Induction of inhibitory factor κBα mRNA in the central nervous system after peripheral lipopolysaccharide administration: An in situ hybridization histochemistry study in the rat

Abstract

In this study we investigate the mRNA expression of inhibitory factor κBα (IκBα) in cells of the rat brain induced by an intraperitoneal (i.p.) injection of lipopolysaccharide (LPS). IκB controls the activity of nuclear factor κB, which regulates the transcription of many immune signal molecules. The detection of IκB induction, therefore, would reveal the extent and the cellular location of brain-derived immune molecules in response to peripheral immune challenges. Low levels of IκBα mRNA were found in the large blood vessels and in circumventricular organs (CVOs) of saline-injected control animals. After an i.p. LPS injection (2.5 mg/kg), dramatic induction of IκBα mRNA occurred in four spatio-temporal patterns. Induced signals were first detected at 0.5 hr in the lumen of large blood vessels and in blood vessels of the choroid plexus and CVOs. Second, at 1–2 hr, labeling dramatically increased in the CVOs and choroid plexus and spread to small vascular and glial cells throughout the entire brain; these responses peaked at 2 hr and declined thereafter. Third, cells of the meninges became activated at 2 hr and persisted until 12 hr after the LPS injection. Finally, only at 12 hr, induced signals were present in ventricular ependyma. Thus, IκBα mRNA is induced in brain after peripheral LPS injection, beginning in cells lining the blood side of the blood–brain barrier and progressing to cells inside brain. The spatiotemporal patterns suggest that cells of the blood–brain barrier synthesize immune signal molecules to activate cells inside the central nervous system in response to peripheral LPS. The cerebrospinal fluid appears to be a conduit for these signal molecules.

One way that the central nervous system (CNS) responds to peripheral immune challenges is by generating its own immune signal molecules (1). The mechanisms by which this phenomenon is induced have not been elucidated. A key regulatory component of intracellular signal pathways in cells of the immune system is the nuclear factor κB (NF-κB)/rel family of transcription factors. These factors respond to numerous immune challenges by activating gene transcription for a wide variety of immune signal molecules (2). Tracing the activation of NF-κB in the CNS after peripheral immune challenges may therefore reveal the extent and neuroanatomical location of the processes that are involved in responding to peripheral immune signals and, in turn, generating centrally derived immune molecules.

The mechanisms by which NF-κB regulates gene expression have recently been reviewed (3–5). Briefly, NF-κB is normally present in the cytoplasm as a dormant complex with an inhibitor, IκB. Upon extracellular stimulation, IκB is phosphorylated and degraded. This leads to the release of free NF-κB, which translocates to the nucleus and binds to specific κB sequences of the DNA to initiate transcription of related genes including immunoreceptors, cytokines, and, interestingly, its own inhibitor, IκB.

Two unique features of the NF-κB/IκB system result from its feedback regulation. First, the activation of transcription by NF-κB also triggers the synthesis of IκB (4). Second, NF-κB-activated transcription is maintained by continuous degradation of IκB (5), which is sustained by ongoing extracellular stimulation (6). Thus, the expression of IκB mRNA parallels both the NF-κB activity and the duration of the activating extracellular stimulation. This temporal parallelism between IκB mRNA expression and the effective external stimulation is different from other transcription factors (e.g., c-Fos) that are transiently expressed only at the onset of cellular stimulation (7). These features may be exploited to observe the localization of cells undergoing NF-κB-regulated gene transcription and infer the presence and duration of extracellular stimuli.

In this study, we measured IκBα mRNA expression by in situ hybridization histochemistry in brain sections of rats after they were injected intraperitoneally (i.p.) with lipopolysaccharide (LPS) or sterile saline to trace the activation of responsive cells that potentially synthesize immune signal molecules in the CNS. Because LPS is a large molecule that is generally thought not to diffuse significantly across the intact blood–brain barrier (BBB) (8), we specially examined the brain levels containing areas with a leaky BBB, such as the circumventricular organs (CVOs) (9). We selected IκBα as a marker of NF-κB activation instead of the other members of the IκB family because it is ubiquitous and responds to most forms of extracellular stimulation (10). Our results show profound LPS-induced elevation of IκBα mRNA levels in selected cells in the CNS. The apparent migration of the observed IκBα mRNA labeling across the BBB into deep cells of the brain suggests pathways by which peripheral LPS may activate the synthesis of immune molecules in CNS.

MATERIALS AND METHODS

Animals and LPS Injection.

Male Sprague–Dawley rats (175–200 gm; Taconic Farms) were group housed and handled daily before experimentation. They were injected i.p. with either LPS (2.5 mg/kg) dissolved in 0.9% saline or 0.9% sterile saline alone. Animals were then killed by decapitation at 0.5, 1, 2, 4, 8, 12, and 24 hr after the injection (n = 3–5 per time point). Five other groups of animals were injected intravenously (i.v.) with 1, 5, 25, or 125 μg/kg or with 2.5 mg/kg of LPS (n = 3–5 per dose) and killed 3 hr after the injection. To control for the effects due to saline injection alone, control animals that received no injection were also killed (n = 3). Injections were timed so that the time of sacrifice fell between 09:00 and 13:00 hr.

Brain Section Collection.

Brains were removed immediately after decapitation, frozen by immersion in 2-methyl butane at −30°C, and stored at −70°C before sectioning. They were then cryostat-cut to 15-μm-thick coronal sections and thaw-mounted onto gelatin-coated slides, dried, and stored at −40°C until further processing. Levels collected were organum vasculosum of the lamina terminalis (OVLT) (−0.02 mm relative to bregma); subfornical organ (SFO) (−0.92 mm); central nucleus of amygdala containing also the arcuate nucleus and median eminence (−3.3 mm); and area postrema (AP) containing the nucleus of the solitary tract (NTS) (−13.7 mm) (11).

In Situ Hybridization.

The in situ hybridization protocols were performed as described previously for ribonucleotide (cRNA) probes (12). First, tissue sections were processed by fixation with 4% formaldehyde solution, acetylation with 0.25% acetic anhydride in 0.1 M triethanolamine-HCl (pH 8.0) solution, dehydration with ethanol, and delipidation with chloroform.

Second, the antisense probes directed against the full-length (1.05 kb) rat IκBα cDNA inserted into the pBluescript plasmid (generously provided by Rebecca Taub, University of Pennsylvania) was transcribed using the Riboprobe System (Promega) with T7 RNA polymerase and [α-³⁵S]UTP (specific activity > 1,000 Ci/mmol; New England Nuclear; 1 Ci = 37 GBq) after linearization with BamHI restriction enzyme (Promega). To control for the specificity of the probe, sense probes of rat IκBα were also generated by transcribing the same plasmid with T3 RNA polymerase after linearization with HindIII restriction enzyme (Promega). Finally, radiolabeled probes were diluted in the riboprobe hybridization buffer and applied to brain sections (500,000 cpm/section). After overnight incubation at 55°C in a humidified chamber, slides containing brain sections were washed first in 20 μg/ml RNase solution and then in 2× standard saline citrate (SSC) and 0.2× SSC (55°C and 60°C) solutions to reduce nonspecific binding of the probe. The slides were then dehydrated with ethanol and air-dried for autoradiography.

Autoradiography.

Slides and ¹⁴C plastic standards containing known amounts of radioactivity (American Radiochemicals, St. Louis) were placed in x-ray cassettes, apposed to film (BioMax MR; Kodak) for 4 days, and developed in an automatic film developer (X-Omat; Kodak). To determine anatomical localization of hybridized probes at the cellular level, sections were dipped in nuclear track emulsion (NTB-2; Kodak), exposed for 2 weeks, developed (D19; Kodak) for 2 min at 16°C, and counterstained with cresyl violet.

Combined Immunohistochemistry and Hybridization Histochemistry.

Four cell type-specific antibodies were used to determine the phenotypes of IκBα mRNA-producing cells: anti-PECAM (platelet and endothelial cell adhesion molecule, generously donated to us by W. F. Hickey, Dartmouth Medical School); anti-GFAP (glial fibrillary acidic protein; ICN); OX-42 (PharMingen), and ED2 (Serotec). These antibodies specifically mark endothelial cells (13), astrocytes, microglia, and perivascular monocytes (14), respectively. The double labeling was carried out as follows. Fresh-frozen slide-mounted brain sections were fixed in 4% formaldehyde for 15 min and rinsed three times in PBS; they were then incubated with the above primary antibodies diluted in PBS (1:400 for anti-PECAM, anti-GFAP, and ED2; 1:2000 for OX-42) for 2 hr. After three rinses in PBS, sections were incubated with the biotinylated secondary antibody (goat anti-mouse IgG, 1:200) for 1 hr, and antibody bindings were visualized by the conventional avidin-biotin immunoperoxidase protocol. These sections were then re-fixed in 4% formaldehyde, treated with 1 μl/ml proteinase K for 15 min, acetylated, and dehydrated. Subsequent localization of IκBα mRNA by in situ hybridization was then carried out as described above.

Data Analysis.

Autoradiographic film images of brain sections and standards were digitized on a Macintosh computer-based image analysis system with image software (Wayne Rasband, National Institute of Mental Health). Light transmittance through the film was measured by outlining the structure on the TV monitor. A density-slice function was applied to each structure to select densities greater than film background and thus measured transmittance confined to the cellular sources of the radioactivity. The density so obtained was used to represent the relative amount of mRNA expression of IκBα.

RESULTS

Representative film autoradiographs of IκBα mRNA hybridization in coronally cut sections at the level of SFO are shown in Fig. 1. No labeling was observed in any sections labeled by sense probes (data not shown). With antisense probe, only faint labeling in SFO, choroid plexus, and meninges was seen in a saline injected animal at 0.5 hr postinjection (Fig. 1A). The same pattern and intensity of IκBα mRNA hybridization were found in all the saline-injected and noninjected animals (data not shown). The levels of IκBα mRNA expression in these animals were measured by densitometry and designated as basal levels. The density of labeling measured in any other conditions will be expressed in parentheses as the percentage of these basal levels. The SEM was typically <10% of the mean levels of radioactivity.

Figure 1.

Representative film autoradiographs at the level of SFO show patterns of IκBα mRNA hybridization in animals killed after saline injection (Sal; A) and at 0.5 hr (B), 2 hr (C), and 12 hr (D) after the LPS injection. The arrows in B and C point to blood vessels. Ch Plx, choroid plexus; 3V, third ventricle; V Ep, ventricular ependyma.

After LPS injection, induction of IκBα was found at several time points. At 0.5 hr, induction of IκBα mRNA was apparent in SFO (122%), choroid plexus (287%), and large blood vessels at the base of the brain (Fig. 1B). At 2 hr, the LPS injection induced maximal levels of IκBα mRNA observed in this experiment (Fig. 1C) in SFO (924%) and choroid plexus (362%). In addition, meninges at the base of the brain, small penetrating blood vessels, and scattered cells inside the brain were now strongly expressing IκBα mRNA. At 4 and 8 hr, expression levels in all the previously stimulated regions except meninges had declined to near the basal levels (data not shown). At 12 hr, IκBα message in SFO (119%), choroid plexus (117%), and blood vessels was still just slightly above the basal levels (Fig. 1D). A previously unaffected structure, the ventricular ependyma, by contrast, was now strongly labeled (Fig. 1D). There was a 97% increase of hybridization signal over the basal level in these cells. The induction of IκBα in the meningeal cells at the base of the brain persisted until this time; i.e., the labeling density was still three times the basal level, and it was three fourths of the maximal level seen at 2 hr. At 24 hr, the IκBα expression in all the brain regions declined toward the basal levels (data not shown).

The global pattern of the induction of IκBα mRNA at other levels, such as OVLT and central nucleus of amygdala containing median eminence, was similar to the pattern seen at SFO level; i.e., IκBα mRNA was first induced in blood vessels, CVOs, and choroid plexus, then in cells of the meninges and brain parenchyma, and finally in ventricular ependyma while previously activated regions declined. However, two interesting differences were observed at the level of AP.

Viewed under low-power dark-field microscopy, the induction of IκBα mRNA at the level of the AP, another CVO, is shown in emulsion-coated sections in Fig. 2. Again, in saline-injected animals labeling was seen at low level in the AP and was not detectable in the surrounding area (basal levels, Fig. 2A). In the AP, IκBα message rose at 0.5 hr, increased further at 1 hr, peaked at 2 hr, and declined to slightly above the basal level by 4 hr (Fig. 2 B–E). At 12 hr, however, strong IκBα labeling reappeared once again (Fig. 2F) before it subsided toward the basal level at 24 hr (data not shown). In the tissue surrounding the AP, induced IκBα message was just detectable at 0.5 hr, became apparent in small penetrating blood vessels and in scattered small cells of the brain parenchyma at 1 hr, intensified to the maximal level in evenly distributed small cells of the brain parenchyma at 2 hr, and declined at 4 hr. At 12 hr, however, strong labeling also reappeared, but this time restricted only to small cells in the area of the NTS.

Figure 2.

Low-power dark-field photomicrographs at the level of AP show IκBα mRNA labeling in animals sacrificed after saline injection (Sal; A), and at 0.5 hr (B), 1 hr (C), 2 hr (D), 4 hr (E), and 12 hr (F) after LPS injection.

Viewed under high-magnification bright-field microscopy, representative large blood vessels and the adjacent arachnoid at the base of the brain are shown in Fig. 3. Sparse labeling was seen in saline injected animals (Fig. 3A). At 0.5 hr, the blood vessels were intensely labeled and the message was restricted to the cells on the lumenal side of the blood vessel (Fig. 3B). At 2 hr, IκBα mRNA labeling was less intense over the lumen of the blood vessel, but cells of the arachnoid had become labeled (Fig. 3C). At 12 hr, the labeling of IκBα mRNA in the blood vessels had returned to the basal level, but the arachnoid remained labeled (Fig. 3D).

Figure 3.

High-magnification bright-field photomicrographs show labeling of IκBα mRNA in blood vessels (BV) at the base of the brain in animals killed after saline injection (Sal; A), and at 0.5 hr (B), 2 hr (C), and 12 hr (D) after the LPS injection. Arrows point to cells of the blood vessels, and arrowheads point to cells in the arachnoid membrane.

Fig. 4 shows high-magnification photomicrographs of IκBα mRNA expression in the choroid plexus after the LPS injection. The expression of IκBα mRNA in saline-injected animals was barely detectable (Fig. 4A). At 0.5 hr, cells of the choroidal blood vessels were labeled, whereas the surrounding choroidal epithelia were not (Fig. 4B). At 2 hr, however, both blood vessels and choroidal ependyma were labeled (Fig. 4C). Labeling declined toward the basal level thereafter (data not shown). Similarly, the IκBα mRNA induction in various CVOs at 0.5 hr was primarily due to the induction of IκBα mRNA in the vascular cells within the CVOs, and the later expression of IκBα mRNA was found in cells of both the blood vessels and the CVOs proper (data not shown).

Figure 4.

High-magnification bright-field photomicrographs show labeling of IκBα mRNA in the choroid plexus in animals killed after saline injection (Sal; A) and at 0.5 hr (B) and 2 hr (C) after the LPS injection. Arrows point to cells of the blood vessels (BV), and arrowheads point to cells in the choroidal ependyma (Ch Ep).

The phenotypes of cells in the brain parenchyma that express IκBα mRNA at 2 hr after the LPS injection are illustrated in Fig. 5. Induced IκBα mRNA is primarily colocalized with endothelial cells (Fig. 5A) and partially with astrocytes (Fig. 5B), but not with microglia (Fig. 5C) or perivascular monocytes (Fig. 5D). IκBα mRNA expression was not detectable in any of these cell types in the brain parenchyma of saline-injected animals.

Figure 5.

High-magnification bright-field color photomicrographs show hybridization of IκBα mRNA (arrows point to clusters of silver grains where cells are positive) and immunohistochemical labeling (brown reaction product) of platelet and endothelial cell adhesion molecule (PECAM) (A), glial fibrillary acidic protein (GFAP) (B), OX-42 (C), and ED2 (D). In the OX-42 and ED2-labeled sections, the silver grains did not colocalize with the immunostaining. The section in D was counter stained with cresyl violet to show that the silver gains are over an adjacent Nissl-stained cell. 

DISCUSSION

The results of the present study show profound induction of IκBα mRNA in the CNS after peripheral LPS administration. Four types of responses are distinguishable. First, cells positioned at the blood side of the BBB, including those on the lumenal side of the blood vessels at the base of the brain, vascular cells of the choroid plexus and CVOs, and cells of the small penetrating blood vessels were IκBα mRNA-positive as early as 0.5–1 hr postinjection. Second, beginning at 1 hr postinjection, endothelial cells and some astrocytes throughout the entire brain were activated, although the highest density of the IκBα mRNA-producing cells were localized in the CVOs and choroid plexus. These two types of responses peaked at 2 hr after the LPS injection and declined thereafter. Third, meningeal cells at the edge of the brain became strongly activated at 2 hr after the LPS injection and persisted for 10 more hours before the response was diminished. Fourth, at 12 hr after the LPS injection, while IκBα mRNA expression in most of the brain regions declined toward the basal levels, newly activated IκBα mRNA-expressing cells appeared in the ventricular ependyma, AP, and NTS. Taken together, the peripheral injection of LPS had induced waves of expression of IκBα mRNA in nonneuronal cells throughout the brain, the induction pattern migrating from the cells of the BBB to cells inside of the BBB. These patterns of IkBα mRNA expression are also observed after either i.p. or i.v. injections of various doses LPS. The lowest effective dose of LPS was 1 μg/kg (unpublished data).

In peripheral cells, NF-κB is known to trigger the expression of numerous genes in the immune system including immunoreceptors, cytokines, adhesion molecules, Rel/NF-κB members, and IκB members (15). Similarly, in dissociated cells of the CNS, NF-κB has recently been shown to regulate the production of cytokines (16, 17) and immunoreceptors (18). The observed pattern of IκBα mRNA induction, therefore, indicates the location and time that cells in the CNS might be activated to produce these various immune signal proteins after the peripheral LPS injection, although the specific genes expressed by the activated cells are not determined.

The onset and duration of the expression of IκBα mRNA, as discussed earlier, reflects the time course of stimulus-induced activation of NF-κB-regulated gene transcription. Of the four types of stimuli currently known to activate NF-κB—pathogens, mitogens, cytokines, and physical stress (3)—LPS (a pathogen and a mitogen) and the cytokines it induces are the likely activating stimuli in the present study. Therefore, the induction of IκBα mRNA also indicates the arrival and continued presence of either LPS or cytokines in the activated brain regions. This is functionally significant because many LPS-induced centrally mediated effects such as fever (19), prolonged slow wave sleep (20), and altered neuroendocrine activity (21) have been attributed to the pro-inflammatory cytokines that LPS induces. Therefore, the following discussion will focus on the pathways, suggested by the present results, by which cytokines might arrive in the CNS after the LPS injection.

A salient finding of the present study is the initial response after the LPS injection in cells of the cerebrovasculature. The early responsiveness of these cells may be expected because they are readily accessible by circulating LPS and cytokines. Cells inside the BBB, by contrast, are not freely accessible by blood-borne LPS (8) and cytokines [although small amounts of circulating cytokines have been reported to be transported across the BBB (22)]. In addition, receptors for LPS (23, 24), tumor necrosis factor α (25), and interleukin 1 (26) have been localized in the endothelial cells of brain vasculature. Moreover, the activation of the NF-κB system in these cells suggests that they respond to the peripheral stimulation by producing their own signal molecules, probably cytokines. This is supported by our recent finding that interleukin 1β mRNA is rapidly synthesized in the CVOs and in brain vascular cells after peripheral LPS injection (27).

At 2 hr after the LPS injection, many astrocytes throughout the entire brain were seen to express IκBα mRNA. This suggests that cytokines signals have now reached brain parenchyma, and, importantly, the responsive astrocytes are probably producing centrally derived cytokines. It is noteworthy that the activation of cells in the brain proper is uniform, independent of the distance from the CVOs (comparing Fig. 2 D to F). This is significant because it has been shown that if large molecules excluded by the BBB were to enter the brain, they will first do so at the CVOs and then diffuse slowly to the vicinity of these structures (9), displaying concentration gradients moving away from the CVOs, similar to that seen in Fig. 2F. The activation pattern we observed in the brain parenchyma at 2 hr after the LPS injection (Figs. 1C and 2D), therefore, is not likely to be produced by LPS and cytokines that arrived by this pathway. The pattern may be best explained by cytokines that were either generated in the cells of brain blood vessels or transported from circulation across the entire BBB. Because only small amounts of certain cytokines (i.e., less than 5% of the blood levels) have been shown to be transported across the BBB (22), it is possible that the majority of the cytokines present in CNS at 2 hr was synthesized by the earlier observed IκBα mRNA-expressing vascular cells. Presumably, these cells secrete cytokines on their ablumenal side. Indeed, we found numerous endothelial cells expressing IkBα mRNA.

The highest local concentrations of IκBα mRNA-expressing cells were seen in the CVOs, choroid plexus, and meninges at 2 hr after the LPS injection. This is probably due to the concentration of cytokine-responsive cells in these sites. For example, it has been shown that type 1 interleukin 1 receptors are highly concentrated in AP, choroid plexus, and meninges (28). In addition, these sites may also receive stimulation by substantial amounts of peripheral cytokines because they possess a leaky BBB (9). The present results suggest that at 2 hr after the LPS injection, the main cytokine-producing cells are located in these sites. Because the CVOs, choroid plexus, and meninges are also bathed by interstitial cerebrospinal fluid (CSF), molecules secreted from these cells will be transported by CSF flow. The persistent activation of the meningeal cells at the base of the brain after this time may result from the stimulation by cytokines flowing in the exiting subarachnoid CSF. On the other hand, cytokines secreted from these cells may also be delivered by the ventricular CSF to brain parenchyma, especially to structures in the close vicinity of the ventricle.

The finding that IκBα mRNA was induced in cells of the ventricular ependyma at 12 hr after the peripheral LPS injection while most of the CVOs and blood vessels were no longer active is very interesting, but puzzling. Because these cells were not stimulated earlier, the extracellular stimuli that are present at 12 hr may be different from those present at earlier time points. It is not clear, however, whether these stimuli arrived from peripheral circulation or from late responsive cells of the CNS. The close vicinity of these cells to the ventricles suggests that ventricular CSF provides a functional conduit for these signal molecules.

Beyond suggesting pathways for the transduction of peripheral immune signals into the brain, the present results may have important implications for the pathogenesis of viral infections of the CNS. This is because the replication of many viruses, including HIV, are regulated by NF-κB (3). It is conceivable that viral proliferation in the CNS could be facilitated when peripheral infections induce NF-κB activation in the cells of the CNS similar to that seen in the present study.