INTRODUCTION

Top Abstract Introduction MATERIAL AND METHODS Results Discussion References

RECIRCULATION, HOMING, and mobilization of lymphoid cells are essential processes for an efficient immune response in vivo. Microenvironmental factors such as adhesion molecules, selectins, integrins, and chemotactic cytokines contribute to the control of cell homing and mobilization, especially regarding the selection of the type of cells that home and leave lymphoid organs (24, 28, 30, 31). However, the primary force that drives immune cell displacement is the cardiovascular and lymphatic circulation. Blood flow is mainly controlled by noradrenergic nerves, which are abundant in lymphoid organs (for review, see Ref. 11). In this way, sympathetic neurotransmitters contribute to control the traffic and mobilization of cells as well as antigen trapping in these organs. On the other hand, there is anatomical and functional evidence suggesting that norepinephrine (NE), the main sympathetic neurotransmitter, may play a role in the control of cell mobilization from lymphoid organs independent of its effect on peripheral vascular resistance. In the spleen, for example, noradrenergic fibers are not confined to the innervation of vascular structures but also penetrate the white pulp, where they establish close contact with lymphoid cells, especially T lymphocytes (11, 21). These cells posses adrenergic receptors (5, 15, 18, 29), and there is abundant literature showing effects of sympathetic neurotransmitters on immune cell activity (Refs. 4, 10, and 16; for review, see Refs. 3 and 22). Furthermore, lymphoid cell mobilization from the spleen to the blood is known to occur during acute stimulation of the sympathetic nervous system (2, 12, 23). On the basis of this, we hypothesized that the sympathetic nervous system may exert a dual control of lymphoid cell mobilization from the spleen by smooth muscle-dependent and -independent mechanisms. The smooth muscle-dependent mechanism would be exerted via regulation of the blood flow in the spleen (this organ does not have external lymphatic circulation) due to the vasoconstrictor effects of sympathetic neurotransmitters (26). Smooth muscle-independent effects would be caused by direct binding of neurotransmitters to adrenergic receptors either in lymphoid cells (5, 15, 18) or in other contractile structures capable of affecting spleen cell mobilization (17, 20, 25). In animals with an intact splenic circulation, these two mechanisms are difficult to dissociate, because both are expected to operate simultaneously and may affect splenic cell mobilization to a different degree and even in an opposite way. To approach this issue, we have used a model of in vivo spleen perfusion in which the innervation of the organ is kept intact and both flow resistance and cell output are monitored simultaneously. Using this model, we show here that NE perfusion and the intrinsic noradrenergic innervation of the spleen can promote lymphoid cell output from this organ in a smooth muscle-independent manner.

	MATERIAL AND METHODS

Top Abstract Introduction MATERIAL AND METHODS Results Discussion References

Animals. Wistar albino male rats (350-400 g body wt) were purchased from Harlan Winkelmann (Borchen, Germany) and individually caged for 1 wk before the start of the experiments. Animals were housed at 22 ± 1°C and 55-60% air humidity on a 12:12-h light-dark cycle and had free access to water and standard food pellets. Experiments were performed according to the written consent of the local ethics commission.

Spleen perfusion. Animals were anesthetized with pentobarbital sodium (60 mg/kg body wt ip Nembutal; Sanofi/Deutsche Tierärzte, Hannover, Germany). In principle, the spleen was prepared for perfusion as originally reported by Groom and Song (14) but modified as described below to minimize splenic ischemia during vessel cannulations. A cannula (1 gauge) connected to a polyethylene catheter (PE-90; Clay Adams, Parsippany, NJ) was inserted into the vena cava at abdominal level. The opposite end of this catheter was expanded and inserted into the splenic vein, as near as possible to the portal vein, and splenocaval flow was opened immediately. Next, the splenic artery was isolated, over a distance of ~0.5 mm, as near to its origin as possible. Great care was taken to minimize tissue damage by dissecting the connecting tissue only at the site of blood vessel cannulation by blunted peeling of the vessel wall. All arterial branches leading to the stomach and pancreas were electrically coagulated (bipolator; Fischer Medical Technology, Freiburg, Germany), and other arterial supply to the spleen from the greater gastric curvature was ligated. When the spleen was clearly supplied only by the splenic artery, this vessel was cannulated with a short, expanded PE-60 catheter (Clay Adams). The resistance of the arterial cannula was 7-12% of the total vascular resistance of a normally perfused spleen. About 2 cm apart from the splenic artery, a T piece connected to a pressure transducer (Statham PD23d) was inserted into the cannulating tube. The hollow space on top of the strain gauges of the pressure transducer provided the bubble trap. About 1 cm apart from the splenic vein, a T piece was inserted into the venous catheter, and an attached tube led to a drop counter that continuously monitored the volume of the outflow. Immediately after the start of the perfusion, the flow to the vena cava was clamped and venous outflow was directed toward the drop counter. The spleen was perfused using a syringe pump with a commercially available saline solution containing 6% hydroxyethyl starch to stabilize the colloidosmotic pressure (Plasmasteril; Fresenius, Bad Homburg, Germany). The fluid was bubbled with oxygen for 5 min and put into a 50-ml glass syringe containing ~5 ml oxygen. The fluid was delivered at a constant flow rate between 0.7 and 0.9 ml · min¹ · g spleen wt¹ and warmed to 37°C by a heat exchanger before flowing near the pressure transducer into the splenic artery. The temperature of the arterial inflow was measured by a thermoelement glued to the T piece before the arterial cannula. Arterial inflow pressure was continuously recorded, and venous outflow pressure was kept constant at 5 mmH₂O. Outflow volumes were registered with a drop counter, and the first 3 ml of the outflow were collected in consecutive fractions of 0.5 ml, the following 7 ml in fractions of 1 ml, and thereafter in 2-ml fractions until a total volume of 30 ml was perfused. The number of erythrocytes and of mononucleated cells was measured with a Coulter counter (model FN; Coulter, Dunstable, England) or determined in a Neubauer chamber when numbers were <1,000 cells/µl. The percentage of polymorphonuclear granulocytes and mononuclear cells was evaluated by stained blood smears (Pappenheim staining). For this purpose, aliquots of the collected splenic venous outflow were concentrated by centrifugation at 1,000 g for 5 min. To check the homogeneity of flow distribution, the spleen weight was determined at the end of each experiment and the organ was stained with Evans blue injected through the splenic artery. Only those experiments in which staining was homogeneous and complete and no organs other than the spleen were tinged were evaluated. This included a total of 55 perfusions.

Drugs. The following drugs were used: NE (Hoechst Pharma, Bad Soden, Germany), phenylephrine (Sigma-Aldrich, Deisenhofen, Germany), isoproterenol (Research Biochemicals International, Natick, MA), phentolamine (Research Biochemicals International), propranolol (ICI/Zeneca, Plankstadt, Germany), papaverine (Paveron; Karlspharma, Kirchheim, Germany), reserpine (Bayer, Leverkusen, Germany), and endotoxin [lipopolysaccharide (LPS) from Escherichia coli 026:B6, TCA extract; Sigma, St. Louis, MO].

Experimental setup. After preparing the spleen as described above, we performed perfusions under basal conditions and at increased or decreased levels of flow resistance. Increased flow resistance was achieved by adding NE to the perfusion medium at a final concentration of 1 or 10 µM. Papaverine, reserpine, propranolol, phentolamine, and LPS were used to reduce the flow resistance. Papaverine acts as a smooth muscle relaxant by inhibiting predominantly smooth muscle phosphodiesterase. Propranolol, a -adrenergic blocker, impedes NE effects on these receptors. Papaverine (10 µg/ml) and propranolol (1 µM) were also added to the perfusion fluid. When NE was combined with papaverine or propranolol, both substances were perfused together in the same fluid. Reserpine, which induces vasodilatation by depletion of NE stores, was injected intraperitoneally at a concentration of 10 mg/kg body wt 24 h before the start of the experiments. The possibility that, under pathophysiological conditions, vascular and nonvascular mechanisms could act synergistically was investigated by treatment of the animals with endotoxin. LPS was injected intraperitoneally at a concentration of 10 µg/kg body wt 1 h before the start of the perfusion, because we have previously reported that vasodilatation and the increase in splenic blood flow reach a maximum under these conditions (27). LPS was also administered at a concentration of 10 µg/l in the perfusion fluid. When LPS was combined with NE, the same procedure was followed, with inclusion of the catecholamine in the perfusion medium.

Evaluation of the results. To evaluate the cellular washout, we calculated the cell concentration in the outflow according to the volume perfused and not the flow rate, as recommended by Groom (13) and Cilento et al. (8). Flow resistance was calculated from the arteriovenous pressure difference and perfusate flow.

Statistics. Results are given as means ± SD. The StatView II computer program from Abacus Concepts was used to calculate the best-fitting curve (Fig. 2). Statistical significance of differences was evaluated by ANOVA followed by Scheffe's F test for multiple comparisons (Fig. 3), or by two-tailed paired t-test (Fig. 5).

	RESULTS

Top Abstract Introduction MATERIAL AND METHODS Results Discussion References

Perfusion allows distinction between blood- and spleen-derived mononucleated cells. The number of erythrocytes and nucleated cells released from the spleen at normal flow resistance (flow 0.88 ± 0.09 ml · g¹ · min¹, pressure 15.07 ± 1.25 kPa) quickly decreased during the perfusion compared with the initial values (Fig. 1). In the first two samples of the perfusate collected, the ratio between erythrocytes and nucleated cells reflected the normal arterial intravascular cell composition, which is ~1,000 erythrocytes per leukocyte, and the ratio between polymorphonuclear and mononuclear cells reflected the composition of circulating blood cells in normal Wistar rats (Table 1). These data show that the cells obtained in outflow samples collected soon after the start of the perfusion derived from the intravascularly suspended and circulating cell pool. After this first and quickly removed cell pool, the erythrocyte/leukocyte ratio in the perfusate changed toward the cell composition of whole spleen suspensions. Furthermore, in contrast to the blood, only a few polymorphonuclear cells were found and most nucleated cells were lymphocytes. This pattern indicates that the cells obtained in the last 27 ml of the perfusate represent a pool of slowly recruitable splenic cells that are most likely derived either from extravascular sites or are adhering to the walls of the vessels. The studies that follow refer to noradrenergic effects on this slowly recruitable cell compartment.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Rat spleens were perfused in vivo under basal conditions as described in MATERIAL AND METHODS. A: concentration of erythrocytes and nucleated cells in perfusate (outflow). Outflow 0 corresponds to determinations in blood of same animals. Each point in curves represents mean ± SD of cells counted in 6 animals. B: ratio between erythrocytes and nucleated cells in outflow.

Inverse correlation between number of nucleated cells washed out from spleen and splenic flow resistance. Maximal changes in spleen flow resistance were obtained by reserpine injection, which causes vasodilatation, and by NE perfusion, which induces vasoconstriction. A moderate decrease in resistance was obtained using the smooth muscle relaxant papaverine at a dose of 10 µg/ml (Fig. 2). These variations in splenic flow resistance affected the number of slowly recruitable nucleated cells obtained in the splenic outflow; i.e., 2 to 3 times more cells were collected during vasodilatation than at normal flow resistance, whereas the cellular washout decreased nearly to zero at high vascular resistance. The relation between the number of slowly mobilized cells from the perfused spleen and the flow resistance fits best (r² = 0.962) to the polynomial second-order equation where c is the number of collected leukocytes in the outflow, and f is flow resistance. This curve was used as a reference for the studies that follow.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Rats were prepared for spleen perfusion as described. In a group of rats, spleen was perfused under basal conditions (Basal; n = 6). Spleen of other rats was perfused either with norepinephrine (NE; 1 µM, n = 6) or papaverine (10 µg/ml, n = 5). Another group of rats was treated with reserpine 24 h before spleen perfusion (n = 6). Total number of cells obtained in outflow and flow resistance were measured simultaneously in each animal. Each symbol corresponds to 1 animal. Data obtained were used to construct a reference curve that fits best to a second-order polynomial regression.

NE induces splenic cell mobilization despite its effects on vascular resistance. The effect of NE administration on cell mobilization from the spleen at different degrees of flow resistance was investigated. The results of the simultaneous evaluation of the number of nucleated cells in the splenic perfusate and the flow resistance in the spleen are shown in Fig. 3. The data of the experiments shown in Fig. 2 on the effect of reserpine, papaverine, and NE are also included for comparison. Propranolol perfusion clearly reduced vascular resistance. However, this treatment did not result in increased cell mobilization from the spleen. When propranolol was perfused simultaneously with NE, the blocker interfered with the increase in vascular resistance induced by the catecholamine. Under these conditions, the number of nucleated cells in the outflow was markedly reduced compared with the control. In contrast, despite the vasoconstriction it induced, perfusion of the -agonist isoproterenol caused a clear increase in splenic cell mobilization relative to the other vasoconstrictors, NE and phenylephrine. Perfusion of the -agonist phenylephrine caused an extreme increase in flow resistance and a decrease in splenic cell outflow. Conversely, the -adrenergic antagonist phentolamine induced an increase in cellular outflow that corresponded to the values expected at the reduced flow resistance observed. In contrast, when NE was administered in combination with phentolamine or papaverine, agents that induce vasodilatation by different mechanisms, NE caused an increase in cell mobilization greater than that expected at that flow resistance. The splenic flow resistance of NE-perfused LPS-treated animals was increased compared with that of rats treated with LPS alone. However, the number of cells mobilized by splenic NE perfusion of LPS-treated rats was increased compared with the values expected at the level of flow resistance attained. This increase was abrogated to a large extent by propranolol. These results are represented in Fig. 4, together with the reference curve. It can be appreciated that, for a given flow resistance, the number of nucleated cells obtained in the spleen perfusate under the experimental conditions described above clearly differs from the values predicted by the curve. This can be better visualized in Fig. 5, in which the expected number of nucleated cells is plotted versus the cell number actually determined.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Rats were prepared for spleen perfusion and subject to different treatments as described in MATERIAL AND METHODS. A: number of nucleated cells collected in spleen outflow. With the exception of groups propranolol-treated, isoproterenol-treated, and lipopolysaccharide (LPS) + NE + propranolol-treated animals, all other groups were statistically different from control (P < 0.05). B: corresponding values of flow resistance. With the exception of groups NE + propranolol-treated, NE + LPS-treated, and NE + LPS + propranolol-treated rats, all other groups were statistically different from control (P < 0.05). Results are expressed as means ± SD. Control, n = 6; reserpine, n = 6; papaverine, n = 5; propranolol, n = 6; phentolamine, n = 2; LPS, n = 6; isoproterenol, n = 4; phenylephrine, n = 2 ; NE (1 µM), n = 6; NE (1 µM) + propranolol, n = 4; NE (1 µM) + phentolamine, n = 3; NE (1 µM) + papaverine, n = 4 ; NE (10 µM) + papaverine, n = 6; NE + LPS, n = 6; NE + LPS + propranolol, n = 3. Statistical significance of differences was evaluated by ANOVA followed by Scheffe's F test for multiple comparisons.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Total number of cells collected in spleen perfusate after treatments indicated in figure and corresponding flow resistance for each individual animal are depicted together with reference curve shown in Fig. 2.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Total number of cells collected in spleen perfusate after different treatments is shown together with cell number predicted for a given level of flow resistance. Each point indicates mean ± SD from number of determinations indicated in Fig. 3. Statistical significance was evaluated by 2-tailed paired t-test. Value obtained experimentally differs significantly from corresponding predicted value for the following treatments: isoproterenol, propranolol, NE + phentolamine, NE + propranolol, NE + papaverine, and NE + LPS.

	DISCUSSION

Top Abstract Introduction MATERIAL AND METHODS Results Discussion References

The in vivo perfusion system used in the studies reported here allows analysis of the effect of endogenous and exogenous substances on the mobilization of resident splenic cells. We first analyzed the cellular composition of the spleen perfusate under conditions of vascular resistance that are comparable to normal physiological levels. After the washout of the cells present in the vascular compartment of the spleen, the relation between mononucleated cells and erythrocytes found in the outflow was similar to that found in spleen cell suspensions. More than 90% of the recovered nucleated cells were lymphocytes, and the proportion of neutrophils in the outflow was decreased compared with the blood composition. This led us to the conclusion that most of the slowly recruitable nucleated cells found in the spleen perfusate derived from the mobilization of parenchymal cells and from cells adherent to the blood vessels of this organ. The discussion that follows refers to this "slow" removable cell compartment. Using this model, we studied possible effects of the sympathetic neurotransmitter NE on cell mobilization by simultaneously measuring, in each individual animal, both flow resistance and splenic cell outflow. This was essential to distinguish the effects of NE on splenic cell mobilization due to its vasoconstrictor actions from other effects that could be independent of its effects on vascular resistance.

Depletion of NE stores or vascular smooth muscle relaxation lead to both a significant increase in the number of cells in the spleen perfusate and a decrease in flow resistance. Conversely, NE administration resulted in a clear decrease in cell outflow paralleled by an increased flow resistance. This illustrates the relevance of the vasomotor tonus for splenic cell mobilization and the contribution of the sympathetic innervation to the control of this process. These results allowed us to construct a reference curve indicating the number of cells that would be mobilized from the spleen under the assumption that flow resistance is the only variable that determines cell outflow. However, the data obtained using NE in combination with other substances suggested that the neurotransmitter can affect splenic cell mobilization independently from its vasoconstrictor actions. As shown in Fig. 3, propranolol induced a significant decrease in flow resistance that was not paralleled by an increase in cell number outflow. When propranolol was perfused in association with NE and flow resistance was normalized, the decrease in cell number in the perfusate was comparable to that obtained by perfusion of NE alone. In contrast, when NE was administered together with papaverine or LPS, agents that decrease flow resistance by a mechanism that does not involve -adrenergic receptors, an increase in spleen cell outflow was detected. Additional evidence for the involvement of -receptors in splenic cell mobilization derived from the results obtained after perfusion of the selective -agonist isoproterenol. This agonist caused a clear increase in splenic cell mobilization despite the moderate vasoconstriction it induced.

The results depicted in Fig. 4 permit better appreciation of these findings by showing the individual values obtained after each treatment together with the reference curve. As shown in Fig. 4, the splenic cell outflow observed in animals treated with propranolol either alone or in association with NE was clearly below that expected at the reduced level of resistance measured. However, NE given in association with the smooth muscle relaxant papaverine caused a significantly higher cell mobilization than expected from the corresponding level of resistance in the spleen. LPS given alone caused a decrease in flow resistance and a proportional increase in cell mobilization. As shown in Fig. 4, animals that received NE and LPS together showed a >200% increase in splenic cell mobilization compared with the values expected at this high flow resistance. It is unlikely that LPS treatment would interfere with adrenergic receptor signaling or expression. In fact, in the present as well as in previous studies (27), we detected an increase in flow resistance induced by NE in LPS-treated rats. Furthermore, as shown here, propranolol blocked to a large extent the increase in splenic cell mobilization that NE caused in LPS-treated rats. Figure 5 shows the deviation of the cell number measured after each treatment from the cell number predicted for a given flow resistance.

How NE can affect splenic cell migration independently of its effect on vascular smooth muscle is at present unknown. -Adrenergic receptors are expressed in T and B lymphocytes, macrophages, and neutrophils (for review, see Ref. 22). These receptors are upregulated during immune cell activation (5, 9), and sympathetic mediators can, depending on the type of immune response, selectively exert inhibitory or stimulatory effects (for review, see Ref. 22). The stimulatory effect of NE on cell mobilization from the spleen could therefore be ascribed to direct effects of the sympathetic mediator on these cells. As mentioned, propranolol administration interfered with cell migration from the spleen during conditions of reduced flow resistance. This strongly suggests that endogenous NE derived from splenic nerve fibers controls lymphoid cell mobilization from this organ through a -adrenergic receptor-dependent mechanism. Furthermore, propranolol also impeded the stimulatory effect of exogenously administered NE on splenic cell migration that should be observed at reduced levels of flow resistance. On the other hand, when the effect of endogenous NE on -adrenergic receptors was blocked by phentolamine, the number of cells mobilized from the spleen corresponded to that expected at reduced levels of flow resistance. In addition, phentolamine perfused together with NE did not significantly affect the cell-mobilizing effect of the catecholamine.

No direct effects of NE on immune cell mobility have been so far reported, although lymphocytes are capable of locomotion under the influence of agents such as colchicine, vinblastin, and anti-IgG (7). Other mechanisms that could underlie the described effect of NE on splenic cell mobilization may concern effects of this neurotransmitter on non-smooth muscle-dependent contractile structures (20, 25) and/or effects on cell adhesion to the endothelium (1, 6). For example, it has been reported that NE causes emptying of about two thirds of erythrocytes from the cat spleen after removal of blood cells from the circulation. This effect was ascribed to the mobilization of erythrocytes from intrasplenic reservoirs and seems to be mediated by effects of NE on non-smooth muscle contractile structures (14). A comparable mechanism could mediate the effects of NE that we have described here on the mobilization of splenic lymphoid cells. An alternative or complementary possibility could be that NE affects the adhesion of lymphocytes to capillary endothelium. This possibility is supported by recent studies showing that catecholamines decrease the adherence of T lymphocytes to endothelial cells without changing the expression of adhesion molecules (1, 6).

The present studies show a dual, opposite effect of NE on cell mobilization from the spleen. On one hand, NE reduces blood flow by increasing flow resistance and thus favors the retention of lymphoid cells in the spleen. On the other, it stimulates lymphoid cell output even at relatively high levels of flow resistance. An effect of NE on splenic cell outflow would not be detectable at levels of vascular resistance that induce an extreme decrease in blood flow. However, local factors, such as nitric oxide and prostaglandins, could oppose NE-induced vasoconstriction, and the stimulatory effect of this neurotransmitter on cell mobilization could therefore be manifested.

The smooth muscle-independent stimulatory effect of NE on splenic cell mobilization that we have described here may be relevant during acute stimulation of the sympathetic nerve system, for example, during increased physical activity. Lymphocytosis is known to occur during physical exercise and after catecholamine administration. Most lymphoid cells displaced to the circulation under these conditions appear to derive from the spleen, because such an effect is not observed in splenectomized individuals (23). As discussed above, the increase in vascular resistance induced by NE is expected to induce an opposite effect, because it should impede lymphoid cell output from the spleen. Thus the stimulatory -adrenergic-mediated effect of NE on cell mobilization described here would explain how during processes such as exercise or acute stress that are paralleled by increased vascular resistance (19), lymphoid cells are mobilized from the spleen.

Perspectives