Intermittent hypoxia-induced cardiorespiratory long-term facilitation: A new role for microglia

G Model

ARTICLE IN PRESS

RESPNB-2633; No. of Pages 9

Respiratory Physiology & Neurobiology xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Respiratory Physiology & Neurobiology journal homepage: www.elsevier.com/locate/resphysiol

Review

Intermittent hypoxia-induced cardiorespiratory long-term facilitation: A new role for microglia Seung Jae Kim a,b , Yeon Jae Kim a , Zohra Kakall a,b , Melissa M.J. Farnham a,b , Paul M. Pilowsky a,b,∗ a b

Department of Physiology, Faculty of Medicine, The University of Sydney, Sydney, New South Wales 2006, Australia The Heart Research Institute, 7 Eliza Street, Newtown, Sydney 2042, Australia

a r t i c l e

i n f o

Article history: Received 2 March 2016 Received in revised form 18 March 2016 Accepted 18 March 2016 Available online xxx Keywords: Phrenic long-term facilitation Sympathetic long-term facilitation Acute intermittent hypoxia Microglia Orexin PACAP

a b s t r a c t Intermittent hypoxia induces plasticity in neural networks controlling breathing and cardiovascular function. Studies demonstrate that mechanisms causing cardiorespiratory plasticity rely on intracellular signalling pathways that are activated by specific neurotransmitters. Peptides such as serotonin, PACAP and orexin are well-known for their physiological significance in regulating the cardiorespiratory system. Their receptor counterparts are present in cardiorespiratory centres of the brainstem medulla and spinal cord. Microglial cells are also important players in inducing plasticity. The phenotype and function of microglial cells can change based on the physiological state of the central nervous system. Here, we propose that in the autonomic nuclei of the ventral brainstem the relationship between neurotransmitters and neurokines, neurons and microglia determines the overall neural function of the central cardiorespiratory system. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The cardiorespiratory network of the brainstem displays an anatomically distinct neural circuitry. Peripheral sensors (neural, oxygen and pH) detect changes in the external milieu relay information via afferent pathways to brainstem centres. These sensory inputs are integrated within brainstem regulatory centres resulting in appropriate changes in the activity of efferent neurons in order to restore homeostasis (Pilowsky and Goodchild, 2002). This neural reflex loop is not hard-wired; rather, there exists a capacity for considerable plasticity. Acute intermittent hypoxia (AIH) causes long-term changes to the way neurons fire in the brainstem nuclei (Xing and Pilowsky, 2010). A well-established concept of this plasticity in the cardiorespiratory regulating pathways of the brain, following intermittent exposures of hypoxia, is termed long-term facilitation (LTF; pLTF in the phrenic motor pathway for breathing, and sLTF for the sympathetic output to the cardiovascular system). Two types of intermittent hypoxia protocols are used for experimental models to study complications arising from the cardiorespiratory system. These include: AIH (Dick et al., 2014,

∗ Corresponding author at: The Heart Research Institute, 7 Eliza Street, Newtown, New South Wales 2042, Australia. E-mail address: [email protected] (P.M. Pilowsky).

2004, 2007; Xing et al., 2013, 2014; Xing and Pilowsky, 2010), and chronic intermittent hypoxia (CIH) (Braga et al., 2006; Leuenberger et al., 2005; Prabhakar et al., 2005; Zoccal et al., 2007, 2008). The physiological importance of LTF in the respiratory and sympathetic pathway remains inconclusive. Communication of signals between neurons is achieved by neurotransmitters. Certain experiences prompt various combinations of endogenous neurotransmitters to be released, and this leads to a change in system behaviour. Additionally, the integrity of neural connections is supported and regulated by microglial cells. Microglia actively interact with neurons to shape the network, and the structure of the central nervous system (CNS) (Kettenmann et al., 2013). Microglia are extremely plastic in nature; the state of the CNS determines the morphological and functional phenotype of the cells. An interplay between neurons and microglia is mediated by various chemicals, such as neurotransmitters (Pocock and Kettenmann, 2007), cytokines, brain-derived neurotrophic factor (BDNF), and transforming growth factor (TGF)-␤ (Biber et al., 2007). Therefore, we hypothesise that neurotransmitters and microglia in the cardiorespiratory system may play a key role in generating neuronal plasticity. The mechanisms underlying the generation of cardiorespiratory plasticity following acute intermittent hypoxia are well-studied, but remain unclear (Dale-Nagle et al., 2010; Devinney et al., 2013; Mitchell et al., 2001; Xing et al., 2013, 2014). Here we review

http://dx.doi.org/10.1016/j.resp.2016.03.012 1569-9048/© 2016 Elsevier B.V. All rights reserved.

Please cite this article in press as: Kim, S.J., et al., Intermittent hypoxia-induced cardiorespiratory long-term facilitation: A new role for microglia. Respir. Physiol. Neurobiol. (2016), http://dx.doi.org/10.1016/j.resp.2016.03.012

G Model RESPNB-2633; No. of Pages 9

ARTICLE IN PRESS S.J. Kim et al. / Respiratory Physiology & Neurobiology xxx (2016) xxx–xxx

2

the known molecular pathways involved in LTF. We suggest that other neurotransmitters are involved in the development of LTF. Furthermore, the role of microglia in supporting the central cardiorespiratory network and the role microglial cells may play in generating intermittent hypoxia-induced LTF will also be discussed. 2. Mechanisms in the generation of pLTF after intermittent hypoxia Long term facilitation in neural activity following intermittent hypoxia conventionally refers to facilitation in the activity recorded from the phrenic nerve. Facilitation in phrenic nerve activity is a form of serotonin- and protein synthesis-dependent plasticity that takes place within the phrenic motor nucleus in the spinal cord (Bach and Mitchell, 1996; Baker-Herman et al., 2004; Fuller et al., 2001). Induction of pLTF is a state of enhanced respiratory neural output, that is independent of increased phrenic nerve activity (Baker and Mitchell, 2000; Xing et al., 2013). Furthermore, the plasticity is mediated by the action of neuromodulators (Dale-Nagle et al., 2010; Dick et al., 2007; Toyama et al., 2009). Although the current understandings of mechanisms giving rise to pLTF remain incomplete, substantial progress has been made, and the key elements will be discussed below. 2.1. Q pathway Acute intermittent hypoxia-induced pLTF is characterised by a prolonged facilitation in respiratory motor activity from the phrenic and hypoglossal (XII) nerves that are regulated by various neurotransmitters (Bach and Mitchell, 1996; Baker and Mitchell, 2000). First, respiratory neural activity is regulated by transmission from serotonin-containing neurons of the raphe nuclei to the spinal cord (Murphy et al., 1995; Pilowsky et al., 1990). Secondly, hypoxia activates the pontine catecholamine neurons in the locus coeruleus and A5 that release noradrenaline in spinal and cranial motornuclei including phrenic nucleus, the nucleus ambiguus and the hypoglossal (XII) nerve (Berkowitz et al., 2005; Gatti et al., 1999; Sun et al., 2002, 2003, 1995). Induction of pLTF by AIH occurs via mechanisms requiring the activation of spinal serotonergic (5HT2 ) receptor subtypes, or brainstem ␣1 -adrenoreceptors that are coupled to the Gq protein (Dale-Nagle et al., 2010; Dick et al., 2007); thus the pathway is referred to as the ‘Q pathway’. The initiation, but not sustained elevation, of pLTF by AIH is blocked by administration of 5HT2 receptor (Fuller et al., 2001) and ␣1 -adrenoreceptor (Neverova et al., 2007) antagonists. Adrenergic ␣1 receptor activation, however, is mechanistically sufficient but not necessary for pLTF (Huxtable et al., 2014). The maintenance in the persistent state of motor discharge requires contributions from intracellular proteins within the spinal phrenic motor neurons. Episodic spinal 5HT2 receptor activation is necessary and sufficient for new synthesis of BDNF (Baker-Herman et al., 2004) from the activated protein kinase C (PKC) ␪ effector protein isoform that is strongly coupled to 5HT2 receptors (Devinney et al., 2015), and an increased expression of the high-affinity TrkB receptor leading to ERK/MAPK signalling (Hoffman et al., 2012). Downstream of ERK, there is an increase in post-synaptic glutamate receptor phosphorylation. This leads to an enhancement in glutamatergic transmission within phrenic motor neurons, causing a robust increase in phrenic nerve activity (Fig. 1). 2.2. S pathway The mechanism by which is pLTF generated is incomplete, however metabotropic receptors coupled to Gs protein, referred to as the ‘S pathway,’ are believed to make a significant contribution.

Hypoxia increases extracellular adenosine levels via active transport and ATP degradation (Gourine et al., 2002). Adenosine release leads to the activation of Gs protein-coupled A2A receptor in the spinal cord. The Gs receptor pathway is mediated by exchange protein directly activated by cAMP (EPAC) effectors (Fields et al., 2015), requiring newly synthesised proteins comprised of immature TrkB isoforms via mTORC1 signalling, and downstream activation of PI3kinase/Akt signalling (Golder et al., 2008) (Fig. 1). Q and S pathways are distinct mechanisms that interact via ‘cross-talk’ inhibition (Hoffman et al., 2010). ‘Cross-talk’ mechanisms between the intracellular Gq and Gs pathways have been frequently demonstrated in past studies (Lai et al., 1997; Zimmermann and Taussig, 1996). Activation of 5-HT7 receptors in the cervical spinal cord elicits pLTF (S pathway), but also constrains 5-HT2 receptor-induced pLTF (Q pathway); this phenomenon is likely to be mediated by divergent cAMP/PKA signalling (Fields et al., 2015). However, some model systems do require PKA activation to induce certain forms of plasticity e.g. CREB-mediated plasticity (Lonze and Ginty, 2002; Silva et al., 1998). Interestingly, the Q and S pathways adopt differential roles that are dependent on the ‘intensities’ of hypoxemic challenges placed on the animals. The ‘intensity’ of intermittent hypoxia stimuli depends on two factors: (a) the level of oxygen in the hypoxic gas mixture (i.e. the% of oxygen, balanced in nitrogen), and (b) the number of cycles, or frequency of the hypoxia (Navarrete-Opazo and Mitchell, 2014), which leads to the overall hypoxemic state of the animal. The S pathway does not contribute to pLTF following ‘moderate’ hypoxia (three 5-min episodes of isocapnic inspired 11% O2 to maintain PaO2 at 45–55 mmHg, separated by 5-min intervals of baseline 51% O2 ) (Hoffman et al., 2010; Nichols et al., 2012), but dominates pLTF following severe AIH (Pa O2 level at 25–30 mmHg) (Nichols et al., 2012). The underlying mechanism is that severe, but not moderate, AIH activates the mTOR protein that is essential for activating PI3-kinase/Akt pathway (Dougherty et al., 2015). A difference in the level of extracellular adenosine accumulation during severe hypoxia may also contribute to the transition, increasing the number of activated A2A receptors. These findings demonstrate that cross-talk inhibition ensures the switch of dominance from one mechanism to the other. Furthermore, the transition between the one pathway to other follows a reference Pa O2 threshold: in that levels above 35 mmHg elicit pLTF via the Q pathway, whereas levels of 30 mmHg or below elicit pLTF via the S pathway (Nichols et al., 2012). Indeed, central serotonin knockout (Tph2−/− ) C57BL/6129SV mice display physical ventilatory LTF following exposure to severe intermittent hypoxia (i.e. lower oxygen concentration and greater frequency of hypoxic bursts), comprised of 4-min periods of 10% oxygen (balanced with nitrogen) interspersed with 4-min recovery intervals, 12 times for 10 days (12th hypoxia lasted for 45 min) (Hickner et al., 2014). Also, smaller animals (mice vs. rats) were used, which cause discrepancies in the relative capacity of accessory respiratory muscles to express plasticity. It is most likely that respiratory facilitation was caused by a shift in dominance from the impaired serotonin-dependent Q pathway to the S pathway. 2.3. Orexin A hallmark of pLTF is pattern sensitivity. Phrenic LTF is elicited by intermittent, but not continuous, hypoxia. In spite of the significant advances made to understand the neuromodulatormediated mechanisms underlying the generation of pLTF, the reasons that pattern-oriented stimuli are crucial remain unclear. Kuwaki and colleagues demonstrated that orexin-containing neurons are activated by distinct pattern-dependent intermittent hypoxic challenges (Terada et al., 2008; Toyama et al., 2009; Yamaguchi et al., 2015). In addition, functional anatomy and in vitro studies show that orexin neurons send dense fibre networks to

Please cite this article in press as: Kim, S.J., et al., Intermittent hypoxia-induced cardiorespiratory long-term facilitation: A new role for microglia. Respir. Physiol. Neurobiol. (2016), http://dx.doi.org/10.1016/j.resp.2016.03.012

G Model RESPNB-2633; No. of Pages 9

ARTICLE IN PRESS S.J. Kim et al. / Respiratory Physiology & Neurobiology xxx (2016) xxx–xxx

3

Fig. 1. A schema summarising the molecular components involved in the mechanisms of respiratory and sympathetic long-term facilitation. Intermittent hypoxia causes activation of orexin neurons in the hypothalamus. Axons from orexin neurons descend, and provide inputs to the serotonergic raphe nucleus, the RVLM, and the spinal cord IML. The release of orexin may be important to the generation of both pLTF and sLTF. Serotonin released from the raphe nucleus act on 5HT2 receptor on spinal phrenic motor neurons. Subsequently, the downstream Q pathway leads to phosphorylation of ERK/MAPK pathway. This phosphorylation causes the enhancement of postsynaptic glutamatergic receptors, and strengthens glutamatergic transmission. Hypoxia also increases extracellular adenosine concentrations, leading to the activation of A2A receptors on spinal phrenic motoneurons. The S pathway is activated, and phosphorylates the Akt pathway. S pathway activation also leads to strengthening of glutamatergic transmission. PACAP may be an important neurotransmitter in the generation of sLTF. PAC1 receptors directly and indirectly activate CREB proteins via a PKA-mediated pathway, and the activation of ␤-adrenoreceptors in the adrenal medulla respectively. Brain and spinal cord maps from Paxinos and Watson (2006).

the raphe nuclei in the brainstem (Brown et al., 2001; Liu et al., 2002) (Fig. 1). Therefore, supramedullary orexin pathways may play a role in mediating the effects of intermittent hypoxia during ventilations. Nakamura et al. (2007) suggested a potential role of orexin in sleep apnoea, where orexin knockout mice displayed a two-fold increase in occurrence index of the frequency of spontaneous sleep apnoeic events during both slow-wave sleep and rapid eye movement sleep, compared to their wild-type littermates as controls. Therefore, orexin may contribute to preserving ventilatory stability and integrity during sleep. Supporting this idea, orexin knockout mice did not express ventilatory LTF for at least 2 h following intermittent hypoxia (Terada et al., 2008). Indeed, facilitation of phrenic motor activity following AIH in orexin neuron-ablated mice is markedly attenuated (Toyama et al., 2009). Although not entirely conclusive, the studies conducted by Kuwaki lab suggest that hypothalamic orexin may be contributing to the pattern-sensitivity of pLTF. It is still unknown how orexin neurons respond to intermittent hypoxic stimuli. Further investigation is required to determine if orexin neurons are intrinsically sensitive to hypoxia, or if there are hypoxia-sensitive inputs that regulate orexin neurons. One possibility may be that hypoxia-sensitive C1 neurons residing in the rostroventrolateral medulla (RVLM) act to excite orexin neurons (Sun and Reis, 1994). A C1-orexin

neuronal connection may represent an indirect route through which intermittent hypoxia causes the release of orexin. Anterograde tracing with viral vectors has demonstrated the projections of C1 neurons to the orexin abundant area of the hypothalamus (Bochorishvili et al., 2014). These cells were predominantly tyrosine hydroxylase (TH)- and phenylethanolamine N-methyltransferase (PNMT)-immunoreactive (ir), and were detected among orexin-ir somata. Furthermore, the PNMT-ir axonal varicosities and orexinir were closely apposed (Bochorishvili et al., 2014). This connection could explain the reason why the RVLM was found to regulate breathing, the circulatory system, and arousal (Abbott et al., 2013a, 2013b). We suggest that a collective effort from supraspinal areas, such as the hypothalamus and the brainstem, may be required to evoke spinal mechanisms to generate pLTF. 3. Sympathetic LTF: pathogenic feature of neurogenic hypertension A key feature of hypertension is a persistent increase in SNA. There is an elevation in the sympathetic activity from medullary cardiovascular nuclei to the heart and blood vessels (Guyenet, 2006). Intermittent hypoxic challenges, such as in sleep apnoea patients, often cause hypertension (Cutler et al., 2004b; Leuenberger et al., 2005). A prolonged increase in sympathetic

Please cite this article in press as: Kim, S.J., et al., Intermittent hypoxia-induced cardiorespiratory long-term facilitation: A new role for microglia. Respir. Physiol. Neurobiol. (2016), http://dx.doi.org/10.1016/j.resp.2016.03.012

G Model RESPNB-2633; No. of Pages 9

ARTICLE IN PRESS S.J. Kim et al. / Respiratory Physiology & Neurobiology xxx (2016) xxx–xxx

4

discharge, or otherwise sLTF (Cutler et al., 2004b; Dick et al., 2007; Leuenberger et al., 2005; Xing and Pilowsky, 2010; Zoccal et al., 2007), lead to hypertension. Additionally, a progressive enhancement in peripheral chemoreflex sensitivity is another critical pathogenic feature of hypertension (Braga et al., 2006; Cutler et al., 2004a; Greenberg et al., 1999; Imadojemu et al., 2007; Xing et al., 2014). The mechanisms leading to the augmentation in sympathetic activity and sensitivity remain unclear. Serotoninergic transmission appears to be important for the development of sLTF (Dick et al., 2007), although this has been discussed in detail elsewhere (Dick et al., 2014; Xing et al., 2014). Here we discuss pituitary adenylate cyclase activating polypeptide (PACAP) and orexin neurotransmitters and their role in the sympathetic responses to AIH. We also highlight possible reasons why these peptide neurotransmitters may be important in intermittent hypoxia-induced sLTF. 3.1. PACAP PACAP is a sympathoexcitatory neuropeptide that exists both centrally and peripherally (Farnham and Pilowsky, 2010). In the CNS, PACAP-38 is the most abundant isoform present. The most centrally abundant PACAP receptor isoform is the PAC1 receptor. PACAP exhibits sympathoexcitatory effects in the brainstem and spinal cord, lasting for at least 60 min (Farnham et al., 2008, 2012). PACAP receptor coupling mechanisms are well understood (Dickson and Finlayson, 2009; Farnham and Pilowsky, 2010). Here, we aim to address the potential role PACAP may play in the generation of sLTF. PAC/VPAC receptors are expressed on cardio-regulatory neurons: including presympathetic neurons of the RVLM, adrenal medullary chromaffin cells, and the intermediolateral (IML) column of the spinal cord (T5/6 ) (Gaede et al., 2012). Activation of the G␣s -coupled PAC1 receptor alters downstream gene transcription and protein phosphorylation. In this process, the activation of PKA and CREB are implicated in increasing synaptic strength (Baxter et al., 2011). It has been demonstrated that hypoxia up-regulates the expression of PACAP receptor proteins (Lam et al., 2012). First indicated by Farnham et al. (2008), intrathecal administration of PACAP to the spinal T5/6 causes prolonged sympathoexcitation and tachycardia. In supraspinal areas, bilateral microinjection of PACAP into the rat RVLM leads to sympathoexcitation and tachycardia (Farnham et al., 2012). Importantly, in both studies, the PAC1/VPAC2 receptor selective antagonist PACAP (6–38), delivered to the T5/6 spinal level, abolishes the sympathoexcitatory response elicited by PACAP (Farnham et al., 2008, 2012). PACAP-induced activation of presympathetic neurons of the RVLM leads to a subsequent release of PACAP at the spinal cord. Sympathetic outflow is dependent partially on the activity of a class of sympathetic efferent neurons that are under chemoreflex control. Activation of the peripheral chemoreflex by hypoxia elevates sympathetic efferent activity, through a direct connection between the NTS and RVLM barosensitive neurons (Guyenet, 2000). Thus, neurons localized in the RVLM may be important in the formation of sLTF (Xing and Pilowsky, 2010). As mentioned earlier, the intrinsic hypoxia-sensing catecholaminergic C1 neurons of the RVLM (Xing et al., 2014) are sympathoexcitatory in nature. More than 80% of spinally projecting C1 neurons of the RVLM express PACAP mRNA (Farnham et al., 2008). This finding suggests that PACAP plays a functional role following the activation of RVLM neurons. Following continuous and intermittent hypoxia (Lam et al., 2012) there is a heightened level of immunohistochemical staining for TH and elevated levels of PACAP, and PAC1 receptors, in the rat carotid body. Hypoxia-driven activation of carotid body neurons is sensed by peripheral chemoreceptors (Abdala et al., 2012; Paton et al., 2013; Zoccal et al., 2008), and results in the projection

of afferent signals terminating on the NTS. Subsequently, pathways emanate from the NTS that synapse onto glutamatergic neurons in the RVLM, where presympathetic neurons relay signals that terminate onto cardiovascular SPN in the IML column of the spinal cord. Although, a sub-threshold dose of PACAP-38 (100 ␮M) does not elicit sympathoexcitation on its own, the same dose of PACAP38 injected intermittently (10 injections interspersed with 5 min intervals; 10 ␮M per injection) into the RVLM unilaterally does cause sLTF (unpublished data; Fig. 2). This sympathoexcitatory effect may occur not only by activation of RVLM neurons, but also through a PACAP signaling cascade that may be an independent and necessary mechanism for eliciting long-term excitation of sympathetic neurons. Excitation of sympathetic neurons by the activation of PAC receptors at the spinal level may be crucial for regulating sympathetic discharge. Evidence indicates that targeting individual PACAP receptor subtypes (PAC/VPAC) results in the differential secretion of catecholamine from the adrenal medulla (Gaede et al., 2012; Inglott et al., 2012). Specifically, intrathecal stimulation of the PAC1 receptor elicits sympathoexcitatory responses and noradrenaline release from the adrenal medulla, resulting in a pressor response that lasts for at least an hour (Inglott et al., 2012). The binding of noradrenaline to beta-adrenergic receptors can also cause phosphorylation of transcription factor CREB, with phosphorylated-CREB measurements peaking 30–60 min (Tamotsu et al., 1995). PACAP-induced activation of G␣s - coupled receptors induces long-lasting plasticity through the phosphorylation of CREB (Baxter et al., 2011). Thus, there may be close relationships between PACAP-induced long-term sympathoexcitation, and phosphorylation of CREB mediated by PAC1 receptors secreting noradrenaline (Fig. 1). 3.2. Orexin Prepro-orexin mRNA knock-out mice have a lower resting mean arterial pressure (∼15 mmHg; MAP) compared to their wild-type littermates (Kayaba et al., 2003). This indicates that tonic release of orexin neuropeptides contribute to cardiovascular regulation. Several studies investigating the effects of orexin revealed a dosedependent depolarization in the RVLM (Chen et al., 2000; Huang et al., 2010; Machado et al., 2002; Shahid et al., 2012), and sympathetic preganglionic neurons (SPN) of the spinal cord (Antunes et al., 2001; Shahid et al., 2011). The sympathetic bulbospinal vasomotor neurons emerging from the RVLM provide the main supraspinal excitatory input for sympathetic vasomotor control (Guyenet, 2006). Orexin-A or -B intracisternal administration causes a sympathoexcitatory effect, and the excitation is primarily mediated by orexin receptor 2 (OX2 R), and to a lesser extent, by the orexin receptor 1 (OX1 R) (Huang et al., 2010). Perfusion of either orexin-A or -B at 100 nM to RVLM cells of neonatal rats causes ∼42% to be depolarized when measured using whole-cell patch-clamp. Application of an OX2 R antagonist (TCS OX2 29) at both 3 and 10 nM causes the number of neuronal depolarization to decrease to 25%, while co-application of both OX2 R and OX1 R antagonists completely ablates orexin-A induced depolarization (Huang et al., 2010). Machado et al. (2002) demonstrated that different concentrations of orexin-A bilateral microinjection (6.35, 12.7, 38.1 ␮M) to the RVLM of conscious rats did not affect the MAP in a dose-dependent manner. Orexin-A dose of 6.35 ␮M significantly increased MAP by 15 mmHg from the baseline level for 6.5 min (Machado et al., 2002). Supporting this finding, TH-ir neurons in the RVLM were frequently co-localized with orexin receptors 1 and 2, and closely apposed to orexin-A ir terminals (Shahid et al., 2012). Bilateral microinjection of orexinA at 50 pmol elicited blood pressure elevation of 42 mmHg, and

Please cite this article in press as: Kim, S.J., et al., Intermittent hypoxia-induced cardiorespiratory long-term facilitation: A new role for microglia. Respir. Physiol. Neurobiol. (2016), http://dx.doi.org/10.1016/j.resp.2016.03.012

G Model RESPNB-2633; No. of Pages 9

ARTICLE IN PRESS S.J. Kim et al. / Respiratory Physiology & Neurobiology xxx (2016) xxx–xxx

5

Fig. 2. (A) Raw trace depicting the splanchnic sympathetic nerve response to a single unilateral 100 ␮M (sub-threshold) microinjection of PACAP-38 in the rostral ventrolateral medulla (green). There was no sympathetic response compared to a single unilateral microinjection of PBS in the rostral ventrolateral medulla (grey). (B) Overlay of experimental traces including a control intrathecal injection of PBS followed by either sub-threshold and intermittent microinjections of PACAP-38 (green) or PBS (grey). Intermittent sub-threshold microinjections of PACAP-38 (10 ␮M each; 5 min apart) significantly elevated splanchnic sympathetic nerve activity compared to PBS control. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

an increase in sympathetic discharge of 45% from baseline activity (Shahid et al., 2012). Orexin-ir fibres are located in the IML column at the thoracolumbar level (T1-L3) of the spinal cord (Date et al., 2000; Van Den Pol, 1999). In vivo intrathecal orexin -A and -B infusions at the spinal T5/6 revealed a consistent increase in MAP and HR (Antunes et al., 2001; Shahid et al., 2011). Intrathecal infusion of orexin-A elicited sympathetically-mediated pressor responses and tachycardia (Shahid et al., 2011). Injections of orexin-A antiserum neutralized the pressor effects induced by orexin-A, while normal serum albumin injections did not generate changes (Antunes et al., 2001). Therefore, it is clear that the MAP response was elicited by orexin-A (Antunes et al., 2001). In vitro results showed that superfusion of orexin-A or orexin-B caused direct neuronal depolarization and/or discharge in the majority of SPN tested (Antunes et al., 2001). Further study is required to determine the role of orexin in the development of sLTF, or in the increased sensitivity of peripheral chemoreceptor response to hypoxia. Similar to the generation of AIH-induced pLTF, orexin neurons may play an important role in sLTF. There are several reasons for this: (1) there are descending orexinergic inputs to both the RVLM and the IML, (2) orexin is a sympathoexcitatory neurotransmitter with long-lasting effects, and (3) orexin neurons in the hypothalamus respond to intermittent hypoxia, possibly from the afferent inputs of the C1 neurons. Future studies should examine whether or not intermittent activation of orexin receptors lead to sLTF, and an enhancement of sensitivity of sympathetic neurons to hypoxia.

4. Microglia may regulate neuronal facilitation in the cardiorespiratory system 4.1. Microglia activation Microglia are the primary immune effector cells in the CNS. These cells exist in two phenotypically differentiated forms: the surveillance state in the normal brain and the ‘activated’ state specialised to operate within the diseased environment. Conversion between the two states is identifiable by the transformation in cell morphology from ramified to amoeboid form, and by the production of cytokines (Abcouwer, 2011). There are two ‘activated’ phenotypes: the ‘inflammatory’ M1 state and the ‘antiinflammatory’ M2 state (Gaikwad and Heneka, 2013). The M1 state is activated by various pro-inflammatory factors, such as lipopolysaccharides (LPS) and interferon-␥ (IFN-␥) (Orihuela et al., 2016). Microglia activate into the M1 state via the nuclear factor kappa B (NF␬B) pathway (Kobayashi et al., 2013), or by the mitogen-activated protein kinase (MAPK/ERK) pathway (Soliman et al., 2012). M1 microglia produces pro-inflammatory cytokines (PICs), such as tumour necrosis factor-␣ (TNF-␣) and interleukin1␤ (IL-1␤) (Cherry et al., 2014). On the other hand, microglia express the M2 phenotype when stimulated by anti-inflammatory factors. M2 microglia are neuroprotective: producing transforming growth factor-␤ (TGF-␤) to promote phagocytosis, neuronal repair, survival, and to suppress the immune response (Hu et al., 2015).

Please cite this article in press as: Kim, S.J., et al., Intermittent hypoxia-induced cardiorespiratory long-term facilitation: A new role for microglia. Respir. Physiol. Neurobiol. (2016), http://dx.doi.org/10.1016/j.resp.2016.03.012

G Model

ARTICLE IN PRESS

RESPNB-2633; No. of Pages 9

S.J. Kim et al. / Respiratory Physiology & Neurobiology xxx (2016) xxx–xxx

6

4.2. Microglia activation by hypoxia As stated above, microglia interact with neurons both under normal and pathological conditions. To date, however, there is no information on the effect of intermittent hypoxia on microglia. Instead, there have been indications that microglial cells exposed to bursts of acute hypoxia, or prolonged hypoxia cause activation of these cells. Wu et al. (2009) showed that Kv1.1 potassium channels are crucial for the release of PICs such as TNF-␣ and IL-1␤; neutralizing the channels using antibodies that suppress PIC release. Two hours of continuous hypoxia increased Kv1.1 protein expression in rats, and both LPS and hypoxia increased Kv1.1 expression and PIC release in cultured microglial BV-2 cells. As LPS is well known to stimulate microglial activation to the M1 phenotype (Orihuela et al., 2016), and M1 microglia are known to produce PICs, these results together suggest that hypoxia stimulates the transformation of microglia to the M1 phenotype. In addition, continuous hypoxia has also been shown to cause microglial activation to the M1 state in cortical brain cells after 3 h, cultured from 1 to 3 day old rat pups (Habib et al., 2014), and when maintained for 8 h in rodent microglial cell cultures (Guo and Bhat, 2006; Park et al., 2002). In the context of cardiorespiratory plasticity, rather than continuous hypoxia, an alternation between bursts of hypoxia and periods of resting state for consolidation are essential to achieving LTF (Poon and Siniaia, 2000). After this phase of intermittent hypoxia stimuli, a memory-like effect of LTF is expressed in the subsequent quiescent phase. Whether or not this concept of intermittent vs. continuous stimulation in producing the LTF in neuronal activity applies to how microglial cells respond is an intriguing area for future study. In vitro studies should confirm if intermittent application of hypoxia or chemical cues can promote microglial cells to transform into the M1 phenotype. Extending from this, cardiorespiratory parameters should be measured using in vivo set up to determine if the inhibition of microglia using agents such as minocycline or doxycycline (Bhandare et al., 2015), prior to intermittent hypoxia produces any changes to previous observations (Xing and Pilowsky, 2010). 4.3. Microglia and pLTF A recent study by Huxtable et al. (2015) suggested that the inflammatory effects of M1 microglia undermine the generation of pLTF. Treating rats with intraperitoneal LPS prior to AIH abolished the development of pLTF. This effect of LPS on pLTF was offset when the rats were pre-treated with intraperitoneal injection of ketoprofen (an anti-inflammatory agent). Ketoprofen is an M1 microglial antagonist (Asanuma et al., 2003), and readily crosses the bloodbrain barrier (BBB) (Netter et al., 1985). LPS does not cross the BBB (Singh and Jiang, 2004; Wang et al., 2014), but activates systemic inflammatory processes. This leads to the production of PICs which can cross the BBB to activate microglia to the M1 phenotype (Qin et al., 2007). Ketoprofen would directly inhibit both systemic PIC production and the M1-activated microglia. Pre-treatment with ketoprofen was shown to restore pLTF in the study above, suggesting that M1 microglia may oppose the formation of pLTF. 4.4. Microglia and the sympathetic pathway Microglial cells are capable of both increasing and decreasing SNA (Bhandare et al., 2015; Dworak et al., 2014). However, the underlying mechanisms remain largely unclear. Chronic SNA elevation is seen following myocardial infarction (MI) (Dworak et al., 2014; Lindley et al., 2004). This is accompanied by microglial activation (Dworak et al., 2014) at brain nuclei involved in sympathetic control (Bardgett et al., 2014; Pilowsky and Goodchild, 2002; Rana et al., 2014). The increased sympathetic outflow and

microglial activation in these nuclei is attenuated at 12wks post-MI in rats that were continuously treated with minocycline by intracerebroventricular infusion, beginning 1wk prior to MI, suggesting that M1 microglia elevate SNA by modulating neuronal activity (Dworak et al., 2014). Applying TNF-␣ or IL-1␤ to the paraventricular nucleus of the hypothalamus increases neuronal activity, leading to increased SNA (Shi et al., 2011). Both TNF-␣ and IL-1␤ are produced by M1 microglia. It is possible that M1 microglia increase SNA by secreting PICs that activate brain nuclei involved in SNA regulation. Another recent study showed that, in rats with elevated sSNA from kainic acid-induced seizures, antagonising microglial activity using minocycline at the IML of the spinal cord further elevated SNA (Bhandare et al., 2015). Contrary to the MI experiment, microglia activity appears to suppress the SNA to some degree during acute seizure. These contradictory findings suggest that microglia may play different roles under acute and chronic conditions, or perhaps there are unknown functions of microglia that need to be elucidated. 4.5. Dendritic pruning by microglia in regulating neuronal activity Apart from affecting neurons through the secretion of factors with transmitters, trophics, or neuroprotective properties, microglial cells regulate neuronal activity by supporting or stripping synapses or the dendritic spines (Kettenmann et al., 2013). These are postsynaptic structural components found on excitatory synapses of principal neurons. These structures contain the molecular machineries for synaptic transmission, and can directly dictate the efficacy of the neuronal network (Araya et al., 2014). All forms of microglia (Chang et al., 2015; Cherry et al., 2014; Tremblay et al., 2011; Wake et al., 2009) mediate synaptic pruning; spines have the ability to respond to plasticity by changing the number and size. Therefore, neuronal activity regulation may be largely dependent on microglial cells to either allow the retention, or the pruning of dendritic spines (Fig. 3). The processes of microglia are highly motile, consistent with their proposed role as vigilant surveyors of the integrity and function of the CNS. Microglial cells also play an important role in synaptic plasticity. They regulate neuronal activity (Ekdahl, 2012) as the ramification of processes make intimate, but transient connections with pre- and postsynaptic elements (Nimmerjahn et al., 2005; Wake et al., 2009), and dendritic spines (Chang et al., 2015). These interactions are consistent in the normal brain, but markedly prolonged under ischemic conditions (Wake et al., 2009). Inducing ischemia in the brain by occluding the middle cerebral artery caused the microglial-synaptic contact duration in the ischemic penumbra (the area <0.5 mm from the ischemic core) increases to 1hr; this was often followed by disappearance of the presynaptic bouton, and therefore the synapse (Wake et al., 2009). Functional state of the CNS, especially the pattern of neuronal activity, is likely to dictate the kinetics of interaction between microglia and synapses. Prolonged contact is observed under pathological conditions. As mentioned earlier, blocking microglia at the level of spinal cord in epileptic rats further enhances the sympathetic discharge (Bhandare et al., 2015). This indicates that microglial cells are necessary for controlling the neural activity, and it is likely that the cells promote neuroprotection under a disease state of the CNS. Therefore, we propose that microglia have the potential to modulate both phrenic and sympathetic neuronal activity during pathological hypoxemic conditions. Hypoxia may cause microglia to engage in inhibitory dendritic pruning, production of excitatory PICs, and other unknown mechanisms to influence synaptic function and plasticity. Evidence on non-immune microglial functions is relatively new, compared to the classic idea of microglia as immune CNS cells. Studying the effects of intermittent hypoxia on

Please cite this article in press as: Kim, S.J., et al., Intermittent hypoxia-induced cardiorespiratory long-term facilitation: A new role for microglia. Respir. Physiol. Neurobiol. (2016), http://dx.doi.org/10.1016/j.resp.2016.03.012

G Model RESPNB-2633; No. of Pages 9

ARTICLE IN PRESS S.J. Kim et al. / Respiratory Physiology & Neurobiology xxx (2016) xxx–xxx

7

Fig. 3. Microglia interact with neurons during normal physiology. Dendrites have spines where excitatory inputs are received from neighbouring neurons. During overexcitation of neurons, microglia can change their shape and chemical phenotype and move towards neurons to remove synaptic spines. This leads to a reduction in excitation so that neuronal membrane potential is maintained within an appropriate dynamic range. This process enables neurons to respond normally to excitatory and inhibitory inputs. The process is reversed when excitation is reduced, allowing dendritic spines and synapses to be restored. More than 40 classical neurotransmitter subtypes are known to be expressed on microglial membranes in addition to receptors for cytokines. Changes in the extracellular environment cause microglia to retract their fimbria, change shape, migrate along concentration gradients, and dramatically change their expression of proteins.

cardiorespiratory neuronal activity, and how microglia may regulate this process may increase our understanding of the mechanistic features of plasticity.

5. Conclusion Considerable advances have been made to understand the mechanisms leading to the generation of cardiorespiratory LTF following intermittent hypoxia. Microglia may make significant contributions to neuroplasticity; regulating the neurons by stripping dendritic spines, or releasing cytokines to control the activity of neurons. These processes are likely to occur during the survey of presynaptic and postsynaptic elements by the ramification of cytoplasmic extensions from microglia, of which the frequency increases under pathological conditions such as ischemia, epilepsy, and possibly during intermittent hypoxia-induced intermittent hypoxemia. In light of current understandings, neurotransmitters and other molecules are essential in the communication of neurons, but also in determining the phenotype of microglia. We propose that a three-way interplay between neurons, neurotransmitters, and microglia is necessary to evoke cardiorespiratory LTF

in its entirety. By targeting multiple molecular and cellular components, we may be able to promote facilitation in respiratory nerves to establish stable breathing during disease states, such as sleep apnoea. Conversely, prolonged enhancement in sympathetic activity, a pathogenic factor of neurogenic hypertension, may also be controlled.

Conflict of interest None.

Acknowledgements Work in the Authors laboratories was supported by the National Health and Medical Research Council of Australia Fellowship (PMP; 1024489), (Grants: 1065485, 1082215) and Australian Research Council (Discovery Early Career Researcher Award (MMJF; DE120100992), and the Heart Foundation of Australia (G 11S 5957). S.J.K. is supported by an Australian Postgraduate Award (APA SC0042) and a Heart Research Institute Scholarship.

Please cite this article in press as: Kim, S.J., et al., Intermittent hypoxia-induced cardiorespiratory long-term facilitation: A new role for microglia. Respir. Physiol. Neurobiol. (2016), http://dx.doi.org/10.1016/j.resp.2016.03.012

G Model RESPNB-2633; No. of Pages 9

ARTICLE IN PRESS S.J. Kim et al. / Respiratory Physiology & Neurobiology xxx (2016) xxx–xxx

8

References Abbott, S.B.G., Coates, M.B., Stornetta, R.L., Guyenet, P.G., 2013a. Optogenetic stimulation of C1 and retrotrapezoid nucleus neurons causes sleep state-dependent cardiorespiratory stimulation and arousal in rats. Hypertension 61, 835–841. Abbott, S.B.G., DePuy, S.D., Nguyen, T., Coates, M.B., Stornetta, R.L., Guyenet, P.G., 2013b. Selective optogenetic activation of rostral ventrolateral medullary catecholaminergic neurons produces cardiorespiratory stimulation in conscious mice. J. Neurosci. 33, 3164–3177. Abcouwer, S.F., 2011. Neural inflammation and the microglial response in diabetic retinopathy. J. Ocul. Pharmacol. Ther. 4, 25–33. Abdala, A.P., McBryde, F.D., Marina, N., Hendy, E.B., Engelman, Z.J., Fudim, M., Sobotka, P.A., Gourine, A.V., Paton, J.F.R., 2012. Hypertension is critically dependent on the carotid body input in the spontaneously hypertensive rat. J. Physiol. 590, 4269–4277. Antunes, V.R., Cristina Brailoiu, G., Kwok, E.H., Scruggs, P., Dun, A.N.J., 2001. Orexins/hypocretins excite rat sympathetic preganglionic neurons in vivo and in vitro. Am. J. Physiol. 281, R1801–R1807. Araya, R., Vogels, T.P., Yuste, R., 2014. Activity-dependent dendritic spine neck changes are correlated with synaptic strength. Proc. Natl. Acad. Sci. U. S. A. 111, E2895–E2904. Asanuma, M., Tsuji, T., Miyazaki, I., Miyoshi, K., Ogawa, N., 2003. Methamphetamine-induced neurotoxicity in mouse brain is attenuated by ketoprofen, a non-steroidal anti-inflammatory drug. Neurosci. Lett. 352, 13–16. Bach, K.B., Mitchell, G.S., 1996. Hypoxia-induced long-term facilitation of respiratory activity is serotonin dependent. Respir. Physiol. 104, 251–260. Baker, T.L., Mitchell, G.S., 2000. Episodic but not continuous hypoxia elicits long-term facilitation of phrenic motor output in rats. J. Physiol. 529, 215–219. Baker-Herman, T.L., Fuller, D.D., Bavis, R.W., Zabka, A.G., Golder, F.J., Doperalski, N.J., Johnson, R.A., Watters, J.J., Mitchell, G.S., 2004. BDNF is necessary and sufficient for spinal respiratory plasticity following intermittent hypoxia. Nat. Neurosci 7, 48–55. Bardgett, M.E., Chen, Q.H., Guo, Q., Calderon, A.S., Andrade, M.A., Toney, G.M., 2014. Coping with dehydration: sympathetic activation and regulation of glutamatergic transmission in the hypothalamic PVN. Am. J. Physiol. 306, R804–R813. Baxter, P.S., Martel, M.A., McMahon, A., Kind, P.C., Hardingham, G.E., 2011. Pituitary adenylate cyclase-activating peptide induces long-lasting neuroprotection through the induction of activity-dependent signaling via the cyclic AMP response element-binding protein-regulated transcription co-activator 1. J. Neurochem. 118, 365–378. Berkowitz, R.G., Sun, Q.J., Goodchild, A.K., Pilowsky, P.M., 2005. Serotonin inputs to laryngeal constrictor motoneurons in the rat. Laryngoscope 115, 105–109. Bhandare, A.M., Mohammed, S., Pilowsky, P.M., Farnham, M.M.J., 2015. Antagonism of PACAP or microglia function worsens the cardiovascular consequences of kainic-acid-induced seizures in rats. J. Neurosci. 35, 2191–2199. Biber, K., Neumann, H., Inoue, K., Boddeke, H.W.G.M., 2007. Neuronal ‘On’ and ‘Off’ signals control microglia. Trends Neurosci. 30, 596–602. Bochorishvili, G., Nguyen, T., Coates, M.B., Viar, K.E., Stornetta, R.L., Guyenet, P.G., 2014. The orexinergic neurons receive synaptic input from C1 cells in rats. J. Comp. Neurol. 522, 3834–3846. Braga, V.A., Soriano, R.N., Machado, B.H., 2006. Sympathoexcitatory response to peripheral chemoreflex activation is enhanced in juvenile rats exposed to chronic intermittent hypoxia. Exp. Physiol. 91, 1025–1031. Brown, R.E., Sergeeva, O., Eriksson, K.S., Haas, H.L., 2001. Orexin A excites serotonergic neurons in the dorsal raphe nucleus of the rat. Neuropharmacology 40, 457–459. Chang, P.K.Y., Khatchadourian, A., McKinney, R.A., Maysinger, D., 2015. Docosahexaenoic acid (DHA): a modulator of microglia activity and dendritic spine morphology. J. Neuroinflammation 12. Chen, C.T., Hwang, L.L., Chang, J.K., Dun, N.J., 2000. Pressor effects of orexins injected intracisternally and to rostral ventrolateral medulla of anesthetized rats. Am. J. Physiol. 278, R692–R697. Cherry, J.D., Olschowka, J.A., O’Banion, M.K., 2014. Are resting microglia more M2? Front. Immunol. 5. Cutler, M.J., Swift, N.M., Keller, D.M., Wasmund, W.L., Burk, J.R., Smith, M.L., 2004a. Periods of intermittent hypoxic apnea can alter chemoreflex control of sympathetic nerve activity in humans. Am. J. Physiol. 287, H2054–H2060. Cutler, M.J., Swift, N.M., Keller, D.M., Wasmund, W.L., Smith, M.L., 2004b. Hypoxia-mediated prolonged elevation of sympathetic nerve activity after periods of intermittent hypoxic apnea. J. Appl. Physiol. 96, 754–761. Dale-Nagle, E.A., Hoffman, M.S., MacFarlane, P.M., Mitchell, G.S., 2010. Multiple pathways to long-lasting phrenic motor facilitation. Adv. Exp. Med. Biol., 225–230. Date, Y., Mondal, M.S., Matsukura, S., Nakazato, M., 2000. Distribution of orexin-A and orexin-B (hypocretins) in the rat spinal cord. Neurosci. Lett. 288, 87–90. Devinney, M.J., Huxtable, A.G., Nichols, N.L., Mitchell, G.S., 2013. Hypoxia-induced phrenic long-term facilitation: emergent properties. Ann. N. Y. Acad. Sci. 1279, 143–153. Devinney, M.J., Fields, D.P., Huxtable, A.G., Peterson, T.J., Dale, E.A., Mitchell, G.S., 2015. Phrenic long-term facilitation requires PKC␪ activity within phrenic motor neurons. J. Neurosci. 35, 8107–8117. Dick, T.E., Hsieh, Y.H., Morrison, S., Coles, S.K., Prabhakar, N., 2004. Entrainment pattern between sympathetic and phrenic nerve activities in the

Sprague-Dawley rat: hypoxia-evoked sympathetic activity during expiration. Am. J. Physiol. 286, R1121–R1128. Dick, T.E., Hsieh, Y.H., Wang, N., Prabhakar, N., 2007. Acute intermittent hypoxia increases both phrenic and sympathetic nerve activities in the rat. Exp. Physiol. 92, 87–97. Dick, T.E., Hsieh, Y.-H., Dhingra, R.R., Baekey, D.M., Galán, R.F., Wehrwein, E., Morris, K.F., 2014. Chapter 10—Cardiorespiratory coupling: common rhythms in cardiac, sympathetic, and respiratory activities, in: Gert Holstege, C.M.B., Hari, H.S. (Eds.), Prog. Brain Res., Elsevier, 191–205. Dickson, L., Finlayson, K., 2009. VPAC and PAC receptors: from ligands to function. Pharmacol. Ther. 121, 294–316. Dougherty, B.J., Fields, D.P., Mitchell, G.S., 2015. Mammalian target of rapamycin is required for phrenic long-term facilitation following severe but not moderate acute intermittent hypoxia. J. Neurophysiol. 114, 1784–1791. Dworak, M., Stebbing, M., Kompa, A.R., Rana, I., Krum, H., Badoer, E., 2014. Attenuation of microglial and neuronal activation in the brain by ICV minocycline following myocardial infarction. Auton. Neurosci. 185, 43–50. Ekdahl, C.T., 2012. Microglial activation-tuning and pruning adult neurogenesis. Front. Pharmacol. 3, 41. Farnham, M.M.J., Pilowsky, P.M., 2010. The role of PACAP in central cardiorespiratory regulation. Respir. Physiol. Neurobiol. 174, 65–75. Farnham, M.M.J., Li, Q., Goodchild, A.K., Pilowsky, P.M., 2008. PACAP is expressed in sympathoexcitatory bulbospinal C1 neurons of the brain stem and increases sympathetic nerve activity in vivo. Am. J. Physiol. 294, R1304–R1311. Farnham, M.M.J., Lung, M.S.Y., Tallapragada, V.J., Pilowsky, P.M., 2012. PACAP causes PAC 1/VPAC 2 receptor mediated hypertension and sympathoexcitation in normal and hypertensive rats. Am. J. Physiol. 303, H910–H917. Fields, D.P., Springborn, S.R., Mitchell, G.S., 2015. Spinal 5-HT7 receptors induce phrenic motor facilitation via EPAC-mTORC1 signaling. J. Neurophysiol. 114, 2015–2022. Fuller, D.D., Zabka, A.G., Baker, T.L., Mitchell, G.S., 2001. Phrenic long-term facilitation requires 5-HT receptor activation during but not following episodic hypoxia. J. Appl. Physiol. 90, 2001–2006 (discussion 2000). Gaede, A.H., Inglott, M.A., Farnham, M.M.J., Pilowsky, P.M., 2012. Catestatin has an unexpected effect on the intrathecal actions of PACAP dramatically reducing blood pressure. Am. J. Physiol. 303, R719–R726. Gaikwad, S.M., Heneka, M.T., 2013. Studying M1 and M2 states in adult microglia. Methods Mol. Cell. Biol., 185–197. Gatti, P.J., Llewellyn-Smith, I.J., Sun, Q.J., Chalmers, J., Pilowsky, P., 1999. Substance P-immunoreactive boutons closely appose inspiratory protruder hypoglossal motoneurons in the cat. Brain Res. 834, 155–159. Golder, F.J., Ranganathan, L., Satriotomo, I., Hoffman, M., Lovett-Barr, M.R., Watters, J.J., Baker-Herman, T.L., Mitchell, G.S., 2008. Spinal adenosine A2a receptor activation elicits long-lasting phrenic motor facilitation. J. Neurosci. 28, 2033–2042. Gourine, A.V., Llaudet, E., Thomas, T., Dale, N., Spyer, K.M., 2002. Adenosine release in nucleus tractus solitarii does not appear to mediate hypoxia-induced respiratory depression in rats. J. Physiol. 544, 161–170. Greenberg, H.E., Sica, A., Batson, D., Scharf, S.M., 1999. Chronic intermittent hypoxia increases sympathetic responsiveness to hypoxia and hypercapnia. J. Appl. Physiol. 86, 298–305. Guo, G., Bhat, N.R., 2006. Hypoxia/reoxygenation differentially modulates NF-␬B activation and iNOS expression in astrocytes and microglia. Antioxid. Redox Signal. 8, 911–918. Guyenet, P.G., 2000. Neural structures that mediate sympathoexcitation during hypoxia. Respir. Physiol. 121, 147–162. Guyenet, P.G., 2006. The sympathetic control of blood pressure. Nat. Rev. Neurosci. 7, 335–346. Habib, P., Slowik, A., Zendedel, A., Johann, S., Dang, J., Beyer, C., 2014. Regulation of hypoxia-induced inflammatory responses and M1-M2 phenotype switch of primary rat microglia by sex steroids. J. Mol. Neurosci. 52, 277–285. Hickner, S., Hussain, N., Angoa-Perez, M., Francescutti, D.M., Kuhn, D.M., Mateika, J.H., 2014. Ventilatory long-term facilitation is evident after initial and repeated exposure to intermittent hypoxia in mice genetically depleted of brain serotonin. J. Appl. Physiol. 116, 240–250. Hoffman, M.S., Golder, F.J., Mahamed, S., Mitchell, G.S., 2010. Spinal adenosine A2A receptor inhibition enhances phrenic long term facilitation following acute intermittent hypoxia. J. Physiol. 588, 255–266. Hoffman, M.S., Nichols, N.L., Macfarlane, P.M., Mitchell, G.S., 2012. Phrenic long-term facilitation after acute intermittent hypoxia requires spinal ERK activation but not TrkB synthesis. J. Appl. Physiol. 113, 1184–1193. Hu, X., Leak, R.K., Shi, Y., Suenaga, J., Gao, Y., Zheng, P., Chen, J., 2015. Microglial and macrophage polarization—new prospects for brain repair. Nat. Rev. Neurol. 11, 56–64. Huang, S.C., Dai, Y.W.E., Lee, Y.H., Chiou, L.C., Hwang, L.L., 2010. Orexins depolarize rostral ventrolateral medulla neurons and increase arterial pressure and heart rate in rats mainly via orexin 2 receptors. J. Pharmacol. Exp. Ther. 334, 522–529. Huxtable, A.G., MacFarlane, P.M., Vinit, S., Nichols, N.L., Dale, E.A., Mitchell, G.S., 2014. Adrenergic ␣1 receptor activation is sufficient, but not necessary for phrenic long-term facilitation. J. Appl. Physiol. 116, 1345–1352. Huxtable, A.G., Smith, S.M.C., Peterson, T.J., Watters, J.J., Mitchell, G.S., 2015. Intermittent hypoxia-induced spinal inflammation impairs respiratory motor plasticity by a spinal p38 MAP kinase- dependent mechanism. J. Neurosci. 35, 6871–6880.

Please cite this article in press as: Kim, S.J., et al., Intermittent hypoxia-induced cardiorespiratory long-term facilitation: A new role for microglia. Respir. Physiol. Neurobiol. (2016), http://dx.doi.org/10.1016/j.resp.2016.03.012

G Model RESPNB-2633; No. of Pages 9

ARTICLE IN PRESS S.J. Kim et al. / Respiratory Physiology & Neurobiology xxx (2016) xxx–xxx

Imadojemu, V.A., Mawji, Z., Kunselman, A., Gray, K.S., Hogeman, C.S., Leuenberger, U.A., 2007. Sympathetic chemoreflex responses in obstructive sleep apnea and effects of continuous positive airway pressure therapy. Chest 131, 1406–1413. Inglott, M.A., Lerner, E.A., Pilowsky, P.M., Farnham, M.M.J., 2012. Activation of PAC1 and VPAC receptor subtypes elicits differential physiological responses from sympathetic preganglionic neurons in the anaesthetized rat. Br. J. Pharmacol. 167, 1089–1098. Kayaba, Y., Nakamura, A., Kasuya, Y., Ohuchi, T., Yanagisawa, M., Komuro, I., Fukuda, Y., Kuwaki, T., 2003. Attenuated defense response and low basal blood pressure in orexin knockout mice. Am. J. Physiol. 285, R581–R593. Kettenmann, H., Kirchhoff, F., Verkhratsky, A., 2013. Microglia: new roles for the synaptic stripper. Neuron 77, 10–18. Kobayashi, K., Imagama, S., Ohgomori, T., Hirano, K., Uchimura, K., Sakamoto, K., Hirakawa, A., Takeuchi, H., Suzumura, A., Ishiguro, N., Kadomatsu, K., 2013. Minocycline selectively inhibits M1 polarization of microglia. Cell Death Dis. 4. Lai, H.L., Yang, T.H., Messing, R.O., Ching, Y.H., Lin, S.C., Chern, Y., 1997. Protein kinase C inhibits adenylyl cyclase type VI activity during desensitization of the A2a-adenosine receptor-mediated cAMP response. J. Biol. Chem. 272, 4970–4977. Lam, S.Y., Liu, Y., Liong, E.C., Tipoe, G.L., Fung, M.L., 2012. Upregulation of pituitary adenylate cyclase activating polypeptide and its receptor expression in the rat carotid body in chronic and intermittent hypoxia. Adv. Exp. Med. Biol. 758, 301–306. Leuenberger, U.A., Brubaker, D., Quraishi, S., Hogeman, C.S., Imadojemu, V.A., Gray, K.S., 2005. Effects of intermittent hypoxia on sympathetic activity and blood pressure in humans. Auton. Neurosci. 121, 87–93. Lindley, T.E., Doobay, M.F., Sharma, R.V., Davisson, R.L., 2004. Superoxide is involved in the central nervous system activation and sympathoexcitation of myocardial infarction-induced heart failure. Circ. Res. 94, 402–409. Liu, R.J., Van Den Pol, A.N., Aghajanian, G.K., 2002. Hypocretins (orexins) regulate serotonin neurons in the dorsal raphe nucleus by excitatory direct and inhibitory indirect actions. J. Neurosci. 22, 9453–9464. Lonze, B.E., Ginty, D.D., 2002. Function and regulation of CREB family transcription factors in the nervous system. Neuron 35, 605–623. Machado, B.H., Bonagamba, L.G.H., Dun, S.L., Kwok, E.H., Dun, N.J., 2002. Pressor response to microinjection of orexin/hypocretin into rostral ventrolateral medulla of awake rats. Regul. Pept. 104, 75–81. Mitchell, G.S., Baker, T.L., Nanda, S.A., Fuller, D.D., Zabka, A.G., Hodgeman, B.A., Bavis, R.W., Mack, K.J., Olson Jr, E.B., 2001. Physiological and genomic consequences of intermittent hypoxia: invited review: intermittent hypoxia and respiratory plasticity. J. Appl. Physiol. 90, 2466–2475. Murphy, S.M., Pilowsky, P.M., Sun, Q.J., Llewellyn-Smith, I.J., 1995. Thyrotropin-releasing hormone-immunoreactive varicosities synapse on rat phrenic motoneurons. J. Comp. Neurol. 359, 310–322. Nakamura, A., Zhang, W., Yanagisawa, M., Fukuda, Y., Kuwaki, T., 2007. Vigilance state-dependent attenuation of hypercapnic chemoreflex and exaggerated sleep apnea in orexin knockout mice. J. Appl. Physiol. 102, 241–248. Navarrete-Opazo, A., Mitchell, G.S., 2014. Therapeutic potential of intermittent hypoxia: a matter of dose. Am. J. Physiol. 307, R1181–R1197. Netter, P., Lapicque, F., Bannwarth, B., Tamisier, J.N., Thomas, P., Royer, R.J., 1985. Diffusion of intramuscular ketoprofen into the cerebrospinal fluid. Eur. J. Clin. Pharmacol. 29, 319–321. Neverova, N.V., Saywell, S.A., Nashold, L.J., Mitchell, G.S., Feldman, J.L., 2007. Episodic stimulation of ␣1-adrenoreceptors induces protein kinase C-dependent persistent changes in motoneuronal excitability. J. Neurosci. 27, 4435–4442. Nichols, N.L., Dale, E.A., Mitchell, G.S., 2012. Severe acute intermittent hypoxia elicits phrenic long-term facilitation by a novel adenosine-dependent mechanism. J. Appl. Physiol. 112, 1678–1688. Nimmerjahn, A., Kirchhoff, F., Helmchen, F., 2005. Neuroscience: resting microglial cells are highly dynamic surveillants of brain parenchyma in vivo. Science 308, 1314–1318. Orihuela, R., McPherson, C.A., Harry, G.J., 2016. Microglial M1/M2 polarization and metabolic states. Br. J. Pharmacol. 173, 649–665. Park, S.Y., Lee, H., Hur, J., Kim, S.Y., Kim, H., Park, J.H., Cha, S., Kang, S.S., Cho, G.J., Choi, W.S., Suk, K., 2002. Hypoxia induces nitric oxide production in mouse microglia via p38 mitogen-activated protein kinase pathway. Mol. Brain Res. 107, 9–16. Paton, J.F.R., Sobotka, P.A., Fudim, M., Engelman, Z.J., Hart, E.C.J., McBryde, F.D., Abdala, A.P., Marina, N., Gourine, A.V., Lobo, M., Patel, N., Burchell, A., Ratcliffe, L., Nightingale, A., 2013. The carotid body as a therapeutic target for the treatment of sympathetically mediated diseases. Hypertension 61, 5–13. Paxinos, G., Watson, C., 2006. The Rat Brain in Stereotaxic Coordinates, 3rd ed. Academic Press Inc, San Diego. Pilowsky, P.M., Goodchild, A.K., 2002. Baroreceptor reflex pathways and neurotransmitters: 10 years on. J. Hypertens. 20, 1675–1688. Pilowsky, P.M., de Castro, D., Llewellyn-Smith, I., Lipski, J., Voss, M.D., 1990. Serotonin immunoreactive boutons make synapses with feline phrenic motoneurons. J. Neurosci. 10, 1091–1098. Pocock, J.M., Kettenmann, H., 2007. Neurotransmitter receptors on microglia. Trends Neurosci. 30, 527–535. Poon, C.-S., Siniaia, M.S., 2000. Plasticity of cardiorespiratory neural processing: classification and computational functions. Respir. Physiol. 122, 83–109.

9

Prabhakar, N.R., Peng, Y.J., Jacono, F.J., Kumar, G.K., Dick, T.E., 2005. Cardiovascular alterations by chronic intermittent hypoxia: importance of carotid body chemoreflexes. Clin. Exp. Pharmacol. Physiol. 32, 447–449. Qin, L., Wu, X., Block, M.L., Liu, Y., Breese, G.R., Hong, J.S., Knapp, D.J., Crews, F.T., 2007. Systemic LPS causes chronic neuroinflammation and progressive neurodegeneration. Glia 55, 453–462. Rana, I., Badoer, E., Alahmadi, E., Leo, C.H., Woodman, O.L., Stebbing, M.J., 2014. Microglia are selectively activated in endocrine and cardiovascular control centres in streptozotocin-induced diabetic rats. J. Neuroendocrinol. 26, 413–425. Shahid, I.Z., Rahman, A.A., Pilowsky, P.M., 2011. Intrathecal orexin A increases sympathetic outflow and respiratory drive, enhances baroreflex sensitivity and blocks the somato-sympathetic reflex. Br. J. Pharmacol. 162, 961–973. Shahid, I.Z., Rahman, A.A., Pilowsky, P.M., 2012. Orexin A in rat rostral ventrolateral medulla is pressor, sympatho-excitatory, increases barosensitivity and attenuates the somato-sympathetic reflex. Br. J. Pharmacol. 165, 2292–2303. Shi, Z., Gan, X.B., Fan, Z.D., Zhang, F., Zhou, Y.B., Gao, X.Y., De, W., Zhu, G.Q., 2011. Inflammatory cytokines in paraventricular nucleus modulate sympathetic activity and cardiac sympathetic afferent reflex in rats. Acta Physiol. 203, 289–297. Silva, A.J., Kogan, J.H., Frankland, P.W., Kida, S., 1998. CREB and memory. Ann. Rev. Neurosci., 127–148. Singh, A.K., Jiang, Y., 2004. How does peripheral lipopolysaccharide induce gene expression in the brain of rats? Toxicology 201, 197–207. Soliman, M.L., Puig, K.L., Combs, C.K., Rosenberger, T.A., 2012. Acetate reduces microglia inflammatory signaling in vitro. J. Neurochem. 123, 555–567. Sun, M.K., Reis, D.J., 1994. Central neural mechanisms mediating excitation of sympathetic neurons by hypoxia. Prog. Neurobiol. 44, 197–219. Sun, Q.J., Pilowsky, P., Llewellyn-Smith, I.J., 1995. Thyrotropin-releasing hormone inputs are preferentially directed towards respiratory motoneurons in rat nucleus ambiguus. J. Comp. Neurol. 362, 320–330. Sun, Q.J., Berkowitz, R.G., Goodchild, A.K., Pilowsky, P.M., 2002. Serotonin inputs to inspiratory laryngeal motoneurons in the rat. J. Comp. Neurol. 451, 91–98. Sun, Q.J., Berkowitz, R.G., Goodchild, A.K., Pilowsky, P.M., 2003. Substance P inputs to laryngeal motoneurons in the rat. Respir. Physiol. Neurobiol. 137, 11–18. Tamotsu, S., Schomerus, C., Stehle, J.H., Roseboom, P.H., Korf, H.W., 1995. Norepinephrine-induced phosphorylation of the transcription factor creb in isolated rat pinealocytes—an immunocytochemical study. Cell Tissue Res. 282, 219–226. Terada, J., Nakamura, A., Zhang, W., Yanagisawa, M., Kuriyama, T., Fukuda, Y., Kuwaki, T., 2008. Ventilatory long-term facilitation in mice can be observed during both sleep and wake periods and depends on orexin. J. Appl. Physiol. 104, 499–507. Toyama, S., Sakurai, T., Tatsumi, K., Kuwaki, T., 2009. Attenuated phrenic long-term facilitation in orexin neuron-ablated mice. Respir. Physiol. Neurobiol. 168, 295–302. Tremblay, M.E., Stevens, B., Sierra, A., Wake, H., Bessis, A., Nimmerjahn, A., 2011. The role of microglia in the healthy brain. J. Neurosci. 31, 16064–16069. Van Den Pol, A.N., 1999. Hypothalamic hypocretin (orexin): robust innervation of the spinal cord. J. Neurosci. 19, 3171–3182. Wake, H., Moorhouse, A.J., Jinno, S., Kohsaka, S., Nabekura, J., 2009. Resting microglia directly monitor the functional state of synapses in vivo and determine the fate of ischemic terminals. J. Neurosci. 29, 3974–3980. Wang, P., You, S.W., Yang, Y.J., Wei, X.Y., Wang, Y.Z., Wang, X., Hao, D.J., Kuang, F., Shang, L.X., 2014. Systemic injection of low-dose lipopolysaccharide fails to break down the blood-brain barrier or activate the TLR4-MyD88 pathway in neonatal rat brain. Int. J. Mol. Sci. 15, 10101–10115. Wu, C.Y., Kaur, C., Sivakumar, V., Lu, J., Ling, E.A., 2009. Kv1.1 expression in microglia regulates production and release of proinflammatory cytokines, endothelins and nitric oxide. Neuroscience 158, 1500–1508. Xing, T., Pilowsky, P.M., 2010. Acute intermittent hypoxia in rat in vivo elicits a robust increase in tonic sympathetic nerve activity that is independent of respiratory drive. J. Physiol. 588, 3075–3088. Xing, T., Fong, A.Y., Bautista, T.G., Pilowsky, P.M., 2013. Acute intermittent hypoxia induced neural plasticity in respiratory motor control. Clin. Exp. Pharmacol. Physiol. 40, 602–609. Xing, T., Pilowsky, P.M., Fong, A.Y., 2014. Mechanism of sympathetic activation and blood pressure elevation in humans and animals following acute intermittent hypoxia. Prog. Brain Res. 209, 131–146. Yamaguchi, K., Futatsuki, T., Ushikai, J., Kuroki, C., Minami, T., Kakihana, Y., Kuwaki, T., 2015. Intermittent but not sustained hypoxia activates orexin-containing neurons in mice. Respir. Physiol. Neurobiol. 206, 11–14. Zimmermann, G., Taussig, R., 1996. Protein kinase C alters the responsiveness of adenylyl cyclases to G protein ␣ and ␤␥ subunits. J. Biol. Chem. 271, 27161–27166. Zoccal, D.B., Bonagamba, L.G.H., Oliveira, F.R.T., Antunes-Rodrigues, J., Machado, B.H., 2007. Increased sympathetic activity in rats submitted to chronic intermittent hypoxia. Exp. Physiol. 92, 79–85. Zoccal, D.B., Simms, A.E., Bonagamba, L.G.H., Braga, V.A., Pickering, A.E., Paton, J.F.R., Machado, B.H., 2008. Increased sympathetic outflow in juvenile rats submitted to chronic intermittent hypoxia correlates with enhanced expiratory activity. J. Physiol. 586, 3253–3265.

Please cite this article in press as: Kim, S.J., et al., Intermittent hypoxia-induced cardiorespiratory long-term facilitation: A new role for microglia. Respir. Physiol. Neurobiol. (2016), http://dx.doi.org/10.1016/j.resp.2016.03.012