Cerebral inflammation involving the release of soluble mediators, infiltration of peripheral immune cells and activation of resident glial cells, is one of the key pathophysiological processes contributing to secondary degeneration following focal traumatic brain injury (TBI). The chemokine CCL2 (also known as macrophage chemoattractant protein-1, or MCP-1) is well recognised for its potent ability to mediate macrophage recruitment and migration to sites of inflammation [¹,²]. In the brain, CCL2 production is rapidly induced by a range of diverse inflammatory conditions, and is associated with the infiltration of blood-derived macrophages and activation of microglia [³,⁴]. Transgenic mice over-expressing Ccl2 in the central nervous system (CNS) exhibit a robust accumulation of macrophages in the brain [⁵,⁶], whereas mice deficient in the Ccl2 gene show reduced leukocyte infiltration after TBI, spinal cord injury and stroke [^7-9]. Detection of elevated CCL2 in the serum and cerebrospinal fluid of severe TBI patients [⁹,¹⁰] corroborates a central role for this chemokine in post-traumatic neuroinflammation.

We recently identified an altered profile of cortical cytokine production in Ccl2-deficient mice subjected to a closed head injury model of focal TBI. Peak levels of interleukin (IL)-1α, IL-1β, IL-6, granulocyte-colony stimulating factor (G-CSF), IL-12(p40), CCL3 and CXCL1 were delayed and significantly exacerbated in Ccl2-/- mice compared to wild-type mice acutely post-injury, whilst the production of other inflammatory mediators including CCL5, interferon-gamma (IFNγ) and IL-2 had reduced production in chemokine-deficient animals [⁹]. These findings appear paradoxical to the delayed neuroprotection shown in Ccl2-/- mice, which had reduced macrophage accumulation and tissue damage associated with improved functional recovery over 4 weeks post-injury. We hypothesise that the altered cytokine network in the brains of Ccl2-/- mice after injury may indicate a previously unrecognised role for CCL2 as a modulator of acute CNS inflammation. Based on astrocytes being the main source of chemokines including CCL2 [^11-14], it is conceivable that CCL2 may exert immunomodulatory effects on this cell type.

In the current study, we aimed to elucidate whether CCL2 regulates immune processes by investigating cytokine production from Ccl2-/- astrocytes compared to wild-type cells in response to inflammatory stimuli in vitro. We demonstrate that primary Ccl2-/- astrocyte cultures secrete exacerbated levels of chemokines and cytokines such as interleukin (IL)-6 compared to wild-type cells in response to lipopolysaccharide (LPS) and IL-1β stimulation. Furthermore, production of IL-6 induced by IL-1β in wild-type astrocytes was suppressed by prior incubation with exogenous CCL2. These results indicate a likely immunomodulatory role for CCL2 in astrocytic cytokine production. In combination with other recently identified functions of this chemokine in neurotransmission [¹⁵,¹⁶] and neuronal cell survival [^17-19], these findings indicate that future application of therapeutics targeting CCL2-mediated leukocyte infiltration following TBI may have tangential effects in the injured brain.

Investigations into inflammatory neuropathologies such as multiple sclerosis and stroke have demonstrated neuroprotection resulting either from interference with the Ccl2 gene, by administration of neutralising anti-CCL2 antibodies, or targeting the CCR2 receptor [⁷,^42-44]. Besides the chemotactic ability of CCL2 to drive the migration of monocytic cells into the brain, recent evidence has shown that CCL2 also has alternative neuroactive properties in the CNS [^17-19,⁴⁵]. Furthermore, research by our group and others have reported alterations in the inflammatory cytokine response and tissue damage in Ccl2-deficient mice after TBI or stroke, prior to the onset of leukocyte infiltration [⁷,⁹]. The purpose of our current study was thus to investigate whether CCL2 can influence astrocyte-specific cytokine secretion following an immune challenge in vitro.

The production of the cytokine IL-6 and chemokines CXCL1 and CCL3 was significantly exacerbated in the supernatant of Ccl2-/- astrocytes following stimulation with LPS, corresponding to the delayed and enhanced secretion detected in the injured mouse brain [⁹]. TNF, whose production was elevated in the brains of Ccl2-/- mice across a time course of 2-24 h after experimental TBI, was also elevated in LPS-stimulated Ccl2-/- astrocytes in vitro. Differences in cytokine protein levels in wild-type and Ccl2-/- astrocyte cultures are likely due to variations in gene expression, as we also detected an enhanced elevation of IL-6 mRNA levels in Ccl2-/- astrocytes compared to wild-type cells.

Interestingly, the macrophage chemoattractant CCL5 was produced at lower levels in the brains of Ccl2-/- mice compared to wild-type after TBI, however, gene-deficient astrocytes had a profile of exacerbated CCL5 production. Particularly considering the low concentration of CCL5 detected in culture supernatant following LPS stimulation compared to the other cytokines measured, it is possible that this chemokine may be primarily produced during neuroinflammation by other CNS cells such as microglia, as has been previously suggested [³⁵,⁴⁶,⁴⁷].

For subsequent experiments we focused specifically on astrocyte production of IL-6, as this multi-functional cytokine is central to CNS inflammation and degeneration following TBI [⁴⁰,⁴¹], and is secreted in vitro in response to mechanical injury and exposure to inflammatory cytokines [²²,²⁶,²⁹,³¹,⁴⁸,⁴⁹]. Consistent with the LPS-stimulated phenotype, a similarly enhanced response by Ccl2-/- astrocytes was evident following stimulation with the pro-inflammatory cytokine IL-1β at both 10 and 50 ng/ml. Interestingly, this characteristic exacerbated response was not evident in Ccl2-/- astrocytes at a lower concentration of IL-1β (1 ng/ml), when wild-type cells produced a higher level of IL-6. This finding may reflect what we observed in vivo, where secretion of IL-6 was similar in wild-type and Ccl2-/- mice early after injury (2-4 h) before a delayed and enhanced response in Ccl2-/- mice occurred at 12 h, corresponding to the time of maximal IL-1β in the lesioned cortex [⁹].

There have been several other studies suggesting that Ccl2-deficiency deregulates inflammatory cytokine synthesis in the CNS, primarily restricted to in vivo rodent models. Contrary to our findings, Hughes and colleagues (2006) reported that IL-6 and IL-1β were significantly reduced in the brains of Ccl2-/- mice after induction of ischaemic stroke [⁷]. Similarly, lower levels of IL-1β and TNF were induced in Ccl2-/- mice compared to wild-type after a striatal injection of LPS [⁵⁰]. A third study also using LPS as the stimulus found that peripheral administration resulted in elevated IL-1β and TNF levels in serum of Ccl2-/- mice, whereas cytokine production in the brain was attenuated [⁵¹]. It is worth noting that all three of these studies considered only one time point (6 h) after stroke or LPS stimulation. Since we have shown here that alterations in cytokine production resulting from Ccl2-deficiency are time and concentration-dependent both in vivo and in vitro, it is possible that these studies do not comprehensively reflect the overall neuroinflammatory response specific to this strain. Importantly, our current work is the first to examine Ccl2-deficient anomalies in cytokine production or synthesis in depth by using an in vitro system of astrocyte activation.

Previous studies have suggested that the early and robust upregulation of CCL2 in neuropathologies such as TBI may indicate a central role for this chemokine in initiating the cytokine response [⁷]. Given that cytokine production in the brain coincides with CCL2 upregulation and most often precedes evidence of microglial activation or macrophage infiltration from the periphery, it is unlikely that the influence of CCL2 on acute cytokine production is dependent on the chemoattraction of leukocytes and subsequent secretion from these cells [⁵¹]. Furthermore, investigation of the macrophage and microglial marker F4/80 in the brains of Ccl2-/- mice early after injury or disease onset has failed to detect any difference from wild-type mice, suggesting that the contribution of CCL2 to acute neuropathology may be independent to its classic actions as a chemokine [⁷,⁵⁰,⁵²]. Given that incubation of cultured astrocytes with rCCL2 alone does not induce production of IL-6, TNF, CXCL1, CCL3 or CCL5 above levels in unstimulated cells, it is also unlikely that CCL2 directly stimulates cytokine production from astrocytes.

Alternatively, based on our findings of exacerbated cytokine production in the brains of injured Ccl2-deficient mice and immune-challenged Ccl2-deficient astrocytes, we hypothesised that CCL2 may act during early immunoactivation in an inhibitory manner. Interestingly, CCL2 has recently been proposed to protect neurons from cell death induced by N-methyl-D-aspartate, glutamate, β-amyloid and human immunodeficiency virus-mediated Tat toxicity [^17-19]. Further prompting our own investigations into alternative properties of CCL2, Rankine and colleagues (2006) previously showed that co-administration of LPS and rCCL2 into the mouse brain did not have a synergistic pro-inflammatory effect as they had hypothesised - instead, the addition of rCCL2 in combination with LPS resulted in a non-significant trend towards reduced IL-6 production.

Also relevant to our findings, Zisman and colleagues showed that treatment with anti-CCL2 antibodies prior to an intraperitoneal LPS challenge increased plasma TNF and IL-12 levels, whilst administration of rCCL2 reduced pro-inflammatory cytokine production by 20-35% [⁵³]. Similarly, in cultured astrocytes we have demonstrated that pre-exposure to rCCL2 was sufficient to reduce the IL-1β-stimulated production of IL-6 in culture supernatant. Furthermore, this inhibitory effect was also evident at a transcriptional level, as the increase in IL-6 mRNA expression induced by IL-1β stimulation was significantly lower in rCCL2-treated cultures. To our knowledge, this is the first in vitro evidence that CCL2 can act on astrocytes in an immunomodulatory manner to down-regulate cytokine expression and production.

One limitation of this study is the relative simplicity of the culture system used. It must be noted that cultured primary astrocytes may respond quite differently in isolation compared to in vivo, where they may be influenced by interactions with adjacent endothelial cells, microglia and extracellular matrix components. However, the consistency between our findings in cultured astrocytes and the mouse cortex in vivo [⁹] supports the use of primary astrocyte cultures as an appropriate model system.

The mechanisms underlying a novel role for CCL2 in CNS cytokine production need to be further elucidated. Whilst CCL2 mediates immune cell chemotaxis by binding to the CCR2 receptor, the presence of CCR2 on astrocytes in the brain remains controversial [¹⁴,^54-57]. Quinones and colleagues recently found that CCL2 treatment in vitro can induce glial cell migration and inhibit the spontaneous apoptosis of mouse astrocytes in a dose-dependent manner. Whilst the latter pro-survival effect was shown to be dependent on signaling via CCR2, residual chemotaxis of Ccr2-/- astrocytes in response to CCL2 may indicate the presence of an alternative receptor for this chemokine on primary astrocytes [⁵⁸]. Whether CCR2 is the primary receptor involved in CCL2-modulation of cytokine production, and which intracellular pathways are responsible for the reduction in cytokine expression observed in IL-1β-stimulated astrocytes pre-exposed to rCCL2, remain unanswered. The use of specific neutralising antibodies, or the recently generated cell-conditional CCL2 knockout mice in which gene deletion is specific to either astrocytes or endothelial cells, may prove essential in answering these questions and eliminating potential confound resulting from global gene deficiency [⁵⁹].

Lastly, the consequences of an altered cytokine environment in the presence or absence of CCL2 remain unclear. Of note, Ccl2-/- mice show long-term improvements in neurological recovery and reduced tissue damage following experimental TBI, despite the aberrations in early cytokine secretion [⁹]. Given the acute nature of cytokine production in the injured brain, it is feasible that the aberrant, exacerbated cytokine profile seen in Ccl2-/- mice may contribute to early post-traumatic degeneration in this strain, which may compensate for the lack of CCL2 and contribute to early leukocyte infiltration. We hypothesise that CCL2 may have differential acute and long-term effects during neuropathology, as has been indicated for TNF [⁶⁰,⁶¹]. A deeper understanding of the consequences of CCL2 deletion in CNS inflammation and cytokine production may allow us to better tailor therapeutic approaches to the appropriate time post-injury. Thus further studies are necessary to elucidate the biological and physiological relevance of these findings.

In the current study we have employed an in vitro model of astrocyte activation to characterise the Ccl2-/- genotype-specific profile of cytokine and chemokine secretion as observed in the injured brains of Ccl2-/- mice. Our findings provide evidence that astrocytes are primarily responsible for exacerbated cytokine response in the absence of CCL2. Furthermore, we present data showing that rCCL2 can inhibit the astrocytic production of inflammatory cytokines in response to IL-1β stimulation. In light of these findings, we propose that CCL2 plays a novel, immunomodulatory role in the inflamed CNS by altering cytokine production in astrocytes, and that this effect is relative to the concentration of the stimulus present. This study highlights the complexity of chemokine - cytokine interactions in the CNS, and suggests that CCL2 may affect multiple cytokine pathways which could participate in neurodegenerative processes. Our data indicate that chemoattraction of immune cells is only one of the functions performed by CCL2 in the injured CNS, and that further investigation is required before targeting of this chemokine as a means of neuroprotection for neuropathologies in which CCL2 upregulation is a characteristic feature.

The authors declare that they have no competing interests.

BDS participated in the project design, performed the astrocyte culture experiments including astrocyte isolation, immunoassays, qPCR, data acquisition and analysis, and drafted and revised the manuscript. TF assisted with the design, coordination and interpretation of qPCR experiments and manuscript revisions. CMK conceived the study, assisted with its design and data analysis, and helped draft the manuscript. All authors read and approved the final manuscript.

Supported by the Victorian Neurotrauma Initiative, Australian Government (Australian Postgraduate Award for BD Semple), Transport Accident Commission and National Health and Medical Research Council.