by Donald Pfaff
With respect to puberty, the Ojeda group has found increasing methylation of the promoters of two crucial genes, Eed and Cbx7, to be associated with decreased transcription as classically expected. On the other hand, trimethylation of histone 3 lysine 4 (H3K4me3), associated with transcriptional activation, and trimethylation of histone 3 lysine 27 (H3K27me3), associated with transcriptional repression, have been implicated at different stages of puberty. As Ojeda states, because we are dealing with repressors, repressors of repressors, and transcriptional activating epigenomics eventually on puberty-activating genes, things are going to get confusing. Perhaps nothing other than such complexity should be expected at this exquisitely regulated period of reproductive development in the female.
Past puberty, in early adulthood, social behaviors change. To begin with, animals of the types used in neuroscience laboratories live in burrows to protect themselves from many types of predators. In particular for the female of the species, the reasons for leaving the burrow would be restricted to foraging (after dark) and finding a mate (about sundown). At that time the female begins to enter what I will describe in a later chapter as a “hormone-dependent behavioral funnel”—a series of communicative and locomotor behaviors that bring the reproductively competent female and the reproductively competent male to the same place at the same time, thus to mate and prolong the species.
The female attracts the male by spreading estrogen-dependent odor cues on the ground. She also exhibits a form of estrogen-dependent locomotor behavior that features extremely fast forward movement followed by a sudden stop. The sudden stop encourages the male, following from the rear, to bump into the female in the proper position for successful mounting (thus to provide the stimuli for lordosis behavior). At that time the mechanisms for the complete behavior (spelled out in Chapters 1 to 4) come into play.
From the beginning to the end of this chain of behaviors in time, transcriptional mechanisms depend on ER-α. For example, as previously mentioned, Thierry Spiteri, Anders Ågmo, and I validated that our small hairpin siRNA (shRNA) actually reduced the number of preoptic area neurons expressing ER-α by about 83 percent (Spiteri et al. 2010, 2012). Females with such few ER-α-expressing cells, in the test for sexual incentive motivation, approached the castrated male as opposed to the normal male incentive more than controls. Moreover, the data from the running wheels showed that females with few preoptic ER-α-expressing neurons failed to show enhanced activity after treatment with estrogen, a type of hormonal treatment that should have increased their total activity and specific courtship behaviors directed toward the male. In a related study, after an 80 percent reduction of the number of neurons expressing ER-α in the VMH, sexual incentive motivation was absent, even after treatment with estradiol and progesterone. Proceptivity and receptivity were also much reduced, while the number of rejections was enhanced. Suppression of the ER-α in the medial amygdala lacked these effects, showing the neuroanatomical specificity of our results. Likewise, the inactive control adeno-associated viral vector failed to modify any behavior. Overall, these results showed that expression from the ER-α gene is crucial for the entire sequence of behavioral events, from the processes leading to the establishment of sexual contact until the accomplishment of copulatory behaviors.
What about maternal behaviors? During the mother’s pregnancy, one of her first duties will be to build an elaborate nest, which has at least three functions: to help keep the newborn babies warm, to keep them out of the sight of predators, and to muffle the ultrasound vocalizations they make. After they are born, they may wiggle out of the nest. A hallmark of laboratory rodent maternal behavior is that the mother will retrieve them to the nest promptly by gently carrying them in her mouth. There she will lick them and groom them in a manner that keeps them warm. As well she will assume a hunched posture over them so that they are not smothered but are kept protected and can nurse.
Some of the mechanisms for these behaviors have been worked out in detail, such as the studies by Michael Numan at Boston College and Robert Bridges at Tufts University. Neurons essential for normal maternal behavior are found in a limited portion of the medial preoptic area, just in front of the hypothalamus. The axons descending from these neurons, axons whose signaling is crucial, travel through the medial forebrain neurons to the midbrain. Descending from the midbrain, the trail gets lost. However, relations between the preoptic area and the limbic forebrain cell groups have been better delineated. Perhaps the most important aspect of this connection is that the performance of maternal behaviors in these small animals must be protected from the disruptive effects of fear and stress.
Early in her career, neuroscientist Margaret McCarthy, now head of Pharmacology at the University of Maryland School of Medicine, put together the literature on the neuropeptide oxytocin, and offered with the hypothesis that an important theme of the results indicated that oxytocin helps to protect biologically adaptive behaviors from interruption by fear and stress (McCarthy et al. 1991). Her hypothesis fit the data on maternal behavior very well.
Rockefeller University graduate student Susan Fahrbach used intracerebroventricular (ICV) infusions of oxytocin, previously studied by Curt Pedersen at the University of North Carolina, to induce the performance of prompt maternal behavior by inexperienced estrogen-treated, ovariectomized virgin rats (Fahrbach, Morrell, and Pfaff 1986). Tests for the effect of ICV oxytocin in maternal behavior latency had included transfer of animals from their home cage to a larger (likely fear-inducing) test cage shortly before oxytocin infusion. If the transfer was too scary, the oxytocin did not work. Fahrbach evaluated the importance of this test feature on peptide-induced short-latency maternal behavior by varying the duration of the pretest cage habituation. The responses of ovariectomized, estrogen-primed rats housed in the test cages 1 week, 2 hours, or 0 hours before oxytocin or saline infusion were compared. It was found that only the rats given two or more hours of pretest cage habituation responded to ICV oxytocin treatment with short-latency maternal behavior. That is, oxytocin could, as conceived by McCarthy, overcome a certain amount of fear-inducing unfamiliarity, but a sudden environmental change just before behavioral assay was too much to overcome.
Most exciting was the opportunity to manipulate gene expression in a neuroanatomically specific way and then measure maternal behavior. Ana Ribeiro, then a postdoctoral researcher working with Sonoko Ogawa in my laboratory, validated the use of an adeno-associated viral vector encoding a small interfering RNA that should block ER expression to achieve more than an 80 percent decrease in ER-α by the neurons among which it is microinjected (Ribeiro et al. 2012). Applying this technique to the medial preoptic area, whose neurons bind estrogens and whose activity is crucial for maternal behavior, she was able to entirely negate maternal behavior. That meant that the females with extremely reduced ER-α expression specifically in the medial preoptic area did not retrieve the baby mice or lick and nurse them, whereas females with the control adeno-associated viral vector applied did these maternal behaviors very well. Remarkably, ER-α suppression did not affect another aspect of maternal behavior, namely, aggression toward a male intruder. Thus, our results established that a specific gene in a specific group of neurons is required for a crucially important natural behavior.
Recapitulating, the normal timing of a chain of natural behaviors starts with a generalized arousal (see Chapter 8) that leads to the type of increased locomotion necessary for courtship behaviors. Following those, sex behavior (lordosis) can occur, resulting naturally in the need for maternal behaviors. All these behaviors have three properties in that they are 1) increased by estrogens, 2) reduced or even eliminated by neuroanatomically specific knockdown of ER-α gene expression, and 3) temporally coordinated with preparations for reproduction (ovulation) by organs outside the brain.
Principles, Universal among Vertebrates, Derived from the Research
Some years ago I put together a formulation of principles of certain instinctive behavioral regulatio
ns (Pfaff, Phillips, and Rubin 2004). Six of them are relevant to the data and logic presented in this book.
1. One hormone can have many effects, and a single hormone can affect entire chains of behaviors. From the phenomena reviewed in Chapters 1 through 4 and their behavioral sequelae, the most obvious example is given by the effects of estrogens on courtship behaviors, sexual behaviors and maternal behaviors. In this chain, the later-named behaviors will not occur if the earlier-named behaviors were not.
As well, other mammalian behavioral systems can illustrate the same principle. Consider stress responses. The neuropeptide CRH not only causes the release of ACTH from the anterior pituitary, but also potentiates anxiety-related behaviors, and affects certain forms of fear learning as well as the autonomic nervous system.
Drinking / thirst provides another example. Angiotensin increases water intake by drinking and induces salt appetite. It has also been associated with memory, perhaps through hippocampal spatial memory mechanisms, and with other cognitive functions.
2. Combinations of different hormones can act together to influence an individual behavior. The converse of the principle named in point 1 above is now addressed in point 2. The behavioral phenomena I laid out in the introduction provides a clear example. That is, in the case of reproductive behavior, a strong example is the ability of progesterone, administered 2 to 4 hours ahead of a behavioral assay to massively amplify the facilitating effects of estrogen treatment (started 24 hours or more before the behavioral assay). Maternal behaviors of laboratory animals offer an even more complicated example. They depend on an orchestrated array of estrogen, progesterone, prolactin, glucocorticoid, and oxytocin.
Hunger provides another example. Long-term malnutrition associated with protein loss brings into play insulin, growth hormone, glucocorticoids, aldosterone, and thyroid hormone for proper eating needed to restore body mass. For short-term regulation of hunger, we need leptin, cholecystokinin, bombesins, glucagon, neuromedin, ghrelin, neuropeptide Y, and somatostatin all to play their roles.
3. Duration of hormone exposure can make a big difference. The simple example is that estrogen treatment must be started 24 hours or more in the ovariectomized female laboratory animal for lordosis behavior to be performed at high levels (see the introduction and Chapter 1). This principle is honored also in two inverse examples. First, Bruce Parsons, a Rockefeller graduate student, showed that long-term (but not short-term) absence of estrogen leads to a reduction in estrogen sensitivity (probably because of decreased effectiveness of ER coactivators). Second, short-term progesterone treatment enhances the effect of estrogen on lordosis, but long term progesterone treatment actually reduces lordosis.
Knowing that estrogens work over long time courses, it is somewhat surprising to read that certain neuropeptides work in exactly the opposite manner. For example, GnRH must be administered in a pulsatile manner to be effective (Chapter 5). Long-term administration causes exactly the opposite effect: suppression of its normal consequences. Growth hormone also normally is released in pulses, and must be, for optimal effect. The gut neuropeptide ghrelin likewise increases in concentration in a pulsatile manner, and operates on food intake that way.
4. Effects of a given hormone can be widespread across the body, central effects consonant with peripheral effects forming coordinated, unified mechanisms. Estrogens and progestins affect the ovaries, uterus, breast tissue, anterior pituitary gland, and brain, all in a manner that provides coordinated mechanisms of reproduction. At the beginning of this chapter I recounted briefly the story of some of the organs involved. Regulation of food intake provides another excellent example. At a minimum, the tissues involved include adipose tissue, stomach, intestines, liver, pancreas, and anterior pituitary and brain tissue. Within the brain, relevant neurons are located as far posterior as nucleus of the tractus solitarius and area postrema, and as far anterior as the hypothalamus and the cerebral cortex itself.
Oxytocin not only regulates parturition through actions on the uterus but also causes the “milk letdown reflex” in breast tissue. In brain tissue it is well known for its effects on reproductive behavior (Chapter 6), but it also facilitates a variety of prosocial behaviors including maternal behavior and even can increase pain thresholds.
5. Hormones can act at all levels of the neuraxis to exert behavioral effects, and the nature of the behavioral effect depends on the site of action. In the service of reproductive behaviors, as covered in Chapter 2 and briefly mentioned previously, going from posterior to anterior in the neuraxis, estrogen effects include those on the pudendal nerve’s distribution of cutaneous receptors, dorsal horn of the spinal cord, the midbrain central grey, the hypothalamus, and the preoptic area. Thyroid hormones provide an excellent example of this principle. As discovered by Björn Vennström at the Karolinska Institutet, virtually all neurons express thyroid hormone receptor α1. Another example is angiotensin, which works throughout the brain to regulate water drinking, blood pressure, heart rate, sympathetic nerve activation, and sodium appetite and can affect cognitive functions. Its sites of action range from the lower brainstem to the forebrain.
Progestins also have obvious effects on mood, acting through midbrain reticular neurons. They depress the activity of neurons associated with arousal on the entire forebrain, perhaps because of their ability to potentiate the effectiveness of the inhibitory transmitter GABA. As illustrated by the work of Donald Stein at Emory University, progestins work in the cerebral cortex to prevent swelling after traumatic brain injury.
6. Neural mechanisms have been conserved to provide adaptive body-brain-behavior coordination. As previously reviewed (Pfaff 1999), several physiological domains related to reproductive behaviors have been conserved from laboratory animal brains into humans. Part of the background for this argument is that the vertebrate body plan was laid down with fish and it has persisted. The popular book Your Inner Fish (Shubin 2008) gives voice to these data. More generally, all mammals shared a common ancestor approximately 140 million years ago. A tremendous number of physiological features have remained in place, as mammals evolved from smaller quadrupeds to small primates to humans themselves. Virtually all the major organs, skeletal structure, and means of reproduction have remained the same. Genomes are approximately the same size and contain large regions of chromosomes wherein genes are ordered in an identical fashion. The proteins I deal with have similar functions between experimental laboratory rodents and humans usually have more than 90 percent identity in their amino acid sequences. The following are several specific examples.
Steroid hormones. The estrogens and progestins effective for regulating behavior in laboratory animals remain chemically identical between lower mammals and humans.
Hormone receptor chemistry. The coding regions of the genes for steroid sex hormone receptors are conceived almost totally across mammalian species, leading to the inference that their chemical functions are essentially identical among laboratory animals and humans. Gene promoters for these same receptors are similar but not identical across species.
Hormone receptor neuroanatomy. The basic limbic-hypothalamic system of ER-expressing nerve cells (reviewed in Chapter 1) was discovered first in rats (Pfaff 1965, 1968) and has been conserved in considerable detail all the way from fish through primates. Naomi Rance at the University of Arizona presented further data indicating that this same system appears in the human brain.
Hypothalamic neuroanatomy. The fundamental neuroanatomy of the hypothalamus and its neuroanatomical connections are conserved in the human brain to a remarkable degree. Individual nerve cell groups remain recognizable. Larger neuroanatomical features remain in the human, such as positioning at the bottom of the brain between major long communicating systems such as the medial forebrain bundle, the fornix, the mammillothalamic tract, and the periventricular system.
Neural connections of hormone-responsive systems. As mentioned in Chapter 1, cell groups of steroid sex hormone-concentrating neurons tend to project
to other cell groups of steroid sex hormone-concentrating neurons. This tendency holds for the human hypothalamus.
Gonadotropin release. GnRH is important because it governs all of mammalian reproduction (see Chapter 5), in humans as in laboratory animals. In women as in female laboratory animals, a surge of GnRH will drive a surge of luteinizing hormone that causes ovulation. The coding region of the GnRH is substantially the same between lower mammals and humans, and the physiologically active decapeptide is identical.
GnRH neuronal migration. As mentioned in Chapter 5, GnRH neurons are not born in the brain itself but instead are born in the olfactory epithelium. They must migrate during development up the nose, along the bottom of the brain, and into the preoptic area. Our discovery was made in the mouse; subsequently, GnRH migration has been confirmed in every vertebrate species studied, ranging from fish through human beings.
Genes in the brain. The fundamental chemistry of DNA and the organization of chromosomes, enhancers, promoters, and coding regions remain essentially the same from mammalian animal brains to human brains.
Molecular biology of neurons. Everything known in recent investigations of transcriptional mechanisms and RNA processing indicates identity of important mechanisms among mammals, including humans. Likewise, matters of protein synthesis, enzymatic mechanisms, structural proteins, and proteins produced for export, in their modern understanding, appear in every respect identical among mammalian species.