Permeability of the blood-brain barrier to soluble cytokine receptors

Permeability of the mouse blood-brain barrier to murine interleukin-2: predominance of a saturable efflux system

Interleukin (IL)-2, a T helper (TH)1 cell-derived glycoprotein with potent neuromodulatory effects, is implicated in the etiology and pathogenesis of various psychiatric and neurological disorders. Paralleling these findings, chronic IL-2 intravenous immunotherapy may induce similar psychopathological outcomes. The findings that acute or repeated injections of IL-2 induce motor and cognitive abnormalities in rodents are consistent with these clinical findings, and raise the possibility that IL-2 crosses the blood-brain barrier (BBB) to alter brain function. However, little is known about the ability of IL-2 to enter the brain or whether its effects vary with the chronicity of IL-2 treatment. Here, we found that radioactively labeled mouse IL-2 (I-IL-2) given intravenously entered the brain at a low rate (Ki=0.142+/-0.044microl/g-min) by a non-saturable process. Repeated injections of either IL-2 or vehicle altered the kinetics of entry without producing a net effect on IL-2 entry. When I-IL-2 was given by brain perfusion, the entry rate greatly increased over 10-fold to 2.2+/-0.805microl/g-min. This suggests a circulating factor is retarding the entry of IL-2 into the brain. A paradoxic increase in the rate of I-IL-2 entry into brain occurred when an excess of unlabeled IL-2 was included in the brain perfusate, suggesting a saturable CNS-to-blood efflux system. Intracerebroventricular injection of I-IL-2 with and without unlabeled IL-2 confirmed the presence of a saturable efflux system. We conclude that IL-2 entry into the brain is low because of the absence of a blood-to-brain transporter and further retarded by circulating factors and a CNS-to-blood efflux system. This is the first description of a saturable CNS-to-blood efflux system for a cytokine. We postulate that this efflux system may protect the brain from circulating IL-2.

Constitutive expression of p55TNFR mRNA and mitogen-specific up-regulation of TNF alpha and p75TNFR mRNA in mouse brain

Serum tumor necrosis factor (TNF) functions as a mediator of the immune-to-brain axis. Numerous TNF receptor-mediated effects on the nervous system are described but the knowledge about the regional and cellular expression of TNF receptor p55TNFR and p75TNFR in vivo is far from being complete. It is unclear whether TNF mediates its neuroimmune effects alone or in combination with other factors, e.g., bacterial mitogens. Here, we investigated the distribution of TNFalpha, p55TNFR, and p75TNFR in normal mouse brain and examined the stimulus-specific effects of lipopolysaccharide (LPS) and staphylococcal enterotoxin B (SEB) on the expression of the cerebral TNF system. Both mitogens caused enhanced TNFalpha serum levels and induced c-fos mRNA in the paraventricular nucleus but exhibited different effects on the cerebral gene expression of the TNF system. LPS but not SEB rapidly induced TNFalpha mRNA in circumventricular organs (CVOs) followed by spreading of TNFalpha mRNA into brain parenchyma close to the CVOs. The p55TNFR gene was constitutively expressed in many neurons with high levels in brainstem motor nuclei and in neurons of the sensory mesencephalic trigeminal nucleus. Moderate levels of p75TNFR mRNA were seen in single cells scattered throughout the brain in a pattern resembling microglia. Neither LPS nor SEB modulated the p55TNFR gene expression in any region or cell type of the brain, and LPS but not SEB induced p75TNFR mRNA in the CVOs. Thus, enhanced TNF serum levels able to stimulate c-fos mRNA expression in the paraventricular nucleus did not necessarily result in a modulation of the cerebral TNF system.

The organum vasculosum laminae terminalis in immune-to-brain febrigenic signaling: a reappraisal of lesion experiments

The organum vasculosum laminae terminalis (OVLT) has been proposed to serve as the interface for blood-to-brain febrigenic signaling, because ablation of this structure affects the febrile response. However, lesioning the OVLT causes many "side effects" not fully accounted for in the fever literature. By placing OVLT-lesioned rats on intensive rehydration therapy, we attempted to prevent these side effects and to evaluate the febrile response in their absence. After the OVLT of Sprague-Dawley rats was lesioned electrolytically, the rats were given access to 5% sucrose for 1 wk to stimulate drinking. Sucrose consumption and body mass were monitored. The animals were examined twice a day for signs of dehydration and treated with isotonic saline (50 ml/kg sc) when indicated. This protocol eliminated mortality but not several acute and chronic side effects stemming from the lesion. The acute effects included adipsia and gross (14% of body weight) emaciation; chronic effects included hypernatremia, hyperosmolality, a suppressed drinking response to hypertonic saline, and previously unrecognized marked (by approximately 2 degrees C) and long-lasting (>3 wk) hyperthermia. Because the hyperthermia was not accompanied by tail skin vasoconstriction, it likely reflected increased thermogenesis. After the rats recovered from the acute (but not chronic) side effects, their febrile response to IL-1beta (500 ng/kg iv) was tested. The sham-operated rats developed typical monophasic fevers ( approximately 0.5 degrees C), the lesioned rats did not. However, the absence of the febrile response in the OVLT-lesioned rats likely resulted from the untreatable side effects. For example, hyperthermia at the time of pyrogen injection was high enough (39-40 degrees C) to solely prevent fever from developing. Hence, the changed febrile responsiveness of OVLT-lesioned animals is given an alternative interpretation, unrelated to febrigenic signaling to the brain.

Effects of glossopharyngeal nerve transection on central and peripheral cytokines and serum corticosterone induced by localized inflammation

Bilateral transection of the glossopharyngeal nerves (GLOx) disrupts the immune-to-brain communication from the posterior oral cavity. The current report tested whether this effect is due to the afferent (sensory) or efferent (parasympathetic motor) components of the nerve. Injection of lipopolysaccharide (LPS) into the soft palate (ISP) of GLOx or sham-operated (SHAM) rats increased the circulating levels of interleukin-1beta (IL-1beta), interleukin-1 receptor antagonist (IL-1ra) and corticosterone (CORT), as well the hypothalamic content of IL-1beta; no difference in circulating levels and hypothalamic content was found between GLOx and SHAM at 2 and 4.5 h after LPS injection. These results indicate that glossopharyngeal neural efferents do not mediate the effects of GLOx on the immune-to-brain communication.

Morphine Regulates Gene Expression of α- and β-Chemokines and Their Receptors on Astroglial Cells Via the Opioid μ Receptor

The brain is a target organ for recreational drugs and HIV-1. Epidemiological data demonstrate that opioid abuse is a risk factor for HIV-1 infection and progression to AIDS. Chemokines and their receptors have been implicated in the neuropathogenesis of HIV-1 infections. However, little is known about the effects of opioids on the expression of chemokines and their receptors (the latter also are HIV-1 coreceptors) by cells of the CNS. Herein we describe the effects of morphine on gene expression of the alpha- and beta-chemokines and their receptors by the astrocytoma cell line U87 and by primary normal human astrocyte (NHA) cultures. U87 cells treated with morphine showed significant down-regulation of IL-8 gene expression, whereas expression of the IL-8 receptor CXCR2 was reciprocally up-regulated as detected by RT-PCR. Treatment of NHAs with morphine suppressed IL-8 and macrophage-inflammatory protein-1beta gene expression, whereas expression of their receptor genes, CCR3 and CCR5, was simultaneously enhanced. These morphine-induced effects on U87 and NHA cells were reversed by the opioid mu receptor antagonist beta-funaltrexamine. Morphine also enhanced the constitutive expression of the opioid mu receptor on astroglial cells. Our results support the hypothesis that opioids play a significant role in the susceptibility of the CNS to HIV-1 infection and subsequent encephalopathy by inhibiting local production of HIV-1-protective chemokines (IL-8 and macrophage-inflammatory protein-1beta) and enhancing expression of HIV-1 entry coreceptor genes (CCR3, CCR5, and CXCR2) within the CNS. These effects of opioids appear to be mediated through the opioid mu receptor that we demonstrated on astroglial cells.

TNFalpha transport across the blood-brain barrier is abolished in receptor knockout mice

The presence of transport systems at the blood-brain barrier (BBB) enables some cytokines in blood to reach specific targets in the brain and spinal cord. The "transporters" function in a way different from conventional receptors, in that cytokines are chaperoned from blood to the CNS rather than being degraded in the specialized endothelial cells composing the BBB. Here we present the first study to determine whether the transporter for tumor necrosis factor-alpha (TNFalpha) is identical to its receptors. Three types of TNFalpha receptor knockout mice were used, and the influx of (125)I-TNFalpha from blood to brain and blood to spinal cord was measured. In either p55 or p75 receptor knockout mice, the influx of (125)I-TNFalpha was significantly, but not completely, decreased in spinal cord, whereas the decrease in brain was not statistically significant. This indicates that both receptors are partially involved in the transport of TNFalpha across the BBB but that neither receptor is the sole transporter. By contrast, in double knockout mice lacking both p55 and p75 receptors, the entry of (125)I-TNFalpha into brain and spinal cord was completely abolished. Therefore, both receptors are necessary for transporting TNFalpha across the BBB. The results clearly demonstrate that the transport of TNFalpha across the BBB is a complicated process involving additive or even synergistic activities of both receptors, thus differing from typical ligand-receptor binding and downstream signal transduction.

Effect of prophylactic intracerebroventricular injection of methylprednisolone on nitrogen and catecholamine excretion in the urine after laparotomy in rats

BACKGROUND

This study was carried out to determine whether the prophylactic injection of glucocorticoid into the intracerebroventricular (i.c.v.) space reduced TNF-alpha and IL-1beta mRNA synthesis in the brain after laparotomy, resulting in a reduction of nitrogen excretion in the urine.

PATIENTS AND METHODS

Male SD rats (body wt., 225-250 g, n = 114) were catheterized into the i.c.v. space on day 0. On day 4, the rats were assigned to four groups: (1) Control, (2) laparotomy (Trauma), (3) intraperitoneal (i.p.) injection of methylprednisolone (MP) plus laparotomy (IPMP), and (4) i.c.v. injection of MP plus laparotomy (ICVMP). Either 3 or 24 h after surgery, the animals were sacrificed. TNF-alpha and IL-1beta mRNA levels in tissues, including the brain cortex and hypothalamus, were measured by RT-PCR. The amounts of nitrogen and catecholamine excretion in the 24-h urine were determined.

RESULTS

The i.p. injection of MP reduced TNF-alpha and IL-1beta mRNA levels in all the tissues 3 h after laparotomy compared with those of the Trauma group. The icv injection of MP prevented elevation of the TNF-alpha and IL-1beta mRNA levels in the brain (cortex, TNF-alpha, ICVMP 0.43 +/- 0.06, P < 0.05, vs Trauma; cortex, IL-1beta, ICVMP 0.25 +/- 0.09, P < 0.05, vs. Trauma; hypothalamus, TNF-alpha, ICVMP 0.31 +/- 0.04, P < 0.05, vs. Trauma; hypothalamus, IL-1beta, ICVMP 0.25 +/- 0.14, P < 0.05, vs. Trauma), but did not inhibit an increase in TNF-alpha and IL-1beta mRNA levels in the liver and skeletal muscle. Both nitrogen and catecholamine excretions in the urine were decreased by ip and by i.c.v. injection of MP compared to those of the Trauma group (nitrogen, ICVMP 559.3 +/- 52.0 mg/day, P < 0.05, vs. Trauma; catecholamine, ICVMP 13.8 +/- 1.8 microg/day, P < 0.05, vs. Trauma).

CONCLUSION

A reduction in TNF-alpha and IL-1beta mRNA synthesis in the brain due to prophylactic injection of MP into the icv space reduced the catabolic response after laparotomy.

Astrogliosis in the adult and developing CNS: is there a role for proinflammatory cytokines?

Astrogliosis, characterized by the enhanced expression of GFAP, represents a remarkably homotypic response of astrocytes to all types of injuries of the CNS, including injuries of the developing CNS. As such, astrocytes serve as microsensors of the injured microenvironment regardless of their location in the CNS. The diversity of insults that engender astrogliosis and the brain-wide nature of the astrocytic response suggest that common injury factors serve as the trigger of this cellular reaction. One prominent theme that has emerged in recent years is that proinflammatory cytokines and chemokines serve as a stimulus for induction of astrogliosis. Here we present a brief critique of this hypothesis based on a review of literature and some of our own recentfindings. Studies of astrocytes, in vitro, clearly indicate that these cell types are responsive to a variety of growth factors, including cytokines and chemokines. A somewhat different picture, however, can be seen from data obtained in vivo. It is true that trauma and diseases of the nervous system, as well as some exposures to neurotoxic chemicals, can be associated with the expression in brain of large varieties of cytokines and chemokines. That these same conditions result in astrogliosis has fostered the circumstantial link between cytokine/chemokine expression and the induction of astrogliosis. Several lines of evidence argue against this view, including (a) suppression of cytokine expression does not suppress gliosis, (b) gliosis can occur in the absence of enhanced expression of cytokines, (c) elevations in brain cytokines can occur in the absence of gliosis and (d) the patterns of cytokine expression in the adult and developing CNS are more consistent with a trophic role for these chemical messengers rather than a role in the induction of inflammation. Enhanced expression of cytokines and chemokines after brain injury appear to be signal transduction events unrelated to the induction of astrogliosis.

The glossopharyngeal nerve as a novel pathway in immune-to-brain communication: relevance to neuroimmune surveillance of the oral cavity

Glossopharyngeal afferents may be the neural channel by which immune challenge of the posterior oral cavity conveys information to the brain. If this is the case, then bilateral transection of the glossopharyngeal nerves (GLOx) should disrupt this communication. Injection of lipopolysaccharide (LPS) or interleukin (IL)-1beta into the soft palate (ISP) of sham-operated rats induced a dose-related febrile response. GLOx significantly attenuated the febrile response induced by ISP injection of both LPS and IL-1beta. In contrast, GLOx did not affect the febrile response when LPS or IL-1beta were injected intraperitoneally, indicating that the effect of GLOx is not systemic. These results provide experimental evidence for a novel neural pathway for immune-to-brain communication.

Differential gene expression during meningeal-meningococcal interaction: evidence for self-defense and early release of cytokines and chemokines

Using microarray technology, we studied the early differential expression of 3,528 genes in human meningothelial cells in response to meningococcal challenge. Thirty-two genes were up-regulated, and four were down-regulated. Those up-regulated included the tumor necrosis factor alpha, interleukin-6 (IL-6), and IL-8 (but not IL-1beta) genes, suggesting that meningeal cells may be a local and early source of these cytokines. Also, a trend in up-regulation of anti-apoptotic genes and down-regulation of pro-apoptotic genes was observed. This is the first evidence that meningothelial cells may mount cytoprotective responses to pathogenic bacteria.

Severity of trauma changes expression of TNF-alpha mRNA in the brain of mice

BACKGROUND

The greater nitrogen loss that occurs with increasing severity of trauma is believed to occur because activation of the hypothalamus-pituitary axis is greater with severe injury. Cytokines in the brain stimulate the hypothalamus-pituitary-adrenal axis. This study was carried out to investigate whether the brain would recognize severity of trauma via TNF-alpha mRNA synthesis in the brain.

METHODS

Male C57BL/6 mice (n = 70, BW: 20-28 g) were randomly assigned into four groups, (1) control (no anesthesia or incision), (2) anesthesia alone, (3) anesthesia plus laparotomy by short incision (short), and (4) anesthesia plus laparotomy by long incision (long). A laparotomy was carried out in the short and long groups by a 1.2-cm vertical incision and by a horizontal plus a vertical incision (2.4 x 2.4 cm), respectively. Exactly either 3 or 24 h after surgery, the animals were decapitated. TNF-alpha mRNA levels in the tissues were determined by semi-quantitative PCR.

RESULTS

Nitrogen and catecholamine excretion were increased in the long wound group compared with the short wound group. Expression of TNF-alpha mRNA in the brain was greater in the long group after surgery than in the control, anesthesia, and short groups (brain, long: 0.150 +/- 0.005; P < 0.01 vs control, anesthesia alone, and short groups), but TNF-alpha levels in the plasma were the same in the short and long groups after surgery.

CONCLUSION

Levels of TNF-alpha mRNA in the brain were enhanced according to the length of the wound probably because of greater neural stimuli from the wound site, and this elevation was involved in the greater nitrogen loss.

Penetration of neurotrophins and cytokines across the blood-brain/blood-spinal cord barrier

Now that peptides are no longer considered too large to cross the blood-brain barrier, attention has turned to the possibility that larger substances like polypeptides might also enter the central nervous system (CNS). This review summarizes evidence showing that many cytokines and neurotrophins not only enter the brain but also enter the spinal cord, sometimes faster than into the brain.

Interleukin-6 and its soluble receptor in serum and cerebrospinal fluid after cerebral trauma

Interleukin-6 (IL-6) and its soluble receptor (sIL-6-R) were measured in cerebrospinal fluid (CSF) and serum of 11 severely head injured patients for up to 3 weeks following trauma. IL-6 increased immediately after injury displaying much higher concentrations in CSF than in serum (n = 11). Differently, median levels of sIL-6-R remained in the normal ranges being 10 times higher in serum than in CSF. However, increased amounts over control levels were found in CSF (n = 7) and intrathecal release of sIL-6-R was also suggested (n = 7). Although no correlation with the extent of cerebral lesion or with clinical outcome was evident, elevation of sIL-6-R in CSF supports a pivotal role for IL-6/sIL-6-R complex in the injured brain.

Cytokines in the neuroendocrine system

Cytokines are important partners in the bidirectional network interrelating the immune and the neuroendocrine systems. These substances and their specific receptors, initially thought to be exclusively present in the immune system, have recently been shown to be also expressed in the neuroendocrine system. Cytokines can modulate the responses of all endocrine axes by acting at both the central and the peripheral levels. To explain how systemic cytokines may gain access to the brain, several mechanisms have been proposed, including an active transport through the blood-brain barrier, a passage at the circumventricular organ level, as well as a neuronal pathway through the vagal nerve. The immune-neuroendocrine interactions are involved in numerous physiological and pathophysiological conditions and seem to play an important role to maintain homeostasis.

Afferent pathways of pyrogen signaling

We and others recently showed that fever induced by intravenously or intraperitoneally injected lipopolysaccharide (LPS) may involve brain signaling via hepatic vagal afferents. This suggests that LPS fever may be initiated by mediators released mainly by cells in the liver, presumably macrophages (Kupffer cells, Kc). To verify this possibility, we disabled the Kc of conscious guinea pigs with gadolinium chloride and monitored their core temperature and associated preoptic prostaglandin E2 (PGE2) responses to i.v. LPS. Gadolinium chloride pretreatment significantly attenuated both the febrile and PGE2 rises, thus supporting the hypothesis. Additionally, fluorescein-labeled LPS was detected in Kc 15 minutes after its i.v. administration. Paradoxically, however, the label was also present in gadolinium chloride-pretreated guinea pigs. Thus, either Kc are not the primary source of pyrogenic mediators or LPS does not provide the stimulus for their production. Because the i.v. injection of LPS elicits virtually immediately the production of complement fragments, and Kc express their receptors and produce various mediators on their activation, we hypocomplemented guinea pigs with cobra venom factor. The core temperature rises produced by i.v. LPS were reduced by complement depletions > 60%. LPS i.v. per se decreased complement, that is, complement was consumed by 12% within 10 minutes. Thus, the onset of LPS fever may involve complement system and Kc activation, but their precise roles await clarification.

Complement reduction impairs the febrile response of guinea pigs to endotoxin

Although it is generally believed that circulating exogenous pyrogens [e.g., lipopolysaccharides (LPS)] induce fever via the mediation of endogenous pyrogens (EP) such as cytokines, the first of these, tumor necrosis factor-alpha, is usually not detectable in blood until at least 30 min after intravenous administration of LPS, whereas the febrile rise begins within 15 min after its administration. Moreover, although abundant evidence indicates that circulating LPS is cleared primarily by liver macrophages [Kupffer cells (KC)], these do not secrete EP in immediate response. This would imply that other factors, presumably evoked earlier than EP, may mediate the onset of the febrile response to intravenous LPS. It is well known that blood-borne LPS very rapidly activates the intravascular complement (C) system, some components of which in turn stimulate the quick release into blood of various substances that have roles in the acute inflammatory reaction. KC contain receptors for C components and are in close contact with afferent vagal terminals in the liver; the involvement of hepatic vagal afferents in LPS-induced fever has recently been shown. In this study, we tested the hypothesis that the initiation of fever by intravenous LPS involves, sequentially, the C system and KC. To test this postulated mechanism, we measured directly the levels of prostaglandin E2 (PGE2) in the interstitial fluid of the preoptic anterior hypothalamus (POA), the presumptive site of the fever-producing controller, of conscious guinea pigs over their entire febrile course, before and after C depletion by cobra venom factor (CVF) and before and after elimination of KC by gadolinium chloride (GdCl3). CVF and GdCl3 pretreatment each individually attenuated the first of the biphasic core temperature (Tc) rises after intravenous LPS, inverted the second into a Tc fall, and greatly reduced the usual fever-associated increase in POA PGE2. We conclude, therefore, that C activation may indeed be pivotal in the induction of fever by intravenous LPS and that substance(s) generated presumably by KC in almost immediate reaction to the presence of LPS and/or C may transmit pyrogenic signals via hepatic vagal afferents to the POA, where they rapidly induce the production of PGE2 and, hence, fever.

Inhibition of tumor necrosis factor and amelioration of brain infarction in mice

Tumor necrosis factor alpha (TNF-alpha) is expressed in the ischemic brain; however, its precise role is not fully understood. We studied the effect of the dimeric form of the type I soluble TNF receptor linked to polyethylene glycol (TNFbp) on focal cerebral ischemia in mice using a permanent middle cerebral arterial occlusion (MCAO) model. TNFbp was applied topically, intravenously, or intraperitoneally. TNFbp binds and inhibits TNF-alpha. The volume of cortical ischemic lesions was measured by means of 2,3,5-triphenyltetrazolium chloride 24 h after MCAO. TNFbp produced a significant reduction in the cortical infarct volume of vehicle-treated animals (p < 0.001). The reduction in the volume of brain damage was 26% in animals that received 3 mg/kg of TNFbp topically. Further analysis of TNF-alpha inhibition following acute brain ischemia is indicated.