Effects of Subfornical Organ Lesions on Sympathetic Nerve Responses to Insulin
Abstract Although insulin exerts potent excitatory effects on the sympathetic nervous system, the mechanisms of insulin-induced activation remain unclear. To demonstrate a central nervous system site of sympathoexcitation, we recently found that destruction of tissues surrounding the anteroventral third ventricle region abolishes elevations in sympathetic nerve activity to intravenous insulin administration. Anteroventral third ventricle lesions may eliminate sympathoexcitation by destroying cell bodies in the lesioned area or by interrupting fibers of passage from the subfornical organ. To determine whether the lesions abolish sympathetic increases by disrupting efferent fibers from the subfornical organ, we measured lumbar sympathetic activity in anesthetized anteroventral third ventricle–lesioned (n=4) and subfornical organ–lesioned (n=12) rats before and during intravenous insulin at 0.13 U/h while maintaining euglycemia. Additional sham-lesioned rats received infusion of insulin (n=10) and the vehicle for insulin (n=10). Insulin administration in sham-lesioned rats elevated lumbar activity from 100% to 171±14% (±SE), whereas vehicle infusion did not alter sympathetic activity (100% to 113±11%). In anteroventral third ventricle–lesioned rats, insulin failed to increase sympathetic nerve activity (100% to 119±14%). Importantly, rats with subfornical organ lesions had increases in nerve activity that were indistinguishable from increases observed in insulin-infused sham-lesioned rats (100% to 163±21%). These findings indicate that whereas the anteroventral third ventricle region itself is crucial for sympathoexcitation to insulin, the subfornical organ and fibers originating from the subfornical organ traversing the anteroventral third ventricle area are not essential in mediating elevations in lumbar sympathetic nerve activity to hyperinsulinemia.
Insulin exerts powerful excitatory effects on the sympathetic nervous system in both humans1 2 and experimental animals.3 4 5 Although insulin-induced sympathoexcitation may play a role in blood pressure regulation and the development of hypertension,6 the underlying mechanisms and sites through which insulin elicits sympathetic activation remain unclear. To demonstrate a central nervous system origin of activation, we recently established that insulin infusion into the third brain ventricle of rats evokes acute increases in lumbar SNA.5 In follow-up studies, we reasoned that if infusion of the hormone into the third cerebral ventricle activates lumbar activity, destruction of tissues surrounding the third ventricle may abolish increases in sympathetic outflow to systemic insulin infusion during euglycemic clamp. To evaluate this possibility, we ablated tissues surrounding the AV3V, a region strongly implicated in blood pressure regulation and sympathetic neural control.4 Destruction of the AV3V abolished increases in lumbar activity to systemic insulin infusion, indicating that structures within or associated with the AV3V are crucial for activation of sympathetic activity in response to hyperinsulinemia.
AV3V lesions may eliminate sympathetic responses to insulin by destroying cell bodies in the ablated area or by interrupting fibers of passage removed from the targeted region. Cell body regions within the lesion include the periventricular nuclei at the preoptic anteroventral hypothalamic level, the OVLT, the ventral portion of the median preoptic nucleus, and the medial edge of the preoptic nuclei.7 The lesion also interrupts a ventrally directed system of efferents originating from the SFO that contact neurons in the median preoptic nucleus and OVLT as well as cell groups in the suprachiasmatic and supraoptic nuclei.8
One way to determine which element of the AV3V is critical for sympathetic responses to insulin is to fractionate the lesion into its component neural systems. For example, if the OVLT were the critical component, it would be expected that selective lesions of the OVLT would be equally effective in abolishing sympathetic responses to hyperinsulinemia. For the present studies, we tested the hypothesis that AV3V lesions abolish sympathoexcitation to insulin by disrupting fibers of passage that arise from the SFO. We focused on the SFO for the following reasons: First, the SFO sends an important set of efferent fibers through the AV3V8 ; second, the SFO is a well-known chemoreceptive circumventricular organ that lacks a blood-brain barrier7 ; third, the SFO contains high concentrations of insulin-specific binding sites9 10 ; and finally, the SFO functions to monitor blood-borne peptide hormones and to transmit this information into the central nervous system, thereby aiding in the regulation of body fluid balance and arterial blood pressure.7 In agreement with a sensory role for the SFO, intravenous injection of angiotensin II stimulates activity in SFO neurons, resulting in vasopressin release and elevations in sympathetic neural outflow.11 12 13 In addition, intravenous endothelin administration increases activity in SFO neurons, and microinjection of endothelin into the SFO causes elevations in blood pressure.14 Taken together, these findings are consistent with the hypothesis that blood-borne signals, such as insulin, activate specific receptors in the SFO, which send efferent signals through the AV3V region that eventually produce increases in peripheral sympathetic neural activity. Accordingly, the purpose of this study was first to determine once more whether lesions of the AV3V would abolish increases in sympathetic activity to hyperinsulinemia, and second to determine whether lesions of the SFO would have the same effect.
Experiments were performed on male Sprague-Dawley rats weighing 280 to 300 g (Harlan Sprague Dawley Inc, Indianapolis, Ind). All procedures were performed in accordance with the University of Iowa and National Institutes of Health guidelines for the care and use of experimental animals.
For SFO lesions, rats were anesthetized with ketamine (Henry Schein, 40 mg/kg IM) and secured in a stereotaxic instrument (Kopf 900, David Kopf Instruments). The scalp was incised on the midline and the skull leveled between bregma and lambda. A 3-mm trephine hole was made at bregma. SFO lesions were made with two or three penetrations of a 0.25-mm insulated nichrome wire electrode and a range of stereotaxic coordinates. The most anterior stereotaxic coordinate was −1.3 or −1.6 mm from bregma in the anterior-posterior plane, with the next penetration(s) being 0.3 mm posterior to the previous one. The electrode was lowered to a point 4.9 to 5.3 mm from the top of the exposed midsagittal sinus, with the next penetration(s) being 0.2 mm lower than the previous. The angle of descent was 6° from the perpendicular at a point 0.5 to 0.7 mm lateral to the middle of the sinus so that retraction of the sinus was avoided. Anodal current was passed through the bare tip at 1 mA for 8 to 10 seconds per penetration. Sham lesions were produced by lowering the electrode to a point 1.0 mm above the coordinates used for SFO lesions, and no current was passed.
For AV3V lesions, rats were anesthetized with ketamine and secured in a Kopf 900 stereotaxic instrument with the skull leveled between bregma and lambda. A lesioning electrode (24-gauge nichrome wire insulated except at the tip) was lowered on the midline 0.3 mm caudal to bregma to a depth 7.5 mm from dura. Anodal current (2 to 3 mA) was passed for 25 to 30 seconds (rectal cathode). In sham-lesioned rats, the electrode was lowered to a point 0.5 mm above the intended target tissue, and no current was passed.
Surgical Procedure for Nerve Recording
After 4 weeks of recovery, rats were prepared for nerve recording during euglycemic infusion of insulin. Anesthesia was induced with 40 mg/kg methohexital sodium IP (Brevital, Eli Lilly Co) and sustained with chloralose (Sigma Chemical Co; 50 mg/kg IV initially, followed by 25 mg/kg per hour IV infusion). The trachea was cannulated and each rat allowed to breathe oxygen-enriched air spontaneously. Body temperature was kept near 37.5°C with a temperature-controlled surgical table. Arterial pressure was monitored with a pressure transducer (Statham P23XL) and displayed continuously on a polygraph (model 7E, Grass Instrument Co). Heart rate was recorded from a linear cardiotachometer (Grass model 7P4). Multifiber recordings of lumbar SNA were obtained as previously described.3 Briefly, a midline abdominal incision was made, and a lumbar sympathetic nerve was placed on a bipolar platinum electrode (Cooner Wire Co) and covered with silicone gel (Sil-Gel 604, Wacker-Chemie). Nerve signals were amplified 20×103 to 100×103 and filtered at low- and high-frequency cutoffs of 100 and 1000 Hz, respectively, with a preamplifier (Grass model P511). The amplified and filtered neurograms were routed to a nerve traffic analyzer (model 706C, University of Iowa Bioengineering) which counted the action potentials that exceeded a threshold voltage set just above the noise level. A counter time bin was set at 1 second so that the impulse frequency for SNA was displayed on the polygraph as the number of spikes collected each second (hertz) as a time-frequency histogram. For each experiment, baseline SNA was set between 40 and 80 Hz.
Hyperinsulinemia With Euglycemic Clamp
Regular insulin (Iletin, Eli Lilly Co; 0.25 U/mL) in 50% rat plasma in isotonic saline was administered through the femoral vein with an infusion pump (model 255, Sage Instruments) at rates of 4.3 and 8.5 μU/min to obtain doses of approximately 0.06 and 0.13 U/h, respectively. Arterial BG levels were measured every 5 minutes before and during insulin infusion with a portable glucometer (Glucometer II, model 5625, Miles Laboratory) that had been calibrated against a glucose analyzer (model 27, Yellow Springs Instrument Co). For maintenance of baseline BG, or euglycemia, 50% glucose in sterile water was infused at variable rates through the jugular vein with an adjustable peristaltic pump (Rabbit Peristaltic Pump, Rainin Instrument Co). Reported BG values were determined with the glucose analyzer, and PI levels were measured by radioimmunoassay.15
The goal of the protocol was to determine the effects of hyperinsulinemia on lumbar SNA in rats with either SFO lesions (n=12) or AV3V lesions (n=4) and in rats with either sham SFO lesions (n=5) or sham AV3V lesions (n=5). In control experiments, the vehicle for insulin was infused in rats with sham lesions to the SFO (n=5) and in rats with sham lesions to the AV3V (n=5). Thus, these experiments consisted of six groups: four receiving insulin and two receiving the vehicle for insulin. In all rats, basal levels of MAP, HR, lumbar SNA, BG, and PI were obtained during a 15-minute control period. These parameters were then monitored during 60 minutes of 0.06 U insulin/h followed by 60 minutes of 0.13 U insulin/h in the four insulin-infused groups or by identical volume infusion of vehicle in the two vehicle-infused groups. PI and BG were obtained at the end of the 120-minute infusion period.
At the end of recording, deeply anesthetized rats were perfused transcardially with physiological saline followed by 10% formalin. Brains were removed and stored in fixative. Frozen 40-μm sections through the area of the lesion in each brain were obtained and stained with cresyl violet. The extent of brain damage was determined by light microscopic examination in a single-blind fashion.
Data were analyzed with appropriate single or repeated measures ANOVA and presented as mean±SE. Post hoc comparisons were made with Fisher’s least significant difference tests. Differences between groups were considered significant at a value of P<.05.
Verification of Lesions
Rats were considered to have SFO lesions if there was approximately 90% to 100% destruction of the nucleus, including destruction of its ventral pole, from which its efferents largely exit. These lesions typically damaged the thalamic periventricular nuclei and ventral hippocampal commissure (Fig 1⇓).
AV3V lesion placement was verified as previously described.16 17 Lesions shared a common area of damage to the periventricular tissue surrounding the optic recess. The lesion consistently destroyed the periventricular nucleus at the preoptic level, the median preoptic nucleus, and the OVLT. Some bilateral damage was usually present at the medial edge of the preoptic nuclei.
Baseline values and responses to the infusion procedure were equivalent in SFO sham-lesioned rats and AV3V sham-lesioned rats receiving the vehicle for insulin; therefore, these two groups were pooled into a single sham-vehicle group (n=10, Table⇓). In a similar fashion, baseline values and responses to the protocol were similar in SFO sham-lesioned rats and AV3V sham-lesioned rats receiving insulin; therefore, these two groups were pooled into a single sham-insulin group (n=10). The rats with SFO lesions receiving insulin (SFO-insulin, n=12) and those with AV3V lesions receiving insulin (AV3V-insulin, n=4) remained the same as described in “Methods.” Comparisons of baseline values across the sham-vehicle, sham-insulin, SFO-insulin, and AV3V-insulin groups revealed no differences in BG, MAP, or HR. However, the SFO-insulin group had significantly greater PI levels compared with the sham-insulin and AV3V-insulin groups.
Responses to Vehicle and Insulin
ANOVA of BG levels revealed a group–by–repeated measures interaction (P<.01), reflecting increasing BG levels during the 8 μL/min infusion period in the sham-vehicle group (P<.01), contrasting with no change in BG in the other experimental groups (Table⇑). PI levels did not change in the sham-vehicle group but increased in a stepwise fashion during the 4 μL/min (P<.01) and 8 μL/min (P<.01) infusion periods in the sham-insulin, SFO-insulin, and AV3V-insulin groups. The increase in PI did not differ among these three groups. The glucose infusion rate necessary to maintain stable levels of glycemia increased in a stepwise fashion in the sham-insulin, SFO-insulin, and AV3V-insulin groups during the 4 μL/min (P<.01) and 8 μL/min (P<.01) infusion periods and did not differ among the groups. Vehicle and insulin infusions did not alter MAP and HR.
Lumbar SNA did not change during vehicle infusions in rats with sham lesions (sham-vehicle group, Fig 2⇓). In sham-lesioned rats receiving insulin infusion (sham-insulin group), lumbar SNA rose significantly during low-dose insulin (P<.01). During high-dose insulin, lumbar SNA continued to rise and increased significantly (P<.001) over levels attained during the low dose. Contrasting with this increase in lumbar SNA, and in agreement with previous findings,4 insulin infusion had no effect on lumbar SNA in rats with AV3V lesions (AV3V-insulin group). In rats with SFO lesions (SFO-insulin group), insulin elicited a rise in lumbar SNA that was indistinguishable from the rise in lumbar SNA observed in rats with sham lesions.
The present studies replicated our previous findings showing that lesions of the AV3V region abolish increases in sympathetic activity to intravenous insulin administration.4 The major new finding is that SFO lesions had no effect on sympathoexcitation to insulin. These data indicate that the SFO itself and fibers originating from the SFO, which traverse and synapse with the median preoptic nucleus and OVLT, both within the AV3V, are not essential in mediating elevations in lumbar SNA to euglycemic hyperinsulinemia. However, it should be recalled that intact sympathetic responses in SFO-lesioned rats do not rule out the possibility that this nuclear group plays a role in mobilizing sympathetic elevations to systemic insulin in unlesioned (ie, normal) rats.
Interruption of sympathoexcitation to insulin by AV3V ablation, if not related to the SFO, could be mediated by several other brain regions. A likely candidate is the OVLT, one of the nuclear groups located within the AV3V region. The OVLT, like the SFO, is a circumventricular organ lacking a blood-brain barrier that monitors blood concentrations of peptide hormones, such as insulin, and transmits this information into the central nervous system.7 In accord with a blood-monitoring role for the OVLT, this region contains receptors for angiotensin II,18 19 atrial natriuretic peptide,19 and relaxin, a member of the insulin family of polypeptide hormones.20 Furthermore, the OVLT contains high concentrations of insulin-specific binding sites in cytoarchitectonically discrete regions9 and sends efferent projections to regions directly controlling sympathetic neural outflow.7 21 22
Lesions of the AV3V region may eliminate increased sympathetic activity in response to hyperinsulinemia by interrupting insulin uptake in the cerebral ventricles. In support of such a possibility, intravenous insulin administration in rats causes increased plasma insulin levels that are matched by parallel increases in CSF insulin.23 Once in the CSF, insulin may bind to receptors located on tanycyte cells lining the surface of the third ventricle.10 It has been postulated that these “CSF contacting neurons” participate in the uptake of insulin from the CSF into neuronal sites behind the brain-CSF barrier.7 10 Baskin and colleagues24 provided direct experimental support for this hypothesis by injecting radiolabeled insulin into the lateral ventricles of rats and found the highest radioactivity from both autoradiographic and microdissection procedures in the periventricular regions lining the third cerebral ventricle. Important in this context is our recent demonstration that insulin infusion into the third cerebral ventricle increases lumbar sympathetic activity in normotensive rats.5 Because AV3V lesions destroy the preoptic periventricular tissues, interruption of sympathoexcitation in the present study may be due to destruction of insulin-transporting tanycytes and consequent elimination of insulin transport from the CSF into the brain.
As a final possibility, it should be recalled that the AV3V region is part of an extensive neural network richly interconnected with other hypothalamic nuclei that receive input from cardiovascular and pulmonary sensory systems.7 AV3V lesions may interrupt fibers of passage from these regions, many of which could be important in generating sympathetic increases to intravenously administered insulin.
In summary, the present study demonstrated that lesions of the AV3V region abolish increases in sympathetic activity to intravenous insulin infusion, whereas lesions of the SFO have no effect on insulin-induced sympathoexcitation. These data indicate that the SFO itself and fibers originating from the SFO are not essential in mediating elevations in lumbar SNA to euglycemic hyperinsulinemia. Thus, by eliminating the SFO, we have ruled out the likelihood that AV3V lesions abolish sympathetic increases to insulin simply by removing critical input from the SFO that is transmitted to or through the AV3V for the mobilization of sympathetic outflow.
Selected Abbreviations and Acronyms
|AV3V||=||anteroventral portion of the third ventricle|
|MAP||=||mean arterial pressure|
|OVLT||=||organum vasculosum of the lamina terminalis|
|SNA||=||sympathetic nerve activity|
This research was supported in part by National Research Service Award HL-08704 and National Institutes of Health Grants HL-14338 and HL-44546 from the National Heart, Lung, and Blood Institute; by research funds from the Department of Veterans Affairs; and by a grant from the City University of New York Internal Awards Program, PSC-CUNY.
- Received June 19, 1996.
- Revision received September 14, 1996.
- Accepted October 29, 1996.
Anderson EA, Hoffman RP, Balon TW, Sinkey CA, Mark AL. Hyperinsulinemia produces both sympathetic neural activation and vasodilation in normal humans. J Clin Invest. 1991;87:2246-2252.
Morgan DA, Balon TW, Ginsberg BH, Mark AL. Nonuniform regional sympathetic nerve responses to hyperinsulinemia in rats. Am J Physiol. 1993;264:R423-R427.
Muntzel MS, Morgan DA, Mark AL, Johnson AK. Intracerebroventricular administration of insulin produces nonuniform increases in sympathetic nerve activity. Am J Physiol. 1994;267:R1350-R1355.
Anderson EA, Mark AL. The vasodilator action of insulin: implications for the insulin hypothesis of hypertension. Hypertension. 1993;21:136-141.
Johnson AK, Gross PM. Sensory circumventricular organs and brain homeostatic pathways. FASEB J. 1993;7:678-686.
Lind RW, Johnson AK. Subfornical organ-median preoptic connections and drinking and pressor responses to angiotensin II. J Neurosci. 1982;2:1043-1051.
McKinley MJ, McAllen RM, Mendelsohn FAO, Allen AM, Chai SY, Oldfield BJ. Circumventricular organs: neuroendocrine interfaces between the brain and the hemal milieu. Front Neuroendocrinol. 1990;11:91-127.
Yalow RS, Berson SA. Immunoassay of endogenous plasma insulin in man. J Clin Invest. 1960;39:1157-1175.
Buggy J, Johnson AK. Pre-optic-hypothalamic periventricular lesions: thirst deficits and hypernatremia. Am J Physiol. 1977;233:R44-R52.
Osheroff PI, Phillips HS. Autoradiographic localization of relaxin binding sites in rat brain. Proc Natl Acad Sci U S A. 1991;88:6413-6417.
Brody MJ, Johnson AK. Role of the anteroventral third ventricle region in fluid and electrolyte balance, arterial pressure regulation, and hypertension. In: Martini L, Ganong WF, eds. Frontiers in Neuroendocrinology. New York, NY: Raven Press Publishers; 1980;6:249-292.
Fink GD, Buggy J, Haywood JR, Johnson AK, Brody MJ. Hemodynamic effects of electrical stimulation of forebrain angiotensin and osmosensitive sites. Am J Physiol. 1978;235:H445-H451.