Insulin-like growth factor-I gene delivery to astrocytes reduces their inflammatory response to lipopolysaccharide

Background Insulin-like growth factor-I (IGF-I) exerts neuroprotective actions in the central nervous system that are mediated at least in part by control of activation of astrocytes. In this study we have assessed the efficacy of exogenous IGF-I and IGF-I gene therapy in reducing the inflammatory response of astrocytes from cerebral cortex. Methods An adenoviral vector harboring the rat IGF-I gene and a control adenoviral vector harboring a hybrid gene encoding the herpes simplex virus type 1 thymidine kinase fused to Aequorea victoria enhanced green fluorescent protein were used in this study. Primary astrocytes from mice cerebral cortex were incubated for 24 h or 72 h with vehicle, IGF-I, the IGF-I adenoviral vector, or control vector; and exposed to bacterial lipopolysaccharide to induce an inflammatory response. IGF-I levels were measured by radioimmunoassay. Levels of interleukin 6, tumor necrosis factor-α, interleukin-1β and toll-like receptor 4 mRNA were assessed by quantitative real-time polymerase chain reaction. Levels of IGF-I receptor and IGF binding proteins 2 and 3 were assessed by western blotting. The subcellular distribution of nuclear factor κB (p65) was assessed by immunocytochemistry. Statistical significance was assessed by one way analysis of variance followed by the Bonferroni pot hoc test. Results IGF-I gene therapy increased IGF-I levels without affecting IGF-I receptors or IGF binding proteins. Exogenous IGF-I, and IGF-I gene therapy, decreased expression of toll-like receptor 4 and counteracted the lipopolysaccharide-induced inflammatory response of astrocytes. In addition, IGF-I gene therapy decreased lipopolysaccharide-induced translocation of nuclear factor κB (p65) to the cell nucleus. Conclusion These findings demonstrate efficacy of exogenous IGF-I and of IGF-I gene therapy in reducing the inflammatory response of astrocytes. IGF-I gene therapy may represent a new approach to reduce inflammatory reactions in glial cells.


Background
As a source of growth factors and of immunologically relevant cytokines and chemokines, astrocytes play a pivotal role in the pathophysiology of neurodegenerative diseases [1][2][3] and in the type and extent of central nervous system immune and inflammatory responses [4]. These responses may promote tissue repair and contribute to recovery of homeostasis under acute neurodegerative conditions. However, sustained inflammatory responses of astrocytes in chronic neurodegenerative diseases may enhance tissue damage through amplification of brain inflammation and consequent neuronal injury [4][5][6][7][8]. Therefore, to limit neuronal cell death under chronic neurodegenerative conditions, it is important to develop tools to control brain inflammatory reactions.
IGF-I is locally produced in the nervous system and it is also actively transported to the brain from plasma through the choroid plexus [9,10]. IGF-I has pleiotropic actions in nervous tissue, influencing neuronal development, synaptic plasticity, neuroendocrine regulation, adult neurogenesis and cognition [11][12][13][14]. IGF-I is also a potent neuroprotective molecule [10,13,15,16] exerting this function in part by reducing brain inflammation [17,18] and reactive astrocytosis [19]. In response to neurodegenerative conditions, astrocytes express IGF-I, probably as an endogenous neuroprotective and anti-inflammatory mechanism [20][21][22][23]. Consequently, the development of methodologies to increase IGF-I production by glial cells is a logical approach to implement anti-inflammatory IGF-I-based therapeutic strategies for neurodegenerative diseases.
We have recently constructed a recombinant adenovirus vector harboring the rat IGF-I gene [24,25]. Using this vector we have shown efficacy of IGF-I gene therapy to increase IGF-I levels in cerebrospinal fluid and to reduce neuronal damage in vivo [24,25]. In the present study we have explored whether exogenous IGF-I and IGF-I gene therapy regulate the inflammatory response of astrocytes.

Adenoviral vectors
A recombinant adenovirus (RAd) vector harboring the rat IGF-I gene (RAd-IGF-I) was constructed as previously described [24] using a variant of the two-plasmid method [26] The cDNA coding for the rat IGF-I gene (kindly donated by Dr. Peter Rotwein, Department of Biochemistry and Molecular Biology, Oregon Health & Science University, Portland, OR), obtained from the mRNA for the IGF-Ib precursor form [27], was placed under the control of the mCMV promoter in order to construct the genome of the desired RAd-IGF-I ( Figure 1). The newly generated RAd was rescued from human embryo kidney 293 (HEK293) cell lysates and plaque purified. It was further purified by ultracentrifugation in a CsCl gradient. Final virus stocks were titrated by a serial dilution plaque assay.
A control RAd vector (RAd-TK/GFP) harboring a hybrid gene encoding the herpes simplex virus type 1 (HSV-1) thymidine kinase fused to Aequorea victoria enhanced green fluorescent protein (a kind gift from Dr. Jacques Galipeau, McGill University, Montreal, Canada) was constructed following the general procedures outlined above (Figure 1). The corresponding gene product, fusion protein TK/GFP, emits green fluorescence with high intensity when excited with 470-nm wideband light [28]. This hybrid gene is also driven by the mCMV promoter. The vector was expanded in 293 cells and purified and titrated as indicated above.

Astrocyte cultures
Astrocyte cultures were prepared by mechanical dissociation of cerebral cortex from newborn CD1 mice [29]. Experimental procedures were approved by our Institutional Animal Use and Care Committee (Spanish National Research Council Animal Experimentation Committee). Cell culture reagents were purchased from Invitrogen (Paisley, UK). The cortex was isolated under a dissecting microscope and cleaned of choroid plexus and meninges. Cell suspensions were filtered through a 70-μm nylon cell strainer into phenol red free Dulbecco's modified Eagle medium (DMEM) containing 10% fetal calf serum and 1% penicillinstreptomycin. After centrifugation, cells were cultured in 75-cm 2 tissue culture flasks at 37°C and 5% CO2. The medium was changed after 4 days in culture and subsequently two times a week for the entire culture period. Cellular confluence was observed 10 days after plating, producing around 4.3 × 10 6 cells per flask, showing a polygonal flat morphology. Enriched astrocyte cultures (90-95% of GFAP immunoreactive cells) were obtained after overnight shaking at 37°C a 280 rpm in a table top shaker (Thermo Forma, Marietta, OH) to minimize oligodendrocyte and microglia contamination. Astrocytes were removed from the flasks by incubation with 0.25% trypsin (type II-S; Sigma-Aldrich, St Louis, MO) and 0.04% EDTA (Sigma) and plated onto poly-L-lysine-coated six-well plates or coverslips in medium supplemented with 10% FBS. Cells were plated at 15,000 cells/cm 2 for immunocytochemistry and at 30,000 cells/cm 2 for gene expression or protein assays.

Cell treatments
One day after seeding, culture medium was changed using serum free medium and the cells were treated for 24 or 72 h with vehicle or IGF-I (50 or 100 nM), and in a second experiment the cells were treated for 24 or 72 h with RAd-IGF-I, RAd-TK/GFP or medium alone (uninfected control cells). The efficacy of infection, assessed with the RAd-TK/GFP vector, was approximately 60% of the cells in the culture. Some cultures were incubated with lipopolysaccharide (LPS, from Escherichia coli 026:B6, Sigma; 500 ng/ml) to induce an inflammatory response [30]. LPS was added for 1 h to determine NFB subcellular localization and for 5 h to assess IGF-I receptor and IGF binding proteins by western blotting, and interleukin 6 (IL6), tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β) and toll-like receptor 4 (TLR4) mRNA levels by quantitative realtime polymerase chain reaction. In some experiments, astrocytes were incubated with the IGF-I receptor tyrosine kinase inhibitor picropodophyllin (PPP, Calbiochem, 100 ng/ml) or vehicle.

IGF-I assay
IGF-I was extracted from supernatants by acid-ethanol cryoprecipitation after incubation for 24 h with RAd-TK/GFP, RAd-IGF-I or medium alone. IGF-I levels were assessed by radioimmunoassay as previously described [24] using antibody UB2-495 (distributed by A. F. Parlow, NHPP, NIDDK). Recombinant human IGF-I (Cell Sciences Inc., Canton, MA, USA) was used as tracer and unlabeled ligand.

Immunocytochemistry
Subcellular distribution of NFB (p65) was assessed by immunocytochemistry after incubation for 24 h with RAd-IGF-I or RAd-TK/GFP and for 1 h with LPS or vehicle. Cells were fixed for 15 min at room temperature in 4% paraformaldehyde, permeabilized for 15 min with 0.3% Triton X-100 in PBS and incubated for 15 min with 10% normal goat serum in PBS, to block unspecific binding of the secondary antibody. Cells were then incubated for 1 h with anti-NFB (p65) mouse monoclonal antibody (BD Biosciences, diluted 1:50 in PBS with 1% BSA). After washing in PBS, cells were incubated with goat anti-mouse Alexa-fluor 594 (Invitrogen, diluted 1:1,000). Cell nuclei were stained with DAPI.

Statistical analysis
Statistical significance was assessed by one-way or twoway analysis of variance (ANOVA) followed by the Bonferroni post hoc test using GraphPad Prism 5 (GraphPad Software, San Diego, CA) or by Student's t-test when only two groups were compared. A probability of P < 0.05 was adopted for statistical significance. Data shown in the figures are the results of four independent experiments. Data are represented as mean ± SEM.

Results
IGF-I increases the expression of IL6, IL-1b and TNF-a and decreases the expression of toll-like receptor 4 (TLR4) in astrocytes under basal conditions mRNA levels of IL6, IL-1β and TNF-α were assessed in astrocytes 24 h after the addition of IGF-I (50 and 100 nM) to the cultures. IGF-I, at a concentration of 50 nM, resulted in a significant increase in mRNA levels for IL6, IL-1β and TNF-α compared to astrocytes treated only with vehicle ( Figure 2). In contrast, 100 nM IGF-I did not significantly affect IL6, IL-1β and TNF-α mRNA levels ( Figure 2). The increase in IL-1β mRNA levels was still detected 72 h after addition of 50 nM IGF-I to the cultures ( Figure 2). However, levels of IL6 and TNF-α were not significantly different between astrocytes treated with IGF-I or vehicle, 72 h after addition of IGF-I (Figure 2). TLR4 mRNA levels were significantly decreased in cultures treated with IGF-I at either 50 nM or 100 nM, compared to cultures treated with vehicle. The effect of IGF-I on TLR4 levels was still detected 72 h after treatment (Figure 2).

IGF-I exerts an anti-inflammatory action on LPS-treated astrocytes
To determine whether IGF-I affects the response of astrocytes to a proinflammatory challenge, we first assessed mRNA levels of IL6, TNF-α, IL-1β and TLR4 in primary astrocyte cultures incubated with LPS (500 ng/ml) or vehicle. Treatment with LPS resulted in a marked increase in mRNA levels of IL6, IL-1β and TNF-α (Figure 3). In contrast, TLR4 mRNA levels were not affected by LPS (Figure 3).
For LPS-treated astrocytes, mRNA levels of IL6, IL-1β and TLR4 were significantly decreased 24 h after administration of 50 or 100 nM IGF-I in comparison to cultures treated with LPS alone (Figure 4). mRNA levels of IL6, IL-1β and TLR4 returned to control levels 72 h after addition of 50 nM IGF-I to the cultures (Figure 4).

RAd-IGF-I increases IGF-I levels in astrocyte cultures
Having established the effects of exogenous application of IGF-I to astrocytes on IL6, IL-1β, TNF-α and TLR4 mRNA levels under basal and LPS-stimulated conditions, we investigated whether similar effects could be obtained by increasing endogenous expression of IGF-I in these cells using an adenoviral vector. To assess the efficacy of IGF-I gene transfer, IGF-I levels were assessed in the culture medium after incubation of the astrocytes for 24 h with RAd-IGF-I, RAd-TK/GFP, or medium alone (uninfected cells; Figure 5). IGF-I levels were significantly higher in the medium of astrocytes incubated with RAd-IGF-I compared to that of astrocytes incubated with RAd-TK/GFP or with medium alone ( Figure 5). In contrast IGF-I receptor levels and IGFBP2 levels were not significantly different in astrocytes incubated for 24 h with RAd-IGF-I, RAd-TK/GFP, or medium alone (data not shown). IGFBP3 was undetectable in the cultures.

RAd-IGF-I increases expression of IL6 and IL-1b and decreases expression of TLR4 under basal conditions
IL6 and IL-1β mRNA levels in astrocytes were significantly increased 24 h after incubation of cultures with RAd-IGF-I, compared to those of uninfected astrocytes or to astrocytes treated with RAd-TK/GFP ( Figure 6). In contrast, 72 h after incubation of the cultures with RAd-TK/GFP, astrocytes showed a significant increase in mRNA levels of IL6 and IL-1β, compared to uninfected astrocytes or to astrocytes treated with RAd-IGF-I ( Figure 6). IL6 and IL-1β mRNA levels 72 h after incubation with RAd-IGF-I were not significantly different from those in uninfected astrocytes ( Figure 6). TNF-α mRNA levels were increased 72 h after incubation with RAd-IGF-I. TLR4 mRNA levels were significantly decreased 24 h after incubation of the cultures with RAd-IGF-I, compared to astrocytes treated with RAd-TK/GFP ( Figure 6).

RAd-IGF-I exerts an anti-inflammatory action on LPStreated astrocytes
After LPS treatment, astrocytes incubated with RAd-IGF-I showed a significant decrease in mRNA levels for IL6, IL-1β and TNF-α compared to astrocytes incubated with RAd-TK/GFP (Figure 7). In addition, IL6 and IL-1β mRNA levels were significantly decreased in cultures incubated with RAd-IGF-I compared to uninfected cultures. In LPS-treated astrocytes, TLR4 mRNA levels were significantly decreased in cultures incubated for 24 h or 72 h with RAd-IGF-I compared to astrocytes treated with RAd-TK/GFP. In addition, TLR4 mRNA levels were significantly decreased in astrocytes incubated with RAd-IGF-I for 24 h compared to uninfected astrocytes.

Role of IGF-I receptor in the effects of RAd-IGF-I
To determine whether the effects of RAd-IGF-I on IL6, IL-1β, TNF-α and TLR4 mRNA levels are mediated by IGF-I receptor, astrocyte cultures were treated with the IGF-I receptor antagonist cyclolignan picropodophyllin (PPP). As shown in figure 8, PPP did not block the  Figure 8G).
In astrocytes treated with LPS, the effects of RAd-IGF-I on IL6, IL-1β, TNF-α and TLR4 mRNA levels were reverted by incubating the cultures for 5 h with PPP ( Figure 8B,D,F,H). In addition PPP increased mRNA levels of IL6 and IL-1β in LPS-treated astrocytes incubated with RAd-TK/GFP or RAd-IGF-I ( Figure 8B,D).

RAd-IGF-I prevents LPS-induced nuclear translocation of the NFB p65 subunit
Since activation of TLR4 results in nuclear translocation of the NFB p65 subunit, we assessed whether RAd-IGF-I affects p65 subcellular distribution in astrocytes. The incubation of astrocytes with RAd-TK/GFP and LPS resulted in a significant increase in p65 immunoreactivity in the cell nucleus compared to astrocytes incubated with RAd-TK/GFP alone. In contrast, p65 nuclear immunoreactivity was not increased by LPS in astrocytes incubated with RAd-IGF-I (Figure 9). Therefore, RAd-IGF-I prevented LPS-induced p65 nuclear translocation.

Discussion
The findings of the present study indicate that IGF-I increases the expression of IL6, IL-1β and TNF-α in cultured astrocytes under basal conditions. In contrast, IGF-I decreases expression of IL6 and IL-1β in cultured astrocytes submitted to a proinflammatory challenge by treatment with LPS. Under basal conditions, the effects of IGF-I on expression of IL6, IL-1β and TNF-α were dependent on IGF-I dose, promoting these cytokine expressions at the lower dose. When challenged with LPS, however, IGF-I at both dosages exerted similar inhibitory effects on the expression of IL6 and IL-1β.
The increases in IL6, IL-1β and TNF-α mRNA levels detected in astrocytes after treatment with IGF-I under basal conditions does not necessarily imply a proinflammatory action of IGF-I. Cytokines such as IL6, IL-1β and TNF-α have different physiological functions, including regulation of neuronal development, ionic homeostasis, neuropeptide release and synaptic plasticity [31][32][33][34]. Therefore, the observed increases in IL6, IL-1β and TNF-α levels under basal conditions may represent a physiological action of IGF-I. In addition, IGF-I reduced expression of TLR4 in both the presence and absence of LPS. TLR4 is a member of the IL-1receptor/ TLR superfamily that is expressed by astrocytes and that is required for an LPS-induced inflammatory response [35][36][37][38]. Therefore the observed decrease in TLR4 expression by IGF-I under basal conditions and after LPS proinflammatory challenge may contribute to a reduced capacity of astrocytes to be activated by proinflammatory molecules such as LPS. Indeed, IGF-I reduced the effects of LPS on mRNA levels of IL6, and IL-1β in astrocytes, in agreement with previous findings showing that IGF-I down-regulates cytokine expression induced by LPS in adult mouse brain [39].
Having established that exogenous IGF-I regulates the expressions of IL6, IL-1β, TNF-α and TLR4 in astrocytes, we then assessed whether an increase in endogenous production of IGF-I by astrocytes might also affect expression of these molecules. The findings of the present study indicate that a recombinant adenoviral vector harboring the gene for rat IGF-I (RAd-IGF-I) is able to increase the production of IGF-I in astrocytes, without affecting IGFBP2 or IGF-I receptor levels. Under basal conditions, RAd-IGF-I resulted in an effect similar to that of IGF-I, increasing the expression of IL6 and IL-1β and decreasing the expression of TLR4. Adenoviral vector infection per se also had an effect, since mRNA levels of IL6 and IL-1β were increased in astrocytes 72 h after incubation with control vector, RAd-TK/GFP. In addition, TNF-α mRNA values. *,**,***, Significant differences (* P < 0.05; **P < 0.01; ***P < 0.001) in comparison to the values of uninfected cells. § , § § , § § § , Significant differences ( § P < 0.05; § § P < 0.01***; § § § P < 0.001) between RAd-TK/GFP and RAd-IGF-I. levels were increased in astrocytes incubated for 72 h with RAd-IGF-I. However, no increase for IL6 or IL-1β was detected at 72 h in astrocytes infected with RAd-IGF-I, suggesting that IGF-I is compensating for some of the proinflammatory actions of the adenoviral vector. Indeed, RAd-IGF-I exerted a clear anti-inflammatory action on LPS-treated astrocytes, decreasing the mRNA levels of IL6, IL-1β, TNF-α and TLR4. Therefore, RAd-IGF-I imitated the effects of IGF-I on LPS-stimulated astrocytes.
These findings demonstrate efficacy of IGF-I gene therapy in reducing the response of astrocytes to an inflammatory challenge. However, as was observed for exogenous IGF-I, the anti-inflammatory action of IGF-I gene therapy was transient, since RAd-IGF-I did not significantly affect levels of IL6, IL-1β, or TNF-α in cultures incubated with the viral vector for 72 h.
Our data suggest that RAd-IGF-I, like IGF-I, reduces the effects of LPS on astrocytes by decreasing the  expression of TLR4. Activation of TLR4 results in translocation of the NFB p65 subunit to the cell nucleus and consequent activation of NFB-mediated transcription of proinflammatory cytokines and chemokines [40][41][42][43][44]. In agreement with the effect of RAd-IGF-I on TLR4 expression, and with the results of a previous study on the effect of IGF-I on astrocytes [18], IGF-I gene therapy prevented LPS-induced translocation of the NFB p65 subunit to the cell nucleus. This suggests that RAd-IGF-I exerts, at least in part, its anti-inflammatory action by downregulation of TLR4 and subsequent inhibition of NFB activity.
The anti-inflammatory action of RAd-IGF-I in LPSstimulated astrocytes may be mediated by IGF-I receptor activation, since it was prevented by incubation of astrocytes with cyclolignan picropodophyllin (PPP), an inhibitor of the IGF-I receptor tyrosine phosphorylation [45][46][47]. However, PPP treatment per se increased IL6 mRNA levels in cells infected with Rad-IGF-I construct or control construct under basal conditions, and increased mRNA levels of IL6 and IL-1β in LPS-treated astrocytes incubated with RAd-TK/GFP or RAd-IGF-I. These actions of PPP cannot be explained simply by IGF-I receptor inhibition, suggesting that PPP may have other effects under the experimental conditions used.

Conclusion
The results of the present study indicate that IGF-I and IGF-I gene therapy are able to reduce inflammatory reactions in glial cells exposed to LPS. This action of IGF-I and IGF-I gene therapy may be relevant for control of reactive gliosis under chronic neurodegenerative conditions, where the glial inflammatory response is thought to have a negative impact on neuronal function and survival.