αKO (red, n = 8) littermates

1427 recorded by radiotelemetry. White and black boxes depict diurnal and

1428 nocturnal periods, respectively. Data are presented as averaged values every

1429 hour ±SEM. Area under the curve followed by unpaired t-test are shown. BP,

1430 blood pressure. HR, heart rate in beats per minute. (D-F) Mean blood

1431 pressure (D), heart rate (E) and subcutaneous temperature (F) alteration in

1432 response to hypoxia in radiotelemetry-implanted TH-HIF-2αKO (red, n = 5)

WT 1433 compared to HIF-2α (black, n = 5). Shift and duration of 10% O2 are

1434 indicated in the graph. White and black boxes represent diurnal and nocturnal

1435 periods, respectively. Data are presented as averaged values every hour

1436 ±SEM. Recovery across the hypoxia period was analysed by one-phase

1437 association curve fitting. BP, blood pressure. HR, heart rate in beats per

1438 minute. (G) Schematic representation of the protocol followed in the Ang II-

1439 induced experimental hypertension model. (H-M) Mean blood pressure and

65 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1440 heart rate recorded from HIF-2αWT (black, n = 8) and TH-HIF-2αKO (red, n = 8)

1441 mice after 7 days (H and K), 14 days (I and L) and 21 days (J and M) of Ang II

1442 infusion. White and black boxes depict diurnal and nocturnal periods,

1443 respectively. Data are presented as averaged values every hour ±SEM. Area

1444 under the curve followed by unpaired t-test are shown. BP, blood pressure.

1445 HR, heart rate in beats per minute.

1446

1447 Figure 5-figure supplement 1. Cardiovascular homeostasis in response to

1448 hypoxia and Ang II-induced experimental hypertension. (A and B) Circadian

1449 oscillations in subcutaneous temperature (A) and physical activity (B) of

1450 radiotelemetry-implanted HIF-2αWT (black, n = 10) and TH-HIF-2αKO (red, n =

1451 8). White and black boxes depict diurnal and nocturnal periods, respectively.

1452 Data are presented as averaged values every hour ±SEM. Area under the

1453 curve followed by unpaired t-test are shown. (C and D) Systolic blood

1454 pressure (C) and diastolic blood pressure (D) alteration in response to

1455 hypoxia in TH-HIF-2αKO (red, n = 5) compared to HIF-2αWT (black, n = 5)

1456 measured by radiotelemetry. Shift and duration of 10% O2 are indicated in the

1457 graph. White and black boxes depict diurnal and nocturnal periods,

1458 respectively. Data are presented as averaged blood pressure values every

1459 hour ±SEM. Recovery across the hypoxia period was analysed by one-phase

1460 association curve fitting. (E) Physical activity recorded by radiotelemetry of

1461 HIF-2αWT (black, n = 5) and TH-HIF-2αKO (red, n = 5) littermates exposed to

1462 10% O2 environment. Shift and duration of 10% O2 are indicated in the graph.

1463 White and black boxes depict diurnal and nocturnal periods, respectively.

1464 Data are presented as averaged blood pressure values every hour ±SEM.

66 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1465 Area under the curve followed by unpaired t-test is shown. (F-Q) Circadian

1466 fluctuations in systolic blood pressure, diastolic blood pressure, subcutaneous

1467 temperature and physical activity of HIF-2αWT (black, n = 5) and TH-HIF-2αKO

1468 (red, n = 5) after 7 days (F-I), 14 days (J-M) and 21 days (N-Q) of Ang II

1469 infusion recorded by radiotelemetry. White and black boxes represent diurnal

1470 and nocturnal periods, respectively. Data are presented as averaged values

1471 every hour ±SEM. Area under the curve followed by unpaired t-test are

1472 shown. ns, non-significant.

1473

1474 Figure 6. Exercise performance and glucose management in mice lacking

1475 CBs. (A-D). Time course showing the increment in O2 consumption (A),

1476 increment in CO2 generation (B), respiratory exchange ratio (C) and

1477 increment in heat production (D) from HIF-2αWT (black, n = 8) and TH-HIF-

1478 2αKO (red, n = 10) mice during exertion on a treadmill. Mice were warmed up

1479 for 5 minutes (5m/min) and then the speed was accelerated 2m/min at 2.5

1480 minutes intervals until exhaustion. Averaged measurements every 2.5

1481 minutes intervals ±SEM are charted. Two-way ANOVA, *p < 0.05, **p < 0.01,

1482 ***p < 0.001, ****p < 0.0001. RER, respiratory exchange ratio. (E) Running

1483 total time spent by HIF-2αWT (black, n = 8) and TH-HIF-2αKO (red, n = 10)

1484 mice to become exhausted. Data are expressed as mean ±SEM. Unpaired t-

1485 test. (F and G) Aerobic capacity (F) and power output (G) measured in HIF-

1486 2αWT (black, n = 8) mice compared to TH-HIF-2αKO (red, n = 10) mice when

1487 performing at VO2max. Data are expressed as mean ±SEM. Unpaired t-test.

1488 *p < 0.05. (H and I) Plasma glucose (H) and lactate (I) levels of HIF-2αWT

1489 (black, before, n = 7; after, n = 12) and TH-HIF-2αKO (black, before, n = 6;

67 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1490 after, n = 12) littermates before and after running to exhaustion. Data are

1491 expressed as mean ±SEM. Two-way ANOVA, *p < 0.05. ***p < 0.001. ns,

1492 non-significant. (J and K) Plasma glucose levels of HIF-2αWT (black, n = 6)

1493 and TH-HIF-2αKO (red, n = 6) mice fed (J, n = 6 per genotype) or fasted over

1494 night (K, HIF-2αWT (n = 8) and TH-HIF-2αKO (n = 10). Data are expressed as

1495 mean ±SEM. Unpaired t-test. (L-M) Time courses showing the tolerance of

1496 HIF-2αWT and TH-HIF-2αKO mice to glucose (L, n = 7 per genotype) insulin

1497 (M, n = 11 per genotype) and lactate (N, n = 7 per genotype) injected bolus.

1498 Data are presented as mean ±SEM. Two-way ANOVA, *p < 0.05, **p < 0.01,

1499 ns, non-significant.

1500

1501 Figure 6-figure supplement 1. Related to Figure 6. Whole-body metabolic

1502 activity of TH-HIF-2αKO in response to fasting-induced hypoglycemia. (A-D)

1503 Oxygen consumption (A, VO2), carbon dioxide generation (B, VCO2),

1504 respiratory exchange rate (C) and heat production (D) of HIF-2αWT (black, n =

1505 7) and TH-HIF-2αKO (red, n = 7) littermates under feed and over night fasting

1506 conditions. White and black boxes represent diurnal and nocturnal periods,

1507 respectively. Data are presented as averaged values every hour ±SEM. Area

1508 under the curve followed by unpaired t-test are shown. ns, non-significant.

1509

1510 Figure 7. mTORC1 activation in CB development and hypoxia-induced CB

1511 neurogenesis. (A) Double TH and p-S6 (mTORC1 activation marker)

1512 immunofluorescence on adult wild type CB slices to illustrate the increased p-

1513 S6 expression after 14 days in hypoxia (10% O2). For clarity, bottom panels

1514 only show p-S6 staining. TH+ outlined area (top panels) indicates the region

68 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1515 used for pixels quantification. Scale bars: 50 µm. (B) Quantification of p-S6

1516 (top) and TH (bottom) mean pixels intensity within the CB TH+ cells of wild

1517 type mice breathing 21% O2 or 10% O2 for 14 days. Each data point depicts

1518 the average of 3-4 pictures. 21% O2, n = 5; 10% O2 14d, n = 9 per genotype.

1519 Data are presented as mean ±SEM. Unpaired t-test, *p < 0.05. (C)

1520 Representative micrographs of TH and p-S6 double immunostaining of HIF-

1521 2αWT and TH-HIF-2αKO embryonic (E18.5) carotid bifurcations. TH+ outlined

1522 area indicates the region used for pixels quantification. Scale bars: 50µm. (D)

1523 Quantitative analysis performed on HIF-2αWT (black, n = 12) and TH-HIF-2αKO

1524 (red, n = 10) micrographs previously immuno-stained for p-S6 and TH. Each

1525 data point depicts the average of 3-4 pictures. Data are presented as mean

1526 ±SEM. Unpaired t-test, **p < 0.01. (E) Double BrdU and TH

1527 immunofluorescence on carotid bifurcation sections from wild type mice

1528 exposed to 10% O2 and injected with rapamycin or vehicle control for 14 days.

1529 Notice the striking decrease in double BrdU+ TH+ cells (white arrowheads) in

1530 the rapamycin treated mice compared to vehicle control animals. Scale bars:

1531 100µm. (F-H) Quantification of total CB volume (F), TH+ cells number (G) and

1532 percentage of double BrdU+ TH+ over the total TH+ cells (H) in wild type mice

1533 breathing 10% O2 and injected with rapamycin or vehicle control for 14 days.

1534 n = 5 mice per treatment. (K) Averaged breaths per minute during the first 2

1535 minutes of acute hypoxia (10% O2) exposure from vehicle or rapamycin

1536 injected mice before and after 7 days of chronic hypoxia (10% O2) treatment.

1537 n = 14 mice per condition and treatment. Data are presented as mean ±SEM.

1538 B, D, F, G and H, unpaired t-test, *p < 0.05, ****p < 0.0001. K, two-way

69 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1539 ANOVA followed by Tukey’s multiple comparison test. ****p < 0.0001, ns,

1540 non-significant.

1541

1542 Figure 7-figure supplement 1. mTORC1 activity in SCG TH+ sympathetic

1543 neurons and effect of rapamycin treatment. (A) Double TH and p-S6

1544 (mTORC1 activation marker) immunofluorescence on adult wild type SCG

+ 1545 slices after 14 days in hypoxia (10% O2). TH outlined area indicates the

1546 region used for pixels quantification. Scale bars: 100 µm. SCG, superior

1547 cervical ganglion. (B) Quantification of p-S6 (red) and TH (green) mean pixels

+ 1548 intensity within the SCG TH area of wild type mice breathing 21% O2 or 10%

1549 O2 for 14 days. Each data point depicts the average of 3-4 pictures. 21% O2,

1550 n = 3; 10% O2, n = 4. Data are presented as mean ±SEM. Unpaired t-test, ns,

1551 non-significant. (C) Representative micrographs of TH and p-S6 double

1552 immunostaining of HIF-2αWT and TH-HIF-2αKO embryonic (E18.5) carotid

1553 bifurcations. SCG TH+ outlined area indicates the region used for pixels

1554 quantification. SCG, superior cervical ganglion. Scale bars: 100 µm. (D)

1555 Quantitative analysis performed on HIF-2αWT (black) and TH-HIF-2αKO (red)

1556 micrographs previously immune-stained for TH (green) and p-S6 (red). Each

1557 data point depicts the average of 3-4 pictures. 21% O2, n = 12; 10% O2, n =

1558 10. Data are presented as mean ±SEM. Unpaired t-test, ns, non-significant.

1559 (E) Double TH and p-S6 immunofluorescence on vehicle control or rapamycin

1560 injected adult wild type SCG slices after 14 days breathing in a 10% O2

1561 atmosphere. Notice the absence of p-S6 signal as a consequence of mTORC

1562 inhibition by rapamycin. SCG, superior cervical ganglion. CB, carotid body.

1563 Scale bars: 100 µm. (F) Haematocrit levels of vehicle control (n = 14) and

70 bioRxiv preprint doi: https://doi.org/10.1101/250928; this version posted January 19, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.

1564 rapamycin (n = 14) injected mice after 7 days at 10% O2. Un-paired t-test, ns,

1565 non-significant.

1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578

71