Expression and characterization of a soluble VEGF receptor 2 protein

Vascular endothelial growth factor A (VEGFA), a polypeptide growth factor, originally discovered as a permeability factor is now accepted as an inducer of angiogenesis and vascular leakage both in vitro and in vivo[1–3]. VEGF acts by binding with its receptors (VEGFR), promoting receptor dimerization and tyrosine auto-phosphorylation. This leads to a kinase cascade in the cell membrane/cytoplasm that transmits a signal to the cell nucleus, inducing a series of changes. VEGFRs are trans-membrane proteins comprised of seven immunoglobulin-like (Ig) extracellular domains, an intracellular tyrosine kinase domain and a trans-membrane sequence.

Members of the VEGFR family include VEGFR1, VEGFR2, and VEGFR3. Among these receptors, VEGFR2 is recognized as the key down regulator for VEGFA’s pathological effects [4]. Vascular leakage, angiogenesis and neuronal degeneration are three features of diabetic retinopathy. VEGFA is a critical driver in these processes, acting as a vascular leakage inducer, angiogenesis mediator and neuron protector in diabetic retinopathy. The VEGF-VEGFR2 pathway was found to play a key role in diabetic retinopathy. Blockage of the VEGF-VEGFR2 pathway may represent a novel therapeutic strategy to reduce vascular leakage and angiogenesis in diabetic retinopathy.

We have previously shown that the function of full length Vegfr2 (long form, 1343 aa) encoded by the rat is similar to humans. A C-terminal truncated version of the protein (991 aa) was found to be encoded by a short form of the vegfr2 gene [3]. We also demonstrated that the long formVEGFR2 is the predominant receptor for VEGF-A in the rat retina, while in comparison the truncated form of VEGFR2 is expressed at very low levels in both normal and diabetic retinas. We further found that the long form of VEGFR2 is the predominant mediator of VEGF-A in the pathogenesis of diabetic retinopathy (DR) and can be significantly inhibited by intravitreal steroid treatment. The short form, which cannot be phosphorylated, does not appear to contribute to the pathogenesis of DR. We proposed that the truncated form of Flk-1 could be used clinically as a dominant negative inhibitor of the effects of VEGF [3].

The extracellular segment of VEGFR2 includes the VEGF ligand binding site. More specifically, the VEGFR2 Ig domain 1 is necessary for VEGF binding, while Ig domains 2–3 ensure a tight bond with VEGF. The receptor employs Ig domain 4 to create the active form of the homologous dimer, whereas Ig domains 5–7 are not closely associated with VEGF binding [5]. Based on our previous findings, we aimed to express a new truncated VEGFR2 by digesting a full-length rat Vegfr2-encoding plasmid to produce a fragment that would encode VEGFR2 Ig domains 1–3 and 5.

Several anti-VEGF/VEGFR therapeutic strategies have been developed, including the use of soluble receptors, VEGF and VEGFR-neutralizing antibodies. Other inhibitors that target both VEGF and VEGFRs such as small molecules or antisense oligodeoxynucleotides that inhibit VEGFR signal transduction have also been explored. When used alone or in conjunction with other drugs, these methods can be utilized to treat malignant tumors and other diseases. Research has shown that soluble high-affinity VEGFR2 fragments are therapeutically effective [6, 7]. Soluble VEGFR2 fragments can bind endogenous VEGF, thereby blocking the action of VEGF and impeding the binding of VEGFR1, VEGFR2, and VEGFR3 to cell membrane ligands, inhibiting the VEGF-induced signaling pathway, and preventing angiogenesis [8].

VEGFR2, the main functional VEGF receptor, is a type III transmembrane protein kinase that was first discovered in 1991 [9]. The VEGFR2 gene locus is found on chromosome 4q11-q12, and it encodes a receptor with a total length of 1,443 amino acids. VEGFR2 is translated into an intracellular non-glycosylated protein of approximately 150 kDa, that is subsequently repeatedly glycosylated, to yield a mature form of approximately 230 kDa that localizes to the cell membrane.

Our previous work demonstrated that though there are two isoforms of VEGFR2 found in rat, the long form of vegfr2 gene share 99.9% identity of gene sequence with human’s. Given that humans have only a single isoform, we have chosen to express and characterize rat truncated VEGFR2 in this study ahead of further in vivo testing.

Catalysis is facilitated by an extracellular structural domain that contains seven immunoglobulin-like domains, a transmembrane structural domain, and a cytoplasmic tyrosine kinase structural domain. Signal transduction is chiefly mediated by the specific binding of VEGF to the extracellular regions I–III of VEGFR2 [10, 11].

Previous studies have only tested a soluble VEGFR2 (sVEGFR2) protein that contained Ig domain 3 (97 amino acids) and have consequently overlooked the regulation of receptor-ligand binding by Ig domains 1 and 2, which reduces the affinity of sVEGFR2 for VEGF. This study employed restriction enzyme digestion in conjunction with the recombinant cloning of a Norway rat Vegfr2 gene restriction fragment into the eukaryotic expression vector pCMV6. The end product was a pCMV6-diagested-Vegfr2 plasmid encoding a form of sVEGFR2 that contained Ig domains 1–3 and 5 of the rat VEGFR2 protein.

Most research on the preparation of sVEGFR proteins has employed recombinant target plasmids to induce target protein expression in an IPTG prokaryotic expression system. However, this process leads to many problems, such as severe cytotoxicity of the expression products, failure to glycosylate the protein product, and failure to achieve correct target protein folding [12, 13]. Reports have indicated that most laboratories relied on Escherichia coli expression systems to induce sVEGFR1 and sVEGFR2 expression and that the anti-angiogenic efficacies of these expression products was less than 40% [14]. This low level was because E. coli usually expresses 50% of the total protein, and rapid bacterial expression leads to either a failure to fold or incorrect folding. These problems are particularly common in proteins that require long expression times such as heterogeneous proteins or those with molecular companions. The transient transfection method ensures that the exogenous genes exist only in free plasmid vectors after target cell transfection. The plasmids do not integrate into the chromosomes. Consequently, the target gene expression products can be obtained in a short time; however, as the cells proliferate, the exogenous genes ultimately disappear. The expression process can continue for several days or up to 2 weeks. This method can greatly shorten the process of recombinant protein production and thus provides a rapid and convenient gene expression method. In addition, we co-transfected, the reporter pCMV-gfp plasmid with the target plasmids for transfection condition optimization.

We initially chose to transfect HEK293 cells in a 48-well plate at 80–90% confluency. However, given the relatively long experimental period, the cells invariably became fully confluent within 24 h of transfection. The cells also underwent apoptosis due to overgrowth during the observation stage, thus influencing subsequent observations and testing. We therefore reduced the required cell density to 50–60% confluency at the time of transfection, although this procedure resulted in lower cell transfection efficiencies and reduced target protein expression. Regarding ELISA analysis, we initially chose to use a VEGFR2 ELISA Kit to test the culture supernatants after the cells had been cultured in DMEM with 10% fetal bovine serum (FBS). However, as the background values were higher in the control group than the experimental group, we concluded that miscellaneous proteins in the FBS had severely interfered with the test results, and therefore, the cultures were performed in serum-free DMEM. Given the need for growth factors in the medium, 1 × 10⁻⁴ v/v EGF was added to ensure cell survival.

Regarding the pCMV6-sVegfr2 plasmid obtained by sub-cloning, we discovered that differences in transfection efficiencies of pCMV6-truncated-rVegfr2 or pCMV6-rVegfr2 in the cells caused different green fluorescent protein expression levels in the cells. This discrepancy might have been due to inhibited GFP expression caused by exogenous target plasmid (pCMV6-truncated-rVegfr2 or pCMV6-rVegfr2) transfection and expression. While the exogenous plasmid transfection efficiencies were largely identical among the experimental groups, the transfection of HEK293 cells with large amounts of pCMV6-rVegfr2 led to reduced transfection efficiency. Transfection with a lower concentration of either pCMV6-truncated-rVegfr2 or pCMV6-rVegfr2 had a protective effect on the cells, inhibiting apoptosis. When the target plasmid concentration was increased, this protective effect was lost but the expressed protein was not cytotoxic. While transfection with pCMV6-truncated-Vegfr2 induced HEK293 cells to express truncated-VEGFR2 protein, the amount of pCMV6-rVegfr2 transfected into the cells did not closely correlate with the amount of truncated-VEGFR-2 secreted into the culture medium. The differences in the results between the transfection and control groups were greatest from 24–30 h after transfection. This period was therefore chosen as the optimal time for culture supernatant collection. Cells in the control and study groups exhibited poor growth at 45 h after plasmid transfection. This may have been due to the release of membrane-associated VEGFR2 protein into the culture medium; the protein fragments were thus degraded as the culture time increased, and the secreted sVEGFR2 in the culture medium tended to decompose. Transient transfection with different exogenous plasmids influenced sVEGFR2 protein expression and secretion both inside and outside cells. pCMV6-truncated-rVegfr2-transfected cells expressed and secreted both the soluble VEGFR2 protein and soluble VEGFR2 protein fragments at high efficiencies. The amount of sVEGFR2 protein secreted by cells in the sVEGFR2 group, as a percentage of all intracellular and extracellular VEGFR2 protein, increased by nearly 10%, compared to that secreted by cells in the soluble VEGFR2 group, and by more than 20% compared to that secreted by cells in the control group. We propose that pCMV6-rVegfr2 induced copious VEGFR2 protein expression on the cell membrane; however, the limited ability of the cell membrane to bind this protein allowed the excess release of receptor protein fragments into the culture medium. In contrast, VEGFR2 protein expression on the cell membrane was lower in the control group, and the positive results obtained when testing the culture supernatant might have been caused by the release of VEGFR2 receptor protein fragments from the cell membrane when the cells underwent apoptosis.

Given the limitations on the number of cells per well, the reduced plate coverage after transfection, and the degradation of sVEGFR2 during the culture process, this study obtained a relatively low target protein yield. Subsequent studies should aim to inhibit protein degradation, ease the cell culture and protein expression restrictions, and perform specific testing of the sVEGFR2 protein with the ultimate goal of purifying and testing its biological activity.

After transfecting cells with pCMV2-trucated-rVegfr2, we were able to obtain serum-free DMEM culture supernatants containing VEGFR2 protein fragments that bound to VEGF. This work will provide a basis for the development of soluble receptors that target endogenous VEGF and treatments for diabetic retinopathy that target VEGFR2.