Opposing Effects of Hypoxia on Catecholaminergic Locus Coeruleus and Hypocretin/Orexin Neurons in Chick Embryos Jeremy P. Landry, Connor Hawkins, Sabrina Wiebe,* Evan Balaban, Maria Pompeiano Department of Psychology, McGill University, Montreal, Quebec, Canada H3A 1B1
Received 14 January 2014; revised 6 April 2014; accepted 14 April 2014
ABSTRACT: Terrestrial vertebrate embryos face a risk of low oxygen availability (hypoxia) that is especially great during their transition to air-breathing. To better understand how fetal brains respond to hypoxia, we examined the effects of low oxygen availability on brain activity in late-stage chick embryos (day 18 out of a 21day incubation period). Using cFos protein expression as a marker for neuronal activity, we focused on two specific, immunohistochemically identified cell groups known to play an important role in regulating adult brain states (sleep and waking): the noradrenergic neurons of the Locus Coeruleus (NA-LC), and the Hypocretin/Orexin (H/O) neurons of the hypothalamus. cFos expression was also examined in the Pallium (the avian analog of the cerebral cortex). In adult mammalian
INTRODUCTION Vertebrate embryos are at risk for hypoxia both prenatally, during brain morphological differentiation and circuit formation, and perinatally, during the *Present address: Department of Translational Medicine and Clinical Pharmacology, Boehringer Ingelheim, Birkerndorfer Strasse 65, Biberach an der Riss D-88400, Germany. Correspondence to: M. Pompeiano ([email protected]). Contract grant sponsor: McGill University start-up funds (MP). Contract grant sponsor: Human Frontiers Science Program; contract grant number: RGP0004/2013 (MP, EB). Contract grant sponsor: Canadian Fund for Innovation; contract grant number: 9908 (EB). Ó 2014 Wiley Periodicals, Inc. Published online 00 Month 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/dneu.22182
brains, cFos expression changes in a coordinated way in these areas. In chick embryos, oxygen deprivation simultaneously activated NA-LC while deactivating H/O-producing neurons; it also increased cFos expression in the Pallium. Activity in one pallial primary sensory area was significantly related to NA-LC activity. These data reveal that at least some of the same neural systems involved in brain-state control in adults may play a central role in orchestrating prenatal hypoxic responses, and that these circuits may show different patterns of coordination than seen in adults. VC 2014 Wiley Periodicals, Inc. Develop Neurobiol 00: 000–000, 2014
Keywords: hypoxia; chick embryo; locus coeruleus; hypocretin/orexin; cFos
transition to air-breathing. In oviparous species like birds where adults physically incubate eggs, sitting on them too tightly can decrease oxygen availability. In mammals, maternal and uterine challenges may result in hypoxic insults to the fetus. Avian embryos, which are free of confounding effects of maternal and placental adaptation during hypoxic exposure, show grossly similar cardio-respiratory physiology (Mortola, 2009), and have grossly similar brain circuitry (Reiner et al., 2004) to mammals. They also display similar responses to hypoxia, including: hypometabolism, reduced thermogenesis, reduced heart rate, reduced respiratory movements, increased blood catecholamines and glucocorticoids, and redistribution of blood circulation to the brain and heart (Mulder et al., 2000; Crossley et al., 2003; Mortola, 1
2
Landry et al.
2009). A similar “immature” reaction to hypoxia has also been described in human infants (Horne, 2013). To better understand whether embryo hypoxia is accompanied by changes in the functional states of the brain, neuronal responses to hypoxia were measured using cFos protein expression as an indicator of neuronal activity in pre-pipped (not yet air-breathing) embryonic day (E) 18 chick embryos. The effects of modest hypoxia (a reduction of oxygen content from the normal 21% to 15%, a level known to be nonlethal; Mortola, 2009) and medium hypoxia (a reduction to 10% oxygen, known to be lethal with more prolonged exposure than used here; Mortola JP, personal communication) were assayed using the activation of two key neuronal groups involved in brainstate control: the noradrenergic Locus Coeruleus (NA-LC) neurons in the brainstem, and the hypothalamic Hypocretin/Orexin (H/O) cell group. In adult mammals, these two groups of cells usually work synergistically to induce (NA-LC) and maintain (H/O) waking states, and typically increase and decrease their activity together in a coordinated fashion (Siegel, 2009); changes in brainstem NA-LC cell activity are also associated with changes in cFos activation of forebrain areas (Cirelli et al., 1996). NA-LC and H/O circuitry may not function in this same integrated and coordinated way in embryos (Balaban et al., 2012); cFos activation in sensory and association areas of the Pallium (avian analog of the cerebral cortex) were also examined to see if changes in brainstem NA-LC cell activity affect telencephalon activation in chick embryos.
METHODS Hypoxia Exposure and Perfusion Fertilized White Leghorn chicken eggs (Gallus gallus, n 5 18, Couvoir Simetin, Mirabel, QC) were incubated in darkness under constant conditions (38 C; 60% relative humidity) with automatic egg rotation every few hours. At day 18 (hatching is at 21 days), embryos were treated in the dark for 4 h, without anesthesia, with one of three conditions: 21% O2 (normoxic controls; n 5 6, three females, three males), 15% O2 (modest hypoxia; n 5 6, four males, two females), and 10% O2 (medium hypoxia; n 5 6, four females, two males). Hypoxia was obtained by placing the eggs into an incubator where a small stream of warmed and humidified N2 was supplied under the control of a flowmeter. An O2 Analyzer (Sable Systems Int., Las Vegas, NV) continuously sampled O2 concentration and displayed this information on a computer monitor. A data logger placed inside the incubator continuously sampled the temperature and humidity. At the end of the hypoxic or norDevelopmental Neurobiology
moxic treatment, embryos were anesthetized in ovo using isoflurane, and then intracardially perfused with Howard Ringer’s solution (123 mM NaCl, 1.5 mM CaCl2, 5 mM KCl; 6.5 mL/min, 2 min) followed by 10% formalin/PBS (150 mM NaCl, 28 mM KH2PO4, 72 mM K2HPO4, pH 7.4; 10 min). The gonads were examined to establish sex, and heads were removed, kept in 10% formalin/PBS at 4 C overnight, and brains were extracted. After cryoprotection in 10% and 30% sucrose/PBS, brains were frozen and cut into 40 mm serial coronal sections at a cryostat (Leica CM3050, Leica Microsystems Canada, Richmond Hill, ON).
Immunohistochemistry All procedures were performed at room temperature and washes were done in PBS (140 mM NaCl, 8 mM Na2HPO4, 2.7 mM KH2PO4, 1.5 mM KCl, pH 7.4). Standard fluorescent double-labeling immunohistochemical techniques were used to study cFos expression in H/O and NA-LC neurons. Antibody solutions were made in blocking solution (10% normal donkey serum, 2% bovine serum albumin, and 0.5% Triton X-100 in PBS). Sections spanning the whole length of the LC (1 section every 240 mm) were incubated for 1 h with blocking solution and then overnight with mouse anti-Tyrosine Hydroxylase (TH; ImmunoStar, 22941; 1:500) and cFos (Santa Cruz Biotechnologies, sc-253; 1:1,000) primary antibodies, together. After PBS washes, sections were incubated for 2 h with Alexa Fluor 594 anti-mouse (Molecular Probes, A21203; 1:500) and Alexa Fluor 488 anti-rabbit (Molecular Probes, A21206; 1:200) antibodies, together. Sections spanning the whole posterior hypothalamus (1 section every 240 mm) were incubated for 1 h with blocking solution and then overnight with goat anti-Orexin A (Santa Cruz Biotechnologies, sc-8070; 1:500) and rabbit anti-cFos (sc-253; 1:1,000) antibodies, together. After PBS washes, sections were incubated for 2 h with Alexa Fluor 594 anti-goat (Molecular Probes, A11058; 1:500) and Alexa Fluor 488 anti-rabbit (A21206; 1:200) antibodies, together. After the secondary antibody incubation, all sections were washed in PBS, counterstained with DAPI (Invitrogen, D1306; 0.1 mg/mL in PBS) for 10 min, washed again in PBS, and coverslipped. cFos expression in the Pallium was evaluated using standard colorimetric immunohistochemical techniques. Antibody solutions were made in blocking solution (10% normal goat serum, 2% bovine serum albumin, and 0.5% Triton X-100 in PBS). A cFos primary antibody (sc-253; 1:2,000), a biotinylated anti-rabbit antibody (Vector Labs, BA1000; 1:500), and the ABC (Vector Labs, PK-6100; 1:200) and DAB (Sigma) revelation system were used.
Quantification Procedures In every animal, every sixth section was stained and examined, meaning approximately seven forebrain, approximately seven hypothalamic, and approximately three brainstem sections were examined from each embryo (the
Hypoxia and Arousal in Chick Embryos LC sections of one modest hypoxia embryo were unusable). Anatomical landmarks were defined according to a standard atlas of the chick brain (Puelles et al., 2007). Co-labeling analyses of cFos in NA-LC neurons and cFos in H/O neurons in the hypothalamus were done at 2003 magnification using a microscope (Leica DM6000 B) and the optical fractionator program in Stereo Investigator (Version 10.55, MicroBrightField, Williston, VT). Contours were manually traced around each region of interest, and the program subdivided these into fields on a grid. Cells within each field were manually tagged according to their labeling characteristics, and these numbers were tallied over all fields. The fields accounted for 80% of the grid space in the H/O areas and 100% in the LC. To determine the percentage of cFosactivated neurons of each type, the total number of neurons per field stained for TH and H/O containing a nucleus (as seen from DAPI labeling) were counted and compared with the number that were co-labeled with cFos in the nucleus. H/O neurons were also subdivided into two anatomically distinct groups: a medial group (also more anterior), which occupies the developing Paraventricular Nucleus and the Periventricular Stratum and a lateral group, within the Posterior Hypothalamic Nucleus and Lateral Hypothalamic area. In rats, the medial and lateral groups receive inputs mostly from the hypothalamus and the brainstem, respectively (Yoshida et al., 2006). The density of cFos immunoreactive cells in the Pallium was measured at 1003 magnification in three areas. Two sensory areas where chosen to contrast a primary sensory area that would be presumably continuously driven by sensory input in embryos kept in the dark (the Basorostral Nucleus, that receives somatosensory inputs), and one that would presumably be less driven by sensory input in the dark (the Entopallial Nucleus, that receives visual inputs). An easily measured association area anatomically removed from both primary sensory areas (the rostral part of the Lateral Pallium) was also examined. Stereo Investigator was used to draw a contour around the regions of interest of these three areas, as defined by the chick brain atlas (Puelles et al., 2007), from which the volume was extrapolated using the contour-area section thicknesses. The number of cFos-labeled cells was determined using the Automatic Object detection program, set to count objects between 10 and 100 mm2 and above a constant luminance threshold. Camera and lighting settings were kept constant.
Statistical Analysis The variables studied here had skewed, asymmetric distributions; some were nominally significantly different from normal distributions and others were not. As statistical tests for normality at these sample sizes have low power, we uniformly used nonparametric statistics because they are the more conservative alternative. All significant effects remained significant if parametric tests were used, and none of the non-significant effects became significant. The Scheirer–Ray–Hare nonparametric two-way ANOVA (Sokal and Rohlf, 2012) was used to compare
3
cFos expression in the NA-LC and H/O neurons. Comparisons between the three treatment groups used Kruskal– Wallis nonparametric ANOVA followed by post-hoc tests corrected for multiple comparisons (Siegel and Castellan, 1988). Results of nonparametric ANOVAs are reported using the H-statistic, with the number of degrees of freedom indicated in parentheses. Differences in the activation level of medial and lateral H/O groups were compared using the nonparametric Wilcoxon matched-pairs sign rank test (Siegel and Castellan, 1988). The Spearman nonparametric correlation coefficient (q) was used for trend analysis. For all statistical analyses, data were considered statistically significant if P < 0.05, using a two-tailed test. An initial analysis of cell densities in the Pallium using a Kruskal–Wallis nonparametric ANOVA indicated that the overall cFos-labeled cell densities varied significantly among these three regions with data from all oxygen levels combined [H(2) 5 32.33, P < 0.0001]; densities were highest in the Lateral Pallium (mean 28,547 6 2,918/mm2), and much lower in the Basorostral Nucleus (mean 1,910 6 568/ mm2) and Entopallial Nucleus (mean 1,439 6 481/mm2); the values in the Lateral Pallium were significantly different from the values in the two other areas, which were not significantly different from each other. To do a single statistical comparison of all data using a nonparametric twoway ANOVA, cFos-labeled cell densities were normalized within each area by dividing all measurements by the mean density of the normoxic group.
RESULTS NA-LC and H/O Neurons Had Similar Overall cFos Expression Levels, But Different Hypoxia Responses The two populations of neurons did not differ in their mean levels of cFos expression over all treatments, but showed a highly significantly different pattern of responses to hypoxia [nonparametric two-way ANOVA, overall differences in percent activation between cell populations: H(1) 5 0.08, P 5 0.77; overall differences between different oxygen concentrations: H(2) 5 2.10, P 5 0.35; interaction between cell population and oxygen concentration: H(2) 5 15.83, P 5 0.0003]. The inverted pattern of responses between the two cell groups was also indicated by the significant negative correlation between their percentage expression values across all embryos [q 5 20.50, n 5 17 (LC slides from one animal being unusable), P < 0.05; Fig. 1(A)].
Hypoxia Increased cFos Expression in NA-LC Neurons There was a significant effect of oxygen level on the percentage of cFos-positive NA-LC neurons Developmental Neurobiology
4
Landry et al.
Figure 1 cFos expression in NA-LC and H/O neurons. (A) The scatterplot illustrates the negative relationship between cFos expressing (cFos1) neurons in NA-LC and H/O neurons. Each dot represents one embryo. (B) The two plots of the data from all individual subjects show the percentage of cFos1 neurons in each population under each experimental condition. Horizontal lines represent mean levels of each group. The * and § symbols above each graph represent the results of statistical tests corrected for multiple comparisons; groups with different symbols are significantly different from each other.
[H(2) 5 9.99, P 5 0.007]. Embryos in the normoxic condition showed a low percentage (mean 5%). Post-hoc tests showed that the percentage of cFosactivated NA-LC cells was not significantly higher than controls in the modest hypoxia group (mean 17%); however, the medium hypoxia group was significantly higher than both the control and modest groups (mean of 46%) [Figs. 1(B) and 2(A–C)].
Hypoxia Decreased cFos Expression in H/O Neurons H/O neurons also showed a significant effect of oxygen level on cFos expression [H(2) 5 6.75, P 5 0.035]. The percentage of Fos-positive H/O neurons was moderately high in normoxia (mean 29%), it was the same at 15% oxygen (mean 29%), but it was significantly lower at 10% oxygen than in both other conditions (mean 8%) [Figs. 1(B) and 2(D–F)]. The medial group of H/O neurons had, over all oxygen conditions combined, an average of 13% higher levels of activation than the lateral group (Wilcoxon test, z 5 3.03, P 5 0.003; Fig. 3). However, there was a very strong correlation between the activity levels of both groups across all experimental conditions (q 5 0.85, n 5 18, P 5 0.0005 with an outlying point included; q 5 0.91, n 5 17, P 5 0.0003 with it excluded; Fig. 3), indicating that they responded in a similar way to oxygen level manipulations. Developmental Neurobiology
Hypoxia Increased cFos Expression in Areas of the Pallium The relative density of cFos-positive cells (after being normalized according to the normoxic mean density in each region) exhibited significant variation between regions of the Pallium and between oxygen concentrations [nonparametric two-way ANOVA, overall differences between regions: H(2) 5 9.51, P 5 0.009; overall differences between different oxygen concentrations: H(2) 5 9.86, P 5 0.008]. The changes between the different oxygen conditions had similar patterns within the three pallial areas [interaction between cell population and treatment: H(4) 5 2.20, P 5 0.70]. The Basorostral Nucleus (with an average of 2.82 6 0.84 times the mean normoxic activation level), and the Lateral Pallium [1.22 6 0.13; Fig. 2(G–I)], had significantly higher densities of cFos-positive cells than the Entopallial Nucleus (0.95 6 0.32), indicative of greater overall responses to hypoxia. The former two areas were not significantly different from each other. As a group, the three areas had a significantly higher response to 10% oxygen (an average of 2.76 6 0.71 times their mean normoxic levels) than to 15% oxygen (0.88 6 0.20) and normoxia (1.00 6 0.22); the latter two densities were not significantly different from each other. To quantify the relationship between NALC cell activity and forebrain activity, Spearman correlation coefficients (Bonferroni-corrected for three
Hypoxia and Arousal in Chick Embryos
5
Figure 2 cFos expression in three examined anatomical regions in the different experimental conditions. The number of NA-LC neurons (in red) showing cFos expression (green nucleus) increases from normoxia (A) to modest hypoxia (B) to medium hypoxia (C). cFos expression (green nucleus) in H/O neurons (red) is similar in normoxia (D) and modest hypoxia (E) and is strongly decreased in medium hypoxia (F). Arrowheads indicate examples of double-labeled NA-LC and H/O neurons. All nuclei are counterstained with DAPI (blue). cFos expression in the Lateral Pallium increases in medium hypoxia (I) with respect to normoxia (G) and modest hypoxia (H). Scale bars are the same within each row, and they are all 30 mm. 4v: fourth ventricle; D: dorsal; L: lateral; M: medial; V: ventral.
comparisons) were calculated. The Basorostral Nucleus had a significant positive correlation with LC activation (q 5 0.66, n 5 15, corrected P 5 0.04); the correlations for both the Lateral Pallium (q 5 0.30, n 5 16, corrected P 5 0.72) and Entopallial Nucleus (q 5 0.43, n 5 16, corrected P 5 0.30) were not significant. In summary, three pallial areas considered as a group had a significantly higher level of normalized cFos activation to medium hypoxia than to modest hypoxia or normoxia; the Basorostral Nucleus and Lateral Pallium both had significantly higher normalized activation to modest hypoxia than the Entopallial nucleus. Only the Basorostral Nucleus had a significant positive correlation with LC activation.
DISCUSSION NA-LC and H/O neurons are both known to be importantly involved in adult brain state regulation. NA-LC neurons, their diffuse projections, and adrenergic receptors are well developed by E18 in chick embryos (Yurkewicz et al., 1981; Dermon and Kouvelas, 1988). We recently studied the development of H/O neurons and receptors in the brain of chick embryos of different ages (Godden KE, Landry JP, Slepneva N, MiguesBlanco V, Pompeiano M. Early expression of Hypocretin/Orexin in the chick embryo brain. Submitted; Cerazy T, Saadat N, Fuchs G, Pompeiano M. Hypocretin/Orexin receptors in the chick embryo brain. In preparation). This work demonstrated that H/O neurons and their receptors are also strongly present by E18. The Developmental Neurobiology
6
Landry et al.
Figure 3 cFos expression in two anatomically separated groups of H/O neurons. The scatterplot illustrates the positive relationship between cFos expression in lateral and medial H/O neurons. The identity line (y 5 x; dotted line) shows that the medial group has a higher percentage of cFos-expressing (cFos1) neurons than the lateral group. The arrow indicates an outlying data point (see text for details).
neuroanatomical substrates by which these two systems could mediate hypoxic effects are, therefore, wellestablished by the age at which the present studies were conducted. The data show that hypoxia affected the activity of both cell groups in late-stage chick embryos prior to air-breathing. cFos-positive NA-LC cells were significantly increased during medium hypoxia relative to both normoxia and modest hypoxia conditions. In contrast, cFos-expressing H/O neurons were significantly decreased during hypoxia. The overall density of cFos-positive cells in the Pallium increased during hypoxia, with a similar pattern in two sensory areas and one association area. However, the pattern of changes over different oxygen concentrations was significantly correlated with NA-LC activation only in a somatosensory area, the Basorostral Nucleus. This contrasted with no correlation in the Entopallial Nucleus, a presumably less-actively used visual sensory area (given experiments run in darkness), or in an association area (Lateral Pallium). Increased LC cFos expression was previously found in fetal sheep exposed to acute hypoxia, but the activated neurons were never chemically identified (Nitsos and Walker, 1999).
Possible Role of NA-LC and H/O Neurons in the Physiological Responses of the Embryo to Hypoxia Many of the previously observed effects of hypoxia on chick embryos can be plausibly explained by increased activation of NA-LC neurons coupled with decreased activation of H/O neurons. However, a contribution Developmental Neurobiology
from other catecholaminergic areas, such as the A1/C1, A5, and A7 groups, is possible (Card et al., 2006; Kanbar et al., 2011; Bruinstroop et al., 2012). Both NA-LC and H/O neurons activate the sympathetic system and inhibit the parasympathetic system (Samuels and Szabadi, 2008; Williams and Burdakov, 2008; Dergacheva et al., 2011; Shahid et al., 2012). As H/O neurons innervate cardiorespiratory regulatory areas in birds (Singletary et al., 2007), reduced heart rate and respiratory movements during hypoxia in chick embryos could be explained by the decreased activity of H/O neurons; hypoxia-induced reductions in metabolism and thermogenesis (Mortola, 2009, 2011) may have a similar explanation (Asakawa et al., 2002; Girault et al., 2012). Chick embryos display increased motor activity in response to hypoxia (Windle and Barcroft, 1938; Gr€ans and Altimiras, 2007; Mortola et al., 2013); this may serve as a distress signal to the hen to change incubation positions (Gr€ans and Altimiras, 2007). H/O neurons are known to stimulate locomotor activity in adult rats (Zheng et al., 2005) and in adult and neonatal birds (da Silva et al., 2008; Katayama et al., 2010). As H/O neuron activity decreases, increased motor activity in hypoxic embryos is more likely to be dependent on the activation of NA-LC neurons, which exert a facilitatory action on spinal motor neurons (Fung et al., 1991; Burgess and Peever, 2013). NA-A7 neurons could also contribute (Bruinstroop et al., 2012), as noradrenergic innervation of the spinal cord is fully established by E18 in chick embryos (Singer et al., 1980).
Possible Circuitry Regulating NA-LC and H/O Activity During Low-Oxygen Exposure in Chick Embryos We hypothesize that hypoxia-sensitive A1/C1 neurons could play an important role in orchestrating brain responses to hypoxia in chick embryos by modulating regulatory brain areas rather than by directly influencing autonomic areas, as also suggested by the distribution of their projections (McKellar and Loewy, 1982; Batten, 1995). In adults, A1/C1 neuron stimulation causes arousal and cardiorespiratory activation (Abbott et al., 2013). In hypoxic embryos, A1/C1 neurons could drive NA-LC activation as in adults (Guyenet et al., 2013). However, the chick embryo cardiorespiratory response to hypoxia (before the start of air-breathing) is the opposite of what is observed postnatally. It is impossible for the embryo to dramatically increase the availability of oxygen through the chorioallantoic membrane (except by alerting the hen). We suggest that the differences between prenatal and postnatal responses could be specifically due to the inhibition of
Hypoxia and Arousal in Chick Embryos
H/O neurons in embryos (we expect H/O neurons to be activated by hypoxia in air-breathing chicks; see Abbott et al., 2013 and Liu et al., 2014. Inhibition of H/O neurons would subserve a neuroprotective role, imposing a global state of reduced brain activity and energy expenditure (Liu et al., 2011). Inhibition of H/O neurons could be initially mediated by A1/C1 neuron activation (Sakurai et al., 2005) even though H/O neurons are known to decrease their activity in the presence of low ATP levels (Liu et al., 2011). There are at least three mechanisms that are possible candidates for the developmental changes in H/O neuron hypoxic responses: (1) a change of sign in the response to catecholamines [the expression of a2 adrenergic receptors (inhibitory) precedes that of a1 and b receptors (both excitatory) in several areas of the chick brain (Dermon and Kouvelas, 1988)]; (2) a change in the sign of GABAergic inputs to H/O neurons (from excitatory to inhibitory); and (3) the development of peripheral, with respect to central A1/C1-dependent, chemoreception [carotid bodies are rather inactive before birth (Purves, 1982), but become postnatally activated by hypoxia, inducing cardiorespiratory activation and arousal (Nattie and Li, 2013; Guyenet and Abbott, 2013)]. The lower cFos expression we saw in the lateral group of H/O neurons with respect to the medial group could be a reflection of the relative immaturity of peripheral chemoreception.
Pallial Activity Paralleling NA-LC Activity In adult rodents, primary sensory areas of the cerebral cortex show the largest changes in cFos expression between sleep and waking and these changes are NALC-dependent (Cirelli et al., 1996). In this study, hypoxia induced significant changes in Pallium cFos expression. The pattern of changes was significantly correlated with NA-LC cFos activity in one somatosensory area likely to have continuously received sensory input, but not in a visual sensory area that was not likely to have been extensively activated by sensory input (given experiments conducted in the dark), nor in an association area. These results suggest that the chick embryo NA-LC system activates at least some forebrain areas, as in adult rodents. Postnatally, arousal is an adaptive reaction to hypoxia, allowing an individual to take behavioral action to counteract negative and possibly fatal effects of hypoxia (Guyenet and Abbott, 2013). Increased cFos expression in the embryo telencephalon during hypoxia would be consistent with this pattern; however, hypoxia appears to trigger an incomplete arousal reaction in embryos because of the simultaneous inhibition of H/O neurons.
7
In summary, this study has begun to elucidate how neural systems involved in brain-state control in adults may centrally orchestrate the pattern of hypoxic responses in embryos toward the end of prenatal development. These results show for the first time that oxygen deprivation activates NA-LC, while simultaneously deactivating H/O neurons of the hypothalamus, and that hypoxia also results in increased activation in at least some areas of the Pallium (analogous to the cerebral cortex in mammals). Future work is needed to elucidate the cellular mechanisms regulating the opposing cardio-circulatory and respiratory responses and the different pattern of brain responses triggered by hypoxia in embryos with respect to post-natal organisms. The authors thank J.P. Mortola for providing the hypoxic eggs, Jae Duke Sheen and Mashiyat Majid for help with the cryostat sectioning, and J.P. Mortola and J. Mogil for comments on the manuscript.
REFERENCES Abbott SBG, Coates MB, Stornetta RL, Guyenet PG. 2013. Optogenetic stimulation of C1 and retrotrapezoid nucleus neurons causes sleep state-dependent cardiorespiratory stimulation and arousal in rats. Hypertension 61:835–841. Asakawa A, Inui A, Goto K, Yuzuriha H, Takimoto Y, Inui T, Katsuura G, et al. 2002. Effects of agouti-related protein, orexin and melanin-concentrating hormone on oxygen consumption in mice. Int J Mol Med 10:523–525. Balaban E, Desco M, Vaquero J-J. 2012. Waking-like brain function in embryos. Curr Biol 22:852–861. Batten TFC. 1995. Immunolocalization of putative neurotransmitters innervating autonomic regulating neurones of cat ventral medulla. Brain Res Bull 37:487–506. Bruinstroop E, Cano G, Vanderhorst VGJM, Cavalcante JC, Wirth J, Sena-Esteves M, Saper CB. 2012. Spinal projections of the A5, A6 (Locus Coeruleus), and A7 noradrenergic cell groups in rats. J Comp Neurol 520:1985–2001. Burgess CR, Peever JH. 2013. A noradrenergic mechanism functions to couple motor behavior with arousal state. Curr Biol 23:1719–1725. Card JP, Sved JC, Craig B, Raizada M, Vasquez J, Sved AF. 2006. Efferent projections of rat rostroventrolateral medulla C1 catecholamine neurons: Implication for the central control of cardiovascular regulation. J Comp Neurol 499:840–859. Cirelli C, Pompeiano M, Tononi G. 1996. Neuronal gene expression in the waking state: A role for the Locus Coeruleus. Science 274:1211–1215. Crossley DA, Burggren WW, Altimiras J. 2003. Cardiovascular regulation during hypoxia in embryos of the domestic chicken Gallus gallus. Am J Physiol Regul Integr Comp Physiol 284:R219–226. da Silva ES, dos Santos TV, Hoeller AA, dos Santos TS, Pereira GV, Meneghelli C, Pezlin AI, et al. 2008. Behavioral Developmental Neurobiology
8
Landry et al.
and metabolic effects of central injections of orexins/hypocretins in pigeons (Columba livia). Regul Pept 147:9–18. Dergacheva O, Philbin K, Bateman R, Mendelowitz D. 2011. Hypocretin-1 (orexin A) prevents the effects of hypoxia/hypercapnia and enhances the GABAergic pathway from the lateral paragigantocellular nucleus to cardiac vagal neurons in the nucleus ambuguus. Neuroscience 175:18–23. Dermon CR, Kouvelas ED. 1988. Binding properties, regional ontogeny and localization of adrenergic receptors in chick brain. Int J Dev Neurosci 6:471–482. Fung SJ, Manzoni D, Chan JY, Pompeiano O, Barnes CD. 1991. Locus coeruleus control of spinal motor output. Prog Brain Res 88:395–409. Girault EM, Yi C-X, Fliers E, Kalsbeek A. 2012. Orexins, feeding, and energy balance. Progr Brain Res 198:47–64. Gr€ans A, Altimiras J. 2007. Ontogeny of vocalizations and movements in response to cooling in chickens fetuses. Physiol Behav 91:229–239. Guyenet PG, Abbott SBG. 2013. Chemoreception and asphyxia-induced arousal. Respir Physiol Neurobiol 188: 333–343. Guyenet PG, Stornetta RL, Bochorishvili G, Depuy SD, Burke PG, Abbott SB. 2013. C1 neurons: The body’s EMTs. Am J Physiol Regul Integr Comp Physiol 305:R187–204. Horne RSC. 2013. Cardio-respiratory control during sleep in infancy. Pediatr Respir Rev. http://dx.doi.org/10.1016/ j.prrv.2013.02.012. Kanbar R, Depuy GH, Stornetta RL, Guyenet PG. 2011. Regulation of visceral sympathetic tone by A5 noradrenergic neurons in rodents. J Physiol 589:903–917. Katayama S, Hamasu K, Shigemi K, Cline MA, Furuse M. 2010. Intracerebroventricular injection of orexin-A, but not orexin-B, induces arousal of layer-type neonatal chicks. Comp Biochem Physiol A Mol Integr Physiol 157:132–135. Liu Z, Jiang L, Zhu F, Fu C, Lu S, Zhou J, Wu X, Bai C, Li S. 2014. Chronic intermittent hypoxia and the expression of orexin and its receptors in the brains of rats. Sleep Biol Rhythms 12:22–29. Liu Z-W, Gan G, Suyama S, Gao X-B. 2011. Intracellular energy status regulates activity in hypocretin/orexin neurons: A link between energy and behavioural states. J Physiol 589:4157–4166. McKellar S, Loewy AD. 1982. Efferent projections of the A1 catecholamine cell group in the rat: An autoradiographic study. Brain Res 241:11–29. Mortola JP. 2009. Gas exchange in avian embryos and hatchlings. Comp Biochem Physiol A Mol Integr Physiol 153:359–377. Mortola JP. 2011. Metabolic and ventilatory sensitivity to hypoxia in avian embryos. Respir Physiol Neurobiol 178: 174–180. Mortola JP, Louis AS, Simeonova M, Toro Velasquez PA. 2013. The motility of the chicken embryo: Energetic cost and effects of hypoxia. Respir Physiol Neurobiol 188:172– 179. Mulder ALM, Golde JMCG, Goor AAC, Giussani DA, Blanco CE. 2000. Developmental changes in plasma catDevelopmental Neurobiology
echolamine concentrations during normoxia and acute hypoxia in the chick embryo. J Physiol 527:593–599. Nattie E, Li A. 2013. Respiration and autonomic regulation and orexin. Progr Brain Res 198:25–46. Nitsos I, Walker DW. 1999. The distribution of FOSimmunoreactive neurons in the brainstem, midbrain and diencephalon of fetal sheep in response to acute hypoxia in mid and late gestation. Dev Brain Res 114:9–26. Puelles L, Paxinos G, Watson C, Martinez S, Martinez-dela-Torre M. 2007. The Chick Brain in Stereotaxic Coordinates: An Atlas Based on Neuromeres. New York: Academic Press. 272 p. Purves MJ. 1982. Respiratory sensitivity before and after birth. Acta Ped Scand 71:529–536. Reiner A, Perkel DJ, Bruce LL, Butler AB, Csillag A, Kuenzel W, Medina L, et al. 2004. Revised nomenclature for avian telencephalon and some related brainstem nuclei. J Comp Neurol 473:377–414. Sakurai T, Nagata R, Yamanaka A, Kawamura H, Tsujino N, Muraki Y, Kageyama H, et al. 2005. Input of orexin/ hypocretin neurons revealed by a genetically encoded tracer in mice. Neuron 46:297–308. Samuels ER, Szabadi E. 2008. Functional neuroanatomy of the noradrenergic locus coeruleus: Its roles in the regulation of arousal and autonomic function part I: Principles of functional organisation. Curr Neuropharmacol 6:235–253. Shahid IZ, Rahman AA, Pilowsky PM. 2012. Orexin and central regulation of cardiorespiratory system. Vitam Horm 89:159–184. Siegel JM. 2009. The neurobiology of sleep. Semin Neurol 29:277–296. Siegel S, Castellan NJJ. 1988. Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill. 399 p. Singer HS, Coyle JT, Vernon N, Kallman CH, Price DL. 1980. The development of catecholaminergic innervation in chick spinal cord. Brain Res 191:417–428. Singletary KG, Deviche P, Strand C, Delville Y. 2007. Distribution of orexin/hypocretin immunoreactivity in the brain of a male songbird, the house finch, Carpodacus mexicanus. J Chem Neuroanat 33:101–109. Sokal RR, Rohlf JF. 2012. Biometry. New York: W.H. Freeman. 937 p. Williams RH, Burdakov D. 2008. Hypothalamic orexins/ hypocretins as regulators of breathing. Expert Rev Mol Med 10:e28. Windle WF, Barcroft J. 1938. Some factors governing the initiation of respiration in the chick. Am J Physiol—Legacy Content 121:684–691. Yoshida K, McCormack S, Espa~na RA, Crocker A, Scammell TE. 2006. Afferents to the orexin neurons of the rat brain. J Comp Neurol 494:845–861. Yurkewicz L, Marchi M, Lauder JM, Giacobini E. 1981. Development and aging of noradrenergic cell bodies and axon terminals in the chicken. J Neurosci Res 6:621–641. Zheng H, Patterson LM, Berthoud HR. 2005. Orexin-A projections to the caudal medulla and orexin-induced cFos expression, food intake, and autonomic function. J Comp Neurol 485:127–142.
Received 14 January 2014; revised 6 April 2014; accepted 14 April 2014
ABSTRACT: Terrestrial vertebrate embryos face a risk of low oxygen availability (hypoxia) that is especially great during their transition to air-breathing. To better understand how fetal brains respond to hypoxia, we examined the effects of low oxygen availability on brain activity in late-stage chick embryos (day 18 out of a 21day incubation period). Using cFos protein expression as a marker for neuronal activity, we focused on two specific, immunohistochemically identified cell groups known to play an important role in regulating adult brain states (sleep and waking): the noradrenergic neurons of the Locus Coeruleus (NA-LC), and the Hypocretin/Orexin (H/O) neurons of the hypothalamus. cFos expression was also examined in the Pallium (the avian analog of the cerebral cortex). In adult mammalian
INTRODUCTION Vertebrate embryos are at risk for hypoxia both prenatally, during brain morphological differentiation and circuit formation, and perinatally, during the *Present address: Department of Translational Medicine and Clinical Pharmacology, Boehringer Ingelheim, Birkerndorfer Strasse 65, Biberach an der Riss D-88400, Germany. Correspondence to: M. Pompeiano ([email protected]). Contract grant sponsor: McGill University start-up funds (MP). Contract grant sponsor: Human Frontiers Science Program; contract grant number: RGP0004/2013 (MP, EB). Contract grant sponsor: Canadian Fund for Innovation; contract grant number: 9908 (EB). Ó 2014 Wiley Periodicals, Inc. Published online 00 Month 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/dneu.22182
brains, cFos expression changes in a coordinated way in these areas. In chick embryos, oxygen deprivation simultaneously activated NA-LC while deactivating H/O-producing neurons; it also increased cFos expression in the Pallium. Activity in one pallial primary sensory area was significantly related to NA-LC activity. These data reveal that at least some of the same neural systems involved in brain-state control in adults may play a central role in orchestrating prenatal hypoxic responses, and that these circuits may show different patterns of coordination than seen in adults. VC 2014 Wiley Periodicals, Inc. Develop Neurobiol 00: 000–000, 2014
Keywords: hypoxia; chick embryo; locus coeruleus; hypocretin/orexin; cFos
transition to air-breathing. In oviparous species like birds where adults physically incubate eggs, sitting on them too tightly can decrease oxygen availability. In mammals, maternal and uterine challenges may result in hypoxic insults to the fetus. Avian embryos, which are free of confounding effects of maternal and placental adaptation during hypoxic exposure, show grossly similar cardio-respiratory physiology (Mortola, 2009), and have grossly similar brain circuitry (Reiner et al., 2004) to mammals. They also display similar responses to hypoxia, including: hypometabolism, reduced thermogenesis, reduced heart rate, reduced respiratory movements, increased blood catecholamines and glucocorticoids, and redistribution of blood circulation to the brain and heart (Mulder et al., 2000; Crossley et al., 2003; Mortola, 1
2
Landry et al.
2009). A similar “immature” reaction to hypoxia has also been described in human infants (Horne, 2013). To better understand whether embryo hypoxia is accompanied by changes in the functional states of the brain, neuronal responses to hypoxia were measured using cFos protein expression as an indicator of neuronal activity in pre-pipped (not yet air-breathing) embryonic day (E) 18 chick embryos. The effects of modest hypoxia (a reduction of oxygen content from the normal 21% to 15%, a level known to be nonlethal; Mortola, 2009) and medium hypoxia (a reduction to 10% oxygen, known to be lethal with more prolonged exposure than used here; Mortola JP, personal communication) were assayed using the activation of two key neuronal groups involved in brainstate control: the noradrenergic Locus Coeruleus (NA-LC) neurons in the brainstem, and the hypothalamic Hypocretin/Orexin (H/O) cell group. In adult mammals, these two groups of cells usually work synergistically to induce (NA-LC) and maintain (H/O) waking states, and typically increase and decrease their activity together in a coordinated fashion (Siegel, 2009); changes in brainstem NA-LC cell activity are also associated with changes in cFos activation of forebrain areas (Cirelli et al., 1996). NA-LC and H/O circuitry may not function in this same integrated and coordinated way in embryos (Balaban et al., 2012); cFos activation in sensory and association areas of the Pallium (avian analog of the cerebral cortex) were also examined to see if changes in brainstem NA-LC cell activity affect telencephalon activation in chick embryos.
METHODS Hypoxia Exposure and Perfusion Fertilized White Leghorn chicken eggs (Gallus gallus, n 5 18, Couvoir Simetin, Mirabel, QC) were incubated in darkness under constant conditions (38 C; 60% relative humidity) with automatic egg rotation every few hours. At day 18 (hatching is at 21 days), embryos were treated in the dark for 4 h, without anesthesia, with one of three conditions: 21% O2 (normoxic controls; n 5 6, three females, three males), 15% O2 (modest hypoxia; n 5 6, four males, two females), and 10% O2 (medium hypoxia; n 5 6, four females, two males). Hypoxia was obtained by placing the eggs into an incubator where a small stream of warmed and humidified N2 was supplied under the control of a flowmeter. An O2 Analyzer (Sable Systems Int., Las Vegas, NV) continuously sampled O2 concentration and displayed this information on a computer monitor. A data logger placed inside the incubator continuously sampled the temperature and humidity. At the end of the hypoxic or norDevelopmental Neurobiology
moxic treatment, embryos were anesthetized in ovo using isoflurane, and then intracardially perfused with Howard Ringer’s solution (123 mM NaCl, 1.5 mM CaCl2, 5 mM KCl; 6.5 mL/min, 2 min) followed by 10% formalin/PBS (150 mM NaCl, 28 mM KH2PO4, 72 mM K2HPO4, pH 7.4; 10 min). The gonads were examined to establish sex, and heads were removed, kept in 10% formalin/PBS at 4 C overnight, and brains were extracted. After cryoprotection in 10% and 30% sucrose/PBS, brains were frozen and cut into 40 mm serial coronal sections at a cryostat (Leica CM3050, Leica Microsystems Canada, Richmond Hill, ON).
Immunohistochemistry All procedures were performed at room temperature and washes were done in PBS (140 mM NaCl, 8 mM Na2HPO4, 2.7 mM KH2PO4, 1.5 mM KCl, pH 7.4). Standard fluorescent double-labeling immunohistochemical techniques were used to study cFos expression in H/O and NA-LC neurons. Antibody solutions were made in blocking solution (10% normal donkey serum, 2% bovine serum albumin, and 0.5% Triton X-100 in PBS). Sections spanning the whole length of the LC (1 section every 240 mm) were incubated for 1 h with blocking solution and then overnight with mouse anti-Tyrosine Hydroxylase (TH; ImmunoStar, 22941; 1:500) and cFos (Santa Cruz Biotechnologies, sc-253; 1:1,000) primary antibodies, together. After PBS washes, sections were incubated for 2 h with Alexa Fluor 594 anti-mouse (Molecular Probes, A21203; 1:500) and Alexa Fluor 488 anti-rabbit (Molecular Probes, A21206; 1:200) antibodies, together. Sections spanning the whole posterior hypothalamus (1 section every 240 mm) were incubated for 1 h with blocking solution and then overnight with goat anti-Orexin A (Santa Cruz Biotechnologies, sc-8070; 1:500) and rabbit anti-cFos (sc-253; 1:1,000) antibodies, together. After PBS washes, sections were incubated for 2 h with Alexa Fluor 594 anti-goat (Molecular Probes, A11058; 1:500) and Alexa Fluor 488 anti-rabbit (A21206; 1:200) antibodies, together. After the secondary antibody incubation, all sections were washed in PBS, counterstained with DAPI (Invitrogen, D1306; 0.1 mg/mL in PBS) for 10 min, washed again in PBS, and coverslipped. cFos expression in the Pallium was evaluated using standard colorimetric immunohistochemical techniques. Antibody solutions were made in blocking solution (10% normal goat serum, 2% bovine serum albumin, and 0.5% Triton X-100 in PBS). A cFos primary antibody (sc-253; 1:2,000), a biotinylated anti-rabbit antibody (Vector Labs, BA1000; 1:500), and the ABC (Vector Labs, PK-6100; 1:200) and DAB (Sigma) revelation system were used.
Quantification Procedures In every animal, every sixth section was stained and examined, meaning approximately seven forebrain, approximately seven hypothalamic, and approximately three brainstem sections were examined from each embryo (the
Hypoxia and Arousal in Chick Embryos LC sections of one modest hypoxia embryo were unusable). Anatomical landmarks were defined according to a standard atlas of the chick brain (Puelles et al., 2007). Co-labeling analyses of cFos in NA-LC neurons and cFos in H/O neurons in the hypothalamus were done at 2003 magnification using a microscope (Leica DM6000 B) and the optical fractionator program in Stereo Investigator (Version 10.55, MicroBrightField, Williston, VT). Contours were manually traced around each region of interest, and the program subdivided these into fields on a grid. Cells within each field were manually tagged according to their labeling characteristics, and these numbers were tallied over all fields. The fields accounted for 80% of the grid space in the H/O areas and 100% in the LC. To determine the percentage of cFosactivated neurons of each type, the total number of neurons per field stained for TH and H/O containing a nucleus (as seen from DAPI labeling) were counted and compared with the number that were co-labeled with cFos in the nucleus. H/O neurons were also subdivided into two anatomically distinct groups: a medial group (also more anterior), which occupies the developing Paraventricular Nucleus and the Periventricular Stratum and a lateral group, within the Posterior Hypothalamic Nucleus and Lateral Hypothalamic area. In rats, the medial and lateral groups receive inputs mostly from the hypothalamus and the brainstem, respectively (Yoshida et al., 2006). The density of cFos immunoreactive cells in the Pallium was measured at 1003 magnification in three areas. Two sensory areas where chosen to contrast a primary sensory area that would be presumably continuously driven by sensory input in embryos kept in the dark (the Basorostral Nucleus, that receives somatosensory inputs), and one that would presumably be less driven by sensory input in the dark (the Entopallial Nucleus, that receives visual inputs). An easily measured association area anatomically removed from both primary sensory areas (the rostral part of the Lateral Pallium) was also examined. Stereo Investigator was used to draw a contour around the regions of interest of these three areas, as defined by the chick brain atlas (Puelles et al., 2007), from which the volume was extrapolated using the contour-area section thicknesses. The number of cFos-labeled cells was determined using the Automatic Object detection program, set to count objects between 10 and 100 mm2 and above a constant luminance threshold. Camera and lighting settings were kept constant.
Statistical Analysis The variables studied here had skewed, asymmetric distributions; some were nominally significantly different from normal distributions and others were not. As statistical tests for normality at these sample sizes have low power, we uniformly used nonparametric statistics because they are the more conservative alternative. All significant effects remained significant if parametric tests were used, and none of the non-significant effects became significant. The Scheirer–Ray–Hare nonparametric two-way ANOVA (Sokal and Rohlf, 2012) was used to compare
3
cFos expression in the NA-LC and H/O neurons. Comparisons between the three treatment groups used Kruskal– Wallis nonparametric ANOVA followed by post-hoc tests corrected for multiple comparisons (Siegel and Castellan, 1988). Results of nonparametric ANOVAs are reported using the H-statistic, with the number of degrees of freedom indicated in parentheses. Differences in the activation level of medial and lateral H/O groups were compared using the nonparametric Wilcoxon matched-pairs sign rank test (Siegel and Castellan, 1988). The Spearman nonparametric correlation coefficient (q) was used for trend analysis. For all statistical analyses, data were considered statistically significant if P < 0.05, using a two-tailed test. An initial analysis of cell densities in the Pallium using a Kruskal–Wallis nonparametric ANOVA indicated that the overall cFos-labeled cell densities varied significantly among these three regions with data from all oxygen levels combined [H(2) 5 32.33, P < 0.0001]; densities were highest in the Lateral Pallium (mean 28,547 6 2,918/mm2), and much lower in the Basorostral Nucleus (mean 1,910 6 568/ mm2) and Entopallial Nucleus (mean 1,439 6 481/mm2); the values in the Lateral Pallium were significantly different from the values in the two other areas, which were not significantly different from each other. To do a single statistical comparison of all data using a nonparametric twoway ANOVA, cFos-labeled cell densities were normalized within each area by dividing all measurements by the mean density of the normoxic group.
RESULTS NA-LC and H/O Neurons Had Similar Overall cFos Expression Levels, But Different Hypoxia Responses The two populations of neurons did not differ in their mean levels of cFos expression over all treatments, but showed a highly significantly different pattern of responses to hypoxia [nonparametric two-way ANOVA, overall differences in percent activation between cell populations: H(1) 5 0.08, P 5 0.77; overall differences between different oxygen concentrations: H(2) 5 2.10, P 5 0.35; interaction between cell population and oxygen concentration: H(2) 5 15.83, P 5 0.0003]. The inverted pattern of responses between the two cell groups was also indicated by the significant negative correlation between their percentage expression values across all embryos [q 5 20.50, n 5 17 (LC slides from one animal being unusable), P < 0.05; Fig. 1(A)].
Hypoxia Increased cFos Expression in NA-LC Neurons There was a significant effect of oxygen level on the percentage of cFos-positive NA-LC neurons Developmental Neurobiology
4
Landry et al.
Figure 1 cFos expression in NA-LC and H/O neurons. (A) The scatterplot illustrates the negative relationship between cFos expressing (cFos1) neurons in NA-LC and H/O neurons. Each dot represents one embryo. (B) The two plots of the data from all individual subjects show the percentage of cFos1 neurons in each population under each experimental condition. Horizontal lines represent mean levels of each group. The * and § symbols above each graph represent the results of statistical tests corrected for multiple comparisons; groups with different symbols are significantly different from each other.
[H(2) 5 9.99, P 5 0.007]. Embryos in the normoxic condition showed a low percentage (mean 5%). Post-hoc tests showed that the percentage of cFosactivated NA-LC cells was not significantly higher than controls in the modest hypoxia group (mean 17%); however, the medium hypoxia group was significantly higher than both the control and modest groups (mean of 46%) [Figs. 1(B) and 2(A–C)].
Hypoxia Decreased cFos Expression in H/O Neurons H/O neurons also showed a significant effect of oxygen level on cFos expression [H(2) 5 6.75, P 5 0.035]. The percentage of Fos-positive H/O neurons was moderately high in normoxia (mean 29%), it was the same at 15% oxygen (mean 29%), but it was significantly lower at 10% oxygen than in both other conditions (mean 8%) [Figs. 1(B) and 2(D–F)]. The medial group of H/O neurons had, over all oxygen conditions combined, an average of 13% higher levels of activation than the lateral group (Wilcoxon test, z 5 3.03, P 5 0.003; Fig. 3). However, there was a very strong correlation between the activity levels of both groups across all experimental conditions (q 5 0.85, n 5 18, P 5 0.0005 with an outlying point included; q 5 0.91, n 5 17, P 5 0.0003 with it excluded; Fig. 3), indicating that they responded in a similar way to oxygen level manipulations. Developmental Neurobiology
Hypoxia Increased cFos Expression in Areas of the Pallium The relative density of cFos-positive cells (after being normalized according to the normoxic mean density in each region) exhibited significant variation between regions of the Pallium and between oxygen concentrations [nonparametric two-way ANOVA, overall differences between regions: H(2) 5 9.51, P 5 0.009; overall differences between different oxygen concentrations: H(2) 5 9.86, P 5 0.008]. The changes between the different oxygen conditions had similar patterns within the three pallial areas [interaction between cell population and treatment: H(4) 5 2.20, P 5 0.70]. The Basorostral Nucleus (with an average of 2.82 6 0.84 times the mean normoxic activation level), and the Lateral Pallium [1.22 6 0.13; Fig. 2(G–I)], had significantly higher densities of cFos-positive cells than the Entopallial Nucleus (0.95 6 0.32), indicative of greater overall responses to hypoxia. The former two areas were not significantly different from each other. As a group, the three areas had a significantly higher response to 10% oxygen (an average of 2.76 6 0.71 times their mean normoxic levels) than to 15% oxygen (0.88 6 0.20) and normoxia (1.00 6 0.22); the latter two densities were not significantly different from each other. To quantify the relationship between NALC cell activity and forebrain activity, Spearman correlation coefficients (Bonferroni-corrected for three
Hypoxia and Arousal in Chick Embryos
5
Figure 2 cFos expression in three examined anatomical regions in the different experimental conditions. The number of NA-LC neurons (in red) showing cFos expression (green nucleus) increases from normoxia (A) to modest hypoxia (B) to medium hypoxia (C). cFos expression (green nucleus) in H/O neurons (red) is similar in normoxia (D) and modest hypoxia (E) and is strongly decreased in medium hypoxia (F). Arrowheads indicate examples of double-labeled NA-LC and H/O neurons. All nuclei are counterstained with DAPI (blue). cFos expression in the Lateral Pallium increases in medium hypoxia (I) with respect to normoxia (G) and modest hypoxia (H). Scale bars are the same within each row, and they are all 30 mm. 4v: fourth ventricle; D: dorsal; L: lateral; M: medial; V: ventral.
comparisons) were calculated. The Basorostral Nucleus had a significant positive correlation with LC activation (q 5 0.66, n 5 15, corrected P 5 0.04); the correlations for both the Lateral Pallium (q 5 0.30, n 5 16, corrected P 5 0.72) and Entopallial Nucleus (q 5 0.43, n 5 16, corrected P 5 0.30) were not significant. In summary, three pallial areas considered as a group had a significantly higher level of normalized cFos activation to medium hypoxia than to modest hypoxia or normoxia; the Basorostral Nucleus and Lateral Pallium both had significantly higher normalized activation to modest hypoxia than the Entopallial nucleus. Only the Basorostral Nucleus had a significant positive correlation with LC activation.
DISCUSSION NA-LC and H/O neurons are both known to be importantly involved in adult brain state regulation. NA-LC neurons, their diffuse projections, and adrenergic receptors are well developed by E18 in chick embryos (Yurkewicz et al., 1981; Dermon and Kouvelas, 1988). We recently studied the development of H/O neurons and receptors in the brain of chick embryos of different ages (Godden KE, Landry JP, Slepneva N, MiguesBlanco V, Pompeiano M. Early expression of Hypocretin/Orexin in the chick embryo brain. Submitted; Cerazy T, Saadat N, Fuchs G, Pompeiano M. Hypocretin/Orexin receptors in the chick embryo brain. In preparation). This work demonstrated that H/O neurons and their receptors are also strongly present by E18. The Developmental Neurobiology
6
Landry et al.
Figure 3 cFos expression in two anatomically separated groups of H/O neurons. The scatterplot illustrates the positive relationship between cFos expression in lateral and medial H/O neurons. The identity line (y 5 x; dotted line) shows that the medial group has a higher percentage of cFos-expressing (cFos1) neurons than the lateral group. The arrow indicates an outlying data point (see text for details).
neuroanatomical substrates by which these two systems could mediate hypoxic effects are, therefore, wellestablished by the age at which the present studies were conducted. The data show that hypoxia affected the activity of both cell groups in late-stage chick embryos prior to air-breathing. cFos-positive NA-LC cells were significantly increased during medium hypoxia relative to both normoxia and modest hypoxia conditions. In contrast, cFos-expressing H/O neurons were significantly decreased during hypoxia. The overall density of cFos-positive cells in the Pallium increased during hypoxia, with a similar pattern in two sensory areas and one association area. However, the pattern of changes over different oxygen concentrations was significantly correlated with NA-LC activation only in a somatosensory area, the Basorostral Nucleus. This contrasted with no correlation in the Entopallial Nucleus, a presumably less-actively used visual sensory area (given experiments run in darkness), or in an association area (Lateral Pallium). Increased LC cFos expression was previously found in fetal sheep exposed to acute hypoxia, but the activated neurons were never chemically identified (Nitsos and Walker, 1999).
Possible Role of NA-LC and H/O Neurons in the Physiological Responses of the Embryo to Hypoxia Many of the previously observed effects of hypoxia on chick embryos can be plausibly explained by increased activation of NA-LC neurons coupled with decreased activation of H/O neurons. However, a contribution Developmental Neurobiology
from other catecholaminergic areas, such as the A1/C1, A5, and A7 groups, is possible (Card et al., 2006; Kanbar et al., 2011; Bruinstroop et al., 2012). Both NA-LC and H/O neurons activate the sympathetic system and inhibit the parasympathetic system (Samuels and Szabadi, 2008; Williams and Burdakov, 2008; Dergacheva et al., 2011; Shahid et al., 2012). As H/O neurons innervate cardiorespiratory regulatory areas in birds (Singletary et al., 2007), reduced heart rate and respiratory movements during hypoxia in chick embryos could be explained by the decreased activity of H/O neurons; hypoxia-induced reductions in metabolism and thermogenesis (Mortola, 2009, 2011) may have a similar explanation (Asakawa et al., 2002; Girault et al., 2012). Chick embryos display increased motor activity in response to hypoxia (Windle and Barcroft, 1938; Gr€ans and Altimiras, 2007; Mortola et al., 2013); this may serve as a distress signal to the hen to change incubation positions (Gr€ans and Altimiras, 2007). H/O neurons are known to stimulate locomotor activity in adult rats (Zheng et al., 2005) and in adult and neonatal birds (da Silva et al., 2008; Katayama et al., 2010). As H/O neuron activity decreases, increased motor activity in hypoxic embryos is more likely to be dependent on the activation of NA-LC neurons, which exert a facilitatory action on spinal motor neurons (Fung et al., 1991; Burgess and Peever, 2013). NA-A7 neurons could also contribute (Bruinstroop et al., 2012), as noradrenergic innervation of the spinal cord is fully established by E18 in chick embryos (Singer et al., 1980).
Possible Circuitry Regulating NA-LC and H/O Activity During Low-Oxygen Exposure in Chick Embryos We hypothesize that hypoxia-sensitive A1/C1 neurons could play an important role in orchestrating brain responses to hypoxia in chick embryos by modulating regulatory brain areas rather than by directly influencing autonomic areas, as also suggested by the distribution of their projections (McKellar and Loewy, 1982; Batten, 1995). In adults, A1/C1 neuron stimulation causes arousal and cardiorespiratory activation (Abbott et al., 2013). In hypoxic embryos, A1/C1 neurons could drive NA-LC activation as in adults (Guyenet et al., 2013). However, the chick embryo cardiorespiratory response to hypoxia (before the start of air-breathing) is the opposite of what is observed postnatally. It is impossible for the embryo to dramatically increase the availability of oxygen through the chorioallantoic membrane (except by alerting the hen). We suggest that the differences between prenatal and postnatal responses could be specifically due to the inhibition of
Hypoxia and Arousal in Chick Embryos
H/O neurons in embryos (we expect H/O neurons to be activated by hypoxia in air-breathing chicks; see Abbott et al., 2013 and Liu et al., 2014. Inhibition of H/O neurons would subserve a neuroprotective role, imposing a global state of reduced brain activity and energy expenditure (Liu et al., 2011). Inhibition of H/O neurons could be initially mediated by A1/C1 neuron activation (Sakurai et al., 2005) even though H/O neurons are known to decrease their activity in the presence of low ATP levels (Liu et al., 2011). There are at least three mechanisms that are possible candidates for the developmental changes in H/O neuron hypoxic responses: (1) a change of sign in the response to catecholamines [the expression of a2 adrenergic receptors (inhibitory) precedes that of a1 and b receptors (both excitatory) in several areas of the chick brain (Dermon and Kouvelas, 1988)]; (2) a change in the sign of GABAergic inputs to H/O neurons (from excitatory to inhibitory); and (3) the development of peripheral, with respect to central A1/C1-dependent, chemoreception [carotid bodies are rather inactive before birth (Purves, 1982), but become postnatally activated by hypoxia, inducing cardiorespiratory activation and arousal (Nattie and Li, 2013; Guyenet and Abbott, 2013)]. The lower cFos expression we saw in the lateral group of H/O neurons with respect to the medial group could be a reflection of the relative immaturity of peripheral chemoreception.
Pallial Activity Paralleling NA-LC Activity In adult rodents, primary sensory areas of the cerebral cortex show the largest changes in cFos expression between sleep and waking and these changes are NALC-dependent (Cirelli et al., 1996). In this study, hypoxia induced significant changes in Pallium cFos expression. The pattern of changes was significantly correlated with NA-LC cFos activity in one somatosensory area likely to have continuously received sensory input, but not in a visual sensory area that was not likely to have been extensively activated by sensory input (given experiments conducted in the dark), nor in an association area. These results suggest that the chick embryo NA-LC system activates at least some forebrain areas, as in adult rodents. Postnatally, arousal is an adaptive reaction to hypoxia, allowing an individual to take behavioral action to counteract negative and possibly fatal effects of hypoxia (Guyenet and Abbott, 2013). Increased cFos expression in the embryo telencephalon during hypoxia would be consistent with this pattern; however, hypoxia appears to trigger an incomplete arousal reaction in embryos because of the simultaneous inhibition of H/O neurons.
7
In summary, this study has begun to elucidate how neural systems involved in brain-state control in adults may centrally orchestrate the pattern of hypoxic responses in embryos toward the end of prenatal development. These results show for the first time that oxygen deprivation activates NA-LC, while simultaneously deactivating H/O neurons of the hypothalamus, and that hypoxia also results in increased activation in at least some areas of the Pallium (analogous to the cerebral cortex in mammals). Future work is needed to elucidate the cellular mechanisms regulating the opposing cardio-circulatory and respiratory responses and the different pattern of brain responses triggered by hypoxia in embryos with respect to post-natal organisms. The authors thank J.P. Mortola for providing the hypoxic eggs, Jae Duke Sheen and Mashiyat Majid for help with the cryostat sectioning, and J.P. Mortola and J. Mogil for comments on the manuscript.
REFERENCES Abbott SBG, Coates MB, Stornetta RL, Guyenet PG. 2013. Optogenetic stimulation of C1 and retrotrapezoid nucleus neurons causes sleep state-dependent cardiorespiratory stimulation and arousal in rats. Hypertension 61:835–841. Asakawa A, Inui A, Goto K, Yuzuriha H, Takimoto Y, Inui T, Katsuura G, et al. 2002. Effects of agouti-related protein, orexin and melanin-concentrating hormone on oxygen consumption in mice. Int J Mol Med 10:523–525. Balaban E, Desco M, Vaquero J-J. 2012. Waking-like brain function in embryos. Curr Biol 22:852–861. Batten TFC. 1995. Immunolocalization of putative neurotransmitters innervating autonomic regulating neurones of cat ventral medulla. Brain Res Bull 37:487–506. Bruinstroop E, Cano G, Vanderhorst VGJM, Cavalcante JC, Wirth J, Sena-Esteves M, Saper CB. 2012. Spinal projections of the A5, A6 (Locus Coeruleus), and A7 noradrenergic cell groups in rats. J Comp Neurol 520:1985–2001. Burgess CR, Peever JH. 2013. A noradrenergic mechanism functions to couple motor behavior with arousal state. Curr Biol 23:1719–1725. Card JP, Sved JC, Craig B, Raizada M, Vasquez J, Sved AF. 2006. Efferent projections of rat rostroventrolateral medulla C1 catecholamine neurons: Implication for the central control of cardiovascular regulation. J Comp Neurol 499:840–859. Cirelli C, Pompeiano M, Tononi G. 1996. Neuronal gene expression in the waking state: A role for the Locus Coeruleus. Science 274:1211–1215. Crossley DA, Burggren WW, Altimiras J. 2003. Cardiovascular regulation during hypoxia in embryos of the domestic chicken Gallus gallus. Am J Physiol Regul Integr Comp Physiol 284:R219–226. da Silva ES, dos Santos TV, Hoeller AA, dos Santos TS, Pereira GV, Meneghelli C, Pezlin AI, et al. 2008. Behavioral Developmental Neurobiology
8
Landry et al.
and metabolic effects of central injections of orexins/hypocretins in pigeons (Columba livia). Regul Pept 147:9–18. Dergacheva O, Philbin K, Bateman R, Mendelowitz D. 2011. Hypocretin-1 (orexin A) prevents the effects of hypoxia/hypercapnia and enhances the GABAergic pathway from the lateral paragigantocellular nucleus to cardiac vagal neurons in the nucleus ambuguus. Neuroscience 175:18–23. Dermon CR, Kouvelas ED. 1988. Binding properties, regional ontogeny and localization of adrenergic receptors in chick brain. Int J Dev Neurosci 6:471–482. Fung SJ, Manzoni D, Chan JY, Pompeiano O, Barnes CD. 1991. Locus coeruleus control of spinal motor output. Prog Brain Res 88:395–409. Girault EM, Yi C-X, Fliers E, Kalsbeek A. 2012. Orexins, feeding, and energy balance. Progr Brain Res 198:47–64. Gr€ans A, Altimiras J. 2007. Ontogeny of vocalizations and movements in response to cooling in chickens fetuses. Physiol Behav 91:229–239. Guyenet PG, Abbott SBG. 2013. Chemoreception and asphyxia-induced arousal. Respir Physiol Neurobiol 188: 333–343. Guyenet PG, Stornetta RL, Bochorishvili G, Depuy SD, Burke PG, Abbott SB. 2013. C1 neurons: The body’s EMTs. Am J Physiol Regul Integr Comp Physiol 305:R187–204. Horne RSC. 2013. Cardio-respiratory control during sleep in infancy. Pediatr Respir Rev. http://dx.doi.org/10.1016/ j.prrv.2013.02.012. Kanbar R, Depuy GH, Stornetta RL, Guyenet PG. 2011. Regulation of visceral sympathetic tone by A5 noradrenergic neurons in rodents. J Physiol 589:903–917. Katayama S, Hamasu K, Shigemi K, Cline MA, Furuse M. 2010. Intracerebroventricular injection of orexin-A, but not orexin-B, induces arousal of layer-type neonatal chicks. Comp Biochem Physiol A Mol Integr Physiol 157:132–135. Liu Z, Jiang L, Zhu F, Fu C, Lu S, Zhou J, Wu X, Bai C, Li S. 2014. Chronic intermittent hypoxia and the expression of orexin and its receptors in the brains of rats. Sleep Biol Rhythms 12:22–29. Liu Z-W, Gan G, Suyama S, Gao X-B. 2011. Intracellular energy status regulates activity in hypocretin/orexin neurons: A link between energy and behavioural states. J Physiol 589:4157–4166. McKellar S, Loewy AD. 1982. Efferent projections of the A1 catecholamine cell group in the rat: An autoradiographic study. Brain Res 241:11–29. Mortola JP. 2009. Gas exchange in avian embryos and hatchlings. Comp Biochem Physiol A Mol Integr Physiol 153:359–377. Mortola JP. 2011. Metabolic and ventilatory sensitivity to hypoxia in avian embryos. Respir Physiol Neurobiol 178: 174–180. Mortola JP, Louis AS, Simeonova M, Toro Velasquez PA. 2013. The motility of the chicken embryo: Energetic cost and effects of hypoxia. Respir Physiol Neurobiol 188:172– 179. Mulder ALM, Golde JMCG, Goor AAC, Giussani DA, Blanco CE. 2000. Developmental changes in plasma catDevelopmental Neurobiology
echolamine concentrations during normoxia and acute hypoxia in the chick embryo. J Physiol 527:593–599. Nattie E, Li A. 2013. Respiration and autonomic regulation and orexin. Progr Brain Res 198:25–46. Nitsos I, Walker DW. 1999. The distribution of FOSimmunoreactive neurons in the brainstem, midbrain and diencephalon of fetal sheep in response to acute hypoxia in mid and late gestation. Dev Brain Res 114:9–26. Puelles L, Paxinos G, Watson C, Martinez S, Martinez-dela-Torre M. 2007. The Chick Brain in Stereotaxic Coordinates: An Atlas Based on Neuromeres. New York: Academic Press. 272 p. Purves MJ. 1982. Respiratory sensitivity before and after birth. Acta Ped Scand 71:529–536. Reiner A, Perkel DJ, Bruce LL, Butler AB, Csillag A, Kuenzel W, Medina L, et al. 2004. Revised nomenclature for avian telencephalon and some related brainstem nuclei. J Comp Neurol 473:377–414. Sakurai T, Nagata R, Yamanaka A, Kawamura H, Tsujino N, Muraki Y, Kageyama H, et al. 2005. Input of orexin/ hypocretin neurons revealed by a genetically encoded tracer in mice. Neuron 46:297–308. Samuels ER, Szabadi E. 2008. Functional neuroanatomy of the noradrenergic locus coeruleus: Its roles in the regulation of arousal and autonomic function part I: Principles of functional organisation. Curr Neuropharmacol 6:235–253. Shahid IZ, Rahman AA, Pilowsky PM. 2012. Orexin and central regulation of cardiorespiratory system. Vitam Horm 89:159–184. Siegel JM. 2009. The neurobiology of sleep. Semin Neurol 29:277–296. Siegel S, Castellan NJJ. 1988. Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill. 399 p. Singer HS, Coyle JT, Vernon N, Kallman CH, Price DL. 1980. The development of catecholaminergic innervation in chick spinal cord. Brain Res 191:417–428. Singletary KG, Deviche P, Strand C, Delville Y. 2007. Distribution of orexin/hypocretin immunoreactivity in the brain of a male songbird, the house finch, Carpodacus mexicanus. J Chem Neuroanat 33:101–109. Sokal RR, Rohlf JF. 2012. Biometry. New York: W.H. Freeman. 937 p. Williams RH, Burdakov D. 2008. Hypothalamic orexins/ hypocretins as regulators of breathing. Expert Rev Mol Med 10:e28. Windle WF, Barcroft J. 1938. Some factors governing the initiation of respiration in the chick. Am J Physiol—Legacy Content 121:684–691. Yoshida K, McCormack S, Espa~na RA, Crocker A, Scammell TE. 2006. Afferents to the orexin neurons of the rat brain. J Comp Neurol 494:845–861. Yurkewicz L, Marchi M, Lauder JM, Giacobini E. 1981. Development and aging of noradrenergic cell bodies and axon terminals in the chicken. J Neurosci Res 6:621–641. Zheng H, Patterson LM, Berthoud HR. 2005. Orexin-A projections to the caudal medulla and orexin-induced cFos expression, food intake, and autonomic function. J Comp Neurol 485:127–142.
