A Lentiviral Gene Therapy Strategy for the In Vitro Production of Feline Erythropoietin

Nonregenerative anemia due to chronic renal failure is a common problem in domestic cats. Unfortunately, administration of recombinant human erythropoietin often only improves anemia temporarily due to antibody development. In this in vitro study, feline erythropoietin cDNA was cloned from feline renal tissue and utilized in the construction of a replication-defective lentiviral vector. The native recombinant feline erythropoietin (rfEPO) sequence was confirmed by sequencing. Upon viral vector infection of human 293H cells, Crandall Renal Feline Kidney cell line and primary feline peripheral blood mononuclear cells, bioactive rfEPO protein was produced. The presence of cellular rfEPO cDNA was confirmed by standard PCR, production of abundant rfEPO mRNA was confirmed by real-time PCR, and secretion of rfEPO protein was demonstrated by Western blot analyses, while rfEPO protein bioactivity was confirmed via an MTT proliferation bioassay. This in vitro study demonstrates the feasibility of a replication-defective lentiviral vector delivery system for the in vitro production of biologically active feline erythropoietin. Anemic cats with chronic renal failure represent a potential in vivo application of a lentiviral gene therapy system.


Introduction
The glycoprotein hormone erythropoietin, made by the peritubular cells of the renal cortex [1,2,3], stimulates the production of red blood cells. Erythropoietin-responsive anemia is a common cause of morbidity in cats affecting 32-65% of cats with chronic renal failure (CRF) [4,5,6]. Injection of exogenous recombinant human erythropoietin (rHuEPO) in these cats often results in the resolution of anemia, improvement in appetite, weight gain, energy level, alertness, playfulness, physical strength and attitude [7,8,9,10,11,12,13]. Multiple rHuEPO products are currently available including epoetin alfa (Epogen, Amgen), epoetin beta (Neo-Recormom, Roche) and darbepoetin alfa (Aranesp, Amgen). These products all have the same humanspecific primary amino acid sequence but differ in the degree of glycosylation, which affects renal clearance, thus influencing the frequency of administration [14]. Multiple adverse effects have been reported to be associated with the use of rHuEPO in cats, including refractory anemia, systemic hypertension, polycythemia, seizures, vomiting, iron deficiency, injection discomfort, cellulitis, cutaneous/mucocutaneous reactions, and arthralgia [12]. Because the amino acid sequence of rHuEPO is only 81.3% homologous to that of feline EPO, antibodies directed against rHuEPO can develop when it is administered to cats [13,15,16]. These antibodies are believed to block rHuEPO bioactivity and potentially cross-neutralize endogenous feline EPO leading to the development of life threatening red cell aplasia (RCA) [17]. Antibodies usually develop within the first few months of rHuEPO administration and a clinically significant immunologic reaction has been reported in 20% to 70% of feline patients receiving rHuEPO [18].
Several therapeutic strategies utilizing species-specific recombinant EPO have been attempted to address the issue of rHuEPO immunogenicity in cats. In an in vitro recombinant feline erythropoietin (rfEPO) strategy, Chinese hamster ovary (CHO) cells were transfected with a construct resulting in the production of biologically active, glycosylated rfEPO protein [19]. In a subsequent in vivo experiment by a different research group, CHO cell-generated rfEPO was utilized in the treatment of cats with nonregenerative anemia attributed to CRF [17]. In treated cats, the hematocrit increased significantly during the first 3 weeks of treatment. However, some of the cats that initially responded appropriately (8 of 26) developed red cell aplasia (RCA) that was refractory to additional treatments. Importantly, the reported cDNA sequence of the rfEPO utilized in this in vivo study had a single nucleotide substitution in the 44th codon (GGG), which resulted in the misincorporation of the glycine amino acid in lieu of glutamic acid. In an analysis of the EPO cDNA sequence of multiple mammalian species, the 44 th codon of the EPO gene, GAG, encodes the glutamic amino acid, manifesting for the existence of strong evolutionary pressure at this locus [15]. It is conceivable that glycine misincorporation in the rfEPO could lead to immunogenicity in treated cats.
Adeno-associated viral vectors expressing recombinant feline erythropoietin (rAAV-rfEPO) have been developed [2,16]. A rAAV-rfEPO vector, when administered intramuscularly to normal healthy cats, resulted in an increase in hematocrit over a 7 week period [16]. In a subsequent study, healthy cats treated with the lower-end vector dose had variable to no response while cats treated with the higher dose had significantly increased hematocrits during the first 3 weeks of rfEPO treatment [2]. Antibodies against the recombinant adeno-associated vector were detected in all of the vector treated cats. The authors speculated that this immune response may have prevented or reduced the success of vector readministration. In the same study, one cat developed RCA and another developed pathologic erythrocytosis. The authors speculated that posttranslational differences may exist between the endogenous and rfEPO synthesis intramuscularly [2]. Nevertheless, the genetic code of the rfEPO cDNA utilized in the construction of this viral vector encoded glycine (44 GGG ) in lieu of glutamic acid (44 GAG ) [2].
The ideal gene delivery system for treatment of CRF-associated non-regenerative anemia would have limited or no secondary immunogenicity (i.e. lack most or all viral genes), stable and native transgene expression and the ability to transduce both dividing and non-dividing cells. Replication-incompetent lentiviral vectors fulfill this list of requirements [20]. Various lentiviruses have been employed as viral vectors including HIV-1, HIV-2, simian immunodeficiency virus, feline immunodeficiency virus, caprine arthritis encephalitis virus, bovine immunodeficiency virus and visna virus [21]. Lentiviral vector pseudotyping with the vesicular stomatitis virus (VSV) surface glycoprotein confers an ability to transduce a broad range of cell types and increases the vector's therapeutic applications. In addition, cell-specific tropism can be individually tailored to the desired target cells via pseudotyping with the appropriate envelope protein [21]. Previous studies have demonstrated long-term stable gene transfer in experimental models in the absence of immune responses [22,23,24,25,26]. A lentivirus expressing EPO cDNA has been tested in both normal [27] and uremic rats [28]. In these studies, lentivirus-treated rats had efficient and sustained EPO secretion. To the authors' knowledge, replication-incompetent lentiviral systems have not been previously evaluated for the production of recombinant rfEPO. In this in vitro study, we demonstrate the effective use of a commercially available HIV-1-based lentiviral vector for the production of biologically active rfEPO in feline and human cell lines and in primary feline peripheral blood mononuclear cells.

Sequence of feEPO cDNA
The feline kidney-derived EPO cDNA was sequenced and is displayed in Figure 1. In this cDNA, the 44 th codon codes for glutamic acid (GAG), consistent with the (almost complete) GenBank fEPO sequence L10606. The complete fEPO sequence was submitted to GenBank under the accession number JQ413414.

Detection of the GFP-EPO Fusion Protein by Fluorescent Microscopy and Western Analysis
293H cells transfected with pGFP-EPO demonstrate abundant cytoplasmic green fluorescence, consistent with expression of the green fluorescence fusion protein (Figures 2a and 2b). The transfection efficiency was determined to be approximately 3% (Figure 2a). Western analysis confirms the production of a GFP fusion protein (primary anti-GFP antibody, Fig. 2c). GFP (28,973 Daltons) and non-glycosylated feline erythropoietin (20,914) have a combined mass of 49,887 Daltons. A band corresponding to ,50,000 can be seen in the transfected 293H cells which is absent for control cells. Three distinct GFP-positive degradation products that fall between 25 and 37 kDa are evident in the lane derived from cells transfected with pGFP-EPO and not in the lane derived from the control cells.
Full Length rfEPO Complimentary DNA Sequence Detected by Standard PCR in Infected PBMC DNA was isolated from lentiviral vector-infected and uninfected feline PBMC, and standard PCR with primers for the full-length EPO cDNA sequence was performed. Appropriate-sized (579 bp) amplicons were present upon agarose gel electrophoresis only in DNA from cells infected with the lentiviral vector (data not shown).

Feline EPO Message Detected by Real-time RT-PCR in PBMC Infected with rfEPO Lentiviral Vector
Total RNA was harvested from feline PBMC infected with the lentiviral EPO vector, reverse transcribed into cDNA, and assayed for a 130 bp fragment of the feline EPO sequence via real-time PCR. Feline EPO transcripts were detected at a level approximately 100-fold greater than housekeeping GAPDH transcripts (9463 copies rfEPO cDNA per copy GAPDH cDNA), demonstrating that feline PBMC infected with a replication-incompetent lentiviral vector coding for rfEPO produce abundant feline EPO mRNA. Feline EPO-specific transcripts were below the limit of detection in uninfected control PBMC.

Discussion
This study was designed to demonstrate the feasibility of a replication-defective HIV-1 based lentiviral vector for the in vitro production of rfEPO in human and feline cell lines as well as in primary feline leukocytes. The results of this study indicate that biologically active rfEPO protein can be produced in vitro utilizing this strategy. A long-term potential application for this gene therapy system is the treatment of cats with non-regenerative anemia secondary to CRF.
The concentration of rfEPO was higher in CRFK and PMBC cells after 3 days of incubation relative to samples derived from cells incubated for 6 days, as demonstrated by figure 3. This may be the result of episomally located, non-integrated viral vector activity which gradually diminished with prolonged incubation due to the loss of the episomal DNA. Alternatively, degradation of rfEPO protein may have been the result of supernatant proteases released form disrupted or dying cells.
In two feline erythropoietin cDNA sequences available in the GenBank database, a single nucleotide mismatch is present in the protein coding region of exon 2 at the 44 th codon (18 th codon in the mature feline EPO, after the signaling peptide is cleaved). One of the GenBank sequences codes for glycine (44 GGG , accession No. U00685) and the other for glutamic acid (44 GAG , accession No. L10606) [17]. Studies utilizing the rfEPO sequence coding for glycine have demonstrated several deleterious effects in experimental animals including RCA, and pathological erythrocytosis [2,17], whereas a study using the rfEPO sequence coding for glutamic acid described no pathology [16]. Whether this mismatch represents focal allelic variation, or alternatively, an artifact introduced during cloning, is controversial. However, in an in vivo study utilizing rfEPO protein coding for glycine (44 GGG ), the feline genomic DNA of 12 recipient cats was amplified and sequenced to establish evidence for allelic variation. The nucleotide sequence at this locus in all 12 cats (24 fEPO alleles) was determined to be 44 GAG , coding for glutamic acid [17]. In a database search of mammalian erythropoietin cDNA sequences (NCBI BLAST), over 50 different species of mammals demonstrate the GAG codon at this site with none coding for GGG. In aggregate, these findings suggest that the ''polymorphism'' identified at the 44 th codon of the feline erythropoietin cDNA most likely represents an artifact introduced during the initial cloning procedures. Importantly, modification of a single amino acid residue in human EPO has been shown to affect the biological function dramatically (two fold reduction in activity) [29].
For in vivo applications, the problem of rEPO immunogenicity is not the only important consideration that needs to be addressed. Regulation of EPO transgene dosage may play a critical role in the future success or failure of rfEPO gene therapy attempts. Transgene regulation can be dramatically affected by the function of the promoter driving expression of the rfEPO transgene. Multiple researchers have chosen to use a powerful constitutive promoter such as the cytomegalovirus immediate-early promoter    [16]. However, In two different studies, healthy cynomologous macaques treated with macaque-derived EPO-expressing AAV vectors developed supraphysiologic levels of EPO and polycythemia [30,31]. In an experiment where the recombinant EPO cDNA was expressed constitutively from a cytomegalovirus promoter, severe anemia (RCA) developed in some animals and polycythemia required repeated therapeutic phlebotomies to maintain nontoxic hematocrits [30]. Profound anemia has also been reported in macaques treated with an AAV vector expressing EPO cDNA driven from a doxycycline-regulated promoter [31]. As a result of studies like these, researchers have developed naked DNA-based gene therapy protocols in rodent models [32,33]. These studies have resulted in correction of anemia without triggering excessive hematopoiesis. Researchers have attempted to regulate the mammalian host hematocrit level by controlling the dosage of the administered recombinant viral vector. Unfortunately, this technique has resulted in an all-or-none phenomenon rather than the clinically desired dose-dependent increase in hematocrit [2]. Excision of the intramuscular rAAV2-rfEPO vector injection site has also proved unreliable for abrogating pathological erythrocytosis of rfEPO [2]. Although the therapeutic window of rfEPO in anemic cats is expected to be large [16], many researchers now feel that clinically useful gene expression of rfEPO should include a mechanism to regulate gene expression to avoid life-threatening anemia or pathological erythrocytosis [2]. An oxygen-regulated gene therapy approach has been successfully employed in erythropoietin-deficient mice [34]. In the same study, mice treated with CMV-EPO constructs acquired fatal polycythemia. The main physiologic stimulus for erythropoietin synthesis is renal hypoxia. When hypoxia is present, hypoxiainducible factor 1 (HIF-1a) is free to bind to hypoxia-response elements of oxygen regulated genes [14]. These response elements control the erythropoietin gene within the kidneys, and binding of HIF-1a stimulates an increase in production of erythropoietin. A constitutively active CMV promoter was utilized in the in vitro lentiviral strategy reported here. As previous in vivo animal studies have indicated that the regulation of EPO expression may play a role in hematopoietic pathology [30,31,34], a more sophisticated mechanism of transcriptional control may be warranted prior to the establishment of in vivo studies.
Insertional mutagenesis induced by viral vectors is an ongoing problem in gene therapy [35]. Vector-induced mutagenesis has been a concern ever since 4 of 9 human patients treated for Xlinked severe combined immunodeficiency developed leukemia following treatment with a retroviral vector [36]. As a result of this, lentiviral vectors with increased safety properties were developed. Nevertheless, proliferative hematopoietic disorders have recently been described in mice [37] and human patients treated with modern lentiviral vectors [36]. This has resulted in the design of ever more sophisticated lentiviral vector safety features including the incorporation of suicide genes, cell or tissue specific promoters, local/regional delivery of viral vectors, locus-targeted transgene integration and the use of ''insulators'' to prevent oncogene activation [36].
To provide an additional measure of control, the replicationdefective lentiviral vector system creates the potential for influencing cell-specific tropism via vector pseudotyping. Selection of certain viral envelope glycoproteins or other proteins facilitates cell targeting to enhance directed gene transfer [21]. For example, cell-specific targeting has been achieved through the use of lentiviral vectors pseudotyped with the Rabies virus glycoprotein [38] or the CD4 receptor [39]. A variety of envelope-like genes are currently available to provide vector-target cell specificity [21]. Optimally, vector targeting of long-lived cells could potentially abrogate the need to continuously readminister the viral vector. An optimal lentiviral vector may include hypoxia response element-regulated expression of rfEPO cDNA and the incorporation of a suicide gene into the vector design. A combination of these various strategies may provide a clinically relevant in vivo method for the treatment of non-regenerative anemia associated with CRF in cats.

Extraction and Cloning of Feline Erythropoietin
Normal feline renal tissue was harvested within 20 minutes of euthanasia from a feline cadaver. The renal tissue was obtained from a cat that was humanely euthanized for reasons unrelated to this experiment, and submitted for the disposal and group cremation at Veterinary Medical Teaching Hospital Pathology at UC Davis. The owners permitted unrestricted use of the cat's organs. The cat was not euthanized for the purposes of this study and the renal tissue was obtained one time only. The humane euthanasia procedure was performed according to the best standards of care. Based on our institution regulations, no IACUC protocol is required for the use of unrestricted organs obtained post mortem. The renal tissue was rinsed with ice cold diethylpyrocarbonate-treated isotonic phosphate buffered saline and approximately 1 gram of renal tissue was subsequently reduced to a fine powder with liquid nitrogen cooled mortar and pestle. Aliquots of frozen renal tissue were dissolved in Trizol reagent (Invitrogen) and total RNA was isolated according to the manufacturer's specifications with the exception that the RNAcontaining aqueous phase was mixed 1:1 with 70% ethanol and passed over an RNeasy mini column (Qiagen) including an on column DNase (Qiagen) digestion step. Messenger RNA was isolated from total RNA using an Oligotex mRNA kit (Qiagen) and was converted into cDNA using SuperScript III (Invitrogen) and random hexamers.
Feline erythropoietin cDNA was amplified using gene-specific primers based upon NCBI accession number NM_001009269.1 (forward 59-ATG-GGG-TCG-TGC-GAA-TGT-CCT-GC; reverse '5-TCA-CCT-GTC-TCC-TCT-TCG-GCA-GGC-CTC). The reaction was catalyzed with Expand High Fidelity Plus blended Taq and proof reading protein (Roche). Reaction conditions were as follows: an initial 94uC dissociation stage for 2.5 min followed by 10 cycles at 94uC for630 s, 59uC 30 s, and 72uC for 45 s and incremented 10 s/cycle thereafter for an additional 25 cycles, followed by a 10 min final elongation at 72uC. The resulting amplicon was electrophoresed and the 579 bp band corresponding to the predicted cDNA molecular weight was extracted using a QIAquick Gel extraction kit (Qiagen). The fresh PCR amplicon was inserted into a green fluorescence protein (GFP) fusion vector (pcDNA3.1/NT-GFP-TOPO, Invitrogen) according to the manufacturer's instructions generating pGFP-EPO. Resulting plasmids were cloned and expanded in TOP10 chemically competent E. coli (Invitrogen) and sequenced by a local vender (Davis Sequencing).

Cells
Human fetal kidney (293H) cells were purchased from GIBCO (Invitrogen), 293FT (Invitrogen), HT1080 (Stratagene) and Crandell Rees feline kidney (CRFK) cells were purchased from ATCC. Whole blood was collected in 5 mL vacutainer tubes containing EDTA from specific pathogen-free (SPF) cats (Feline Research Laboratory, Center for Companion Animal Health, UC Davis), and peripheral blood mononuclear cells (PBMC) were harvested by density gradient centrifugation through Ficoll-Hypaque (Sigma).

Erythropoietin-GFP Fusion Protein Transfection and Visualization
Fluorescent microscopy was utilized to determine if cells transfected with pGFP-EPO vector appropriately expressed the GFP fusion protein. 293 H cells grown in 75 cm 2 flasks were trypsinized and seeded into 60 mm diameter culture dishes (Nunclon) and allowed to progress to 80% confluence. Cells were grown in DMEM (Gibco) containing 20% fetal bovine serum (FBS, Hyclone), 1X non-essential amino acids (Gibco), penicillin (100 U/mL) and streptomycin (100 mg/mL) antibiotics (HyClone). Cells were transfected with pGFP-EPO according to the manufacturer's instructions (Lipofectamine 2000, Invitrogen). Seventy-two hr post-transfection, the cells were dislodged with trypsin and transferred to glass cover slips pre-coated with human recombinant fibronectin (18.75 mg/22 mm 2 ). Cells were allowed to attach overnight, the coverslips were washed with PBS and then fixed for 30 minutes in 2.5% paraformaldehyde in PBS. The fixative was changed with 0.1% Triton-X-100 in PBS containing 1 mg/ml of DAPI (Invitrogen), the reaction was allowed to progress for 5 min, followed by 365 min washes with 0.1% Triton-X-100 in PBS. Coverslips were mounted with Prolong Gold antifade reagent (Molecular Probes). Fluorescent microscopy was performed using an Olympus BX61 microscope coupled to a Penguin 600CL/150CL digital camera system (Pixera Corporation). Photography was completed within 96 hours post-transfection.

Lentiviral Vector Construction
The feEPO cDNA was transferred from pGFP-EPO without the GFP cDNA, to the lentiviral expression vector pLenti6/V5 (Invitrogen) using PCR technology. The following primers were employed, forward 59-CACC-ATG-GGG-TCG-TGC-GAA-TGT-39, reverse '5-TCA-CCT-GTC-TCC-TCT-TCG-GCA-G-39. The reverse primer codes for a STOP signal (TGA) at the 39 end of EPO coding sequence. PCR conditions were the same as the cloning reaction with the exception that the annealing temperature was dropped to 54uC. After the initial amplification, the resulting amplicon was blunted with Pfu (Stratagene), agarose gel purified as previously specified, and inserted into the pLenti6/ V5 directional TOPO vector (Invitrogen) as specified by the manufacturer. The resulting plasmid, containing the rfEPO cDNA but lacking the GFP gene (named pLenti rfEPO , Figure 1) was propagated and selected in Stbl3 E. coli (Invitrogen). Potential rearrangement in the LTR regions was checked by digesting the plasmid with Afl II and XhoI restriction enzymes. The plasmids were sequenced by a local vender (Davis Sequencing).
Lentivirus was titered using HT1080 cells as recommended by Invitrogen. Virus was added at 1/100, 1/1000, and 1/10,000 dilution in volumes of 50 mL; a control was included which was devoid of virus. After 24 hours, media was replaced with 4.5 ml DMEM and again 24 hours later with the addition of a selecting agent, 10 mg Blasticidin/ml. Selection progressed until all cells in the control flask were dead and then colonies were counted in the flasks receiving virus. The total yield for 2610 cm plates was 6610 7 viral particles.

Nucleic Acid Detection and Quantification in Feline PBMC
SPF feline PBMC were grown in RPMI media (Gibco) with 10% FBS (Hyclone), 100 U recombinant human IL2/ml (Hoffman-La Roche), and containing penicillin (100 U/mL) and streptomycin (100 mg/mL) antibiotics (HyClone). Cells were plated in 12-well tissue culture dishes and infected with the lentiviral EPO vector at an MOI of 10. Uninfected PBMC propagated in parallel were used as a negative control. Two days post infection, cells were harvested by centrifugation and DNA was isolated via QIAamp DNA mini kit (Qiagen) while RNA was isolated using TRIzol reagent (Invitrogen) according to the manufacturers' protocols. Isolated RNA was DNase treated (TURBO DNase, Ambion) according to the manufacturer's protocol, and further purified using a commercial RNA clean-up protocol (RNeasy kit, Qiagen). Approximately 1 mg RNA was reverse-transcribed into cDNA (First Strand cDNA Synthesis System, Origene); a control sample was run without reverse transcriptase.
Standard PCR was performed on DNA samples with the forward and reverse primers used for lentiviral vector construction (above), and amplicons were electrophoresed on agarose gel post-PCR to visualize bands corresponding to the full-length EPO cDNA sequence. Real time PCR assays were performed on cDNA samples in triplicate with Real Mastermix SYBR Rox (5 Prime) on a 7300 Real Time PCR System (Applied Biosystems), and subsequently analyzed with the 7300 system software. Both infected and uninfected control PBMC cDNA were assayed for a 130 bp amplicon at the 39 terminus of the EPO coding sequence (forward primer 59-AACCTCTGCTGCTCCACTCC-39, reverse primer 59-TCACCTGTCTCCTCTTCGGCAG-39), as well as feline housekeeping gene product GAPDH (forward primer 59-AAATTCCACGGCACAGTCAAG-39, reverse primer 59-TGATGGGCTTTCCATTGATGA-39) in parallel. Negative controls (no template and reverse transcriptase negative samples) were run simultaneously to ensure there was no nucleic acid contamination. EPO and GAPDH Ct values were converted to copy number using standard curves generated by serial dilution of plasmid DNA containing the cloned EPO or GAPDH sequences.

Western Blot
Human 293H and feline CRFK cells were grown in DMEM media (HyClone) with 10% fetal bovine serum (FBS, Gemini Bioproducts), and feline PBMC were grown in RPMI media (HyClone) with 10% FBS and 100 U/mL recombinant human IL2 (Hoffman-La Roche); both cultures contained penicillin (100 U/mL) and streptomycin (100 mg/mL) antibiotics (Hy-Clone). Cells were plated in 25 cm 2 cell culture flasks (293H, CRFK) or 12-well tissue culture dishes (PBMC), and were infected with the lentiviral EPO vector at a multiplicity of infection (MOI) of both 1 and 10. Uninfected cells were used as controls.
Clarified supernatants were harvested at 6 days (293H) or 3 and 6 days (CRFK and PBMC) post-infection, mixed 1:1 with 2x loading buffer (Hoefer Scientific Instruments), boiled for 2 min, and loaded onto 10-20% SDS-PAGE gels (Expedeon). Aliquots of each 293H cell supernatant were concentrated using 3 kDa centrifugal filters (Milipore) prior to use. Five microliters (10 U) of commercially available human EPO (Epoetin alpha, EPOGEN) were used as a positive control, and protein size was determined with a molecular weight marker (Expedeon). Samples were electrophoresed and electrotransfered to a PVDF membrane (Biorad) according to the gel system manufacturer's instructions (PAGEgel). The PVDF membrane was rinsed twice in trisbuffered saline with 0.1% Tween-20 (TBS/T) wash buffer, blocked for 1 hr with 5% nonfat milk/TBS/T (blocking buffer), and washed three more times before being incubated overnight in 1:1000 polyclonal rabbit anti-human EPO antibody (Accurate Chemical & Scientific Corporation) in blocking buffer at 4uC (as above). The membrane was washed three times, and incubated for 1 hr in a 1:10,000 dilution of goat anti-rabbit peroxidase conjugated antibody (Thermo Scientific) in blocking buffer, washed three more times, and incubated in ECL reagent (Pierce) for 1 min. Autoradiographs were created by exposing ECL-treated membranes to Scientific Imaging Film (Kodak) for 1-10 min, and developed with a medical film processor (Konica Minolta Medical & Graphic).
Western analyses for the GFP-erythropoietin fusion protein were conducted as follows. Cells were cultured and transfected as described previously. Cells in 60 mm diameter dishes were washed with ice cold PBS 2x and then dislodged by scraping with Costar Cell lifters (Corning Incorporated). Cells were sedimented at 4006g for 3 minutes and then disrupted with 150 mM NaCl, 50 mM Tris (pH, 7.4), 1% NP40, 0.25% Sodium Deoxycholate, 0.1% SDS, 1x Protease Inhibitor cocktail Set I (Calbiochem), 1 mM NaF, and 1 mM activated Na 3 VO 4 (RIPA). The reaction was allowed to progress for 1 hour on ice and the cell debris was pelleted at 16,0006g for 30 minutes. The protein concentration was determined for both naïve 293H (control) and transfected cell supernatants, and adjusted with RIPA to the same protein concentration per volume. Protein lysates were mixed with 46Laemmli sample buffer and boiled for 3 min. Lysates were applied at 75 mg per lane and separated on an 11% T 2.75% C SDS-PAGE gel (1406120.561.5 mm) with a 4% stacking gel. Proteins were transferred to PVDF overnight at 10uC. The membrane was blocked for 1 hr with 5% blotting grade non-fat milk (BioRad) in TBS (blocking buffer) and subsequently incubated overnight at 4uC with mouse anti-GFP antibody (UC Davis/NINDS/NIMH NeuroMab Facility) diluted 1:1000 in blocking buffer. The blot was subsequently washed 36with TBS and then incubated for 1 hr with sheep HRP-anti-mouse antibody (GE Healthcare, 1/5000) diluted in blocking buffer. The blot was washed 16with 0.05% Tween-20 in TBS followed with 36with TBS. Film detection was as previously described except ECL Plus Western Blotting Detection System was employed (GE Health Care) coupled with Hyperfilm (Amersham).

MTT Bioassay
To demonstrate bioactivity of recombinant feline erythropoietin, 293H, CRFK, and feline PBMC were grown in OPTIMEM (GIBCO) media with penicillin/streptomycin antibiotics as above, and infected with the lentiviral EPO vector at an MOI of 1 and 10 in the presence of polybrene (5 mg/ml, Sigma). Supernatants were harvested at 6 days (293H) or 3 and 6 days (CRFK and PBMC) post-infection; uninfected cell supernatants were used as a control. Human TF-1 erythroblastic cells (ATCC), were plated in a 96-well tissue culture dish at 5610 4 cells/well in RPMI media with 5% FBS and penicillin/streptomycin antibiotics (as above), and incubated with clarified supernatants for 48 hrs. TF-1 cells+1 U/mL human EPO (Epoetin alpha, EPOGEN) were used as a positive control, and TF-1 cells+fresh media, or fresh media alone, were used as negative controls. After the 48 hour incubation, an MTT cell proliferation assay (ATCC bioproducts) was performed per the manufacturer's protocol, and the optical density of the wells was determined at 595 nm using a Thermomax microplate reader (Molecular Devices). The MTT assay was performed in triplicate for each sample, supernatant or control. One-way ANOVA with Tukey's multiple comparison post-test was performed using GraphPad Prism version 5.04 for Windows (GraphPad Software, San Diego, CA). A standard curve was created using serial dilutions of human EPO from 0.01 U/mL to 2 U/mL, and the MTT assay was performed as described previously.

Statistics
Quantifiable MTT bioassay data was performed in triplicate, means and standard deviations were determined (GraphPad). For an experimental set of data, a global one-way ANOVA test was performed. Where significant global differences were detected, Tukey's multiple comparison tests were made. Differences were determined to be significant for p values less than 0.05.