0306-4522/90$3.00+ 0.00 Pergamon Pressplc Q 1990IBRO
Neuroscience Vol. 34, No. 2, pp. 433-440, 1990 Printed in GreatBritain
THE SECRETION OF CATECHOLAMINES, CHROMOGRANIN A AND NEUROPEPTIDE Y FROM THE ADRENAL MEDULLA OF THE CAT VIA THE ADRENOLUMBAR VEIN AND THORACIC DUCT: DIFFERENT ANATOMIC ROUTES BASED ON SIZE S.W.
CARMICHAEL,*S.
L. STODDARD,~D. T. O’CONNORJ T. L. YAKF.H§ and G. M. TYCEJ/
Departments of *Anatomy, $Neurosurgery and //Physiology and Biophysics, Maya Clinic~Foundation, Rochester, MN 55905, U.S.A. TDepartment of Anatomy, Indiana Unive~ity School of Medicine, Fort Wayne, IN 46805, U.S.A. fDepartment of Medicine, School of Medicine, University of California and Veterans Administration Medical Center, San Diego, CA 92161, U.S.A. Abstract-Secretion of the adrenal medulla was stimulated in nine cats by insulin-induced hypoglycemia. Levels of catecholamines (mol. wt 1533183), neuropeptide Y (mol. wt 4254) and chromogranin A (mol. wt 48,000) were measured in concurrently collected samples of adrenolumbar venous blood and thoracic duct lymph for up to 4 h following insulin administration. Insulin-induced hypoglycemia elicited an increase in the secretion of catecholamines, which reached peak levels in the adrenolumbar venous plasma at 1.5-2 h and in the lymph at 2.5 h. Although catecholamines were the most numerous measured molecules in the lymph, levels of norepinephrine and epinephrine were 75-250-fold less than those found in the adrenolumbar venous plasma. Neuropeptide Y in the adrenolumbar venous plasma reached peak levels between 1 and 1.5 b; at this time approximately 20% of the peak venous amount was detected in the lymph. Chromogranin A was found in approximately equal amounts in both plasma and lymph; the peak level in the plasma occurred at 1.5-2 h, while that in the lymph was reached at 2-3 h. We suggest that the size of a molecule influences the route it takes following exocytosis from the chromatlin vesicle, Smaller molecules such as catecholamines may pass directly into the circulation, while larger molecules such as chromogranin A may be temporarily sequestered in the interstitial space before passing into the lymph, and hence into the circulation.
cells of the adrenal medulla release the contents of the chromaffin vesicle \da exocytosis. These contents include catecholamines, proteins (including chromogranin A), peptides (including neuropeptide Y), nucleotides and ions.” While the m~hanisms of stimulus-secretion coupling leading up to exocytosis have been well studied,29 the pathway taken by the various molecules following release from the cell is not known. The fact that catecholamines such as epinephrine exert their effects via the bloodstream has been appreciated since the experiments of Oliver and Schifer.25 Dopamine /?hydroxylase (DBH), one protein known to be released during exocytosis of the chromaffin vesicle, can also be detected in serum.36 Additionally, at least some of the D/IH released by components of the sympathetic nervous system reaches the vascular system via l~p~atics. ‘,I7 Pinardi et af. (Ref. 27, Fig. 6) demonstrated that, under conditions of hemorrhagic hypotension in the dog, at least half the adrenal output of D/?H (U/min) entered the circulation via the thoracic duct lymph.
The chromaffin
Abbreviations: CA, catecholamine; CgA, chromogranin; DA, dopamine; DBH, dopamine p-hydroxylase; EDTA, ethylenediaminetetra-acetate; NE, norepinephrine; NPY, neuropeptide Y; RIA, radioimmunoassay.
More recently, O’Connor et at.24 reported that the adrenal medulla is the principal source of increments in plasma chromogranin A in human subjects who were tested under conditions of insulin-induced hypoglycemia. In these studies the peak levels of chromogranin A in the plasma lagged behind the peak level of catecholamines by about 1 h, suggesting that this protein, like DPH, may enter the bloodstream via lymphatics. The present study was undertaken to examine the potential routes taken by secretory products from the adrenal medulla, specifically the adrenal vein (adrenolumbar vein in the cat) and thoracic duct. Based on the observations of Pinardi et aLz’ and O’Connor et a1.,24it was hypothesized that molecules of different size would take different routes; smaller molecules might enter the venous system directly, while larger molecules might accumulate in the interstitial fluid and enter lymphatics. Therefore, molecules of different sizes [catecholamines: epinephrine, mol. wt 183.2; norepinephrine (NE), mol. wt 169.2; dopamine (DA), mol. wt 153.2; neuropeptide Y (NPY), mol. wt 4254; and chromogranin A (CgA), mol. wt 48,000] were measured in the adrenolumbar venous plasma and thoracic duct lymph after insulin-induced hypoglycemia.
433
S. W.
434 EXPERIMENTAL Mongrel
CARMICHAEL
PROCEDURES
cats [rV = 9; mean body weight (kg) 2 S.E.M. = 3.83 + 0.251 were anaesthetized by hafothane inhalation. Resptratory effort was abolished with pancuronium bromide (Pavulon) and the animals were maintained by positive pressure ventilation with 1.5-2.0% halothane in oxygen. Body temperature was maintained at 37..5-385°C with heating pads. The left adrenolumbar vein was exposed through a lateral abdominal incision. The vessel was cannulated from a lateral approach (20 ga IV placement unit) using a method adapted from Hume and Nelson;” tension on a suture loop around the adrenolumbar vein between the adrenal gland and the inferior vena cava (or renal vein) caused blood flowing through the adrenal gland to be diverted into the cannula. A cannula (PE 90) placed in the femoral vein was used to withdraw peripheral blood samples and for the administration of fluid and drugs, while another cannula (PE 90) in the femoral artery permitted the continuous monitoring of heart rate and blood pressure. A thoracotomy was performed by sectioning three ribs to the left of the sternum. Careful dissection exposed the thoracic duct caudal to the junction of the left subclavian and internal jugular veins. The thoracic duct was cannulated (20 ga IV placement unit); lymph was collected continuously through this cannula. Blood volume was maintained with a Plasmalyte drip into the femoral vein. Additionally, red blood cells and lymph were returned after each sampling period. Lymph collected between samples was also returned. Following completion of the surgical procedures, the animals were heparinized (200 U/kg body wt) and allowed to stabilize, at which time baseline adrenolumbar venous. femoral venous and i~phati~ samples were collected. All samples were collected in chilled, untreated plastic tubes. Femoral venous blood samples were centrifuged and levels of glucose in the plasma determined by enzymatic assay (Dri-Stat Glucose-HK Endpoint Reagent; Beckman). Adrenolumbar venous and lymph samples were divided into small and large aliquots of, respectively, l/3 and 2/3 of the total volume collected. Chromogranin A was determined in the smaller samples by a homologous human CgA radioimmunoassay (RIA) which is a rapid, sensitive modification” of previously published procedures.‘9.2’ ~eliminary experiments demonstrated that feline CgA showed parallel crossreactivity (tracer displacement) in the human CgA RIA (data not shown). Albumin concentration was also measured in plasma and lymph samples using the Bromcresol Green calorimetric method (Sigma; Diagnostic Kit No. 631). The larger aliquots of adrenolumbar venous blood and lymph were mixed with EDTA (0.5 ml; 15%). NPY in the plasma of the adrenolumbar venous sample and in lymph was determined by radioimmunoassay.~ For the measurement of catecholamines, sodium metabisulfite (20 $/ml of a 5% solution) was added to the samples of adrenolumbar venous plasma and lymph. Catecholamines were measured by high performance liquid chromatography with electrochemical detection.5 All plasma and lymph samples were frozen at -80°C until assayed. After the collection of the baseline samples, adrenal medullary secretion was stimulated by the induction of hypoglycemia by intravenous injection of insulin (IU/kg). Samples of approximately 5 ml were collected at 0.25, 0.5, 0.75, 1, 1.5, 2, 2.5, 3, 3.5 and 4 h after injection of insulin. In five cats, both the rate of blood flow from the adrenolumbar vein and the rate of lymph flow from the thoracic duct were determined during the sampling periods. In three cats, both adrenal glands were isolated after collection of the 4 h sample by tightly ligating all affluent and effluent connections and cauterizing the pedicle; lymph samples were collected at 0.5 and 1.5 h after these procedures. The Wilcoxon Test for paired observations and a one-way analysis of variance were used, as indicated. to evaluate the
ei
ul
data. Differences between groups were considered significant if P < 0.05. RESULTS
The level of glucose in the systemic plasma fell to approximately 50 mg/dl within 30 min of insulin administration and remained below this level for the remainder of the experiment (Fig. 1). Catecholamines in adrenolumbar wnous plasma and ~~~~h
Catecholamines were more numerous on a molar basis than either NPY or CgA in both the adrenolumbar venous plasma and the lymph, as measured in pmol/ml (Figs 24). The concentration of all cateGLUCOSE AFlER
zoJ s 0
IN THE FIlMORAL
INSULIN
ADMINISTRATION
2
/ TIME
IN HOURS
VEIN
(1 U/KG)
3
4
AFTERINSULIN
Fig. 1. Level of glucose in plasma from femoral vein (mg/ deciliter). Values are mean -I-S.E.M. (N = 9).
TIME IN HOURS
AFTER INSULIN
Fig. 2. Levels of cat~holam~nes (NE, top; epinephrine, middle; DA, lower) in prnol/ml in adrenolumbar venous plasma (left Y-axes) and thoracic duct lymph (right Y-axes). Value are mean +S.E.M. for venous levels and mean -S.E.M. for lymphatic levels (N = 9).
435
Different size molecules secreted from adrenal medulla Table 1. Baseline levels of autocoids (pmol/ml) in adrenolumbar venous plasma and thoracic duct lymph
Norepinephrine (N = 9) ~pinephrine (N = 9) Dopamine (N = 9) Total catecholamines Neuropeptide Y (N = 9) Chromogranin A (N = 8)
Lymph 2.17+0.60 0.40 i 0.22 0.73 + 0.51 3.30 f 0.71 0.15+0.04 0.20 * 0.05
NEUROPEPTIDE
ADRENOLUMBAR
VEIN
Y
VS THORACIC
DUCT
Plasma 20.15f 14.35 26.52 & 12.27 0.51 + 0.14 47.18 + 20.76 0.21 f 0.04 0.23 f 0.04
Values are mean + S.E.M. eholamines in the adrenolumbar vein increased 30 mm after insulin administration and remained elevated for the duration of the experiment (Fig. 2). Insulin also preferentially increased the secretion of epinephrine from the adrenal medulla, as was indicated by an alteration of the ratio of NE:epinephrine at baseline from 1: 1.3 up to 1: 5.3. The changes in the NE : epinephrine ratio were significantly different from baseiine at 1.5 and 2.0 h. NE and epinephrine reached peak levels in the adrenolumbar vein at I .5-2.0 h after insulin injection; peak levels of NE and epinephrine in the thoracic duct lymph tended to occur later, at 2.5 h. Baseline levels of NE and epinephrine were notably greater in the adrenolumbar vein (nine- and 66-fold, respectively), while DA in the thoracic duct was l.Cfold higher than in the adrenolumbar vein (Table 1). However, following insulin-induced hypoglycemia, the levels of NE and epinephrine in the adrenolumbar vein exceeded the values in the lymph even further, rising up to 75-fold greater for NE and about 250fold greater for epinephrine (note the different right side Y-axis scales in Fig. 2, top and middle). DA in the thoracic duct fell to less than 1 pmol/ml (Fig. 2, bottom). The baseline samples of adrenolumbar venous plasma contained approximately equal proportions of NE and epinephrine (43 and 56%, respectively), while baseline lymph samples contained 66% NE and 12% epinephrine, in addition to 22% DA. Following insulin-induced hypoglycemia, the relative proportions of catecholamines changed in both plasma and lymph. Adrenolumbar venous plasma epinephrine increased to 84% of the total catecholamines between 45 and 90 min, while thoracic duct epinephrine increased to 60% of the total catecholamines at 90 min. Later in the experiment, between 3 and 4 h, epinephrine still comprised about 80% of the catecholamines in the adrenolumbar venous plasma, while NE in the lymph increased to between 60 and 70%, perhaps reflecting increased peripheral sympathetic activation.
TIME IN HOURS
Fig. 3. Levels of NPY (pmol/ml) in adrenolumbar venous plasma and thoracic duct lymph. Values are mean +S.E.M. as in Fig. 2 (N = 9). C~romogra~i~ A in adreno~~~ar
Y in ffdre~o~~rn~ar venous plasma and
&mph
The mean levels of NPY peaked between 1 and 1.5 h in the adrenolumbar venous plasma, while NPY in the thoracic duct gradually declined throughout the experiment (Fig. 3).
venous plasma and
lymph
Chromogranin A was found in approximately equal concentrations in both the adrenolumbar venous plasma and thoracic duct lymph (Fig. 4). Peak levels of CgA in the adrenolumbar vein were reached at 1.5-2 h following insulin injection; levels at these times were significantly greater than baseline (P < 0.002). In contrast, peak levels of CgA in the thoracic duct lagged behind the plasma, being reached at 2-3 h; levels at 2.5 and 3 h were significantly higher than baseline (P = 0.05). In the five cats in which the rate of lymph and adrenolumbar venous blood flow rates were calculated, CgA reached its peak secretion (pmol/min) at 1 h in the adrenolumbar vein, while a three-fold higher peak was reached in the thoracic duct at 2.5 h (Fig. 5). Neither adrenolumbar venous nor lymph flow, as measured by ANOVA, changed significantly during the course of the experiment. Expression of catecholamine and NPY data as pmolfmin did not reveal different patterns of release than those shown in Figs 2 and 3, with the data expressed as pmol/ml.
CHROMOGRANIN ADRENOLUMBAR
2
VEIN
A
VS THORACIC
ADRENOLUMBAR .-.THORACIC DUCT
o-o
8 P
~e~ropeptide
AFTER INSULIN
0.1 1
0
TIME
2
IN HOURS
3
AFTER
OUCl
VEIN
I
4
INSULIN
Fig. 4. Levels of chromogranin A (pmol/ml) in adrenolumbar venous plasma and thoracic duct lymph. Values are mean +S.E.M., as in Fig. 2 (N = 9). Levels at 1.5 and 2 h (vein) and 2.5 and 3 h (duct) are significantly greater than baseline.
436
S. W. CARMICHAEL et al. CHROMOGRANIN
A SECRETION
RAIE
00’ 0
1
TIME
2
IN HOURS
4
3
AFTER
INSULIN
Fig. 5. Secretion rate of chromogranin A (pmol/min) in adrenolumbar venous plasma and thoracic duct lymph. Values are mean +S.E.M. for lymphatic levels and mean
-S.E.M.
for venous levels (iv = 5).
In the three cats in which the adrenal glands were isolated following the 4 h sample, the mean (+S.E.M.) level of CgA in the thoracic duct at 4 h was 0.28 f 0.09 pmol/ml. The CgA content dropped at 30 and 90 min after adrenal isolation to, respectively, 0.10 + 0.06 and 0.01 + 0.05 pmoI/ml lymph. Albumin
in adrenolumbar
oenous plasma
and lymph
A non-parametric r-test compared the mean levels of albumin in plasma and lymph for all samples for each cat (N = 9). The mean albumin concentration in plasma of 1.51 & 0.11 mg/dl was significantly greater (P < 0.002, two-tailed) than the mean albumin concentration in all lymph samples for each cat (0.60 f 0.10 mg/dl). A one-way analysis of variance revealed that the concentration of albumin in both plasma and lymph declined significantly during the course of the experiment (P = 0.03 and P = 0.00, respectively). The “CgA quotient” in both plasma and lymph was determined by expressing pgrn CgA/mg albumin. The mean CgA quotient for lymph in all samples for each cat (3.79 & 0.53 pgrn CgA/mg albumin) was significantly greater (P < 0.002) than the CgA quotient for plasma in all samples for each cat (1.24 f 0.24; N = 8). In contrast to the changes of albumin concentration, the CgA quotient increased significantly during the course of the experiment in both plasma (P < 0.002) and lymph (P d 0.001). A Wilcoxon test determined the lymph CgA quotient to be significantly greater than the plasma CgA quotient at 2.5 and 3.0 h (4.52 + 0.75 vs 1.73 + 0.55 and 6.05 + 1.65 vs 1.45 k 0.62, respectively); these were times at which CgA reached its peak levels in thoracic duct lymph.
DISCUSSION
Insulin-induced hypoglycemia has been shown to be a strong and relatively specific stimulus for secre-
tion from the adrenal medulla, although stimulation of other components of the sympathetic nervous system may occur.34 In the present study hypoglycemia was induced following insulin administration; an increase in catecholamines in the adrenolumbar venous plasma was observed when plasma glucose levels fell to 50 mg/dl. The increase in the adrenal secretion of epinephrine as shown by significant changes in the molar ratio of NE:epinephrine at 1.5 and 2.0 h established that insulin provided a potent stimulus for adrenal medullary activation. Measurement of albumin was employed in this study as an index of the efficacy of thoracic duct cannulation and, hence, lymph collection. Human plasma contains about 4.5 g/d1 of albumin,” which is several fold greater than the concentration found in lymph. ” Thus , our observed ratio of albumin in plasma: lymph of approximately 2.5 establishes that we were indeed collecting fluid from the lymphatic system. The significant decrease in albumin concentration in both plasma and lymph throughout the course of the experiment was most likely caused by hemodilution secondary to the addition of intravenous fluids. Pathway
,fkr
secreted
adrenal
medullary
cate-
cholamines
In line with previously established theories, most of the catecholamines secreted from the adrenal medulla entered the circulation directly via the adrenolumbar vein (Fig. 2). Somewhat surprising, however, was the observation that the levels of catecholamines in the thoracic duct, while up to 200 orders of magnitude lower than those in the adrenolumbar vein, also peaked in response to insulin-induced hypoglycemia. The peak of catecholamines in the lymph was approximately 0.5 h later than that observed in the vein. These data suggest that most, but not all, of the catecholamine molecules released from the adrenal chromaffin cell exit the adrenal gland via the adrenolumbar vein. A minor fraction of adrenal medullary catecholamines accumulated in the interstitial space and, after a lag period, entered the lymphatic system. Although we do not propose that the catecholamines traveling through the thoracic duct have a specific physiological function, it is of interest that smooth muscle is found in the thoracic duct; the motility of this muscle is affected by adrenergic agonists and antagonists.” Pathways
.fkr secreted
adrenal
medullary
neuropep-
tide Y
NPY in the cat is believed to be co-stored in chromaffin vesicles with epinephrine.15 Most of the NPY secreted from the adrenal gland entered the circulation via the adrenolumbar vein (Fig. 3). However, at the time NPY peak in the vein (l-l.5 h), between 17 and 24% of the peak amount was detected in the thoracic duct. In contrast, generally less
Different size molecules secreted from adrenal medulla than 1% of catecholamines were found in the thoracic duct at peak times. It is of interest that the NPY peak in the adrenolumbar vein was not reflected in the lymph, perhaps due to rapid metabolism of this peptide. Pathways for mogranin A
secreted
adrenal
medullary
chro-
In the three cats in which the adrenal glands were isolated from the circulation after 4 h, levels of CgA dropped sharply, suggesting that the adrenal glands were the primary source of CgA detected in the thoracic duct in these experiments. Further analysis of our data is based upon this assumption. Peak levels of CgA in the thoracic duct lagged behind those in the adrenolumbar vein by about 1 h (Fig. 4). Interestingly, the peak levels in the vein and duct were approximately equal, suggesting that about the same amount of this protein leaves the adrenal gland by each route. It is important to note, however, that the absolute level of CgA in plasma or lymph is related to the flow rate of these substances. Thus, consideration of the CgA data on the basis of rate of flow of adrenolumbar blood or lymph (Fig. 5) indicated that greater amounts of CgA were leaving the adrenal medulla via the lymph than the adrenolumbar venous blood. Furthermore, the CgA quotient (pgm CgA/mg albumin) was significantly greater in the lymph than in the adrenolumbar venous plasma, when these values were compared in each cat. This observation suggests a preferential transport of CgA into the lymph, rather than merely an increased transport of proteins (such as albumin). The general increase in the CgA quotient throughout the experiment in the light of decreasing albumin values further argues for the favored movement of this protein into the lymph. The central vein of the adrenal gland drains both the medulla and cortex before emptying into the adrenolumbar vein. Thus, it is not possible for us to distinguish definitively between the cortex and medulla as the source of the substances measured in this study. However, catecholamines, NPY and CgA are well recognized contents of adrenal chromaffin vesicles.37 Further, immunohistological observations have established that CgA is present in the adrenal medulla, but not in the cortex.20 Thus we have assumed that the substances measured in the current study were secreted from the adrenal medulla. Comparison
of present
results with a previous study
It is of particular interest to compare our results with the observations of Pinardi et al*’ These investigators measured levels of catecholamines and dopamine /I-hydroxylase (DfiH) in the adrenolumbar vein and thoracic duct of dogs following hemorrhagic hypotension. In agreement with Pinardi et al., we found that baseline levels of catecholamines in adrenolumbar venous plasma were about 15-fold
431
greater than levels in thoracic duct lymph. However, the peak catecholamine levels reported by Pinardi and colleagues were only about four-fold greater in the adrenolumbar vein than in the thoracic duct, while we noted catecholamine levels 130-l 88-fold greater in adrenolumbar venous plasma at the time of peak response. The difference in these observations could be due to the use of different animals (dog vs cat), or the use of different stressors. Insulin-induced hypoglycemia is known to specifically activate the adrenal medulla,34 while hemorrhagic hypotension has more widespread effects on the sympathetic nervous system. It is of interest that Pinardi et al. observed the catecholamine peak at the same time in the adrenolumbar vein and thoracic duct, while in our animals the catecholamine peak in lymph lagged behind that in the adrenolumbar vein by 30-60 min. The protein measured by Pinardi et al. D/?H (mol. wt 64,86213), is notably larger than the protein, CgA (mol. wt 48,000), measured in the present study. At the time of peak levels, we found approximately equal amounts of CgA in the adrenolumbar venous plasma and lymph (Fig. 4), while Pinardi et al. (Ref. 27, Figs 3 and 5) found about twice as much D/?H in the lymph as in the adrenolumbar venous plasma. Of particular interest is the observation that reinfusion of blood into Pinardi’s animals resulted in a significant increase in thoracic duct DfiH. Pinardi and colleagues suggested that the reinfusionbound increase in D/?H was caused by a “wash out” of this protein from the interstitial space into the lymph. We suggest that the previous observations of Pinardi et al.*’ support the hypothesis that large molecules secreted from the adrenal medulla use the lymph, at least in part, as a pathway to the circulation. Comparison mogranin A
between
catecholamines
and
chro -
In the present study the peak levels of catecholamines and CgA occurred at the same time in the adrenolumbar vein. In contrast, insulin-induced hypoglycemia in human volunteers was accompanied by a 60-90 min delay of the plasma CgA peak after the catecholamine peak.” Chromaffin vesicles contain approximately 600 mM catecholamines (CA) and 2 mM CgA, giving a vesicular ratio of 300 : 1 (CA : CgA).4 The CA :CgA ratio in the adrenolumbar venous plasma baseline samples in the present study was 205 : 1, reasonably in agreement with previous reports. However, at the time of peak levels of catecholamines and CgA in adrenolumbar venous plasma, this ratio increased to 1369 : 1. In contrast, the baseline ratio of CA: CgA in lymph was 16.5:1, while the ratio at peak levels (2.5 h) was 14.5 : 1. Taken together these data suggest that there is a preferential transport of CgA into the lymph. It is also possible that, by the time the catecholamines reached the lymph, considerable reuptake or degradation had occurred, thus explaining
438
S. W. CARMICHAELet al.
the decrease in the CA : CgA ratio in lymph from that found in the chromaffin vesicle. The notably elevated levels of catecholamine compared to CgA in the adrenolumbar venous blood at the time of peak response might indicate a preferential transport of catecholamines into the bloodstream under conditions of hypoglycemic stress. Lymphatic drainage of the adrenal medulla The existence of lymphatic drainage from the adrenal medulla is controversial. Lymphatic capillaries have been demonstrated in the adrenal medulla of man3’ and cow,’ while other investigators have denied their presence in both manI and rat.33 Sulyok et a13’ demonstrated that, in the dog, reduction of adrenal venous outflow enhanced release of corticoids into the thoracic duct. These reports suggest either that the lymphatic drainage of the adrenal gland varies with species or that lymphatics are generally present but the methods to detect or demonstrate them are uncertain. Although we know of no description of lymphatic structures in the adrenal medulla of the cat, our results support the presence of lymphatic drainage in this animal. The absence of CgA in the thoracic duct lymph following the isolation of the adrenal glands combined with the identification of CgA in only the medullary portion of the adrenal glands clearly point to the adrenal medulla as the source of CgA measured in our cats. Size appears to influence the route used by a secreted ndrenal medullury molecule Studies employing the isolated retrogradely perfused bovine adrenal gland have established that catecholamines and CgA are co-released from the chromaffin cell.‘4.3H However, the methodology of these studies, i.e. cutting slits into the capsule of the gland to facilitate retrograde drainage, has prevented disclosure of the route taken by these secreted substances in uiro. The present results suggest that the size of a molecule released at the surface of the adrenal chromaffin cell influences the route that it takes to enter the vascular system. Smaller molecules, such as catecholamines, appear to take a direct route into the venous system draining the adrenal medulla. Thus, catecholamines may have an immediate effect on physiological processes since they enter the systemic circulation with minimal delay. Larger molecules, such as NPY and CgA, appear to be partially shielded from direct entry into the vein. The observation that more CgA than NPY was diverted from the vein into the lymph suggests that the larger molecule, the more it is shielded. The morphological barriers between the chromaffin cell and a vascular lumen are substantial, including endothelial cells, a continuous basement membrane, and a subendothelial space of variable width.6 Furthermore, the transcapillary movement of proteins is very slow. ‘* Intercellular junctions between
capillary endothelial cells may only open to 5-6 nm.26 Although fenestrated capillaries are found in the adrenal medulla, the diameter of the fenestrae, at least in the rat, is only about 50 nm.’ Thus, CgA, with an exceptionally high Stokes radius of about 8 nm,22 may not be able to pass readily into the vascular lumen. In fact, fenestrated capillaries may not be more permeable to macromolecules than nonfenestrated capillary endothelium; thus fenestrae may not be protein transport sites.28 In the present study, albumin (Stokes radius, 3.6 nm2*), a protein notably smaller than CgA, appeared to be transported differently across the capillary endothelium as suggested by the increased ratio of CgA to albumin, even under conditions of likely hemodilution at later time points in the experiment. Based on the observations presented in this study, we hypothesize that some of the larger molecules accumulate in the interstitial space, eventually move into lymphatic capillaries, and hence are returned to the circulation. The lag of the thoracic duct CgA peak behind the adrenolumbar venous CgA peak (Figs 4 and 5) is consistent with this hypothesis.
CONCLUSION
While the rapid translocation of catecholamines into the vascular system may result in an obvious physiological advantage to the organism, we may only speculate on the advantage of slower access of NPY and CgA to the circulation. Such speculation is based on the lack of clearly established physiological roles for NPY and CgA. Recently, NPY was shown to be secreted from the adrenal medulla in response to hemorrhagic hypotension.’ Additionally, NPY may potentiate the action of NE at adrenergic receptors.* Chromogranin A-derived peptides have been shown to exert a feedback control on chromaffin cell secretory activity.3’ Theoretically, such control of adrenal medullary secretion could be exerted over an extended time period (several minutes to a few hours) and then diminish as CgA is gradually eliminated into the venous and lymphatic systems. Perhaps the function of lymphatic-carried NPY and CgA will be clarified as the physiological functions of these molecules and their degradation products are revealed. Acknowledgements-The Kettelkamp for preparation
authors thank MS Julia of the graphs, and the following personnel for their expert performance of the various assays used in this study: MS Sharon Chinnow (catecholamine assays), MS Laurie Bale (glucose assays), MS Annie Chen (chromogranin A and albumin assays), and MS Diane Roddy (NPY assay). We also thank Drs William Engeland, Jose-Maria Trifaro and Richard Weinshilboum for helpful comments on this manuscript. This work was supported in part by grants to Dr Tyce- (NS.17858) and Dr -&Ionnor (Veterans Administration, NIH [DK-36,400; HL-35.0181. and the American Heart Association).
Different
size molecules
secreted
from adrenal
medulla
439
REFERENCES I. Aerg H. E., Hansson H. E., Wetterberg L., Ross S. B. and Friien 0. (1974) Dopamine-b-hydroxylase in human lymph. Life Sci. 14, 65-71. 2. Allen J. M., Yeats J. C., Adrian T. E. and Bloom S. R. (1984) Radioimmunoassay of neuropeptide Y. Reg. Pep?. 8, 61-70. 3. Alvarez-Morujo A. J. and Alvarez-Morujo A. (1983) Scanning electron microscopy of lymphatic system in the adrenal glands of the bull in three different situations. Morfol. Norm. Patol. Histol. 7A, 361-370. 4. Apps D. K. (1982) The reuptake of amines by secretory granules. Trends biochem. Sci. 7, 153-156. 5. Bove A. A., Dewey J. D. and Tyce G. M. (1984) Increased conjugated dopamine in plasma after exercise therapy. J. Lab. clin. Med. 104, 77-85. 6. Carmichael S. W., Brooks J. C., Malhotra R. K., Wakade T. D. and Wakade A. R. (1989) Ultrastructural demonstration of exocytosis in the intact rat adrenal medulla. J. Electron Microsc. Tech. 12, 316-322. 7. Elfvin L.-G. (1965) The ultrastructure of the capillary fenestrae in the adrenal medulla of the rat. J. ulfrastrucf. Res. 12, 687-704. 8. Evtquoz D., Waeber B., Corder R., Nussbergers J., Gaillard R. and Brunner H. R. (1987) Markedly reduced blood pressure responsiveness in endotoxemic rats: reversal by neuropeptide Y. Li& Sci. 41, 2573-2580. 9. Gaumann D., Yaksh T., Roddy D., Michner S., Nelson D. and Lucas D. (1986) Adrenal release of [Metlenkephalin and other neuropeptides during hypotensive hemorrhage in halothane anesthetized cat. Sot. Neurosci. Abstr. 12, 1484. 10. Guyton A. C. (1986) Textbook of Medical Physiology, 7th edn. W. B. Saunders, Philadelphia. 11. Hargens A. R., Tucker B. J. and Blantz R. C. (1977) Renal lymph protein in the rat. Am. J. Physiol. 233, F269-273. 12. Hume D. M. and Nelson D. H. (1955) Adrenal cortical function in surgical shock. Surg. Forum 5, 568-575. 13. Lamouroux A. A., Vigny A., Faucon Biguet N., Darmon M. C., Franck R., Henry J.-P. and Mallet J. (1987) The primary structure of human dopamine /I-hydroxylase: insights into the relationship between the soluble and membrane-bound forms of the enzyme. Eur. molec. Biol. Org. J. 6, 3931-3937. 14. Ledbetter F. H., Kilpatrick D., Sage H. L. and Kirshner N. (1978) Synthesis of chromogranins and dopamine fi-hydroxylase by perfused bovine adrenal glands. Am. J. Physiol. 235, FA75486. 15. Lundberg J. M., Hofelt T., Hemsen A., Theodorsson-Norheim E., Pernow J., Hamberger B. and Goldstein M. (1986) Neuropeptide Y-like immunoreactivity in adrenaline cells of adrenal medulla and in tumors and plasma of pheochromocytoma patients. Regl. Pep?. 13, 169-182. 16. Merklin R. J. (1966) Suprarenal gland lymphatic drainage. Am. J. Anat. 119, 359-374. 17. Ngai S. H., Dairman W., Marchelle M. and Spector S. (1974) Dopamine fi-hydroxylase in dog lymph-effect of sympathetic activation. Life Sci. 14, 2431-2439. 18. Noer I. and Lassen N. A. (1979) Evidence of active transport (filtration?) of plasma proteins across the capillary walls in muscle and subcutis. Acta physiol. stand. Suppl. 463, 105-l 10. 19. O’Connor D. T. and Bernstein K. N. (1984) Radioimmunoassay of chromogranin A as a measure of exocytotic sympathoadrenal activity in normal subjects and patients with pheochromocytoma. New Engl. J. Med. 311, 764779. 20. O’Connor D. T., Burton D. W. and Deftos L. J. (1983) Chromogranin A: immunohistology demonstrates its universal occurrence in normal polypeptide hormone producing endocrine glands. Z,ife Sci. 33, 1657-1663. 21. O’Connor D. T. and Deftos L. J. (1986) Secretion of chromogranin A by peptide producing endocrine neoplasms. New Engl. J. Med. 314, 1145-1151. 22. O’Connor D. T. and Frigon R. P. (1984) Chromogranin A, the major catecholamine storage vesicle soluble protein. J. biol. Chem. 259, 3237-3247. 23. O’Connor D. T., Pandian M. R., Carlton E., Cervenka J. H. and Hsiao R. .I. (1989) Rapid measurement of circulating human chromogranin A: in vitro stability, exploration of the neuroendocrine character of neoplasm, and assessment of the effects of organ failure. Clin. Chem. 35, 1631-1637. 24. O’Connor D. T., Pandian M. R., Cervenka J., Mezger M. S. and Parmer R. J. (1987) What is the source and disposition of chromogranin A (CgA) in normal human plasma? (Abstract.) C/in. Res. 35, 60SA. 25. Oliver G. and Schler E. A. (1894) On the physiological action of extract of the suprarenal capsules. J. Physiol., Lond. 16, i-iv. 26. Palade G. E., Simionescu M. and Simionescu N. (1979) Structural aspects of the permeability of the microvascular endothelium. Acta physiol. stand. Suppl. 463, 1 l-32. 27. Pinardi G., Talmaciu R. K., Santiago E. and Cubeddu L. X. (1979) Contribution of adrenal medulla, spleen and lymph, to the plasma levels of dopamine /l-hydroxylase and catecholamines induced by hemorrhagic hypotension in dogs. J. Pharmac. exp. Ther. 209, 176-184. 28. Renkin E. M. (1977) Multiple pathways of capillary permeability. Circulation Res. 41, 736-743. 29. Rosenheck K. and Lelkes P. I. (eds) (1987) Stimulus-Secretion Coupling in Chromafin Cells, Vols I and II. CRC Press, Boca Raton. 30. Sapin M. R. (1959) On adrenal intraorgan lymphatic system in man. Arkhiv. Anat. Histol. i Embryol. 36, 52-59. 31. Simon J.-P., Bader M.-F. and Aunis D. (1988) Secretion from chromaffin cells is controlled by chromogranin A-derived peptides. Proc. natn. Acad. Sci. U.S.A. 85, 1712-1716. 32. Sulyok A., Monos E., Koltay E., Bede A. and Kovach A. G. B. (1969) Effect of local venous congestion on adrenocortical function. Acta physiol. Acad. Sci. hung. 35, 313-320. 33. Verhofstad A. A. J. and Lensen W. F. J. (1973) On the occurrence of lymphatic vessels in the adrenal gland of the white rat. Acra anat. 84, 475483. 34. Viveros 0. H., Arqueros L., Connett R. J. and Kirshner N. (1969) Mechanism of secretion from the adrenal medulla. IV. The fate of the storage vesicles following insulin and reserpine administration. Molec. Pharmac. 5, 69-82. 35. Wang G. and Zhong S. (1985) Experimental study of lymphatic contractility and its clinical importance. Ann. Plast. Surg. 15, 278-284.
440
S. W. CARMICHAEL et al.
36. Weinshilboum R. and Axelrod J. (1971) Serum dopamine-beta-hydroxylase activity. Circulation Res. 2S, 3077315. 37. Winkler H., Apps D. K. and Fischer-Colbrie R. (1986) The molecular function of adrenal chromaffin granules: established facts and unresolved topics. Neuroscience 18, 261-290. 38. Winkler H., H&nag1 H., Schiif J. A. L., Hiitnagl H. and Zur Nedden G. (1971) Bovine adrenal medulla: synthesis and secretion of radioactively labelled catecholamines and chromogranins. Naunyn-Schmiedeberg’s Arch. Pharmac. 271, 193-203. (Accepted 17 July 1989)
Neuroscience Vol. 34, No. 2, pp. 433-440, 1990 Printed in GreatBritain
THE SECRETION OF CATECHOLAMINES, CHROMOGRANIN A AND NEUROPEPTIDE Y FROM THE ADRENAL MEDULLA OF THE CAT VIA THE ADRENOLUMBAR VEIN AND THORACIC DUCT: DIFFERENT ANATOMIC ROUTES BASED ON SIZE S.W.
CARMICHAEL,*S.
L. STODDARD,~D. T. O’CONNORJ T. L. YAKF.H§ and G. M. TYCEJ/
Departments of *Anatomy, $Neurosurgery and //Physiology and Biophysics, Maya Clinic~Foundation, Rochester, MN 55905, U.S.A. TDepartment of Anatomy, Indiana Unive~ity School of Medicine, Fort Wayne, IN 46805, U.S.A. fDepartment of Medicine, School of Medicine, University of California and Veterans Administration Medical Center, San Diego, CA 92161, U.S.A. Abstract-Secretion of the adrenal medulla was stimulated in nine cats by insulin-induced hypoglycemia. Levels of catecholamines (mol. wt 1533183), neuropeptide Y (mol. wt 4254) and chromogranin A (mol. wt 48,000) were measured in concurrently collected samples of adrenolumbar venous blood and thoracic duct lymph for up to 4 h following insulin administration. Insulin-induced hypoglycemia elicited an increase in the secretion of catecholamines, which reached peak levels in the adrenolumbar venous plasma at 1.5-2 h and in the lymph at 2.5 h. Although catecholamines were the most numerous measured molecules in the lymph, levels of norepinephrine and epinephrine were 75-250-fold less than those found in the adrenolumbar venous plasma. Neuropeptide Y in the adrenolumbar venous plasma reached peak levels between 1 and 1.5 b; at this time approximately 20% of the peak venous amount was detected in the lymph. Chromogranin A was found in approximately equal amounts in both plasma and lymph; the peak level in the plasma occurred at 1.5-2 h, while that in the lymph was reached at 2-3 h. We suggest that the size of a molecule influences the route it takes following exocytosis from the chromatlin vesicle, Smaller molecules such as catecholamines may pass directly into the circulation, while larger molecules such as chromogranin A may be temporarily sequestered in the interstitial space before passing into the lymph, and hence into the circulation.
cells of the adrenal medulla release the contents of the chromaffin vesicle \da exocytosis. These contents include catecholamines, proteins (including chromogranin A), peptides (including neuropeptide Y), nucleotides and ions.” While the m~hanisms of stimulus-secretion coupling leading up to exocytosis have been well studied,29 the pathway taken by the various molecules following release from the cell is not known. The fact that catecholamines such as epinephrine exert their effects via the bloodstream has been appreciated since the experiments of Oliver and Schifer.25 Dopamine /?hydroxylase (DBH), one protein known to be released during exocytosis of the chromaffin vesicle, can also be detected in serum.36 Additionally, at least some of the D/IH released by components of the sympathetic nervous system reaches the vascular system via l~p~atics. ‘,I7 Pinardi et af. (Ref. 27, Fig. 6) demonstrated that, under conditions of hemorrhagic hypotension in the dog, at least half the adrenal output of D/?H (U/min) entered the circulation via the thoracic duct lymph.
The chromaffin
Abbreviations: CA, catecholamine; CgA, chromogranin; DA, dopamine; DBH, dopamine p-hydroxylase; EDTA, ethylenediaminetetra-acetate; NE, norepinephrine; NPY, neuropeptide Y; RIA, radioimmunoassay.
More recently, O’Connor et at.24 reported that the adrenal medulla is the principal source of increments in plasma chromogranin A in human subjects who were tested under conditions of insulin-induced hypoglycemia. In these studies the peak levels of chromogranin A in the plasma lagged behind the peak level of catecholamines by about 1 h, suggesting that this protein, like DPH, may enter the bloodstream via lymphatics. The present study was undertaken to examine the potential routes taken by secretory products from the adrenal medulla, specifically the adrenal vein (adrenolumbar vein in the cat) and thoracic duct. Based on the observations of Pinardi et aLz’ and O’Connor et a1.,24it was hypothesized that molecules of different size would take different routes; smaller molecules might enter the venous system directly, while larger molecules might accumulate in the interstitial fluid and enter lymphatics. Therefore, molecules of different sizes [catecholamines: epinephrine, mol. wt 183.2; norepinephrine (NE), mol. wt 169.2; dopamine (DA), mol. wt 153.2; neuropeptide Y (NPY), mol. wt 4254; and chromogranin A (CgA), mol. wt 48,000] were measured in the adrenolumbar venous plasma and thoracic duct lymph after insulin-induced hypoglycemia.
433
S. W.
434 EXPERIMENTAL Mongrel
CARMICHAEL
PROCEDURES
cats [rV = 9; mean body weight (kg) 2 S.E.M. = 3.83 + 0.251 were anaesthetized by hafothane inhalation. Resptratory effort was abolished with pancuronium bromide (Pavulon) and the animals were maintained by positive pressure ventilation with 1.5-2.0% halothane in oxygen. Body temperature was maintained at 37..5-385°C with heating pads. The left adrenolumbar vein was exposed through a lateral abdominal incision. The vessel was cannulated from a lateral approach (20 ga IV placement unit) using a method adapted from Hume and Nelson;” tension on a suture loop around the adrenolumbar vein between the adrenal gland and the inferior vena cava (or renal vein) caused blood flowing through the adrenal gland to be diverted into the cannula. A cannula (PE 90) placed in the femoral vein was used to withdraw peripheral blood samples and for the administration of fluid and drugs, while another cannula (PE 90) in the femoral artery permitted the continuous monitoring of heart rate and blood pressure. A thoracotomy was performed by sectioning three ribs to the left of the sternum. Careful dissection exposed the thoracic duct caudal to the junction of the left subclavian and internal jugular veins. The thoracic duct was cannulated (20 ga IV placement unit); lymph was collected continuously through this cannula. Blood volume was maintained with a Plasmalyte drip into the femoral vein. Additionally, red blood cells and lymph were returned after each sampling period. Lymph collected between samples was also returned. Following completion of the surgical procedures, the animals were heparinized (200 U/kg body wt) and allowed to stabilize, at which time baseline adrenolumbar venous. femoral venous and i~phati~ samples were collected. All samples were collected in chilled, untreated plastic tubes. Femoral venous blood samples were centrifuged and levels of glucose in the plasma determined by enzymatic assay (Dri-Stat Glucose-HK Endpoint Reagent; Beckman). Adrenolumbar venous and lymph samples were divided into small and large aliquots of, respectively, l/3 and 2/3 of the total volume collected. Chromogranin A was determined in the smaller samples by a homologous human CgA radioimmunoassay (RIA) which is a rapid, sensitive modification” of previously published procedures.‘9.2’ ~eliminary experiments demonstrated that feline CgA showed parallel crossreactivity (tracer displacement) in the human CgA RIA (data not shown). Albumin concentration was also measured in plasma and lymph samples using the Bromcresol Green calorimetric method (Sigma; Diagnostic Kit No. 631). The larger aliquots of adrenolumbar venous blood and lymph were mixed with EDTA (0.5 ml; 15%). NPY in the plasma of the adrenolumbar venous sample and in lymph was determined by radioimmunoassay.~ For the measurement of catecholamines, sodium metabisulfite (20 $/ml of a 5% solution) was added to the samples of adrenolumbar venous plasma and lymph. Catecholamines were measured by high performance liquid chromatography with electrochemical detection.5 All plasma and lymph samples were frozen at -80°C until assayed. After the collection of the baseline samples, adrenal medullary secretion was stimulated by the induction of hypoglycemia by intravenous injection of insulin (IU/kg). Samples of approximately 5 ml were collected at 0.25, 0.5, 0.75, 1, 1.5, 2, 2.5, 3, 3.5 and 4 h after injection of insulin. In five cats, both the rate of blood flow from the adrenolumbar vein and the rate of lymph flow from the thoracic duct were determined during the sampling periods. In three cats, both adrenal glands were isolated after collection of the 4 h sample by tightly ligating all affluent and effluent connections and cauterizing the pedicle; lymph samples were collected at 0.5 and 1.5 h after these procedures. The Wilcoxon Test for paired observations and a one-way analysis of variance were used, as indicated. to evaluate the
ei
ul
data. Differences between groups were considered significant if P < 0.05. RESULTS
The level of glucose in the systemic plasma fell to approximately 50 mg/dl within 30 min of insulin administration and remained below this level for the remainder of the experiment (Fig. 1). Catecholamines in adrenolumbar wnous plasma and ~~~~h
Catecholamines were more numerous on a molar basis than either NPY or CgA in both the adrenolumbar venous plasma and the lymph, as measured in pmol/ml (Figs 24). The concentration of all cateGLUCOSE AFlER
zoJ s 0
IN THE FIlMORAL
INSULIN
ADMINISTRATION
2
/ TIME
IN HOURS
VEIN
(1 U/KG)
3
4
AFTERINSULIN
Fig. 1. Level of glucose in plasma from femoral vein (mg/ deciliter). Values are mean -I-S.E.M. (N = 9).
TIME IN HOURS
AFTER INSULIN
Fig. 2. Levels of cat~holam~nes (NE, top; epinephrine, middle; DA, lower) in prnol/ml in adrenolumbar venous plasma (left Y-axes) and thoracic duct lymph (right Y-axes). Value are mean +S.E.M. for venous levels and mean -S.E.M. for lymphatic levels (N = 9).
435
Different size molecules secreted from adrenal medulla Table 1. Baseline levels of autocoids (pmol/ml) in adrenolumbar venous plasma and thoracic duct lymph
Norepinephrine (N = 9) ~pinephrine (N = 9) Dopamine (N = 9) Total catecholamines Neuropeptide Y (N = 9) Chromogranin A (N = 8)
Lymph 2.17+0.60 0.40 i 0.22 0.73 + 0.51 3.30 f 0.71 0.15+0.04 0.20 * 0.05
NEUROPEPTIDE
ADRENOLUMBAR
VEIN
Y
VS THORACIC
DUCT
Plasma 20.15f 14.35 26.52 & 12.27 0.51 + 0.14 47.18 + 20.76 0.21 f 0.04 0.23 f 0.04
Values are mean + S.E.M. eholamines in the adrenolumbar vein increased 30 mm after insulin administration and remained elevated for the duration of the experiment (Fig. 2). Insulin also preferentially increased the secretion of epinephrine from the adrenal medulla, as was indicated by an alteration of the ratio of NE:epinephrine at baseline from 1: 1.3 up to 1: 5.3. The changes in the NE : epinephrine ratio were significantly different from baseiine at 1.5 and 2.0 h. NE and epinephrine reached peak levels in the adrenolumbar vein at I .5-2.0 h after insulin injection; peak levels of NE and epinephrine in the thoracic duct lymph tended to occur later, at 2.5 h. Baseline levels of NE and epinephrine were notably greater in the adrenolumbar vein (nine- and 66-fold, respectively), while DA in the thoracic duct was l.Cfold higher than in the adrenolumbar vein (Table 1). However, following insulin-induced hypoglycemia, the levels of NE and epinephrine in the adrenolumbar vein exceeded the values in the lymph even further, rising up to 75-fold greater for NE and about 250fold greater for epinephrine (note the different right side Y-axis scales in Fig. 2, top and middle). DA in the thoracic duct fell to less than 1 pmol/ml (Fig. 2, bottom). The baseline samples of adrenolumbar venous plasma contained approximately equal proportions of NE and epinephrine (43 and 56%, respectively), while baseline lymph samples contained 66% NE and 12% epinephrine, in addition to 22% DA. Following insulin-induced hypoglycemia, the relative proportions of catecholamines changed in both plasma and lymph. Adrenolumbar venous plasma epinephrine increased to 84% of the total catecholamines between 45 and 90 min, while thoracic duct epinephrine increased to 60% of the total catecholamines at 90 min. Later in the experiment, between 3 and 4 h, epinephrine still comprised about 80% of the catecholamines in the adrenolumbar venous plasma, while NE in the lymph increased to between 60 and 70%, perhaps reflecting increased peripheral sympathetic activation.
TIME IN HOURS
Fig. 3. Levels of NPY (pmol/ml) in adrenolumbar venous plasma and thoracic duct lymph. Values are mean +S.E.M. as in Fig. 2 (N = 9). C~romogra~i~ A in adreno~~~ar
Y in ffdre~o~~rn~ar venous plasma and
&mph
The mean levels of NPY peaked between 1 and 1.5 h in the adrenolumbar venous plasma, while NPY in the thoracic duct gradually declined throughout the experiment (Fig. 3).
venous plasma and
lymph
Chromogranin A was found in approximately equal concentrations in both the adrenolumbar venous plasma and thoracic duct lymph (Fig. 4). Peak levels of CgA in the adrenolumbar vein were reached at 1.5-2 h following insulin injection; levels at these times were significantly greater than baseline (P < 0.002). In contrast, peak levels of CgA in the thoracic duct lagged behind the plasma, being reached at 2-3 h; levels at 2.5 and 3 h were significantly higher than baseline (P = 0.05). In the five cats in which the rate of lymph and adrenolumbar venous blood flow rates were calculated, CgA reached its peak secretion (pmol/min) at 1 h in the adrenolumbar vein, while a three-fold higher peak was reached in the thoracic duct at 2.5 h (Fig. 5). Neither adrenolumbar venous nor lymph flow, as measured by ANOVA, changed significantly during the course of the experiment. Expression of catecholamine and NPY data as pmolfmin did not reveal different patterns of release than those shown in Figs 2 and 3, with the data expressed as pmol/ml.
CHROMOGRANIN ADRENOLUMBAR
2
VEIN
A
VS THORACIC
ADRENOLUMBAR .-.THORACIC DUCT
o-o
8 P
~e~ropeptide
AFTER INSULIN
0.1 1
0
TIME
2
IN HOURS
3
AFTER
OUCl
VEIN
I
4
INSULIN
Fig. 4. Levels of chromogranin A (pmol/ml) in adrenolumbar venous plasma and thoracic duct lymph. Values are mean +S.E.M., as in Fig. 2 (N = 9). Levels at 1.5 and 2 h (vein) and 2.5 and 3 h (duct) are significantly greater than baseline.
436
S. W. CARMICHAEL et al. CHROMOGRANIN
A SECRETION
RAIE
00’ 0
1
TIME
2
IN HOURS
4
3
AFTER
INSULIN
Fig. 5. Secretion rate of chromogranin A (pmol/min) in adrenolumbar venous plasma and thoracic duct lymph. Values are mean +S.E.M. for lymphatic levels and mean
-S.E.M.
for venous levels (iv = 5).
In the three cats in which the adrenal glands were isolated following the 4 h sample, the mean (+S.E.M.) level of CgA in the thoracic duct at 4 h was 0.28 f 0.09 pmol/ml. The CgA content dropped at 30 and 90 min after adrenal isolation to, respectively, 0.10 + 0.06 and 0.01 + 0.05 pmoI/ml lymph. Albumin
in adrenolumbar
oenous plasma
and lymph
A non-parametric r-test compared the mean levels of albumin in plasma and lymph for all samples for each cat (N = 9). The mean albumin concentration in plasma of 1.51 & 0.11 mg/dl was significantly greater (P < 0.002, two-tailed) than the mean albumin concentration in all lymph samples for each cat (0.60 f 0.10 mg/dl). A one-way analysis of variance revealed that the concentration of albumin in both plasma and lymph declined significantly during the course of the experiment (P = 0.03 and P = 0.00, respectively). The “CgA quotient” in both plasma and lymph was determined by expressing pgrn CgA/mg albumin. The mean CgA quotient for lymph in all samples for each cat (3.79 & 0.53 pgrn CgA/mg albumin) was significantly greater (P < 0.002) than the CgA quotient for plasma in all samples for each cat (1.24 f 0.24; N = 8). In contrast to the changes of albumin concentration, the CgA quotient increased significantly during the course of the experiment in both plasma (P < 0.002) and lymph (P d 0.001). A Wilcoxon test determined the lymph CgA quotient to be significantly greater than the plasma CgA quotient at 2.5 and 3.0 h (4.52 + 0.75 vs 1.73 + 0.55 and 6.05 + 1.65 vs 1.45 k 0.62, respectively); these were times at which CgA reached its peak levels in thoracic duct lymph.
DISCUSSION
Insulin-induced hypoglycemia has been shown to be a strong and relatively specific stimulus for secre-
tion from the adrenal medulla, although stimulation of other components of the sympathetic nervous system may occur.34 In the present study hypoglycemia was induced following insulin administration; an increase in catecholamines in the adrenolumbar venous plasma was observed when plasma glucose levels fell to 50 mg/dl. The increase in the adrenal secretion of epinephrine as shown by significant changes in the molar ratio of NE:epinephrine at 1.5 and 2.0 h established that insulin provided a potent stimulus for adrenal medullary activation. Measurement of albumin was employed in this study as an index of the efficacy of thoracic duct cannulation and, hence, lymph collection. Human plasma contains about 4.5 g/d1 of albumin,” which is several fold greater than the concentration found in lymph. ” Thus , our observed ratio of albumin in plasma: lymph of approximately 2.5 establishes that we were indeed collecting fluid from the lymphatic system. The significant decrease in albumin concentration in both plasma and lymph throughout the course of the experiment was most likely caused by hemodilution secondary to the addition of intravenous fluids. Pathway
,fkr
secreted
adrenal
medullary
cate-
cholamines
In line with previously established theories, most of the catecholamines secreted from the adrenal medulla entered the circulation directly via the adrenolumbar vein (Fig. 2). Somewhat surprising, however, was the observation that the levels of catecholamines in the thoracic duct, while up to 200 orders of magnitude lower than those in the adrenolumbar vein, also peaked in response to insulin-induced hypoglycemia. The peak of catecholamines in the lymph was approximately 0.5 h later than that observed in the vein. These data suggest that most, but not all, of the catecholamine molecules released from the adrenal chromaffin cell exit the adrenal gland via the adrenolumbar vein. A minor fraction of adrenal medullary catecholamines accumulated in the interstitial space and, after a lag period, entered the lymphatic system. Although we do not propose that the catecholamines traveling through the thoracic duct have a specific physiological function, it is of interest that smooth muscle is found in the thoracic duct; the motility of this muscle is affected by adrenergic agonists and antagonists.” Pathways
.fkr secreted
adrenal
medullary
neuropep-
tide Y
NPY in the cat is believed to be co-stored in chromaffin vesicles with epinephrine.15 Most of the NPY secreted from the adrenal gland entered the circulation via the adrenolumbar vein (Fig. 3). However, at the time NPY peak in the vein (l-l.5 h), between 17 and 24% of the peak amount was detected in the thoracic duct. In contrast, generally less
Different size molecules secreted from adrenal medulla than 1% of catecholamines were found in the thoracic duct at peak times. It is of interest that the NPY peak in the adrenolumbar vein was not reflected in the lymph, perhaps due to rapid metabolism of this peptide. Pathways for mogranin A
secreted
adrenal
medullary
chro-
In the three cats in which the adrenal glands were isolated from the circulation after 4 h, levels of CgA dropped sharply, suggesting that the adrenal glands were the primary source of CgA detected in the thoracic duct in these experiments. Further analysis of our data is based upon this assumption. Peak levels of CgA in the thoracic duct lagged behind those in the adrenolumbar vein by about 1 h (Fig. 4). Interestingly, the peak levels in the vein and duct were approximately equal, suggesting that about the same amount of this protein leaves the adrenal gland by each route. It is important to note, however, that the absolute level of CgA in plasma or lymph is related to the flow rate of these substances. Thus, consideration of the CgA data on the basis of rate of flow of adrenolumbar blood or lymph (Fig. 5) indicated that greater amounts of CgA were leaving the adrenal medulla via the lymph than the adrenolumbar venous blood. Furthermore, the CgA quotient (pgm CgA/mg albumin) was significantly greater in the lymph than in the adrenolumbar venous plasma, when these values were compared in each cat. This observation suggests a preferential transport of CgA into the lymph, rather than merely an increased transport of proteins (such as albumin). The general increase in the CgA quotient throughout the experiment in the light of decreasing albumin values further argues for the favored movement of this protein into the lymph. The central vein of the adrenal gland drains both the medulla and cortex before emptying into the adrenolumbar vein. Thus, it is not possible for us to distinguish definitively between the cortex and medulla as the source of the substances measured in this study. However, catecholamines, NPY and CgA are well recognized contents of adrenal chromaffin vesicles.37 Further, immunohistological observations have established that CgA is present in the adrenal medulla, but not in the cortex.20 Thus we have assumed that the substances measured in the current study were secreted from the adrenal medulla. Comparison
of present
results with a previous study
It is of particular interest to compare our results with the observations of Pinardi et al*’ These investigators measured levels of catecholamines and dopamine /I-hydroxylase (DfiH) in the adrenolumbar vein and thoracic duct of dogs following hemorrhagic hypotension. In agreement with Pinardi et al., we found that baseline levels of catecholamines in adrenolumbar venous plasma were about 15-fold
431
greater than levels in thoracic duct lymph. However, the peak catecholamine levels reported by Pinardi and colleagues were only about four-fold greater in the adrenolumbar vein than in the thoracic duct, while we noted catecholamine levels 130-l 88-fold greater in adrenolumbar venous plasma at the time of peak response. The difference in these observations could be due to the use of different animals (dog vs cat), or the use of different stressors. Insulin-induced hypoglycemia is known to specifically activate the adrenal medulla,34 while hemorrhagic hypotension has more widespread effects on the sympathetic nervous system. It is of interest that Pinardi et al. observed the catecholamine peak at the same time in the adrenolumbar vein and thoracic duct, while in our animals the catecholamine peak in lymph lagged behind that in the adrenolumbar vein by 30-60 min. The protein measured by Pinardi et al. D/?H (mol. wt 64,86213), is notably larger than the protein, CgA (mol. wt 48,000), measured in the present study. At the time of peak levels, we found approximately equal amounts of CgA in the adrenolumbar venous plasma and lymph (Fig. 4), while Pinardi et al. (Ref. 27, Figs 3 and 5) found about twice as much D/?H in the lymph as in the adrenolumbar venous plasma. Of particular interest is the observation that reinfusion of blood into Pinardi’s animals resulted in a significant increase in thoracic duct DfiH. Pinardi and colleagues suggested that the reinfusionbound increase in D/?H was caused by a “wash out” of this protein from the interstitial space into the lymph. We suggest that the previous observations of Pinardi et al.*’ support the hypothesis that large molecules secreted from the adrenal medulla use the lymph, at least in part, as a pathway to the circulation. Comparison mogranin A
between
catecholamines
and
chro -
In the present study the peak levels of catecholamines and CgA occurred at the same time in the adrenolumbar vein. In contrast, insulin-induced hypoglycemia in human volunteers was accompanied by a 60-90 min delay of the plasma CgA peak after the catecholamine peak.” Chromaffin vesicles contain approximately 600 mM catecholamines (CA) and 2 mM CgA, giving a vesicular ratio of 300 : 1 (CA : CgA).4 The CA :CgA ratio in the adrenolumbar venous plasma baseline samples in the present study was 205 : 1, reasonably in agreement with previous reports. However, at the time of peak levels of catecholamines and CgA in adrenolumbar venous plasma, this ratio increased to 1369 : 1. In contrast, the baseline ratio of CA: CgA in lymph was 16.5:1, while the ratio at peak levels (2.5 h) was 14.5 : 1. Taken together these data suggest that there is a preferential transport of CgA into the lymph. It is also possible that, by the time the catecholamines reached the lymph, considerable reuptake or degradation had occurred, thus explaining
438
S. W. CARMICHAELet al.
the decrease in the CA : CgA ratio in lymph from that found in the chromaffin vesicle. The notably elevated levels of catecholamine compared to CgA in the adrenolumbar venous blood at the time of peak response might indicate a preferential transport of catecholamines into the bloodstream under conditions of hypoglycemic stress. Lymphatic drainage of the adrenal medulla The existence of lymphatic drainage from the adrenal medulla is controversial. Lymphatic capillaries have been demonstrated in the adrenal medulla of man3’ and cow,’ while other investigators have denied their presence in both manI and rat.33 Sulyok et a13’ demonstrated that, in the dog, reduction of adrenal venous outflow enhanced release of corticoids into the thoracic duct. These reports suggest either that the lymphatic drainage of the adrenal gland varies with species or that lymphatics are generally present but the methods to detect or demonstrate them are uncertain. Although we know of no description of lymphatic structures in the adrenal medulla of the cat, our results support the presence of lymphatic drainage in this animal. The absence of CgA in the thoracic duct lymph following the isolation of the adrenal glands combined with the identification of CgA in only the medullary portion of the adrenal glands clearly point to the adrenal medulla as the source of CgA measured in our cats. Size appears to influence the route used by a secreted ndrenal medullury molecule Studies employing the isolated retrogradely perfused bovine adrenal gland have established that catecholamines and CgA are co-released from the chromaffin cell.‘4.3H However, the methodology of these studies, i.e. cutting slits into the capsule of the gland to facilitate retrograde drainage, has prevented disclosure of the route taken by these secreted substances in uiro. The present results suggest that the size of a molecule released at the surface of the adrenal chromaffin cell influences the route that it takes to enter the vascular system. Smaller molecules, such as catecholamines, appear to take a direct route into the venous system draining the adrenal medulla. Thus, catecholamines may have an immediate effect on physiological processes since they enter the systemic circulation with minimal delay. Larger molecules, such as NPY and CgA, appear to be partially shielded from direct entry into the vein. The observation that more CgA than NPY was diverted from the vein into the lymph suggests that the larger molecule, the more it is shielded. The morphological barriers between the chromaffin cell and a vascular lumen are substantial, including endothelial cells, a continuous basement membrane, and a subendothelial space of variable width.6 Furthermore, the transcapillary movement of proteins is very slow. ‘* Intercellular junctions between
capillary endothelial cells may only open to 5-6 nm.26 Although fenestrated capillaries are found in the adrenal medulla, the diameter of the fenestrae, at least in the rat, is only about 50 nm.’ Thus, CgA, with an exceptionally high Stokes radius of about 8 nm,22 may not be able to pass readily into the vascular lumen. In fact, fenestrated capillaries may not be more permeable to macromolecules than nonfenestrated capillary endothelium; thus fenestrae may not be protein transport sites.28 In the present study, albumin (Stokes radius, 3.6 nm2*), a protein notably smaller than CgA, appeared to be transported differently across the capillary endothelium as suggested by the increased ratio of CgA to albumin, even under conditions of likely hemodilution at later time points in the experiment. Based on the observations presented in this study, we hypothesize that some of the larger molecules accumulate in the interstitial space, eventually move into lymphatic capillaries, and hence are returned to the circulation. The lag of the thoracic duct CgA peak behind the adrenolumbar venous CgA peak (Figs 4 and 5) is consistent with this hypothesis.
CONCLUSION
While the rapid translocation of catecholamines into the vascular system may result in an obvious physiological advantage to the organism, we may only speculate on the advantage of slower access of NPY and CgA to the circulation. Such speculation is based on the lack of clearly established physiological roles for NPY and CgA. Recently, NPY was shown to be secreted from the adrenal medulla in response to hemorrhagic hypotension.’ Additionally, NPY may potentiate the action of NE at adrenergic receptors.* Chromogranin A-derived peptides have been shown to exert a feedback control on chromaffin cell secretory activity.3’ Theoretically, such control of adrenal medullary secretion could be exerted over an extended time period (several minutes to a few hours) and then diminish as CgA is gradually eliminated into the venous and lymphatic systems. Perhaps the function of lymphatic-carried NPY and CgA will be clarified as the physiological functions of these molecules and their degradation products are revealed. Acknowledgements-The Kettelkamp for preparation
authors thank MS Julia of the graphs, and the following personnel for their expert performance of the various assays used in this study: MS Sharon Chinnow (catecholamine assays), MS Laurie Bale (glucose assays), MS Annie Chen (chromogranin A and albumin assays), and MS Diane Roddy (NPY assay). We also thank Drs William Engeland, Jose-Maria Trifaro and Richard Weinshilboum for helpful comments on this manuscript. This work was supported in part by grants to Dr Tyce- (NS.17858) and Dr -&Ionnor (Veterans Administration, NIH [DK-36,400; HL-35.0181. and the American Heart Association).
Different
size molecules
secreted
from adrenal
medulla
439
REFERENCES I. Aerg H. E., Hansson H. E., Wetterberg L., Ross S. B. and Friien 0. (1974) Dopamine-b-hydroxylase in human lymph. Life Sci. 14, 65-71. 2. Allen J. M., Yeats J. C., Adrian T. E. and Bloom S. R. (1984) Radioimmunoassay of neuropeptide Y. Reg. Pep?. 8, 61-70. 3. Alvarez-Morujo A. J. and Alvarez-Morujo A. (1983) Scanning electron microscopy of lymphatic system in the adrenal glands of the bull in three different situations. Morfol. Norm. Patol. Histol. 7A, 361-370. 4. Apps D. K. (1982) The reuptake of amines by secretory granules. Trends biochem. Sci. 7, 153-156. 5. Bove A. A., Dewey J. D. and Tyce G. M. (1984) Increased conjugated dopamine in plasma after exercise therapy. J. Lab. clin. Med. 104, 77-85. 6. Carmichael S. W., Brooks J. C., Malhotra R. K., Wakade T. D. and Wakade A. R. (1989) Ultrastructural demonstration of exocytosis in the intact rat adrenal medulla. J. Electron Microsc. Tech. 12, 316-322. 7. Elfvin L.-G. (1965) The ultrastructure of the capillary fenestrae in the adrenal medulla of the rat. J. ulfrastrucf. Res. 12, 687-704. 8. Evtquoz D., Waeber B., Corder R., Nussbergers J., Gaillard R. and Brunner H. R. (1987) Markedly reduced blood pressure responsiveness in endotoxemic rats: reversal by neuropeptide Y. Li& Sci. 41, 2573-2580. 9. Gaumann D., Yaksh T., Roddy D., Michner S., Nelson D. and Lucas D. (1986) Adrenal release of [Metlenkephalin and other neuropeptides during hypotensive hemorrhage in halothane anesthetized cat. Sot. Neurosci. Abstr. 12, 1484. 10. Guyton A. C. (1986) Textbook of Medical Physiology, 7th edn. W. B. Saunders, Philadelphia. 11. Hargens A. R., Tucker B. J. and Blantz R. C. (1977) Renal lymph protein in the rat. Am. J. Physiol. 233, F269-273. 12. Hume D. M. and Nelson D. H. (1955) Adrenal cortical function in surgical shock. Surg. Forum 5, 568-575. 13. Lamouroux A. A., Vigny A., Faucon Biguet N., Darmon M. C., Franck R., Henry J.-P. and Mallet J. (1987) The primary structure of human dopamine /I-hydroxylase: insights into the relationship between the soluble and membrane-bound forms of the enzyme. Eur. molec. Biol. Org. J. 6, 3931-3937. 14. Ledbetter F. H., Kilpatrick D., Sage H. L. and Kirshner N. (1978) Synthesis of chromogranins and dopamine fi-hydroxylase by perfused bovine adrenal glands. Am. J. Physiol. 235, FA75486. 15. Lundberg J. M., Hofelt T., Hemsen A., Theodorsson-Norheim E., Pernow J., Hamberger B. and Goldstein M. (1986) Neuropeptide Y-like immunoreactivity in adrenaline cells of adrenal medulla and in tumors and plasma of pheochromocytoma patients. Regl. Pep?. 13, 169-182. 16. Merklin R. J. (1966) Suprarenal gland lymphatic drainage. Am. J. Anat. 119, 359-374. 17. Ngai S. H., Dairman W., Marchelle M. and Spector S. (1974) Dopamine fi-hydroxylase in dog lymph-effect of sympathetic activation. Life Sci. 14, 2431-2439. 18. Noer I. and Lassen N. A. (1979) Evidence of active transport (filtration?) of plasma proteins across the capillary walls in muscle and subcutis. Acta physiol. stand. Suppl. 463, 105-l 10. 19. O’Connor D. T. and Bernstein K. N. (1984) Radioimmunoassay of chromogranin A as a measure of exocytotic sympathoadrenal activity in normal subjects and patients with pheochromocytoma. New Engl. J. Med. 311, 764779. 20. O’Connor D. T., Burton D. W. and Deftos L. J. (1983) Chromogranin A: immunohistology demonstrates its universal occurrence in normal polypeptide hormone producing endocrine glands. Z,ife Sci. 33, 1657-1663. 21. O’Connor D. T. and Deftos L. J. (1986) Secretion of chromogranin A by peptide producing endocrine neoplasms. New Engl. J. Med. 314, 1145-1151. 22. O’Connor D. T. and Frigon R. P. (1984) Chromogranin A, the major catecholamine storage vesicle soluble protein. J. biol. Chem. 259, 3237-3247. 23. O’Connor D. T., Pandian M. R., Carlton E., Cervenka J. H. and Hsiao R. .I. (1989) Rapid measurement of circulating human chromogranin A: in vitro stability, exploration of the neuroendocrine character of neoplasm, and assessment of the effects of organ failure. Clin. Chem. 35, 1631-1637. 24. O’Connor D. T., Pandian M. R., Cervenka J., Mezger M. S. and Parmer R. J. (1987) What is the source and disposition of chromogranin A (CgA) in normal human plasma? (Abstract.) C/in. Res. 35, 60SA. 25. Oliver G. and Schler E. A. (1894) On the physiological action of extract of the suprarenal capsules. J. Physiol., Lond. 16, i-iv. 26. Palade G. E., Simionescu M. and Simionescu N. (1979) Structural aspects of the permeability of the microvascular endothelium. Acta physiol. stand. Suppl. 463, 1 l-32. 27. Pinardi G., Talmaciu R. K., Santiago E. and Cubeddu L. X. (1979) Contribution of adrenal medulla, spleen and lymph, to the plasma levels of dopamine /l-hydroxylase and catecholamines induced by hemorrhagic hypotension in dogs. J. Pharmac. exp. Ther. 209, 176-184. 28. Renkin E. M. (1977) Multiple pathways of capillary permeability. Circulation Res. 41, 736-743. 29. Rosenheck K. and Lelkes P. I. (eds) (1987) Stimulus-Secretion Coupling in Chromafin Cells, Vols I and II. CRC Press, Boca Raton. 30. Sapin M. R. (1959) On adrenal intraorgan lymphatic system in man. Arkhiv. Anat. Histol. i Embryol. 36, 52-59. 31. Simon J.-P., Bader M.-F. and Aunis D. (1988) Secretion from chromaffin cells is controlled by chromogranin A-derived peptides. Proc. natn. Acad. Sci. U.S.A. 85, 1712-1716. 32. Sulyok A., Monos E., Koltay E., Bede A. and Kovach A. G. B. (1969) Effect of local venous congestion on adrenocortical function. Acta physiol. Acad. Sci. hung. 35, 313-320. 33. Verhofstad A. A. J. and Lensen W. F. J. (1973) On the occurrence of lymphatic vessels in the adrenal gland of the white rat. Acra anat. 84, 475483. 34. Viveros 0. H., Arqueros L., Connett R. J. and Kirshner N. (1969) Mechanism of secretion from the adrenal medulla. IV. The fate of the storage vesicles following insulin and reserpine administration. Molec. Pharmac. 5, 69-82. 35. Wang G. and Zhong S. (1985) Experimental study of lymphatic contractility and its clinical importance. Ann. Plast. Surg. 15, 278-284.
440
S. W. CARMICHAEL et al.
36. Weinshilboum R. and Axelrod J. (1971) Serum dopamine-beta-hydroxylase activity. Circulation Res. 2S, 3077315. 37. Winkler H., Apps D. K. and Fischer-Colbrie R. (1986) The molecular function of adrenal chromaffin granules: established facts and unresolved topics. Neuroscience 18, 261-290. 38. Winkler H., H&nag1 H., Schiif J. A. L., Hiitnagl H. and Zur Nedden G. (1971) Bovine adrenal medulla: synthesis and secretion of radioactively labelled catecholamines and chromogranins. Naunyn-Schmiedeberg’s Arch. Pharmac. 271, 193-203. (Accepted 17 July 1989)
