Inhibition of NUCKS Facilitates Corneal Recovery Following Alkali Burn

Corneal wound healing involves a complex cascade of cytokine-controlled cellular events, including inflammatory and angiogenesis responses that are regulated by transcriptional chromatin remodeling. Nuclear Ubiquitous Casein and cyclin-dependent Kinase Substrate (NUCKS) is a key chromatin modifier and transcriptional regulator of metabolic signaling. In this study, we investigated the role of NUCKS in corneal wound healing by comparing its effects on corneal alkali burn in NUCKS knockout (NKO) and NUCKS wild-type (NWT) mice. Our data showed that following alkali-injury, inhibition of NUCKS (NKO) accelerated ocular resurfacing and suppressed neovascularization; the cytokine profile of alkali burned corneas in NKO mice showed suppressed expression of inflammation cytokines (IL1A & IL1B); upregulated expression of antiangiogenic factor (Pigment Epithelium-derived Factor; PEDF); and downregulated expression of angiogenic factor (Vascular Endothelial Growth Factor, VEGF); in vitro, following LPS-induced NFκB activation, NKO corneal cells showed reduced expression of IL6, IP10 and TNFα. In vitro, corneal epithelial cells showed reduced NF-κb activation on silencing of NUCKS and corresponding NFκB-mediated cytokine expression was reduced. Here, we illustrate that inhibition of NUCKS played a role in cytokine modulation and facilitated corneal recovery. This reveals a potential new effective strategy for ocular burn treatment.

factors (IL1A and IL1B) and angiogenic factor (VEGF) and anti-angiogenic factor (PEDF). Our in vitro intracellular data revealed that upon stimulation with lipopolysaccharide (LPS; LPS-induced-NFκ B activation), NKO group showed reduced expression on IL6, IP10 and TNFα compared with NWT group. In addition, silencing of NUCKS and stimulation with LPS resulted in reduced NFκ B signaling activation and reduced expression of cytokines downstream of the NFκ B pathway.

Inhibition of NUCKS Accelerates Corneal Resurfacing Following Alkali Injury in Vivo.
Genotypes of NWT and NKO mice were confirmed as described previously (Supplementary Figure S1) 13 . Alkaline injury was induced within a confined circular area ~2-mm in diameter in the central corneal region in NWT and NKO mice. The effects of inhibition of NUCKS on corneal wound healing following alkali injury were investigated. Bright field and fluorescein images of the corneal healing process were recorded at specific post injury time points (day 0, 3, 7 and 14) (Fig. 1). Using the slit lamp system, a slit beam illumination was generated for the examination of oedema and corneal opacity. Our bright field data showed that faster recovery was observed in the NKO group than the NWT group. On day 14, corneas of the NKO group were transparent whereas those of the NWT group still showed oedema (Fig. 1A, upper panel of NKO; no oedema; upper panel of NWT; ***severe oedema; *oedema). Green fluorescein staining was used in the slit lamp examination, corneal defeat stained green (the (A) Bright field images (upper rows of NWT and NKO columns) and fluorescein images (lower rows of NWT and NKO columns) were captured by slit lamp biomicroscopy. Representative bright field images are shown for NWT (i, iii, vii, xi and xv) and NWT mice and (ii, v, ix, xiii, xvii) at specific time points (pre-injury, days 0, 3, 7 and 14). NKO mice exhibited less severe oedema than NWT mice at the end point, day 14. (NWT*oedema, N = 4; NKO no oedema, N = 4; Scale bars: 2 mm). Representative fluorescein images are shown for NWT (iv, viii, xii and xvi) and NKO mice (vi, x, xiv and xviii). Corneal defeat recovery in NKO mice was significantly faster than that for NWT mice at the end point, post injury day 14. (B) Percentage of retained corneal defeat was calculated as follows: the area of defeat retained at a specific time point divided by the area of defeat measured on post injury day 0 (initial point following injury). On day 3, NWT mice showed 96.56 ± 3.89% and NKO mice 27.75 ± 6.61% defeat, indicating that NKO mice achieved better initial healing. As early as day 7, NKO mice showed no defeats 0 ± 0% compared with NWT mice who showed 84.03 ± 9.67% defeat. At the end point on day 14, NWT mice retained 50.18 ± 7.05% defeat and NKO mice exhibited no defeats 0 ± 0%. (*P < 0.01; NWT N = 4; NKO N = 4; Scale bars: 2 mm). defeat at the time immediate after injury, day 0, was defined as 100%) (Fig. 1A). Corneal recovery was notably faster in the NKO group compared with the NWT group. NWT corneas showed positive staining (corneal defeats) on days 3, 7 and 14 (Fig. 1A, lower panel in NWT). In contrast, NKO corneas showed minimal staining on day 3 and no stain was found (stained negative) on day 7 (Fig. 1A, lower panel in NKO). The percentage of retained defeat for NWT and NKO mice respectively was 100 ± 0% and 100 ± 0% on day 0; 96.56 ± 3.89% and 27.75 ± 6.61% on day 3; 84.03 ± 9.67% and 0 ± 0% % on day 7 and 50.18 ± 7.05% vs 0 ± 0% on day 14 ( Fig. 1B; *P < 0.01; NKO vs NWT).

Inhibition of NUCKS Promotes Healing in Vitro. Corneal epithelial marker cytokeratin K3 (K3) and
conjunctival epithelial cell marker cytokeratin K19 (K19) were used to confirm the identity of the cultures isolated from the corneas of NWT and NKO groups. Our data showed that both cultures were positively stained for K3 (~100%), but negatively stained for K19 ( Fig. 2A and B respectively), indicating that both of the isolated cell cultures were corneal epithelial cells.
To examine the effects of inhibition of NUCKS on cell migration, a classic scratch wound healing assay 14 was performed. By scraping the cultures (NWT and NKO) with a 10 μ l pipette tip, a cell-free space was generated (Fig. 3A). Our data demonstrated faster wound closure in the NKO group. The percentage of wound retained for NWT and NKO mice respectively was 100 ± 0% and 100 ± 0% at 0 hour; 79.44% ± 3.5 and 66.75 ± 4.5% at 8 hours; 74.80 ± 3.5% and 49.20 ± 2.8% at 12 hours and 58.82 ± 2.2% and 0.03 ± 0.5% at 24 hours (*P < 0.01; NKO vs NWT).
Inhibition of NUCKS Reduced the Expression of Inflammatory Cytokines, IL1A &IL1B, Following Alkali Injury in Vivo. Following corneal alkali injury, corneas were harvested at various time points from NWT and NKO groups. Since IL1 is a master gene that regulates inflammation and other diverse cellular responses in corneal recovery 2 and the expression of IL1A and IL1B peaks at day 7 post alkali injury 15,16 , we chose to study IL1A and IL1B at day 7. Our data also showed full recovery of NKO corneas by day 14 that was consequently taken as the end-point in our cytokine study. The relative level of expression of NKO was calculated as follows: all the expression levels in NWT and NKO mice were normalized to the corresponding expression levels of GAPDH. The relative level of expression at a particular time point for NKO mice was taken as the ratio of expression in NKO mice compared with that in NWT mice at that particular time point. The baseline level pre-injury, which was around 1, indicated that expression in NKO mice was similar to that of NWT mice at the initial time point (normal condition without injury); on the contrary at subsequent time points a lower relative expression than baseline in NKO mice indicated downregulation of cytokine expression in NKO, and vice versa. Our data showed that in NKO mice. the expression of IL1A and IL1B was reduced (reduced by 3.33 fold and 2.34 fold respectively for IL1A and IL1B; **P < 0.001) on day 7 compared to NWT mice. On day 14, the expression level of NKO for IL1A and IL1B returned to baseline level ( Fig. 4A and B). Taken together, inhibition of NUCKS is associated with precise regulated expression of inflammatory cytokines, IL1A and IL1B, following injury.

Inhibition of NUCKS Reduced Angiogenic Response Following Alkali Injury in Vivo.
The effect of NUCKS deletion on corneal neovascularization in vivo was analyzed. Alkali was administered to the central region of the corneas in NWT and NKO mice and the neovascularization process monitored for 14 days (at the time point: post injury days 0, 7, and 14). The total area of blood vessels was recorded and analyzed with software 'Image Processing and Analysis in Java' (Image J; Wayne Rasband, National Institute of Mental Health, Bethesda, Maryland, USA) for both the NWT and NKO groups (Fig. 5A). Our data showed that a reduced corneal angiogenic response was observed in the NKO group compared with the NWT group. In the NWT group, a greater response of corneal angiogenesis was observed on days 7 to 14; on the contrary in the NKO group corneal angiogenesis was reduced on day 7 and gradually diminished on day 14 (Fig. 5A). The mean area of corneal neovascularization for NWT and NKO mice was 1.017 ± 0.124 mm 2 and 0.466 ± 0.125 mm 2 respectively on day 7; and 0.868 ± 0.066 mm 2 and 0.341 ± 0.043 mm 2 on day 14 (*P < 0.01; Fig. 5B).

Inhibition of NUCKS Downregulated Expression of Angiogenic VEGF and Upregulated Expression of Anti-Angiogenic Factors (PEDF) Following Alkali Injury in Vivo.
We investigated the role of NKO in the angiogenic response of the cornea following alkali-burn by comparing the particular responses in NKO and NWT mice. Corneas were harvested at particular time points following injury, and RNA extracted, converted for cDNA and analyzed using real time RT-PCR. The expression of anti-angiogenic and angiogenic cytokines, PEDF and VEGF, respectively, was examined. As for our previous data in Fig. 1, NKO mice showed full recovery on day 14, thus day 14 was set as the end point. The relative level of expression by NKO mice was calculated as follows: all expression levels of NWT and NKO mice were normalized to the corresponding expression level of GAPDH. The relative level of expression at a particular time point for NKO mice was taken as the ratio of expression of NKO mice compared with that of NWT mice at that particular time point. The baseline level pre-injury, which was around 1, indicated that expression by NKO mice was similar to that of NWT mice at the initial time point (normal condition without injury); at subsequent time points, there was a lower relative expression by NKO mice compared with baseline indicating downregulation of cytokine expression. In NKO mice, mRNA expression of PEDF was upregulated (2.63 fold; **P < 0.001) in corneas on day 7, and returned to the basal level on day 14 ( Fig. 6A) compared to NWT mice. In addition, the level of expression of VEGF remained the same as baseline on day 7 and was downregulated on day 14 (1.86 fold; **P < 0.001) in NKO mice corneas (Fig. 6B) compared to NWT mice. Our data demonstrate that inhibition of NUCKS precisely regulated the expression level of antiangiogenic and angiogenic factors to enable good recovery of injured corneas.
NUCKS regulates NF-κB Activation. The relationship between NFκ B and NUCKS was examined in vitro. Corneal epithelial cells were transfected with luciferase construct that reflects the cellular activities of the corresponding signaling pathways. NUCKS-overexpressed cells, compared with null-overexpressed cells (control), showed enhanced NFκ B activities upon TNF activation (Fig. 7A). Similarly, overexpression of NUCKS also enhanced lipopolysaccharide (LPS)-induced NFκ B activation (Fig. 7Bi) and LPS-induced-interferon-stimulated    (Fig. 7Diii). We also found that upon silencing of NUCKS and LPS stimulation, the phosphorylation of IKK, p65 proteins and Iκ Bα (PIκ B; a key for the activation NFκ B cascade) was impaired 45 mins post-LPS stimulation until the end point of 120 minutes (Fig. 7E). With the above data, we demonstrated that NUCKS regulated NFκ B and ISRE activation, the expression of the particular NFκ B-mediated cytokines and the phosphorylation of particular proteins critical for the signaling pathway in ocular epithelial cells.  Fig. 8Ai and ii, a control experiment was performed with the NFκ B inhibitor, SC154, and corneal epithelial cells in vitro. Expression of PIκ B and TNFα (Fig. 8Bi-iii) was tested on LPS-treated or LPS + SC154 treated corneal epithelial cells. The NKO mice showed similar effects with the NFκ B inhibitor, SC-154, demonstrating reduced expression of PIκ B and TNFα . Next, we investigated the effects of inhibition of NKO on expression of inflammatory cytokines (IL-6, IL-10, INFγ , IP10, TNFα and IL-12), upon LPS-induced NFκ B activation. The NKO group showed reduced expression of various pro-inflammatory cytokines (IL6 17 , IP10 18 and TNFα 19 ) compared with the NWT group ( Fig. 8Ci and ii). (Full-length western blot images are presented in Supplementary Figure S2).

Discussion
Cellular events, such as cell movement, apoptosis and cell proliferation, followed by scar formation are critical to the corneal recovery process 1,2 . Each specific cellular check-point in the recovery process is precisely maintained by a balance of the specific cytokines that are expressed in a timely manner and facilitate the delicate process of corneal healing 2 . Among the various pathways involved in controlling cytokine regulation, the NFκ B signaling pathway is widely recognized and studied. NFκ B is a principal transcription factor that controls various biological processes, including corneal wound healing, inflammation 20 , angiogenesis, apoptosis and the stress response [21][22][23][24][25] . NFκ B is also thought to be involved in several ocular surface disorders, including chemical injury, microbial infections, dry eye, pterygium, ultraviolet radiation-induced injury and corneal graft rejection 26 . We have reported previously that a regulator of NFκ B, RAP1 4,27 , plays a role in angiogenesis and the inflammatory response 28 and corneal recovery is enhanced via deletion of RAP1. Recently, genome-wide ChIP sequencing proved that NUCKS could bind more than 1000 genes involved in inflammation 12 . It is possible that NUCKS, as a transcription factor, is involved in the regulation of NFκ B activation, the release of NFκ B-mediated-cytokine and the particular cellular events driven by these particular cytokines for inflammation and angiogenesis, thus resulting in an optimal clinical recovery of burnt corneas 29,30 .
Our data demonstrated in vivo that deletion of NUCKS facilitated corneal resurfacing and reduced angiogenic responses following alkali burn. Furthermore, we showed that inhibition of NUCKS in vivo modulated inflammatory and angiogenic cytokines during the wound healing process. On LPS-induced NFκ B activation of isolated corneal epithelial cells from NWT and NKO mice, NKO mice showed reduced expression of PIκ B and inflammatory cytokines in vitro compared with the NWT group. Upon either stimulation of TNF, LPS or polyIC, overexpression of NUCKS enhanced NFκ B and ISRE activation, whereas silencing of NUCKS reduced expression of NFκ B-mediated cytokines (TNF, IL8, MCP1) and reduced the phosphorylation of IKK2, P65 and Iκ Bα in vitro in corneal epithelial cells. Others have reported that NUCKS is a key chromatin modifier and transcriptional regulator of a number of signaling pathways and is involved in regulation of chronic inflammation in metabolic syndrome 9,10 . NUCKS also plays a role in cell protection against undesirable factors 31 and cell proliferation, apoptosis and cell movement 8 . In this study, our data provided further evidence that NUCKS enhanced NFκ B activation, and modulated the expression of NFκ B-mediated cytokines, regulated NFκ B-mediated phosphorylation of IKK2, P65 and Iκ Bα and modulated the various specific inflammatory and angiogenic cytokines.
In our model of alkali injury in NKO mice, expression of the inflammatory cytokines, IL1A and IL1B was downregulated. IL1A is the master gene in wound healing and released first for epithelialization and in response to inflammation 1,32 . Meanwhile IL1B plays a role in various cellular events such as fibrosis, cell proliferation, apoptosis, and differentiation 33,34 . Our data on the downregulation of these inflammatory cytokines, IL1A and IL1B, were shown to follow the fine healing process in vivo. Similar mechanisms of precise regulation of inflammatory cytokines in good recovery have also been reported previously 35,36 .
Angiogenesis is another crucial factor during the healing process. The fine balance of angiogenic and anti-angiogenic cytokines (eg. VEGF and PEDF) is important for the survival, apoptosis of corneal endothelial cells 2,5,6 . In our study, PEDF was highly upregulated on day 7 and returned to baseline on day 14; expression of VEGF was repressed on day 14. A synchronized expression pattern was demonstrated during the healing process in the NKO group in vivo. Our data demonstrated a precise regulation of the expression of specific inflammatory and angiogenic cytokines in the corneal recovery process following inhibition of NUCKS. This is consistent with our previous report on RAP1, a NFκ B regulator, that controls cytokine homeostasis and that is critical for fine control of various cellular events in corneal healing 28 . Here, our data showed a similar mechanism of cytokine-modulated recovery by controlled activation of NFκ B, similar to previous findings 4, 28,35,36 .
In summary, this is a first study to describe the role of NUCKS in corneal wound healing, inflammation and angiogenesis and links NUCKS to the activation of NF-κ B and modulation of various cytokines. Our findings shed light on NUCKS (silencing of NUCKS) as a new therapeutic target for fine recovery of the injured cornea.

Materials and Methods
The Transgenic Mouse Lines NWT and NKO. All

Corneal Angiogenic Response Following Alkali Injury in NWT and NKO Mice (in Vivo).
(1) To investigate the degree of angiogenic response in corneal healing, bright view microphotography was performed (NikonSM2800) (Chinetek Scientific Cat No. Infinity 1-3C) 28    Briefly, corneal epithelial cells in culture were detached using Accutase ® Solution (Sigma Cat. No. A6964), then washed with sorting buffer (PBS with 1% FBS/BSA and 2 mM EDTA, keep 4 °C). Cells (10 6 -10 7 ) were fixed, washed, resuspended again in 100 ul sorting buffer and incubated for 15 mins at room temperature with specific antibodies. Cells were then washed twice with sorting buffer and suspended (cell density = 10 6 cells/ml) for flow cytometry analysis 38  Statistical Analysis. The mean value of the samples belonging to either NWT or NKO groups are presented ± SD; unpaired student's t-test was used to determine whether a difference existed between two groups. P-values ≤ 0.05 were considered statistically significant. The tests were two-tailed. Standard deviation (± s.d.) was applied for the error bars.