Background
Chronic inflammation is an inherent ongoing process in the progression of Alzheimer’s disease [
1‐
3]. However, the specific mechanisms by which sustained inflammatory reactivity contributes to the progressive neuronal degeneration underlying loss of cognition in Alzheimer’s disease (AD) brain are not well understood. Some evidence suggests limited benefits of non-steroidal anti-inflammatory drugs (NSAIDS) [
4] with the relatively small extent of drug efficacy attributed to previous deterioration in cognitive function in AD individuals prior to medication. Another possibility is that inflammatory reactivity in AD brain is manifest from activation of multiple pathways other than cyclooxygenase-dependent activity targeted by NSAIDS.
A critical component of inflammatory response is a chemokine-mediated mobilization of microglia in response to peptide deposition [
3,
5‐
7]. A spectrum of chemokines contributes to inflammatory responses in disease [
8,
9], with some evidence suggesting a prominent role for interleukin-8 (IL-8) in AD pathology. Gene microarray analysis has shown that IL-8 exhibits the largest increase in expression of any inflammatory factor in human microglia incubated with amyloid-beta (Aβ
1–42) [
10]. This same group also reported dose-dependent increases in production of IL-8 in human microglia stimulated with peptide [
11]. Elevated cerebrospinal fluid (csf) levels of IL-8 have been documented in AD brain relative to controls [
12]. Interestingly, IL-8 has been reported to potentiate Aβ
1–42-induced expression and production of a number of pro-inflammatory cytokines in cultured human microglia [
13].
Immunostaining for the IL-8 receptor CXCR2 has demonstrated receptor association with neuritic plaques in AD tissue [
5,
14]. However, CXCR2 also ransduces IL-8-dependent cellular inflammatory chemokine responses in the periphery and brain. In the former case, the receptor is expressed by infiltrating neutrophils in chronic obstructive pulmonary disease (COPD) with inhibition of CXCR2-mediated inflammatory responses effective in attenuating lung damage [
15]. Prominent CXCR2 activity in activated microglia has been reported in damaged brain with antagonism of receptor effective in reducing inflammation and promoting recovery in lesioned spinal cord [
16], following traumatic brain insult [
17] and in animal tumor models [
18].
At present, pharmacological modulation of CXCR2 has not been examined in animal models of AD. We posited that given the high levels of IL-8 in AD brain that pharmacological inhibition of CXCR2 could serve as a novel strategy to protect neurons exposed to inflammatory microenvironments. To examine this hypothesis, we have used the compound SB332235, a selective inhibitor of CXCR2 in macrophage cells [
15,
19], as a receptor antagonist to attenuate microglial inflammatory reactivity induced by Aβ
1–42 intrahippocampal injection. Specifically, SB332235 has been examined in vivo as a modulator of CXCR2 cell-specific association, gliosis, microglial chemotactic response, oxidative stress factors, and neuronal viability.
Discussion
This study presents novel findings for enhanced expression of the chemokine IL-8 receptor CXCR2 in human AD brain and in ML region of dentate gyrus in Aβ1–42-injected rat hippocampus. Evidence is presented in the AD animal model indicating upregulation of CXCR2 may be linked with microglial-mediated responses which in turn are correlated with neuronal damage in inflamed brain. In essence, deposition of Aβ1–42 induces a microglial chemotactic response involving upregulation of CXCR2 and its ligand, IL-8. A net migration of microglia is manifest in clustering of cells in the vicinity of peptide leading to cell activation and subsequent production of an assemblage of pro-inflammatory mediators. Our findings suggest microglial-derived oxidative species and lipid peroxidation could contribute to oxidative stress damage to GCL neurons with pharmacological inhibition of CXCR2 efficacious in blocking inflammatory reactivity and attenuating neuronal damage.
The demonstration of upregulated CXCR2 in AD vs ND cortical brain tissue served as a rationale for the design of animal model experiments. Importantly, cortical brain tissue from AD individuals demonstrated areas of CXCR2 co-localization with activated microglia. Similar results were obtained in hippocampal brain sections in the few cases where tissue was available. The cell-specific association of CXCR2 supports the possibility that microglial-mediated inflammatory responses may be involved in AD pathology. Involvement of CXCR2 activation in inflamed brain is consistent with the finding that the receptor ligand, IL-8, is reported as the most highly upregulated factor from Aβ
1–42-stimulated human microglia [
10]. Interestingly, IL-8 priming of human microglia subsequently exposed to Aβ
1–42 has been found to enhance cellular production of a host of inflammatory factors including pro-inflammatory cytokines [
13].
The intrahippocampal injection of Aβ
1–42 is an AD animal model characterized by enhanced inflammatory reactivity with pharmacological block of microgliosis correlated with increased viability of GCL neurons [
25,
26,
30,
38]. In the present work, expression of CXCR2 and IL-8 showed similar time-dependent (1–7 days) increases following Aβ
1–42, relative to controls (PBS and reverse peptide), intrahippocampal injection. Both receptor and ligand expressions were maximal at 3 days post-peptide injection and remained elevated at 7 days post-injection. Immunohistochemical staining exhibited similar results with Aβ
1–42 injection yielding a fivefold increase in CXCR2 expression with Aβ
1–42, relative to PBS, injection (time point of 3 days post-injection). Results from Western blot assay showed consistent trends in CXCR2 expression with duration of peptide injection with CXCR2 levels maximum at 3 days post-peptide injection.
At 3 days post-peptide injection, considerable extents of CXCR2 immunoreactivity were co-localized with microglia with lesser association of receptor with astrocytes (Fig.
3). Pharmacological antagonism of CXCR2 by SB332235 was examined with an initial focus on drug effects on gliosis at 3 days subsequent to intrahippocampal injection of Aβ
1–42. A marked enhancement for both microgliosis and astrogliosis was evident in ML region of dentate gyrus compared with PBS or reverse peptide application (Fig.
4). Treatment of peptide-injected animals with SB332235 significantly inhibited microgliosis but was ineffective in attenuating astrogliosis. It can be noted that contributions from CXCR2-(+)ve neurons would be minimized in sections isolated from the ML region. In addition, the absence of myeloperoxidase (MPO) immunoreactivity (data not shown) indicated that CXCR2-mediated neutrophils did not contribute to inflammatory responses in Aβ
1–42-injected rat brain.
Microglial chemotaxis is a rapid inflammatory response to Aβ deposition in the AD model [
7,
26]. In this work, we measured net migration of microglia in a single quadrant in the immediate vicinity of peptide deposits in ML. Double staining was then used to determine CXCR2 and glial ir within 300 μm of Aβ
1–42 deposits (Fig.
5). Animal treatment with SB332235 was examined for localized effects on CXCR2 area density and microgliosis and astrogliosis. Both CXCR2 area density and microglial ir were significantly attenuated by SB332235 administration with no effects of the compound on astrocytic responses. Although this component of study does not directly target chemotactic processes, the results suggest efficacy for SB332235 in inhibiting microglial responses and CXCR2-dependent activity nearby peptide.
Peptide-injected (3 days) rat brain exhibited a considerable loss of GCL neurons compared to PBS or reverse peptide injection (Fig.
6a, b). Treatment of Aβ
1–42-injected rats with SB332235 conferred a significant degree of neuroprotection as shown by NeuN staining in the GCL region of dentate gyrus. Previous work using this animal model has demonstrated that drug actions which inhibit microgliosis are correlated with enhancement in numbers of GCL neurons [
6,
26]. We also examined if lipid peroxidation could contribute to neurotoxicity by assessing 4-HNE ir in the GCL region. Overall, levels of 4-HNE were markedly elevated with Aβ
1–42, and absent with PBS or Aβ
42–1, intrahippocampal injection (all results obtained at 3 days post-injection). Animal treatment with SB332235 markedly inhibited 4-HNE levels in peptide-injected brain.
Previous work has demonstrated peptide-stimulated microglia as a prominent source of superoxide radical [
35,
36]. Oxidative stress induced by superoxide species could be involved in the lipid peroxidation damage to neurons [
34]. To examine this possibility, HEt ir was determined in the ML region of dentate gyrus. This region was chosen to correspond to the areas of microglial and astrocytic responses. Superoxide was not detectable in controls (PBS or Aβ
42–1) at 3 days post-injection; however, considerable HEt ir was evident in Aβ
1–42-injected brain. Treatment of peptide-injected animals with SB332235 was effective in attenuating levels of the superoxide marker. The neuroprotection conferred by SB332235 is consistent with previous results showing inhibition of microgliosis as a mechanism enhancing neuronal viability in the peptide-injected animal model but does not rule out possible direct effects of the CXCR2 antagonist on GCL neurons.
As noted above, direct intrahippocampal injection of Aβ
1–42 serves as an AD animal model which exacerbates inflammatory reactivity. The model appears to be characterized as one in which an acute insult evolves into a chronic inflammatory perturbation in a relatively short time. The injection of peptide has particular utility in correlating effects of pharmacological modulation of microgliosis with viability of neurons. Validation of the model has been considered in terms of a comparison of cellular responses and processes with properties characteristic of AD brain tissue [
39]. This comparison has shown similarities in a number of features including microglial and astroglial responses, abnormalities in microvasculature, and leakiness in BBB. Neuronal loss apparent in the AD model is the correlate of cognitive dysfunction in AD brain.
Our in vivo results provide evidence for efficacy of SB332235 at a time point associated with maximal expression for both CXCR2 and IL-8. However, the RT-PCR data suggest that both receptor and ligand expressions may remain elevated over longer times. In AD brain, expression of CXCR2 (Fig.
1) and IL-8 [
12] is increased compared to levels in controls. In this case, the CXCR2 antagonist may have utility in reducing chronic inflammatory activity over extended times.
It is important to note that beneficial effects of microglial response and activation have been reported in AD brain [
40‐
42]. Indeed, previous work on chemokine receptors in Tg 2576 mice has demonstrated that attenuation of Ccr2 in microglia was associated with abnormal accumulation of Aβ and increased mortality of animals [
43]. Conversely, knockout of chemokine receptor Cx3cr1 was found to confer neuroprotection in a mouse model of AD [
44]. Overall, a manifold of microglial-mediated inflammatory pathways is active in peptide-stimulated brain [
1,
45] with diverging negative or positive effects on the viability of the neurovascular unit [
46].
Our findings suggest the relevance in using transgenic animal models to examine pharmacological inhibition of CXCR2 as a strategy to enhance cognitive function. Such studies would reflect the effects of a progressive buildup of peptide deposits over time, rather than direct injection of amyloid, to more closely mimic chronic inflammation in AD brain. It should be emphasized that a number of chemokines, their receptors, and a host of non-chemokine factors could contribute to inflammatory reactivity in the progression of AD pathology. We suggest the merits in using a cocktail delivery of drugs as a strategy to examine effects for modulation of multiple components of chronic inflammation in treatment of the disease.
Competing interest
The authors declare that they have no competing interests.
Authors’ contributions
JKR designed and conducted the research experiments and analyzed the immunofluorescence data. TC and HBC conducted the RT-PCR and Western blot studies and analyzed the data. NJ analyzed and interpreted the immunofluorescence staining results from AD and ND brain tissue. JGM designed the overall research program, analyzed and interpreted the data, and drafted the manuscript. All authors read and approved the final manuscript.