Neuregulin-1 inhibits neuroinflammatory responses in a rat model of organophosphate-nerve agent-induced delayed neuronal injury

Neuregulin-1 inhibits DFP-induced microglial activation

DFP is structurally and toxicologically similar to the OP nerve agents and thus is used as an OP nerve agent stimulant in experimental animal models [29,30]. We previously demonstrated that rats injected with DFP at 9 mg/kg, i.p., experience seizures and exhibited significant delayed neurodegeneration in multiple brain regions [9]. Microglial activation is a characteristic brain inflammatory response induced following OP nerve agent intoxication [20,21]. Under normal physiological conditions, resting microglia display a ramified state; however, when activated, microglia undergo a morphological transformation from the resting ramified state to an amoeboid shape. To determine the effects of acute DFP intoxication on microglia, brain sections from rats injected with vehicle or DFP in the absence or presence of NRG-1 were immunostained for CD11b, a biomarker of microglia [33]. Microglia in the superficial layers of cortex (Figure 1A) and lateral dorsal thalamus (Figure 1B) in control animals displayed the characteristic ramified morphology of resting microglia. Acute intoxication with DFP caused microglial activation, as indicated by the increased size of the cell body, a thickening of proximal processes, decreased ramification of distal branches and/or amoeboid shaped cell bodies of CD11b immunopositive cells (Figure 1C, E). NRG-1 treatment prevented the DFP-induced morphological changes of microglial cells in those brain regions (Figure 1D, F) as CD11b immunopositive cells were morphologically similar to microglia in control brains. Dual labeling with CD11b and FJB showed that in the thalamus of animals acutely intoxicated with DFP, activated microglia were detected in areas of brain injury (Figure 1G). However, neither activated microglia nor injured neurons were present in the thalamus of DFP intoxicated animals treated with NRG-1 (Figure 1H). We previously showed that DFP administration resulted in neuronal injury in the CA1, CA3, and dentate gyrus of the hippocampus which was attenuated by NRG-1 [9]. Similarly, microglial activation was also seen in the hippocampus, including the dentate gyrus shown here, following DFP intoxication (Figure 2A, C). The microglial response to DFP was attenuated by NRG-1 treatment (Figure 2B, D). NRG-1 treatment did not prevent microglial activation in the amygdala, medial dorsal thalamus, reunion area of thalamus or piriform cortex of the DFP intoxicated animal (data not shown), brain regions in which neuronal injury was previously shown to not be protected by NRG-1 [9].

To confirm the data based on morphological criteria of resting versus activated microglia, brain sections were also immunostained using the ED-1 antibody, which is a specific biomarker of macrophages and activated microglia [33]. There was virtually no ED1 immunoreactivity in the brains of vehicle control animals, including dorsal lateral thalamus (Figure 3A). However, there were numerous ED-1 immunopositive macrophages/activated microglia detected in brain regions of DFP-intoxicated animals. ED-1 immunopositive cells were amoeboid shaped (characteristic of activated macrophages and microglia) and distributed throughout the dorsal lateral thalamus (Figure 3B) and the superficial layers of cortex (Figure 3D). However, in DFP-intoxicated animals pre-treated with NRG-1, only a few ED-1 immunopositive cells were seen in injured regions of the dorsal lateral thalamus (Figure 3C) and no ED-1 immunopositive cells were observed in the cortex (Figure 3E). Dual labeling for ED-1 and FJB showed that activated microglia were associated with areas of neuronal injury in the cortex (Figure 3F) of DFP-intoxicated animals. Neither activated microglia nor injured neurons were observed in the cortex of DFP-intoxicated animals pre-treated with NRG-1 (Figure 3G). Similar to observations of CD11b immunolabeling, there was no change in the density of ED1 immunopositive cells in the amygdala, reunion area of thalamus, medial dorsal thalamus, and piriform cortex (data not shown), which were previously shown to be non-responsive to the neuroprotective effects of NRG-1 [9]. DFP induced ED-1 immunoreactivity in the hippocampus (Figure 4A), which was significantly reduced by NRG-1 pre-treatment (Figure 4B). Labeling of adjacent hippocampal sections of the dentate gyrus with ED-1 and FJB indicated that ED-1 immunoreactivity was detected both within and outside the hilar region where there were significant numbers of FJB stained neurons (Figure 4C) but in the hippocampi of DFP-intoxicated rats treated with NRG-1, there was no FJB labeling (Figure 4D).

NRG-1 prevents DFP-induced hippocampal gene expression

To examine gene expression profiles, total RNA was isolated from the hippocampus of each experimental group (control, DFP, and DFP + NRG-1) at 24 h post-DFP injection and interrogated using the rat genome U230 2.0 GeneChip that contains approximately 30,000 probe sets (transcripts). To assess inter-experiment differences in the intensity of gene expression, replicated experiments were compared for each experimental group and analyzed by principle component analysis (PCA). Similarity between data points in the PCA was determined by the distance between the points in the three-dimensional plot generated, with shorter distance indicating increased similarity (Figure 5A). Each data point represents one animal’s gene expression profile. Control, DFP, and DFP + NRG-1 gene expression profiles clustered within treatment groups, but the control group was well isolated from the other two treatment groups and the profiles for DFP and DFP + NRG-1 treatment groups were similar. To examine the changes in gene expression, we compared the gene expression profiles of the vehicle controls to the DFP animals and the gene expression profiles of the DFP animals to the DFP + NRG-1 animals. Figure 5B is a heat map illustrating that many genes were induced following DFP administration and that a very small subset of those genes were attenuated by NRG-1 treatment.

The set of genes induced by two-fold or more at 24 h following DFP administration (1,298 transcripts) and that were suppressed 50% or more by treatment with NRG-1 (41 transcripts) was imported into the IPA software to examine functions and pathways associated with this subset of genes. Analysis of the top ten canonical pathways associated with the dataset in IPA shows that most of these pathways are associated with immune function (Figure 6). Immune cell trafficking (agranulocytes and granulocyte adhesion and diapedesis), cytokine (IL-17 and IL-6) signaling, and acute phase response signaling were ranked among the most significant pathways, consistent with a key role for inflammation in mediating the delayed neural injury associated with OP neurotoxicity. The full list of genes that passed the fold change and false discovery rate cutoffs, including names and the GenBank accession numbers, is provided in Table 1. The DFP-induced genes that were significantly attenuated by NRG-1 treatment included the cytokine IL-1β, chemokines CXCL2 and CXCL11, and leukocyte cell derived chemotaxin-1 (LECT1). DFP increased the expression of paraoxonase-1 (PON1), an enzyme involved in the hydrolysis of organophosphates and NRG-1 suppressed the induction of PON-1 by DFP.

Transcript cluster ID	Gene symbol	Description	FC OP vs. control	FC OP/NRG-1 vs. OP
1367794_at	A2m	Alpha-2-macrolobulin	2.34	−2
1369625_at	Aqp1	Aquaporin 1	5.1	−17.93
1383946_at	Cldn 1	Claudin 1	3.94	−5.7
1375933_at	Cldn 2	Claudin 2	2.44	−25.97
1393891_at	Col8a 1	Collagen, type VIII, alpha 1	2.73	−5.42
1374172_at	Col8a 2	Collagen, type VIII, alpha 2	3.37	−19.49
1370391_at	Crabp 2	Cellular retinoic acid binding protien 2	2.18	−2.04
1379365_at	Cxcl 11	Chemokine (C-X-C motif) ligand 11	17.41	−2.13
1368760_at	Cxcl 2	Chemokine (C-X-C motif) ligand 2	2.67	−2.61
1367700_at	Fmod	Fibromudolin	2.02	−8.91
1387889_at	Folr 1	Folate receptor 1 (adult)	3.9	−12.25
1368337_at	Glycam	Glycosylation-dependent cell adhesion molecule 1	3.81	−7.01
1393840_at	Gnat 2	Guanine nucleotide protein (G protein), alpha transducing activity polypeptide 2	2.68	−2.52
1391915_at	Hspa 9	Heat shock protein 9	2.02	−2.13
1398256_at	II 1b	Interleukin 1 beta	3.08	−2.13
1387050_s_at	Kng 1	Kininogen 1; kininogen 1-like 1; kininogen 2	3.46	−3.31
1387868_at	Lbp	Lipopolysaccharide binding protien	3.31	−2.62
1387164_at	Lect 1	Leukocyte cell derived chemotaxin 1	4.8	−5.56
1393210_at	LOC100361383	Extracellular matrix protein 2-like	2.11	−2.09
1375465_at	Otx 2	Homeobox protein OTX2-like; orthodenticle homeobox 2	2.51	−17.47
1375908_at	Mpzl 2	Myelin protein zero-like 2	4.06	−4.16
1368302_at	Msx1	Msh homeobox 1	2.04	−5.63
1371849_at	Nt5dc2	5′-nucleotidase domain 2	2.13	−4.9
1385248_a_at	Ogn	Osteoglycin	2.38	−2.16
1375367_at	Pdlim	PDZ and LIM domain 2	2.67	−3.73
1370068_at	Pla2g5	Phospholipase A2, group V	3.01	−7.5
1371050_at	Pon1	Paraoxonase 1	2.48	−3.99
1370012_at	Ptgis	Prosta landin l2 (prostacyclin) synthase	2.75	−2.74
1380334_at	Rbm47	RNA binding motif protein 47	2.11	−3.46
1385500_at	RGD1561795	Similar to RIKEN cDNA 1700012B09	2.96	−5.52
1387125_at	S100a9	S100 calcium binding protein A9	6.6	−2.88
1377034_at	Serpin1a	Serine (or cysteine) proteinase inhibitor, clade B, member 1a	3.07	−2.2
1385005_at	Slc22a8	Solute carrier family 22 (organic anion transporter), member 8	2.12	−3.45
1368606_at	Slco1a5	Solute carrier organic transporter family, member 1a5	4.34	−48.37
1367998_at	Slpi	Secretory leukocyte peptidaase inhibitor	5.12	−2.32
1390525_a_at	Stra6	Stimulated by retinoic acid gene 6	2.64	−4.31
1384522_at	Sytl3	Synaptotagmin-like 3	2	−2.06
1383606_at	Tc2n	Tandem C2 domains, nuclear	4.01	−3.59
1387013_at	Tmem27	Transmembrane protein 27	4.53	−34.51
1388557_at	Tubb4b	Tubulin, beta 4B class IVb	2.24	−4.21

NRG-1 reduces DFP-induced mRNA levels of pro-inflammatory cytokines in the brain

To confirm the microarray results indicating an effect of NRG-1 on pro-inflammatory cytokine mRNA expression, we used real-time RT-PCR to determine the levels of IL-1β, IL-6, and TNFα in hippocampal tissues 24 h following DFP and NRG-1 administration. Microarray results showed that IL-1β and IL-6, but not TNFα, were upregulated by DFP. Pretreatment with NRG-1 reduced levels of DFP-induced IL-1β mRNA by more than 50% of the levels observed in DFP intoxicated animals. DFP-induced levels of IL-6 mRNA were also reduced by NRG-1 but by less than the 50% cutoff, and NRG-1 treatment did not affect levels of DFP-induced TNFα mRNA. Similarly, real-time RT-PCR results showed that IL-1β and IL-6 were strongly induced by DFP in the brain while relatively little was seen in control brain tissues (Figure 7A, B). Consistent with the microarray results, DFP-induced IL-1β and IL-6 mRNA levels were reduced by NRG-1. TNFα levels were unchanged by either DFP or NRG-1 at 24 h post-DFP injection (Figure 7C).

Although there has been progress with efforts to treat patients with acute symptoms following OP exposure, strategies to prevent the subsequent OP-induced delayed neuronal injury are currently unavailable. It is well established that OP intoxication results in the stimulation of pro-inflammatory responses that lead to neuronal injury and increased neurological impairment [2,15-19]. OPs initiate inflammatory responses in the injured brain that progress for days after the onset of symptoms. Experimental studies show that inflammatory reactions in the brain can enhance neuronal excitability and impair cell survival, and conversely, some anti-inflammatory treatments reduce brain pathology in animal models of acute CNS injury [2,34]. Therefore, interventions that are aimed at decreasing neuroinflammation have potential as therapeutic agents for treating humans poisoned with OP neurotoxins and nerve agents.

Recent studies from our lab showed that NRG-1 treatment significantly reduced DFP-induced delayed neuronal damage and oxidative stress in rats when administered up to 1 h following DFP intoxication [9,30]. One of the early consequences of OP intoxication is the activation of brain microglia [2,20,21,35], therefore, in this study, we examined whether NRG-1 could regulate DFP-induced microglial activation in the brain. Our data demonstrated that NRG-1 blocked DFP-induced morphological changes in microglia associated with activation and subsequent delayed neuronal damage. The region specific anti-inflammatory effects of NRG-1 are consistent with our findings showing selective DFP-induced neuronal injury and neuroprotection by NRG-1 [9,30]. We proposed that the effects could be the result to several factors, including access of exogenously administered NRG-1 to various brain regions, selective vulnerability of specific neurons and/or and regional differences in NRG-1 receptor expression.

To elucidate the molecular mechanisms used by NRG-1 to block microglia activation and neuronal injury, we examined gene expression profiles in hippocampal brain tissues following DFP intoxication and NRG-1 treatment. DFP increased the expression of nearly 1,300 transcripts by twofold or more, however only 41 of those genes were suppressed 50% or more by NRG-1. Bioinformatic analysis of the NRG-1 suppressed genes with IPA software indicated that immune cell trafficking, hepatic fibrosis/hepatic stellate cell activation, acute phase response, and IL-6 signaling were among the top canonical pathways associated with the anti-inflammatory and neuroprotective effects of NRG-1. These findings are similar to studies that analyzed gene expression profiles of the prefrontal cortex from rats 24 h after soman exposure [36].

Activated brain microglia release pro-inflammatory cytokines following OP poisoning, which are toxic to neurons [2,22-27]. Inflammatory cytokines, such as IL-1α, IL-1β, IL-6, and TNFα are induced in the rodent brain following OP intoxication [22-27]. IL-1β is primarily expressed by activated brain microglia following OP exposure [25]. NRG-1 attenuated the induction of IL-1β in addition to the chemokines CXCL2, CXCL11, and LECT1. Interestingly, NRG-1 attenuated the DFP-induced upregulation of paraoxonase-1 (PON1), an enzyme involved in the hydrolysis of organophosphates. The serum concentration of PON-1 is known to be mediated by inflammatory responses [37]. NRG-1 increased brain protein levels of IL-10, an anti-inflammatory cytokine that represses the expression of IL-1, IL-6, and TNFα by activated macrophages [38]. Consistent with this finding, we recently showed that NRG-1 blocked pro-inflammatory responses, induced anti-inflammatory cytokines, and increased survival in a mouse model of cerebral malaria [39]. NRG-1 suppressed the levels of TNFα, IL-6, IL-1α, and CXCL10, while inducing the anti-inflammatory cytokines IL-5 and IL-13. NRG-1 was effective despite having no effect on parasite load during the treatment, suggesting that NRG-1 can alter the immune state to prevent cellular damage following acute CNS injuries.

NRG-1 has previously been shown to affect microglial and macrophage activation [40-46]. NRG-1 reduced free radical release from cultured mouse microglial cells [40], and studies from our laboratory showed that NRG-1 suppressed cyclooxygenase-2 (COX-2) expression in activated human monocyte/macrophage cell cultures [41]. NRG-1 also stimulated microglial proliferation and chemotaxis in a rat model of peripheral nerve injury (PNI) [42,43]. Microglial activation and NRG-1 receptor activation following PNI peaked around 3 days after injury suggesting that microglial activation by NRG-1 may be associated with neuronal repair rather than acute neurotoxicity. The neuropathological and neuroinflammatory sequelae of acute OP poisoning are similar to those observed in other acute CNS injuries, such as stroke, brain trauma, and status epilepticus [15-19,47,48]. Studies from our laboratory and others demonstrated that administration of NRG-1 reduced delayed ischemic brain damage and improved functional recovery in a rat middle cerebral artery occlusion (MCAO) stroke model [32,49-52]. NRG-1 prevented macrophage/microglial activation, reactive astrogliosis, neuronal apoptosis, and pro-inflammatory cytokine expression following stroke [41,50,52]. Taken together, these studies suggest that the neuroprotective efficacy of NRG-1 in DFP-induced brain injury, ischemic stroke, and cerebral malaria might be explained, at least in part, by regulating the immune response and inflammatory mediators.

Recent clinical studies have demonstrated the utility of NRG-1 in human patients with congestive heart failure [53,54]. Phase I clinical studies in China (Chinese Clinical Trial: ChiCTR-TRC-00000414) and Australia (Australian New Zealand Clinical Trials Registry: ACTRN12607000330448) showed that NRG-1 was safe in both healthy and heart failure patients. Recombinant human NRG-1 was used in phase II clinical trials investigating its efficacy in patients with chronic heart failure in both the Australian and Chinese studies [53,54]. In these studies, patients received placebo or NRG-1 at a dose of 0.3 to 1.2 μg/kg/day intravenously for 10 days, in addition to standard drug therapies. During a follow-up period 11 to 90 days after study initiation, NRG-1 significantly improved heart function in patients and the effective doses were shown to be safe and tolerable. Three additional clinical trials to determine the ability of NRG-1 to improve cardiac function after heart failure have been initiated in the US (ClinicalTrails.gov identifiers NCT01258387; NCT01944683; NCT01251406). The doses of NRG-1 that showed efficacy for treating heart failure are near the doses used in our rat stroke and OP studies. In these studies, we showed that NRG-1 was neuroprotective after a single i.a. administration of NRG-1. However, we expect that i.v. administration of NRG-1 will similarly provide neuroprotection as previously demonstrated in other models of acute brain injury [51,55].

In conclusion, our data demonstrate that treatment with NRG-1 blocks the activation of microglia by DFP and decreases DFP-induced pro-inflammatory expression in brain tissues. Our results suggest that NRG-1 protects neurons against DFP-induced delayed cell death by inhibiting toxic pro-inflammatory responses. These findings indicate that NRG-1 has enormous clinical potential and could lead to the development of effective medical countermeasures to facilitate better emergency treatment and protection of civilians and military personnel following exposure to OP nerve agents.