Age and gender effects on serotonin-dependent plasticity in respiratory motor control

Abstract

We review recent evidence indicating that serotonin-dependent plasticity in respiratory control is influenced by age and gender. Specifically, respiratory long-term-facilitation following intermittent hypoxia decreases with age in male rats, but increases in female rats. We speculate about a possible relationship between age and gender effects on serotonin-dependent plasticity in upper airway motoneuron pools and the prevalence of obstructive sleep apnea.

Introduction

Little is known about the effects of age on respiratory control. Nonetheless, there is growing evidence that aging causes structural, neurochemical and functional changes in the respiratory system (Schlenker and Goldman, 1985, Fukuda, 1991, Fukuda, 1992). The control of breathing is also influenced by gender-specific events such as changes in the estrus cycle, pregnancy, and following menopause (Regensteiner et al., 1989, Tatsumi et al., 1995, Tatsumi et al., 1997). However, there is little evidence concerning the neural mechanisms whereby age and/or gender influence respiratory control. Within the respiratory neural control system, an obvious link with gender and aging arises from the effects of sex hormones on neuromodulatory systems, in particular serotonin. In this review, we provide a general overview of the role played by serotonin in the control of breathing, and how the serotonergic system alters with aging. We will then discuss how gonadal hormones influence serotonin. With this background, the effects of age and gender will be considered inasmuch as they influence a model of serotonin-dependent plasticity in respiratory motor control, respiratory long-term-facilitation (LTF) following intermittent hypoxia. We conclude by speculating about a possible link between age and gender effects on LTF and the prevalence of sleep disordered breathing.

Serotonin (5HT) plays a pivotal role in the control of breathing (for reviews see Bonham, 1995, Bianchi et al., 1995, McCrimmon et al., 1995). Brain and spinal cord regions involved in respiratory control receive 5HT input primarily from the medullary raphe nuclei including nucleus raphe magnus, pallidus, and obscurus (Holtman, 1988, Li et al., 1993, Manaker and Tischler, 1993). Caudal raphe neurons are activated during hypoxia, and presumably release 5HT in the vicinity of respiratory premotor and motoneurons (Erickson and Millhorn, 1994, Teppema et al., 1997, Kinkead et al., 2001). Serotonin has an excitatory effect on upper airway and phrenic motoneurons (Berger et al., 1992, Lindsay and Feldman, 1993, Arita et al., 1995, Hilaire et al., 1997, Di Pasquale et al., 1997). Several 5HT receptor subtypes may be involved. Using the reverse transcriptase–polymerase chain reaction and highly selective molecular probes, Okabe et al. detected mRNA for five 5HT receptor subtypes in the hypoglossal nucleus including 1B, 2A, 2C, 3 and 7 (Okabe et al., 1997). Spinal motoneurons have been shown immunocytochemically to express 1A, 1B, 2A and 2C subtypes (Rekling et al., 2000), and phrenic nuclei express transcripts for 2A and 2C (Basura et al., 2001). However, serotonergic facilitation of excitatory drive in respiratory motoneurons is thought to be largely mediated by 5HT₂ receptors (Kubin et al., 1992, Lindsay and Feldman, 1993).

Changes in serotonergic modulation of respiratory motoneurons associated with early development and with aging can be viewed as two ends of a continuum. Breathing must adapt to changes that occur throughout the life of an animal including growth and development, exercise, obesity, pregnancy, age, injury and disease. For many years, the adult respiratory control system was thought to be immutable, but there is a growing body of evidence to suggest that, like many other neural systems, respiratory control exhibits plasticity (McCrimmon et al., 1995, Powell et al., 1998, Mitchell et al., 2001). One model of 5HT-dependent plasticity in respiratory motor control is LTF. LTF is a long lasting increase in respiratory motor output following either electrical stimulation of chemoafferent neurons (Millhorn et al., 1980a, Millhorn et al., 1980b, Hayashi et al., 1993, Fregosi and Mitchell, 1994) or intermittent hypoxia (Bach and Mitchell, 1996, Kinkead and Mitchell, 1999, Fuller et al., 2000, Mitchell et al., 2001). Three-five-minute episodes of hypoxia can elicit LTF, whereas continuous hypoxia of a similar duration is ineffective (Baker and Mitchell, 2000). LTF requires the activation of 5HT receptors during episodic hypoxia (Fuller et al., 2001a), and it can be abolished by pretreatment with 5HT receptor antagonists (methysergide, ketanserin), a 5HT depleter (p-chloro-phenylalanine) or a 5HT neurotoxin (5-7, dihydroxytryptamine) (Millhorn et al., 1980a, Millhorn et al., 1980b, Fregosi and Mitchell, 1994, Bach and Mitchell, 1996, Kinkead and Mitchell, 1999). The specific 5HT receptor involved appears to be of the 5HT₂ family; although not yet firmly established, the weight of evidence is in favor of 5HT_2A receptors (Kinkead and Mitchell, 1999, Fuller et al., 2001a, Mitchell et al., 2001), since ketanserin has a 30–100-fold greater affinity for 5HT_2A over 5HT_2C receptors (Hoyer et al., 1994, Barnes and Sharp, 1999), and 5HT_2A (vs. _2C) receptors predominate on phrenic motoneurons (Basura et al., 2001). We have used LTF as a model to study age- and gender-specific changes in a form of serotonin-dependent plasticity in respiratory motor control, as LTF can itself exhibit plasticity (Kinkead et al., 1998, Mitchell et al., 2001, Ling et al., 2001).

There are numerous reports of changes in the serotonergic system with increasing age. While the number of serotonergic neurons in raphe nuclei does not appear to decrease with age (van Luijtelaar et al., 1992), there are reports of alterations (increase or decrease, depending on the brain region) in other components of the system including 5HT and 5HIAA concentration (the major metabolite of 5HT), terminal density, receptor density and binding characteristics, and the serotonin reuptake transport protein (Shih and Young, 1978, Brunello et al., 1985, Moretti et al., 1987, Halpern et al., 1989, Goudsmit et al., 1990, van Luijtelaar et al., 1992, Bigham and Lidow, 1995). In contrast, in some brain regions, the serotonergic system appears to be unaffected by age (Ponzio et al., 1982, Herrera et al., 1991, Bigham and Lidow, 1995, Palego et al., 1997, Nyakas et al., 1997). Clearly, age-related changes are specific to each neuroanatomical region, and this heterogeneity may help explain apparently conflicting data in different systems.

Overall, little attention has been paid to age-associated changes in serotonergic modulation of respiratory motor control. Age-related decreases in srotonergic innervation have been described in the hypoglossal nucleus of old rats relative to young rats (Behan and Brownfield, 1999), and preliminary data show an increase in 5HT-immunoreactivity in the hypoglossal nucleus in middle-aged female rats (Behan, unpublished observations). Ko et al. (1997) have reported an age-associated decrease in the serotonergic innervation of spinal segments near the phrenic motor nucleus. Aging may differentially affect serotonergic modulation of phrenic and hypoglossal motoneurons, despite their common serotonergic input (Manaker et al., 1992). In neonatal rats, for example, hypoglossal and spinal motoneurons respond differently to 5HT, probably due to differences in 5HT receptor expression in these two motoneuron populations (Takahashi and Berger, 1990, Berger et al., 1992, Talley et al., 1997).

The serotonergic system differs with gender. In all brain regions thus far examined, levels of the 5HT precursor tryptophan, 5HT and 5HIAA/5HT ratios are higher in females than males (Carlsson and Carlsson, 1988). There are also gender differences in the expression of serotonin receptors in the rat brain (Zhang et al., 1999). Many studies in rats and primates are complicated by the influence of the estrus cycle, and also by hormone manipulation. Nonetheless, it is generally agreed that there is greater 5HT activity in females than in males (for review, see Rubinow et al., 1998). In females, serotonin levels vary throughout the estrus cycle, presumably in response to changes in hormone levels. The rat has a 4–5 day estrus cycle in which four distinct phases can be recognized by changes in the vaginal epithelium: proestrus (∼12 h), estrus (∼12 h), metestrus (∼21 h) and diestrus (∼57 h) (Hebel and Stromberg, 1986). Circulating levels of estrogen and progesterone vary throughout the cycle, with high progesterone and progressively increasing estrogen during diestrus and high estrogen and progesterone levels during proestrus (Freeman, 1994). Using in vivo microdialysis, lowest levels of 5HT were measured in the hypothalamus of female rats in estrus and highest levels in diestrus (Gundlah et al., 1998).

The influence of estrogen on the serotonergic system is variable and complex. Estrogen increases tryptophan hydroxylase (a key enzyme in the synthesis of 5HT) in the dorsal raphe of macaque monkeys (Pecins-Thompson et al., 1996), although there is little effect in rats (Alves et al., 1998). Estrogen also affects 5-HT reuptake, although primates and rats again respond differently (Bethea et al., 1998, Grasse et al., 1993). Confounding factors in these studies include the time-course of hormonal effects in rats and monkeys, time between ovariectomy and intervention, the duration of hormone treatment, and region-specific effects (Rubinow et al., 1998). Estrogen treatment has been shown to influence select 5HT receptors, notably 5HT_1A and 5HT_2A receptors (Bethea et al., 1998, Sumner et al., 1999). These estrogen effects are rapid. For example, 5HT_2A levels increase in association with the estrogen surge between diestrus and proestrus in rats, a period of hours (Sumner and Fink, 1997). Studies in humans also show that estrogen influences the serotonergic system in a gender-specific manner, including 5HT synthesis and 5HT_2A receptor binding (Ellenbogen et al., 1996, Blum et al., 1996, Biver et al., 1996, Nishizawa et al., 1997, Rubinow et al., 1998, Moses et al., 2000).

The effect of estrogen on target cells in the brain is mediated primarily by two forms of estrogen receptor (ER), the classical α form and the more recently discovered β form (Pfaff and Keiner, 1973, Simerly et al., 1990, Kuiper et al., 1996). When activated, ERs translocate to the nucleus where they can modulate the expression of genes encoding a wide variety of proteins including neurotransmitters and their receptors (Alves et al., 1998). Whereas there are no data on ERs in phrenic or hypoglossal nuclei, in situ hybridization studies have shown that the medullary raphe nuclei, particularly nucleus raphe pallidus and obscurus contain both ERα and ERβ receptors (Shughrue et al., 1997). Estrogen can also affect other elements of the serotonergic system. In the primate dorsal raphe nucleus, estrogen has been shown to increase the expression of tryptophan hydroxylase (Bethea et al., 1998). In the rat, estrogen caused a 2–3-fold increase in 5HT_2A receptor mRNA and a 50% increase in serotonin reuptake transporter mRNA expressing cells in the dorsal raphe nucleus (McQueen et al., 1997, Sumner and Fink, 1997). Even in the absence of detectable intracellular ERs, neurons can still respond to estrogen, possibly by novel membrane receptors, but also by interaction with other second messenger systems (Toran-Allerand et al., 1999, Singh et al., 1999). The neuroprotective effects of estrogen in the brain have received considerable attention in recent years, partially due to the beneficial effects of estrogen replacement therapy in the treatment of certain age-related neurodegenerative diseases (Behl and Holsboer, 1999, McEwen and Alves, 1999). Estrogen and/or progesterone, or perhaps a critical ratio of these hormones may be the key factor in the control of respiration via effects on neuromodulatory systems such as 5HT.

Changes in estrogen levels are normally accompanied by changes in levels of circulating progesterone. It has been known for some time that high circulating levels of progesterone can stimulate ventilation (Skatrud et al., 1978, Strohl et al., 1981, White et al., 1983, Regensteiner et al., 1989, Bayliss and Milhorn, 1992, Tatsumi et al., 1995). Furthermore, in women, a combination of progesterone and estrogen is more effective in increasing the hypoxic ventilatory response than progesterone alone (Tatsumi et al., 1995). Like estrogen, progesterone can bind to intracellular receptors and influence gene expression. However, it is not clear how progesterone exerts its effects on respiration. Progesterone receptors are found in a number of brain regions, and progesterone receptor mRNA (but not protein) has been identified in the rat hypoglossal nucleus (Kastrup et al., 1999). Progesterone receptors are also present in the dorsal raphe, although estrogen is necessary for their expression (Bethea et al., 1998).

Testosterone may also play a role in age and gender effects on the serotonergic system. Castration in adult rats reduces 5HT synthesis in dorsal raphe neurons (Long et al., 1983), and results in region-specific and serotonin receptor-specific changes that can be reversed with testosterone replacement (Zhang et al., 1999). Androgen receptors have been identified in the hypoglossal nucleus in both male and female rats (Yu and McGinnis, 2000). In the brain, testosterone is converted to dihydrotestosterone by 5α reductase, and to estrogen by an aromatase (Celotti et al., 1991). While testosterone can act through its specific steroid receptor, its effects on 5HT_2A receptors appear to depend primarily on conversion to estrogen (Fink et al., 1998). As with estrogen, plasma testosterone levels decrease significantly with age in rats and humans (Goudsmit et al., 1990, Mitchell et al., 1995, Ferrini and Barrett-Connor, 1998). The relative lack of response to testosterone in older animals may be due to regulation of testosterone receptors, or to alterations in aromatase activity in the brain (Balthazart and Ball, 1998).