by Donald Pfaff
Thus, knowing that (a) estrogens increase transcription rates in both OT and OT receptor systems; and (b) OT working through its OT receptor facilitates lordosis; it follows that (c) one way in which estrogenic hormones increase lordosis behavior is by revving up the transcription of OT and the OT receptor genes.
Enkephalin and the Delta Opioid Receptor
Gary Romano, whose laboratory skills I lauded earlier, not only wanted to study estrogen effects on the opioid peptide gene for preproenkephalin (PPE) but also to determine whether there is a sex difference in steroid hormone regulation of PPE gene expression. Slot blot hybridization analyses of RNA isolated from the VMH indicated that estrogen treatment increased the PPE mRNA levels in the ventrolateral portion of the ventromedial hypothalamic nucleus (VL-VM) of ovariectomized female rats (2.2-fold) but had no measurable effect on the PPE mRNA levels in gonadectomized males (Romano et al. 1990). Gary’s work was replicated and extended a few years later in my laboratory by Vanya Quiñones-Jenab (Quiñones-Jenab, Ogawa, et al. 1996).
In these experiments we studied the effects of estrogen treatment on PPE mRNA expression in female ovariectomized Swiss Webster mice after 0, 1, 6, 12, 24, or 48 hours using the in situ hybridization technique. The surprising aspect of the results was the amount of time required for a maximal effect; VMH neurons showed a three times increase in PPE mRNA levels after 48 hours of estrogen treatment when compared with ovariectomized control females. Andrea Lauber, working with Gary, performed what might be considered some of the most important experiments of all for the purposes of this book (Lauber et al. 1990a,b). That is, the hormone effects on PPE mRNA allowed us an opportunity to compare a brain region-specific molecular change with a quantifiable behavior in an animal by individual animal basis. Slot blots were used to measure PPE mRNA levels in the VMH as a function of the dose of estrogen administered to ovariectomized rats. Every rat used had been characterized for the ability to display lordosis behavior. Estradiol treatment led to a monotonic dose-dependent increase in PPE mRNA level in the VMH.
Lordosis behavior, assessed by two different types of quantitative assays, also increased monotonically with estradiol dose. The data indicated that an apparent threshold level of PPE mRNA in VMH coincided with the display of reproductive behavior. One potential reason for this came from the ultrastructural work done by Katie Commons when she joined the laboratory years later (Commons and Pfaff 2001).
Because enkephalin is thought to act in several brain areas to modulate the activity of GABAergic neurons, we studied the ultrastructural morphology and relationship between neurons containing these neurochemicals using dual-labeling immunocytochemistry in ovariectomized rats, half of which received estrogen replacement. Immunolabeling for enkephalin was almost always detected within axon terminals, while GABA immunoreactivity was more often localized to cell bodies and dendrites.
Crucially, axon terminals containing enkephalin immunolabeling provided a major innervation to soma or dendrites containing GABA. That is, over one-third of the axon terminals in contact with GABA-immunoreactive dendrites contained enkephalin. Furthermore, these GABA-immunoreactive dendrites accounted for a fifth of the somatodendritic processes associated with enkephalin-containing axon terminals. These findings support the hypothesis that enkephalin may act in the VMH by inhibiting GABAergic neurons, which could result in the disinhibition of neural circuits relevant for lordosis. Thus, we think that the estrogenic induction of PPE has one of its major sources of behavioral import in the disinhibition of VMH neurons as they signal to the midbrain as shown in Chapter 2.
There are fewer data on estrogens and δ-opioid receptors (primarily responsible for enkephalin signaling). The most important come from the laboratory of Theresa Milner at Cornell Medical College. Her earliest observations suggested increased δ-opioid receptor internalization and trafficking from dendrites toward cell bodies of hippocampal neurons in females at a time in their cycle when estrogen levels are high. In such females there was also increased colocalization with corticotropin-releasing hormone receptors, better capacity for long-term potentiation of electrical excitability, and more δ-opioid receptor (immunoreactivity) in dendritic spines. Whether these data would hold up for hypothalamic neurons remains to be determined.
What about lordosis behaviors? When Jim Pfaus joined the laboratory we knew that previous studies suggested that opioid receptor agonists infused into the lateral ventricles can facilitate (through δ-receptors) the lordosis behavior of ovariectomized rats treated with estrogen and a low dose of progesterone. In an elaborate experiment, we found that hypothalamic application of either δ-receptor agonist [D-Pen2,5]-enkephalin hydrate (DPDPE) or U-50488H [2-(3,4-dichlorophenyl)-N-methyl-N-[(1R,2R)-2-pyrrolidin-1-ylcyclohexyl]acetamide;methanesulfonic acid] increased lordosis quotients and lordosis magnitudes, and that the facilitation of lordosis behavior by δ-receptor agonists is independent of progesterone treatment (Pfaus and Pfaff, 1992). Anne Etgen’s laboratory got similar results and further showed that pretreatment with the selective δ-opioid receptor antagonist naltrindole (NTDL) blocked DPDPE effects on lordosis behavior. As expected, Anne’s results showed the specific importance of the VMH. As an additional point she showed that the PR antagonist RU486 blocked receptive (lordosis) and proceptive (“courtship”) behaviors induced by DPDPE. Laboratories in Japan replicated these findings and further showed differences in efficacy among different chemical forms of enkephalin.
The necessity of the enkephalin stimulation in the VMH for high levels of lordosis behavior was proved when Arnaud Nicot joined my laboratory. That is, to assess the physiological role of hypothalamic opioid expression in lordosis behavior, we synthesized a 16-mer oligodeoxynucleotide directed toward PPE mRNA and microinjected it the VMH of estradiol-primed ovariectomized rats (Nicot et al. 1997). Estradiol-induced lordosis behavior was observed in response to a stud male 2 days thereafter. Antisense injections near the ventrolateral portion of the VMH resulted in a significant reduction in lordosis quotient compared with the control treatments. The results were specific to the VMH. We also validated the effectiveness of our antisense VMH treatment both by checking enkephalin immunoreactive levels determined by radioimmunoassay and by using in situ hybridization. Thus, we could say with confidence that enkephalin gene expression in the VMH contributes causally to lordosis behavior.
In summary, because (a) estrogens heighten PPE gene expression, and (b) enkephalin helps to cause lordosis behavior; it follows that (c) one way in which estrogens facilitate lordosis behavior is through their activation of PPE gene transcription.
Neuronal Nitric Oxide Synthase
Solomon Snyder, chief of neuroscience at the Johns Hopkins School of Medicine, had already had a sparkling career marked by several high-profile discoveries when he found out that nitric oxide (NO) can act as a gaseous transmitter. This was unheard of. As a gaseous neurotransmitter NO, in part, works in tandem with glutamatergic neurons to transmit neuronal excitation.
Of course, we wanted to investigate this novel transcriptional system with respect to estrogen action in the brain. Neuronal nitric oxide synthase (nNOS) is an isoform of the enzyme responsible for the synthesis of the gaseous NO.
First, Sandra Ceccatelli, now a professor at the Karolinska Institutet, studied expression and estrogen regulation of the genes for NO-synthesizing enzymes (NO synthase, NOS) using in situ hybridization (Ceccatelli et al. 1996). Brains were sectioned and hybridized with antisense riboprobes for neuronal NOS, macrophage NOS, and endothelial NOS. In the hypothalamus, mRNA was clearly detectable only for the neuronal NOS with the probes used. A strong hybridization signal was observed in the hypothalamic cell group of greatest interest—at the top of our lordosis circuit. Quantitative analysis showed an increase in neuronal NOS mRNA in the VMH of the ovariectomized rats treated with estradiol benzoate. The increase was mainly in the ventrolateral aspect of the VMH, exactly where the most ER-expressing cells are located. Ilya Rachman working with my longtime collaborator ce
ll biologist Rochelle Cohen replicated and extended Ceccatelli’s results (Rachman et al. 1998). Short-term estrogenic regulation specifically of neuronal nNOS mRNA in the ventrolateral subdivision of the VMH was demonstrated using in situ hybridization. Estrogen-treated animals showed a significantly greater signal in the same portion of the VMH as Sandra Ceccatelli had reported. As a retrograde messenger, NO may mediate some of estrogen’s actions on various inputs to VMH neurons.
In parallel work, Ilya Rachman studied the distribution of the enzymes NADPH diaphorase (an NO marker) and NOS in the VMH (Rachman, Pfaff, and Cohen 1996). Some, but not all, neurons in the ventrolateral subdivision of the VMH contained both NADPH diaphorase and brain NOS, as demonstrated by colocalization of these two enzymes in individual cells of this area. That NADPH diaphorase and brain NOS were found in estrogen-binding cells was shown by colocalization of NADPH diaphorase and ER and brain NOS and ER at the light and ultrastructural levels, respectively. Thus, both of these enzymes are in the right place in the hypothalamus to be subject to estrogenic influence (as reported earlier) and important for lordosis (as covered later).
nNOS re lordosis: Samuel McDonald McCann was one of the most effervescent figures in American endocrinology. At international meeting after meeting, he was the loud, happy, hard-drinking Yank who could entertain every scientist present late into the evening. With respect to nNOS, as he planned his work he already knew the importance of NO in the neuroendocrine control of reproduction. Indeed, building on the pioneering work of Solomon Snyder of the Johns Hopkins School of Medicine, who had proved NO as a transmitter, McCann provided the evidence that the excitatory transmitter glutamic acid works together with NO to stimulate the release of luteinizing hormone-releasing hormone (now known as GnRH; see Chapter 5) for the ovulation-causing pituitary release of luteinizing hormone (LH). Because it makes sense that female reproductive behavior, exposing the female to predation, should be coupled with ovulation, McCann and his colleagues explored the role of NO in lordosis and demonstrated that third ventricle injection of the NO donor sodium nitroprusside, a treatment what would bathe the VMH in NO, increased female reproductive behavior; this in turn could be blocked by inhibitors of NOS.
The leading molecular endocrine laboratory of Bert O’Malley at Baylor College of Medicine followed up McCann’s work by studying the mating behaviors of female rats after administration of an inhibitor of NOS, NG-monomethyl-L-arginine, into the cerebral ventricle adjacent to the hypothalamus. This NOS blocker prevented estrogen and progesterone-facilitated lordosis, while a control injection of NG-monomethyl-D-arginine, which does not inhibit NOS, did not inhibit lordosis under the same experimental conditions. Further, microinjection into the 3V of sodium nitroprusside, which spontaneously releases NO, facilitated lordosis. O’Malley’s team concluded, in agreement with McCann, that the NOS / NO system, in addition to fostering ovulation, boosts female mating behavior.
Thus, (a) estrogens increase transcription from the NOS gene; and (b) NO activity is both sufficient and necessary to increase lordosis; so it follows that (c) one of the transcriptional systems through which estrogens act to increase female reproductive behavior is the NOS pathway.
Estrogens Also Trigger Growth Processes in Hypothalamic Neurons
Consonant with estrogen-stimulated mRNA and protein synthesis introduced at the beginning of this chapter (and consistent with the requirements for lordosis behavior), estrogens increase synthesis of ribosomal RNA (rRNA), and the resulting synthetic processes lead to cell biological signs of neuronal growth. An amplified hormone-dependent signal likely results.
Katherine J. Jones entered my laboratory shouting. The younger sister of a bunch of loud boys, she had learned that she had to shout in order to be heard at all. My office in Smith Hall at Rockefeller was then about 50 meters from the elevator, but I nevertheless could hear Kathy clearly as soon as she got off the elevator at our floor.
Kathy began by collaborating with molecular biologist Dona Chikaraishi of Duke University to use Dona’s ribosomal RNA probes with in situ hybridization to demonstrate estrogenic stimulation of rRNA synthesis in VMH neurons (Jones et al. 1986). The background of our thinking was that in tissues throughout the body important for reproduction, sex steroids make those tissues grow. And in many non-neural steroid-sensitive tissues, hormonal regulation of the polymerase I system has been shown to be a major aspect of the mechanisms by which those steroids alter cellular functions. At that time, no effect of steroids on nucleolar gene expression had been reported for any neuronal regions.
In our case we used estrogen-free ovariectomized female rats to quantify effects of 6 hours, 24 hours, or 15 days of estrogen treatments on rRNA synthesis in the brain. At 6 hours a highly significant 70 percent increase due to estrogen was observed; at 24 hours a significant 60 percent increase was seen. Surprisingly, at 15 days we saw no effect. I note the size of these hormone-induced inductions because, typically, in the brain changes in transcription rates are not as large as in peripheral tissues. The results were robust whether we quantified grains per neuron or grains per unit area (Jones et al. 1986).
Kathy was dealing with the “polymerase I” system, one of three RNA polymerase systems discovered by my Rockefeller colleague Robert Roeder. This system is extremely important in cells with high protein synthetic rates and metabolic rates such as neurons. Given the importance of the rRNA system and the strength of her initial findings (as mentioned), Kathy wanted to follow up by determining “precursor / product” relations in transcription from the DNA that encodes rRNA (Jones et al. 1990). That is, the initial transcript (“precursor”) is relatively short-lived compared with mature, stable (“product”) rRNA. Thus, two types of ribosomal DNA probes were used for in situ hybridization of the estrogen effect: rDNA with the short-lived external transcribed spacer region (precursor) and rDNA of stable 18S RNA coding region (product). Already at 30 minutes of estrogen exposure, the estrogen-treated females had significantly greater precursor rRNA synthesized than control (Jones et al. 1990). The result was specific in that we did not see results in other parts of the hypothalamus. Likewise, at the protein synthesis level, only three proteins out of 39 whose synthesis was elevated by estrogens in the VMN were also among the proteins affected in the preoptic area (Jones et al. 1988). The VMH result at 30 minutes comprises the fastest genomic effect of hormone actions in the brain.
Thinking back now to all of Kathy Jones’s work in the laboratory, it turned out, in a most logical fashion, that at very early time points, the estrogen effect could be seen on the amount of precursor rRNA per VMH neuron. At much later times, after the onset of estrogen treatment, the major effect is on mature, stable product rRNA.
Ultrastructure
I had the courage to begin electron microscopic work because of a collaboration I struck up with cell biologist Dr. Rochelle Cohen, a postdoctoral researcher in the Rockefeller University laboratory of Phillip Siekevitz, a member of the Nobel Prize–winning cell biology team headed by George Palade. With Rochelle’s extremely discriminating methodology we were able to show massive elaboration of the rough endoplasmic reticulum—the “protein synthetic machinery” of individual VMH neurons caused by estrogen treatment (Cohen and Pfaff 1981). When Bob Meisel came to the laboratory he replicated and extended Rochelle’s work (Meisel and Pfaff, 1985). The percentage of VMH neurons with what we called “stacked” endoplasmic reticulum went from 21 percent in the ovariectomized control females to 51 percent in females that had received estrogen treatment for 15 days, a more than two times increase. Further, Bob’s results were specific to the VMH. And the ultrastructural results in the VMH were tightly correlated, on an animal-by-animal basis, with the amount of lordosis behavior.
Most relevant to these rRNA results is the function of the nucleolus, the primary site of rRNA synthesis. From studies of several cell types it was known that separation or segregation of nucleolar components can occur when the demand for rRNA by a cell is greater than its synthesis.
Thus, we were alert to the appearance of the nucleolus when we compared VMH neurons of ovariectomized control animals with those from animals exposed to estrogens for 15 days. We began with light microscopy at the highest magnifications possible and then moved on to study ultrathin sections stained by sodium tungstate (Cohen, Chung, and Pfaff 1984). We concentrated on the portion of the VMH that I had earlier determined has the highest concentration of ER-α expressing neurons. Already at the light microscope level we noticed protuberances on the surfaces of some neuronal nucleoli. In fact, these protuberances were more than two times more frequent in the VMH of estrogen-treated animals than in controls.
Figure 3.3. Top: Effect of very short-term estrogen treatment (2 hours). (A) A representative ventromedial hypothalamus (VMH) neuronal nucleus, estrogen free. (B) After only 2 hours of estrogen exposure, the nucleus (N) and nucleolus (Nu) are larger and the nucleus more spherical. Fewer clumps of heterochromatin are scattered throughout the nucleus. In the lower right, a truncated view of a major change: an increase in stacked endoplasmic reticulum within an enlarged cytoplasm. Bottom: Effect of a longer estrogen treatment, behaviorally effective. (C) A representative neuronal nucleus from VMH, estrogen-free control. (D) After a long and discontinuous behaviorally effective estrogen treatment, not only is the nucleus more spherical with fewer dark heterochromatin clumps scattered within the nucleoplasm, but also there is an obvious mass of nucleolus-associated chromatin. (Adapted from Jones, Pfaff, and McEwen 1985.)
In every matched pair of animals, the estrogen-treated animal had more nucleolar protuberances than the control in that pair. Then we moved on to the ultrastructural examination. As expected, we found cell nuclei whose nucleoli had aggregations of electron-dense material corresponding to the surface features of nucleoli quantified at the light microscopic level. At high magnification, this material could be seen to be separated from the main part of the nucleolus by a narrow gap, penetrated by stands of this electron-dense material that connected it to the main portion of the nucleolus. We became very interested in this phenomenon, so we then used the sodium tungstate staining method, which differentiates between RNA- and DNA-containing structures in ultrathin sections. The protuberant surface feature was shown to be stained more densely than the nucleolus proper, supporting the interpretation that the protuberance contains DNA and thus could be considered to be nucleolus-associated chromatin, obviously there as a mechanisms for increased rates of synthesis of rRNA (consistent with our VMH data mentioned). Based on the cell biological literature, the obvious interpretation was that VMH neurons exposed adequately to estrogen face a demand for rRNA to support the increased rates of protein synthesis necessary in turn for lordosis behavior.