EFFECTS OF TUMOR NECROSIS FACTOR-ALPHA ON DORSAL VAGAL
COMPLEX NEURONS THAT EXERT REFLEX CONTROL OF THE
GASTROINTESTINAL TRACT
DISSERTATION
Presented in Partial Fulfillment of the Requirements for
the Degree Doctor of Philosophy in the Graduate
School of The Ohio State University
By
Gregory S. Emch, B. S.
*****
The Ohio State University
2002
Dissertation Committee: Approved by:
Richard C. Rogers, Advisor ______
Georgia A. Bishop Advisor
Susan P. Travers Neuroscience
ABSTRACT
The results of the experiments presented in this thesis have shown that injection of tumor necrosis factor-alpha (TNF) into the dorsal vagal complex [DVC; made up of the area postrema (AP), the sensory nucleus of the solitary tract (NST), and the dorsal motor nucleus of the vagus (DMN)] has mixed results on neuronal activity in this medullary brainstem area. In the
NST, microinjection of TNF causes a significant and dose-dependent increase in neuronal firing rate (FR) as compared to injection of saline controls. Subsequently, some NST neurons exhibit a potentiated response to afferent stimulation following pre-exposure of the neurons to TNF.
Conversely, microinjection of TNF significantly inhibits the FR of most neurons in the DMN. Immunohistochemical studies show that the protein product of the proto-oncogene c-Fos (a marker of neuronal activation) is increased in response to systemic administration of lipopolysaccharide (LPS; bacterial cell coat component that induces endogenous production of TNF) in the DVC. Additionally, protein expression is independent of the integrity of the vagus nerve(s). That is, surgical section of the vagi does not inhibit the increase in number of c-Fos labeled neurons. Direct injection of TNF into the
ii
NST causes a significant elevation of Fos-labeled neurons, and protein expression is dependent on glutamate neurotransmission since glutamate receptor antagonists abolish any significant increase in Fos-positive neurons evoked by injection of TNF alone. Therefore, it was concluded that tumor necrosis factor-alpha causes gastrointestinal stasis by removing cholinergic excitation to the stomach. TNF alters vago-vagal reflexes by acting directly on neurons at the level of the dorsal vagal complex in the medullary brainstem.
iii
ACKNOWLEDGMENTS
I would like to express my gratitude to Dr. Richard Rogers, my advisor of more than 4 years. He was interested in my progress on a daily basis, both in the lab and in my classes. Rick gave me excellent training in surgical techniques as well as in writing techniques. He also allowed me to work at my own pace and was very excited about my data. I am forever thankful to him for his excellent mentoring.
I would also like to thank Dr. Gerlinda Hermann, my other mentor in the laboratory. Gerlinda was also very excited about my day-to-day work and was also an excellent teacher. Gerlinda was the best editor a student could ask for when submitting manuscripts for publication. Gerlinda was also a very good friend and was always there to give advice on any problem, whether it was academic or personal.
Thank you to the other members of my dissertation committee, Dr.
Georgia Bishop and Dr. Susan Travers. I appreciate your time spent in reviewing my dissertation and for administering my oral defense. Also, thanks to Georgia for serving as my official advisor for the final months of my graduate school experience.
iv
VITA
Born: 11-24-74, Youngstown, OH
1997 – Present Ohio State University College of Medicine, Columbus, OH Graduate Student
1999 – 2001 Ohio State University Neuroscience Department Pre-doctoral Fellow
1993 – 1996 Youngstown State University Honors Program, Youngstown, OH Biology/Chemistry, BS, Cum Laude
PUBLICATIONS
Emch, GS, Hermann, GE, Rogers, RC. TNF-a induces c-Fos generation in the nucleus of the solitary tract that is blocked by NBQX and MK-801. Soc for Neurosci Abs. 634.9, 2001. Rogers, RC, Hermann, GE, Emch, GS, Travagli, RA, and Browning, KN. TNF-alpha acts directly on circuit elements in the dorsal vagal complex to control gastric motility. Brain Gut Odyssey Meeting, 2001. Rogers, RC, Emch, GS, and Hermann, GE. Central effects of TNF on dorsal vagal complex neurons. Euro Winter Brain Conference Abs. 86, 2001 Emch, GS, Hermann, GE, and Rogers, RC. Tumor necrosis factor-alpha: Effects on identified neurons of the dorsal vagal complex. Soc Neurosci Abs. 674.14, 1999. Emch, GS, Hermann, GE, and Rogers, RC. TNF activates gastric vago-vagal reflex NST neurons and potentiates their responsiveness to vagal afferent input. AGA Annual Meeting. Abs. 100130, 1999. Emch GS, Hermann, GE, and Rogers, RC. TNF-a inhibits physiologically identified dorsal motor nucleus neurons in vivo. Brain Res, Submitted.
v
Emch GS, Hermann, GE, and Rogers, RC. TNF-a induced c-Fos generation in the nucleus of the solitary tract is blocked by NBQX and MK-801. Am J Physiol 281: R1394-R1400, 2001. Hermann, GE, Emch, GS, Tovar, CA, and Rogers, RC. c-Fos generation in the dorsal vagal complex following systemic endotoxin is not dependent on the vagus nerve. Am J Physiol 280: R289-R299, 2001. Emch, GS, Hermann, GE, and Rogers, RC. TNF-a activates solitary nucleus neurons responsive to gastric distention. Am J Physiol 279: G582-G586, 2000.
FIELDS OF STUDY
Major Field: Neuroscience
- Neurophysiology - Immunocytochemistry - Pharmacology
vi TABLE OF CONTENTS
Page:
Abstract……………………………………………………………………………..ii
Acknowledgments…………………………………………………………………iv
Vita………………………………………………………………………………….v
List of Figures……………………………………………………………………..xii
Chapter 1: Introduction…………………………………………………………..1
Illness Response…………………………………………………………………..2
Tumor necrosis factor-alpha………………………………………………………3
Vago-vagal reflex control of digestion……………………………………………5
Afferent limb………………………………………………………………..5
NST………………………………………………………………………….6
DMN…………………………………………………………………………8
Nucleus ambiguous………………………………………………………..9
Chemical specificity of afferent connections to the DVC………………9
Examples of vago-vagal reflexes…………………………………………9
Modulation by central pathways………………………………………….10
Modulation by peripheral factors…………………………………………11
vii TNF and the DVC………………………………………………………………….13
TNF signal transduction…………………………………………………………..14
Cellular production of TNF………………………………………………………..16
TNF effects on glia………………………………………………………………...17
TNF effects on neurons…………………………………………………………...18
Central vs. peripheral activation by cytokines…………………………………..19
Hypothesis………………………………………………………………………….20
CHAPTER 2: TNF activates solitary nucleus neurons responsive to gastric distention……………………………………………………………………………21
Introduction…………………………………………………………………………21
Methods…………………………………………………………………………….24
Chemicals………………………………………………………………….24
Electrode construction…………………………………………………….24
Data collection……………………………………………………………..25
Gastric stimulation…………………………………………………………25
Surgical preparations……………………………………………………...26
Experimental design……………………………………………………….26
Statistical analysis…………………………………………………………28
Results………………………………………………………………………………28
Dose dependency of NST firing on TNF………………………………...29
Potentiation of NST response to gastric stimulation…………………...31
Discussion………………………………………………………………………….31
viii
CHAPTER 3: TNF inhibits physiologically identified dorsal motor nucleus neurons in vivo…………………………………………………………………….36
Introduction………………………………………………………………
Methods…………………………………………………………………………….38
Chemicals………………………………………………………………….38
Pipette construction……………………………………………………….38
Data Collection…………………………………………………………….39
Gastric stimulation…………………………………………………………39
Surgical Preparations……………………………………………………..40
Experimental Design………………………………………………………40
Statistical Analysis…………………………………………………………42
Results………………………………………………………………………………42
Discussion………………………………………………………………………….45
CHAPTER 4: c-Fos generation in the dorsal vagal complex after systemic
endotoxin is not dependent on the vagus nerve………………………………..47
Introduction…………………………………………………………………………47
Methods…………………………………………………………………………….52
Chemicals…………………………………………………………………..52
Experimental Design………………………………………………………53
Animals……………………………………………………………………..54
Surgical Preparations……………………………………………………..54
ix
Histological processing for c-Fos protein………………………………..56
Counting c-Fos nuclei in the dorsal medulla……………………………57
TNF assay………………………………………………………………….57
Analysis…………………………………………………………………….58
Experiment 1: Intact Vagi / Inactin (thiobutabarbitol) anesthesia…….59
Experiment 2: Left Cervical Vagotomy…………………………………..60
Experiment 3: Bilateral Cervical Vagotomy……………………………..60
Plasma TNF levels………………………………………………………...61
Results………………………………………………………………………………61
Experiment 1: c-Fos labeling in the NST, DMN, and Area Postrema in the vagus-intact, Inactin anesthetized rat……………………………………………61
Experiment 2: Effects of LPS on c-Fos labeling in the NST, DMN, and Area
Postrema in rats with left cervical vagotomy……………………………………64
Experiment 3: Effects of LPS on c-Fos labeling in the NST, DMN, and Area
Postrema in rats with bilateral cervical vagotomy……………………………..67
Plasma TNF levels……………………………………………………………….70
Discussion…………………………………………………………………………72
CHAPTER 5: TNF induced c-Fos generation in the nucleus of the solitary tract is blocked by NBQX and MK-801………………………………………..76
Introduction……………………………………………………………………….76
Methods…………………………………………………………………………..80
Animals……………………………………………………………………80
Chemicals…………………………………………………………………80 x
Surgical Preparation……………………………………………………..82
Experimental Design…………………………………………………….83
Histological processing…………………………………………………..84
Counting c-Fos labeled cells…………………………………………….85
Analysis…………………………………………………………………….86
Results………………………………………………………………………………86
Experiment 1: c-Fos labeling in the NST induced by TNF…………….86
Experiment 2: Effects of glutamate antagonists on NST c-Fos labeling
in response to co-injections with TNF…………………………………...89
Experiment 3: Effects of glutamate antagonists on NST c-Fos labeling
in response to co-injections with OT…………………………………….89
Discussion………………………………………………………………………….93
Chapter 6: Perspectives………………………………………………………….98
Summary……………………………………………………………………98
Future Directions……………………………………………
BIBLIOGRAPHY……………………………………………………..…………..106
xi LIST OF FIGURES
FIGURES PAGE
1. Effects of gastric distension, PBS, and tumor necrosis factor-a (TNF) microinjection on a single neuron from the nucleus of the solitary tract (NST; A-G are the same neuron). A: raw oscillograph record of the identification of a gastric distension-related neuron in the NST. Antral distension (bottom) produces a brisk increase in firing rate (FR) that is phase locked to the stimulus. Time scale bar = 5 s. B: integrated rate-meter record of the same event as A. Inset: 20 superimposed NST spikes. At lower time scale, bar = 10 s; inset scale bars = 200 mV/2.5 ms. C: control injection of PBS (3 nl at arrow) has no effect on FR; same neuron as in B. Inset: 20 superimposed NST spikes showing that the counted event is unitary; same waveform as the spike in B can be seen. At lower time scale, bar = 1 min; inset scale bars = 200 mV/2.5 ms. D: effect of TNF (0.03 fmol in 3 nl) injected on the NST neuron. At lower scale, bar = 1 min; inset scale bars = 200 mV/2.5 ms. E: effect of gastric distension on the NST unit 30 min after exposure to and recovery from the TNF injection in D. Note greatly potentiated response that significantly outlasts the stimulus. At lower time scale, bar = 10 s; inset scale bars = 200 mV/2.5 ms. F: drawing depicting summary of Neurobiotin-labeled recording sites. Marked area represents the spread of all NST neurons recorded in the study. Scale bar = 0.3 mm. AP, area postrema; mNST, medial solitary nucleus; dmn, dorsal motor nucleus of the vagus; st, solitary tract. G: micrograph of the Neurobiotin injection site verifying the neuron's location in the medial NST (arrow). Scale bar = 0.3 mm. cc, Central canal………………………………………………………………………...30
2. Dose-dependent effect of TNF on NST FR and potentiation of neuronal response to gastric distension. Each bar represents the FR in pulses per second (pps) compiled in 1-min epochs. A: dose-dependent NST response to TNF was very steep. NST neurons xii exposed to 0.3 fmol TNF did not recover from excitation (data not shown). Decreasing the TNF dose to 0.0003 fmol had no effect on NST neuronal firing (data not shown). Top: application of 0.003 fmol TNT does not induce a significant increase in NST neuronal FR at 1, 2, 3, and 4 min postinjection. Bottom: application of 0.03 fmol TNF induces a significant increase in NST FR at all times sampled after TNF microinjection compared with PBS. *P < 0.01 from Dunnett's test. B: TNF potentiation of neuronal response to gastric distension. FR sampled at 1, 2, and 3 min after gastric distension is not different from 1 min predistension in neurons exposed to PBS. After exposure to TNF, gastric distension triggered prolonged activation after inflation was terminated. FR at 1 min postdistension in TNF neurons was significantly greater than FR at 1 min postdistension in PBS neurons. *P < 0.05, paired t-test…………………………………………32
3. TNF inhibits gastric DMN neurons. A, Raw spike record shows a reduction in neuronal activity (upper trace) during a gastric antral balloon distention (lower trace). Inset: 10 superimposed oscilloscope traces corresponding to neuronal spikes in A. B, Expanded rate-meter record of the spike activity of DMN neuron exposed to 3nl of PBS, followed by a delayed application of TNF (0.03 fmoles). C, TNF causes a strong suppression of neuronal firing within 30 seconds of administration. D, Inhibition lasted for an average of 8.3 + 6.4 minutes (mean + S.E.M.) before neurons returned to basal FR. uV = microvolts, PBS = phosphate buffered saline, pps = pulses per second, sec = seconds, TNF = tumor necrosis factor…………………43
4. Graphic representation of the effect of TNF on DMN firing rate. Each column represents the FR in pulses per second (pps) sampled at 1 minute epochs following either PBS or TNF microinjection. At time = 0 minutes, PBS was microinjected into the DMN, and the resultant change in FR was recorded. PBS microinjection did not affect DMN neuronal firing at each time sampled (time = 0, 1 minute post-PBS, or 2 minutes post-PBS). TNF strongly inhibited DMN firing at 1, 2, 3, and 4 minutes post-TNF injection as compared to basal FR (Dunnett’s post test; *P<0.01). FR = firing rate, pps = pulses per second, PBS = phosphate buffered saline, TNF = tumor necrosis factor, min = minutes……………………………………………………………………..44
5. Experiment 1: c-Fos-activated neurons in the brainstem of vagally intact, Inactin anesthetized rats that received either intravenous PBS or LPS. A) Intravenous LPS induced a significant (p<0.0001) elevation in c- Fos-activated NST neurons in Inactin anesthetized, vagus - intact rats. The distribution of c-Fos activated neurons within the NST did xiii
not show any intrinsic “sidedness” (p=0.32), i.e., the distribution was symmetric within the NST. B) LPS exposure also resulted in significant increases in the number of c-Fos positive neurons in the DMN (p<0.01), but the absolute numbers are very small. C) LPS significantly increased (p<0.005) c-Fos labeling in the AP…..62
6. Micrographs of original coronal histological sections through the NST at the level of the area postrema. c-Fos production in response to systemic (iv) challenge of either PBS (A) or 1000ug/kg LPS (B) is demonstrated by the dark staining nucleoli. LPS evokes a substantial, bilaterally symmetrical, increase in c-Fos production in the AP, NST, and DMN in the Inactin anesthetized rat. Scale bar = 0.5mm. AP = area postrema; CC = central canal; DMN = dorsal motor nucleus of the vagus; mNST = medial portion of the nucleus of the solitary tract; st = solitary tract; 12 = hypoglossal nucleus……………63
7. Experiment 2: c-Fos-activated neurons in the brainstem of left, cervical, vagotomized that received either intravenous PBS or LPS. A) Intravenous LPS challenges demonstrated a significant increase in the number of c-Fos labeled nuclei in the NST. This increase in c- Fos activated cells was seen regardless of dose of LPS administered or side of brainstem sampled (F = 9.71, df = 3,17, p = 0.0006; Dunnett’s post hoc * = p<0.05). In these unilaterally vagotomized preparations, the number of c-Fos activated neurons showed a small but consistent difference in the pattern of distribution (i.e., left vs right); this effect was even observed in the PBS control group. (F = 40.9, df = 1,17; p = 0.0001). B) LPS exposure also resulted in increases in the number of cFos positive neurons in the DMN, but the absolute numbers are, comparatively, very small and not statistically significant (F = 2.206; df = 3,17; p = 0.1298). C) All doses of LPS significantly increased cFos labeling in the AP (F = 4.634; df = 3,16; p = 0.0162; * Dunnett’s post-test p<0.05)……..65
8. Micrographs of c-Fos production in response to systemic (iv) challenge of either PBS (A), 25 ug/kg LPS (B), 100ug/kg LPS (C), or 1000ug/kg LPS production (D) in the dorsal vagal complex of rats with left cervical vagotomy. Transection of the left cervical vagal trunk resulted in a subtle, but consistent, reduction in number of c-Fos activated neurons in the left NST of all groups tested (i.e., even the PBS controls). Scale bar = 0.5mm…………………………………………………………………….…66
xiv
9. Experiment 3: c-Fos-activated neurons in the brainstem of bilateral, cervical, vagotomized that received PBS or LPS (iv or ip). A) Animals with bilateral cervical vagotomy demonstrated a significant increase in c-Fos labeled neurons in the NST in response to 25ug/kg LPS; regardless of the route of administration (F = 15.24, df = 2,12, p = 0.0005; * = Dunnett’s post test p < 0.05). Similar to the observations in the vagally intact group, there was no “sidedness” in the distribution of c-Fos activated neurons in the NST. B) Although LPS exposure also resulted in increases in the number of c-Fos positive neurons in the DMN comparable in number to those seen in the intact and unilaterally vagotomized groups, this response was not statistically significant (p = 0.7876). C) LPS significantly increased c-Fos labeling in the AP; however, only the LPS (IV) group was statistically significant relative to the PBS group (F = 5.108; df = 2,10; p = 0.0296; * Dunnett’s post-test p<0.05)…………………………………………………………………….68
10. Micrographs of cFos production in the dorsal vagal complex of rats with bilateral cervical vagotomy. Upper photo is an example of cFos production in response to intravenous PBS. Middle photo is example of response to 25ug/kg LPS (iv); bottom photo is an example of response to 25 ug/kg LPS (ip). Bilaterally vagotomized rats are still capable of responding with an increase in cFos production in neurons of the DVC following systemic (either iv or ip) challenge with endotoxin. Scale bar = 0.5mm………………………………………………………69
11. ELISA determination of plasma TNF levels. All three doses of LPS (25, 100, or 1000 ug/kg b.w.) elicited significant production of circulating TNF in Inactin anesthetized rats regardless of route of administration (i.e., intravenous or intraperitoneal) or integrity of the vagal nerve trunks (i.e., intact, unilateral, or bilateral vagotomy). Kruskal-Wallis test p = 0.0001; * = Dunnett’s post test p < 0.05…………………………………………………………………….…71
12. Experiment 1: Effect of TNF on c-Fos expression in the NST. Graphic representation of c-Fos labeled NST cell counts expressed as mean + SEM. A dose of 3.4fgrams TNF induced c-Fos counts that were significantly different than those obtained from animals injected with PBS (*P<0.001, Selected Bonferroni’s test). Decreasing the dose 10-fold (0.34fg) did not produce statistically significant c-Fos labeling, whereas increasing the dose 10-fold (3.4fg) resulted in significant c-Fos expression in the NST (*P<0.001, Selected Bonferroni’s test)…………………………………………………………………….....87
13. Photomicrographs of coronal histological sections through the NST at the level of the area postrema. c-Fos labeled neurons are xv
characterized by their dark stain, a minimal 6mm diameter, and the presence of nucleoli. A, In PBS injected rats (20nl), few c-Fos positive cells were present (ap = area postrema, nst = nucleus of the solitary tract, dmn = dorsal motor nucleus of the vagus, st = solitary tract). B, Injection of TNF (3.4fg) induces a significant increase in c-Fos expression. C, Co-injection of TNF with the AMPA receptor antagonist, NBQX results in little c-Fos production. D, Co-injection of TNF with the NMDA antagonist, MK-801 causes scarce c-Fos labeling. Scale bar = 700mm…………………………………………………………….……….88
14. Experiment 2: Graphic representation of the effect of glutamate antagonists on c-Fos cell counts (mean + SEM) induced by TNF. Microinjection of TNF (3.4fg) results in significant c-Fos labeling in the NST compared to PBS (Selected Bonferroni’s test, *P<0.001). Inclusion of the AMPA antagonist, NBQX, along with TNF in the injection pipette inhibited significant c-Fos production. Similarly, when the NMDA antagonist, MK-801 was injected with TNF, there was no significant increase in NST c-Fos expression………………………….90
15. Photomicrographs of c-Fos induction by microinjection of OT in the presence and absence of glutamate receptor antagonists. A, PBS microinjection (20nl) results in scarce c-Fos expression in the NST. (ap = area postrema, nst = nucleus of the solitary tract, dmn = dorsal motor nucleus of the vagus, st = solitary tract). B, OT injection induces significant c-Fos labeling in the NST. C, Co-injection of OT with the AMPA antagonist, NBQX does not affect c-Fos production induced by OT. D, Microinjection of OT and the NMDA antagonist, MK-801 generates c-Fos expression that is not different than that produced by OT alone. Scale bar = 700mm………………………………………91
16. Experiment 3: Effect of glutamate receptor antagonists on c-Fos expression in the NST induced by OT. Each bar represents c-Fos positive cell counts expressed as mean + SEM. Microinjection of 20ng of OT into the NST induces a significant increase in c-Fos labeled cells compared to control (*P<0.05, Selected Bonferroni’s test). Injection of either NBQX or MK-801 with OT did not significantly inhibit c-Fos production in the NST……………………………………………………92
xvi
CHAPTER 1
INTRODUCTION
Digestion of food and subsequent absorption of nutrients is carried out in the gastrointestinal (GI) tract. For the most part, this is accomplished through the functioning of the enteric nervous system, which is relatively autonomous of the central nervous system (CNS). Although it does contain connections with the sympathetic and parasympathetic nervous systems, these are relatively sparse as compared to the large number of enteric neurons (Wood 1987). Disruption of enteric-CNS connections has little effect on functioning of the large and small intestines; however, the stomach and esophagus are more significantly impaired (Rogers 1995, Rogers 1996).
Therefore, potions of the enteric nervous system are subject to significant descending CNS control, most notably through vago-vagal reflexes. In healthy individuals, the GI tract provides needed energy for daily activities.
However, illness may disturb the digestive process to provide a variety of symptoms of GI distress. There is evidence that these symptoms are induced by the immune system’s influence over vago-vagal reflexes. The aim of this
1
report is to characterize this immune influence over the CNS. Specifically, modulation of vagal reflexes by the cytokine tumor necrosis factor-alpha will be addressed.
Illness response
There are a variety of physiological, behavioral, and neural changes that occur during inflammation, injury, or infection. These have been called
“illness responses” since they form an integrated set of adaptations that promote survival (Watkins 1995). These responses include increased sleep, fever, decreased activity, decreased food intake, formation of taste aversions to novel foods, nausea, and vomiting. These responses are protective to the host organism in that they promote elimination of pathogens. For example, fever slows the replication of pathogens and accelerates inflammatory reactions and the replication of immune cells (Moltz 1993), while gastrointestinal symptoms limit further ingestion of pathogens. In chronic disease states, such as carcinoma, sepsis, or AIDS, this immune response, i. e., gastrointestinal distress, contributes to wasting. These disorders do not target the gut directly; rather the symptoms are mediated by action in the
CNS.
The illness responses are all produced by cytokines during inflammation, injury, or infection. These symptoms can be induced by peripheral administration of the cytokines: interleukin-1, (IL-1), IL-6, and TNF.
For example, systemic injection of TNF causes increased body temperature
(Goehler 1999), decreased gastric motility (Hermann 95), decreased food
2
intake, nausea, and vomiting (Kemeny 1990). These symptoms can also be
prevented by introduction of antibodies directed against the cytokines or by
antagonists to the cytokine receptors (Watkins 1995).
Illness responses can also be induced by toxins that elaborate the
production of cytokines, such as lipopolysaccharides. Septic shock can be
induced in animal models by systemic injection of LPS. In this paradigm, LPS
induces cytokine release from immune effector cells, and TNF has been
implicated as the primary mediator of the symptoms that result (Galanos
1993). When administered to canines, LPS causes a long-lasting
suppression of gastric emptying, gastric acid secretion, and colonic transit as
well as vomiting (Cullen 1995). In the rat, injection of LPS is sufficient to
suppress gastric motility (Hermann 1999).
Tumor necrosis factor-alpha
Tumor necrosis factor-a (TNF) was initially characterized by its
production of hemorrhagic necrosis of tumors in vivo and its cytotoxicity of
tumor cells in vitro (Carswell 1975). The active form of TNF is usually a
homotrimer composed of identical 17-kDa protein subunits (Perez 1996).
This cytokine is produced primarily by macrophages and lymphocytes within
and outside of the CNS during the inflammatory response (Nathan 1987,
Powrie 1993). Cytokines, such as IL-1, and CNS pathogens, such as LPS
can induce TNF production in many different cell types (Pan 1997).
Depending on the concentration of TNF at the cellular, tissue, or
systemic levels, the net biological effects of this cytokine can be beneficial or
3
injurious to the host (Kunkel 1989, Pfizenmaier 1992, Tracey 1992). At low concentrations, TNF participates in the inflammatory response by altering capillary permeability and induces adhesion molecule expression in endothelial cells for immune cell trafficking (Collins 1986, Pober 1986). At higher TNF levels, such as in septic shock, this cytokine plays a role in the initiation of the cytokine cascade responsible for the inflammatory response, and it induces a variety of symptoms associated with illness (Pan 1997,
Dinarello 1992, Tracey 1991, Tracey 1992).
The use of recombinant human TNF (rhTNF) to treat patients with advanced cancer causes nausea and vomiting (Kemeny 1990, Hersch 1991,
Ijzermans 1992, Muggia 1992). In addition, transplant patients are commonly treated with the murine antibody, OKT3 (antihuman CD3), which induces a tremendous systemic release of cytokines, including TNF (Chatenoud 1989).
These patients may develop high fever, nausea, vomiting, and diarrhea.
Injection of anti-TNF monoclonal antibody before anti-CD3 administration suppressed nausea, vomiting, and diarrhea (Charpentier 1992, Ferran 1991).
The connection between immune activation, i. e., elaboration of TNF, and gastrointestinal neural control has been the subject of previous research from our laboratory. These data imply a direct role for TNF in the suppression of gastric motility. Here, TNF injected into the dorsal medullary brainstem abolished a centrally stimulated increase in gastric motility in a dose
4
dependent manner (Hermann 1995). This study directly linked TNF to an area involved in gastrointestinal control, i. e. vago-vagal reflex control of digestion.
Vago-vagal reflex control of digestion
Afferent limb
The typical vago-vagal reflex consists of three components: the afferent limb, the nuclei contained in the dorsal vagal complex, and the efferent limb. The afferent limb is composed of chemo- and mechanosensory receptors in the gastric antrum and duodenum that send information via vagal afferent fibers carried in the vagus nerve, the cell bodies of which are contained in the nodose ganglia (Mei 1983). The neurons contained in these ganglia are bipolar in nature, with one axon projecting centrally and another projecting to the gut (Mei 1983). The receptors are believed to be a majority of free nerve endings (Mei 1983) as well as some Pacinian-corpuscle like nerve endings (Mei 1983). These receptors transmit information about the tension of the gastrointestinal tract as well as the chemical content of the gut.
Afferent information is transmitted through the vagus nerve centrally, with vagal fibers entering the dorsolateral brainstem and traveling in the solitary tract before terminating in the area postrema (AP) (Kalia 1980, Kalia 1982,
Rinaman 1989), the nucleus of the solitary tract (NST) (Kalia 1980, Kalia
1982, Rinaman 1989), and the dorsal motor nucleus of the vagus (DMN)
(Rinaman 1989).
5
NST
The second leg of the vago-vagal reflex consists of the DVC nuclei in the medulla. The NST is the main recipient of first order visceral and gustatory information from the vagus, glossopharyngeal, trigeminal, and facial nerves (Sawchenko 1983). The NST is somatotopically organized with information from the vagus represented primarily in the medial NST
(Sawchenko 1983).
The NST is connected to a variety of brainstem and forebrain structures. This nucleus projects to preganglionic vagal motor neurons of the
DMN and the nucleus ambiguous (NA) (Beckstead 1980, Bystrzycka 1985).
In addition, projections from the NST to the intermediolateral cell column have been described in the cat (Sawchenko 1983). Projections to these three groups of neurons arise primarily from the caudal portion of the NST
(Beckstead 1980, Bystrzycka 1985, Rogers 1995, Rogers 1996). These pathways are involved in the reciprocal alterations of sympathetic and parasympathetic outflow, such as the baroreceptor reflex as well as gastric vago-vagal reflexes.
The NST sends projections to motor nuclei of cranial nerves that supply the face and tongue. These include projections from the rostral and caudal NST to nuclei of cranial nerves V, VII, and XII (Norgren 1978,
Sawchenko 1982). The NST sends visceral afferent information to reticular
6
areas that control oro-facial motor neurons. These pathways are involved in reflex behavior associated with feeding, such as the acceptance or rejection of food based on taste (Grill 1978).
The NST also projects to a group of “relay” nuclei in the brainstem.
The parabrachial nucleus receives extensive input from the NST with the medial part of the nucleus receiving gustatory information from the rostral
NST and the lateral part receiving visceroceptive information from the caudal
NST (Sawchenko 1983). The parabrachial nucleus then projects back to the
NST, to the same areas that receive input from the NST, and to part of the ventrobasal complex of the thalamus (Sawchenko 1983). There is also a direct connection between the NST and the thalamus that was described in the primate (Beckstead 1980).
Other relay areas that receive inputs from the NST are the A1 and A5 catecholamine groups cell groups in the brainstem. The A1 group projects to the DVC, locus coeruleus, and the hypothalamus (paraventricular nucleus)
(Sawchenko 1982), whereas the A5 group projects to the NST, DMN, NA, and the intermediolateral cell column (Loewy 1979). Both of these groups of projections are involved in cardiovascular control (McKellar 1982, Sawchenko
1981, Sawchenko 1982).
A number of forebrain nuclei receive inputs from the NST. Among these are nuclei of the hypothalamus including the arcuate nucleus, the lateral hypothalamic area, and the paraventricular nucleus. The bed nucleus of stria terminalis and the central amygdaloid nucleus, and other limbic nuclei
7
are also interconnected with the NST. These structures coordinate autonomic functions with anticipated changes in behavior, e. g., suppression of food intake in response to a full gut (Rogers 1995, Rogers 1996). With the exception of motor nuclei of cranial nerves V, VII, X, and XII, all nuclei that the
NST projects to projects back to the DVC and can thus influence the activity of vagal motor outflow or the afferent information transmitted through the NST itself (Sawchenko 1983).
DMN
The dorsal motor nucleus of the vagus lies directly below the NST in the caudal medulla. The DMN receives a large number of projections from the NST, which then sends vagal efferent projections to virtually all regions of the gastrointestinal tract (Altschuler 1993, Berthoud 1990a, 1991, 1990b,
Ewart 1988, Kirchgessner 1989, Leslie 1982, Roman 1987). Also, extensive projections to this nucleus arise from the lateral hypothalamus, the paraventricular nucleus (PVH), A1 and A5 groups, as well as the central amygdaloid nucleus (Sawchenko 1983).
There are numerous neurotransmitters located within the DMN. One major pathway from the DMN to the enteric nervous system involves preganglionic cholinergic fibers that synapse on nicotinic postganglionic enteric neurons, which then synapse on muscarinic receptors in the target tissue. This pathway is excitatory on the target tissue. Other neurotransmitters that exist within the DMN are enkephalin, substance P, VIP, galanin, and glutamate (Lundberg 1979, Roman 1987, Senba 1991). There
8
are also vagal efferent connections made with inhibitory enteric neurons.
Here, the neurotransmitter is of the non-adrenergic and non-cholinergic
(NANC) type (Roman 1987). Possible candidate neurotransmitters include
VIP, nitric oxide, and adenosine (Lundberg 1979, Meulemans 1993).
Nucleus ambiguous
The nucleus ambiguous, along with the DMN, provides vagal preganglionic fibers to the viscera. There is considerable overlap in visceral innervation between the DMN and the NA (Kalia 1980b). The NA lies in the path of many ascending and descending fiber systems making it difficult to discern terminal endings onto this nucleus with light microscopy (Sawchenko
1983). However, it is known that there are massive projections from the NST
(Sawchenko 1983) and input from the medial parabrachial nucleus
(Sawchenko 1983).
Chemical specificity of afferent connections to the DVC
Massive projections arising from the paraventricular nucleus of the hypothalamus terminate among the nuclei in the DVC. Immunoreactivities that have been demonstrated in these projections include those of corticotropin-releasing factor, somatostatin, enkephalin, vasopressin, dopamine, oxytocin, neurotensin, and substance P (Sawchenko 1983).
Examples of vago-vagal reflexes
There are several documented examples of gastric vago-vagal control reflexes (Abrahamsson 1973, Abrahamsson 1984, Glise 1984, Kosterlitz
1968, McCann 1992, Rogers 1995, 1996). Some of these reflexes control the
9
passage of food into and out of the stomach, e. g. duodeno-gastric reflex, while others are involved in secretion of digestive fluid, e.g., gastric distention- secretion reflex. The gastric “accommodation” reflex involves a relaxation in the proximal gut in response to luminal distention and mechanoreceptor activation in the antrum. Vagal afferents release glutamate thus exciting neurons in the medial NST. The NST sends inhibitory connections to DMN neurons that control the proximal stomach (Rogers 1995, Rogers 1996). The inhibition of DMN efferents creates a state of low pressure in the proximal stomach that causes a retropulsion of chyme from the antrum to the corpus
(Abrahamsson 1973a, b, Roman 1987, Scratcherd 1989). This mechanism is responsible for trituration of large food particles during digestion.
The “receptive relaxation” reflex reduces intragastric pressure and increases gastric volume in anticipation of a bolus of swallowed food (Miolan
1984, Roman 1987). Here, vagal afferent receptors in the wall of the esophagus are activated in response to esophageal distention, thus activating inhibitory NANC vagal efferents innervating the proximal gut (Miolan 1984,
Roman 1987). This reflex is important in relaxing the stomach so that it can receive larger amounts of food. This reflex is also accompanied by inhibition of cholinergic excitation to the stomach (Miolan 1984, Roman 1987).
Modulation by central pathways
Vago-vagal reflexes are subject to substantial modulation by the CNS.
One major pathway involves the nucleus raphe obscurus (nRO), which makes many connections with the DVC (Tache 1993, 1994). Neurons in the nRO
10
containing serotonin (5HT) and thyrotropin releasing hormone (TRH) project to the DVC (Merchanthaler 1988, Tache 1993, Tork 1985), which has been shown to have receptors for both transmitters (Manaker 1985, 1993, Thor
1992). Picomolar amounts of TRH applied in the ventricles or directly to the dorsal vagal complex evokes massive increases in gastric motility and gastric acid secretion (Garrick 1989, Hornby 1989, Ishikawa 1988, Rogers 1985a,
1987), and this effect is mimicked by nRO stimulation (Rogers 1995, Rogers
1996). 5HT synergizes the TRH-induced gastric activation (McCann 1988,
McTigue 1992).
When applied directly to DVC neurons in vivo, picomolar amounts of
TRH inhibit NST neurons, whereas DMN neurons are strongly excited
(McCann 1988). In this way, TRH overrides the normal negative feedback nature of vago-vagal reflexes by inhibiting the sensory limb (the NST) and by activating the excitatory efferent limb (Rogers 1995, 1996). In vitro brainslice preparations reveal that DMN neurons are rapidly depolarized, causing bursts of action potentials (McCann 1989). This regulation of gastrointestinal function by TRH is probably part of the homeostatic response to cold exposure, wherein augmentation of gastrointestinal function leads to increased core body temperature in preparation for increased food intake
(Rogers 1995, 1996).
Modulation by peripheral factors
Although the DVC is modulated by central descending pathways, it is also subject to modulation by peripheral factors. Previous anatomical work
11
(Rogers 1993, Shapiro 1985) has shown that dendrites of the NST and DMN
enter the area postrema, a well-established chemosensory structure.
Because the dendrites of the sensory (NST) and the motor (DMN)
components of this area are completely merged, this region is referred to as
the dorsal vagal “complex”. The blood supply to the DVC consists of
fenestrated capillaries and this structure is essentially devoid of the blood-
brain barrier (Broadwell 1993, Gross 1990, Whitcomb 1990). Thus, the DVC
may be in a position to monitor blood-borne and CSF-borne factors (e.g., peptide hormones, IgG, serum albumin, and other large peptides) (Rogers
1995, Rogers 1996). Horseradish peroxidase (>60,000 MW) can be seen in all circumventricular organs and adjacent brain parenchyma, such as the
DVC, within 10 minutes after IV injection of HRP (Broadwell 1993).
Therefore, both NST and DMN neurons are continuously exposed to circulating peptides and proteins (Broadwell 1993, Gross 1990).
Previous work has shown that neurons of the NST and the DMN are subject to modulation by the peptide hormones, peptide YY (PYY) and pancreatic polypeptide (PP). These large molecular weight peptides (>5 kD) are produced peripherally yet have specific sites of action within the DVC
(Chen 1995, Chen 1997, McTigue 1993, McTigue1997, McTigue1995).
PP is released from endocrine cells in the pancreas at low rates during fasting and its release is strongly exaggerated during anticipation of feeding and after a meal (Rogers 1995, Rogers 1996). In the feeding response, the presence of fats in the stomach and the small intestine activates
12
chemosensitive vagal afferents that, in turn, activate vagal reflex components resulting in PP release (Rogers 1995, Rogers 1996). Once in the circulation,
PP reduces pancreatic enzyme secretion through a vagally dependent mechanism and increases motility and gastric acid secretion through its action at the DVC (McTigue 1993, McTigue1995, Rogers 1995, Rogers
1996).
PYY is released from endocrine secreting cells after a meal or after perfusion of the intestine with fatty acids (Chen 1995). PYY delays gastric emptying, decreases pancreatic exocrine secretion, and suppresses motility in the proximal gut, the jejunum, and the colon through vagal mechanisms
(Chen 1995, Chen 1997). PYY acts through the Y2 receptor to directly inhibit parasympathetic efferent neurons in the DMN (Chen 1997). This mechanism is responsible for the ileal brake effect in which fatty acids in the ileum cause a profound decrease in gastric motility.
TNF and the DVC
Given the precedence of peptide hormones modulating centrally mediated gastrointestinal function, it was hypothesized that the DVC may be the site of TNF action as well. In support of this hypothesis, the brainstem has been shown to have a high density of TNF binding sites (Kinouchi 1991).
Previous studies have demonstrated that endogenous production of TNF in response to intravenous administration of LPS is sufficient to suppress gastric motility (Hermann 1999). Additionally, an earlier study demonstrated that
TNF injected unilaterally into the DVC abolished a centrally stimulated vagally
13
dependent increase in gastric motility in a dose dependent manner (Hermann
1999). The rapid onset of the centrally injected TNF effect on gastric motility, i.e., within 30 seconds of application, suggested that TNF could be directly modulating the firing rate of neurons in the NST and/or the DMN (Hermann
1999). If this hypothesis is correct, it is still unknown how TNF could be affecting neurons in the DVC. A review of TNF signal transduction may provide some clues.
TNF signal transduction
There are two subtypes of receptors for TNF: P55 and P75 (Rothe
1992). Both are transmembrane proteins that belong to the low affinity neurotrophin receptor gene superfamily (Hsu 1993, Rothe 1992). Both subtypes have been identified in the CNS, although the cellular distributions have not been characterized yet (Pan 1997). TNF binding studies reveal that binding capacity is greatest in the brainstem and detectable in the cortex, thalamus, and basal ganglia. Both receptor subtypes are present in neurons and glia (Boka 1994).
A number of signal transduction pathways have been linked to the P55 and P75 receptors (Beyaert 1994, Kronke 1990, Rothe 1992). G-protein mediated activation of phospholipase C has been demonstrated resulting in protein kinase C activation (PKC) and phosphorylation of multiple cytosolic proteins, activation of NFkB that binds to the NFkB gene element, and binding to the AP-1 site with immediate early gene expression. G-protein mediated activation of protein kinase A has been shown with subsequent
14
cAMP production and activation of cAMP- response element (CREB) in the
target gene. TNF activates phospholipase A2, thus producing arachidonic
acid leading to leukotrienes and prostaglandin synthesis. The
sphingomyelinase pathway is activated by TNF, and nitric oxide, free radicals,
and ceramide are produced. Additionally, activation of these receptors has
been linked to phosphorylation of other membrane receptors by protein
kinases.
The two subtypes of TNF receptors are regulated independently,
although some cytokines can upregulate both receptors in culture (Bebo
1995, Tartaglia 1991). However, there is functional cooperation by the two
subtypes. The P75 receptor is more sensitive to membrane-bound-TNF
(Grell 1995), and therefore may serve to pass the bound ligand to P55, thus regulating signal transduction induced by P55 (Lindvall 1990,Tartaglia
1993b). The P55 receptor is involved in both cell proliferation (Barna 1993) as well as a Fas-linked cytotoxicity pathway (Tartaglia 1993a) induced by
TNF. However, the P75 receptor is required for apoptotic cell death (Heller
1992,Vandenabeele 1995). TNF induced apoptotic pathways involve an increase in c-Myc and decrease in bcl-2 oncoproteins, as well as free radical formation (Klefstrom 1994, Talley 1995, Vandenabeele 1995). In summary,
TNF activates diverse signal transduction pathways through the receptors, leading to a modification of cell proliferation or death.
15
Cellular production of TNF
During inflammation TNF is produced primarily by macrophages
(Nathan 1987). However, it is also produced by lymphocytes (Powrie 1993).
Production of TNF by these two cell types takes place both centrally and peripherally (Pan 1997). Therefore, during inflammation when the permeability of the BBB is increased, circulating immune effector cells as well as TNF can enter and induce further central TNF production (Pan 1997).
Microglia, or activated macrophages, produce TNF during CNS injury (Perry
1993). Astrocytes have been shown to produce TNF in response to lipopolysaccharide and other cytokines, such as interleukin-Ib (Lee 1993,
Sharif 1993). Other inducers of TNF in astrocytes include interferon-g (Nitta
1994), radiation (Chiang 1991), and viruses (Sierra 1993). Production of TNF by astrocytes can depend on differences in developmental stage of the animal (fetal vs. adult) and synergism between different cytokines (Chung
1991), as well as strain differences between rats (Heuschling 1995). Under physiological conditions, neurons seem to be the principal cell type that produces TNF (Pan 1997). TNF-like immunoreactivity in the CNS has been mapped, and a high density of labeling has been demonstrated in the ventral pons and medulla, the hypothalamus, bed nucleus of stria terminalis, and the raphe nuclei and is abundant in pathways that innervate autonomic and neuroendocrine-related regions (Pan 1997). TNF expression is also increased after ischemic injury (Liu 1994). Massive production of TNF is induced in neurons depending on the nature, extent, and locus of the insult 16
(Pan 1997). TNF produced by glia seems to be important in affecting neuronal activities in a paracrine manner, whereas TNF produced by neurons serves in a protective or damaging manner to neighboring cells (Pan 1997).
TNF effects on glia
TNF promotes growth in differentiated astrocytes whereas no effect is noted in undifferentiated astrocytes (Oh 1993). TNF increases proliferation in astrocytoma cell lines (Lachman 1987). TNF is a mitogen in both human and bovine adult astrocytes in culture, whereas in those from human fetal brains, the same doses inhibit proliferation (Barna 1990, Moretto 1993, Selmaj 1990).
In microglia, TNF induces proliferation in mixed cultures but not in isolated cultures (Ganter 1992), suggesting that other factors may play a role in the cellular response (Pan 1997). Microglia are more responsive than astrocytes in IL-6 production following LPS challenge. In addition, TNF causes protein kinase C dependent IL-6 production in microglia but not in astrocytes
(Sawada 1992, Sparacio 1992).
The effect of TNF on oligodendrocytes, by contrast, is usually detrimental. In myelinated aggregating cultures of mouse spinal cord tissue,
TNF results in necrosis of oligodendrocytes and damage to myelin 18-24 h later. A neutralizing antibody to TNF blocks the oligodendrocyte damage, suggesting that the effect is a specific, rather than a toxic, response (Selmaj
17
1988). TNF can activate microglia to produce nitric oxide, thus leading to free
radical formation within oligodendrocytes and eventual cell death (Mitrovic
1994).
Taken together, current in vitro studies indicate that the three types of
glial cells respond differently to TNF. For oligodendrocytes, TNF is generally
cytotoxic, whereas astrocytes and microglia may show either growth arrest or promotion, as well as other transient chemical changes. The overall reaction
depends on the stage of differentiation and coexisting factors, including the other types of glial cells.
TNF effects on neurons
TNF interacts with multiple neurotransmitters, and participates in
synaptic activity. Besides its well-known effect as an endogenous pyrogen on hypothalamic neurons, TNF may modulate the responsive state of C12 adrenergic receptors to tricyclic antidepressants and potentiate the release of norepinephrine in hippocampal slices (Ignatowski 1994). TNF also increases the frequency of spontaneous miniature synaptic current through a presynaptic mechanism (Grassi 1994), augments evoked synaptic activity, and inhibits long-term potentiation in hippocampal slices (Tancredi 1992).
TNF potentiates glutamate neurotoxicity in human fetal cell culture, and the NMDA receptor antagonist MK-801 can block this potentiation.
Incubation with TNF reduces 3 H-glutamate uptake and glutamine synthetase activity in astrocytes, indicating interference with the metabolism of glutamate
18
(Chao 1994). The effect of TNF is decreased by AMPA receptor antagonists
(Gelbard 1993). These results suggest a negative role for excess TNF in
CNS development by facilitating excitotoxicity.
There are also reports, however, that TNF protects neurons against neurotoxins. Amyloid b-peptide, which may be involved in the pathogenesis of Alzheimer's disease, increases the production of free radicals and intracellular calcium and is related to neuronal degeneration. TNF attenuates both the accumulation of peroxide and elevation of calcium by glutamate in the presence of amyloid peptide, and this protective effect may be mediated by a pathway involving NFKB (Barger 1995). Similarly, TNF is beneficial in the maintenance of calcium homeostasis in that it reduces the neuronal injury resulting from deprivation of glucose and the cytotoxicity of excitatory amino acids. TNF is produced in the substantia nigra of Parkinson's disease, but whether it contributes to neuronal degeneration or serves as a protective agent is uncertain (Boka 1994, Mogi 1994). Thus, paradoxical effects of TNF have been reported from different experimental models and neurons of different origins.
Central vs. peripheral activation by cytokines
It is possible that TNF gains access to the CNS at the DVC to activate a wide variety of signal transduction pathways in order to evoke changes in neuronal firing, thus providing gastroinhibition. However, the possibility exists that TNF evokes changes centrally by acting in the periphery. There is a body of evidence that suggests an alternate pathway by which information
19
about cytokines may evoke changes in CNS functioning (Blatteis 1998,
Fleshner 1998, Gaykema 1998, Goehler 1999, Milligan 1997, Sehic 1996). In
these studies, vagal afferents are implicated as containing elements
responsive to cytokines, and the resultant activation of these elements is
thought to mediate illness responses that ensue. However, while
subdiaphragmatic vagotomy blunts some of the cytokine-induced effects,
specific deafferentiation of vagal pathways does not block symptoms, such as
suppression of food intake or the febrile response (Caldwell 1999, Kapas
1998, Porter 1998, Schwartz 1997). Therefore, it is possible that both the
vagal afferent and direct DVC mechanisms operate in parallel to monitor the
portal circulation and the general systemic circulation, respectively.
Hypothesis
TNF is responsible for gastroinhibition due to its actions in the DVC. It
is possible that TNF may act directly at the DVC circumventricular organ, that
TNF activates receptors on the vagus nerve, or both. According to previous
data from our laboratory, we hypothesized that tumor necrosis factor-alpha causes gastrointestinal stasis by removing cholinergic excitation to the stomach. Here, the action of TNF would mimick the effect of activating afferent receptors in the GI tract. That is, TNF alters vago-vagal reflexes by acting directly on neurons at the level of the dorsal vagal complex in the medullary brainstem. The studies outlined in this dissertation have been designed to prove this hypothesis through electrophysiological and immunohistochemical methods.
20
CHAPTER 2
TNF activates solitary nucleus neurons responsive to gastric distention.
INTRODUCTION
Tumor necrosis factor-alpha is liberated as part of the immune response to antigenic challenge, carcinogenesis, and radiation therapy.
Previous studies have implicated elevated circulating levels of this cytokine in the gastric hypomotility associated with these disease states. Our earlier studies suggest that a site of action of TNF may be within the medullary dorsal vagal complex. In this study, we describe the role of TNF as a neuromodulator affecting neurons in the nucleus of the solitary tract that are involved in vago-vagal reflex control of gastric motility. The results presented herein suggest that TNF may induce a persistent gastric stasis by functioning as a hormone that modulates intrinsic vago-vagal reflex pathways during illness.
Cytokines are released by activated macrophages and lymphocytes as part of the immune response to antigenic challenge. Elevation of tumor necrosis factor in the systemic circulation has been correlated with anorexia, nausea, vomiting, and gastrointestinal stasis (Hersch 1991, Kapas 1992), which could imply an immune to nervous system communication during
21
illness. Recent evidence suggests that the nucleus of the solitary tract in the
medulla oblongata may be one locus for TNF action to control gastrointestinal
function (Hermann 1995, 1999).
The solitary nucleus receives a host of gastrointestinal mechano- and
chemosensory information via the vagus nerve (Berthoud 1992, Rogers 1995,
1996). General visceral afferent signals in the vagus uniformly excite
second-order NST neurons, which, in the main, inhibit dorsal motor nucleus
neurons that provide tonic excitatory vagal input to the stomach (Rogers
1995, 1996, Zhang 1992). This vago-vagal reflex (i.e., the accommodation
reflex) control of the stomach involves a negative feedback signal produced
by the NST. Additionally, the dorsal vagal complex (DVC) possesses the
characteristics of a circumventricular organ and is devoid of the blood-brain
barrier (Broadwell 1993, Gross 1990, Whitcomb 1990). Previous anatomical
work (Rogers 1993, Shapiro 1985) has also shown that dendrites of the NST
and DMN enter the area postrema, a well-established chemosensory
structure.
Thus the NST may be in a position to monitor bloodborne and
cerebrospinal fluid-borne factors (Rogers 1995, Rogers 1996). Indeed,
previous work in our laboratory has shown that neurons of the NST and the
DMN are subject to modulation by the hormones peptide YY and pancreatic
polypeptide. These peptides are produced peripherally yet have specific sites of action within the DVC (Chen 1997, McTigue 1997).
22
The brain stem has been shown to have a high density of TNF binding sites (Kinouchi 1991). Given the anatomical characteristics of the DVC and the precedence of peptide hormones modulating centrally mediated gas- trointestinal function, it was hypothesized that the NST may be the site of TNF action as well.
Our previous studies (Hermann 1999) have demonstrated that endogenous production of TNF in response to intravenous administration of the bacterial cell wall component lipopolysaccharide is sufficient to suppress gastric motility. Our earlier study (Hermann 1995) demonstrated that TNF injected unilaterally in the DVC abolished a thyrotropin-releasing hormone-stimulated vagally dependent increase in gastric motility in a dose-dependent manner. The rapid onset of the centrally injected TNF effect on gastric motility, i.e., within 30 s of application, suggested that TNF could be directly modulating the firing rate of neurons in the NST and/or the DMN.
Preliminary studies by our laboratory show that expression of the immediate-early gene product c-Fos in the NST in response to LPS challenge is significantly elevated 90 min after either intravenous or intraperitoneal injection. This increase in c-Fos production is independent of an intact vagal pathway (Emch 1999).
Because TNF powerfully inhibits gastric motility when placed in the
DVC, we hypothesized that this peptide would activate NST neurons involved in the accommodation reflex. In this way, TNF would suppress gastric motility and tone by mimicking the effects of the activation of gastrointestinal
23
afferents. Therefore, in the present study, NST neurons that form the sensory portion of the gastric accommodation reflex were identified using neurophysiological methods described by McCann and Rogers (1992) and
Zhang et al. (1992) and were exposed to microinjection of TNF.
METHODS
Chemicals
Animals were anesthetized with thiobutabarbitol (Inactin; Sigma
Chemicals, St. Louis, MO) dissolved to a concentration of 100 mg/ml in saline solution (0.9% NaCl) and administered at a dose of 100 mg/kg body wt ip. A
2 M NaCl solution was used to record extracellular neuronal potentials.
Neurobiotin (2%; Vector, Burlingame, CA) was included in the recording pipette to iontophoretically mark the recording site. PBS (124 mM NaCl, 26 mM NaHC03, and 2 mM KH2PO4; 304 mosmol; pH = 7.4) was used as a vehicle injection control. Recombinant rat TNF (R&D Systems, Minneapolis,
MN) was dissolved in PBS to a concentration of 10-6 M, divided into 25-ml aliquots, and stored at –700C until use. Stored aliquots of TNF were further diluted with PBS such that the microinjection electrode contained 10-8, 10-9, or
10-10 M TNF. Neurobiotin-injected brain stem slices were reacted with Vector
ABC and SG reagents (Vector) to visualize the marked recording site.
Electrode construction
Triple-barrel glass micropipettes were constructed such that one pipette was used for recording (tip diameter of ~ 1mm), whereas the other two
24
were available for drug delivery (combined tip diameter of 10-20 mm). The
recording pipette was filled with 2 M NaCl and 2% Neurobiotin. The other
pipettes were filled with either PBS or TNF in PBS. The electrode array was
placed in a stereotaxic electrode carrier, which was oriented at an
approximate 20o rostral angle. The injection pipettes were connected to a
micropressure injection apparatus as described by Chen et al. (1997).
Data collection
Extracellular neuronal activity was recorded via a silver-silver chloride
wire placed in the recording pipette. Extracellular neuronal potentials from
the recording electrode were amplified (10,000 X) and band-pass filtered
(300-10,000 Hz). Signals were then displayed on an oscilloscope and recorded on magnetic tape. Extracellular spike potentials were also processed on-line by a window discriminator-rate meter circuit. The resulting neuronal firing rate (FR) along with the gastric distension stimulation signal were displayed on a chart recorder. Figures were generated by processing the recorded data with an IBM-PC based RC Electronics (Goleta, CA) waveform analysis system.
Gastric stimulation
A small gastric stimulation balloon was constructed as described previously (McCann 1992). The balloon catheter was connected to one port of the dome of a Statham P23 pressure transducer. The balloon was fully
25
inflated by injecting 3 ml of air into the other port with a 5-ml syringe (McCann
1992). The gastric distension signal was transmitted from the pressure transducer to the polygraph and magnetic tape.
Surgical preparations
Male Long-Evans rats (Charles River, Wilmington, MA, USA) weighing
200-600 g were provided with food and water ad libitum and were kept on an approximate 12:12-h day-night cycle. Before each experiment, the animal was food deprived for approximately 18 h. The animal was deeply anesthetized with Inactin, and the trachea was cannulated to maintain an open airway. The gastric balloon was secured in the antrum of the stomach according to previous protocol (McCann 1992). The animal was placed in a stereotaxic frame, and the dorsal surface of the brain stem was exposed as described previously (McCann 1992). All experimental protocols were performed according to guidelines set forth by the National Institutes of Health and were approved by the Ohio State University Institutional Laboratory Ani- mal Care and Use Committee.
Experimental design
The pipette array was advanced into the DVC in search of NST neurons responsive to gastric distension. Initial stereotaxic coordinates were
0.3 mm anterior to the calamus scriptorum and 0.3-0.5 mm lateral to the midline of the area postrema (McCann 1992). A hydraulic microdrive (David
Kopf Instruments, Tujunga, CA) was used to advance the array at 10-mm increments. At each electrode advancement through the medullary brain
26
stem, the gastric balloon was momentarily inflated and deflated. Neurons briskly activated by gastric distension were located between 250 and 600 mm below the brain stem surface.
Once a cell was identified as a gastric inflation-related NST neuron,
PBS (3 nl) was micropressure injected from the attached pipette, and the spontaneous activity was monitored for at least 5 min. If neuronal activity was altered by PBS microinjection (as a consequence of nonspecific volume effects), it was rejected from further consideration. If PBS injection had no effects on the FR of the neuron, the same volume of TNF was then injected
(e.g., 3 nl of 10-8 M TNF was delivered for a total dose of 0.03 fmol). This dose of TNF delivered to the NST is also the threshold dose for TNF inhibition of gastric motility when applied to the DVC (Hermann 1995). With the use of this dose as a reference point, additional doses of TNF were tested on other cells by increasing or decreasing the concentration 10-fold.
Cells considered responsive to TNF must demonstrate a change in FR of a minimum of 50% relative to basal levels of activity (Chen 1997, McCann
1992). At the end of the recording session, iontophoretic current was applied to the recording pipette (100-1,000 nA positive direct current) to eject
Neurobiotin to mark the recording site. The amount of TNF-PBS applied was verified by inspecting the movement of the meniscus with a small microscope
(Chen 1997). Animals were transcardially perfused with saline and 4%
27
paraformaldehyde. The brain stems were removed, postfixed overnight,
sectioned at 50 mm, and reacted with Vector ABC and SG reagents for the
demonstration of Neurobiotin-labeled NST neurons (Rogers 1999).
Statistical analysis
The effect of TNF on NST firing was examined by comparing the peak
FR after TNF injection with basal FR (i.e., post-PBS). FR comparisons were
made at 1-min epochs for 1 min before and 1, 2, 3, and 4 min post-PBS or
post-TNF injection using a repeated-measures one-way ANOVA followed by
Dunnett's posttest. Time of activation, i.e., the duration that the neuron produced action potentials at a rate >50% above basal FR, was analyzed by a one-way ANOVA and a post hoc Dunnett's test. Potentiation by TNF of
NST responsiveness to gastric distension was evaluated by comparing the peak FR after inflation in neurons previously exposed to TNF with the peak
FR after inflation following PBS. Comparisons were made (i.e., pre- TNF vs. post- TNF of a given neuron) at 1-min epochs after inflation by using paired t-tests. Statistical significance required P values <0.05.
RESULTS
NST neurons located between 250 and 600 mm below the brain stem surface were neurophysiologically identified and exposed to PBS and one of three concentrations of TNF; a representative injection site is included in Fig.
1G. Injection of TNF from the pipette array typically increased neuronal firing
28
in the NST within 30 s of administration (Fig. 1D). Duration of activation varied with TNF dose (0.03 fmol = 8.8 + 2.1 min; 0.003 fmol = 2.4 ± 1.3 min).
Dose dependency of NST firing on TNF
A dose of 0.03 fmol of TNF provided the ideal stimulus in that all identified neurons (N = 14 from 11 animals) subjected to this dose responded to TNF within 30 s of application, all were activated by TNF, and all recovered from TNF-induced activation (Fig.1). If the dose was increased 10-fold (0.3 fmol), NST neurons did not recover from excitation (data not shown). If the dose was decreased 10-fold (0.003 fmol), a more subtle activation by TNF was documented. At this dose, approximately one-half (4:10 from 7 animals) of the neurons elicited a response, implying that the lower dose of 0.003 fmol is close to the ED50 for TNF to elicit changes in NST FR. Decreasing the
TNF dose to 0.0003 fmol had no effect on NST neuronal firing (data not shown). In contrast, a dose of 0.03 fmol elicited a significant increase in neuronal FR at 1, 2, 3, and 4 min after microinjection (Fig. 2A) as well as an activation time that was significantly greater than in controls in all neurons tested.
29
Figure 1. Effects of gastric distension, PBS, and tumor necrosis factor-a (TNF) microinjection on a single neuron from the nucleus of the solitary tract (NST; A-G are the same neuron). A: raw oscillograph record of the identification of a gastric distension-related neuron in the NST. Antral distension (bottom) produces a brisk increase in firing rate (FR) that is phase locked to the stimulus. Time scale bar = 5 s. B: integrated rate-meter record of the same event as A. Inset: 20 superimposed NST spikes. At lower time scale, bar = 10 s; inset scale bars = 200 mV/2.5 ms. C: control injection of PBS (3 nl at arrow) has no effect on FR; same neuron as in B. Inset: 20 superimposed NST spikes showing that the counted event is unitary; same waveform as the spike in B can be seen. At lower time scale, bar = 1 min; inset scale bars = 200 mV/2.5 ms. D: effect of TNF (0.03 fmol in 3 nl) injected on the NST neuron. At lower scale, bar = 1 min; inset scale bars = 200 mV/2.5 ms. E: effect of gastric distension on the NST unit 30 min after exposure to and recovery from the TNF injection in D. Note greatly potentiated response that significantly outlasts the stimulus. At lower time scale, bar = 10 s; inset scale bars = 200 mV/2.5 ms. F: drawing depicting summary of Neurobiotin-labeled recording sites. Marked area represents the spread of all NST neurons recorded in the study. Scale bar = 0.3 mm. AP, area postrema; mNST, medial solitary nucleus; dmn, dorsal motor nucleus of the vagus; st, solitary tract. G: micrograph of the Neurobiotin injection site verifying the neuron's location in the medial NST (arrow). Scale bar = 0.3 mm. cc, Central canal.
30
Potentiation of NST response to gastric stimulation
NST cells exhibit a stereotypic response to gastric distension stimulation (Fig. 1, A and B). That is, NST neurons demonstrate an increased
FR tightly locked to gastric balloon inflation. Once mechanoreceptor activa- tion is terminated, NST neurons quickly return to basal FRs (McCann 1992,
Zhang 1992). Activation by TNF altered this relationship in 6 of 14 NST neurons so tested at 0.03 fmol. After these neurons recovered from initial activation after TNF microinjection, i.e., return to basal FR, subsequent balloon inflation triggered prolonged activation (Figs. 1E and 2B). The potentiation response ranged from 1 to 7 min in duration and could be elicited for up to 90 min after recovery from exposure to TNF.
DISCUSSION
Our previous study (Hermann 1995) demonstrated that injecting subfemtomolar amounts of TNF into the DVC suppressed centrally stimulated gastric motility. The present study showed that identified gastric NST neu- rons are excited by TNF. The observation that TNF can affect the sensory component of this vago-vagal reflex circuit may account for the rapid and large reductions in gastric motility observed in our previous studies (Hermann
1995). The principal action of NST neurons activated by gastrointestinal distension appears to be the inhibition of DMN neurons providing excitation to the stomach (McCann 1992, Rogers 1999, Zhang 1992). Therefore, the activation of gastric-related neurons by TNF would be expected, given the
31
Figure 2. Dose-dependent effect of TNF on NST FR and potentiation of neuronal response to gastric distension. Each bar represents the FR in pulses per second (pps) compiled in 1-min epochs. A: dose-dependent NST response to TNF was very steep. NST neurons exposed to 0.3 fmol TNF did not recover from excitation (data not shown). Decreasing the TNF dose to 0.0003 fmol had no effect on NST neuronal firing (data not shown). Top: application of 0.003 fmol TNT does not induce a significant increase in NST neuronal FR at 1, 2, 3, and 4 min postinjection. Bottom: application of 0.03 fmol TNF induces a significant increase in NST FR at all times sampled after TNF microinjection compared with PBS. *P < 0.01 from Dunnett's test. B: TNF potentiation of neuronal response to gastric distension. FR sampled at 1, 2, and 3 min after gastric distension is not different from 1 min predistension in neurons exposed to PBS. After exposure to TNF, gastric distension triggered prolonged activation after inflation was terminated. FR at 1 min postdistension in TNF neurons was significantly greater than FR at 1 min postdistension in PBS neurons. *P < 0.05, paired t-test.
32
potent central nervous system (CNS) effect of TNF to suppress gastric motility via a vagal efferent route (Hermann 1995). Results from our laboratory show that DMN neurons appear to be inhibited by TNF, perhaps as a consequence of inhibitory input from the NST (Emch 1999; Emch 2000, see chapter 3).
It is likely that TNF generated as a consequence of infection or other proinflammatory processes inhibits gastric motility by directly affecting the sensitivity of gastric vagal control circuitry in the medulla. This vago-vagal reflex circuit is outside the blood-brain barrier and therefore is readily accessible to large circulating peptides (Broadwell 1993, Gross 1990,
Whitcomb 1990). Additionally, TNF increases vascular-brain permeability
(Lotan 1994, Megyeri 1992), and TNF may gain access to the brain through a specific transport system (Gutierrez 1993).
The results from the present study support the proposition that TNF alters autonomic function as a consequence of direct peptide action on regulatory circuits in the medulla. Other routes for TNF modulation of brain stem autonomic control are possible. Studies by Sehic and Blatteis (Sehic
1996) indicate a role of the afferent vagus in the transduction of information about systemic cytokine production, which results in the onset of fever.
However, this pathway seems to be specific for the detection of cytokines in the abdominal cavity and is not involved in the detection of cytokines in the circulation (Sehic 1996). Results that suggest that the afferent vagus is involved in the initiation of cytokine-evoked visceral aversion behavior (Bret-
33
Dibat 1995) are controversial. Schwartz et al. (1997) have shown that cytokine-induced visceral aversion behavior is not altered by specific afferent vagotomy. These authors concluded that cytokine-induced behavioral aversions must be produced as a consequence of direct action in the CNS.
The cellular mechanisms by which TNF rapidly changes NST excitability are not known. Two types of TNF receptors (TNFR) have been identified, the P75 TNFR and the P55 TNFR (Carlson 1998, Haviv 1998).
Although the neuronal distributions of these receptors have not yet been characterized (Pan 1997), the highest density of TNF-binding sites is in the medullary brain stem (Kinouchi 1991). Studies of P55 transduction mechanisms in other cells show that this receptor can rapidly activate (within
1-2 min) a sphingomyelinase pathway that produces ceramide (Hannun 1994,
Perry 1998). Intracellular ceramides can, in turn, increase intracellular calcium levels and regulate protein kinase cascades, which can modulate properties of membrane ion channels (Furukawa 1998), and may be involved in the augmented response to afferent stimulation. However, very little is presently known about the specifics of how cytokines can rapidly alter the excitability of neurons.
In summary, this study shows that NST neurons are strongly activated by TNF. Sensory afferent activation of NST neurons produces gastroinhibition as a consequence of their action on vagal motor neurons.
Therefore, our study suggests that centrally acting TNF probably produces
34
gastroinhibition by activating NST neurons. Thus in states of elevated circulating levels of TNF, sensory activation of NST neurons may more readily evoke gastroinhibition.
35
CHAPTER 3
TNF inhibits physiologically identified dorsal motor nucleus neurons in vivo
INTRODUCTION
Activated macrophages and lymphocytes release cytokines as part of the immune response to infection. These peptides can enter the central nervous system at specialized regions that lack the blood-brain barrier to evoke physiological changes associated with illness. One such region is the dorsal vagal complex in the medulla oblongata (Broadwell 1993, Gross 1990,
Shapiro 1985, Whitcomb 1990).
The DVC consists of the sensory nucleus of the solitary tract and the dorsal motor nucleus of the vagus. These nuclei comprise the final common pathway of the vago-vagal reflex circuits that control gastric motility (Rogers
1995, Rogers 1996). Elevated circulatory levels of the early cytokine tumor necrosis factor-alpha are correlated with nausea, vomiting, and gastrointestinal stasis (Cerami 1985, Hersch 1990, Kemeny 1992, McCann
1992, Patton 1987). Therefore, it is possible that TNF evokes changes in
DVC neuronal firing that translates to general gastrointestinal malaise.
Our previous studies (Hermann 1999) have demonstrated that IV administration of the bacterial cell wall component, lipopolysaccharide (LPS)
36
potently suppresses gastric motility and that this effect is critically dependent on the production of TNF. That is, if TNF production was blocked, then gastric motility was not suppressed by intravenous LPS. Our earlier study demonstrated that TNF injected unilaterally into the DVC abolished a centrally stimulated vagally dependent increase in gastric motility in a dose dependent manner (Hermann 1995). Given the rapidity of the centrally injected TNF effect on gastric motility, i.e., within 30 seconds of application to the DVC, this suggested that TNF could be directly and rapidly affecting the firing rate of neurons in the DVC. Indeed, our subsequent study (Emch 2000; see chapter
1) showed that gastric-distention related NST neurons are strongly excited by subfemtomolar doses of TNF.
In the present study, DMN neurons that form the motor portions of the gastric accommodation reflex were identified using neurophysiological methods described previously (Emch 2000, McCann 1992, Zhang 1992).
Most vago-vagal control of gastric motility and tone is inhibitory. That is, activation of vagal afferents by distention of the stomach, intestine, or esophagus results in a marked reduction in gastric motility and tone (McCann
1992, Zhang 1992). Activated vagal afferents excite NST neurons, which in turn inhibit DMN neurons that are the penultimate source of tonic cholinergic activation of the stomach. Removal of this tonic cholinergic drive reduces gastric motility and tone.
Because of the magnitude and duration of gastric motility suppression induced by injection of TNF into the DVC (Hermann 1995) and the fact that
37
NST neurons are strongly activated by TNF (Emch 2000), we hypothesized that this peptide may also strongly inhibit DMN neurons identified as part of the accommodation reflex. In this way, TNF would suppress gastric motility and tone by mimicking the effects of the activation of gastrointestinal afferents.
METHODS
Chemicals
The long acting anesthetic, Inactin (RBI), was dissolved to a concentration of 100 mg/ml in saline solution (0.9% NaCl) and administered at a dose of 100 mg/kg body weight, ip. A 2 M NaCl solution was used to record extracellular neuronal potentials. The recording pipette also contained
2% Neurobiotin (Vector) to iontophoretically mark the recording site.
Phosphate buffered saline solution was used as a vehicle injection control.
Recombinant rat TNF (R&D Systems) was dissolved in PBS to a concentration of 10-6 M and was divided into 25-ml aliquots and stored at –70o
C until use. Stored aliquots of TNF were further diluted with PBS such that the microinjection electrode contained 10-7 M, 10-8 M, or 10-9 M concentrations of TNF.
Pipette construction
Triple barrel micropipettes were constructed as described elsewhere
(Emch 2000). One pipette was utilized for recording neuronal action potentials while the other two were available for drug delivery. The recording
38
pipette was filled with 2M NaCl and 2% Neurobiotin. The other pipettes were
filled with either PBS or one of the three concentrations of TNF in PBS. The
electrode array was placed in a stereotaxic electrode carrier, which was
oriented at an approximate 20o rostral angle. The injection pipettes were
connected to a micropressure injection apparatus (Chen 1997). Extracellular
neuronal activity was recorded via a silver-silver chloride wire placed in the
recording pipette.
Data Collection
Extracellular neuronal potentials from the recording electrode were
amplified (10,000 x) and band-pass filtered (300 – 10,000Hz). Signals were
then displayed on an oscilloscope and recorded on a FM-VCR. Extracellular
spike potentials were also processed on-line by a window discriminator-rate meter circuit. The resulting neuronal firing rate along with the gastric distention stimulation signal were displayed on a chart recorder. Figures were generated by processing the recorded data with an IBM-PC based RC
Electronics waveform analysis system.
Gastric stimulation
A small gastric stimulation balloon was constructed from the tip of a latex surgical glove finger and attached to a 6-inch section of 0.065-inch O.D. silastic tubing. The balloon catheter was connected to one port of the dome of a Statham P23 pressure transducer. The balloon was fully inflated by injecting 3 cc of air into the other port with a 5-cc syringe (Emch 2000,
McCann 1992).
39
Surgical Preparations
Male Long Evans rats (Charles River) weighing 200-600 g were
provided with food and water ad libitum and kept on an approximate 12 hour
day-12 hour night cycle. The animal was anesthetized with Inactin, and the
trachea was cannulated in order to maintain an open airway. A laparotomy
was performed, and the balloon was inserted through a small incision in the
greater curvature of the stomach. After positioning the balloon in the antrum,
the gastric incision was sutured closed. The laparotomy incision was then
closed, the animal was placed in a stereotaxic frame, and a craniotomy was
performed to expose the dorsal medullary surface. Placement of the gastric
stimulation balloon and exposure of the dorsal surface of the brainstem were
performed as described previously (Emch 2000). All experimental protocols
were carried out according to guidelines set forth by the National Institutes of
Health and were approved by the Ohio State University Institutional
Laboratory Animal Care and Use Committee.
Experimental Design
The pipette array was advanced into the brainstem in order to identify neurons of the DMN. Initial stereotaxic coordinates were: 0.3 mm anterior to the calamus scriptorum and 0.3 - 0.5 mm lateral to the presumptive midline of the area postrema (McCann 1992). A hydraulic micro-drive (David Kopf
Instruments) was used to advance the array at 10-micron increments. At each electrode advancement through the medullary brainstem, the gastric balloon
40
was momentarily inflated and deflated. Neurons responding to gastric
distention were located at 550 – 900 microns below the brainstem surface.
Basal activity of DMN neurons typically demonstrated a very regular
rhythm (McCann 1992, Travagli 1991). Our criterion for identification of DMN
neurons is a minimum 50% reduction in firing frequency during the period of
gastric balloon distention. Typically, however, DMN neurons respond to
gastric distention by stopping firing completely. This classification of gastric
distention-sensitive dorsal vagal complex neurons has been shown to be
highly reliable [McCann 1992, Zhang 1992).
Once a cell was identified as a gastric related DMN neuron, PBS (3nL)
was micropressure injected from the attached pipette, and the spontaneous
activity was monitored for 1-2 minutes. These nano-injection methods are
described in detail elsewhere (Emch 2000). If neuronal activity was altered
by PBS microinjection (as a consequence of non-specific volume effects), it was rejected from further consideration. If PBS injection had no effects on the firing rate (FR) of the neuron, one of the three concentrations of TNF (3nl) was injected using the same pressure and duration settings as those for the preceding PBS injection. The FR was then observed for several minutes after injection to determine if the neuron was responsive to TNF. A cell was considered “responsive” if its FR was altered by at least 50% [Chen 1997,
McCann 1992).
41
Statistical Analysis
The effect of TNF on DMN firing was examined by comparing peak FR
post TNF injection with basal FR. FR was compared at one minute epochs at
the time of PBS microinjection, 1 and 2 minutes post-PBS injection, at the
time of TNF injection, and 1, 2, 3, and 4 minutes post TNF injection using a
repeated measures one-way analysis of variance followed by Dunnett’s post
test. Statistical significance was set at P < 0.05.
RESULTS
A total of 45 DMN neurons (Fig.3) were neurophysiologically identified
and exposed to TNF. To eliminate the possibility of tachyphylaxis, each
neuron was exposed to only one concentration of TNF. Most of the
responsive DMN cells were located at approximately 600 mm below the
brainstem surface. Of the 45 identified DMN cells, 36 were inhibited by TNF,
5 were activated, and 4 did not respond to TNF.
A dose of 0.03 fmoles (i. e., 3nl of 10-8M = 0.03 fmoles) of TNF
inhibited 34 identified neurons (Fig.4), while activating 5 identified neurons.
At this dose, DMN cells were completely inhibited for 8.3 + 6.4 minutes (mean
+ SEM). If the dose of TNF was increased 10-fold (0.3 fmoles), DMN neuronal firing was suppressed and these cells did not recover from inhibition
42
Figure 3. TNF inhibits gastric DMN neurons. A, Raw spike record shows a reduction in neuronal activity (upper trace) during a gastric antral balloon distention (lower trace). Inset: 10 superimposed oscilloscope traces corresponding to neuronal spikes in A. B, Expanded rate-meter record of the spike activity of DMN neuron exposed to 3nl of PBS, followed by a delayed application of TNF (0.03 fmoles). C, TNF causes a strong suppression of neuronal firing within 30 seconds of administration. D, Inhibition lasted for an average of 8.3 + 6.4 minutes (mean + S.E.M.) before neurons returned to basal FR. uV = microvolts, PBS = phosphate buffered saline, pps = pulses per second, sec = seconds, TNF = tumor necrosis factor
43
Figure 4. Graphic representation of the effect of TNF on DMN firing rate. Each column represents the FR in pulses per second (pps) sampled at 1 minute epochs following either PBS or TNF microinjection. At time = 0 minutes, PBS was microinjected into the DMN, and the resultant change in FR was recorded. PBS microinjection did not affect DMN neuronal firing at each time sampled (time = 0, 1 minute post-PBS, or 2 minutes post-PBS). TNF strongly inhibited DMN firing at 1, 2, 3, and 4 minutes post-TNF injection as compared to basal FR (Dunnett’s post test; *P<0.01). FR = firing rate, pps = pulses per second, PBS = phosphate buffered saline, TNF = tumor necrosis factor, min = minutes
44
(N = 2). If the dose was decreased 10-fold (0.003 fmoles), there was no
significant effect of TNF on DMN neuronal firing (N = 4). This tight
relationship between dose of TNF and DMN neuronal response is similar to
that documented in previous electrophysiological studies (Emch 2000).
DISCUSSION
The results of the present study demonstrate that a substantial majority of (84%) DMN neurons that form the motor portion of the gastric accommodation reflex are rapidly and potently inhibited by TNF. In most cases, (38/45), this inhibition persisted for multiple minutes with most cells recovering from TNF inhibition. Only a few identified DMN cells (5/45) were activated by TNF. Our earlier gastric motility study demonstrated that injecting sub-femtomolar amounts of TNF into the DVC suppressed centrally
stimulated gastric motility in a dose-dependent manner (Hermann 1995). Our
more recent electrophysiological study demonstrated that identified NST
neurons are strongly and directly activated by TNF (Emch 2000). The
observation that TNF can affect both sensory and motor components of this
vago-vagal reflex circuit may account for the rapid and large reductions in
gastric motility observed in our previous studies (Hermann 1995). Taken
together, these studies strongly support the hypothesis that TNF has a direct
effect on DVC neuronal circuitry.
45
Given the pleotrophic nature of TNF action on cellular transduction mechanisms, it is difficult to predict how TNF acts to inhibit DMN neurons. It is possible, however, that TNF exerts its anti-motility effects on vago-vagal reflex circuitry by acting at multiple sites. One possibility is that TNF increases the efficiency of vagal afferent synaptic input to NST neurons, which, for the most part, cause synaptic inhibition of DMN neurons (Emch
2000, McCann 1992, Miolan 1978, Roman 1987, Zhang 1992). The resulting withdrawal of cholinergic input to the stomach reduces motility and tone.
Preliminary studies (Emch 2001 abs, Hermann 2001 abs) suggest that one type of TNF receptor (p55) is present on vagal afferent fibers in the solitary nucleus. However, one cannot rule out the possibility that TNF can cause changes in the cellular excitability of the DMN itself. Preliminary in vitro neurophysiological studies performed under synaptic blockade suggest that
TNF may also act directly on receptors on neurons in the DMN (Browning
2001).
The current study, in conjunction with our previous work (Emch 2000,
Hermann 1995), suggests that TNF suppresses gastric motility by acting on gastric vago-vagal reflex circuit elements in the medulla. These observations provided a mechanistic explanation of the gastric stasis, nausea, appetite reduction, and potential vomiting that is associated with a variety of pathophysiological conditions.
46
CHAPTER 4 c-Fos generation in the dorsal vagal complex after systemic endotoxin is not
dependent on the vagus nerve
INTRODUCTION
Cytokines are released by activated macrophages and lymphocytes as part of the immune response to antigenic challenge, injury, or irradiation.
Elevation of the early proinflammatory cytokine TNF in the systemic circulation has been correlated with anorexia, nausea, vomiting, and gastrointestinal stasis (Cerami 1985, Kapas 1992, Kemeny 1990). This correlation between elevated plasma cytokine levels and changes in physiological state associated with illness implies a communication between the immune and nervous systems. Recent evidence suggests that the dorsal vagal complex (DVC) in the medulla oblongata may be one locus for TNF action to control gastrointestinal function (Emch 2000, Hermann 1995, 1999).
The DVC consists of the sensory nucleus of the solitary tract, the dorsal motor nucleus of the vagus, and the area postrema. These nuclei comprise the final common pathway of the vago-vagal reflex circuits that control gastric motility (Rogers 1995, 1996). This medullary brainstem area
47
has been identified as possessing the characteristics of a circumventricular
organ and is essentially devoid of a blood-brain barrier (Broadwell 1993,
Gross 1990, Whitcomb 1990). In addition, previous anatomical work (Rogers
1999, Rogers 1993, Shapiro 1985) demonstrated that dendritic endings of
neurons in both the NST and the DMN penetrate the area postrema and the
floor of the fourth ventricle. These anatomical characteristics place the DVC
in a position to monitor blood-borne and CSF-borne factors and to change vagally mediated autonomic functions accordingly (Chen 1995, McTigue
1997, Rogers 1995, 1996). The brainstem has a high density of TNF binding sites (Kinouchi 1991) and is in a position to monitor blood-borne peptides.
Therefore, it was hypothesized that the DVC may be the site of circulating
TNF action to provoke gastric stasis and the other prodromata of illness such as nausea and emesis.
Endogenous production of TNF can be readily elicited by systemic administration of the bacterial cell coat component, lipopolysaccharide (LPS)
(Waage 1987). Our previous studies (Hermann 1999) have demonstrated that
TNF production in response to systemic (i.e., intravenous) LPS is sufficient to suppress centrally-stimulated increases in gastric motility. Our earlier study demonstrated that TNF injected unilaterally into the DVC abolished a centrally stimulated and vagally dependent increase in gastric motility in a dose dependent manner (Hermann 1995). The rapidity of the centrally injected
TNF effect on gastric motility, i.e., within 30 seconds of application to the
DVC, suggested that TNF could directly and rapidly affect the firing rate of
48
neurons in the DVC. Electrophysiological studies by Emch et al. (2000, see chapter 1) have shown that neurons of the NST which form the sensory limb of a vago-vagal gastroinhibitory reflex (Rogers 1999) are strongly activated by doses of TNF previously shown to effect gastroinhibition (Hermann 1995).
Work by Sehic and Blatteis (Blatteis 1998, Sehic 1996) and others
(Fleshner 1998, Gaykema 1998, Goehler 1999, Milligan 1997) describe a potential alternate pathway by which information about immune activation may be transmitted to the CNS. There is evidence suggesting that vagal afferents, especially those in the hepatic branch, contain receptive elements responsive to cytokine or complement levels (Blatteis 1998, Simmons 1998).
The mechanism implied is similar to that responsible for the integrative physiological and behavioral actions of cholecystokinin (CCK). Here, CCK, released by duodenal enterocytes activates vagal afferent fibers that, in turn, produce a suppression of gastric motility as well as food intake (Ritter 1994,
Schwartz 1997). The hypothesis had been made that systemic levels of cytokines are monitored by vagal afferents in the periphery and their activation is responsible for illness behaviors and physiologic responses such as fever and gastrointestinal malaise (Fleshner 1998, Gaykema 1998,
Goehler 1999, Milligan 1997). However, the role for vagal afferents in the transmission of information about peripheral cytokine release provoking gastrointestinal malaise and other illness behaviors has been called into question recently.
49
That is, elimination of vagal afferent pathways does not block the suppression
of food intake, malaise inducing, somnogenic, or febrile effects of cytokines
(Caldwell 1999, Kapas 1998, Porter 1998, Schwartz 1997).
However, it is possible that both the vagal afferent and direct NST
mechanisms operate in parallel to monitor the portal circulation and the
general systemic circulation, respectively. This hypothesis is supported by
some reports that CNS effects (i.e., fever, illness behavior, etc.) of either low
dose intravenous or intraperitoneal LPS may be blunted by vagotomy, while
the same CNS effects following high dose intravenous LPS administration are
not blocked by vagotomy (Caldwell 1999, Kapas 1998, Romanovsky 1998).
We decided to test the hypothesis that vagal pathways are important to
the CNS signaling of peripheral cytokine production by using the generation
of the protein product of the proto-oncogene c-Fos as an anatomical identification of functionally activated neurons (Rinaman 1993) in the nucleus of the solitary tract. Previous studies (Elmquist 1996, Tkacs 1997) have shown that LPS-induced cytokine generation produces a significant increase in c-Fos labeling of neurons in the NST, i.e., the neurons in the medulla which receive vagal afferent projections. However, this earlier work did not establish whether the NST neurons were activated directly by circulating cytokine action or by vagal afferent pathways. It should be noted, however, that studies by Gaykema (1998) showed that subdiaphragmatic vagotomy abolished c-Fos expression in vagal sensory ganglia after intraperitoneal administration of LPS but only attenuated c-Fos expression in these nuclei
50
when LPS was administered intravenously. These results suggest that
different or redundant pathways are employed to inform the CNS about
peripheral levels of cytokines.
The majority of the hepatic vagal afferents (i.e., the most likely
peripheral afferent target for cytokines – Simons 1998) and the usual
complement of general visceral afferents from the thorax and abdomen
ascend via the left cervical vagal trunk (Berthoud 1992, Rogers 1983).
Therefore, we propose the following hypotheses: A) If intact vagal pathways
are critical to the transmission of information to the NST concerning
peripheral cytokine generation, then section of the left cervical vagal trunk should eliminate c-Fos generation in the NST on the side of section. B) If
direct action of cytokines at the NST is primary, then vagal transection should
have no effect on NST c-Fos generation and the numbers of c-Fos labeled nuclei on both sides of the NST should be comparable. C) If vagal afferent inputs modulate NST neurons that are, themselves, sensitive to levels of cytokines, then vagal section will reduce but not eliminate c-Fos generation by the NST.
These studies were designed to investigate three specific questions.
First, it was necessary to establish that Inactin anesthesia would not interfere with c-Fos-activation of NST neurons in response to endotoxin challenge and subsequent cytokine production. The second set of experiments was designed to determine whether transection of the left cervical vagal trunk
(with its predominance of hepatic afferent components), would eliminate c-
51
Fos-activation in the NST indicating that vagal afferents were the critical conduits signaling the CNS of peripheral immune activation. The third set of experiments were performed on bilateral, cervical vagotomized rats, to determine if loss of all vagal afferents (and efferents) would prevent c-Fos- activation of NST neurons in response to endotoxin challenge via either intravenous or intraperitoneal routes.
METHODS
Chemicals
Rats were anesthetized with Inactin (100mg/ml; 100mg/kg, ip) dissolved in saline. This thiobutabarbitol compound has been shown not to interfere with brainstem autonomic reflexes or with the generation of cytokines following the administration of LPS (Buelke-Sam 1978, Kotanido
1996). Endogenous production of TNF was induced by systemic administration of lipopolysaccharide. LPS was derived from Escherichia coli serotype 0111:B4 (Sigma; Waage 1987) and suspended in phosphate buffered saline (PBS, pH 7.4).
Histological processing of the medullary brainstem for c-Fos production required: primary c-Fos antibody (Oncogene Science Diagnostics, Inc.,
Cambridge, MA; AB-5; rabbit c-Fos, 1:20000) and biotinylated goat, anti- rabbit IgG (Vector Labs Inc., Burlingame, CA, 1:600). Amplification of
52
antibody-antigen reactions required incubation with Vector elite avidin - biotin
-peroxidase complex (Vector Labs, Inc.; 1:600 in PBS) followed by Vector SG
peroxidase detection reagents (Vector Labs, Inc.).
Experimental design
Experiment one was designed to establish that thiobutabarbitol,
Inactin, anesthesia would not interfere with c-Fos-activation of NST neurons
in response to endotoxin challenge and subsequent cytokine production.
Studies were performed in Inactin anesthetized, vagally intact rats that were
exposed to equivalent intravenous volumes (0.1ml/100 gm b.w.) of either PBS
(n=4) or LPS (1000ug/kg b.w.; n=6).
Experiment two was designed to determine whether transection of the
left cervical vagal trunk (with its predominance of hepatic afferent
components), would eliminate c-Fos-activation in the NST. These studies
were performed in unilateral (i.e., left), cervical vagotomized rats that were exposed to one of four drug groups: (a) PBS (n=4); (b) 1000ug/kg LPS dose
(n=6); (c) 100ug/kg LPS dose (n=6); or (d) 25ug/kg LPS dose (n=6).
Experiment three was designed to determine if loss of all vagal
afferents (and efferents) would prevent c-Fos-activation of NST neurons in
response to endotoxin challenge via either intravenous or intraperitoneal
routes. These studies were performed in bilateral cervical vagotomized rats
that were exposed to one of the following: (a) intravenous PBS (n=4); (b)
intravenous 25ug/kg LPS (n=6); or (c) intraperitoneal 25ug/kg LPS (n= 5).
53
Animals
Male Long-Evans rats (Charles River) were maintained in a temperature controlled vivarium with a 12hr day:night cycle. Animals had ad libitum access to food and water. All experimental procedures were performed according to guidelines set forth by the National Institutes of Health and were approved by the Ohio State University Institutional Laboratory
Animal Care and Use Committee.
Surgical preparations
Rats were anesthetized with Inactin. All subjects received tracheal cannulae to ensure the maintenance of an open airway for the duration of the experiment. Animals assigned to intravenous studies were equipped with sterile jugular cannulae. Depending on the assigned vagal status, each animal received one of three surgical manipulations: a) exposure of cervical vagi without sectioning of the vagal trunks, i.e., intact (n = 10); b) left cervical vagal trunk section (n = 22); or c) bilateral cervical trunk section (n=15). In the rats receiving bilateral cervical vagotomies, the left vagus was cut approximately 20 minutes prior to the section of right vagus. Although the rats developed apneustic breathing after the section of the remaining vagus, all survived without auxiliary ventilation (Rogers 1979).
All three experiments used systemic administration of LPS to induce endogenous cytokine production. Our “high” dose of LPS (1000ug/kg b.w.) has been shown to be sufficient to induce substantial TNF secretion
(Kotanidou 1996) and to produce a significant gastric stasis under similar
54
anesthetic conditions (Hermann 1999). This dose also produces a modest but consistent hypotension. Our “intermediate dose” (100ug/kg b.w.) has been shown to elicit anorexic effects (Porter 1998, Schwartz 1997) or fever
(Milligan 1997) in awake rats. Lastly, our “low dose” (25 ug/kg) has been shown to effectively elevate plasma TNF levels (Givalois 1994), produce fever, elicit c-Fos expression in the central amygdala, but does not provoke hypotension (Tkacs 1997). The effects (or lack thereof) of LPS on blood pressure at these doses (i.e., 25, 100, or 1000ug/kg) were verified under
Inactin anesthesia in our preliminary studies (data not shown).
Endogenous TNF production reaches maximal plasma levels within 90 minutes of systemic administration of LPS (Waage 1987). Studies by
Rinaman (1993) have demonstrated that maximal nuclear c-Fos immunoreactivity is present approximately 60-90 minutes following the presence of the presumptive stimulus. Therefore, comparable to other studies on c-Fos activation within the CNS following systemic exposure to endotoxin (Sagar 1995, Tkacs 1997), survival time prior to perfusion was selected to be three hours after systemic injections of either PBS or LPS to maximize c-Fos activation expression. At the end of three hours, rats were given a 0.2 ml bolus intravenous injection of lidocaine to stop respiration and cardiac function. The chest cavity was opened and a blood sample by ventricular puncture was taken for subsequent enzyme-linked immunosorbent assay (ELISA) verification of TNF production. Animals were then
55
transcardially perfused with PBS followed by 4% paraformaldehyde in PBS.
The brainstems were then removed to a solution of 4% paraformaldehyde
and 20% sucrose in PBS to post fix for 16 hours.
Histological processing for c-Fos protein
Brainstems were sectioned on a freezing microtome at 50 microns
thickness; sections were collected in PBS. After rinsing in PBS, sections
were treated with1% sodium borohydride to reduce the fixative remaining in
the tissue. After rinsing in PBS, tissue sections were incubated for 1 hour on
a shaker in 10% normal sheep serum plus 0.3% triton-X in PBS to block non-
specific binding of the primary c-Fos antibody. After rinsing, tissue sections
were incubated in primary c-Fos antibody (Oncogene AB-5; rabbit c-Fos,
1:10000) in 0.3% triton in PBS for 16 hours at room temperature with gentle agitation. Tissue sections were rinsed and incubated in biotinylated goat, anti- rabbit IgG (Vector, 1:600) for 1 hour. Sections were rinsed and reacted with
Vector elite avidin - biotin -peroxidase complex (1:600 in PBS) for one hour, followed by Vector SG peroxidase detection reagents. Specificity of the c-Fos immunocytochemical reaction was verified by omitting the c-Fos antibody from randomly selected sections. Sections were rinsed, mounted on glass slides, dried, cleared in Hemo-De (Fisher Sci., Pittsburgh, PA) and coverslipped with Entellan (Electron Microscopy Sciences, Fort Washington,
PA).
56
Counting c-Fos nuclei in the dorsal medulla
c-Fos labeled nuclei were counted manually with the aid of an MD2
Microscope Digitizer (Minnesota Datametrics Corp., St. Paul, MN) encoder
attached to the stage of a Leitz Dialux Microscope. Inclusion of c-Fos-labeled
neurons required that the nuclei be at least six microns in diameter and had to
exhibit a nucleolus. These criteria guaranteed that staining artifacts and
nuclear fragments would not be included in the count. c-Fos stained nuclei were counted without knowledge of the experimental condition and counts were verified by a second observer. The agreement between counts of the two observers was within 5%. c-Fos-activation of medullary neurons was analyzed at four specific coronal levels for each animal: 0.5 mm posterior to the calamus scriptorum, the level of the calamus, the level of the area postrema (0.5 mm anterior to calamus) and the level of the anterior NST (1.0 mm anterior to calamus). The cumulated number of activated cells was totaled for the right and left side of each animal’s brain for statistical analysis.
The distribution of labeled nuclei from the NST, area postrema, and DMN regions were analyzed separately.
TNF assay
Plasma TNF was determined by an enzyme-linked immunosorbent assay (ELISA) for rat TNF (R & D Systems). Fifty microliter plasma samples, in duplicate, were incubated at room temperature in microwells pre-coated with monoclonal anti-rat TNF antibody. After a 2-hr incubation, each well was
57
aspirated and washed with Wash Buffer; this process was repeated four times. One hundred microliters of antibody against rat TNF conjugated to horseradish peroxidase were added to each well. After a 2-hr incubation, each well was aspirated and washed with Wash Buffer; this process was repeated four times. One hundred microliters of a tetramethylbenzidine peroxidase substrate were added to all wells. After a 30 minute incubation at room temperature, the reaction was stopped by addition of hydrochloric acid.
The optical absorbance of each well was read within 30 minutes using a microplate reader set to 450nm. Absorbance values were converted to TNF concentrations by comparison with a simultaneously generated standard curve. The limits of detection per well of this assay kit were 12.5-800 pg/ml; the inter- and intra-assay variabilities were 8.8 and 2.1%, respectively
(manufacturer’s data).
Analysis
Although animals were randomly assigned to one of the three surgical groups and experiments were run simultaneously, the analyses of c-Fos data were segregated according to surgical condition, i.e., (a) intact vagi, (b) left cervical vagotomy, and (c) bilateral cervical vagotomy, for two reasons. Most importantly, these three surgical manipulations result in rats with very different physiological states (e.g., apneustic versus normal breathing) that may be reflected in basal (i.e., PBS challenge) conditions of c-Fos activation of DVC neurons. Secondly, these three experiments were designed to address different aspects of the DVC response to systemic endotoxin challenge.
58
Experiment 1: Intact Vagi / Inactin (thiobutabarbitol) anesthesia
Previous studies of CNS c-Fos production following LPS have been performed in unanesthetized rodents (Elmquist 1996, Sagar 1995, Tkacs
1997). Studies have shown that different anesthesia types may affect the immune response to LPS challenge, e.g., the reduction of TNF production under urethan anesthesia (Hermann 1999, Kotanidou 1996). Therefore, it was necessary to first establish that vagally intact, Inactin anesthetized rats were capable of inducing c-Fos synthesis in response to intravenous LPS challenge.
Secondly, recent data suggest that the significant cytokine sensitive sensory pathway to the CNS arises from hepatic vagal afferents (Simons
1998) which are represented asymmetrically within the two cervical vagal trunks and the NST. That is, the large majority of hepatic vagal afferents travel in the left cervical vagus and terminate in the left medial NST (Berthoud
1992, Rogers 1983). Therefore, the second aim of this experiment was to determine whether there was any intrinsic left versus right “sidedness” to the distribution of c-Fos in the brainstem of rats with both vagi intact.
The cumulated number of activated cells (i.e., total number of c-Fos- activated neurons from the four coronal sections) was totaled for the right and left side of each animal’s brain within either the NST or DMN. The NST and
DMN c-Fos count results were independently subjected to Student’s t-tests.
Area postrema c-Fos counts were obtained from the single sample section that contained this midline structure. Given that left versus right
59
“sidedness” was not an issue with this structure, c-Fos-activated cell counts of
PBS versus LPS challenged rats were analyzed using a Student’s t-test.
Experiment 2: Left Cervical Vagotomy
Animals received left cervical vagotomy to eliminate vagal connections with half of the brainstem (i.e., the half that may receive a physiologically significant hepatic afferent projection). These unilaterally vagotomized rats were challenged with one of three doses of intravenous LPS or PBS. The cumulated number of activated cells (i.e., total number of c-Fos-activated neurons from the four coronal sections) was totaled for the right and left side of each animal’s brain within either the NST or DMN. The NST and DMN c-
Fos count results were independently subjected to a repeated measures analysis of variance (i.e., left and right sides from the same animal; PBS vs
LPS groups; Motulsky 1995). In the event of a significant p-value (i.e., p<0.05), Dunnett’s post tests were used. Given that the AP is a midline structure, c-Fos-activated cell counts in the AP of PBS versus LPS challenged rats were analyzed by one-way ANOVA.
Experiment 3: Bilateral Cervical Vagotomy
It could be argued that any c-Fos label observed in the dorsal medulla ipsilateral to the unilateral vagotomy might be attributed to afferent activity from the remaining intact vagal trunk. Therefore, in this experiment, rats received bilateral cervical vagotomies to totally eliminate vagal connections with the NST. These animals received either intravenous PBS or our lowest
LPS dose (25 ug/kg, b.w.) via either intravenous or intraperitoneal routes. As
60
in the previous experiment, the cumulated number of activated cells (i.e., total number of c-Fos-activated neurons from the four coronal sections) was totaled for the right and left side of each animal’s brain within either the NST or DMN. The NST and DMN c-Fos count results were independently subjected to a repeated measures ANOVA (i.e., left and right sides from the same animal; PBS vs LPS groups; Motulsky 1995). c-Fos-activated cell counts of the AP of PBS versus LPS challenged rats were analyzed by one- way ANOVA.
Plasma TNF levels
Plasma TNF levels were analyzed according to the systemic challenge
(i.e., PBS or different doses of LPS; n = 35) by using the Kruskal-Wallis test for non-parametric samples. Statistical significance was defined as an overall
P<0.05; Dunn’s multiple comparison post-tests were applied.
RESULTS
Experiment 1: c-Fos labeling in the NST, DMN, and Area Postrema in the vagus-intact, Inactin anesthetized rat
Systemic LPS challenge induces a significant rise in NST c-Fos labeling in vagus-intact, Inactin anesthetized rats (t=11.3, df=8, p<0.0001;
Figs. 5A and 6). This elevation in NST c-Fos count following intravenous LPS is symmetrical in intact, non-vagotomized rats; i.e., there is no intrinsic difference in distribution (“sidedness”) of NST neurons c-Fos-activated in response to systemic endotoxin challenge (p = 0.3223). If hepatic afferents
61
Figure 5. Experiment 1: c-Fos-activated neurons in the brainstem of vagally intact, Inactin anesthetized rats that received either intravenous PBS or LPS.
A) Intravenous LPS induced a significant (p<0.0001) elevation in c-Fos- activated NST neurons in Inactin anesthetized, vagus - intact rats. The distribution of c-Fos activated neurons within the NST did not show any intrinsic “sidedness” (p=0.32), i.e., the distribution was symmetric within the NST. B) LPS exposure also resulted in significant increases in the number of c-Fos positive neurons in the DMN (p<0.01), but the absolute numbers are very small. C) LPS significantly increased (p<0.005) c-Fos labeling in the AP.
62
Figure 6. Micrographs of original coronal histological sections through the NST at the level of the area postrema. c-Fos production in response to systemic (iv) challenge of either PBS (A) or 1000ug/kg LPS (B) is demonstrated by the dark staining nucleoli. LPS evokes a substantial, bilaterally symmetrical, increase in c-Fos production in the AP, NST, and DMN in the Inactin anesthetized rat. Scale bar = 0.5mm.
AP = area postrema; CC = central canal; DMN = dorsal motor nucleus of the vagus; mNST = medial portion of the nucleus of the solitary tract; st = solitary tract; 12 = hypoglossal nucleus
63
(which ascend predominantly within the left cervical vagal trunk to terminate
within the left NST) were the principal pathway by which systemic exposure to
endotoxin provoked c-Fos activation of brainstem neurons, then one might
have expected more NST neurons to be labeled on the left as opposed to the
right side.
Systemic LPS challenge produced a small but significant increase in
the numbers of DMN neurons containing c-Fos nuclear staining (t = 3.612, df
= 8; p = 0.0069; Fig. 5B and 6). Again, there is no difference in numbers of c-
Fos labeled nuclei between right and left sides of the brainstem.
c-Fos labeling of the area postrema was also significantly increased by
LPS challenge in the vagus-intact rat ([PBS] 15.3 + 7.3 neurons versus [LPS]
106.0 + 18.1 neurons; t = 4.658; df = 6; p = 0.0035; Fig. 5C and 6).
Experiment 2: Effects of LPS on c-Fos labeling in the NST, DMN, and
Area Postrema in rats with left cervical vagotomy
Animals with left cervical vagotomy demonstrated a significant
increase in c-Fos activation of NST neurons, regardless of dose of
intravenous LPS or side of brainstem sampled (Fig 7A and 8; F = 9.71; df =
3,17; p = 0.006; Dunnett’s post-test, p < 0.05). In these unilaterally vagotomized preparations, the number of c-Fos activated neurons showed a small but consistent difference in the pattern of distribution (i.e., left vs right).
That is, across all groups, including the PBS controls, the right (i.e., non-
vagotomized) side of each brainstem contained somewhat more c-Fos
activated NST cells than the corresponding left side (F = 40.909, df = 1,17, p
64
Figure 7. Experiment 2: c-Fos-activated neurons in the brainstem of left, cervical, vagotomized that received either intravenous PBS or LPS. A) Intravenous LPS challenges demonstrated a significant increase in the number of c-Fos labeled nuclei in the NST. This increase in c-Fos activated cells was seen regardless of dose of LPS administered or side of brainstem sampled (F = 9.71, df = 3,17, p = 0.0006; Dunnett’s post hoc * = p<0.05). In these unilaterally vagotomized preparations, the number of c-Fos activated neurons showed a small but consistent difference in the pattern of distribution (i.e., left vs right); this effect was even observed in the PBS control group. (F = 40.9, df = 1,17; p = 0.0001). B) LPS exposure also resulted in increases in the number of cFos positive neurons in the DMN, but the absolute numbers are, comparatively, very small and not statistically significant (F = 2.206; df = 3,17; p = 0.1298). C) All doses of LPS significantly increased cFos labeling in the AP (F = 4.634; df = 3,16; p = 0.0162; * Dunnett’s post-test p<0.05).
65
Figure 8. Micrographs of c-Fos production in response to systemic (iv) challenge of either PBS (A), 25 ug/kg LPS (B), 100ug/kg LPS (C), or 1000ug/kg LPS production (D) in the dorsal vagal complex of rats with left cervical vagotomy. Transection of the left cervical vagal trunk resulted in a subtle, but consistent, reduction in number of c-Fos activated neurons in the left NST of all groups tested (i.e., even the PBS controls). Scale bar = 0.5mm.
66
= 0.0001). This “sidedness” was statistically significant as a consequence of the consistency of this observation across all groups as opposed to the actual magnitude of the difference within any particular group. Indeed, the effect of vagotomy was not significant at any individual dose.
The number of c-Fos labeled cells in the DMN was increased with LPS challenge. However, due to the overall small number of DMN cells activated, this increase was not statistically significant (Fig. 7B and 8; F = 2.206, df =
3,17, p = 0.1298).
There was significant increase in number of c-Fos labeling of neurons in the area postrema in response to all doses of LPS challenge (Fig. 7C and
8; F = 4.634, df = 3,16, p = 0.0162; Dunnett’s post test p < 0.05).
Experiment 3: Effects of LPS on c-Fos labeling in the NST, DMN, and
Area Postrema in rats with bilateral cervical vagotomy
Animals with bilateral cervical vagotomy (i.e., devoid of vagal connections with the CNS) still demonstrated a highly significant elevation in the numbers of c-Fos - labeled neurons in the NST. This response to our low dose (25ug/kg, b.w.) of LPS was evident whether delivered via either the intravenous or intraperitoneal route (Fig. 9A and 10; F = 15.24, df = 2,12, p =
0.0005; Dunnett’s post test *p < 0.05).
The number of DMN neurons showing c-Fos activation following endotoxin challenge (via either iv or ip routes) was quite comparable to the numbers seen in either the unilaterally vagotomized or vagally intact groups.
67
Figure 9. Experiment 3: c-Fos-activated neurons in the brainstem of bilateral, cervical, vagotomized that received PBS or LPS (iv or ip). A) Animals with bilateral cervical vagotomy demonstrated a significant increase in c-Fos labeled neurons in the NST in response to 25ug/kg LPS; regardless of the route of administration (F = 15.24, df = 2,12, p = 0.0005; * = Dunnett’s post test p < 0.05). Similar to the observations in the vagally intact group, there was no “sidedness” in the distribution of c-Fos activated neurons in the NST. B) Although LPS exposure also resulted in increases in the number of c-Fos positive neurons in the DMN comparable in number to those seen in the intact and unilaterally vagotomized groups, this response was not statistically significant (p = 0.7876). C) LPS significantly increased c-Fos labeling in the AP; however, only the LPS (IV) group was statistically significant relative to the PBS group (F = 5.108; df = 2,10; p = 0.0296; * Dunnett’s post-test p<0.05). 68
Figure 10. Micrographs of cFos production in the dorsal vagal complex of rats with bilateral cervical vagotomy. Upper photo is an example of cFos production in response to intravenous PBS. Middle photo is example of response to 25ug/kg LPS (iv); bottom photo is an example of response to 25 ug/kg LPS (ip). Bilaterally vagotomized rats are still capable of responding with an increase in cFos production in neurons of the DVC following systemic (either iv or ip) challenge with endotoxin. Scale bar = 0.5mm
69
However, due to the relatively high number of c-Fos activated DMN neurons
in the PBS group, this response is not significant (Figs. 9B and 10; p =
0.7876).
Lastly, the number of c-Fos activated cells in the area postrema
following endotoxin challenge was significantly increased in the bilateral,
cervical vagotomized groups (Figs. 9C and 10; F = 5.108, df = 2,10, p =
0.0296). Dunnett’s post test revealed that only the LPS (iv) group was
statistically significant relative to the PBS group.
Plasma TNF levels
Blood samples were obtained at approximately 180 minutes post-
injection of either PBS or one of the different doses of LPS, i.e., immediately
prior to transcardial perfusion for histological processing. All three doses of
LPS (25, 100, or 1000 ug/kg b.w.) elicited significant production of circulating
TNF-a in Inactin anesthetized rats regardless of route of administration (i.e.,
intravenous or intraperitoneal) or integrity of the vagal nerve trunks (i.e.,
intact, unilateral, or bilateral vagotomy). (Fig. 11; n = 35; Kruskal-Wallis test p
= 0.0001; Dunn’s post test p<0.05).
Although there was not a significant difference between the amounts of
TNF elicited by the different doses of LPS across the various groups, there was a correlation (Spearman r = 0.6908; p <0.0001) between levels of plasma
TNF and number of c-Fos activated neurons in the NST.
70
Figure 11. ELISA determination of plasma TNF levels. All three doses of LPS (25, 100, or 1000 ug/kg b.w.) elicited significant production of circulating TNF in Inactin anesthetized rats regardless of route of administration (i.e., intravenous or intraperitoneal) or integrity of the vagal nerve trunks (i.e., intact, unilateral, or bilateral vagotomy). Kruskal-Wallis test p = 0.0001; * = Dunnett’s post test p < 0.05.
71
DISCUSSION
These studies demonstrated that, even in Inactin anesthetized rats,
neurons in the NST and AP demonstrate a significant increase in c-Fos
nuclear protein labeling following systemic challenge of endotoxin (i.e., via
either intravenous or intraperitoneal routes) regardless of the integrity of the
vagus nerves (i.e., intact, unilateral, or bilateral, cervical vagotomy). This
increase occurs at both hypotensive and non-hypotensive doses of LPS, suggesting that the c-Fos labeling is a primary effect of LPS-induced cytokine production on the nervous system, as opposed to a primary effect on peripheral vasculature which is signaled by baroreceptive afferents (Curtis
1999).
Additionally, from the unilateral vagotomy experiments, our data indicate that the vagus nerve exerts a subtle effect on the responsiveness of the NST to any afferent information. That is, in the unilateral vagotomized groups (Experiment 2), c-Fos labeling was modestly but consistently depressed on the vagotomized side of the brainstem (i.e., left side) at all doses of LPS as well as in the control, PBS, group. Although one cannot rule out the possibility that cutting the vagus removes an important pathway regarding information concerning peripheral cytokine levels (Blatteis 1998,
Gaykema 1998, Simons 1998), it is more likely that this uniform reduction in c-Fos labeling in the NST is due to removal of a significant source of general vagal afferent excitation. That is, general visceral afferent input to the NST is glutamatergic and this tonic glutamatergic input is responsible for a majority of
72
the tonic excitation that NST neurons receive (Rogers 1995, Rogers 1996).
Therefore, removal of vagal inputs to the NST, regardless of the afferent modality, will reduce the excitability of NST neurons (Palkovits 1995).
Nevertheless, the effects of LPS exposure were certainly detectable by the
NST regardless of the connectivity of vagal afferents.
Our studies employed unilateral and bilateral cervical vagotomies (i.e., caudal to the nodose ganglia). One might propose that nodose neurons may still detect LPS-related signals and communicate them to the CNS through their intact synaptic inputs to the DVC. However, studies by Gaykema (1998) have shown that subdiaphragmatic vagotomy (i.e., even more caudal vagotomies) abolished c-Fos expression in vagal sensory ganglia after intraperitoneal administration of LPS and attenuated c-Fos expression in these nuclei when LPS was administered intravenously. Furthermore, studies disrupting vagal afferents rostral to the nodose (Porter 1998, Schwartz 1997) does not block the anorexia produced by peripheral LPS. These results suggest that different or redundant pathways are employed to inform the CNS about peripheral levels of cytokines.
c-Fos labeling of cells in the area postrema parallels that of the NST. It is possible that the area postrema is the principal CNS detector of cytokines elicited by the LPS challenge. That is, NST labeling may only be a consequence of excitatory inputs from the area postrema (van der Kooy
1983). However, morphological and physiological studies do not support the concept of the area postrema acting only as a specialized chemosensor with
73
the NST acting only as a processor of general visceral afferent information
from the vagus. Rather, the NST and area postrema share several
morphological and functional features: (a) both nuclei receive primary vagal
afferent inputs (Rogers 1983), (b) both nuclei are vascularized by fenestrated
capillaries (Gross 1990), and (c) dendrites from the NST (and DMN) are
intermingled within the area postrema (Shapiro 1985). It is likely that the c-
Fos labeled NST and AP neurons share a common sensitivity to cytokines
and that area postrema neurons modulate NST and DMN excitability.
Data from our plasma TNF assay support the conclusion that these
brainstem neurons are activated by cytokines elicited by the systemic LPS
challenge in that there is a positive correlation between the number of c-Fos labeled neurons and the amount of TNF produced. Furthermore, neurophysiological studies in our laboratory (Emch 2000) show that NST neurons responsible for coordinating reflex inhibition of gastric function
(Rogers 1993, McCann 1992, Zhang 1992) are directly activated by sub- femtomole doses of the early cytokine TNF. Parallel studies (Hermann 1995) have also shown that sub-femtomole doses of TNF delivered to the DVC area containing these NST neurons produce a profound reduction in gastric motility. Together, these data support the view that neurons in the dorsal vagal complex, particularly the NST, are intrinsically sensitive to the effects of
TNF. One physiological result of NST activation would be gastroinhibition
(Rogers 1995, Rogers 1996).
74
TNF suppresses gastric motility as part of the constellation of signs and symptoms of the illness behavior of inflammatory disease. These results provide us with a tentative CNS mechanism for this effect. TNF released as part of the cytokine cascade probably gains access to the dorsal vagal complex through fenestrated capillaries. It is now well known that the excitability of neurons in this important autonomic integrative zone can be controlled by circulating peptides, as well as by vagal afferent and descending
CNS afferent influences (Rogers 1995, 1996, Whitcomb 1990). The NST is in position to directly transduce this “hormonal” signal into changes in excitability, though it is extremely likely that NST activity (perhaps, also the
DMN) is indirectly modulated by chemosensor neurons in the area postrema which, in turn, synapse on NST or DMN neurons (van der Kooy 1983).
75
CHAPTER 5
TNF induced c-Fos generation in the nucleus of the solitary tract is blocked by
NBQX and MK-801
INTRODUCTION
Gastric-distention related neurons of the nucleus of the solitary tract are strongly excited by subfemtomolar doses of the cytokine tumor necrosis factor-alpha (Emch 2000). Additionally, after exposure to TNF, these neurons exhibit potentiated responsiveness to subsequent afferent stimulation. Given that vagal afferent connections with NST neurons are predominantly glutamatergic (McCann 1990, Smith 1998, Torrealba 1991), it is possible that
TNF modulates responsiveness of NST neurons through a modification of glutamate neurotransmission.
Much evidence suggests that glutamate is the primary neurotransmitter released by vagal sensory neurons at the level of the NST (McCann 1990,
Smith 1998, Torrealba 1991). Microinjection of exogenous glutamate into this brainstem area evokes physiological changes such as cardiovascular responses, respiratory modulation, and esophageal contractions involved in swallowing (Rossiter 1990, Spencer 1986). Electrophysiological studies demonstrated that identified gastric-distention related NST neurons respond
76
similarly to iontophoretic application of glutamate or natural stimulation.
Responses to either stimulation could be blocked by the glutamate
antagonist, kynuretic acid (McCann 1990). Axon terminals that contact the
NST show glutamate immunoreactivity in all solitary subnuclei (Lin 1999,
Treece 1998). Immunohistochemical and pharmacological studies have
shown that the N-methyl-D-aspartate (NMDA) and alpha-amino-3-hydroxy-5-
methylisoxazole-4-poprionic acid (AMPA), are present in the various solitary
subnuclei, and each carries a discrete functional role (Ambalavanar 1998).
The neural circuitry of the dorsal vagal complex (DVC), made up of the
sensory NST and the dorsal motor nucleus of the vagus (DMN), is involved in
the control of physiological functions that may be affected during illness.
Cytokines elicit a variety of physiological changes associated with illness,
e.g., fever, nausea, and vomiting, through modulation of the central nervous
system (CNS). The route by which these peptides evoke changes in CNS
activity has been the subject of debate. Some investigators have proposed
that vagal afferent fibers propagate information about peripheral cytokine
generation centrally (Fleshner 1998, Gaykema 1998, Goehler 1999). Other
studies have implicated direct entry of cytokines at the DVC, since it exhibits
the characteristics of a circumventricular organ, i. e., fenestrated capillary
network and absence of functional blood-brain barrier (Caldwell 1999, Gill
1999, Hermann 2001, Houpt 1997, McCaughy 1997). Regardless of the route of activation of TNF in the DVC, the majority of vagal afferents utilize glutamate at the level of the NST (McCann 1990, Smith 1998). Therefore, the
77
possibility exists that cytokines, such as TNF, may modulate the efficacy of glutamate-mediated activity of neurons in this area. Indeed, several studies have shown that cytokines can increase glutamate release in the NST (Li
2000, Malenka 1988) as well as alter the firing rate of NST neurons that are coupled to vagal afferent stimulation, i. e., gastric balloon distention (Emch
2000).
Our previous study utilized immunocytochemical identification of the immediate early gene product, c-Fos, as a marker for neuronal activation in the DVC (Hermann 2001). Peripheral administration (iv or ip) of LPS caused a significant increase in c-Fos in the DVC. The number of neurons within the
DVC that expressed c-Fos activation after peripheral administration of LPS correlated with plasma levels of TNF. This presumptive activation of DVC neurons did not require intact vagal pathways, suggesting that peripherally generated TNF-a acts directly on these neurons (Hermann 2001). Our earlier gastric motility study (Hermann 1995) revealed that direct nanoinjection of
TNF into the DVC inhibited centrally stimulated motility in a dose-dependent manner. Given that those studies were performed in animals pretreated with dexamethasone, production of additional cytokines or the initiation of the cytokine cascade in response to TNF was not a factor. Our electrophysiological studies (Emch 2000) have shown that microinjection of
TNF onto identified neurons of the NST can rapidly and directly activate these cells as well as modify their responsiveness to afferent stimulation. Taken
78
together, these studies strongly support the hypothesis that TNF directly
affects DVC neuronal circuitry.
The current study was designed to: 1.) definitively demonstrate that
TNF, and not some other cytokine, was responsible for c-Fos activation of
NST neurons, 2.) test the hypothesis that TNF activation of NST cells is
dependent on glutamate transmission, and 3.) determine if glutamate
antagonism retards the effectiveness of other agonists known to activate the
NST (Mascarruci 1998) and cause gastric relaxation (Rogers 1986). It is
possible that TNF acts presynaptically (e. g., increased glutamate release) or
that it modulates the postsynaptic action of glutamate. Either way, blockade
of glutamate receptors should result in less c-Fos expression than with
injections of TNF alone. Therefore, we examined the induction of the c-Fos
protein in the NST in response to nano-injections of TNF or in response to co- injection of TNF with the AMPA receptor antagonist, NBQX (Randle 1992) or the NMDA antagonist, MK-801 (Wong 1986).
Since practically all NST neurons receive glutamatergic inputs, it is
possible that blocking AMPA or NMDA receptors may produce a non-specific inhibition that affects the potency of any agonist operating on the NST. To
evaluate this possibility, we took advantage of the fact that oxytocin (OT)
directly activates NST neurons and also produces a potent gastroinhibition by
operating on vago-vagal reflex circuitry (Alberi 1997, McCann 1990,
Raggenbass 1998, Rogers 1986). If blockade of glutamate transmission with
79
NBQX or MK-801 has a uniform effect to reduce NST excitability, then the numbers of c-Fos expressing NST neurons will be reduced following either
TNF or OT.
METHODS
Animals
Forty-six male Long-Evans rats (Simonsen Labs, Gilroy, CA, USA) weighing 200-500 g were provided with food and water ad libitum and kept on an approximate 12 hour day-12 hour night cycle. All experimental protocols were performed according to guidelines set forth by the National Institutes of
Health and were approved by the Ohio State University Institutional
Laboratory Animal Care and Use Committee.
Chemicals
Animals were anesthetized with Inactin (Sigma-RBI) dissolved to a concentration of 100 mg/ml in saline solution and administered at a dose of
200 mg/kg body weight, ip. Phosphate buffered saline solution (PBS; 124 mM NaCl, 26 mM NaHCO3, 2 mM KH2PO4; 304 mosmol; pH = 7.4) was used as a solvent diluent as well as a vehicle injection control. Recombinant rat
TNF (R&D Systems) was dissolved in PBS to a concentration of 3 x 10-6 M, divided into 25-ml aliquots, and stored at –70o C until use. Oxytocin (Bachem
Bioscience Inc., King of Prussia, PA, USA) was dissolved in PBS to a concentration of 1mM, divided into 25-ml aliquots, frozen and stored. The concentration of OT used in this study was chosen since it was shown to 80
evoke changes in DVC neuronal firing (McCann 1990) and suppress gastric
motility (Rogers 1986) in our previous work. The AMPA glutamate receptor
antagonist, 1,2,3,4-Tetrahydro-6-nitro-2,3-dioxo-benzo[f]quinoxaline-7- sulonamide disodium (NBQX; Sigma-RBI) was diluted to a concentration of
1mM, aliquoted at 25-ml, and frozen until use. The NMDA glutamate receptor antagonist, (5R,10S)-(+)-5-Methyl-10,11-dihydro-5H-dibenzo[a,d]cyclo- hepten-5,10-imine hydrogen maleate (MK-801; Sigma-RBI) was reconstituted in PBS to a concentration of 1mM and was stored at 4o C until use. Stored aliquots of TNF were further diluted with PBS such that the injection pipette contained 10-7 M, 10-8 M, or 10-9 M TNF. These concentrations were chosen since they have been shown to excite single NST neurons and significantly suppress gastric motility in our previous studies (Emch 2000, Hermann 1995).
NBQX was further diluted to a final concentration of 0.1mM in the injection pipette (Li 2000). MK-801 was diluted such that the injection pipette contained a concentration of 0.67mM (Zheng 1999).
Histological processing of the medullary brainstem for c-Fos production required primary c-Fos antibody (Oncogene Science Diagnostics, Inc.,
Cambridge, MA; AB-5; rabbit c-Fos, 1:20000) and biotinylated goat, anti- rabbit IgG (Vector Labs Inc., Burlingame, CA, 1:600). Amplification of antibody-antigen reactions required incubation with Vector elite avidin - biotin
-peroxidase complex (ABC) (Vector Labs Inc.; 1:600 in PBS) followed by
Vector SG peroxidase detection reagents (Vector Labs Inc.).
81
Surgical Preparation
The animal was deeply anesthetized with Inactin, and the trachea was cannulated in order to maintain an open airway. The animal was placed in a stereotaxic frame, the occipital skull plate was removed, and the dura and arachnoid meninges were resected exposing the dorsal surface of the brainstem.
Injection pipettes were made from 1.5mm OD glass capillary tubes
(Radnoti Glass Technology, Inc., Monrovia, CA) pulled to a point with a
Narishinge PE2 puller and beveled with the tip opening approximately 20 – 30 microns in diameter. The injection pipette was advanced with a hydraulic micro-drive (David Kopf Instruments) into the NST at stereotaxic coordinates of: +0.4mm anterior to the calamus scriptorum, +0.4mm lateral to the midline posterior tip of the calamus, and 400 microns below the brainstem surface
(Emch 2000). All injections were unilateral and were made using micropressure application techniques as previously described (McCann
1990). All injections had a total volume of 20nl.
Maximal nuclear c-Fos activity occurs approximately 60 – 90 minutes following the presumptive stimulus (Rinaman 1993). Therefore, resultant c-
Fos activation was evaluated at 90 minutes post-injection. At this time, the animals were tested to ensure a deep plane of anesthesia. The chest cavity was quickly opened, and the animal was transcardially perfused with PBS followed by 4% paraformaldehyde in PBS. Brainstems were then removed and postfixed in 4% paraformaldehyde – 20% sucrose – PBS overnight.
82
Experimental Design
Experiment one was designed to investigate whether neurons in the
NST express c-Fos in response to nanoinjections of TNF directly into the
brainstem and whether that expression is dependent on the dose of TNF.
Experimental groups received unilateral nanoinjections of one of the following
conditions:
(a) PBS (n=5)
(b) 10-7 TNF = 2femtomoles = 34picograms (n=5)
(c) 10-8 TNF = 0.2fmoles = 3.4pg (n=6)
(d) 10-9 TNF = 0.02fmoles = 0.34pg (n=5).
These concentrations of TNF were chosen as a result of our previous studies
of the effects of the cytokine on dorsal vagal complex circuits controlling
gastric motility (Emch 2000).
Experiment two was designed to explore whether TNF induced
excitation of the NST is dependent on glutamate transmission. Experimental
groups received unilateral nanoinjections of one of the following conditions:
(a) TNF (3.4pg) plus NBQX (67pg) (n=5)
(b) TNF (3.4pg) plus MK-801 (4.5ng) (n=5)
Our previous neurophysiological study showed that TNF at 10-8 M caused the
“ideal” NST neuronal response to TNF in that all cells were activated, all cells recovered from TNF induced activation, and some cells exhibited potentiated responses to subsequent afferent stimulation after recovery from initial
83
activation (Emch 2000). Therefore, this dose of TNF was chosen for the
current study to be potentially blocked by the glutamate antagonists.
In experiment three, we were interested in investigating whether
glutamate neuromodulation is specific to TNF or whether glutamate may set
the background sensitivity of the NST to generalized effectors. Previous
studies have shown that OT directly excites neurons in the DVC (Alberi 1997,
McCann 1990, Raggenbass 1998) and that the DVC contains receptors
specific for OT (Alberi 1997). Experimental groups received unilateral
nanoinjections of one of the following conditions:
(a) OT = 20pmoles = 20ng (n=5)
(b) OT (20ng) plus NBQX (67pg) (n=5)
(c) OT (20ng) plus MK-801 (4.5ng) (n=5)
Histological processing
Three hypotheses were tested utilizing a total of nine experimental groups. Animals were assigned to the groups at random; therefore the c-Fos processing across groups was also performed at random. Controls were run throughout the duration of all three experiments. Brainstems were sectioned on a freezing microtome at 40 microns thickness, and sections were collected in PBS. Sections were treated with 1% sodium borohydride to reduce remaining fixative in the tissue. Tissue sections were incubated in 10% normal sheep serum plus 0.3% triton-X in PBS to block non-specific binding of the primary c-Fos antibody. The tissue was then incubated in primary c-
Fos antibody (Oncogene AB-5; rabbit c-Fos, 1:20000) plus triton-X in PBS for
84
17 hours at room temperature with gentle agitation. The following day, the
tissue was incubated in biotinylated goat, anti-rabbit IgG (Vector, 1:600) for 1
hour. Sections were then reacted with Vector elite ABC (1:600 in PBS) for
one hour followed by Vector SG peroxidase detection reagents. Specificity of
the c-Fos immunocytochemical reaction was verified by omitting the c-Fos
antibody from randomly selected sections. Sections were rinsed, mounted on
gelatin-coated glass slides, dried, cleared in Hemo-De (Fisher Sci.) and
coverslipped with Entellan (Electron Microscopy Sciences).
Counting c-Fos labeled cells
c-Fos labeled nuclei in the NST were counted manually with the aid of an MD2 Microscope Digitizer (Minnesota Datametrics Corp.) encoder attached to the stage of a Leitz Dialux Microscope. Inclusion of c-Fos-labeled neurons required that their stained nuclei be at least six microns in diameter and that they exhibit a nucleolus. These criteria guaranteed that staining artifacts and nuclear fragments would not be included in the count (Hermann
2001). c-Fos stained nuclei were counted without knowledge of the experimental condition and a second observer verified counts. The
agreement between counts of the two observers was within 5%. Counts were
made from one histological section that corresponded to site of the injection.
The site of injection was taken to be the 40-micron section that exhibited
maximal c-Fos labeled cells for a given animal.
85
Analysis
c-Fos counts among all nine experimental groups were analyzed using
a one-way analysis of variance (ANOVA). Statistical significance was
determined at p<0.05. Experiment one sought to establish a relationship
between dose of TNF microinjection and the resultant number of c-Fos
labeled neurons. Experiment two was designed to see if co-injection of the
glutamate receptor antagonists (NBQX or MK-801) would affect the NST c-
Fos expression evoked by TNF. Experiment three was performed to examine the effect of NBQX and MK-801 on c-Fos induction evoked by OT. Therefore, subsequent post hoc comparisons were made by selected Bonferroni tests between relevant groups. All data are presented as means + SEM.
RESULTS
Experiment 1: c-Fos labeling in the NST induced by TNF
Compared with results obtained in PBS injected rats, TNF induced a dramatic increase in c-Fos labeling when injected into the NST. Indeed, PBS injection did not appear to produce any increase in NST c-Fos labeling over what we have obtained in rats receiving no NST injections (Hermann 2001).
An overall ANOVA revealed a significant effect of treating experimental animals with the various protocols outlined in the current study (df = 9, P <
0.0001). Microinjection of TNF induced a significant rise in NST c-Fos labeling in rats subjected to a dose of 3.4pg TNF as compared to vehicle
86
Figure 12. Experiment 1: Effect of TNF on c-Fos expression in the NST. Graphic representation of c-Fos labeled NST cell counts expressed as mean + SEM. A dose of 3.4fgrams TNF induced c-Fos counts that were significantly different than those obtained from animals injected with PBS (*P<0.001, Selected Bonferroni’s test). Decreasing the dose 10-fold (0.34fg) did not produce statistically significant c-Fos labeling, whereas increasing the dose 10-fold (3.4fg) resulted in significant c-Fos expression in the NST (*P<0.001, Selected Bonferroni’s test).
87
Figure 13. Photomicrographs of coronal histological sections through the NST at the level of the area postrema. c-Fos labeled neurons are characterized by their dark stain, a minimal 6mm diameter, and the presence of nucleoli. A, In PBS injected rats (20nl), few c-Fos positive cells were present (ap = area postrema, nst = nucleus of the solitary tract, dmn = dorsal motor nucleus of the vagus, st = solitary tract). B, Injection of TNF (3.4fg) induces a significant increase in c-Fos expression. C, Co-injection of TNF with the AMPA receptor antagonist, NBQX results in little c-Fos production. D, Co-injection of TNF with the NMDA antagonist, MK-801 causes scarce c- Fos labeling. Scale bar = 700mm.
88
injection control (205.1 + 29.5 vs. 62.0 + 10.2; Selected Bonferroni’s multiple
comparison test, t = 4.647, P < 0.001; Figs. 12 and 13). Increasing the dose
to 34pg TNF also yielded a significant increase in c-Fos expression in the
NST above PBS (213.6 + 25.0; t = 4.522, P < 0.001; Fig. 12). Injection of
0.34pg TNF produced an increase in NST c-Fos expression that was not statistically significant (124.8 + 25.7; P > 0.05; Fig. 12). This relationship
parallels the results obtained in our electrophysiological studies (Emch 2000).
It is of interest to note that in all experimental groups, c-Fos expression was
not confined to the side of the brainstem in which injections were made
(right), although labeling was always most dense on this side.
Experiment 2: Effects of glutamate antagonists on NST c-Fos labeling in
response to co-injections with TNF.
When TNF (3.4pg) was co-injected with the AMPA glutamate receptor antagonist, NBQX, into the brainstem, NST c-Fos levels were significantly lower than when this dose of TNF was injected alone (79.0 + 10.8 vs. 213.6 +
25.0; t = 3.891, P < 0.01; Figs. 13 and 14). Co-injection of TNF with the
NMDA glutamate receptor antagonist, MK-801, also caused a significant
reduction in c-Fos expression in the NST compared to TNF alone (104.8 +
13.3 vs. 213.6 + 25.0; t = 3.256, P < 0.05; Figs. 13 and 14).
Experiment 3: Effects of glutamate antagonists on NST c-Fos labeling in
response to co-injections with OT
Microinjection of OT (20ng) into the medulla induced a significant rise
in c-Fos labeled cells in the NST as compared to PBS injected rats (171.5 +
89
Figure 14. Experiment 2: Graphic representation of the effect of glutamate antagonists on c-Fos cell counts (mean + SEM) induced by TNF. Microinjection of TNF (3.4fg) results in significant c-Fos labeling in the NST compared to PBS (Selected Bonferroni’s test, *P<0.001). Inclusion of the AMPA antagonist, NBQX, along with TNF in the injection pipette inhibited significant c-Fos production. Similarly, when the NMDA antagonist, MK-801 was injected with TNF, there was no significant increase in NST c-Fos expression.
90
Figure 15. Photomicrographs of c-Fos induction by microinjection of OT in the presence and absence of glutamate receptor antagonists. A, PBS microinjection (20nl) results in scarce c-Fos expression in the NST. (ap = area postrema, nst = nucleus of the solitary tract, dmn = dorsal motor nucleus of the vagus, st = solitary tract). B, OT injection induces significant c-Fos labeling in the NST. C, Co-injection of OT with the AMPA antagonist, NBQX does not affect c-Fos production induced by OT. D, Microinjection of OT and the NMDA antagonist, MK-801 generates c-Fos expression that is not different than that produced by OT alone. Scale bar = 700mm.
91
Figure 16. Experiment 3: Effect of glutamate receptor antagonists on c-Fos expression in the NST induced by OT. Each bar represents c-Fos positive cell counts expressed as mean + SEM. Microinjection of 20ng of OT into the NST induces a significant increase in c-Fos labeled cells compared to control (*P<0.05, Selected Bonferroni’s test). Injection of either NBQX or MK-801 with OT did not significantly inhibit c-Fos production in the NST.
92
16.7 vs. 62.0 + 10.2; t = 3.064, P < 0.05; Figs. 15 and 16). In contrast to our
results in experiment 2, co-injection of OT with either NBQX (241.3 + 55.7) or
MK-801 (163.8 + 36.7) did not result in significantly different c-Fos counts
than in rats treated with OT alone (171.5 + 16.7) (P > 0.05; Figs. 15 and 16).
Again, c-Fos labeled cells were not restricted to the side of the brainstem in which injections were performed in any experimental group.
DISCUSSION
These studies demonstrated that nano-injections of TNF into the brainstem increased the expression of the immediate early gene c-Fos in the
NST. The results also suggest that NST activation (as estimated by c-Fos production) is dependent on glutamate neurotransmission, since inclusion of the AMPA receptor antagonist, NBQX or the NMDA antagonist, MK-801 along with TNF in the injection pipette suppressed the increase in c-Fos production induced by TNF alone. Lastly, glutamate does not appear to set the general
background excitability for NST activation. That is, neither NBQX nor MK-801
diminished c-Fos expression evoked by OT.
The effect of concentration of TNF on NST neurons is consistent with
the results of our electrophysiological studies (Emch 2000). In this study,
concentrations of 10-8 M and 10-7 M induced excitation of identified NST neurons, whereas a concentration of 10-9 M TNF failed to activate these cells
(Emch 2000). In the current study, the same concentrations of TNF (10-8 M
93
and 10-7 M) induced significant production of c-Fos in the NST, whereas the
lowest dose (10-9 M) produced a modest but non-statistically significant increase in c-Fos expression. Although labeling was most dense on the side
of the injection, c-Fos positive cells were present on both sides of the
brainstem. This is not surprising given the intrinsic projections between
solitary nuclei (Whitehead 2000), as well as the possibility of drug diffusion.
The data presented in this study reveal that both AMPA and NMDA receptors probably play a role in TNF activation of NST neurons, since antagonists for both receptors block c-Fos activation in the NST in response to TNF. An earlier study by Wan et al. (1993) described that MK-801 inhibited c-Fos production in the NST in response to peripheral injections of LPS.
However, it is still unknown how TNF modulates glutamatergic transmission.
It is possible that TNF causes an increase in glutamate release presynaptically. Using dialysis techniques, Mascarucci et al. (1998) demonstrated short-term enhancement (within 60 minutes) of glutamate release in the NST in response to intraperitoneal injections of LPS or interleukin-1b. Glutamate release in the LPS preparation was simultaneous with elevated plasma levels of TNF (Mascarucci 1998). Glutamate levels remain elevated in the NST for up to 3 hours after iv LPS injection (Lin 1999).
TNF may also modulate the postsynaptic action of glutamate at the level of the NST. This type of relationship exists with corticotropin-releasing factor (CRF) in the cerebellum. Here, CRF amplifies both spontaneous and glutamate-induced single-unit activity of Purkinje cells (Bishop 1990, 1992).
94
Other investigators have shown that NMDA and AMPA receptors in the
DVC play a role in gastrointestinal function. Sivarao et al. (1999)
demonstrated that both receptors are involved in modulating intragastric
pressure and motility. NMDA receptors play a role in satiety since direct
injection of MK-801 into the NST (as well as fourth ventricular infusion) delay
satiety (Treece 1998, Zheng 1999). One of these studies presented
conflicting data on c-Fos expression in response to gastric distention.
Although distention produced significant increases in c-Fos production in the
medial NST, fourth ventricular infusion of MK-801 did not decrease this protein expression (Zheng 1999). In addition, injection of MK-801 increased
c-Fos labeling in the NST in the absence of gastric distention (Zheng 1999).
One may expect that MK-801 would have opposite effect on c-Fos production in this experimental paradigm. Perhaps ventricular infusion of MK-801 causes inhibition of glutamatergic transmission elsewhere in the CNS, thus removing the NST from tonic inhibitory influences. This disinhibition would, in turn, lead to an excitation of NST neurons, thus an increase in c-Fos positive cells. In the current study, MK-801 inhibited TNF induced c-Fos production in the NST and not OT induced c-Fos production. Perhaps this is because local injections of the antagonist do not allow for more widespread inhibition of glutamate neurotransmission.
The results presented here suggest that the relationship between TNF and glutamate may be specific. That is, blocking of glutamatergic transmission in the NST blunted TNF activation of NST neurons, while having
95
no effect on OT activation of NST neurons. It has been demonstrated that
OT-ergic neurons are present in the DVC and that there are receptors specific
to OT in this region (Alberi 1997, Rinaman 1998). Other work has shown that
OT directly excites NST neurons in vitro and in vivo (McCann 1990,
Raggenbass 1998). The current study corroborates these findings by
showing that injection of OT into the NST increases c-Fos expression. This
OT-induced c-Fos activation was not dependent on glutamate receptors since
inclusion of either NBQX or MK-801 with OT did not decrease the number of c-Fos positive cells in the NST.
Our previous study revealed that TNF directly excites NST neurons that are responsive to gastric distention (Emch 2000). Furthermore, these
NST neurons that were pre-exposed to TNF exhibited a potentiated response to afferent stimulation. The current study shows that NMDA (and AMPA) receptors are involved in TNF activation of NST neurons. Perhaps the facilitation recorded in the neurophysiological study is similar to other known mechanisms of synaptic plasticity, e. g., long-term potentiation (LTP). NMDA receptors are required for LTP generation in the CA1 region of the hippocampus, (the area in which LTP is most often studied), since MK-801 and other NMDA antagonists inhibit LTP (Bliss 1993). In this system, the
AMPA receptor is initially responsible for depolarization of the cell. Once a threshold voltage is reached, the NMDA receptor is removed from inhibition and calcium is allowed to enter the cell, which is required for LTP induction
(Malenka 1988). In our neurophysiological preparation, AMPA receptor
96
activation was probably involved in the fast response to TNF, whereas NMDA receptors were probably involved in sustained activation, and perhaps potentiation. Additional experiments are required to investigate these hypotheses further.
97
Chapter 6
PERSPECTIVES
Summary
The experiments outlined in this dissertation imply a direct effect of
TNF on brainstem circuits that exert reflex control over the gastrointestinal tract. Several previous studies have shown that TNF evokes gastroinhibition when injected into the DVC and when peripheral production of the cytokine was induced. However these studies did not definitively show a direct effect of TNF on the neurons in the NST and the DMN. To examine this possibility, several experimental designs were utilized and described in the previous chapters. The results of these studies have shown that peripheral (and possibly central) sources of TNF act directly on neurons in the brainstem to inhibit cholinergic drive to the stomach. Direct injection of TNF into the DVC activates neurons in the nucleus of the solitary tract that were previously identified as being part of the accommodation reflex. Some NST neurons that were previously exposed to TNF exhibited a facilitated response to afferent stimulation. Similarly, direct injection of TNF inhibits previously identified neurons in the dorsal motor nucleus of the vagus. According to these data, it
98 is implied that TNF mimics the effects of activating vagal afferent receptors in the stomach wall that would translate to gastroinhibition in the proximal gut.
Gastroinhibition is an obligatory prodrome of emesis and is commonly perceived as nausea, and DVC neuronal circuitry is involved the CNS motor program of emesis (Rogers 95, 96). Therefore, it is plausible to assume that
TNF’s effects on the DVC are at least partly responsible for the nausea, vomiting, and gastrointestinal malaise seen during illness. The fact that TNF can affect both sensory and motor components of the accommodation reflex lends support to the effectiveness of this cytokine in generating these symptoms of illness. Stimulation of endogenous production of TNF, through peripheral injection of the endotoxin, is sufficient to suppress gastrointestinal motility (Hermann). LPS injection also causes activation of single neurons in the solitary nucleus, as shown through c-Fos immunocytochemical methods.
Therefore, according to these data, peripheral TNF can gain access to the
DVC parenchyma, thus influencing single unit neuronal activity and producing gastroinhibition.
The studies outlined in chapters 2, 3, and 4 show that TNF affects single neurons in the DVC. However, little is presently known about how TNF specifically activates neurons in the NST. Therefore, the experiments in chapter 5 were performed. The results show that some type of neuromodulatory relationship exists between TNF and glutamate in the DVC.
Microinjection of TNF into the medial NST causes a significant rise in c-Fos labeled cells. Administering glutamate receptor antagonists along with TNF
99
abolishes this Fos expression. This relationship between TNF and glutamate
seems to be specific since Fos expression induced by injection of another
NST agonist, oxytocin, is unaffected by co-injection with the glutamate antagonists. It is unknown whether TNF causes an increase in glutamate release presynaptically or whether TNF causes a postsynaptic increase in neuronal excitability.
These c-Fos immunocytochemical studies suggest that NMDA (and
AMPA) receptors are involved in TNF mediated c-Fos production in the NST.
Our previous work has shown that NST responsiveness to afferent input is increased after exposure to TNF. A large number of circulating agents can act in the DVC to produce nausea, vomiting, and a long-lasting aversion to food (Rogers 95, 96). This is an interesting observation given that cytokines have been shown to produce conditioned visceral aversion behavior (Bret-
Dibat 1995). Perhaps the potentiating effect of TNF on NST neurons is critical
to the production of such long-term changes in the responsiveness to visceral afferent input that may change behavior.
The NST sends long ascending connections with forebrain structures that are involved in the integration of autonomic, neuroendocrine, and behavioral responses to specific stimuli, such as the central nucleus of the amygdala (Sawchenko). The amygdaloid nucleus, in turn, projects to the insular cortex, a temporal lobe area that has been implicated in conditioned taste aversion (CTA) behavior (Bermudez-Rattoni 1991, Keifer 1990). CTA is a model of learning and memory in which an animal acquires aversion to a
100
novel taste when it is followed by digestive malaise (Escobar 1998). NMDA
receptors are required for CTA production as well as LTP induction in the
insular cortex (Escobar 1998). More recently, induction of LTP in amygdaloid
projections to the insular cortex was directly linked to CTA behavior (Escobar
2000).
The NST has indeed been linked to CTA in several experimental
paradigms. McCaughey et al. (1997) showed that a burst of activity occurs in
the NST in response to presentation of the conditioned stimulus to CTA
conditioned rats. Houpt et al. (1997) demonstrated c-Fos induction in the NST after CTA acquisition, and protein expression was not dependent on the integrity of the vagus nerve(s). Our data suggest a mechanism of enhanced responsiveness to afferent input after exposure to TNF. Potentiation in the
NST may be translated to the amygdala, which is then transmitted to the insular cortex for UP induction and subsequent CTA acquisition. Such plasticity exhibited at multiple levels in the CNS would efficiently establish visceral aversion behavior in response to infection.
Future Directions
The studies outlined in this dissertation have provided insight into the role of TNF in vago-vagal control of gastrointestinal function. However, these studies have left room for several new experiments.
101
Extracellular recordings
The present electrophysiological studies have shown that TNF directly
influences neurons that are involved in both the afferent (NST) and efferent
(DMN) limbs of the accommodation reflex. Additionally, some NST neurons
were sensitized to further afferent stimulation. Therefore, endogenous
production of TNF may also cause facilitated electrophysiological responses
to afferent stimulation. Here, a gastric stimulation balloon is placed in the
stomach, and the brainstem is exposed as outlined in earlier chapters. LPS is
utilized to generate TNF in the circulation. Following 90 minutes, a single
barrel microelectrode is used to record extracellular potentials from neurons
in the NST. Signals are recorded and displayed as described in previous
chapters. In this preparation, the pre-exposure of NST neurons to circulating
TNF may set a background for increased excitability in which gastric inflation causes neurons to fire at sustained elevated rates. This would add physiological relevance to the possibility that TNF causes potentiation in the
NST that is transferred to CNS areas associated with visceral aversion behavior, i. e. the central amygdala and the insular cortex. It would also be of interest to record from these areas to see if potentiation exists there following
LPS treatment. c-Fos immunocytochemistry
The present c-Fos immunocytochemical studies have shown that both centrally injected and peripherally produced TNF activates neurons in the
DVC. However, the phenotype of the cells being activated is unknown.
102
Accordingly, double-labeling immunocytochemistry experiments can be
performed. To do this, the same doses of TNF can be microinjected into the
DVC of Inactin anesthetized rats. After 90 minutes, animals are transcardially
perfused with PBS followed by 4% paraformaldehyde. Brainstems are
removed and postfixed overnight in a 4% paraformaldehyde-20% sucrose
solution. Brainstems are then processed as outlined in chapters 4 and 5.
However, following exposure of the tissue to the chromagen and appearance
of c-Fos labeled cells, the tissue is processed for the presence of various
neurotransmitters. Again tissue is treated with a blocking serum before
exposure to the primary antibody. Candidate antibodies include those for
tyrosine hydroxylase, GABA, and neuronal-nitric oxide synthase. Tissue
sections are incubated overnight and then rinsed and subjected to an
appropriate secondary antibody. Sections are then rinsed and exposed to the
Vector ABC solution for magnification of the signal. Final detection of labeled
cells is carried out by using the Vector Nova Red solution. In this way, c-Fos labeled nuclei are stained black, while cytoplasmic domains are stained red for the phenotype of the neuron.
Stained cells are then counted using the same apparatus and methods that were described in chapters 4 and 5. Briefly, cells designated as “c-Fos positive” must display nucleoli and must have nuclei that are at least 6 microns in diameter. The total number of cells that are c-Fos positive is counted as well as the total number of phenotypically stained cells is counted.
Also, The total number of double-labeled cells is counted. In this way, the
103
percentage of TNF activated cells with a specific phenotype can be obtained
from the raw counts as well as the percentage of cells of a particular
phenotype that is responsive to TNF. Various statistical analyses can then be
performed from these data.
In vivo microdialysis
The current immunocytochemical studies have shown that co-injection of glutamate antagonists along with TNF into the NST suppresses c-Fos activation that is induced with injection of TNF alone. However, it is not known how glutamate neurotransmission is specifically affected by TNF. It is possible that the change in excitability is presynaptic, i.e., through increased glutamate release, postsynaptic, or both. To investigate the possibility of a presynaptic change, in vivo microdialysis studies can be performed.
In vivo microdialysis allows for infusion of a solution containing drugs, neurotransmitters, or other peptides with simultaneous collection of neurotransmitters that are released in response. Using this experimental paradigm TNF can be infused into the NST, and the resulting dialysate solution that is collected can be analyzed for glutamate. Long Evans rats are anesthetized with Inactin, and the brainstem is exposed as in all previous work. 1-mm length microdialysis probes with a membrane diameter of 100 microns are inserted into the NST at a 45o-rostral angle so that most of the membrane surface can lie in the NST. After inserting the probe into the brain tissue artificial cerebrospinal fluid is constantly infused into the brainstem.
Dialysates are not collected from the probe for the first 90 minutes of infusion
104
to allow for a tissue equilibration period. After 90 minutes of baseline
collections, TNF is constantly transfused into the inlet of the probe. In this
way, samples can be analyzed for a baseline level of glutamate versus the
level that is present during and after TNF is infused into the brainstem.
Samples are collected for another 120 minutes to allow for full diffusion of
TNF through the probe as well as to allow for any response to TNF infusion.
Dialysates are collected into vials appropriate for high-pressure liquid
chromatography analysis. Samples are collected at set intervals, e.g., every
fifteen minutes and at a constant flow rate, e.g., 0.5mm/minute. Samples can
then be processed and analyzed the content of glutamate.
TNF rapidly activates NST neurons when microinjected into the DVC.
If TNF does this through an increase in glutamate release presynaptically, then a spike in glutamate release should be observed shortly after TNF infusion into the brainstem. This can be investigated by comparing samples collected from the dialysis probe during baseline with samples collected after infusion of TNF.
The experiments outlined above would provide further insight into the role of TNF in affecting vago-vagal reflex control of digestion. Hopefully, they would reveal more specifically how a cytokine, such as TNF, can cause a
change in neuronal excitability by acting directly on neurons in the DVC that
control gastrointestinal motility. Perhaps these experiments would cause
further, more targeted inquiries into TNF signal transduction in the brain.
105 BIBLIOGRAPHY
1. Abrahamsson, H. Vagal relaxation of the stomach induced from the gastric antrum. Acta. Physiol. Scand. 89: 406-414, 1973.
2. Abrahamsson, H. and H. Glise. Sympathetic nervous control of gastric motility and interaction with vagal activity. Scand. J. Gastroenterol. 19: Suppl. 89: 83-87, 1984.
3. Abrahamsson, H. and G. Jansson. Vago-vagal gastro-gastric relaxation in the cat. Acta Physiol. Scand. 88: 289-295, 1973.
4. Adachi, A., M. Kobashi, and M .Funahashi. Glucose sensors in the medulla oblongata. Japanese J. Physiol. 47(Suppl 1): S37-8, 1997.
5. Alberi S, Dreifuss JJ, and Raggenbass M. The oxytocin-induced inward current in vagal neurons of the rat is mediated by G protein activation but not by an increase in the intracellular calcium concentration. Eur J Neurosci 9: 2605-2612, 1997.
6. Altschuler, S. M., J. Eseardo, R. B. Lynn, and R. R. Miselis. The central organization of the vagus nerve innervating the colon of the rat. Gastroenterology 104: 502-509, 1993.
7. Ambalavanar R, Ludlow CL, Wenthold RJ, Tanaka Y, Damirjian M, and Petralia RS. Glutamate receptor subunits in the nucleus of the tractus solitarius and other regions of the medulla oblongata in the cat. J Comp Neurol 402: 75-92, 1998.
8. Barger, S. W.; D. Horster, K. Furukawa, Y. Goodman, J. Kriegistein, and M. P. Matson. Tumor necrosis factor alpha and beta-peptide toxicity: evidence for involvement of a kappa B-binding factor and attenuation of peroxide and Ca2+ accumulation. Proc. Natl. Acad. Sci. USA 92: 9328-9332, 1995.
9. Barna, B. P., G. H. Barnett, B. S. Jacobs, and M. L. Estes. Divergent responses of human astrocytoma and non-neoplastic astrocytes to tumor 106
10. necrosis factor alpha involve the 55 kDa tumor necrosis factor receptor. J. Neuroimmunol. 43:185-190, 1993.
11. Barna, B. P., M. L. Estes, B. S. Jacobs, S. Hudson, R. M. Ranscohoff. Human astrocytes proliferate in response to tumor necrosis factor alpha. J. Neuroimmunol. 30: 239-243, 1990.
12. Bebo, B. F. Jr. and D. S. Linthicurn. Expression of mRNA for 55-kDa and 75-kDa tumor necrosis factor (TNF) receptors in mouse cerebrovascular endothelium: effects of interleukin-Ib, interferon-g, and TNFa on cultured cells. J. Neuroimmunol. 62: 161-167,1995.
13. Bermudez-Rattoni F. and McGaugh JL. Insular cortex and amygdala lesions differentially affect acquisition of inhibitory avoidance and conditioned taste aversion. Brain Res 549, 165-170, 1991.
14. Beyaert, R. and W. Fiers. Molecular mechanisms of tumor necrosis factor- induced cytotoxicity. What we do understand and what we do not. Fedn. Eur. Biochem. Socs. Lett. 340: 9-16, 1994.
15. Beckstead, R. M., J. R. Morse, and R. Norgren. The nucleus of the solitary tract in the monkey: projections to the thalamus and brain stem nuclei. J Comp. Neurol. 190: 259-282, 1980.
16. Berthoud, H. R., N. R. Carlson, and T. L. Powley. Topography of efferent vagal innervation of the rat gastrointestinal tract. Am. J Physiol. 260(29): R200-R207,1991.
17. Berthoud, H. R., A. Jedrzejewska, and T. L. Powley. Simultaneous labeling of vagal innervation of the gut and afferent projections from the visceral forebrain with Dil injected into the dorsal vagal complex in the rat. J Comp. Neurol. 65-79, 1990.
18. Berthoud, H. R., M. Kressel, & W. L. Neuhuber. An anterograde tracing study of the vagal innervation of rat liver, portal vein, and biliary system. Anat. Embryol. 186: 431-442, 1992.
19. Berthoud, H. R. and T. L. Powley. Morphology and distribution of efferent vagal innervation of rat pancreas as revealed with anterograde transport of Dil. Brain Res. 553: 336-341, 1991.
20. Bishop GA. Neuromodulatory effects of coricotropin releasing factor on cerebellar purkinje cells: an in vivo study in the cat. Neuroscience 39: 251- 257, 1990.
107
21. Bishop GA and JS King. Differential modulation of purkinje cell activity by enkephalin and corticotropin releasing factor. Neuropeptides 22: 167-174, 1992.
22. Blatteis, CM, E Sehic, and S Li. Afferent pathways of pyrogen signaling. Ann. NY Acad Sci. 856: 95-107, 1998.
23. Bliss TVP and Collingridge GL. A synaptic model of memory: long-term potentiation in the hippocampus. Nature 361: 31-39, 1993.
24. Bret-Dibat, J. L., R. M. Bluthe, S. Kent, R. W. Kelley, & R. Dantzer. Lipopolysaccharide and interleukin-1 depress food-motivated behavior in mice by a vagal-mediated mechanism. Brain Behav. Immun. 9: 242-246, 1995.
25. Boka, G., P. Anglade, D. Wallach, and F. Javoy-Agrid. Immunocytochemical analysis of tumor necrosis factor and its receptors in Parkinson’s disease. Neurosci. Lett. 172: 151-154, 1994.
26. Broadwell, R. D. and M. V. Sofroniew. Serum proteins bypass the blood- brain barriers for extracellular entry into the central nervous system. Exp. Neurol. 120: 245-263, 1993.
27. Browning, K. N., R. A. Travagli, and R. C. Rogers. Effects of tumor necrosis factor alpha (TNF-a) on rat dorsal motor nucleus of the vagus (DMV) neurons, Neurosci. Abs. 31: 634.10, 2001.
28. Buelke-Sam, J, JF Holson, JJ Bazare, and JF Young. Comparative stability of physiological parameters during sustained anesthesia in rats. Lab. Anim. Sci. 28:157-162, 1978.
29. Bystrzycka, E. K. and B. S. Nail. Brain stem nuclei associated with respiratory, cardiovascular and other autonomic functions. In: The Rat Nervous System: Hindbrain and Spinal Cord, edited by G. Paxinos. Sydney: Academic Press, Inc. 1985, p. 95-110.
30. Caldwell, FT Jr., DB Graves, and BH Wallace. Humoral vs. neural pathways for fever production in rats after administration of lipopolysaccharide. J. Trauma 47: 120-129, 1999.
31. Carlson, N. G., A. Bacchi, S. W. Rogers, & L. C. Gahring. Nicotine blocks TNF-alpha-mediated neuroprotection to NMDA by an alpha-bungarotoxin- sensitive pathway. J. Neurobiol. 35: 29-36,1998.
108
32. Carswell, E. A., L. J. Old, and R. L. Dassei. An endotoxin-induced serum factor that causes necrosis of tumors, Proc. Natl Acad. Sci. USA 72: 3666-3670, 1975.
33. Cerami, A, Y Ikeda, N LeTrang, JP Hotez, and B Beutler. Weight loss associated with an endotoxin-induced mediator from peritoneal macrophages: the role of cachetin (tumor necrosis factor). Immunol. Lett. 11:173-177, 1985.
34. Chai, LY and MT Lin. Effects of heating and cooling the spinal cord and medulla oblongata on thermoregulation in monkeys. J. Physiol. 225:297- 308, 1972.
35. Chao, C. C. and S. Hu. Tumor necrosis factor-alpha potentiates glutamate neurotoxicity in human fetal brain cell cultures. Devl. Neurosci. 16: 172- 179, 1994.
36. Charpentier, B., C. Hiesse, O. Lantz, C. Ferran, S. Stephens D. O'Shaugnessy M. Bodmer G. Benoit J. F. Bach and L. Chatenoud. Evidence that antihuman tumor necrosis factor monoclonal antibody prevents OKT3-induced acute syndrome. Transplantation 54: 997-1002, 1992.
37. Chatenoud, L., C. Ferran, A. Reuter, et al. Systemic reaction to the monoclonal antibody OKT3 in relation to serum levels of tumor necrosis factor and interferon gamma. N. EngI. J. Med. 320: 1420, 1989.
38. Chen, C. H. and R. C. Rogers. Central inhibitory action of peptide YY on gastric motility in rats. Am. J. Physiol. 269: R787-R792, 1995.
39. Chen, C. H. and R. C. Rogers. Peptide YY and the Y2 agonist PYY-(13- 36) inhibit neurons of the dorsal motor nucleus of the vagus. Am. J. Physiol. 273: R213-R218, 1997.
40. Chiang, C. S. and W. H. McBride. Radiation enhances tumor necrosis factor alpha production by murine brain cells. Brain Res. 59: 265-269, 1991.
41. Chung, I. Y., J. G. Norris, E. N. Benveniste. Differential tumor necrosis factor a expression from experimental allergic encephalo- myelitis-susceptible and resistant rat strains. J. Exp. Med. 173: 801 –811, 1991.
42. Collins, T. L., L. A. Lapierre, W. Fiers, J. L. Strominger, and J. S. Pober. Recombinant tumor necrosis factor increases mRNA levels and surface
109
expression of HLA-A, B antigen in vascular endothelial cells and dermal fibroblasts in vitro. Proc. Natl. Acad. Sci. USA 83: 446-450,1986.
43. Cullen J. J., D. K. Caropreso, and K. S. Ephgrave. Effect of endotoxin on canine gastrointestinal motility and transit. J. Surg. Res. 58:90-95, 1995.
44. Curtis KS, JT Cunningham, and CM Heesch. Fos expression in brain stem of pregnant rats after hydralazine-induced hypotension. Am. J. Physiol. 277.2:R532-540, 1999.
45. Dean, JB, DA Bayliss, JT Erickson, WL Lawing, and DE Millhorn. Depolarization and stimulation of neurons in nucleus tractus solitarii by carbon dioxide does not require chemical synaptic input. Neurosci. 36(1):207-16, 1990.
46. Dinarello, C. A. Role of interleukin-1 and tumor necrosis factor in systemic responses to infection and inflammation. In: Inflammation: basic and clinical correlates, 2nd Ed. Edited by Gallin, Goldstein, and Snyderman. Raven Press, NY, pp 211-32, 1992.
47. Elmquist, JK, TE Scammell, CD Jacobson, And CB Saper. Distribution of Fos-like immunoreactivity in the rat brain following intravenous lipopolysaccharide administration. J. Comp. Neurol. 371: 85-103, 1996.
48. Emch, G. S., G. E. Hermann, and R.C. Rogers. TNF-a activates solitary nucleus neurons responsive to gastric distention. Am. J. Physiol. 279: G582-G586, 2000.
49. Emch, G. S., G. E. Hermann, and R. C. Rogers. TNF-a induces c-Fos generation in the nucleus of the solitary tract that is blocked by NBQX and MK-801. Soc. Neurosci. Abs. 31: 634.9, 2001.
50. Emch, G. S., G. E. Hermann, and R. C. Rogers. Tumor necrosis factor- alpha: Effects on identified neurons of the dorsal vagal complex. Soc. Neurosci. Abs. 25: 674.14, 1999.
51. Escobar ML, Alcocer I, and Chao V. The NMDA receptor antagonist CPP impairs conditioned taste aversion and insular cortex long-term potentiation in vivo. Brain Res 812: 246-251, 1998.
52. Escobar ML and Bermudez-Rattoni F. Long-term potentiation in the insular cortex enhances conditioned taste aversion retention. Brain Res 852: 208-212, 2000.
110
53. Ewart, W. R., M. V. Jones, and B. F. King. Central origin of vagal nerve fibres innervating the fundus and corpus of the stomach in rat. J Auton. Nerv. Syst. 25: 219-231, 1988.
54. Ferran, C., K. Sheehan, R. Schreiber, J. F. Bach, and L. Chatenoud. Anti- TNF abrogates the cytokine-related ant-CD3 induced syndrome. Transplant. Proc. 23: 849-850, 1991.
55. Fleshner, M., LE Goehler, BA Schwartz, M. McGorry, D. Martin, SF Maier, and LR Watkins. Thermogenic and corticosterone responses to intravenous cytokines (IL-1beta and TNF-alpha) are attenuated by subdiaphragmatic vagotomy. J. Neuroimmunol. 86: 134-141, 1998.
56. Furukawa, K. & M. P. Mattson. The transcription factor NF-kappaB mediates increases in calcium currents and decreases in NMDA- and AMPA/kainate-induced currents induced by tumor necrosis factor-alpha in hippocampal neurons. J. Neurochem. 70: 1876-1886, 1998.
57. Galanos, C and M. A. Fredenberg. Mechanisms of endotoxin shock and endotoxin hypersensitivity. Immunobiology 187:346-56,1993.
58. Ganter, S., H. Northoff, D. Martnel, and P. J. Gebicke-Harter. Growth control of cultured microglia. J. Neurosci. Res. 33: 218-230, 1992.
59. Garrick, R., R. Stephens, T. Ishikawa, A. Sierra, A. Avidan, H. Weiner, and Y. Tache. Medullary sites for TRH analogue stimulation of gastric contractility in the rat. Am. J Physiol. 256(19): G1O11 –G1O15, 1989.
60. Gaykema, R. P., L. E. Goehler, F. J. Tilders, J. G. Bol, M. McGorry, M. Fleshner, S. F. Maier, and L. R. Watkins. Bacterial endotoxin induces fos immunoreactivity in primary afferent neurons of the vagus nerve. Neuroimmunomodulation 5:234-240, 1998.
61. Gelbard, H. A., K. A. Dzenko, D. DiLoreto, C. del Cerro, M. del Cerro, and L. G. Epstein. Neurotoxic effects of tumor necrosis factor alpha in primary human neuronal cultures are mediated by activation of the glutamate AMPA receptor subtype: implications for AIDS neuropathogenesis. Devl. Neurosci. 15: 417-422, 1993.
62. Gill CF, Madden JM, Roberts BP, Evans LD, and King MS. A subpopulation of neurons in the rat rostral nucleus of the solitary tract that project to the parabrachial nucleus express glutamate-like immunoreactivity. Brain Res 821: 251-262, 1999.
111
63. Givalois, L., J. Dornand, M. Mekaouche, M. D. Solier, A. F. Bristow, G. Ixart, P. Siaud, I. Assenmacher, and G. Barbanel. Temporal cascade of plasma level surges in ACTH, corticosterone, and cytokines in endotoxin- challenged rats. Am. J. Physiol., 266:R164-R170, 1994.
64. Glise, H. and H. Abrahamsson. Reflex inhibition of gastric motility: pathophysiological aspects. Scand. J. Gastroenterol. 19, Suppl. 89: 77-82, 1984.
65. Goehler, LE, RP Gaykema, KT Nguyen, JE Lee, FJ Tilders, SF Maier, and LR Watkins. Interleukin 1-beta in immune cells of the abdominal vagus nerve: a link between the immune and nervous systems? J. Neurosci. 19: 2799-2806, 1999.
66. Grassi, F., A. M. Mileo, L. Monaco, A. Punturieri, A. Santoni, and F. Eusebi. TNF-alpha increases the frequency of spontaneous miniature synaptic currents in cultured rat hippocampal neurons. Brain Res. 659: 226-230, 1994.
67. Grell, M., E. Douni, and H. Wajant. The transmembrane form of tumor necrosis factor is the prime activating ligand of the 80 kDa tumor necrosis factor receptor. Cell 83, 793-802 1995.
68. Grill, H. J. and R. Norgren. Chronically decerebrate rats demonstrate satiation but not bait shyness. Science 201: 267-269, 1978.
69. Gross, P. M., K. M. Wall, J. J. Pang, S. W. Shaver, and D. S. Wainman. Microvascular specializations promoting rapid interstitial solute dispersion in nucleus tractus solitarius. Am. J. Physiol. 259: G687-G691, 1990.
70. Gutierrez, E. G., W. A. Banks, and A. J. Kastin. Murine tumor necrosis factor alpha is transported from blood to brain in the mouse. J. Neuroimmunol. 47: 169-176, 1993.
71. Hannun, Y.A. The sphingomyelin cycle and the second messenger function of ceramide. J. Biol. Chem. 269: 3125-3128, 1994.
72. Haviv, R. & R. Stein. The intracellular domain of P55 tumor necrosis factor receptor induces apoptosis which requires different caspases in naïve and neuronal PC12 cells. J. Neurosci. Res. 52: 380-389, 1998.
73. Heller, R. A., K. Song, N. Fan, and D. J. Chang. The p7O tumor necrosis factor receptor mediates cytotoxicity. Cell 70: 47-56, 1992.
112
74. Hermann GE, Emch GS, Tovar CA, and Rogers RC. c-Fos generation in the dorsal vagal complex after systemic endotoxin is not dependent on the vagus nerve. Am J Physiol Regulatory Integrative Comp Physiol 280: R289-R299, 2001.
75. Hermann, G. E. and R.C. Rogers. TNFa-suppression of gastric motility relieved by TNFR:FC adsorbant construct in the dorsal vagal complex. Soc. Neurosci. Abs. 31: 634.11, 2001.
76. Hermann, G. E. and R. C. Rogers. Tumor necrosis factor-alpha in the dorsal vagal complex suppresses gastric motility. Neuroimmunomodulation 2: 74-81, 1995.
77. Hermann, G. E., C. A. Tovar, and R. C. Rogers. Induction of endogenous tumor necrosis factor-a: suppression of centrally stimulated gastric motility. Am. J. Physiol. 276: R59-R68, 1999.
78. Hersch, E. M., B. S. Metch, F. M. Muggia, T. D. Brown, R. P. Whitehead, G. T. Budd, J. J. Rinehart, E. D. Crawford, J. D. Bonnet, and B. C. Behrans. Phase II studies of recombinant TNF in patients with malignant disease: a summary of the Southwest Oncology Group experience. J. Immunother. Emphasis Tumor Immunol. 10: 426-431, 1991.
79. Heuschling, P., C. Faber, and E. Morga. Astrocytes express regional heterogeneity after lipopolysaccharide induction of the immunoreactive phenotype. Soc. Neurosci. Abstr. 21: 885, 1995.
80. Hornby, P. J., C. D. Rossiter, S. V. Pineo, W. P. Norman, E. K. Friedman, S. Benjamin, and R. A. Gillis. TRH: immunocytochemical distribution in vagal nuclei of the cat and physiological effects of microinjection. Am. J Physiol. 257(20): G454-G462,1989.
81. Houpt TA, Berlin R, and Smith GP. Subdiaphragmatic vagotomy does not attenuate c-Fos induction in the nucleus of the solitary tract after conditioned taste aversion expression. Brain Res 747: 85-91, 1997.
82. Hsu, K. C. and M. V. Chao. Differential expression and ligand binding properties of tumor necrosis factor receptor chimeric mutants. J. Biol. Chem. 268: 16430-16436, 1993.
83. Ignatowski, T. A., and R. N. Spengler. Tumor necrosis factor a: pre- synaptic sensitivity is modified after antidepressant drug administration. Brain Res. 665: 293-299, 1994.
113
84. Ijzermans, J. N., M. Scheringa, G. P. van der Schelling, R. A. Geerling, R. L. Marquet, and J. Jeekel. Injection of recombinant tumor necrosis factor directly into liver metastases: an experimental and clinical approach. Clin. Exp. Metastasis 10: 91-97, 1992.
85. Ishikawa, T., H. Yang, and Y. Tache. Medullary sites of action of the TRH analogue, RX 77368, for stimulation of gastric acid secretion in the rat. Gastroenterology 95: 1470-1476, 1988.
86. Kalia, M. and M.-M. Mesulam. Brain stem projections of sensory and motor components of the vagus complex in the cat: the cervical vagus and nodose ganglion. J Comp. Neurol. 193: 435-465, 1980.
87. Kalia, M. and M.-M. Mesulam. Brain stem projections of sensory and motor components of the vagus complex in the cat II: Laryngeal, tracheobronchial, pulmonary, cardiac and gastrointestinal branches. J. Comp. Neurol. 193: 467-508, 1980.
88. Kalia, M. and J. M. Sullivan. Brainstem projections of sensory and motor components of the vagus nerve in the rat. J Comp. Neurol. 211: 248-264, 1982.
89. Kapas, L, M. K. Hansen, H. Y. Chang, and J. M. Krueger. Vagotomy attenuates but does not prevent the somnogenic and febrile effects of lipopolysaccharide in rats. Am. J. Physiol. 274: R406-R411, 1998.
90. Kapas, L., L. Hong, A. B. Cady, M. R. Opp, A. E. Postlethwaite, J. M. Seyer, & J. M. Krueger. Somnogenic, pyrogenic, and anorectic activities of tumor necrosis factor-a and TNF-a fragments. Am. J. Physiol. 263: R708-R715, 1992.
91. Keifer SW. Neural mediation of conditioned food aversions. Ann NY Acad Sci 443: 100-109, 1990.
92. Kemeny, N, B Childs, W Larchian, K Rosado, and D Kelsen. A phase II trial of recombinant tumor necrosis factor in patients with advanced colorectal carcinoma. Cancer 66: 659-663, 1990.
93. Kinouchi, K., G. Brown, G.Pasternak, & D. B. Donner. Identification and characterization of receptors for tumor necrosis factor-a in the brain. Biochem. Biophys. Res. Commun. 181: 1532-1538, 1991.
114
94. Kirchgessner, A. L. and M. D. Gershon. Identification of vagal efferent fibers and putative target neurons in the enteric nervous system of the rat. J Comp. Neurol. 285: 38-53, 1989.
95. Klefstrom, J., I. Vastrik, E. Saksela, J. Valle, M. Eilers, and K. Alitalo. c-Myc induces cellular susceptibility to the cytotoxic action of TNF-alpha. Eur. Molec. Biol. Org. J. 13: 5442-5450, 1994.
96. Kobashi, M and A Adachi. Convergence of hepatic osmoreceptive inputs on sodium-responsive units within the nucleus of the solitary tract of the rat. J. Neurophysiol. 54 (2):212-219, 1985.
97. Kosterlitz, H. W. Intrinsic and extrinsic nervous control of motility of the stomach and intestines. In: Handbook of Physiology, Alimentary Canal. Vol. 4: Motility. Edited by C. F. Code, American Physiological Society, 1968.
98. Kotanidou, A, AM Choi, RA Winchurch, L Otterbein, and HE Fessler. Urethan anesthesia protects rats against lethal endotoxemia and reduces TNFa release. J. Appl. Physiol. 81(5):2304-2311, 1996.
99. Kronke, M., S. Schutze, and P. Scheurich. Tumor necrosis factor signal transduction. Cell. Signal. 2: 1-8, 1990.
100. Kunkel, S. L., D. G. Remick, R. M. Streiter, and J. W. Larrick. Mechanisms that regulate the production and effects of tumor necrosis factor. Critical Rev. Immunol. 9:93-117, 1989.
101. Lachman, L. B., D. C. Brown, and C. A. Dinarello. Growth promoting effect of recombinant interleukin-I and tumor necrosis factor for a human astrocytoma cell line. J. Immunol.138: 2913-2916, 1987.
102. Lee, S. C., W. Liu, D. W. Dickson, C. F. Brosnan, and J. W.Berman. Cytokine production by human fetal microglia and astrocytes. Differential induction by lipopolysaccharide and IL-I beta. J. Immunol. 150: 2659-2667, 1993.
103. Leslie, R. A., D. G. Gwyn, and D. A. Hopkins. The central distribution of the cervical vagus nerve and gastric afferent and efferent projections in the rat. Brain Res. Bull. 8: 37-43, 1982.
104. Li S and Stys PK. Mechanisms of ionotropic glutamate receptor- mediated excitoxicity in isolated spinal cord white matter. J Neurosci 20: 1190-1198, 2000.
115
105. Lin H-C, Wan F-J, Kang B-H, Wu C-C, and Tseng C-J. Systemic administration of lipopolysaccharide induces release of nitric oxide and glutamate and c-Fos expression in the nucleus tractus solitarii of rats. Hypertension 33: 1218-1224, 1999.
106. Lin L-H, Emson PC, and Talman WT. Apposition of neuronal elements containing nitric oxide synthase and glutamate in the nucleus tractus solitarii of rat: a confocal microscopic analysis. Neuroscience 96: 341-350, 2000.
107. Lindvall, L., M. Lantz, U. Gullberg, I. Olsson. Modulation of the constitutive gene expression of the 55 kd tumor necrosis factor receptor in hematopoietic cells. Biochem. Biophys. Res. Commun. 172: 557-563, 1990.
108. Liu, T., R. K. Clark, and P. C. McDonnell. Tumor necrosis factor-alpha expression in ischemic neurons. Stroke 25: 1481-1488, 1994.
109. Loewy, A. D., S. McKellar, and C. B. Saper. Direct projections from the A5 catecholamine cell group to the intermediolateral cell column. Brain Res. 174: 309-314, 1979.
110. Lotan, M., A. Solomon, S. Ben-Bassat, & M. Schwartz. Cytokines modulate the inflammatory response and change permissiveness to neuronal adhesion in injured mammalian central nervous system. Exp. Neurol. 126: 284-290, 1994.
111. Lundberg, J. M., T. Hokfelt, J. Kewenter, G. Pettersson, H. Ahlman, R. Edin, A. Dahlstrom, G. Nilsson, L. Terenius, K. Uvnas-Wallensten, and S. Said. Substance P-, VIP-, and enkephalin-like immunoreactivity in the human vagus nerve. Gastroenterology 77: 468-471, 1979.
112. Malenka RC, Kauer JA, Zucker RS, and Nicoll RA. Postsynaptic calcium is sufficient for potentiation of hippocampal synaptic transmission. Science 242: 81–84, 1988.
113. Manaker, S., A. Winokur, W. H. Rostene, and T. C. Rainbow. Autoradiographic localization of thyrotropin-releasing hormone receptors in the rat central nervous system. J Neurosci. 5(1): 167-174, 1985.
114. Manaker, S. and P. C. Zucchi. Effects of vagotomy on neurotransmitter receptors in the rat dorsal vagal complex. Neuroscience 52(2): 427-441, 1993.
116
115. Mascarucci P, Perego C, Terrazzino S, and De Simoni MG. Glutamate release in the nucleus tractus solitarius induced by peripheral lipopolysaccharide and interleukin-1 beta. Neuroscience 86: 1285-1290, 1998.
116. McCann, M. J., G. E. Hermann, and R. C. Rogers. Dorsal medullary serotonin and gastric motility: enhancement of effects by thyrotropin-releasing hormone. J Auton. Nerv. Syst. 25: 35-40, 1988.
117. McCann, M. J., G. E. Hermann, and R. C. Rogers. Thyrotropin-releasing hormone: effects on identified neurons of the dorsal vagal complex. J Auton. Nerv. Syst. 26: 107-112, 1989.
118. McCann, M. J. and R. C. Rogers. Functional and chemical anatomy of a gastric vago-vagal reflex. In: Innervation of the Gut: Pathophysiological Implications. Boca Raton: CRC, 1994, p. 81-92.
119. McCann, M. J. and R. C. Rogers. Impact of antral mechanoreceptor activation on the vago-vagal reflex in the rat: functional zonation of responses. J Physiol. 453: 401-411, 1992.
120. McCann M. J. and Rogers R. C. Oxytocin excites gastric-related neurones in rat dorsal vagal complex. J Phsiol (Lond) 428: 95-108, 1990.
121. McCaughey SA, Giza BK, Nolan LJ, and Scott TR. Extinction of a conditioned taste aversion in rats: II. Neural effects in the nucleus of the solitary tract. Physiol Behav 61: 373- 379, 1997.
122. McKellar, S. and A. D. Loewy. Efferent projections of the A1 catecholamine cell group in the rat: an autoradiographic study. Brain Res. 241: 11-29, 1982.
123. McTigue, D. M., N. K. Edwards, and R. C. Rogers. Pancreatic polypeptide in the dorsal vagal complex stimulates gastric acid secretion and motility in the rat. Am. J. Physiol. 265: G1169-G1176, 1993.
124. McTigue, D. M., G. E. Hermann, and R. C. Rogers. Effect of pancreatic polypeptide on rat dorsal vagal complex neurons. J. Physiol. 499.2: 475- 483, 1997.
125. McTigue, D. M. and R. C. Rogers. Pancreatic polypeptide stimulates gastric motility through a vagal-dependent mechanism in rats. Neurosci. Lett. 188: 93-96, 1995.
117
126. McTigue, D. M., R. C. Rogers, and R. L. Stephens Jr. Thyrotropin-releasing hormone analogue and serotonin interact within the dorsal vagal complex to augment gastric acid secretion. Neurosci. Lett. 144: 61-64, 1992.
127. Megyeri, P., C. S. Abraham, P. Temesvari, J. Kovacs, T. Vas, & C. P. Speer. Recombinant human tumor necrosis factor a constricts pial arterioles and increases blood-brain barrier permeability in newborn piglets. Neurosci. Lett. 148: 137-140, 1992.
128. Mei, N. Recent studies on intestinal vagal afferent innervation. Functional implications. J Auton. Nerv. Syst. 9: 199-206, 1983.
129. Merchenthaler, I., V. Csernus, C. Csontos, P. Petrusz, and B. Mess. New data on the immunocytochemical localization of thyrotropin-releasing hormone in the rat central nervous system. Am. J Anat. 181: 359-376, 1988.
130. Meulemans, A. L., L. F. Helsen, and I A. I Schuurkes. Role of NO in vagally-mediated relaxations of guinea-pig stomach. Arch. Pharmocol. 347: 225-230, 1993.
131. Milligan, E. D., M. M. McGorry, M. Fleshner, R. P. A Gaykema, L. E. Goehler, L. R. Watkins, and S. F. Maier. Subdiaphragmatic vagotomy does not prevent fever following intracerebroventricular prostaglandin E2: further evidence for the importance of vagal afferents in immune-to-brain communication. Brain Res. 766:240-243, 1997.
132. Miolan, J. P. and C. Roman. Discharge of efferent vagal fibers supplying gastric antrum: indirect study by nerve suture technique. Am. J. Physiol. 235: E366-E373, 1978.
133. Miolan, J. P. and C. Roman. The role of oesophageal and intestinal receptors in the control of gastric motility. J Auton. Nerv. Syst. 10: 235-241, 1984.
134. Mitrovic, B., F. C. Martin, and A. C. Charles. Neurotransmitters and cytokines in CNS pathology. Progr. Brain Res. 103: 319-330, 1994.
135. Mogi, M., M. Harada, P. Riederer, and H. Narabayashi. Tumor necrosis factor-alpha (TNF-alpha) increases both in the brain and in the cerebrospinal fluid from Parkinsonian patients. Neurosci. Lett. 165: 208-210, 1994.
118
136. Moltz, H. Fever: causes and consequences. Neurosci. Biobehav. Rev. 17: 237-269, 1993.
137. Moretto, G., A. Yoo, and S. U. Kim. Human astrocytes and cytokines: tumor necrosis factor alpha and interferon gamma do not promote astrocytic proliferation. Neurosci. Lett. 151: 17-20, 1993.
138. Motulsky, H. Repeated measures ANOVA. In: Intuitive Biostatistics. Oxford Univ. Press, 1995, pp.255-62.
139. Muggia, F. M., T. D. Brown, P. J. Goodman, J. S. Macdonald, E. M. Hersh, T. R. Fleming, and L. Leichman. High incidence of coagualopathy in phase II studies of recombinant tumor necrosis factor in advanced pancreatic and gastric cancers. Anticancer Drugs 3: 211-217, 1992.
140. Nathan, C. F. Secretory products of macrophages. J. Clin. Invest. 79: 319 – 326, 1987.
141. Nitta, T., M. Ebato, K. Sato, and K. Okumura. Expression of tumour necrosis factor-alpha, -beta and interferon-gamma genes within human neuroglial tumour cells and brain specimens. Cytokine 6: 142. 171-180, 1994.
143. Norgren, R. Projections from the nucleus of the solitary tract in the rat. Neuroscience 3: 207-218, 1978.
144. Oh, Y. J., G. J. Markelonis, and T. H. Oh. Effects of interleukin-1beta and tumor necrosis factor-alpha on the expression of glial-fibrillary acidic protein and transferrin in cultured astrocytes. Glia 8: 71-96, 1993.
145. Palkovits, M, E Mezey, M Fodor, D Ganten, U Bahner, H Geiger, and A Heidland. Neurotransmitters and neuropeptides in the baroreceptor reflex arc: connections between the nucleus of the solitary tract and the ventrolateral medulla oblongata in the rat. Clin. Exptl. Hypertension 17 (1- 2):101-13, 1995.
146. Pan, W., J. E. Zadina, R. E. Harlan, J. T. Weber, W. A. Banks, A. J. Kastin. Tumor necrosis factor-alpha: a neuromodulator in the CNS. Neurosci. Biobehav. Rev. 21: 603-613, 1997.
147. Patton, J., P. Peters, J. McCabe, D. Crase, S. Hansen, A. Chen and D. Liggit. Development of partial tolerance to the gastrointestinal effects of high doses of recombinant tumor necrosis factor in rodents. J. Clin. Invest. 80: 1587-1596, 1987.
119
148. Perez, C., I. Albert, K. DeFay, N. Zachariades, L. Gooding, and M. Knegier. A nonsecretable cell surface mutant of tumor necrosis factor (TNF) kills by cell-to-cell contact. Cell 63: 251-258, 1996.
149. Perry, S. W., J. A. Hamilton, L. W. Tjoelker, G. Dbaibo, K. A. Dzenko, L. G. Epstein, Y. Hannun, J. S. Whittaker, S. Dewhurst, & H. A. Gelbard. Platelet-activating factor receptor activation. J. Biol. Chem. 273: 17660- 17664, 1998.
150. Perry, V. H., P. B. Andersson, and S. Gordon. Macrophages and inflammation in the central nervous system. Trends Neurosci. 16: 268-273, 1993.
151. Pfizenmaier, K., A. Himmler, S. Schutze, P. Scheurich, and M. Kronke. TNF receptor's and TNF signal transduction. In: Tumor necrosis factors: The molecules and their emerging role in medicine. (Ed.) Beutler, Raven Press, New York, pp. 439-472, 1992.
152. Pober, J. S., M. A. Gimbrone, L. A. Lapierre, D. L. Mendrick, W. Fiers, R. Rothelein, and T. A. Springer. Overlapping patterns of activation of human endothelial cells by interleukin-1, tumor necrosis factor, and immune interferon. J. Immunol. 137: 1893-1896, 1986.
153. Porter, MH, BJ Hrupka, W Langhans, and GJ Schwartz. Vagal and splanchnic afferents are not necessary for the anorexia produced by peripheral IL-1 beta, LPS, and MDP. Am. J. Physiol. 275:R384-389, 1998.
154. Powrie, F., and R. L. Coffman. Cytokine regulation of T-cell function: potential for therapeutic intervention. Immunol. Today 14: 270-274, 1993.
155. Raggenbass M, Alberi S, Zaninetti M, Pierson P, and Dreifuss JJ. Vasopressin and oxytocin action in the brain: cellular neurophysiological studies. Prog Brain Res 119: 263-273, 1998.
156. Randle JC, Guet T, Cordi A, and Lepagnol JM. Competitive inhibition by NBQX of kainate/AMPA receptor currents and excitatory synaptic potentials: importance of 6-nitro-substitution. Eur J Pharmacol 215: 237- 244, 1992.
157. Rinaman L. Oxytocinergic inputs to the nucleus of the solitary tract and dorsal motor nucleus of the vagus in neonatal rats. J Comp Neurol 399: 101-109, 1998.
120
158. Rinaman, L., J. P. Card, J. S. Schwaber, and R. R. Miselis. Ultrastructural demonstration of a gastric monosynaptic vagal circuit in the nucleus of the solitary tract in rat. J Neurosci. 9(6): 1985-1996, 1989.
159. Rinaman, L., J. G. Verbalis, E. M. Striker, and G. E. Hoffman. Distribution and neurochemical phenotypes of caudal medullary neurones activated to express c-FOS following peripheral administration of cholecystokinin. J. Comp. Neurol. 338:475-490, 1993.
160. Ritter, R. C., L. A. Brenner, and C. S. Tamura. Endogenous CCK and peripheral neural substrates of intestinal satiety. Ann. NY Acad. Sci., 713:255-67, 1994.
161. Rogers, R. C. and G. E. Hermann. Central connections of the hepatic branch of the vagus nerve: a horseradish peroxidase histochemical study. J. Autonomic Nerv. Sys. 7:165-174, 1983.
162. Rogers, R. C. and G. E. Hermann. Dorsal medullary oxytocin, vasopressin, oxytocin antagonist, and TRH effects on gastric acid secretion and heart rate. Peptides 6: 1143-1148, 1985.
163. Rogers, R. C. and G. E. Hermann. Oxytocin, oxytocin antagonist, TRH, and hypothalamic paraventricular nucleus stimulation effects on gastric motility. Peptides 8: 505-513, 1987.
164. Rogers, R.C., G. E Hermann, & R. A Travagli. Brainstem pathways responsible for oesophageal control of gastric motility and tone in the rat. J. Physiol. 514.2: 369-383, 1999.
165. Rogers, R. C. and M. J. McCann. Intramedullary connections of the gastric region in the solitary nucleus: a neurobiotin injection histochemical tracing study in the rat. J. Auton. Nerv. Syst. 42: 119-130, 1993.
166. Rogers, R. C., D. M. McTigue, and G. E. Hermann. Vagal control of digestion: Modulation by central neural and peripheral endocrine factors. Neurosci. Biobehav. Rev. 20: 57-66, 1996.
167. Rogers, R.C., D.M. McTigue, and G.E. Hermann. Vago-vagal reflex control of digestion: Afferent modulation by neural and ‘endoneurocrine’ factors. Am. J. Physiol. 268: G1-G10, 1995.
168. Rogers, R. C., D. Novin, and L. L. Butcher. Electrophysiological and neuroanatomical studies of hepatic portal osmo-and sodium-receptive afferent projections within the brain. J. Autonomic Nerv. Sys. 1:183-202, 1979.
121
169. Roman, C. and J. Gonella. Extrinsic control of digestive tract motility. In: Physiology of the Gastrointestinal Tract, edited by L. R. Johnson. New York: Raven Press, 1987, p. 507-553.
170. Romanovsky, A. A., C. T. Simons, V. A. Kulchitsky, N. Sugimoto, and M. Szekely. Vagus nerve in fever: recent developments. Ann. NY Acad. Sci. 856:298-9, 1998.
171. Rossiter CD, Norman WP, Jain M, Hornby PJ, Benjamin S, and Gillis RA. Control of lower esophageal sphincter pressure by two sites in dorsal motor nucleus of the vagus. Am J Physiol Gastrintest Liver Physiol 259: G899-G906, 1990.
172. Rothe, J., G. Gehr, H. Loetscher, and W. Lesslauer. Tumor necrosis factor receptors – structure and function. Immunol. Res. 11: 81- 90, 1992.
173. Sagar, SM, KJ Price, NW Kasting, and FR Sharp. Anatomic patterns of FOS immunostaining in rat brain following systemic endotoxin administration. Brain Res. Bull. 36: 381-392, 1995.
174. Sawada, M., A. Surumura, and T. Martmouchi. TNF alpha induces IL-6 production by astrocytes but not by microglia. Brain Res. 583: 296-299, 1992.
175. Sawchenko P. E. Central connections of the sensory and motor nuclei of the vagus nerve. J Auton Nerv Syst 9: 13-26, 1983.
176. Sawchenko, P. E. and L. W. Swanson. Central noradrenergic pathways for the integration of hypothalamic neuroendocrine and autonomic responses. Science 214: 685-687, 1981.
177. Sawchenko, P. E. and L. W. Swanson. The organization of noradrenergic pathways from the brainstem to the paraventricular and supraoptic nuclei in the rat. Brain Res. Rev. 4: 275-325, 1982.
178. Schwartz, G. J., C. R. Plata-Salaman, & W. Langhans. Subdiaphragmatic vagal deafferentiation fails to block feeding-suppressive effects of LPS and IL-1b in rats. Am. J. Physiol. 273: R1193-R1198, 1997.
179. Scratcherd, T., D. Grundy, L. Rudge, J. S. Ball, and R. G. Clark. Reflex control of gastric motility by stimuli acting from within the stomach. In: Nerves and the Gastrointestinal Tract, edited by M. V. Singer and H. Goebell. Lancaster: MTP Press, Ltd. 1989, p. 373-381.
122
180. Sehic, E. & C. M. Blatteis. Blockade of lipopolysaccharide-induced fever by subdiaphragmatic vagotomy in guinea pigs. Brain Res. 726: 160- 166, 1996.
181. Selmaj, K. W. M. Farooq, W. T. Norton, C. S. Raine, and C. F. Brosnan. Proliferation of astrocytes in vitro in response to cytokines: A primary role for tumor necrosis factor. J. Immunol. 144: 129-135, 1990.
182. Selmaj, K. W. and C. S. Raine. Tumor necrosis factor mediates myelin and oligodendrocyte damage in vitro. Ann. Neurol. 23: 339-346, 1988.
183. Senba, E., T. Kaneko, N. Mizuno, and M. Tohyama. Somato-, branchio- and viscero-motor neurons contain glutaminase-like immunoreactivity. Brain Res. Bull. 26: 85-97, 1991.
184. Shapiro, R. E. and R. R. Miselis. The central organization of the vagus nerve innervating the stomach of the rat. J. Comp. Neurol. 238: 473-488, 1985.
185. Sharif, S. F, R. J. Hariri, V. A. Chang, P. S. Barie, R. S. Wang, and J. B. Ghajar. Human astrocyte production of tumour necrosis factor alpha, interleukin-l beta, and interleukin-6 following exposure to lipopolysaccharide endotoxin. Neurol. Res.15: 109-112, 1993.
186. Sierra. A. and N. Rubio. Theiler's murine encephalomyelitis virus induces tumour necrosis factor-alpha in murine astrocyte cell cultures. Immunology 78: 399-404, 1993.
187. Simons, CT, VA Kulchitsky, N. Sugimoto, LD Homer, and AA Romanovsky. Signaling the brain in systemic inflammation: which vagal branch is involved in fever genesis? Am. J. Physiol. 275: R63-R68, 1998.
188. Sivarao DV, Krowicki ZK, Abrahams P, and Hornby PJ. Vagally- regulated gastric motor activity: evidence for kainate and NMDA receptor mediation. Eur J Pharamacol 368: 173-182, 1999.
189. Smith BN, Dou P, Barber WD, and Dudek FE. Vagally evoked synaptic currents in the immature rat nucleus tractus solitarii in an intact in vitro preparation. J Physiol (Lond) 512: 149-162, 1998.
190. Sparacio, S. M., Y. Zhang, J. Vilcek, and E. N. Benveniste. Cytokine regulation of interleukin-6 gene expression in astrocytes involves activation of an NF-kappa B-like nuclear protein. J. Neuroimmunol. 39: 231-242, 1992.
123
191. Spencer SE and Talman WT. Modulation of gastric and arterial pressure by nucleus tractus solitarius in rat. Am J Physiol Regulatory Integrative Comp Physiol 250: R996-R1002, 1986.
192. Tache, Y. and H. Yang. Role of medullary TRH in the vagal regulation of gastric function. In: Innervation of the Gut.- Pathophysiological Implications, edited by Y. Tache, D. L. Wingate, and T. F. Burks. Boca Raton: CRC Press, 1994, p. 67-88.
193. Tache, Y., H. Yang, and M. Yoneda. Vagal regulation of gastric function involves thyrotropin-releasing hormone in the medullary raphe nuclei and dorsal vagal complex. Digestion 54: 65-72, 1993.
194. Talley, A. K., S. Dewhurst, and S. W. Perry. Tumor necrosis factor alpha-induced apoptosis in human neuronal cells: protection by the antioxidant N-acetylcysteine and the genes bcl-2 and crmA. Molec. Cell. Biol. 15: 2359-2366, 1995.
195. Tancredi, V., G. D’Arcangelo, and F. Grassi. Tumor necrosis factor alters synaptic transmission in rat hippocampal slices. Neurosci. Lett. 146: 176-178, 1992.
196. Tartaglia, L. A., T. M. Ayres, G. H. Wong, and D. V. Goeddel. A novel domain within the 55 kd TNF receptor signals cell death. Cell 74: 845-853, 1991.
197. Tartaglia, L. A., D. Pennica, and D. V. Goeddel. Ligand passing the 75- kDa tumor necrosis factor (TNF) receptor recruits TNF for signaling by the 55-kDa TNF receptor. J. Biol. Chem. 268: 18542-18548, 1993.
198. Tartaglia, L. A., R. F. Weber, I. S. Figari, C. Reynolds, M. A. Palladino, and D. V. Goeddel. The different receptors for tumor necrosis factor mediate distinct cellular responses. Proc. Natl. Acad. Sci. USA 88: 9292- 9296, 1991.
199. Tkacs, NC, J Li, and AM Strack. Central amygdala Fos expression during hypotensive or febrile, nonhypotensive endotoxemia in conscious rats. J. Comp. Neurol. 379:592-602, 1997.
200. Thor, K. B., A. Blitz-Siebert, and C. J. Helke. Autoradiographic localization of 5HT1 binding sites in autonomic areas of the rat dorsomedial medulla oblongata. Synapse 10: 217-227 1992.
124
201. Tork, I. Raphe nuclei and serotonin containing systems. In: The Rat Nervous System: Vol 2 Hindbrain and Spinal Cord, edited by G. Paxinos. Sydney: Academic Press, Inc. 1985, p. 43-78.
202. Torrealba F and Muller C. Ultrastructure of glutamate and GABA immunoreactive axon terminals of the rat nucleus tractus solitarius, with a note on infralimbic cortex afferents. Brain Res 820: 20-30, 1999.
203. Tracey, K.J. and A. Cerami. Tumor necrosis factor and regulation of metabolism in infection: role of systemic versus tissue levels. Proc. Soc. Exp. Biol. and Medicine 200: 233-239, 1992.
204. Tracey, K. J., and A. Cerami. Pleiotropic effects of TNF in infection and neoplasma: beneficial, inflammatory, catabolic, or injurious. In: Tumor necrosis factors: structure, function, and mechanisms of action. Edited by Aggarwal and Vilcek. Marcel Dekker, NY, pp. 431-452, 1991.
205. Travagli, R. A., R. A. Gillis, C.D. Rossiter and S. Vicini. Glutamate and GABA-mediated synaptic currents in neurons of the rat dorsal motor nucleus of the vagus. Am. J. Physiol. 260: G531-G536, 1991.
206. Treece BR, Covasa M, Ritter RC, and Burns GA. Delay in meal termination follows blockade of N-methyl-D-aspartate receptors in the dorsal hindbrain. Brain Res 810: 34-40, 1998.
207. Vandenabeele, P., W. Declercq, B. Vanhaesebroeck, J. Grooten, and W. Fiers. Both TNF receptors are required for TNF-mediated induction of apoptosis in PC60 cells. J. Immunol. 154: 2904-2913, 1995.
208. Van der Kooy, D and L. Y. Koda. Organization of the projections of a circumventricular organ: the area postrema in the rat. J. Comp. Neurol. 219:328-338, 1983.
209. Waage, A. Production and clearance of tumor necrosis in rats exposed to endotoxin and dexamethasone. Clin. Immunol. Immunopathol. 45:348- 355, 1987.
210. Wan W., L. Janz, C. Y. Vriend, C. M. Sorensen, A. H. Greenberg, and D.M. Nance. Differential induction of c-Fos immunoreactivity in hypothalamus administration of endotoxin. Brain. Res. Bull. 32: 581-587, 1993.
211. Watkins, L. R., S. F. Maier, and L. E. Goehler. Immune activation: the role of pro-inflammatory cytokines in inflammation, illness responses and pathological pain states. Pain 63: 289-302, 1995.
125
212. Whitcomb, D. C., I. L. Taylor, and S. R Vigna. Characterization of saturable binding sites for circulating pancreatic polypeptide in the rat brain. Am. J. Physiol. 259: G687-G691, 1990.
213. Whitehead M. C., A. Bergula, and K. Holliday. Forebrain connections to the rostral nucleus of the solitary tract in the hamster. J. Comp. Neurol. 422: 429-447, 2000.
214. Wong E.H.F., Kemp JA, Priestley T, Knight AR, Woodruff GN, and Iversen LL. The anticonvulsant MK-801 is a potent NMDA antagonist. Proc. Natl. Acad. Sci. USA 83: 7104-7108, 1986.
215. Wood, J. A. Physiology of the enteric nervous system. In: Physiology of the Gastrointestinal Tract, edited by L. R. Johnson. New York: Raven Press, 1987, p. 67-109.
216. Zhang, X., R. Fogel, & W. E. Renehan. Physiology and morphology of neurons in the dorsal motor nucleus of the vagus and the nucleus of the solitary tract that are sensitive to distention of the small intestine. J. Comp. Neurol. 323: 432-448, 1992.
217. Zheng H, Kelly L, Patterson LM, and Berthoud H-R. Effect of brain stem NMDA-receptor blockade by MK-801 on behavioral and Fos responses to vagal satiety signals. Am J Physiol Regulatory Integrative Comp Physiol 277: R1104-R1111, 1999.
126