Thymosin β₄ inhibits TNF-α-induced NF-κB activation, IL-8 expression, and the sensitizing effects by its partners PINCH-1 and ILK

Abstract

The mechanisms by which thymosin β 4 (Tβ₄) regulates the inflammatory response to injury are poorly understood. Previously, we demonstrated that ectopic Tβ₄ treatment inhibits injury-induced proinflammatory cytokine and chemokine production. We have also shown that Tβ₄ suppresses TNF-α-mediated NF-κB activation. Herein, we present novel evidence that Tβ₄ directly targets the NF-κB RelA/p65 subunit. We find that enforced expression of Tβ₄ interferes with TNF-α-mediated NF-κB activation, as well as downstream IL-8 gene transcription. These activities are independent of the G-actin-binding properties of Tβ₄. Tβ₄ blocks RelA/p65 nuclear translocation and targeting to the cognate κB site in the proximal region of the IL-8 gene promoter. Tβ₄ also inhibits the sensitizing effects of its intracellular binding partners, PINCH-1 and ILK, on NF-κB activity after TNF-α stimulation. The identification of a functional regulatory role by Tβ₄ and the focal adhesion proteins PINCH-1 and ILK on NF-κB activity in this study opens a new window for scientific exploration of how Tβ₄ modulates inflammation. In addition, the results of this study serve as a foundation for developing Tβ₄ as a new anti-inflammatory therapy.—Qiu, P., Kurpakus Wheater, M., Qiu, Y., Sosne, G. Thymosin β₄ inhibits TNF-α-induced NF-κB activation, IL-8 expression, and the sensitizing effects by its partners PINCH-1 and ILK.

Thymosin β 4 (Tβ₄) is a water-soluble, 43-aa polypeptide with a molecular mass of 4.9 kDa (1). Tβ₄ is ubiquitously expressed and highly conserved across species (2). Although historically considered to function solely as the major monomeric G-actin-sequestering protein in the cytoplasm, an ever-growing number of studies provide evidence to highlight the multifunctional intracellular and extracellular roles of Tβ₄ (3, 4). While the mechanism of Tβ₄ entry into cells is still unclear, and no cell membrane receptors for Tβ₄ are known, its extracellular activities have been widely observed as a secreted peptide in wound and oral fluids (5–7). Tβ₄ was also identified as an essential paracrine factor from both Akt-modified mesenchymal stem cells and embryonic endothelial progenitor cells in the heart (8, 9).

Recently, several proteins were identified as intracellular binding partners for Tβ₄, including focal adhesion proteins [LIM protein PINCH-1 and the ankyrin repeat (AR) protein ILK], hMLH1, Ku80, and stabilin-2 (10–13). Tβ₄ is now known to promote wound healing, tissue regeneration, and cytoprotection in the cornea, heart, skin, gingiva, and nervous system (9, 10, 14–23). In addition to accelerating wound repair, Tβ₄ has anti-inflammatory and anti-septic shock activities, as well as antimicrobial and antistaphylococcal biofilm activities (14, 15, 24–28). Since the mechanisms of anti-inflammatory activity are not completely understood, Tβ₄ is the focus of investigations into its novel anti-inflammatory function by modulating immune regulatory cells, inflammatory signaling mediators, and the critical transcription factor, NF-κB (29–31).

TNF-α is one of the most potent proinflammatory cytokines. TNF-α plays a prominent role in a number of autoimmune diseases, including rheumatoid arthritis, systemic lupus erythematosis, and inflammatory bowel disease, as well as a number of ocular and corneal inflammatory diseases. Biological agents developed to target TNF-α signaling have been approved for clinical therapy (32–34). The key factor for TNF-α-mediated inflammation is NF-κB, a transcription factor responsible for both innate and adaptive immune responses. NF-κB dysregulation is implicated in the pathogenesis of many inflammatory diseases (35–37).

TNF-α activates NF-κB and stimulates proinflammatory gene expression primarily through the canonical signaling cascade that results in IκB (the key NF-κB inhibitor) degradation via a sequential activation of the receptor proximal signaling adaptor complex (TRADD-TRAF2/5-RIP1-TAK1/MEKK3) and the IKK complex (IKKα/β and the modulator Nemo). The released dimeric NF-κB enters the nucleus to target its cognate DNA (κB site; ref. 38). Negative feedback loops for activated NF-κB signaling ensure the proper termination of NF-κB-mediated immune responses for the resolution of inflammation and avoidance of further tissue damage by uncontrolled inflammation (39). In contrast to the well-characterized system for initiating NF-κB signaling, the terminating processes for NF-κB activation are less well understood. The AR proteins, the IκB family, and the LIM-domain protein PDLIM2 share a remarkable functional similarity in targeting and in silencing the NF-κB subunit RelA/p65 directly through their protein-protein interactions (40–42).

NF-κB is bound by actin and actin-associated proteins and is distributed along with actin-containing structures in different cellular regions, such as focal adhesions, stress fibers, and the nuclear matrix (43–45). Functional studies first connected the modulation of the dynamic balance between actin monomers and polymers to the regulation of NF-κB activity (45–47). Focal adhesion complex proteins contribute directly to TNF-α signaling transduction. Additional intracellular NF-κB activators or inhibitors anchored into both the nuclear matrix and cytoplasmic stress fibers have been shown to physically interact with NF-κB (48–51).

The mechanisms of the anti-inflammatory properties of Tβ₄ remain poorly understood. Considering that Tβ₄ is a major intracellular monomeric G-actin-sequestering molecule that also interacts with the focal adhesion proteins PINCH-1 and ILK, we investigated whether the anti-inflammatory properties of Tβ₄ are related to its association with actin and these intracellular binding partners. In this report, we extend our previous findings that Tβ₄ inhibits TNF-α-mediated NF-κB activation and provide novel evidence for the negative regulation of Tβ₄ on the activation of NF-κB subunit RelA/p65, as well as the expression of the downstream proinflammatory gene IL-8. We provide evidence that Tβ₄ directly targets the NF-κB subunit RelA/p65 and inhibits the sensitizing effects of its intracellular binding partners, PINCH-1 and ILK, in an actin-independent manner.

MATERIALS AND METHODS

Reagents, materials, and plasmids

Synthetic Tβ₄ was obtained as kind gift from RegeneRx Biopharmaceuticals (Rockville, MD, USA). Primary human corneal epithelial cells (HCECs) and adult human epidermal keratinocytes (HEKas) were obtained from Cascade Biologics (Portland, OR, USA). The immortalized HCEC line HCET, the rat artery smooth cell line A7r5, and the COS-7 cell lines were purchased from the American Type Culture Collection (ATCC; Manassas, VA, USA). Primary human corneal fibroblasts were provided by Fu-Shin Yu (Wayne State University, Detroit, MI, USA). The human conjunctival epithelial cell line HCO597 was generously provided by Sherry Ward (Gillette Medical Evaluation Laboratories, Gillette Co., Gaithersburg, MD, USA).

The pIL-8 promoter plasmid (−135 to +46) Luc and the pIL-8 promoter plasmid (−135 to +46/ΔNF-κB) Luc were gifts of Lawrence S. Young (University of Birmingham, Birmingham, UK). pEGFP-N1-Tβ₄ and pEGFP-N1-Tβ_4m7A plasmids were kindly provided by Czeslaw S. Cierniewski (Nencki Institute of Experimental Biology, Warsaw, Poland). pFlag-CMV-2-ILK (aa1–452) plasmid, pFLAG-CMV-2-ILK-ANK1-deletion (aa 66–452) plasmid, and pFLAG-CMV-2-ILK-ANK repeats (aa 1–230) plasmid, as well as GFP-tagged PINCH-1 and truncated mutants, were kindly provided by Chuanyue Wu (University of Pittsburgh, Pittsburgh, PA, USA). The pEGFP-p65 construct was obtained from Johannes A. Schmid (Medical University, Vienna, Austria). The expressed GFP-p65 was reported to be fully functional both in vitro and in vivo similar to wild-type p65 (52). The pEGFP-IκBα construct was provided by Michael R. H. White (University of Liverpool, Liverpool, UK). The pcDNA 3.1-profilin-1 was kindly provided by Marc I. Diamond (University of California at San Francisco, San Francisco, CA, USA).

QuantiTech primer assays for IL-8 and GAPDH, as well as Mini- and Maxi-Plasmid Kits, were purchased from Qiagen (Valencia, CA, USA). SuperScript III reverse transcriptase and TaqDNA polymerase were obtained from Invitrogen (Carlsbad, CA, USA). The RT² qPCR-grade RNA isolation kit, RT² first-strand kit, and RT² SYBR green/fluorescein qPCR master mix were ordered from SuperArray Bioscience (Frederick, MD, USA). PathDetect pNF-κB-Luc cis-Reporter plasmid was purchased from Stratagene (La Jolla, CA, USA). All other reagents were purchased from Sigma-Aldrich Chemical (St. Louis, MO, USA).

Cell culture, drug treatments, and transfection

The culture conditions for HCET, HCO597, and rat pulmonary artery smooth muscle (PAC-1) cells were described previously (17, 20, 53). COS-7, A7r5, and HEKa cells were grown according to the supplier's protocols. Human corneal stromal fibroblasts were cultured in DMEM supplemented with 10% FBS. Prior to transfection, subconfluent cells in 24-well plates were incubated in serum-free medium overnight. Cells were cotransfected with reporter plasmids and equal amounts of effecter plasmids by using Lipofectamine/Plus reagent (Invitrogen), according to the supplier's manual. After 24 h of transfection, cells were pretreated with 0.5 μM jasplakinolide (JP; Invitrogen), 5 μM cytochalasin D (CytoD; Enzo Life Sciences, Farmingdale, NY, USA), and 2.5 μM latrunculin A (LatA; Enzo Life Sciences) for 1 h, and then treated with 10 ng/ml TNF-α (R&D Systems. Minneapolis, MN, USA) for 16 to 18 h. Following washes with PBS, cell lysates were collected using 200 μl/well 1× report lysis buffer (Promega, Madison, WI, USA) and then mixed with 100 μl luciferase assay medium (Promega). Luciferase activity was measured in 20 μl cell lysate using the SpectraMax M5 (Molecular Devices Corp., Sunnyvale, CA, USA). Each transfection assay sample was analyzed in triplicate, and the data presented are representative of ≥3 independent experiments. Luciferase activity was normalized by using the dual-luciferase assay system (Promega) or by quantifying protein concentrations in cell lysates using the Coomassie plus protein assay (Thermo Scientific, Rockford, IL, USA). Luciferase assays were statistically analyzed using ANOVA and a Tukey post hoc test with significance set at P < 0.05.

For stable transfection of GFP-Tβ₄ in HCET cells, cells transfected by pEGFP-N1-Tβ₄ were selected by treatment with 40 μg/ml geneticin (Invitrogen) in selection medium (KGM-2/F12, 1:1, with 12% FBS) for ≥7 d. Cells that survived the treatment were pooled, and overexpression of GFP- Tβ₄ was confirmed by Western blot analysis.

Real-time RT-PCR, ELISA, and chromatin immunopreciptation (ChIP)

For real-time RT-PCR, total RNA preparation and IL-8 mRNA quantification were performed using SuperArray kits and primers (Qiagen), according to the manufacturer's manuals. For detection of IL-8 by ELISA (Quantikine human IL-8 immunoassay kit; R&D Systems), 400 μl of conditioned medium was collected from each well of a 24-well plate containing HCET cells and centrifuged at 850 g for 5 min to remove unattached cells prior to analysis. The data are representative of 3 independent experiments.

ChiP assay was performed according to previously described protocols (53). Formaldehyde-crosslinked DNA-protein complexes were precipitated using rabbit polyclonal anti-NF-κB p65 (Santa Cruz Biotechnology, Santa Cruz, CA, USA). The released DNA fragments, after reverse cross-linking by heating, were analyzed using the following primer pairs: IL-8 promoter proximal κB site (−146 to +62), 5′-GAAGTGTGATGACTCAGGTTTGC-3′ and 5′-GTTGGTTTTCTTCCTGGCTCT-3′; NF-κB-like site (+1996 to +2416), 5′-AGGGTTGCCAGATGCAATAC-3′ and 5′-AAACCAAGGCACAGTGGAAC-3′; upstream control site (−3589 to −3319), 5′-AACCCAGGTGAGAGCTGAGA-3′ and 5′-CCCACGCCATAGGAATTTTA-3′; downstream control site (+3236 to +3396), 5′-TGCTTCCCCTTAGCATTTTG-3′ and 5′-ACAGTGGGGAGTTGAAAACATT-3′.

Immunofluorescence staining and confocal microscopy

HCET cells were seeded onto coated glass coverslips, placed in wells of a 24-well plate, and grown to subconfluency. Cells were transfected with 1 μg/well of pEGFP-N1-Tβ₄ and pEGFP-N1-Tβ_4m7A, as described above. At 24 h after transfection, cells were washed with prewarmed PBS 3 times and fixed with 1% paraformadehyde for 10 min at room temperature. After washing 4 times with PBS, cells were permeabilized in 0.2% Triton X-100 for 5 min and blocked by incubating with 1× Blocker BSA/TBS (Thermo Scientific) for 30 min.

For indirect fluorescence staining of RelA/p65, 1 μg/ml of mouse monoclonal NF-κB p65 antibody (Santa Cruz Biotechnology) was added and incubated overnight at 4°C. After washing with PBS 3 times, cells were incubated with 5 μg/ml Alexa Fluor 594 chicken anti-mouse IgG (Invitrogen) for 1 h at room temperature. Following washes with PBS, the coverslips were taken out of the wells of the 24-well plate and mounted on microscope slides using Prolong Gold Antifade Reagent (Invitrogen). Photoimages were taken at room temperature using an Axioplan 2 Imaging Apotome (Carl Zeiss, Oberkochen, Germany) and processed with AxioVision 4.7 software (Zeiss).

For double-staining experiments, subconfluent HCET cells on coverslips were first fixed by 4% paraformadehyde and permeabilized in 0.2% Triton X-100. Samples were then sequentially incubated with a mixture of NF-κB p65 mouse monoclonal and Tβ₄ rabbit polyclonal antibodies (Santa Cruz Biotechnology) and Alexa Fluor 647 goat anti-mouse IgG and Alexa Fluor 488 donkey anti-rabbit IgG (Invitrogen). Immunofluorescent images were captured at room temperature using the Leica TCS SP2 spectral confocal and multiphoton system (Leica Microsystems, Wetzlar, Germany). A series of optical sections through each cell was taken at vertical steps between 0.5 and 0.75 μm.

RESULTS

Overexpression of Tβ₄ suppresses the dose-dependent effects on NF-κB binding activities both after treatment with TNF-α and enforced expression of the NF-κB subunit RelA/p65

In contrast to numerous studies demonstrating the effects of Tβ₄ on various cellular processes, such as migration, differentiation, apoptosis, and transformation after exogenous treatment, relatively fewer reports exist describing the functional effects of overexpressing Tβ₄ on cell mobility, apoptosis, cell signaling transduction, and gene expression (9, 10, 12, 54–57) Here, we assessed the effects of Tβ₄ and other related functional proteins on TNF-α-induced NF-κB activation by transiently coexpressing relevant proteins with a luciferase reporter driven by a 5-tandem-repeat κB site from the murine CXCL10 promoter. Our previous findings demonstrated that exogenous treatment of Tβ₄ suppressed TNF-α-induced NF-κB activation in HCECs. We first confirmed that TNF-α stimulates NF-κB binding activity in a dose-dependent manner, and overexpression of Tβ₄ remarkably inhibited NF-κB binding activity both in resting cells and in cells treated with different doses of TNF-α (Fig. 1A).

Figure 1.

Tβ₄ suppresses TNF-α-induced NF-κB activation. A) Luciferase activities in HCET lysates were tested from cells transiently transfected with NF-κB luciferase reporter and an empty mammalian expression vector pcDNA3.1 (control). In the Tβ₄-enforced expression group, cells were transfected with an NF-κB luciferase reporter and a Tβ₄-expressing plasmid based on pcDNA-3.1 vector. Inset: the NF-κB luciferase reporter is composed of a cDNA sequence for the firefly luciferase enzyme and a promoter sequence of the 5 tandem repeated κB sites (sequence bolded and underlined) from a murine CXCL10 promoter. After 24 h of transfection, cells were treated or not with indicated doses of human recombinant TNF-α for 18 h prior to assay for luciferase activity. B) Luciferase activities were tested from HCET cells cotransfected with DNA vectors, including NF-κB luciferase reporter, GFP-RelA/p65, and pcDNA3.1-Tβ₄. Ratios of transfected DNA vectors encoding for RelA/p65 (maximum activity group) and Tβ₄ were 4:1, 2:1, and 1:1, respectively. Luciferase activity was normalized against the Renilla luciferase reporter (transfection control) or against protein concentrations in cell lysates (cell quantity control). Luciferase activities are expressed as arbitrary units related to the reference activity of 100 from cells transfected with NF-κB luciferase reporter alone. Luciferase activity (arbitrary units) is represented as means ± se. Experiments were repeated independently 3 times with similar results. C) HEKa, HCO597, COS-7, A7r5, and PAC-1 cells were transfected and treated with TNF-α (10 ng/ml) following the same protocol used in HCET cells.

We further tested the hypothesis that the inhibitory effect of Tβ₄ on TNF-α-induced NF-κB activation may act directly on the function of the transcriptional factor NF-κB. Our results establish that enforced expression of RelA/p65 alone causes a similar stimulatory effect on luciferase activity after TNF-α treatment. This stimulation could be blocked by increased Tβ₄ concentration (Fig. 1B). The inhibitory effect on NF-κB activation by increasing intracellular Tβ₄ levels was also detected in several other cell lines, including COS-7 cells and human dermal keratinocytes (HEKa line) and conjunctival epithelial cells (HCO597 line). However, Tβ₄ does not show this inhibitory function in rat vascular smooth muscle cell lines, such as A7r5 and PAC-1 (Fig. 1C). These critical observations that Tβ₄ efficiently targets the NF-κB subunit RelA/p65 in the dermal and ocular surface (corneal and conjunctival) epithelial cells support the notion that the anti-inflammatory action of Tβ₄ may be a major factor in its ability to promote wound healing after skin and eye injuries.

Tβ₄, but not profilin-1, inhibits NF-κB binding activity, and this inhibitory effect is not dependent on its G-actin-sequestering property

We next investigated whether the ability of Tβ₄ to inhibit NF-κB-binding function depends on its G-actin-binding properties. To this end, we used a 2-pronged approach. First, we compared the functional inhibition by overexpressing wild-type Tβ₄ and its mutant Tβ_4M7A, which lacks G-actin binding activity due to replacement of the actin-binding motif, ₁₇KLKKTET₂₃, with 7 alanines (12). In a second approach, we used LatA to assay its effects on Tβ₄-mediated NF-κB inhibition. LatA is an actin-binding sponge toxin previously described specifically to interrupt actin binding to Tβ₄ (58).

As shown in Fig. 2A, overexpression of Tβ₄ either alone or with GFP labeling demonstrated the same inhibitory effect on NF-κB-binding function, indicating that attachment of GFP to the Tβ₄ polypeptide did not interfere with its inhibitory capabilities. Contrary to our expectations, overexpression of the Tβ_4M7A mutant inhibited NF-κB binding activity even more strongly than the wild-type protein. Further, we overexpressed another major G-actin-sequestering protein, profilin-1, to test whether it shares the inhibitory effects of Tβ₄ on TNF-α-mediated NF-κB activation. We found that enforced expression of profilin-1 does not inhibit NF-κB binding activity. In addition, cotransfection of profilin-1 with Tβ₄ does not hinder the ability of Tβ₄ to inhibit TNF-α-induced NF-κB activation. Taken together, these results suggest that the actin-binding domain of Tβ₄ does not contribute to its inhibitory effects on NF-κB activation. Tβ₄ may be unique among the actin-binding proteins in its ability to interfere with NF-κB binding to its cognate DNA in targeted gene promoter regions.

Figure 2.

Tβ₄ inhibits NF-κB binding activity in a G-actin-binding-independent manner. A) Both baseline and TNF-α-induced NF-κB binding activities were suppressed by overexpressing Tβ₄, GFP-Tβ₄, and GFP-Tβ_4(M7A), but not profilin-1. Inset: the major G-actin binding frame (KLKKTET) in Tβ_4(M7A) was replaced by 7 alanines. The NF-κB luciferase reporter was cotransfected with DNA plasmids with each mammalian expression vector for indicated proteins for 24 h, and then treated with TNF-α (10 ng/ml) for 18 h. *P = 0.0011, **P < 0.0001 vs. GFP. B) Actin-targeting drug LatA, but not CytoD or JP, strengthened the inhibitory effects of TNF-α-mediated NF-κB activation by Tβ₄. HCET cells with (shaded bars) or without (open and solid bars) pEGFP-N1-Tβ₄ transfection were pretreated with 2.5 μM LatA, 5 μM CyotD, or 0.5 μM JP for 1 h and then treated with TNF-α (10 ng/ml) for 18 h. Compared to untreated control, LatA treatment resulted in a significant (P<0.0001 for all comparisons) decrease in luciferase activity after TNF-α-treatment, both in non-GFP-Tβ₄-expressing cells (solid bars) and GFP-Tβ₄-expressing cells (shaded bars). Luciferase activity (arbitrary units) is represented as means ± se. Experiments were repeated independently 3 times with similar results.

Because we found that Tβ₄ inhibited NF-κB activation by TNF-α without requiring binding to actin, we also used LatA to interfere with the binding between Tβ₄ and G-actin, and tested the consequential effects on Tβ₄-mediated NF-κB inhibition (Fig. 2B). While other actin-targeting drugs, such as CytoD and JP, strongly stimulated NF-κB activation in HCECs, treatment with LatA did not similarly influence NF-κB binding activity. Conversely, LatA suppressed TNF-α-mediated stimulatory effects. It is noteworthy that there was a synergistic effect on the inhibition of the TNF-α-induced NF-κB activation when Tβ₄-overexpressing cells were treated with LatA. Overexpression of Tβ₄ also completely blocked the stimulatory effects on NF-κB activation by either CytoD or JP. On the basis of these findings, we next explored the potential interactions between Tβ₄ and the NF-κB subunit RelA/p65.

Tβ₄ colocalizes with the NF-κB subunit RelA/p65 in both the cytoplasm and nucleus in resting and TNF-α-stimulated cells

To assess the potential interaction between Tβ₄ and the RelA/p65 with and without TNF-α treatment, we observed the intracellular distribution and colocalization of both Tβ₄ and RelA/p65 in HCET cells via indirect immunofluorescence and confocal microscopy (Fig. 3A). In cells without TNF-α treatment, RelA/p65 localized to the cytoplasm, whereas Tβ₄ distributed in both the cytoplasm and nucleus. The distribution of Tβ₄ in HCET cells was similar to previous studies, which showed a pronounced nuclear staining of Tβ₄ in the human cervical cancer cell line SiHa and the mammary carcinoma cell line MCF-7 (57, 59). Interestingly, we also discovered that Tβ₄ and RelA/p65 colocalized in the cytoplasm, especially in the perinuclear areas. After TNF-α treatment for 30 min, the majority of RelA/p65 transferred to the nucleus and colocalized with Tβ₄. The exception to this was the nucleolus, where RelA/p65 was excluded. The observation of Tβ₄ and p65 colocalization hints at the possibility that Tβ₄ interferes with NF-κB function via their physical interaction.

Figure 3.

Tβ₄ directly targets the NF-κB subunit RelA/p65. A) Confocal microscopy reveals that Tβ₄ and RelA/p65 colocalize in cytoplasmic compartments before, and in the nucleus after, TNF-α treatment. HCET cells were treated with TNF-α (10 ng/ml) for 1 h and then hybridized with primary antibodies to Tβ₄ and RelA/p65. Orange signal (merge) represents areas of colocalization of Tβ₄ (α-Tβ₄, green), and RelA/p65 (α-RelA/p65, red). B) Overexpression of GFP-Tβ₄ in HCET cells blocks TNF-α-mediated RelA/p65 nuclear translocation. HCET cells were transfected with plasmids encoding GFP or fusion protein GFP-Tβ₄. With (+TNF-α) or without (−TNF-α) 1 h TNF-α treatment, the localization of GFP or GFP-Tβ₄ fusion proteins (GFP, green) and RelA/p65 (α-RelA/p65, red) was recorded using the Apotome fluorescence microscope. Following TNF-α stimulation, RelA/p65 translocates to the nucleus (arrows) in 50–60% of cells. In contrast, overexpression of Tβ₄ inhibits RelA/p65 translocation to the nucleus (arrowheads). C) Similar blocking effects were observed in HCET cells transfected with plasmids encoding fusion protein GFP-Tβ_4(M7A). Following TNF-α stimulation, RelA/p65 translocates to the nucleus (arrows). In contrast, overexpression of Tβ_4(M7A) inhibits RelA/p65 translocation to the nucleus (arrowheads). Localization experiments were repeated independently 3 times with similar results. Images are representative. Scale bars = 20 μm.

Overexpression of Tβ₄ or Tβ_4M7A mutant blocks TNF-α-induced nuclear translocation of RelA/p65

Previously, we found that exogenous addition of Tβ₄ to HCECs suppressed TNF-α-mediated RelA/p65 nuclear translocation (30). Here, we extended these findings by overexpressing Tβ₄ in these cells to define the endogenous function on NF-κB activation after TNF-α stimulation. While overexpression of GFP alone did not block TNF-α-mediated NF-κB nuclear translocation, GFP-tagged Tβ₄ completely blocked the TNF-α stimulatory effect on the nuclear translocation of RelA/p65 (Fig. 3B). To test whether this blocking effect by Tβ₄ is related to its G-actin-binding function, we also overexpressed GFP-tagged Tβ_4M7A in HCECs and examined its effects on RelA/p65 nuclear translocation. We found that similar to wild-type Tβ₄, GFP-tagged Tβ_4M7A also blocked the RelA/p65 nuclear translocation after TNF-α treatment (Fig. 3C). These results are further evidence that Tβ₄ interferes with NF-κB activation independent of its actin-binding function.

Overexpression of Tβ₄ attenuates the sensitizing effects on NF-κB activation by its intracellular binding partners, PINCH-1 and ILK, in response to TNF-α treatment

Tβ₄ forms a complex with the focal adhesion proteins PINCH-1 and ILK in cardiomyocytes, thereby activating ILK/Akt signaling to protect heart muscle injury following myocardial infarction (10). Considering the fact that these Tβ₄ intracellular binding partners contain common NF-κB negative regulatory modules (LIM domains in PINCH-1 and ARs in ILK), we hypothesized that Tβ₄-mediated inhibition of NF-κB activation by TNF-α might be connected to these intracellular binding partners. Unexpectedly, our results showed that overexpression of either PINCH-1 or ILK in corneal epithelial cells enhanced NF-κB binding activities and sensitized the NF-κB response to TNF-α stimulation. Interestingly, these enhanced NF-κB activities caused by the elevated levels of intracellular PINCH-1 and ILK were significantly decreased (P<0.001) by overexpressing Tβ₄. Similar compromising effects were also detected by enforced expression of the mutant Tβ_4M7A (Fig. 4A). Taken together with the colocalization studies in Fig. 3, these results show that Tβ₄ may directly target p65 and thereby interfere with PINCH-1/ILK-mediated p65 activation.

Figure 4.

Tβ₄ suppresses the sensitizing effects of TNF-α-mediated NF-κB activation by its intracellular binding partners, PINCH-1 and ILK. A) In order to determine the effects of overexpressing PINCH-1 and ILK on NF-κB activation, as well as their influence by Tβ₄ or Tβ₄(M_7A)_, the NF-κB luciferase reporter was cotransfected with plasmids encoding for the indicated proteins in HCET cells. Equal amounts of plasmids were used in transfections for the effecter GFP-Tβ₄ or GFP-Tβ₄(M_7A) and for PINCH-1 or ILK. Plasmids of pEGFP and pcDNA3.1 were used to balance transfected DNA amounts in control groups. Transfection and TNF-α treatment followed procedures described in Fig. 1. Luciferase activity (arbitrary units) is represented as means ± se. B, C) HCET cells were transfected with NF-κB luciferase reporter and expression vectors carrying coding sequences for various functional domains in PINCH-1 (B) and ILK (C). Relative luciferase activity vs. baseline wild-type PINCH-1 and ILK (arbitrary units) is represented as means ± se. Experiments were repeated independently 3–6 times (A) or ≥3 times (B, C) with similar results. Results are representative.

Tβ₄ binds both to the LIM4 and LIM5 domains of PINCH-1, as well as to the ILK N-terminal AR. This same ILK N-terminal AR also binds to PINCH-1 via its LIM1 domain (60). The interaction among the functional domains of these three molecules in regulating NF-κB activity is unknown. To further identify the potential contribution of these functional domains in response to TNF-α-mediated NF-κB activation, we transiently overexpressed a combination of recombinant proteins representing different functional domains of PINCH-1 and ILK in corneal epithelial cells treated with TNF-α. Compared to the functional effects of the intact PINCH-1 molecule and different functional domains on NF-κB activation, we found that the PINCH-1 LIM1–2 domain demonstrated significantly weaker effects, while the strongest stimulatory effects were found by the PINCH-1 LIM1–4 domain (P<0.001 for all comparisons). Moreover, no significant differences in the inhibitory effects of overexpressed Tβ₄ were observed on NF-κB activation whether the enforced expression of Tβ₄ was unbound to the PINCH-1 LIM1–2 domains or to intact PINCH-1 (Fig. 4B). These results provide further support to propose that the Tβ₄-suppressive effects on p65 are independent of its binding to PINCH-1.

We further characterized the effects of the ILK N-terminal ankyrin domain on NF-κB activation after TNF-α-stimulation. As shown in Fig. 4C, the intact ILK molecule increases NF-κB activation after TNF-α-stimulation (P<0.0001), and most of the stimulatory effects can be attributed to its C-terminal kinase domain, while the N-terminal ankyrin domain was ineffective. However, no significant differences were detected comparing Tβ₄'s inhibitory effects when overexpressing intact ILK, as well as the fragments of ILK N-terminal and C-terminal functional domains. Taken together, these novel findings demonstrate that Tβ₄ intracellular binding partners, PINCH-1 and ILK, sensitize the NF-κB activation response to TNF-α-stimulation and suggest an important potential role for these molecules in proinflammatory conditions. Considering the ability of Tβ₄ to bind to both the PINCH/ILK complex and NF-κB RelA/p65 subunit, our findings suggest that Tβ₄ may be targeting the proinflammatory signaling mediated by PINCH/ILK but not the PINCH/ILK molecules themselves to suppress inflammation.

Tβ₄ suppresses NF-κB-mediated downstream IL-8 gene expression after TNF-α-stimulation

In previous studies, we demonstrated that exogenous Tβ₄ modulated inflammatory mediators in vivo (14, 15, 25). Among the various proinflammatory cytokines and chemokines inhibited in the cornea by Tβ₄ treatment after alkali injury, MIP-2, the murine homologue for IL-8, was significantly reduced (15). Work by others has demonstrated that the IL-8 promoter is activated primarily by the NF-κB homodimer p65-p65, especially in corneal epithelial cells after TNF-α-stimulation (61). In addition, we previously demonstrated that Tβ₄ inhibits p65 activation and nuclear translocation (30).

Since IL-8 is a prominent functional end point of the NF-κB canonic signaling pathway and plays a critical role in orchestrating polymorphonuclear (PMN) cell influx to sites of inflammation, we further investigated the effects of Tβ₄ interference with Rel/A/p65 on endogenous IL-8 gene expression (mRNA production, protein secretion, promoter activation, and RelA/p65 promoter targeting) in corneal epithelial cells (62). As demonstrated in Fig. 5, exogenous Tβ₄ treatment significantly suppressed IL-8 protein secretion into culture medium from primary corneal epithelial cells at all TNF-α treatment time-points (Fig. 5A). Similar effects were also observed in primary human corneal fibroblasts (Fig. 5B). We next created a stable cell line overexpressing GFP-tagged Tβ₄ in corneal epithelial cells. Similar to exogenous treatment, overexpressed Tβ₄ in corneal epithelial cells also suppressed IL-8 protein levels induced by TNF-α stimulation (Fig. 5C).

Figure 5.

Tβ₄ suppresses TNF-α stimulated IL-8 expression. A, B) Primary HCECs (A) or corneal fibroblasts (B) were pretreated (solid bars) or not (shaded bars) with Tβ₄ (1 μg/ml) for 1 h prior to stimulation with TNF-α (10 ng/ml) for the indicated times. IL-8 secretion into culture medium was assayed by ELISA. At all treatment time points measured, Tβ₄ significantly suppressed IL-8 secretion caused by TNF-α stimulation. IL8 secretion by cells treated with Tβ₄ only was negligible and is not shown. *P < 0.01. C) Stably enforced expression of GFP-Tβ₄ in HCET cells significantly decreased TNF-α-induced IL-8 secretion compared to control (GFP). IL-8 concentration in culture medium (pg/ml) is represented as means ± se. Experiments were repeated independently 6 times with similar results. Results are representative.

To determine whether down-regulating IL-8 transcription is responsible for Tβ₄ suppression of TNF-α-stimulated IL-8 production, we first performed real-time RT-PCR to measure IL-8 mRNA transcription levels after TNF-α treatment (Table 1). At every time point assayed, we observed a ≥4-fold increase of IL-8 mRNA levels. A peak production of >24-fold IL-8 gene transcription was seen at 60 min after TNF-α treatment. Pretreatment with Tβ₄ before TNF-α exposure significantly decreased IL-8 mRNA levels to 1/2 to 2/3 of those observed with cells treated with TNF-α only. All of the time points measured showed significant inhibition by Tβ₄ (both with and without TNF-α treatment), except for the 6-h time point when the effects of Tβ₄ had diminished. Our results show that both ectopic treatment and overexpression of Tβ₄ significantly decrease TNF-α-mediated IL-8 production in human corneal cells, both epithelial and fibroblasts.

Table 1.

Tβ₄ suppresses TNF-α-mediated IL-8 mRNA production in HCET cells

Treatment	Time (min)	Fold changes	P, control vs. treated	P, TNF-α vs. TNF-α/Tβ4
TNF-α	15	4.1 ± 0.231	0.0055	0.0014
TNF-α/Tβ4	15	2.3 ± 0.276	0.0171	0.0014
TNF-α	30	21.4 ± 0.953	0.0022	0.0293
TNF-α/Tβ4	30	14.7 ± 0.603	0.0006	0.0293
TNF-α	60	42.2 ± 1.462	0.0039	0.0058
TNF-α/Tβ4	60	16.5 ± 1.650	0.0039	0.0058
TNF-α	180	6.3 ± 0.736	0.0185	0.0041
TNF-α/Tβ4	180	3.6 ± 0.907	0.1033	0.0041
TNF-α	360	4.0 ± 0.376	0.0157	0.1967
TNF-α/Tβ4	360	2.5 ± 0.473	0.0866	0.1967

Real-time PCR. Data are expressed as fold change compared to medium control set to 1.0. Data are derived from 3 independent experiments.

We next determined whether the inhibitory effect of Tβ₄ on IL-8 mRNA levels was related to its suppression of NF-κB-mediated IL-8 promoter activation and its G-actin-sequestering property. Comparing luciferase activities from both the wild-type IL-8 promoter and the IL-8 promoter with a mutated κB site, our results demonstrate that the inhibitory effects of Tβ₄ on TNF-α-stimulated IL-8 promoter activation are NF-κB-binding dependent but are not related to Tβ₄ interaction with G-actin (Fig. 6A).

Figure 6.

Tβ₄ interferes with the targeting of NF-κB subunit RelA/p65 onto the proximal region of IL-8 promoter. A) Inset: IL-8 promoter luciferase reporter (IL8-P₁₃₅-LUC) is composed of a 180-bp sequence containing the proximal promoter region from the IL-8 gene and a cDNA fragment encoding the firefly luciferase enzyme. Mutated luciferase reporter (IL8-P135-ΔNF-κB-LUC) contains a mutated κB site in the proximal region and cannot be targeted by RelA/p65. HCET cells were transiently transfected with either IL-8 promoter luciferase reporter or the mutated reporter. After 18 h of treatment with TNF-α (10 ng/ml), cell lysates were used for luciferase activity analysis. Luciferase activity (arbitrary units) is represented as means ± se. Experiments were repeated independently ≥3 times with similar results. B) ChIP analysis of targeting of endogenous RelA/p65 into the 7 kb of genomic DNA sequences containing the IL-8 gene. a) IL-8 proximal promoter region (−146 to +62 bp) with 3 other control regions was detected for the potential occupation by RelA/p65 with or without TNF-α treatment. HCET cells were treated with TNF-α (10 ng/ml) for 1 h, and DNA isolated from heat-reversed crosslinked chromatin fragments was used as PCR templates. b, c) PCR products were analyzed using agarose gel electrophoresis. Oligo primers were designed and used for PCR that specifically targeted the proximal promoter regions (b) and other control regions (c). Signals of input control DNA represent 0.5% of cell lysate. The results are representative of 3 independent experiments.

Lastly, we used ChIP to investigate whether Tβ₄ interferes with RelA/p65 binding to the endogenous IL-8 promoter after TNF-α stimulation (Fig. 6B). We compared RelA/p65 localization in the IL-8-promoter κB site to various upstream and downstream control sites in response to TNF-α treatment. TNF-α significantly promoted RelA/p65 binding to the κB site of the proximal IL-8-promoter region. Tβ₄ treatment abolished this TNF-α-mediated binding activity. These results provide the first evidence that the molecular mechanism of how Tβ₄ suppresses TNF-α-promoted IL-8 production is through blocking RelA/p65 targeting to the IL-8-promoter κB site.

DISCUSSION

Previously, we demonstrated that in addition to promoting corneal wound healing, topical Tβ₄ treatment also potently blocked corneal PMN infiltration and decreased the expression of key proinflammatory proteins after injury (14–16, 25). Mechanistically, Tβ₄ inhibited TNF-α-mediated NF-κB nuclear translocation and activation (30). Despite advances in characterizing the anti-inflammatory properties of Tβ₄, the potential functional targets of Tβ₄ remain unknown. Here, we focused on elucidating the molecular mechanisms of how Tβ₄ interferes with NF-κB activity in response to TNF-α treatment, as well as on its downstream effects on the expression of the proinflammatory IL-8 gene. Further, we explored the potential coordination between Tβ₄ and its intracellular binding partners, PINCH-1 and ILK, in modulating NF-κB activity.

Our findings can be summarized as follows: Tβ₄ colocalizes with RelA/p65 in both resting and TNF-α-stimulated cells, and its overexpression interferes with TNF-α-induced RelA/p65 nuclear translocation and DNA binding activity; overexpression of Tβ₄, but not profilin-1, inhibits TNF-α-induced NF-κB activation; Tβ₄ suppresses TNF-α-stimulated IL-8 mRNA and protein production by inhibiting gene promoter activation, as well as by blocking RelA/p65 targeting to its cognate κB site in the proximal region of the IL-8 gene promoter; overexpression of PINCH-1 and ILK enhances NF-κB activity and sensitizes the activated NF-κB responses to TNF-α-mediated stimulation, while overexpression of Tβ₄ compromises these events; and the ability of Tβ₄ to inhibit TNF-α-stimulated NF-κB binding activity and the sensitizing effects by PINCH-1/ILK, as well as IL-8 promoter activation, is independent of its G-actin-sequestering function. These original findings define the potential clinical therapeutic applications of Tβ₄ as an effective inhibitor of TNF-α-mediated inflammatory diseases. Moreover, as the clinical therapeutic potential of Tβ₄ is rapidly expanding in many diverse areas, such as ocular and dermal wound healing, cardiac and neural tissue regeneration after myocardial infarction, stroke, and multiple sclerosis, we believe that the regulatory effect of Tβ₄ on NF-κB activity may provide a new mechanistic platform for the its beneficial actions in these diverse pathologies (30, 63, 64).

Tβ₄ is a major intracellular G-actin-sequestering protein with novel anti-inflammatory capabilities, which localizes in both the cytoplasm and nuclei of cornea epithelial cells (3, 64). A main priority of this study was to determine whether the immunosuppresive function of Tβ₄ depends on its G-actin binding activity. The function of cytoplasmic actin and actin-binding proteins, including Tβ₄, have been well characterized in determining multiple cellular processes, such as cell adhesion, migration, endocytosis, organelle trafficking, cytokinesis, and apoptosis (65). Recent progress in the field has demonstrated the relevance of actin and actin-binding proteins in nuclear function by modulating gene transcription, chromatin remodeling, and signal transduction (66, 67). Therefore, this functional diversity has demonstrated that actin and actin-binding proteins, previously thought of solely as static cell-building materials, are novel regulatory molecules with capabilities of bridging intracellular signaling and modulating gene expression profiles. Recent studies provide new insights, suggesting that G-actin-sequestering proteins play novel functions in modulating cellular immuno-responses. Plasma Tβ₄ and another G-actin-binding protein, gelsolin, act as “actin scavenging” agents in counteracting actin toxicity and inflammatory responses in bacterial endotoxin-induced sepsis (13, 27). In vivo studies showed that attenuated expression of profilin-1 reduced LDL-induced inflammation and, therefore, prevented atheroma-related lesions in the mouse aorta (68).

Interestingly, emerging evidence also links actin and actin-binding protein to interacting with the NF-κB subunit RelA/p65 and thereby controlling cellular inflammatory responses to various environmental insults. Cellular actin and the actin-binding protein, α-actinin-4, were identified physically to interact with RelA/p65 (44). Functionally, the capabilities of RelA/p65 to shuttle between the cytoplasm and nucleus and to specifically bind to its cognate κB sites are modulated by actin-targeting compounds (50), and overexpression of actin dynamics-regulating proteins, such as the small RhoGTPases and its downstream target cofilin-1 (50). However, other studies imply that the actin-binding properties of Tβ₄ may not contribute to its anti-inflammatory function. Specifically, the modifications in the Tβ₄ molecule, such as oxidation at the Met-6 residue, as well as the alternative splicing variant containing an extra 6 or 7 aa and more methionines at the N-terminal, enhance its immunosuppressive functions (24). Considering that this sulfoxidation can abolish Tβ₄ actin-binding capacity, these findings may hint to the potential functional separation of Tβ₄ G-actin-sequestering and immunosuppressive properties (69).

By using both mutated Tβ₄ lacking actin-binding capacity and the G-actin-sequestering chemical, LatA, we demonstrate that the inhibitory effects of Tβ₄ on TNF-α-induced RelA/p65 and IL-8 promoter activation are not related to its actin-binding capacity. The hexapeptide motif ₁₇KLKKTET₂₃ of Tβ₄ is now accepted to be the major actin-binding domain and shows similar functional effects to Tβ₄ in promoting angiogenesis, cell migration, and exocytosis (70, 71). Cells transfected with the mutant Tβ_4M7A showed similar actin cytoskeleton morphology without disturbing actin polymerization when compared to wild-type Tβ₄ (12). Utilizing this mutant to separate Tβ₄ inhibitory effects on RelA/p65 activation from its G-actin-sequestering property, we demonstrated that the mutant Tβ_4M7A exhibited an even stronger inhibition on NF-κB binding activity compared to wild-type Tβ₄. We also confirmed this mechanistic function by Tβ₄ using LatA, which was identified specifically to target Tβ_4, but not profilin, from G-actin (58). Prior studies showed that LatA successfully abolished Tβ₄'s G-actin-binding function in determining the phenotype and invasiveness of transformed fibroblasts (72). Similar to our findings, treatment by LatA showed a stronger inhibition on endotoxin-induced RelA/p65 nuclear translocation and proinflammatory ICAM-1 gene expression when compared to other actin-targeting modulators (73). Taken together, the unique function on NF-κB activity by the mutant Tβ_4M7A and by LatA suggests that releasing Tβ₄ from bound G-actin may favor the formation of a Tβ₄-RelA/p65 complex, thereby enhancing Tβ₄-mediated inhibition of NF-κB activity after TNF-α stimulation. Future studies should be directed to evaluate quantitatively the competitive binding dynamics of Tβ₄ and the mutant Tβ_4M7A to G-actin or to RelA/p65, as well as with and without LatA treatment.

We next defined the potential coordination by Tβ₄ with its intracellular binding partners in suppressing NF-κB activity, especially the ARs and LIM domain containing proteins PINCH-1 and ILK. Similar to the IκB molecule, other AR proteins, such as NF-κB1 (p105) and B2 (p100), RelA-associated inhibitor (RAI), gankyrin, and nucling, are well characterized and play key inhibitory roles by directly targeting the NF-κB molecule (40, 41, 74–77). For example, the 6 or 7 ARs in IκB proteins facilitate binding to the NF-κB molecule and inhibit NF-κB function by blocking nuclear localization and IκBα binding to the NF-κB dimer and reducing IκB degradation (40, 41).

In addition to AR-containing proteins, the LIM-domain protein PDLIM2 represents another mechanistic system to terminate activated NF-κB by targeting, sequestering, and degrading intranuclear RelA/p65. PDLIM2, an ubiquitin E3 ligase, specifically targets intranuclear p65 and promotes p65 polyubiquitination through its LIM domain. PDLIM2-deficient mice suffer from an uncontrolled NF-κB-mediated immune response (42). Since NF-κB inhibition is linked to AR and to LIM-domain proteins and these domains are found in the Tβ₄ intracellular binding partners, PINCH-1 and ILK, we hypothesized that Tβ₄ may recruit PINCH-1 and ILK to interfere with NF-κB function through their AR and/or LIM domains. Unexpectedly, our data showed that the enforced expression of PINCH-1 and ILK enhanced NF-κB activation and sensitized the response to TNF-α-stimulation. Interestingly, all of these elevated NF-κB activities by ILK and PINCH-1 could be abolished by overexpressing Tβ₄. Further, functional domain analysis indicated that LIM1–4 domains of the PINCH-1 molecule, as well as the ILK C-terminal kinase-domain, enhance their effects on NF-κB activation. Conversely, the fragments of LIM1–2 domains in PINCH-1 attenuated TNF-α-mediated NF-κB activation. Also, this stimulatory effect by TNF-α is completely abolished by the N-terminal AR fragment in ILK. Interestingly, the compromised function of enforced expressed Tβ₄ on the sensitizing effects of NF-κB activation by PINCH/ILK is not dependent on their intracellular interaction.

Our data indicate that there may be binding between complexes of Tβ₄-actin and Tβ₄-RelA/p65 and that similar binding competition may also exist between Tβ₄-ILK-PINCH and Tβ₄-RelA/p65. This competition model is supported by others who showed that the selective binding by Tβ₄ to ILK but not G-actin initiated cell signaling to promote cell mobility (78). Similarly, the specific binding of Rsu-1, a RAS suppressor, to the ILK-PINCH complex affects the transformed phenotype of cells and their invading ability (79). Future studies aimed at dissecting the interactions among Tβ₄, G-actin, RelA/p65, and PINCH-ILK should elucidate the mechanisms connecting the NF-κB transcriptional regulatory network to the cytoskeleton and integrin adhesion complex. These studies will facilitate translation of our basic scientific findings with Tβ₄ into safe and effective therapeutic regimens for the treatment of inflammatory mediated disorders.