Morpholino-Mediated Increase in Soluble Flt-1 Expression Results in Decreased Ocular and Tumor Neovascularization

Background Angiogenesis is a key process in several ocular disorders and cancers. Soluble Flt-1 is an alternatively spliced form of the Flt-1 gene that retains the ligand-binding domain, but lacks the membrane-spanning and intracellular kinase domains of the full-length membrane bound Flt-1 (mbFlt-1) protein. Thus, sFlt-1 is an endogenous inhibitor of VEGF-A mediated angiogenesis. Synthetic mopholino oligomers directed against splice site targets can modulate splice variant expression. We hypothesize that morpholino-induced upregulation of sFlt-1 will suppress angiogenesis in clinically relevant models of macular degeneration and breast cancer. Methods and Findings In vivo morpholino constructs were designed to target murine exon/intron 13 junction of the Flt-1 transcript denoted VEGFR1_MOe13; standard nonspecific morpholino was used as control. After nucleofection of endothelial and breast adenocarcinoma cell lines, total RNA was extracted and real-time RT-PCR performed for sFlt-1 and mbFlt-1. Intravitreal injections of VEGFR1_MOe13 or control were done in a model of laser-induced choroidal neovascularization and intratumoral injections were performed in MBA-MD-231 xenografts in nude mice. VEGFR1_MOe13 elevated sFlt-1 mRNA expression and suppressed mbFlt-1 mRNA expression in vitro in multiple cellular backgrounds (p<0.001). VEGFR1_MOe13 also elevated sFlt/mbFlt-1 ratio in vivo after laser choroidal injury 5.5 fold (p<0.001) and suppressed laser-induced CNV by 50% (p = 0.0179). This latter effect was reversed by RNAi of sFlt-1, confirming specificity of morpholino activity through up-regulation of sFlt-1. In the xenograft model, VEGFR1_MOe13 regressed tumor volume by 88.9%, increased sFlt-1 mRNA expression, and reduced vascular density by 50% relative to control morpholino treatment (p<0.05). Conclusions Morpholino oligomers targeting the VEGFR1 mRNA exon/intron 13 junction promote production of soluble FLT-1 over membrane bound FLT-1, resulting in suppression of lesional volume in laser induced CNV and breast adenocarcinoma. Thus, morpholino manipulation of alternative splicing offers translational potential for therapy of angiogenic disorders.


Introduction
Angiogenesis, though a fundamental physiologic process, is a key pathogenetic feature of numerous disease states. At present, therapeutic strategies have limited potential largely due to the fact that the underlying mechanisms of angiogenesis are incompletely understood. Work to elucidate the full complement of mediators and mechanisms important for angiogeneis, and further apply this knowledge in such a way to alter disease progression continues to be the foremost goal within the field.
Identification of VEGF as a critical mediator of vessel growth has been an important step to understanding the human condition in terms of the underlying molecular events. VEGF has been shown to be necessary and sufficient for ocular neovascularization [1][2][3][4]. In fact, transgenic mice which over-express human VEGF show widespread ocular neovascularization [5][6]. Additionally, pioneering work by Judah Folkman in the early 1970's demonstrated that solid tumor growth required VEGF expression [7]. However, it is not fully known what specific molecular mediators regulate VEGF expression. Soluble Flt-1, first described by Kendall and Thomas in 1993, is an alternatively spliced form of the Flt-1 gene, also referred to as VEGF-receptor 1 [8]. This alternative splicing event occurs within intron 13 such that sFlt-1 contains the ligand-binding domain, but lacks the membranespanning and intracellular kinase domains of the full-length membrane bound Flt-1 (mbFlt-1) protein [9]. As suggested by its structure, sFlt-1 is a potent endogenous inhibitor of VEGF Ainduced angiogenesis [9]. While the full complement of sFlt-1 expression and function has not been described, it has been shown to be both necessary and sufficient for maintenance of the avascular cornea [10]. In addition, recent work has demonstrated a role for modulation of sFlt-1 in the development and treatment of a form of pathologic ocular neovascularization termed choroidal neovasculization (CNV), via modulation of VEGF ( [11][12], unpublished data). Choroidal neovascularization is characterized by choroidal capillary growth through Bruch's membrane beneath the retinal pigmented epithelial (RPE) cell layer. This vascular pathology is most classically seen in exudative or ''wet'' age-related macular degeneration, the leading cause of vision loss in the western world [13][14]. Furthermore, sFLT-1 has also been shown to reduce VEGF expression and tumor vascularity in breast adenocarcinoma xenografts [15]. Thus, sFlt-1 represents a potential therapeutic target to reduce aberrant blood vessel growth over a spectrum of disease.
Current therapies targeting both CNV and tumor vasculature are focused on inhibiting the new vessel growth, and include such modalities as photocoagulation, photodynamic therapy, anti-VEGF intraocular injections, as well as systemic administration of anti-VEGF monoclonal antibody. These approaches have shown promise; however, they induce retinal damage, require repeated intraocular administration, or have recently been contraindicated for use in the case of systemic anti-VEGF therapy for breast adencarcinoma [4,13,[16][17]. Additionally though certainly progress has been made, incomplete efficacy and recurrence is commonly seen with all modalities. Development of novel therapeutic techniques to either augment or circumvent our current treatments is necessary to improve both efficacy and the risk profile.
In this work, we describe the use of morpholino oligomers promoting the expression of soluble Flt-1 as a means to determine its potential for therapeutic use in disorders characterized by aberrant neovascularization. This technology has human precedent in Duchene Muscular Dystrophy, where morpholino technology targeting the dystrophin gene has shown efficacy in splice site alteration and disease modification in both animal and human trials [18][19][20]. Our work demonstrates the use of morpholino oligomers alters the splicing of VEGFR1 such that production of soluble Flt-1 is favored both in vitro and in vivo. Furthermore, we demonstrate the efficacy of increased sFlt-1 expression in treatment of aberrant blood vessel growth in ocular and cancer disease models. Thus, morpholino technology holds promise for use in the context of human disease.

Results
VEGFR1_MOe13 promotes a shift from membrane VEGFR1 to soluble VEGFR1 in vitro In order to demonstrate the utility of morpholino constructs for modulating sFLT-1 expression, morpholino oligomers were designed targeting the FLT-1 mRNA exon13-intron13 junction (VEGFR1_MOe13) or intron13_exon14 junction (VEGFR1_ MOi13). The canonical Flt-1 gene consists of thirty exons in human and mouse. Full-length mRNA from all exons produces mbFLT-1. By contrast, sFLT-1 utilizes a polyadenylation site within intron13. Therefore, interaction between the morpholino constructs and VegfR1 pre-mRNA is predicted to influence the alternative splicing event such that production of soluble FLT-1 is favored. To directly measure the relationship between membrane bound and soluble FLT-1 in the presence of VEGFR1_MOe13 and VEGFR1_ MOi13, human umbilical vein endothelial cells (HUVEC) were electroporated with targeting or standard morpholino oligomers designed against human VEGFR1. Using this technique MO constructs were found to sufficiently access the nuclear compartment (Fig. 1a). Forty-eight hours after electroporation, total RNA was harvested and soluble VEGFR1 and membrane VEGFR1 mRNA expression was assessed using real-time PCR. We found that VEGFR1_MOe13, VEGFR1_MOi13, and a combination of a combination of one-half dose of both VEGFR1_MOe13 and VEGFR1_MOi13 significantly decreased membrane VEGFR1 mRNA and increased soluble VEGFR1 mRNA ( Fig. 1b-c; Figure S1). To determine if the VEGFR1 morpholino constructs also affected soluble FLT-1 expression at the protein level, soluble VEGFR1 protein in HUVEC culture medium was measured by ELISA in the presence or absence of the VEGFR1 morpholino constructs or control mopholino. VEGFR1_MOe13 increased soluble VEGFR1 protein production as compared with standard morpholino or control PBS conditions (Fig. 1d). In contrast, VEGFR1_MOi13 were less effective, whereas combined delivery of both VEGFR1_MOi13 and VEGFR1_MOe13 showed intermediate response, possibly due to competitive interference or reduced efficacy of one morpholino as compared with the other. Thus, morpholino targeting of the FLT-1 mRNA exon13-intron13 junction using VEGFR1_MOe13 is most efficacious for increasing sFLT-1 and decreasing mbFLT-1 expression in vitro.
To ensure that this effect was not cell line specific, we sought to validate our findings in other cellular backgrounds. Although soluble FLT-1 is predominantly a product of endothelial cells per its role in angiogenesis, it is also expressed by a number of other cell types including tumor cells [15,[21][22][23][24]. Thus, we sought to determine whether expression of the VEGFR1_MOe13 in MCF7 and MBA-MD-231 breast adenocarcinoma cells would increase soluble FLT-1 levels. As demonstrated in figure 2, electroporation of VEGFR1_MOe13 directed against the human FLT-1 transcript increases soluble FLT-1 RNA and decreases membrane FLT-1 RNA in both adenocarcinoma lines ( Fig. 2a-d).
VEGFR1_MOe13 increases sFlt/mFlt mRNA ratio in the mouse retina and suppresses laser-induced choroidal neovascularization volume in vivo In order to directly test the efficacy of the VEGFR1_MOe13 for both in vivo activity as well as predicted effect on the process of angiogenesis we adopted the well established murine model of laser-induced choroidal neovascularization which induces significant CNV lesions 1 week after laser injury [25][26]. We hypothesized that expression of VEGFR1_MOe13 in vivo would both increase soluble Flt-1 levels and lead to suppression of laserinduced CNV. To first evaluate the effectiveness of VEGFR1_ MOe13 to modulate sFlt-1 levels in vivo, we examined the sFlt/ mFlt mRNA ratio in the mouse retina 24 hours after injection with PBS, vivo-standard_MO, or vivo-VEGFR1_MOe13 designed to target murine sFlt-1 (the ''vivo'' denotes modification allowing MO construct to enter cells in vivo as demonstrated [27]). We found that intra-vitreol injection of vivo-VEGFR1_MOe13 leads to a significant increase of sFlt/mFlt mRNA ratio as compared with PBS or vivo-standard_MO injection (Fig. 3a). Thus, VEGFR1_MOe13 expression is sufficient to increase soluble Flt-1 expression in vivo.
To determine whether VEGFR1_MOe13 expression in vivo could suppress development of CNV in the setting of laser insult, vivo-VEGFR1_MOe13, vivo-standard_MO or PBS were injected intra-vitreously on day 1 and day 4 after laser photocoagulation. One week after laser photocoagulation, eyes were enucleated and the degree of CNV volume was measured by confocal microscope after isolectin GS-IB 4 vasculature staining. CNV volumes were quantified using confocal microscopy. Murine eyes treated with intra-vitreal vivo-VEGFR1_MOe13 displayed a statistically significant decrease in CNV volume as compared with eyes treated with either vivo-standard morpholino or PBS controls ( Fig. 3b-d).

Short hairpin RNA-mediated sFlt-1 knock-down ''rescues'' the CNV phenotype in vivo
In order to demonstrate that the measured effect of reduced CNV following intra-vitreol injection of VEGFR1_MOe13 was specific for an increase in sFlt-1 expression, we knocked-down sFlt-1 expression AAV2_shsFlt encoding short hairpin RNA(shRNA) targeting sFlt-1 mRNA (unpublished data; [10]). Intra-vitreol injections were performed using PBS, AAV2_shNEG (non specific shRNA) or AAV2_shsFlt (shRNA targeting sFlt-1) and laser photocoagulation performed 2 weeks later. In a consistent fashion with prior studies, on day 1 and 4 following photocoagulation PBS, vivo-standard_MO or vivo-VEGFR1_MOe13 constructs were injected into pretreated eyes. We hypothesized that if increased sFlt-1 was sufficient for suppression of the laser-induced CNV phenotype, co-expression of AAV2_shR-NA_sFlt would reverse this effect. In agreement with this hypothesis, we observed that pre-treatment with AAV2_shsFlt results in reversal of VEGFR1_MOe13-mediated CNV suppression (Fig. 4). Thus, increased soluble Flt-1 expression is sufficient to at least partially mediate CNV suppression in the setting of laser-induced injury. In order to demonstrate that anti-angiogenic activity of the VEGFR1_MOe13 construct is not limited to the ocular compartment, we sought to measure its efficacy in the setting of malignancyassociated neovascularization. Tumor vasculature is a rapidly emerging therapeutic target. As such, this model represents an attractive context in which to apply technologies designed to inhibit neovascularization. Within the context of malignancy, breast adenocarcinoma is known to demonstrate marked dependence on VEGF signaling for sustained neovascularization and growth [28][29]. In fact, VEGF inhibition has been shown to reduce tumor growth in both the experimental as well as the clinical setting [17,30]. We hypothesized that treatment of MBA-MD-231 human breast adenocarcinoma xenograft tumors with vivo-VEGFR1_ MOe13 would result in increased levels of soluble Flt-1 and a subsequent decrease in neovascularization and tumor regression. To directly test this hypothesis, female nude mice were inoculated as described. Xenografts were permitted to grow for 14 days. Tumors were then directly injected with either murine vivo-VEGFR1_ MOe13 or vivo-standard morpholino. Injections and tumor volume assessments were performed bi-weekly for a duration of 4 weeks. We found that treatment of xenograft tumors with vivo-VEGFR1_ MOe13 led to tumor regression when compared with standard morpholino treatment (p = 0.04) (Fig. 5a). Furthermore, murine sFlt-1 mRNA transcript levels were increased and mbFlt-1 mRNA levels decreased in treatment tumors at the conclusion of the 4 week treatment period as compared with control tumors when assessed using real-time PCR (Fig. 5b-c). Finally, to determine whether vascular density was decreased in tumors treated with viv-VEGFR1_MOe13 injection, tumor vasculature was stained with GS-IB 4 and vessel density quantified using fluorescence microscopy following the 4 week treatment period. Tumors treated with the vivo-VEGFR1_MOe13 construct demonstrated a statistically significant decrease in vascular density (Fig. 5d). Interestingly, these results were achieved using VEGFR1_MOe13 targeting the murine Flt-1 transcript, while treatment with a VEGFR1_MOe13 construct targeting human FLT-1 did not demonstrate tumor regression in vivo ( Figure S2). This may indicate that host vasculature is important for tumor growth,, suggesting that human sFLT-1 produced by tumor cells does not effectively inhibit murine VEGF signaling. However, analysis of the human and murine morpholino sequences demonstrates significant overlap between the murine morpholino construct with the human VEGFR1 sequence. The same is not true for the human morpholino with respect to the murine VEGFR1 mRNA sequence. Therefore, this cross-reactivity could account for our findings as could overall decreased efficacy of the human morpholino construct. The latter is unclear given the fact that in vitro this sequence increases soluble FLT-1 very effectively. Further studies are needed to fully understand this.

Discussion
Neovascularization is a common pathological process underpinning numerous disease states. Elucidation of key molecular mediators has allowed for the development of therapeutic strategies targeting the underlying molecular processes. Soluble Flt-1 first described in 1993, is an alternatively spliced form of the Flt-1 gene [4][5]. The morpholino oligomer VEGFR1_MOe13 is designed to target the Flt-1 mRNA exon13-intron13; therefore, interaction between VEGFR1_MOe13 and VegfR1 pre-mRNA is predicted to influence the alternative splicing event such that production of soluble Flt-1 is promoted.
Herein we demonstrate that expression of VEGFR1_MOe13 both in vitro and in vivo results in elevated levels of soluble Flt-1 and decreased membrane bound Flt-1. Furthermore, these data show that VEGFR1_MOe13-mediated increase in soluble Flt-1 within the retina can prevent choroidal neovascularization following laser photocoagulation. Finally, our data indicate that this phenotype is attributable to increased sFlt-1 expression as co-treatment with sFlt-1-RNAi negates this effect. Thus, soluble Flt-1 expression is both necessary and sufficient for suppression of laser-induced CNV. Taken together these data indicate that modulation of soluble Flt-1 expression in the clinical setting has potential therapeutic value. This represents great promise when considering the over 7 million patients in the US alone with non-exudative age related macular degeneration (ARMD) currently ''at risk'' for development of CNV.
Angiogenesis in the form of aberrant neovascularization is fundamental to the pathophysiology of other disease states as well. Notably, new blood vessel formation is absolutely required for sustained solid tumor growth [7]. As noted, breast adenocarcinoma is dependent on sustained neovascularization both in animal and human studies [28][29]. As a ''proof of principle'' for broad  PLoS ONE | www.plosone.org applicability of morpholino-mediated Flt-1 modulation in the treatment of neovascular disease, we demonstrate that intratumoral injection of VEGFR1_MOe13 leads to increased intratumoral levels of sFlt-1 and decreased mbFlt-1.. Furthermore, breast adenocarcinoma xenografts expressing elevated levels of sFlt-1 demonstrate a blunted neovascular response and regress once established as compared with control tumors. Thus, modulation of soluble Flt-1 expression using morpholino technology represents a therapeutic tool with broad applicability across a spectrum of neovascular disease.
Taken together, these data indicate that morpholino expression is a viable tool for modulating the expression of FLT-1, i.e., the balance between membrane and soluble forms of this transcript. In our system, these data indicate that morpholino interference at the Flt-1 mRNA exon13-intron13 junction leads to the greatest amount of soluble FLT1 production. The VEGFR1_MOe13 construct is effective in two independent disease models, choroidal neovascularization and breast adenocarcinoma. In both, aberrant vasculature and disease burden are reduced. Given the highly divergent roles of FLT1 isoforms, morpholino oligos show promise in a variety of physiologic settings and disease states to act as an ''exogenous switch'' to modulate angiogenesis.
These data demonstrate the efficacy of morpholino technology to influence FLT-1 gene expression at the mRNA level both in vitro and in vivo setting. Recent application of splice-blocking morpholinos in the setting of DMD has demonstrated disease-modifying effects in both animal and human trials [18][19][20]. Therefore, these constructs hold promise as true ''bench to bedside'' tools. Within the setting of neovascularization, use of morpholino technology represents a shift from current therapies which are predominantly based on monoclonal antibody approaches. The evolving potential for morpholino technology to be targeted to areas of pathology following systemic administration may lend this technology to broader appeal and utility [31]. Additionally, combining morpholino-mediated sFLT-1 modulation with other known anti-angiogenic strategies is a promising area of investigation. For example, morpholino-based inhibition of other angiogenic factors such as FGF and PDGF remains an interesting question, as these factors have been shown to promote intra-ocular neovascularization and inhibition shown to reduce neovascular pathology [32][33][34]. Furthermore, synergy between these pathways has been shown to promote tumor vascularity and metastasis in murine studies [35][36]. Therefore, steric inhibition of splice sites required for proper expression of these factors may represent yet another role for morpholino technology in the treatment of neovascular disease. Thus morpholino technology currently represents an effective mechanism for treatment of neovascular pathology across a spectrum of disease with high translational potential and exciting future applications.

Constructs and Reagents
In vivo morpholino constructs were designed to target murine intron/exon 13 junction of the FLT-1 transcript denoted VEGFR1_MOe13 and VEGFR1_MOi13. Constructs were chemically modified such that they would easily diffuse to access the intracellular compartment [27]. For use in xenograft assays, VEGFR1_MOe13 morpholino constructs were suspended in sterile PBS and used at a concentration of 400 ng/dose. Each dose is equal to 50 ul total volume. A standard morpholino targeting the murine beta-globin subunit2 was used as a control and prepared at the same concentration. Morpholino sequences are as follows.

Laser induced choroidal neovascularization
Laser photocoagulation (532 nm, 150 mV, 100 ms, 100 mm; NIDEK MC-4000) was performed on both eyes (2 to 5 spots per eye) as described [24][25]. After enucleating the eyes, sclera/ choroid/RPE complex were fixed in 4% paraformaldehyde for 2 hours at 4uC. After blocking in 5% FBS/PBS with 0.02% tritonX-100 and 2 mM MgCl2, samples were stained with 5 mg/ ml Alexa488 and Alexa546 conjugated isolectin GS-IB 4 (Invitrogen Corporation, Carlsbad, USA) overnight. After washing the samples were flat mounted on glass slides. CNV volume was measured by scanning laser confocal microscopy (Olympus America Inc., Center Valley, USA). These animal studies were performed under IACUC protocol approval number 09-03005 approved by the Committee on the Ethics of Animal Experiments at the University of Utah. All interventions were performed either under sodium pentobarbital anesthesia or after animals were humanely euthanized.
Xenograft analysis 3610 6 MBA-MD-231 human breast adenocarcinoma cells were injected subcutaneously into the flanks of female nude mice. Following a standardized 2 week inoculation period, xenograft tumors were injected biweekly for a total duration of 4 weeks with VEGFR1_MOe13 or standard morpholino control. These constructs were prepared as detailed above. Tumor growth was assessed using digital calipers with bi-weekly measurements correlating with injection period. (Volume mm 3 = width 2 6 length/2. These animal studies were performed under IACUC protocol approval number 09-03005 approved by the Committee on the Ethics of Animal Experiments at the University of Utah. All sFLT-1 Moropholino Decreases Neovascularization interventions were performed either under sodium pentobarbital anesthesia or after animals were humanely euthanized.
After electroporation of F-MO(10 ul of 1 mM) or STD MO into HUVECs, the cells were plated to 8 well glass slide (nunc, Rochester, NY) coated with collagen. 24 hours later, the cells were fixed with 4% parafolmaldehyde/PBS, stained with DAPI, and observed with the Zeiss Axiovert 200 inverted fluorescence microscope.

Statistical analysis
Data are presented as mean +/2SEM. Statistical analysis was performed using student T test. A p-value 0.05 was considered significant. Figure S1 The Soluble/Membrane FLT-1 ratio increases following electroporation of VEGFR1_MOe13 into HUVEC.

Supporting Information
HUVECs were electroporated with VEGFR1_MOe13, VEGFR1_ MOi13, a combination of VEGFR1_MOe13 and VEGFR1_ MOi13, Standard_MO. All morpholino sequences were designed to target the human VEGFR1 transcript. mbFLT-1 mRNA and sFLT-1 mRNA expression were assessed using real time PCR. Values were normalized to GAPDH mRNA and normal HUVEC was used as 1.0. (TIF) Figure S2 Intra-tumoral injection of VEGFR1_MOe13 targeting human FLT-1 does not induce xenograft tumor regression. MB-MDA-231 breast cancer cells were grown as xenografts in female nude mice for 2 weeks prior to beginning intra-tumoral injection treatment with either a standard morpholino or VEGFR1_MOe13 morpholino targeting the human FLT-1 transcript. Change in tumor volume was assessed following a 4 week treatment course. p = 0.3 (TIF)