We report here that MCP-1-/- mice show a decreased upregulation of brain cytokine and chemokine production and decreased activation of brain immune cells after systemic LPS treatment compared to wild-type MCP-1+/+ mice. These decreased responses were correlated with lower serum corticosterone levels in the knockout mice. In contrast, there were significantly higher levels of the pro-inflammatory cytokines IL-1β and TNF-α in the serum of LPS-injected MCP-1-/- mice compared to LPS-injected MCP-1+/+ mice. These data suggest that during endotoxemia MCP-1 is necessary for full activation of the brain immune cells and production of endogenous brain inflammatory mediators. Further, the lower levels of corticosterone, a potent immunosuppressant, in the serum of MCP-1-/- mice may be one factor involved in the increased immune response in the periphery in these mice.
The mechanistic relationship between MCP-1 and the production of peripheral inflammatory mediators is not completely understood. Previous studies have shown that MCP-1 knockout mice or wild type mice treated with MCP-1 neutralizing antibodies show a reduction in the serum levels of the anti-inflammatory cytokine IL-10 and increased levels of the pro-inflammatory cytokine TNF-α after LPS injection [15, 27]. Further, when recombinant mouse MCP-1 is administered to mice receiving LPS, IL-10 levels are increased and TNF-α levels are decreased in the serum . Our data showing increased levels of IL-1β and TNF-α in the serum of LPS-injected MCP-1-/- mice compared to wild-type mice are consistent with these previous studies, and suggest that MCP-1 may be involved in controlling the balance between production of pro-inflammatory and anti-inflammatory mediators by peripheral immune cells.
Another mechanism by which MCP-1 may influence peripheral inflammatory responses is by affecting the production of immune regulating glucocorticoids. Glucocorticoids like corticosterone are produced by the adrenal gland after activation of the HPA axis and exert powerful anti-inflammatory properties . Glucocorticoids can both suppress production of inflammatory mediators such as IL-1β, TNF-α, prostaglandins, free oxygen radicals, and nitric oxide, and also upregulate expression of anti-inflammatory cytokines such as IL-10 and TGF-β [3, 33]. Further, adrenalectomized mice show increased mortality and increased plasma levels of IL-1β and TNF-α after LPS treatment and this mortality can be reversed by treatment with glucocorticoids . The lower levels of corticosterone we detected in the serum of LPS-injected MCP-1-/- mice may contribute to the increased levels of serum cytokines seen in these mice, but this idea needs to be further tested.
In addition to its effects on peripheral inflammatory responses, LPS administration can lead to upregulation of detrimental neuroinflammatory responses. Recent studies have shown that a single injection of a high dose of LPS (5 mg/kg) causes a decrease in neuronal numbers in the substantia nigra 10 months after injection and is correlated to increased microglial activation and TNF-α production . Other studies have found long-term losses in neurons, cognitive deficits, or decreases in hippocampal size after peripheral LPS exposure [35, 36]. Although a number of previous studies have examined the relationship between MCP-1 and inflammatory cytokine production in the peripheral immune system, less is known about the role of MCP-1 in CNS inflammation after LPS administration. Our data showing that a peripheral injection of LPS induces an increase in brain MCP-1 levels are consistent with a previous report showing an increase in brain MCP-1 protein levels at 1 hour after a 3 mg/kg i.p. LPS injection . In other studies, MCP-1 mRNA was found to be rapidly expressed starting at 30 minutes in areas lacking the BBB and along blood vessels , and MCP-1 mRNA expression appeared throughout the brain parenchyma starting by 3 hours and peaking at 6–8 hours after systemic LPS administration . Further, in the brain both immune cells like macrophages/microglia  and non-immune cells like endothelia  and astrocytes  can produce MCP-1 after peripheral LPS. These data suggest a pattern of LPS-induced MCP-1 expression that begins in the periphery and spreads to the brain, first in areas that are in contact with peripherally circulating inflammatory mediators and then throughout other areas of the brain. Other inflammatory mediators also follow a similar site-specific pattern of expression after peripheral LPS, including cyclooxygenase-2, TNF-α, IL-1β, inducible nitric oxide synthase and Iκ Bα, where these molecules are first upregulated in the choroid plexus, circumventricular organs, and/or near blood vessels [7, 38–41].
Whether LPS-induced increases in MCP-1 in brain are mechanistically coupled to changes in brain inflammation has not been explored in detail. The results reported here showing that MCP-1-/- mice have decreased brain inflammation after systemic LPS administration compared to wild-type mice suggest that MCP-1 has an important role in activating the brain during peripheral endotoxemia. There are several possible mechanisms by which MCP-1 can regulate neuroinflammation and the production of inflammatory mediators in the CNS.
First, MCP-1 in the brain may act as a chemoattractant to recruit microglia or other immune cells to areas of pathology or inflammation; once activated near the site of pathology, the recruited cells can produce more pro-inflammatory mediators thus increasing inflammation. This possibility is supported by observations that MCP-1 is chemotactic to microglia in vitro , and that injection of recombinant MCP-1 into the striatum or hippocampus of mice results in a significant increase in numbers of microglia/monocytes near the injection site compared to vehicle-injected mice [43, 44]. Other animal models also support a chemoattractant role for MCP-1. For example, in rodent stroke models, inhibition of MCP-1 using knockout mice or anti-MCP-1 gene therapy results in fewer activated microglia and astrocytes and significantly smaller infarcts [22, 45], while overexpression of MCP-1 is associated with increased chemoattraction of immune cells and larger infarct volumes . MCP-1 has also been implicated in immune cell recruitment in several other CNS disorders such as demyelinating diseases, lesion models, and cortical injury [23, 47, 48]. Our findings of decreased activation of microglia as measured by CD45 staining in the LPS-treated MCP-1-/- mice that correlates with a decrease in brain inflammatory cytokines suggest that MCP-1 is necessary for recruitment and activation of microglia. Future studies are necessary to further delineate the detailed mechanisms, such as flow cytometry to quantify cell infiltration into the brain and MCP-1 production.
Second, MCP-1 may be required for glia to mount an effective neuroinflammatory response, either through altering the balance between anti-inflammatory and pro-inflammatory responses towards a more pro-inflammatory profile or in "priming" glial cells toward a more inflammatory profile in the brain. Our results so far with the MCP-1-/- mice do not indicate a major increase in anti-inflammatory mediators in the brain. For example, IL-10 levels were not different in the LPS-injected MCP-1-/- mice versus the LPS-injected MCP-1+/+ mice. The idea of a MCP-1 "priming" mechanism driving an inflammatory profile comes from studies showing that when LPS is injected directly into the striatum of MCP-1-/- mice, there is a decrease in inflammatory cytokine production that is not correlated with activation or recruitment of microglia or brain immune cells . The authors propose that low levels of MCP-1 along with other inflammatory mediators in the brain result in "priming" of the glial cells to become more responsive to later inflammatory insults, and the total inhibition of MCP-1 in the knockout model renders the glia less responsive . We are currently investigating this possibility with glial cell cultures.
Third, MCP-1 may also be important in regulation of the BBB, and disruption of this regulation might lead to increased permeability of the BBB and more leukocyte infiltration from the periphery. Support for this idea comes from the finding that MCP-1 can increase the BBB permeability to FITC-albumin  and modulate the expression of tight junction proteins in endothelial cells of the BBB [20, 49]. It is possible that MCP-1 is involved in increasing BBB permeability after an LPS insult, thereby allowing infiltration of immune cells. The lack of MCP-1 in the knockout mice would then result in less brain inflammation because of fewer infiltrating immune cells. We have not directly tested this possibility, but have observed that 4 of 5 MCP-1+/+ mice injected with LPS had what appeared to be leukocyte infiltration into the brain parenchyma, whereas only 1 of 4 MCP-1-/- mice showed these cells. These data indicate that the lack of MCP-1 may confer a protective effect against alterations in BBB permeability after LPS treatment.