An anti-TNF-α antibody mimetic to treat ocular inflammation

Infliximab is an antibody that neutralizes TNF-α and is used principally by systemic administration to treat many inflammatory disorders. We prepared the antibody mimetic Fab-PEG-Fab (FpFinfliximab) for direct intravitreal injection to assess whether such formulations have biological activity and potential utility for ocular use. FpFinfliximab was designed to address side effects caused by antibody degradation and the presence of the Fc region. Surface plasmon resonance analysis indicated that infliximab and FpFinfliximab maintained binding affinity for both human and murine recombinant TNF-α. No Fc mediated RPE cellular uptake was observed for FpFinfliximab. Both Infliximab and FpFinfliximab suppressed ocular inflammation by reducing the number of CD45+ infiltrate cells in the EAU mice after a single intravitreal injection at the onset of peak disease. These results offer an opportunity to develop and formulate for ocular use, FpF molecules designed for single and potentially multiple targets using bi-specific FpFs.


Results
Intravitreal administration of Infliximab suppresses EAU. Infliximab is an IgG 1 , chimeric monoclonal antibody developed to bind to human TNF-α to inhibit its interaction with TNF-receptors. It was not clear from the literature describing systemic use of infliximab in different preclinical models, whether it can bind to mouse TNF-α and therefore enable us to demonstrate efficacy in the EAU model [16][17][18] . Ocular administration of infliximab has been shown to provide significant retinal and corneal protection in an mouse model of alkali injury 19 , as well as modulating choroidal neovascularization and endotoxin-induced inflammation in rat and rabbit experimental models respectively 20,21 . In the current study we therefore sought to confirm the binding of infliximab to both murine (Fig. 1a) and human ( Figure S1A) recombinant TNF-α by surface plasmon resonance (SPR) prior to conducting an in vivo assessment using the EAU model.
As in previous pre-clinical studies, we then utilised the highly susceptible B10.RIII mouse strain in which the immunising regimen produces consistent moderate disease severity 22 . Using topical endoscopic fundal imaging (TEFI) 23,24 , clinical changes which correlate with significant inflammatory cell tissue infiltrate are evident from around day 10 post immunization. During the acute period these are progressive changes that include a swollen optic nerve, peri-vascular infiltrate and vasculitis, vitritis and retinal detachment. These resolve by day 28 leaving a persistent chorioretinal inflammatory infiltrate for months 23 . In the current experiments, TEFI enabled us to screen and select experimental groups of mice that displayed early signs of disease onset, namely raised optic nerve or early vasculitis.
Groups of selected mice were treated on day 10 with a single intravitreal administration of infliximab (15 μ g/ eye) or control vehicle. The clinical appearance was assessed daily, and the retinal infiltrate examined and enumerated by flow cytometry at 96 hours following dose administration. Clinically at this time point, control treated mice display typical disease features with raised optic disc, vasculitis and choroidal lesions, whereas mice receiving infliximab only exhibited low grade disease (Fig. 1b,c). Flow cytometric analyses of the retinas of the infliximab mice on day 14 display significantly reduced CD45 positive infiltrate (Fig. 1d), reflecting reductions in CD4, CD8, CD11b and Ly6G positive populations ( Figure S2A). Overall, there is a 55% reduction of the total CD45 + infiltrate (with 68%, 62%, 38% & 64% reductions in CD4, CD8, CD11b and Ly6G populations respectively) in mice receiving infliximab compared to control animals.

Synthesis of endotoxin-free infliximab mimetic, FpF
Infliximab . An IgG antibody mimetic called an FpF (Fab-PEG-Fab) was then prepared using the Fabs from infliximab (FpF infliximab ) and reagent 1 (Fig. 2). The robust suppression of clinical disease evident following intravitreal administration of infliximab, provided the rationale to prepare and evaluate FpF infliximab . The Fabs in both an IgG antibody and FpF are at each terminus of a flexible chain that is topologically similar (Fig. 2a,c) 25 . The bis-sulfone moieties in reagent 1 readily undergo elimination to the corresponding mono-sulfone moieties, depicted as PEG-di(mono-sulfone) 2 (Fig. 2b). Conjugation occurs by a sequence of addition-elimination reactions to insert a stable 3-carbon methylene bridge between the two thiols of an accessible disulfide bond in a Fab ( Figure S3) 26 . Site-specific conjugation with the two cysteine thiols from the Fab disulfide occurs in the region where the Fab is normally bound to the hinge in an IgG antibody. The thiol ether bonds in the re-bridged disulfide are more stable than the original disulfide bond.
Using endotoxin free reagents and working under sterile conditions, FpF infliximab was prepared (Fig. 3a). First, infliximab was proteolytically digested with immobilized papain in the presence of cysteine monohydrate (Fig. 3a, lane 2), before elution of the purified Fab infliximab over protein A (lane 3). The Fc fragment appeared at a slightly higher molecular weight on SDS-PAGE than the Fab infliximab due to glycosylation present on the CH2 domain of the Fc fragment. Fab infliximab was then treated with dithiothreitol (DTT) to reduce the accessible disulfide (lane 4).
After removal of DTT by elution over a PD-10 column, the Fab infliximab solution was incubated with PEG di(mono-sulfone) reagent 2 (0.9 equiv.) to obtain FpF infliximab that was then purified by ion-exchange chromatography (lanes 5-6), which was confirmed by silver stain (lane 7). No unconjugated Fab infliximab or the intermediate, (mono-sulfone)-PEG-Fab infliximab , were observed. The endotoxin levels of both the purified Fab infliximab after proteolytic digestion of infliximab and the purified FpF infliximab were less than 0.005 EU/mL (EndoZyme assay). There was no heavy-light chain dissociation or aggregation of the FpF infliximab when stored at 4 °C for 480 days (lane 8).
FpF infliximab displayed slower dissociation rate constant compared to infliximab. SPR analysis demonstrated FpF infliximab displayed concentration dependent binding to human histidine-tag TNFα ( Figure S1B). The kinetic rate constants and relative binding affinities (KD) to human TNF-α were calculated using a 1:1 binding model ratio, showing that FpF infliximab and infliximab share similar K D values (Fig. 3c). While the dissociation (k d ) rate constants was approximately 5 times slower in FpF infliximab compared to infliximab, association rate constant (k a ) appeared to be faster in infliximab. This is consistent for FpFs 25 and other bivalent mimetics 27 that have been evaluated by SPR. The slower dissociation rate constant suggested a tighter interaction between FpFs and TNFα , and potentially improved therapeutic efficacy for FpF infliximab . The SPR analysis confirmed FpF infliximab binding to murine TNF-α and warranted an in vivo assessment starting at a concentration of 100 μ g/mL FpF infliximab (Fig. 3b).
In vitro evaluation of Fc mediated uptake. Antibody uptake into the retinal pigment epithelium (RPE) 28 can occur via neonatal Fc-receptors expressed in RPE cells 29,30 . Uptake of infliximab and FpF infliximab were assessed in vitro using a murine B6RPE-07 cell line 31 in which cell cultures exhibit cobblestone-like morphology similar to primary mouse RPE cultures. The cells also express RPE markers, tight junction and adhesion molecules, and show polarized phenotype when cultured in collagen-coated membrane 31 and have been used as an in vitro RPE model in many recent functional studies [32][33][34][35] . We used RT-qPCR to measure Fcgrt mRNA from samples of cultured B6RPE-07 cells, to confirm expression of the transmembrane domain of the neonatal Fc receptor in this cell line (Fig. S4). Confluent RPE cell monolayers were treated separately with 250 μ g/ml of each molecule, and cell Scientific RepoRts | 6:36905 | DOI: 10.1038/srep36905 lysates examined for presence of anti-human IgG Fab fragment. Western blot analysis indicated that intracellular infliximab was detectable in murine B6RPE-07 cells at 4 and more strongly by 24 hrs. In contrast, incubation of FpF infliximab did not result in any detectable Fab protein indicating no intracellular uptake (Fig. 3d).
FpF Infliximab is efficacious in EAU. Using the same local administration regimen as described for infliximab, immunized mice displaying the initial signs of clinical disease were selected for treatment groups on day 10, with animals receiving a single intravitreal administration of FpF infliximab (15 μ g/eye) or control vehicle. In eyes receiving FpF infliximab only the initial disease changes to the optic disc with some early signs of vasculitis were evident contrasting the expected inflammatory changes and disease severity observed in EAU controls (Fig. 4a,b). At day Resonance (SPR) binding analysis, confirming that infliximab can bind to murine TNF-α using a NTA chip. (b-d) Mice were immunized for EAU and eyes monitored using TEFI from day 10 onward to select experimental mice displaying clinically evident disease. Groups of mice were injected via intravitreal route with 15 μ g infliximab or vehicle control (EAU) on day 10. Eyes were enucleated on day 14, and retinal infiltrate characterized and quantified; (b) Representative TEFI images and (c) combined total disease scores demonstrating the difference in clinical disease progression between treatment groups. In EAU control eyes typical disease progression with signs of raised optic disc, vasculitis and severe inflammation; In infliximab treated eyes, only raised optic disc and initial signs of vasculitis are evident. (d) Graph showing total CD45 + infiltrate numbers from individual eyes. *P < 0.05, **P < 0.005, ***P < 0.0005; Data presented as means ± SEM, and representative of two independent experiments.
Scientific RepoRts | 6:36905 | DOI: 10.1038/srep36905 14, flow cytometric assessment of retinal infiltrate demonstrates an equally significant reduction of CD45 + infiltrate in mice receiving FpF infliximab treatment (Figs 4c and S2B) as compared to infliximab. Overall, there is a 74% reduction of the total CD45+ infiltrate, comprising 69%, 62%, 74% and 82% reductions in CD4, CD8, CD11b and Ly6G populations respectively in mice treated with FpF infliximab compared to control animals. Proteolytical digestion of IgG1 yield the Fabs fragment which are then treated with DTT and then linked to each end of PEG di(mono-sulfone) reagent 2 to make FpF, (b) PEG di(mono-sulfone) reagent 2 is synthesised from PEG di(bis-sulfone) reagent 1 through elimination of toluene sulfinic acid anion (c) The FpF infliximab is synthesized in a way to compare directly to the starting infliximab. This strategy is applicable on different kind of monoclonal IgGs such as human, humanised and chimeric IgGs.  1-7). FpF infliximab stored at 4 °C for 480 days did not display dissociation from the PEG linker or aggregate formation (lane 8), (b) SPR binding analysis demonstrates that FpF infliximab binds murine TNF-α using NTA chip, (c) Kinetic affinity (K D ) and rate constants (k a , k d ) for infliximab and FpF infliximab to recombinant human TNF-α . FpF has similar binding affinity to full IgG, but the dissociation rate constants are 5× slower for FpF infliximab, (d) Whole cell lysates prepared from B6RPE-07 cells treated with infliximab or FpF infliximab , for 4 & 24 hrs, were probed and the presence of the molecules detected using anti-human IgG Fab. As a negative control, cells were incubated with media alone; and as a positive control preparations of either molecule were applied to the lane.

Discussion
There is an ever-growing need to develop targeted therapies for use in the eye. Ideally, therapies are needed that are directed to multiple specific targets. We are developing recombinant-conjugation strategies focused on the development of multifunctional and bispecific proteins that are difficult to make entirely by recombinant means alone. For organ specific delivery such as to the eye, there is also a need to ensure new biologically based medicines remain stable throughout the planned duration of action. In this study we started with an anti-TNF-α FpF to determine its ability to achieve biological efficacy in terms of suppressing retinal infiltrate in a clinically relevant, aggressive model of ocular inflammation. TNF-α is involved in the pathogenesis of inflammation, and it has also important roles in progression of undesirable angiogenesis in the eye in the neovascular age-related macular degeneration (AMD) and proliferative vitreoretinopathy diseases 36,37 . In this proof-of-concept study we demonstrate for the first time that infliximab suppressed the acute phase of inflammation by reducing the number of CD45+ infiltrate cells in the eyes of the EAU mice. In vivo, infliximab and FpF infliximab are equally efficacious as determined by clinical observation while significantly suppressing the infiltration of leukocytes into the retina during peak EAU. This aspect of the study was required before FpF molecules directed to multiple targets can be evaluated. FpF infliximab synthesis from Fabs obtained by the enzymatic digestion of infliximab allowed the direct comparison of infliximab and FpF infliximab . The flexible hydrophilic PEG scaffold replaced the infliximab Fc. The inflammatory reaction after IVT injection of infliximab may be due to Fc related effects or due to antibody degradation or aggregation 38 . An FpF cannot cause localised inflammatory reactions due to the antibody Fc effector functions, which is important when considering treatments for an acute inflammatory condition such as uveitis. Antibody Fc mediated recycling that prolongs the duration of action during systemic use is not critical for ocular pharmacokinetics 39 and may to some extent remove antibodies from within the eye 30, [40][41][42][43] .
Since the Fab has an accessible disulfide near the hinge region, each Fab in an FpF is conjugated in essentially the same region that native Fabs are bound to the hinge region in IgG antibodies. The thiol-ether bonds of a re-bridged disulfide are more stable than the starting disulfide. The PEG scaffold is more stable than the single polypeptide chains of the IgG hinge, which are susceptible to cleavage reactions 44,45 , and the PEG would reduce the propensity for aggregation 46,47 compared to an IgG. It is thought that the potential for increased stability of FpF infliximab will provide opportunities to make easier-to-use high concentration formulations for direct intravitreal ocular administration than is possible with IgG antibodies. This also raises the possibility of a requirement for less frequent injections compared to the standard antibody.
As TNF-α is a non-covalent trimer in circulation, it is not possible to maintain the trimeric structure of immobilised TNF-α after the regeneration step required for each SPR cycle to remove bound antibody. Since more realistic and comparative results are expected when the ligand, rather than the antibody, is immobilised onto a sensor chip 48 we used histidine-tagged TNF-α for immobilisation onto a metal affinity chip (nitrilotriacetic acid; NTA). Each regeneration cycle caused removal of the TNF-α , but it was possible to obtain reproducible results by re-immobilising the same amount of TNF-α at the beginning of each SPR sample cycle.
Infliximab and FpF infliximab both displayed binding to mouse TNF-α at high concentration. Binding of both infliximab and FpF infliximab to human TNF-α was also observed, although binding was achieved at lower concentrations than with mouse. Binding of infliximab to mouse TNF-α was important to establish because there have been conflicting observations about infliximab's binding with mouse TNF-α 16,49 . Although the concentrations needed for binding to mouse TNF-α were higher than for human TNF-α , this level of binding was sufficient to evaluate anti-TNF-α molecules in the EAU model. The binding kinetics suggested that FpF infliximab has comparable affinity to infliximab, but slower association and dissociation rates. Slower association rates may be important in the design of new high affinity therapies that display decreased off-target effect. Slower dissociation rates may also be important to ensure that once bound to a ligand, an antibody does not in effect become a slow releasing depot for that ligand. Certainly there are applications were a decreased dissociation rate may be a viable strategy to increase efficacy by increasing the residence time within the target organ or tissue 50 . A potential disadvantage of decreased dissociation rates would be if a target cell surface receptor had agonist function.
The outer blood-retinal barrier, located between the choroid/Bruch's membrane, exterior to the sub-retinal space and neural retina is formed by the retinal pigment epithelium (RPE), a cell monolayer interconnected by tight junctions. As a selective barrier, the RPE permits intraocular migration of leukocytes as well as a variety of physiological functions including the phagocytosis of shed photoreceptor outer segments and transport of molecules to and from the retina 51,52 . Recent data shows that therapeutic anti-VEGF mAbs (bevacizumab and afilbercept) can be transported intracellularly across RPE cells, implicating these cells as important protagonists in the clearance of antibodies from the sub-retinal space into the choroid 28 . Uptake in the RPE is a general feature of Fc-fragment containing molecules such as IgG 28 , via Fc-receptors expressed in RPE cells 29,30 . Using in vitro RPE cells, no sub-retinal uptake was observed for FpF infliximab . The avoidance of RPE uptake and uptake to other cell types such as retinal endothelium 53 could be important to minimize toxicity for longer acting dosage forms of biologics.
In summary, we have demonstrated that both infliximab and FpF infliximab suppress intraocular inflammation by reducing the number of CD45+ infiltrate cells in the EAU mice model. Clinically, both molecules appeared to be equally effective in modulating the acute inflammatory response that characterizes EAU. FpF infliximab displayed similar binding affinity (K D ) with slower association and dissociation rates compared to infliximab. There was no Fc mediated RPE cellular uptake of FpF infliximab . These characteristics along with a greater potential for FpF stability are being further investigated to determine potential clinical advantages in developing new treatments using the FpF approach that can bind to two different targets.

. Purified
Scientific RepoRts | 6:36905 | DOI: 10.1038/srep36905 Fab infliximab was isolated from the digestion mixture using ET-free protein A midi kit (Generon, Maidenhead, UK) according to the manufacturer's instructions. The endotoxin level of purified Fab inflix preparations was less than 0.005 EU/mL as determined using EndoZyme assay.
Under sterile conditions, Fab infliximab (2.4 mg/mL, 6.0 mg in 2.5 mL PBS, pH 7.3) was incubated with dithiothreitol (DTT) (1.0 mg/mL, 2.5 mg) at ambient temperature without shaking for 30 min. DTT was removed by elution over a new PD-10 column, and the protein was buffer exchanged into the conjugation buffer (20 mM sodium phosphate, 10 mM EDTA, pH 7.4). PEG-di(mono-sulfone) 2 (0.9 eq, 10 kDa) 25 was added (1.08 mg) to the reduced Fab infliximab solution (6.0 mg in 3.3 mL). The endotoxin level of the PEG reagent was determined to be less than 0.005 EU/mL. The solution was incubated at ambient temperature for approximately 3 h without shaking while maintaining sterility. FpF infliximab was purified using a single step HiTrap Macrocap SP cation exchange column (IEC-Macrocap SP, 5.0 mL). The IEC-Macrocap SP column was first washed with NaOH (1 M) and then ET free water before sample loading. After sample loading, the column was first eluted with 100% acetate buffer A for 10 min followed by a 30 min linear gradient using 100% acetate buffer B containing 1.0 M NaCl with a flow-rate of 1.5 mL/min. Fractions (1.5 mL) were analysed by SDS-PAGE and the FpF infliximab fractions pooled. Using viva-spin (Turbo 15, MW cut off 30 kDa), the FpF infliximab was concentrated to 4.0 mg/mL in a total volume of 0.2 mL. The concentration of the purified FpF infliximab was calculated by micro BCA assay using infliximab as standard. The endotoxin level of the concentrated FpFinfliximab was determined to be less than 0.005 EU/mL using the EndoZyme assay according to manufacturer's instructions (Hyglos, Munich, Germany).

Surface Plasmon Resonance (SPR) Binding studies. Sensor chip NTA (nitrilotriacetic acid) is made
of carboxymethylated dextran which is covalently immobilised with NTA. The surface of sensor chip NTA was first activated with nickel (Ni +2 ) solution to create the nickel chelated NTA group which allowed capture of histidine-tagged TNFα . Samples (FpF infliximab or infliximab) were passed over the surface and the binding kinetics were determined. Between runs, the NTA surface was regenerated with EDTA solution to refresh the sensor chip for a new analysis cycle, starting with fresh NiCl 2 solution and histidine-tag TNFα . Conditions to run the kinetic study were as follows; flow rate of 30 μ L/min, 200 second contact time for NiCl 2 solution (0.4 mL), 250 second contact time for human histidine-tag TNFα (5 μ g, 1.0 mL), 150 second contact time for series of infliximab solutions (0.2 mg/mL-0.0625 mg/mL), and FpF Infliximab solutions (0.2 mg/mL-0.0625 mg/mL). The NTA chip was regenerated using EDTA solution (350 mM, 1.0 mL) for 60 seconds.
For local administration of infliximab or FpF infliximab , intravitreal injections were performed on day 10 post-immunization. In brief, the eye was proptosed and held in position with a pair of forceps, while 15 μ g of infliximab or FpF infliximab diluted in 2 μ l PBS was injected using a 33-gauge hypodermic needle (Esslab, Essex, UK). The injection site was treated with chloramphenicol and globe reposited.
EAU Clinical Assessment. Using a method adapted from Paques et al. 54 an endoscope with a 5 cm long teleotoscope of 3 mm outer diameter (1218AA; Karl Storz, Tuttlingen, Germany) was connected to a Nikon D80 digital camera with a 10-million pixel charge-coupled device image sensor and Nikkor AF 85/F1.8 D objective (Nikon, Tokyo, Japan), with an additional + 4.00 dioptre magnifying lens. Through pupils dilated with topical tropicamide 1% and phenylepherine 2.5% (Minims, from Chauvin Pharmaceuticals, UK), and topical oxybropucaine 0.4% (Minims) and Viscotears (Novartis Pharmaceuticals, UK) for corneal anesthesia, images were obtained by direct corneal contact with the endoscope. Images were processed using Adobe Photoshop (Adobe Corporation, Mountain View, CA). Using an adapted clinical grading system, fundal images were scored according to inflammatory changes to the optic disc and retinal vessels in addition to retinal lesions and structural damage 24 . All scores were added together to calculate a final disease score.
Isolation and flow cytometric assessment of retinal infiltrate. Each eye was dissected in 100 μ l ice-cold HBSS with aqueous, vitreous and retina extracted using a limbal incision, lens extraction and transfer into a 1.5 mL tube. The tissue was mechanically dissociated by rapping the tube across a standard rack ten times before transfer into a 96-well 60 μ m cell strainer plate (Merck Millipore, UK). This was centrifuged at 1200 rpm for 5 minutes, the retinal supernatants was aspirated and the remaining cell pellet transferred into a 96-well V-bottom plate for immuno-staining.
Cells were incubated with 24G2 cell supernatant for 10 minutes at 4 °C before incubation with fluorochromeconjugated monoclonal antibodies against cell surface markers including, CD4 [RM4-5], CD8 [53-6.7], CD11b [M1/70], Ly6G [1A8] and CD45 [30-F11] at 4 °C for 20 minutes. Cells were resuspended in 7AAD, and dead cells excluded from analysis by gating on 7AAD negative cells. Cell suspensions were acquired using a 3-laser BD TM LSR-II flow cytometer (BD Cytometry Systems, Oxford, UK). Analysis was performed using FlowJo software (Treestar, San Carlos, California). Cell numbers were calculated by reference to a known cell-standard, as previously reported 55 . Briefly, splenocytes at a range of known cell concentrations were acquired using a fixed and stable flow rate for 1 minute. Based on total cell number acquired during this time, a standard curve was generated and used to interpolate cell concentrations of ocular infiltrating cells acquired at the same flow rate and time.
Cell culture and intracellular uptake assay. A spontaneously transformed mouse retinal pigment epithelium (RPE) cell line B6-RPE07 31 was cultured in DMEM medium supplemented with 10% heat-inactivated fetal calf serum, 2% L-glutamine, 1 mM sodium pyruvate, 60 μ M 2-mercaptoethanol, 100 U/ml penicillin and 100 μ g/ml streptomycin (complete medium) at 37 °C in an atmosphere of 5% CO 2 . RPE cells were passaged with a split ratio of 1:5 using 0.05% trypsin-EDTA (Life Technologies, Paisley, UK), and allowed to recover for 2 days in complete medium prior to experiments.
Confluent monolayers of B6RPE-07 cells in 24-well plates were treated once with 250 μ g infliximab or FpF infliximab for 4 or 24 hrs. Following treatment, B6RPE-07 cells were washed twice with ice-cold PBS. Cells were lysed directly from the plate using 200 μl CelLytic M solution supplemented with protease inhibitor cocktail (both from Sigma), and incubated on ice with gentle shaking for 10 minutes. Finally, these whole cell lysates were centrifuged for 10 minutes, at 4 °C at 13,000 rpm and the supernatant lysate collected and stored at − 80 °C. To ensure equal loadings, a bicinchoninic acid (BCA) protein assay was performed prior to SDS-PAGE, and samples prepared with 5× Laemilli buffer. For Western blot, 10 μ g of cell lysate loaded per well and separated on a 8-16% SDS-PAGE gel (Biorad), and proteins transferred to a nitrocellulose membrane. Following blocking in 5% milk/ TBS/Tween-20, the membrane was subjected to analysis using the anti-human IgG (Fab specific)-peroxidase antibody (1:1000) (A0293; Sigma) and β -actin antibody (Cell Signaling) followed by a HRP conjugated polyclonal anti-mouse IgG (1:2000) before visualization using the chemiluminescent method (GE Healthcare Life Sciences).
Total RNA from B6RPE-07 cells, and ex-vivo mouse retina and corneal tissue was isolated using TRIzol reagent (Life Technologies), treated with RQ1 RNase-free DNase before cDNA synthesis using the ImProm-II TM Reverse Transcription System (Promega, Southampton, UK). cDNA was amplified using the Power SYBR ® Green

Statistical analyses. Data was analyzed with unpaired Student's t-test (GraphPad Prism software, San
Diego, CA). Data are generated as mean ± SEM and representative of at least 2 independent experiments. Values were considered statistically significant at *p < 0.05, **p < 0.005, ***p < 0.0005.