How the Vertebrate Brain Regulates Behavior

by Donald Pfaff

  Members of the G-protein coupled receptor superfamily of membrane proteins, α1B-ARs are located on the cell surface. α1B-ARs can couple to a variety of G-proteins and second-messenger systems, but it is generally agreed that the primary pathway involves activation Gq/11, eliciting second messenger signals, including Ca2+, diacylglycerol, and inositol triphosphate.

  Ion channels. Norepinephrine uses both L-type calcium channels and reduced conductance of potassium channels (A currents) to signal through Gq proteins, and thus to activate phospholipase C and protein kinase C. In contrast, glutamate, acting through NMDA (N-methyl-D-aspartate) receptors coupled directly to ion channels, triggers an inward sodium current. Acetylcholine, acting through muscarinic receptors in the VMH, will excite VMH neurons through the inward flow of sodium ions. As will be noted here, histamine, acting through H1 receptors, works through the inhibition of a potassium leak current to depolarize VMH neurons. All these ion channel routes contribute to the VMH facilitation of lordosis.

  Calcium channels, especially the L-type channels, are targets of modulation by norepinephrine and α1-adrenergic agonists, but the majority of these studies have been carried out in cardiac myocytes. In VMH, both L- and N-type calcium channels contribute to the phenylephrine (PHE) response, although N-type calcium channels predominate. PHE also increases low-voltage-activated currents in the majority of VMH neurons. These results are important, as an increase in calcium entry can be expected to modulate second messenger systems and increase the release of neurotransmitters as well as modulate spike frequency by facilitating action potential firing. Although these effects may be important, in the VMH PHE-mediated membrane excitability is not solely dependent on an increase in intracellular calcium (Lee et al. 2008). The effect of adrenergic stimulation on VMN neurons is probably also due to effects on K+ conductances.

  It follows that the α-adrenergic agonist PHE depolarizes VMH neurons, in part by reducing membrane conductance for K+ via an A-type K+ channel (Lee et al. 2008). As shown by Anne Etgen, this pathway is potentiated by estradiol in two ways: by increasing levels of mRNA encoding the α1B-adrenergic receptor and by increasing protein kinase C. These results show a role for the noradrenergic system in modulating potassium channels, and this may provide a powerful way to increase the excitability of hormone-dependent VMH neurons that govern female sexual behavior.

  Thus, the intracellular signal transduction pathway for mediating the facilitation of lordosis can be outlined as follows: norepinephrine binds to α1B-adrenoceptor causing conformational change → activates α-subunit of Gαq/11 → stimulates phosphatidyl-inositol-specific phospholipase C-β1 in the plasma membrane → hydrolyzes membrane-bound phosphatidylinositol 4,5-biphosphate (PIP2) → generates second messengers inositol triphosphate and diacylglycerol → mobilizes intracellular Ca2+ and activates protein kinase C, respectively → phosphorylates a variety of cellular substrates to achieve the final functional results (Kow and Pfaff, 1998). As noted previously, we also have gathered some information about the ion channels involved.

  With respect to lordosis, Lee-Ming Kow showed clearly that α1-adrenergic agents act on the VMH to increase female reproductive behavior (Kow, Weesner, and Pfaff 1992). Lee microinjected into the VMH α1-agonists, methoxamine (MA) and phenylephrine (PhE), and various control agents. In ovariectomized rats treated with estrogen, infusion of MA, PhE, or a β-agonist isoproterenol into the lateral ventricle, or bilateral infusions of MA or PhE into the VMH significantly facilitated lordosis. Conversely, intra-VMH infusion of the α1-antagonist prazosin inhibited lordosis. The phenomenon was specific: intra-VMH infusion of isoproterenol or an α2-agonist clonidine had no effect. Neither was the intra-VMH infusion of MA effective if 1) the rats were not primed with estrogen, 2) the tips of the cannulae were outside the VMH, or 3) it was preceded by an intra-VMH infusion of the α1B-antagonist, chloroethyl clonidine. In parallel experiments with microelectrodes, he confirmed that, to go along with the lordosis-facilitating effect of α1-activation, the actions of MA and PhE on the electrical activity of single neurons are to increase firing rates.

  Our behavioral work extended the work of the late Robert Moss at Southwestern Medical School in Dallas. He showed that administering either adrenalin or noradrenaline into the VMH increased lordosis, an effect prevented by receptor blockers.

  Because (a) estrogens increase synthetic activity in adrenergic systems in the brain, and (b) activation of adrenergic receptors at the top of the lordosis circuit in VMH increase lordosis; it follows that (c) one way in which estrogens increase lordosis is through heightened adrenergic synapse activation in the VMH.

  Histamine

  Histamine (HA) effects on VMH neurons (to promote lordosis) work differently. HA is well known to play an important role in electrophysiological actions of HA mediated by H1 that are predominantly excitatory, signaling through Gq linked to phospholipase C and then protein kinase C.

  Ion channels. In the mouse VMH, HA-induced depolarization was mediated by H1, but not H2, receptors and was not affected by the blockade of sodium or calcium channels, but was abolished by potassium channel blockade (Zhou et al. 2007). Further analyses indicated that HA depolarization was due to an inhibition of a potassium leakage current (Zhou et al. 2007).

  Both estrogens (Pfaff 1999, 2005) and HA are known to increase arousal (see Chapter 8). Estrogens potentiate the depolarizing actions in VMH due to HA administration by inhibiting outward K+ currents (probably the delayed rectifier) that would lead to depolarization.

  Muscarinic Cholinergic Receptor

  As reviewed by the authoritative Anne Etgen, estrogen treatment of ovariectomized rats increases the activity of choline acetyltransferase, the rate-limiting enzyme in acetylcholine synthesis. Altogether, the estrogen effect can be detected at the mRNA, protein, and enzymatic activity levels of analysis.

  Release of acetylcholine in the hypothalamus is increased by estrogens as well. As far as muscarinic receptors are concerned, Kathie Olsen, working with the neuroendocrine pioneer Richard Whalen, demonstrated the estrogen-dependent enhancement of muscarinic agonist 3H-quinuclidinyl benzilate ([3H]QNB) binding in the preoptic area as compared with ovariectomized, estrogen-free controls. A significant sex difference was found in the ability of estrogen to induce [3H]QNB—estrogen was ineffective in altering [3H]QNB binding in either brain region of castrated males. This muscarinic binding goes along with our electrophysiology: Lee-Ming Kow and I used extracellular single-neuron recording in brain tissue slices through the VMH to show that estrogen treatment greatly increased the number of neurons responding to acetylcholine, and that effect depended on muscarinic receptors (Kow and Pfaff 1995; Kow et al. 1995).

  That is, we analyzed the effects of muscarinic agents on the single-neuron activity of VMH neurons recorded in brain tissue slices of estrogen-primed female rats. All the agonists tested, including acetylcholine (ACh), oxotremorine-M (OM), carbachol (CCh), and McN-A-343 (McN), evoked primarily excitation (80–100 percent), some inhibition (0–20 percent), or occasional biphasic responses (0–8 percent). By comparing the response magnitude and the effectiveness in evoking a response, the rank order for evoking excitation, the primary response, was found to be OM > CCh > ACh approximately McN, which is consistent with that (OM > CCh > McN) for facilitating lordosis reported by others. The consistency and frequency of occurrence suggest that the excitatory electric action of the muscarinic agonists is related to their facilitatory behavioral effect.

  Experiments with antagonists selective for M1 (pirenzepine), M2 (AF-DX 116 [otenzepad]), and M3 (4-DAMP [[3H]4-diphenylacetoxy-N-methyl-piperidine methiodide] and p-F-HHSiD) indicate that muscarinic excitations are mediated by M1 and / or M3, but not M2. Because M1 receptors have been shown to be neither sufficient nor necessary to mediate the muscarinic facilitation, the M3 receptor may be crucially involved in this behavioral effect. Autoradiographic assays of binding to [3H]4-DAMP with or without pirenzepine and AF-DX 116 also indicate the presence of M3 receptors in th
e VMH. Altogether, the electrophysiology links VMH muscarinic synaptic transmission to lordosis.

  With respect to lordosis behavior, Gary Dohanich and his colleagues, as well as other laboratories, demonstrated facilitation of lordosis by several receptor agonists that mimic acetylcholine: carbachol, bethanechol, eserine, oxotremorine, pilocarpine, and acetylcholine itself. We extended Dohanich’s work by showing that the antagonist N-methyl scopolamine inhibited lordosis (Kaufman, McEwen, and Pfaff 1988). In the matched the literature on other muscarinic receptor antagonists, atropine, hemicholinium, and scopolamine all blocked lordosis. Further, we used tritiated N-methyl scopolamine to show that when we put the muscarinic blocker into the VMH it did not diffuse beyond the VMH.

  Thus, in the VMH, stimulation through muscarinic cholinergic receptors is sufficient and necessary to facilitate lordosis. Given that (a) estrogens increase cholinergic excitation to VMH neurons, and (b) such cholinergic activity fosters lordosis, it follows that (c) one way in which estrogens facilitate lordosis is through ramping up synthetic activity to produce more VMH excitation through cholinergic receptors.

  Gonadotropin-Releasing Hormone and Gonadotropin-Releasing Hormone Receptor

  Although we often think of neurotransmitters as achieving point-to-point neuronal signaling, neuropeptides, composed of several amino acids, are usually thought of in a more inclusive manner, comprising “systems” with particular, individual neurobiological themes. We dealt effectively with three of them, the most famous of which, gonadotropin-releasing hormone (GnRH), regulates all aspects of reproductive physiology.

  Gene expression for the ultimately integrative neuropeptide GnRH was crucial to our thinking about reproductive physiology. Joel Rothfeld entered the laboratory as one of the best distance runners we had ever seen and with a determination to learn in situ hybridization. His results showed that estrogen treatment raised the GnRH levels by an average of 63 percent (Rothfeld et al. 1989).

  By the mid-1990s we had been doing in situ hybridization in the brain for more than 10 years and wanted to put the technique to good use on a subject we expected to be difficult: the receptor for GnRH (Quiñones-Jenab, Jenab, et al. 1996). The GnRH system is crucial in regulating the reproductive system of female vertebrates. Thus it was compelling to analyze the estrogenic regulation of the GnRH receptor mRNA at the cellular level in female rats.

  Northern blot analysis detected three species (5.0, 4.5, and 1.4 kilobases) of GnRH receptor mRNA in pituitary tissues. The GnRH receptor mRNA levels of all three of these three species were increased by estrogen. Via in situ hybridization we observed a 3.5 times increase in GnRH receptor mRNA levels after 48 hours of estrogen treatment when compared with ovariectomized rats (12 hours of estrogen treatment did not change the GnRH receptor mRNA levels). Similar increases in GnRH receptor mRNA levels by estrogen were also found in female rat pituitary tissue. In situ hybridization analysis identified clusters of anterior pituitary cells that expressed the GnRH receptor mRNA. The estradiol effect depends on increased mRNA levels in these clusters, and a significant increase in the number of pituitary cells that expressed GnRH receptor was observed after 48 hours of estrogen treatment.

  GnRH neurons themselves do not express substantial levels of ER mRNA (Shivers et al. 1983). Some of the best information available tells us that the effect of estrogen on GnRH expression is mediated by ER-dependent GABA (γ-aminobutyric acid) neurons nearby.

  Several laboratories have confirmed and extended Joel Rothfeld’s results. One group concluded that enhanced GnRH receptor mRNA expression observed on the day of proestrus is largely due to the actions of estrogen, and other groups have agreed. Physiological concentrations of estradiol increase the steady-state levels of GnRH receptor mRNA in a dose-dependent manner.

  I note that estrogen-induced elevations of the ligand GnRH and of the mRNA for its receptor could produce a multiplicative hormone effect.

  What about lordosis? I got the idea for a direct GnRH effect on the behavior about 11:30 on a Wednesday night and had ordered the animals by 9:00 the next morning. The effect was that, clearly, GnRH injection elevated lordosis behavior responses (Pfaff, 1973), a phenomenon quickly replicated by the authoritative neuroendocrinologist Robert Moss in Dallas (for more detail, see Chapter 5).

  The conclusion: (a) estrogens increase GnRH indirectly and GnRH receptor transcription directly, and (b) GnRH increases lordosis behavior; it follows that (c) one way in which estrogens increase lordosis behavior is by signaling through the GnRH system.

  Oxytocin and Oxytocin Receptor

  Sookja Kim Chung came to the laboratory after receiving her Ph.D. working with Rochelle Cohen at the University of Chicago, after a successful college career in South Korea that included participation on the national volleyball team. She was strong, as indicated by her work with oxytocin (OT) in addition to her ultrastructural work. She knew that oxytocinergic signaling was important throughout the reproductive axis (see Chapter 5) and that OT-containing synapses could be found in the VMH. In fact, after OT administration during in vitro electrophysiological recordings, it was clear that OT increased the firing of action potentials by VMH neurons (Kow et al. 1991).

  As a result of these data, Sookja decided to use in situ hybridization to study OT gene expression in preoptic and hypothalamic neurons as well as to determine estrogenic influences on OT expression (Chung, McCabe, and Pfaff 1991). The hybridization findings were confirmed by immunocytochemistry. Using these techniques, we documented OT expression in the medial preoptic area, the nucleus of the anterior commissure, the periventricular neurons, (importantly) the paraventricular and supraoptic nuclei, the perifornical nuclei, as well as the bed nucleus of the stria terminalis. In a new set of experiments, estrogen-free control ovariectomized rats were compared with short-term (2 days) or long-term (2 months) estrogen-treated rats. The most striking estrogen effects came from the preoptic area neurons, in which either short- or long-term estrogen treatment caused an approximately 75 percent increase in the number of grains per cell (i.e., OT mRNA levels per cell). Additional analyses used frequency distributions of the number of grains per cell. In both the supraoptic nucleus and in the nucleus of anterior commissure such graphs revealed subpopulations of neurons in the estrogen-treated groups (compared with controls) that “pushed the curves to the right”—revealed subpopulations with particularly high concentrations of estrogen-dependent OT mRNA expression.

  To pursue the OT system further, we analyzed the potential estrogen effects on the OT receptor mRNA levels (Quiñones-Jenab et al. 1997) in some areas integral to the limbic-hypothalamic system discovered (as described in Chapter 1), namely, the VMH, posterior medial nucleus of amygdala, arcuate nucleus, caudate putamen, CA1 region of the hippocampus, anterior pituitary, and uterine tissue of ovariectomized female rats. Via in situ hybridization we observed a 4.4 times increase in OT receptor mRNA levels in the VMH after 48 hours of estrogen treatment when compared with ovariectomized rats. No other place in the brain had such a large estrogen effect.

  Later I claim the potential multiplicative actions of estrogens acting at both the levels of ligand transcription and corresponding receptor transcription. Here, the estrogenic effects on OT expression could be multiplied in their effectiveness by estrogenic effects on OT receptor expression.

  Because pioneers such as Charles Pedersen at the University of North Carolina had shown OT administration to the hypothalamus to be strongly facilitatory and in some cases essential for lordosis, it remained for us to replicate and extend his work. Michael Schumacher at Rockefeller (now an INSERM leader in Paris), for example, specified the relations of progesterone effects and OT effects on female reproductive behavior (Schumacher et al. 1990).

  So we thought that OT receptor activation would be required for high levels of lordosis behavior. To prove this, we used infusions of antisense oligodeoxynucleotides directed against OT receptor mRNA, microinjected directly into the VMH (McCarthy et al. 1994). Control infusions consisted
of a scrambled-sequence oligo that had little or no homology to known mRNAs. The OT receptor antisense oligo infusion significantly reduced lordosis frequency and intensity in females primed with estrogen. There were also a significantly greater number of male-rejection behaviors exhibited by antisense-oligo-infused estrogen-treated females versus controls. These behavioral results are supported by our electrophysiological findings.

  Lee-Ming Kow recorded single-neuron activity from the lordosis-relevant VMH in hypothalamic slices to characterize the electrophysiological actions of OT (Kow et al. 1991). To examine the effects of ovarian steroids on OT actions, we used brain slices prepared from ovariectomized rats either treated with estrogen or not, and some slices were treated with progesterone in vitro. OT affected the activity of a large number of VMH—and of those neurons affected, 94 percent responded with excitation.

  This predominant stimulatory action of OT is consistent with its lordosis-facilitating effect, because increases in the activity of VMH neurons are generally associated with the facilitation of lordosis. Pharmacological analyses with selective OT agonists and antagonists as well as structurally related peptides showed that the excitatory action of OT is, as expected, mediated by OT receptors. Estradiol modulated several aspects of OT transmission. First, it increased neuronal responsiveness to OT, especially at the lowest concentration used (0.2 nM). In addition, it caused neuronal responses to OT to correlate significantly with responses to acetylcholine and norepinephrine, which also can act on the ventromedial hypothalamus to facilitate lordosis. In summary, electrophysiology reinforces our applied genomic results.