Biosynthesis, Characterization, and Efficacy in Retinal Degenerative Diseases of Lens Epithelium-derived Growth Factor Fragment (LEDGF1–326), a Novel Therapeutic Protein*

Background: There is no FDA-approved drug therapy for blinding retinal degenerative diseases such as retinitis pigmentosa and dry age-related macular degeneration. Results: LEDGF1–326 increased viability of retinal cells subjected to mutant rhodopsin aggregation stress and reduced loss of photoreceptors, thereby improving retinal function. Conclusion: LEDGF1–326 reduced retinal degeneration. Significance: LEDGF 1–326 is a new protein therapeutic to treat vision loss. For vision-threatening retinitis pigmentosa and dry age-related macular degeneration, there are no United States Food and Drug Administration (FDA)-approved treatments. We identified, biosynthesized, purified, and characterized lens epithelium-derived growth factor fragment (LEDGF1–326) as a novel protein therapeutic. LEDGF1–326 was produced at about 20 mg/liter of culture when expressed in the Escherichia coli system, with about 95% purity and aggregate-free homogeneous population with a mean hydrodynamic diameter of 9 ± 1 nm. The free energy of unfolding of LEDGF1–326 was 3.3 ± 0.5 kcal mol−1, and melting temperature was 44.8 ± 0.2 °C. LEDGF1–326 increased human retinal pigment epithelial cell viability from 48.3 ± 5.6 to 119.3 ± 21.1% in the presence of P23H mutant rhodopsin-mediated aggregation stress. LEDGF1–326 also increased retinal pigment epithelial cell FluoSphere uptake to 140 ± 10%. Eight weeks after single intravitreal injection in Royal College of Surgeons (RCS) rats, LEDGF1–326 increased the b-wave amplitude significantly from 9.4 ± 4.6 to 57.6 ± 8.8 μV for scotopic electroretinogram and from 10.9 ± 5.6 to 45.8 ± 15.2 μV for photopic electroretinogram. LEDGF1–326 significantly increased the retinal outer nuclear layer thickness from 6.34 ± 1.6 to 11.7 ± 0.7 μm. LEDGF1–326 is a potential new therapeutic agent for treating retinal degenerative diseases.

acids 1-324) of LEDGF acts as a chromatin tether and binds LEDGF to DNA, whereas the C terminus (amino acids 325-530) binds to the HIV-1 integrase and promotes transcription and replication of HIV (27,28). In vivo administration of C-terminal LEDGF 325-530 potently inhibited HIV replication by competing with endogenous LEDGF for binding to HIV integrase (26). Second, purification of full-length LEDGF in bulk quantities has never been reported, possibly due to inherent instability. Because GST-LEDGF protein degrades to smaller fragments during its biosynthesis and purification (29), attempts were made to stabilize GST-LEDGF with heparin. In the presence of 71 mg/ml heparin in the culture medium, the full-size GST-LEDGF in purified protein fraction increased to 56% from 32% in controls. In a previous study, we identified that gene delivery of LEDGF  , an N-terminal fragment of LEDGF, can reduce P23H rhodopsin aggregation and promote cell survival (30). Due to the absence of the C-terminal domain of LEDGF in LEDGF  , the possibility of HIV-1 integration is expected to be minimized with LEDGF  .
In this study, we cloned, synthesized, and purified LEDGF  protein for the first time in stable and nondegraded form and scaled up its production to large quantities. Further, we established the biophysical properties of LEDGF  and assessed its ability to reduce in vitro P23H rhodopsin aggregation-mediated retinal cell damage. Finally, the ability of LEDGF  to reduce retinal degeneration in the RCS rat model for dry AMD and RP was determined.

Preparations of LEDGF 1-326 DNA Construct
Gene encoding LEDGF  protein was designed to clone into pET-28a (ϩ) vector (Novagen, Madison, WI). Briefly, LEDGF  gene was amplified from the pEGFP-LEDGF plasmid using the forward primer 5Ј-AGTAGTGGATCCAT-GACTCGCGATTTCAAAC-3Ј and reverse primer 5Ј-AAT-AATAAGCTTTCACTGCTCAGTTTCCATTTGTTC-3Ј. Thereafter, the purified LEDGF  gene and pET-28a (ϩ) vector were digested using HindIII and BamHI restriction enzymes and ligated overnight at 4°C using T4 DNA ligase. Competent Escherichia coli DH5␣ cells were transformed with the ligation product as per the manufacturer's protocol. Insertion of LEDGF  in pET-28a (ϩ) vector (named as pLEDGF  ) was confirmed by PCR screening, restriction digest, and finally by DNA sequencing. Purity and the size of the recombinant DNA were analyzed using 2% agarose gel. The colony showing positive PCR signal and correct sequence was cultured further, and the bacterial glycerol stock was made and stored at Ϫ80°C for all future use.

Cloning and Expression of LEDGF 1-326
For protein biosynthesis, pLEDGF 1-326 plasmid was isolated from E. coli DH5␣ colony and transformed in E. coli BL21(DE3) as per the manufacturer's protocol. LEDGF  was expressed and purified in shake flask culture under the controlled addition of isopropyl-␤-D-thiogalactoside. Cells were harvested and lysed, and crude LEDGF  was collected as soluble cell lysate.

Purification of LEDGF 1-326 Using Fast Protein Liquid Chromatography (FPLC)
LEDGF  was solely expressed in soluble fraction as determined by SDS-PAGE. LEDGF  was purified using two-step fast protein liquid chromatography (FPLC), first based on charge (cation exchange) and then based on size (gel filtration). Briefly, cation exchange SP Sepharose beads were packed in XK 16/20 column, and the soluble cell lysate was loaded. The column was washed with buffer A (25 mM Tris-HCl, pH 7.0, 1 mM EDTA, 1 mM PMSF, and 5% sucrose). The nonspecifically and loosely bound impurities were eluted using a gradient of NaCl. Fractions containing high amount of protein were pooled together, dialyzed using dialysis buffer (25 mM Tris, pH 7.0, and 0.1% sucrose), and lyophilized for 48 h. The lyophilized protein was resolubilized in 2 ml of deionized water and further purified using prepackaged S-200 gel filtration column in buffer B (25 mM Tris-HCl, pH 7.0, and 100 mM NaCl). Fractions containing the pure LEDGF 1-326 were pooled together and dialyzed in the dialysis buffer (25 mM Tris-HCl, pH 7.0, and 0.1% sucrose) for 48 h at 4°C with three buffer exchanges. The dialyzed LEDGF 1-326 was lyophilized, aliquoted, and stored at Ϫ80°C for all future purposes.

Physical Characterization
Size Exclusion Chromatography (SEC) HPLC-The lyophilized protein was dissolved in deionized water to final concentration of 500 g/ml and filtered through 0.22-m polyvinylidene difluoride (PVDF) filters. The protein was size-separated in an Agilent Bio SEC-3 column using 25 mM Tris buffer containing 1 mM CaCl 2 , pH 7.0, at 25°C with a flow rate of 1 ml/min. Retention time was averaged from four chromatograms.
Matrix-assisted Laser Desorption/Ionization Time-of-Flight (MALDI-TOF) Mass Spectrometry-Protein homogeneity and molecular weight were confirmed by 4800 Plus MALDI TOF/ TOF TM (AB Sciex, Framingham, MA) mass spectrometry. The protein sample was dissolved into a solution of standard MALDI matrix sinapinic acid, spotted, and dried onto the metal target plate. Data were collected as total ion current from 1000 laser shots of 5900 intensity.
Dynamic Light Scattering (DLS)-The homogeneity and size of the LEDGF 1-326 protein was analyzed using Nano ZS (Malvern, Westborough, MA). Briefly, lyophilized protein sample was dissolved in deionized water to get final protein concentration of 1 mg/ml. The hydrodynamic diameter of LEDGF 1-326 using the dynamic light scattering technique with data collection at 173°backscatter angle was obtained. Measurement was an average of 13 scans. Data represents mean of n ϭ 8.
Biophysical Characterization-For biophysical characterization, the protein was extensively dialyzed in 25 mM phosphate buffer, pH 7, to remove Tris-HCl and sucrose and filtered through a 0.22-m PVDF syringe filter. Spectra obtained were analyzed using either Origin 8.5 (OriginLab Corp., Northampton, MA) or SigmaPlot 11.0 (SYSTAT Software, Inc., Chicago, IL). The data were fitted using Equations 1 and 2, defined by Scholtz et al. (31) as below to determine the ⌬G, m-value, and [urea]1 ⁄ 2 .
where y F°a nd y U°a re the intercepts, m F and m U are the slopes of the pre-and post-transition phase baselines, and the m-value is the slope of the transition phase. ⌬G is the free energy change at any particular urea concentration, and it varies linearly with urea concentration and is used to estimate ⌬G(H 2 O). ⌬G(H 2 O) is defined as the Gibbs free energy of a protein in the absence of urea at 25°C. R is the universal gas constant, and T is the temperature of the sample.
[urea]1 ⁄ 2 is the concentration of urea at which LEDGF 1-326 is 50% unfolded. Data represent mean of duplicate studies.
Circular Dichroism (CD)-To determine the secondary structures of LEDGF 1-326 and its conformational stability parameters, far-ultraviolet CD spectra were recorded. Briefly, a 500 g/ml protein sample was placed in 1-mm quartz cuvette, and spectra were recorded at a scan speed of 0.5 s/time point, step size of 1 nm, and a bandwidth of 4 nm from 200 to 280 nm using a Chirascan CD instrument (Applied Photophysics Ltd.). All scans were done in triplicate. The native LEDGF 1-326 spectrum thus obtained was deconvoluted using CDNN 2.1 software (Dr. Gerald Bohm, Martin-Luther-University at Halle-Wittenberg, Germany) to get the percentage of secondary structures present in native LEDGF 1-326 protein.
LEDGF 1-326 chemical denaturation was performed at various urea concentrations. Briefly, 300 g/ml protein was incubated with 0 -6 M urea in 25 mM phosphate buffer, pH 7.0, for 24 h. CD signal was recorded as mentioned above. The conformational stability parameters of LEDGF 1-326 were determined by plotting the CD signal at 230 nm as a function of urea concentration as we obtained the maximum CD signal difference between the folded and unfolded protein spectrum at this wavelength. Similarly, to investigate the thermal stability of LEDGF 1-326 , 500 g/ml LEDGF 1-326 was subjected to heat denaturation from 25 to 90°C in smooth ramp mode at ramp rate of 1°C/min. Because major changes were seen at 222 nm, the CD signal at this wavelength was used to determine the melting point (T m ).
Fluorescence Spectroscopy-The steady state fluorescence spectroscopy was done to determine the tertiary structure perturbation. The protein sample (final concentration 300 g/ml) was incubated with various concentrations of urea solution (0 -6 M) in 25 mM phosphate buffer, pH 7.0, for 24 h. The intrinsic tryptophan fluorescence spectra were recorded from 300 to 400 nm, at 280-nm excitation wavelength, with an increment of 1 nm using SpectraMax M5 (Molecular Devices, Downingtown, PA). The conformational stability parameters of LEDGF 1-326 were determined by plotting the fluorescence intensity ratio at 340/356 nm as a function of urea concentration. All intensity values were corrected for buffer effects and inner filter effects.

Functional Characterization
In Vitro Assay-ARPE-19 cells were maintained as described earlier (30). For cell viability assay, 10,000 cells were plated in 96-well plate and incubated for 24 h. After 24 h, the serumcontaining medium was aspirated out. The test groups (pP23H-Rhoϩ LEDGF 1-326 ) were transiently transfected with pP23H-Rho plasmid (1 g/ml) using a 1:3 ratio of Lipofectamine 2000 (LP-2000) in serum-free medium as per the manufacturer's protocol. After 6 h of transfection, the medium was aspirated out, and cells were treated with increasing amount of LEDGF  . No cells (just the medium), cells with no LP-2000, and cells with LP-2000 were also maintained as control.
MTT Assay-After 48 h, the medium was aspirated out, and 200 l of fresh serum-free medium was added. 20 l of MTT reagent (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, 5 mg/ml in PBS, pH 7.4) was added to each well, and further incubation was done for 3 h at 37°C. The MTT-containing medium was aspirated out, and the formazan crystals formed were dissolved in 200 l of dimethyl sulfoxide (DMSO). The absorbance of the color developed was measured at 570 nm using a SpectraMax M5. The percentage of viability of groups was calculated with reference to the control group containing cells with no LP-2000. All groups were repeated with n ϭ 4.
Live/Dead Cell Count Assay-ARPE-19 cells were treated similar to MTT assay as stated above. At the end of LEDGF 1-326 treatment period, cells were washed with PBS. The cells were labeled with a combination of plasma membrane permeant (Hoechst 33342), a plasma membrane-impermeable molecule (BOBO TM 3), and a nuclear dye (4Ј,6-diamidino-2-phenylindole, dihydrochloride; DAPI). Hoechst 33342 was used to label cell nuclei, whereas BOBO TM 3 was used to label dying or dead cells. The cells were visualized using the Operetta high content imaging system. Cell count was obtained using the automated software tool in the Operetta instrument. The number and percentage of live cells were calculated by subtracting the dead cell count from the "all cell" count.
Phagocytic Assay-ARPE-19 cells were seeded in 24-well plates and transfected with 20 pM/ml MERTK siRNA (Santa Cruz Biotechnology Inc., Dallas, TX), using siRNA transfecting agent (Santa Cruz Biotechnology) for 6 h. The transfecting medium was removed, and cells were further incubated in serum-free medium for 24 h. Cells transfected only with the transfecting medium and no MERTK siRNA were maintained as control. Cells were washed once and treated with 0.05, 0.5, or 5 g/ml LEDGF 1-326 for 24 h, and then phagocytosis of 2 m particles was monitored. Briefly, 100 g/ml 2-m blue Fluo-Spheres (Life Technologies) was incubated with cells for 3 h.
Thereafter, cells were washed twice with cold PBS, pH 7.4, followed by two washes of cold PBS, pH 5.0, to remove adherent FluoSpheres. Cells were lysed using 1% Triton X, and the fluorescence of the particles in the cell lysate was measured using 350 nm excitation and 430 nm emission. Cells transfected with only transfecting agent without siRNA were taken as control for particle uptake. Cells with no particle treatment were used for background fluorescence measurements.

In Vivo Efficacy Assay
Animal Maintenance-Homozygous RCS rat breeders were generously provided by Dr. Jeffrey Olson (University of Colorado Anschutz Medical Campus, Aurora, CO). Thereafter, the RCS rat colony was maintained in the animal facility of the University of Colorado Anschutz Medical Campus and with the approval of the Institutional Animal Care and Use Committee (IACUC). The experiments were carried as per the Association for Research in Vision and Ophthalmology (ARVO) statement for the Use of the Animals in Ophthalmic and Vision Research.
Electroretinography-At 4 weeks of age, rats were darkadapted for at least 30 min. Thereafter, the animal was prepared for electroretinogram (ERG) under dim red light. Briefly, rat was anesthetized with intraperitoneal injection of a mixture of 80 mg/kg of ketamine and 12 mg/kg of xylazine. The pupil was dilated with 0.5% tropicamide (Akorn, Lake Forest, IL) and was kept moist using 2.5% hypromellose (Akorn). Thereafter, the animal was placed on a heated water jacket stabilized at 37°C. A reference electrode (LKC Technologies Inc., Gaithersburg, MD) was inserted into the tail and cheek of the animal. A DTL Plus electrode (LKC Technologies Inc.) was placed across the cornea of each eye. The animal was exposed to brief flashes of 0.4 log cd-s/m 2 with an interval of 10 s between each flash, and scotopic ERGs were recorded. Thereafter, the animal was lightadapted for 3 min with a background light of 30 cd/m 2 . Photopic ERG was recorded at the same intensity of flash but with background light on. At least three ERGs were averaged to get a single ERG for each animal. 2 l of 0.25, 0.5, or 2.5 mg/ml sterile-filtered LEDGF 1-326 was given intravitreally in one eye, and vehicle was given in the contralateral eye. ERGs were recorded every 2 weeks for 8 weeks after intravitreal injection, i.e. until rats were 12 weeks age. Data are presented for n ϭ 3 or 4.
Histology-At the end of the study, i.e. on the 12th week, eyes were enucleated after ERG measurements and fixed in Davidson's fixative (2% of 37-40% formaldehyde, 35% ethanol, 10% glacial acetic acid, and 53% distilled water) for 24 h at room temperature. The eyes were then stored in 70% ethanol for subsequent serial dehydration and embedment in paraffin. Three vertical sections of 6 m thick were cut from the nasal to the temporal side at the optical nerve (500 m apart) on a standard microtome. Gross retinal morphology was assessed by light microscope following hematoxylin/eosin staining of tissue sections. The thickness of outer nuclear layer (ONL) and inner nuclear layer (INL) was measured methodically using the Aperio ImageScope software version 11.1.2.760. Because the photoreceptor cell protection may be uneven across the retina, every 500 m from the superior edge to the inferior edge in each section was analyzed, and the average of three sections was done for each point. Data represent the average of three eyes.
Immunofluorescence-For immunofluorescence, after removal of paraffin, eye sections were processed through the following sequential steps at room temperature, unless otherwise indicated. Antigen was retrieved by boiling the sections at 80°C for 15 min. After blocking the nonspecific binding, sections were incubated with mouse anti-rhodopsin (1D4) primary antibody at 4°C overnight followed by 30-min incubation with Alexa Fluor 594-conjugated donkey anti-mouse IgG and DAPI. Finally, eye sections were washed and mounted by SuperMount H (Biogenex, San Ramon, CA) mounting medium to prevent rapid loss of fluorescence. The fluorescence was visualized using a confocal microscope (Nikon Eclipse C1) at 20ϫ optical zoom. The excitation-emission wavelengths used for DAPI and Alexa Fluor were 408 -450/35 and 637-605/75 nm, respectively. Images were captured using the Nikon EZ-C1 software version 3.40.

Statistical Analysis
All data are represented as the mean Ϯ S.D. An independent samples Student's t test or one-way analysis of variance followed by Tukey's post hoc analysis (SPSS, version 11.5; SPSS, Chicago, IL) was performed for comparisons between the two or multiple experimental groups, respectively. The differences were considered statistically significant at p Յ 0.05.
Protein estimation indicated that about 20 mg of protein was obtained per liter of the shake flask culture.
Bioinformatics Analysis of LEDGF 1-326 -Bioinformatics analysis of LEDGF 1-326 sequence using SIB ExPASy portal (32) indicated its theoretical molecular mass to be 36.9 kDa. The computed isoelectric point (pI) of LEDGF 1-326 was 9.23, with 73 positively charged (arginine and lysine) and 63 negatively charged (aspartic acid and glutamic acid) amino acid residues. The theoretical molar extinction coefficient was 15,470 M Ϫ1 cm Ϫ1 at 280 nm in water. Based on its N-terminal amino acid methionine, its half-life in mammalian cells was predicted to be 30 h. The N-, O-, and C-glycosylation of LEDGF 1-326 was predicted using NetNGlyc 1.0, NetOGlyc 3.1, and NetCGlyc 1.0 servers, respectively. According to the output of NetNGlyc, LEDGF 1-326 has only one potential site for N-glycosylation at 103 amino acids. However, the probability of this site to be glycosylated is 0.6, indicating low likelihood of N-glycosylation. According to NetOGlyc 3.1 and NetCGlyc 1.0 server outputs, LEDGF 1-326 is unlikely to be O-or C-glycosylated. Thus, bac-terial biosynthesis of protein may not alter LEDGF 1-326 potency due to differences in glycosylation status.
LEDGF  Is Purified to Near Homogeneity as a Random Coiled Protein-The purity of LEDGF 1-326 protein, determined by SEC-HPLC ( Fig. 2A), indicated 95% pure LEDGF 1-326 with a retention time of 10.63 Ϯ 0.06 min. To further investigate whether LEDGF 1-326 self-associates to form any higher molecular weight oligomers, DLS was performed (Fig. 2B). DLS indicated a homogeneous population of LEDGF 1-326 , with a mean hydrodynamic diameter of 9 Ϯ 1 nm.
The molecular weight of LEDGF 1-326 was confirmed by MALDI-TOF mass spectrometry. The major peak obtained in MALDI-TOF spectrum was at 40,314.32 and 80,663.19 m/z (mass to charge) ratio (Fig. 2C). MALDI-TOF indicated that LEDGF 1-326 has a molecular mass of 40.314 kDa, which was equivalent to its theoretical molecular mass, indicating that the protein is not glycosylated. However, a second peak at 80,663 m/z was also seen, which indicated that LEDGF 1-326 may exist as a dimer. To investigate the existence of the dimers, SDS-PAGE of LEDGF 1-326 was run under reducing and nonreducing conditions (Fig. 2D) To investigate the secondary structure of LEDGF 1-326 , farultraviolet CD spectrum of the native LEDGF 1-326 was analyzed (Fig. 2E). The CD signal remained negative from 280 to 200 nm. A very low ellipticity above 210 nm and negative band below 200 nm indicated that LEDGF 1-326 may be primarily composed of random coils (33). To further dissect the secondary structure of LEDGF 1-326 , the CD spectrum was deconvoluted using CDNN 2.1 software ( Table 1). Assuming that the spectrum obtained is the linear combination of the individual spectrum of the component secondary structure elements and the noise due to the aromatic chromophores and prosthetic groups, LEDGF 1-326 was predicted to be 45.1% randomly coiled. The ␤-turn was about 21.2%, and there were 15% parallel ␤-sheets and 16% antiparallel ␤-sheets. The contribution from the ␣-helix was about only 16%. The three-dimensional structure of LEDGF 1-326 native protein was predicted to be predominantly a random coil, using the I-Tasser (Iterative Threading  Assembly Refinement) protein modeling server (Fig. 2F). The LEDGF 1-326 predicted model had the confidence score of Ϫ3.18 and template modeling score of 0.36 Ϯ 0.12, and root mean square deviation was equal to 14.1 Ϯ 8 Å. LEDGF  Is Conformationally Stable-To investigate the conformational stability of LEDGF 1-326 in water, the perturbation in the tertiary structure due to chemical denaturation was measured by the intrinsic fluorescence of tryptophan molecules present in LEDGF   (Fig. 3A). Emission spectrum of native LEDGF 1-326 protein, in absence of urea, had a max at 340 nm and ⌬1 ⁄ 2 (half-width of ⌬) of 56 nm (Fig. 3A, panel i). As the concentration of urea increased from 0 to 5 M, quenching in the fluorescence signal as well as red shift was seen. The signal decreased slowly until 0.9 M urea concentration was reached. Thereafter, there was a sharp decrease in the fluorescence signal until 2.3 M urea concentration was reached. Beyond this concentration, the decrease in the fluorescence signal was minimal. The max of LEDGF 1-326 shifted to 356 nm, and ⌬1 ⁄ 2 was 71 nm at 5 M urea. When the ratio of LEDGF 1-326 fluorescence signal at 340 -356 nm was plotted as a function of urea concentration, a sigmoidal curve was obtained (Fig. 3A, panel ii). Using Equations 1 and 2 (described under "Experimental Procedures"), ⌬G(H 2 O) of LEDGF 1-326 was estimated to be 3.24 Ϯ 0.48 kcal mol Ϫ1 , the m-value was estimated to be 1.70 Ϯ 0.22 kcal mol Ϫ1 M Ϫ1 , and [urea]1 ⁄ 2 was estimated to be 1.81 Ϯ 0.02 M, indicating that LEDGF 1-326 might be a stable protein.
Far-ultraviolet CD spectroscopy was performed to investigate the perturbation in the secondary structures of LEDGF  in the presence of urea (Fig. 3B). The CD signal of the LEDGF 1-326 was traced against the wavelength at each urea concentration (Fig. 3B, panel i). The CD signal continuously became more negative as the concentration of urea was increased. When CD signal at 230 nm was plotted as a function of urea concentration (Fig. 3B, panel ii) Thermal stability of LEDGF 1-326 was determined using farultraviolet CD spectroscopy (Fig. 3C). The CD signal in the presence of heat as a denaturant was measured from 215 to 250 nm (Fig. 3C, panel i). As the temperature of the LEDGF 1-326 solution was increased, the negative dip at about 235 nm increased. The CD signal followed the same pattern as chemical denaturation, a pre-transition phase between ϳ30 and 35°C, followed by a transition phase between ϳ35 and 55°C, followed by a post-transition phase from ϳ55 to 70°C (Fig. 3C, panel ii). When these data were fitted using a global fit analysis equation, the T m (the melting temperature) of LEDGF 1-326 obtained was 44.8 Ϯ 0.2°C, indicating that LEDGF 1-326 will possibly be stable at 25°C (room temperature).
LEDGF  Rescues ARPE-19 Cells from Aggregation-mediated Stress-LEDGF 1-326 efficacy to rescue ARPE-19 cells from protein aggregation-mediated stress was measured by MTT assay (Fig. 4A). Initially, the ability of LEDGF 1-326 to increase the viability of ARPE-19 cells in the absence of any stress was investigated (Fig. 4A, No Stress). There was no significant difference in the cell viability in untreated and 0.001-50 g/ml LEDGF 1-326 -treated cells following 48 h of treatment. At the highest dose of LEDGF 1-326 (50 g/ml), the cell viability was 108.14 Ϯ 5.63% (right-most bar) as compared with 100 Ϯ 13.19% for untreated cells (left-most bar), which was not significantly different. However, in pP23H-Rho-transfected ARPE-19 cells having aggregation stress, LEDGF 1-326 behaved differently (Fig. 4A, Aggregation Stress). Cells expressing P23H mutant rhodopsin showed a decline in cell viability to 48.25 Ϯ 5.62% (bar 2). This loss in cell viability could be attributed to the toxic effect of expression and accumulation of aggregationprone P23H mutant rhodopsin protein within the cells. When cells expressing P23H mutant rhodopsin (bars 3-9) were treated with increasing amounts of LEDGF 1-326 , an increase in the cell viability was seen. Even at a concentration as low as 0.001 g/ml, LEDGF 1-326 increased the cell viability of ARPE-19 cells from 48.3 Ϯ 5.6 to 77.0 Ϯ 10.2%. At and beyond this point, the cell viability remained significantly higher as compared with the pP23H-Rho-transfected group.
To confirm the results of MTT assay, fluorescence imaging (Fig. 4B) and cell counting (Fig. 4C) were also performed. Based on fluorescence images, the cells were counted automatically by high throughput analysis software. Groups treated with LEDGF 1-326 in the presence of P23H-Rho aggregation stress indicated a greater number of cells per frame as compared with the untreated group, indicating the ability of LEDGF 1-326 to prevent the loss of cells due to aggregation stress (Fig. 4B). Under P23H-Rho aggregation stress, the live cell count decreased to 38 Ϯ 6% (Fig. 4C). LEDGF 1-326 increased the live cell count significantly from 38 Ϯ 6 to 118 Ϯ 16% in a dose-dependent manner in 24 h. Further, a single treatment of LEDGF 1-326 was effective until day 7 for all doses ranging from 0.005 to 5 g. P23H-Rho-expressing cells treated with LEDGF 1-326 (blue lines) indicated a significantly higher number of live cells than the untreated group (red line) on day 7.
LEDGF  Reduces the Functional and Morphological Loss of Photoreceptors-LEDGF 1-326 efficacy to reduce the loss of visual function of photoreceptors was investigated in RCS rats by monitoring the ERGs (Fig. 6A). In dark-adapted (scotopic) ERG, the b-wave amplitude of 4-week-old untreated and treated rats ranged from 180.17 Ϯ 27.42 to 216.60 Ϯ 35.30 V (base ERG) before intravitreal injection was administered (Fig.  6A, panel i). At 2 weeks, after intravitreal injection, the b-wave amplitude ranged from 65.80 Ϯ 15.44 to 91.13 Ϯ 13.94 V. There was no significant difference in the untreated and LEDGF 1-326 -treated groups. However, beyond 2 weeks, the b-wave amplitude reduction was less in the LEDGF 1-326treated groups. At 8 weeks after the intravitreal injection, the b-wave amplitudes of the untreated group and groups treated with 0.5, 1.0, and 5 g of LEDGF 1-326 were 9.40 Ϯ 4.57, 32.43 Ϯ 10.34, 37.93 Ϯ 0.60, and 57.63 Ϯ 8.81 V, respectively. The b-wave amplitude of LEDGF 1-326 -treated groups was significantly (p Ͻ 0.05) higher than the untreated group. A dose-dependent delay in the decline of b-wave amplitude was seen for the LEDGF 1-326 -treated groups. With increasing doses of LEDGF 1-326 , the loss of b-wave amplitude was reduced.
Morphological analysis of retinal layer was done (Fig. 6B) for groups treated with 5 g of LEDGF 1-326 to determine the thickness of the various layers of retinal cells. Representative H&Estained images (Fig. 6B, panel ii) of wild type healthy Sprague-Dawley rat, the untreated RCS rat, and LEDGF 1-326 -treated RCS rat indicated ϳ10 -11-, 0 -1-, and 3-4-layer-thick ONL, respectively. The outer nuclear layer contains the nuclear bodies of the photoreceptors. For both control and LEDGF 1-326treated groups, the ONL was thinnest at the optic nerve head (ONH) region and thickened as it reached the periphery (Fig.  6B, panel iii). The LEDGF 1-326 -treated group consistently indicated thicker ONL as compared with the corresponding control group at every data point with significantly thicker ONL at 0.5, and 2 mm away from the ONH at the superior side and at 1.0 and 2.0 mm away from the ONH at the inferior side. INL was thicker near the ONH and thinner near the periphery (Fig. 6B, panel iv). The INL was significantly thicker at 1.0 and 2 mm away from the ONH at the superior side and at 1.0 mm away from the ONH at the inferior side in the LEDGF 1-326 -treated group.
The ability of LEDGF 1-326 to reduce the loss of rod photoreceptors was determined by immunofluorescence (Fig. 7). Immunolabeling of eye tissue sections with anti-rhodopsin antibody stained rod photoreceptors as red. The ONL and INL were stained blue with DAPI. A thick red band visualized in the LEDGF 1-326 -treated group (right panel) indicated the presence of large numbers of rod photoreceptors. In the untreated group, this band was almost undetectable. Further, similar to H&E staining (Fig. 6B, panel ii), the ONL was thicker (blue fluorescence) in the LEDGF 1-326 -treated group as compared with the untreated group.

DISCUSSION
In this study, we successfully cloned and expressed LEDGF 1-326 in large quantities and characterized its biophysical properties, determined its in vitro ability to enhance phagocytic activity and ameliorate protein aggregation-mediated cellular stress, and determined its in vivo efficacy in rescuing photoreceptors from degeneration. As elaborated below, our in vitro studies indicated that LEDGF  improves phagocytic activity and is effective in rescuing retinal pigment epithelial cells from death due to aggregationmediated stress. Further, our in vivo studies indicated that intravitreally injected LEDGF 1-326 can reduce retinal degeneration for at least 8 weeks in RCS rats.
In our previous study, we identified LEDGF 1-326 gene as a potential candidate for rescuing retinal pigment epithelial cells from P23H mutant rhodopsin aggregation-induced stress (30). The objective of this study was to develop a protein therapy based on LEDGF  . To develop LEDGF 1-326 as a therapeutic agent, it should be produced in tens of milligrams, which in itself is a daunting task. We designed a cloning strategy to express LEDGF 1-326 as a cytoplasmic protein in pET-28a (ϩ), a bacterial cell-based system, and confirmed the successful cloning by sequence analysis for the correct reading frame, right orientation of the gene, and no mutagenesis during the cloning procedure (Fig. 1A). Utilizing two-step purification, we were able to produce ϳ95% pure protein (Fig. 1B). Because LEDGF 1-326 is not highly likely to undergo glycosylation based on molecular modeling, and based on our observed efficacy, bacterial biosynthesis of LEDGF 1-326 is adequate for therapeutic activity. However, the role of other post-translational modifications and an alternation in activity of our LEDGF 1-326 as compared with native protein cannot be ruled out. SEC-HPLC of purified LEDGF 1-326 indicated 5% higher molecular weight species of LEDGF 1-326 ( Fig. 2A). Because it has been known that the presence of large aggregates can trigger immunogenic reactions in vivo (34), we also investigated the possible formation of high molecular weight aggregates during purification of LEDGF 1-326 protein, using DLS (Fig. 2B). Interestingly, the number of mean size measurements gave a single narrow size distribution of LEDGF 1-326 , indicating the absence of higher molecular weight aggregates. Because MALDI-TOF analysis indicated that LEGF 1-326 may exist as a monomer and/or dimer (Fig. 2C), nonreducing SDS-PAGE was utilized to confirm the existence of dimers (Fig. 2D). Higher molecular weight bands were evident in the gel, indicating the possible existence of a dimeric form of LEDGF  . Possibly a dynamic equilibrium exists between the monomeric and the dimeric forms of LEDGF  . However, the ratio of dimer to monomer is currently not established. The far-ultraviolet CD of LEDGF   (Fig. 2E) indicated that LEDGF 1-326 is predominantly a random coiled structure. Deconvolution of the CD signal further strengthened this view.
To establish a protein's therapeutic value, it is very important to understand its biophysical nature, which includes the conformational stability of the protein. Protein conformational stability is contributed by various environmental factors including pH, ionic strength, and temperature. Such information is useful in developing a stable formulation of the protein (35). In this study, fluorescence and CD were used to determine conformational stability of LEDGF   (Fig. 3). The results obtained from fluorescence spectroscopy for LEDGF   (Fig. 3A) indicated that tryptophan residues of native LEDGF 1-326 are partially exposed to the aqueous environment (36). LEDGF 1-326 free energy of unfolding, ⌬G(H 2 O), was positive, indicating that the unfolding of LEDGF 1-326 is unfavorable in the absence of any denaturant. The melting temperature of LEDGF 1-326 was predicted to be about 44°C, indicating that LEDGF 1-326 is stable at room temperature. ⌬G(H 2 O), m-value, [urea]1 ⁄ 2 , and T m together defined the conformational and thermal stability of LEDGF  .
RPE cells are known to accumulate rod outer segments containing rhodopsin with aging in human eyes (37). P23H rhodopsin-containing rod outer segment accumulation has also been implicated in vivo in neovascularization of RPE (38). Thus, our study utilizing mutant rhodopsin in ARPE-19 cells is pertinent to the in vitro simulation of retinal degeneration models. The in vitro efficacy data indicated significant decrease in the cell viability of ARPE-19 cells in the presence of P23H mutant rhodopsin protein (Fig. 4). This decrease in the cell viability was previously related to aggregation-mediated stress caused by the expression of P23H mutant rhodopsin and the associated aggregates in the cellular environment (39). Interestingly, upon treatment with LEDGF 1-326 , ARPE-19 cell viability increased significantly. LEDGF 1-326 protein did not alter the viability of ARPE-19 cells in the absence of stress, indicating that LEDGF 1-326 is more active in stressed conditions. Because MTT assay is indicative of mitochondrial respiratory state, but not a direct measure of cell viability, we performed high throughput cell counting using the Operetta High Content Analysis System. We did not see many dead cells in the P23H-Rho group. However, the cell count in the P23H-Rho group was significantly lower than the cells not expressing P23H-Rho, suggesting that rhodopsin aggregation arrested cell proliferation. In the presence of LEDGF 1-326 , the number of cells per frame increased as compared with controls, indicating that LEDGF 1-326 acts as a proliferating agent in the presence of protein aggregation stress. We previously reported that LEDGF 1-326 reduces oligomers of P23H rhodopsin as well as wild type rhodopsin, whereas increasing their monomers in a dose-dependent manner (30). LEDGF  in this earlier study did not affect the total rhodopsin protein expression at low doses. However, at high doses, there was a decrease in the total rhodopsin protein content with no changes in mRNA levels, suggesting that LEDGF 1-326 may be targeting rhodopsin aggregates to degradation pathways, in addition to preventing/disrupting rhodopsin aggregates.
Because RPE cells are specialized to phagocytose photoreceptor outer segments and other cellular debris (40), the phagocytic activity of ARPE-19 cells was determined in the presence of LEDGF  . Depletion of MERTK receptors in the retinal pigment epithelium is known to inhibit phagocytosis in RPE (40). However, MERTK siRNA transfection did not reduce the phagocytosis or uptake of FluoSpheres by ARPE-19 cell in our experimental conditions. It could be possible that under the conditions of the experiment, MERTK was not sufficiently depleted in ARPE-19 cells. Interestingly, LEDGF 1-326 significantly increased ARPE-19 phagocytic activity irrespective of MERTK siRNA transfection. This result is of high interest as this is the first time that LEDGF  has been shown to enhance phagocytosis.
In the absence of any exact replicate animal model for dry AMD, RCS rat is a useful model in understanding retinal degenerations. RCS rats exhibit primary genetic defect in the RPE due to deletion of MERTK gene, leading to loss of phagocytic activity and toxic accumulation of cell debris. The ultimate pathology in these animals is photoreceptor degeneration (41). Histopathology and ERG are established methods to monitor photoreceptor degeneration and retinal functional loss, respectively. B-waves of ERG specifically indicate photoreceptor health. Therefore, we used histopathology and ERG to monitor the ability of LEDGF 1-326 to reduce retinal degeneration. The decline in the b-wave amplitude in scotopic ERG represents loss or degeneration of rod and cone photoreceptors. In photopic ERG, the rods are bleached, so the response is mostly from the cone photoreceptors. The amplitude of the b-wave reflects the number of functionally active photoreceptors. Single intravitreal injection of LEDGF 1-326 significantly reduced the loss of b-wave amplitude, indicating reduction in the functional loss of photoreceptors for a period of at least 2 months from the day of treatment (Fig. 6A). Further, the functional protection of photoreceptors indicated by ERG correlated with the morphological protection of photoreceptors indicated by histology (Fig.  6B) and immunofluorescence (Fig. 7). LEDGF 1-326 rescued photoreceptors, as evidenced by the thicker ONL and thicker band of rhodopsin-stained photoreceptors. It also protected INL, which is made up of amacrine and bipolar cells, from degeneration, as evident in histological analysis. N-terminal LEDGF 1-326 has a stress-related element binding domain (30), which is capable of activating stress-related proteins (42). Despite some efforts in the current study, the mechanism as to how exactly LEDGF  works has yet to be investigated at the molecular level and is beyond the scope of this study. However, based on existing knowledge, the putative mechanisms of action of LEDGF 1-326 are summarized in Fig. 8. We speculate that LEDGF 1-326 may reduce both oxidative as well as endoplasmic reticulum stress (generated by protein aggregation) by up-regulating stress response proteins. It may also improve the phagocytic activity of RPE, thereby improving the survival of photoreceptors. Because a close association has been suggested (43) between protein aggregation stress (RP) and oxidative stress (dry AMD), a molecule like LEDGF 1-326 can be a universal therapeutic protein to treat multiple retinal degenerative diseases. Further, because many neurodegenerative diseases including Alzheimer, Parkinson, and Huntington diseases have been linked to protein aggregation, a possibility exists where LEDGF 1-326 could also be useful in some of these diseases as a therapeutic intervention.

CONCLUSIONS
In conclusion, in this study, we were able to biosynthesize and purify large quantities of LEDG 1-326 in aggregate-free, highly pure form. LEDGF 1-326 was found to be a conformationally and thermally stable protein at 25°C. LEDGF 1-326 was able to prevent the loss of cell viability due to aggregation-mediated stress at concentrations ranging from 0.001 to 50 g/ml. A single intravitreal dose of LEDGF 1-326 was effective in reducing retinal degeneration in RCS rats for over 2 months. Thus, LEDGF 1-326 is a potential therapeutic agent for retinal degenerative disorders.