Co-culture of Retinal and Endothelial Cells Results in the Modulation of Genes Critical to Retinal Neovascularization

Background

Neovascularization (angiogenesis) is a multistep process, controlled by opposing regulatory factors, which plays a crucial role in several ocular diseases. It often results in vitreous hemorrhage, retinal detachment, neovascularization glaucoma and subsequent vision loss. Hypoxia is considered to be one of the key factors to trigger angiogenesis by inducing angiogenic factors (like VEGF) and their receptors mediated by hypoxia inducible factor-1 (HIF-1α) a critical transcriptional factor. Another factor, nuclear factor kappa B (NFκB) also regulates many of the genes required for neovascularization, and can also be activated by hypoxia. The aim of this study was to elucidate the mechanism of interaction between HRPC and HUVEC that modulates a neovascularization response.

Methods

Human retinal progenitor cells (HRPC) and human umbilical vein endothelial cells (HUVEC) were cultured/co-cultured under normoxia (control) (20% O₂) or hypoxia (1% O₂) condition for 24 hr. Controls were monolayer cultures of each cell type maintained alone. We examined the secretion of VEGF by ELISA and influence of conditioned media on blood vessel growth (capillary-like structures) via an angiogenesis assay. Total RNA and protein were extracted from the HRPC and HUVEC (cultured and co-cultured) and analyzed for the expression of VEGF, VEGFR-2, NFκB and HIF-1α by RT-PCR and Western blotting. The cellular localization of NFκB and HIF-1α were studied by immunofluorescence and Western blotting.

Results

We found that hypoxia increased exogenous VEGF expression 4-fold in HRPC with a further 2-fold increase when cultured with HUVEC. Additionally, we found that hypoxia induced the expression of the VEGF receptor (VEGFR-2) for HRPC co-cultured with HUVEC. Hypoxia treatment significantly enhanced (8- to 10-fold higher than normoxia controls) VEGF secretion into media whether cells were cultured alone or in a co-culture. Also, hypoxia was found to result in a 3- and 2-fold increase in NFκB and HIF-1α mRNA expression by HRPC and a 4- and 6-fold increase in NFκB and HIF-1α protein by co-cultures, whether non-contacting or contacting.

Treatment of HRPC cells with hypoxic HUVEC-CM activated and promoted the translocation of NFκB and HIF-1α to the nuclear compartment. This finding was subsequently confirmed by finding that hypoxic HUVEC-CM resulted in higher expression of NFκB and HIF-1α in the nuclear fraction of HRPC and corresponding decrease in cytoplasmic NFκB and HIF-1α. Lastly, hypoxic conditioned media induced a greater formation of capillary-like structures (angiogenic response) compared to control conditioned media. This effect was attenuated by exogenous anti-human VEGF antibody, suggesting that VEGF was the primary factor in the hypoxic conditioned media responsible for the angiogenic response.

Conclusions

These findings suggest that intercellular communications between HRPC and HUVEC lead to the modulation of expression of transcription factors associated with the production of pro-angiogenic factors under hypoxic conditions, which are necessary for an enhanced neovascular response. Our data suggest that the hypoxia treatment results in the up-regulation of both mRNA and protein expression for VEGF and VEGFR-2 through the translocation of NFκB and HIF-1α into the nucleus, and results in enhanced HRPC-induced neovascularization. Hence, a better understanding of the underlying mechanism for these interactions might open perspectives for future retinal neovascularization therapy.

Cell Culture

Hypoxia Treatments

Monolayer cultures were incubated in a hypoxic environment by the following protocol. Briefly, media were pre-equilibrated with a hypoxic gas mixture (1%O₂, 5%CO₂, and 94%N₂). Cultures were replenished with the hypoxic medium, placed in the hypoxia incubator and maintained for 24 h to 48 h at 37°C. Control monolayer cultures were maintained in a normoxic environment, which refers to standard cell culture in a humidified incubator (20% O₂, 5% CO₂, and 75% N₂). All cells maintained in normoxic condition showed no morphological changes by light microscopy. The trypan blue exclusion test of cell viability showed cells to be >98% viable, and could then be sub-cultured normally. Cells incubated under normoxia conditions from the same batch, and subsequently sub-cultured, were used as control.

Preparation of HRPC/HUVEC conditioned medium (HRPC-CM/HUVEC-CM)

In a separate set of experiments, we evaluated the biological activity of conditioned medium (CM) obtained from HRPC (as HRPC-CM) and HUVEC (as HUVEC-CM) in normoxia and hypoxia condition [Figure 2]. HRPC cells were seeded in a 75 cm2 flasks with DMEM/F-12 with above mentioned supplements. HUVEC cells were cultured in EBM-2 medium with above mentioned supplements for 24 h. The cells were grown to 80% confluence then washed twice with the respective serum-free medium. Finally, the medium was replaced again with serum-free medium for another 24 h in normoxia and hypoxia condition. The CM from both cells under normoxia and hypoxia were collected, centrifuged at 1000 rpm for 10 min to remove cellular components and filtered through 0.2-μm filters (Millipore, Billerica, MA) before use. Medium from dishes under normoxic conditions was used as a control. The aliquots were stored at -80°C in a freezer.

Figure 2   ( Jump back to reading ◀ )

Quantifying secreted VEGF protein levels by ELISA

Changes in environmental factors in the cells culture can be differentially regulated by diverse signals. Therefore, we thought it was important to determine how hypoxia and co-culture interact in the regulation of VEGF secretion. Conditioned medium were collected, as described above, for the determination of secreted VEGF levels by ELISA. VEGF was assessed using commercially available human VEGF-ELISA kit (R&D Systems, Minneapolis, MN), performed according to manufacturer's instructions and normalized to protein content (BioRad, CA). The experiments were repeated three times in triplicates and are shown as mean ± SEM. The level of VEGF is expressed as pictograms of VEGF protein per milligram protein.

Capillary-like Tube Formation Assays

The effects secreted factors from HRPC-CM under normoxia and hypoxia were examined in the formation of capillary tube-like structures by HUVEC. The matrigel assay is a frequently used method for detection of angiogenesis in vitro. Briefly, the growth factor reduced Matrigel (BD Biosciences, San Jose, CA) was added to pre-chilled 48 well culture plates and allowed to polymerize at 37° C for 2 h. HUVECs (4 × 10⁵ cells/well) were plated in the Matrigel-coated plates and the cells cultured in HRPC-CM from normoxia and hypoxia conditioned for 24 h at 37°C with 5% CO₂. To study the role of VEGF in capillary tube formation, parallel cultures were treated with 20 ng/ml neutralizing anti-VEGF antibody, which were added at the time of HUVEC seeding. VEGF (50 ng/ml) was used as the positive control. Angiogenesis (capillary cord formation) were evaluated after culturing for 24 and 48 h and resulting tube-like capillary structures were examined by using an Olympus TMS inverted phase contrast microscope. Photomicrographs were documented with an Olympus digital camera.

Evaluating the expression of angiogenic protein by Western blot

The total cellular protein was isolated and used to examine the expression of angiogenic factors such as: VEGF, VEGF receptors (VEGFR-1 and VEGFR-2), and transcriptional factors (HIF-1α and NFκB). All these proteins play a major role in regulation of the retinal neovascularization. Briefly, after normoxia/hypoxia incubation for 24 h, media were collected and cells were washed with the PBS. The total proteins were extracted from the cell layers using M-PER Mammalian Protein Extraction Reagent (Pierce; CA) containing a protease inhibitor mixture (Sigma, CA). The protein extracts were centrifuged for 15 min at 15,000 rpm at 4° C and the supernatant were saved for further analysis. Total protein concentration of each sample was determined by the Bradford method as per manufacturer protocol (BioRad, CA) with bovine serum albumin as a standard. 30 μg total cellular proteins were analyzed on SDS/PAGE under reduced conditions. Before loading, all cell lysates were added to an equal volume of Laemmli sample buffer and heated at 95°C for 5 min, and samples were resolved by SDS-PAGE. The fractionated proteins were transferred to a polyvinylidene difluoride (PVDF) membrane. The membrane was blocked with 5% nonfat milk (BioRad, CA) in Tris buffered saline (TBS) with 0.05% Tween-20 (TBS-T) for 1 h at RT followed by incubation for 2 h with the primary antibodies in 2% milk in TBS-T. The membrane was washed three times with TBS-T for 10 min, and then incubated with the HRP-conjugated secondary antibodies, diluted in 2% milk in TBS-T, for 1 h, followed by three washes of 15 min in TBS-T. All membranes were probed with antibodies to β-actin to correct for loading and transfer differences among samples. The membranes were visualized by an enhanced chemiluminescence (ECL) Western Blotting Analysis System (Amersham).

Evaluating the expression of angiogenic transcriptions by RT-PCR

Total RNA was extracted from cultures (control monolayer and co-cultured under normoxia and hypoxia conditions) using the extraction solution Triazol reagent (Invitrogen, CA) according to the supplier's instructions. The purity and concentration of RNA were quantitated by absorbance measurement at 260/280 nm. The quality of RNA was also checked by the agarose gel electrophoresis. Total RNA (1 μg from each sample) was run on 1% agarose gel, then stained with ethidium bromide, and examined the 28S and 18S rRNA bands on a UV transilluminator. No significant degradations were observed in any total RNA samples. The extracted RNA was then analyzed by RT-PCR for VEGF, VEGF receptor-2 (VEGFR-2), and transcriptional factors (NF_kB and HIF-1α). The transcription of the housekeeping gene 18S RNA was used as internal controls (Ambion Inc. TX). The list of primers for RT-PCR to provided in Table 1. The reverse transcriptase (RT) reaction was performed with 1 μg of RNA, random hexamers, and Superscript II reverse transcriptase (Invitrogen, CA), according to the manufacturer's instructions. RT-PCR was conducted using the Gene Amp PCR System 2400 (Applied Biosystems) under the conditions described in the figure legends. The reaction mixture was first denatured at 95°C for 10 min, and the PCR conditions applied for 30 cycles. The amplified products were analyzed by electrophoresis on 1.8% agarose gels, and the ethidium bromide stained bands were visualized by UV transillumination. All images were saved in TIFF format. The levels of expression were quantitated by densitometer scanning. The target PCR products were analyzed by determining the intensity of band by gel-pro software (USA) which is a PC version of the image analysis software.

Expression and Translocation of NFκB and HIF-1α

We also examined the expression and translocation of NFκB and HIF-1α in HRPC cultured in HUVEC-CM (normoxia/hypoxia) by western blot analysis and immunofluorescence studies.

HRPC-CM induces neovascular responses by HUVEC

The effect of HRPC-CM from hypoxia or normoxia treated cells on the induction of capillary-like structures (angiogenesis or neovascularization) in vitro was investigated using a well-established model - the matrigel in vitro angiogenesis assay. Control HUVEC (Figure 3A) cultures were grown on matrigel-coated plates for 24 h in serum free conditions. Treatment of HUVEC with normoxic HRPC-CM showed very weak neovascular response (Figure 3B) although somewhat blunted compared to that induced by hypoxic HRPC-CM (Figure 3D). This effect was also attenuated by the simultaneous treatment with anti-VEGF antibodies (Figure 3C). On the other hand, HUVEC treated with hypoxic HRPC-CM for 24 h was found to have the highest neovascular response (Figure 3D) which was attenuated by the simultaneous treatment with anti-VEGF antibodies (Figure 3E). This suggests that the primary factor in the HRPC-CM responsible for the neovascular response may be VEGF.

Figure 3   ( Jump back to reading ◀ )

VEGF secretion by individual cultures and co-cultures

We measured the levels of VEGF released into the culture media of individual cultures of HRPC and HUVEC as well as co-cultures of these cell lines, following cultures under hypoxic conditions for 6, 12, 24 and 48 h. Control cultures were treated the same except they were maintained in normoxic conditions. Co-cultures were setup as shown in Figure 1. Media from the various cultures were collected and analyzed for the levels of total VEGF by ELISA. We found that hypoxia clearly enhanced VEGF secretion by all the cultures (Figure 4). For all cultures, the VEGF level in the media shows a time-dependent increase from 6 to 48 hours. At the 24 h time point, there was a five-fold increase in VEGF level for HRPC (Figure 4A) and HUVEC (Figure 4B). Our co-culture data showed an 8- to 10-fold increase for the non-contacting or contacting co-cultures (Figure 4C & D) for hypoxia treated cultures compared to normoxia controls. There was no significant difference noted for VEGF secretion by cells in co-culture with contacting or non-contacting between the two cell lines.

Figure 4   ( Jump back to reading ◀ )

Effect of hypoxia on VEGF and VEGFR-2 expression

We also examined the transcriptional expression of VEGF and VEGFR-2 as well as the expressions VEGF and VEGFR-2 protein by cultures maintained in hypoxia for 24 hr. Figure 5A and 5B show that hypoxia treatment of either HRPC or HUVEC cultured alone resulted in a 2- to 2.5-fold higher expression of VEGF and VEGFR-2 mRNA. However, hypoxia was able to induce about a six-fold higher expression of VEGF and a four-fold higher expression of VEGFR-2 by co-cultures of HUVEC and HRPC. There were no significant differences noted for VEGF and VEGFR-2 expression when the two cell lines were co-cultured with contacting or non-contacting set-ups. These findings for co-cultures appear to be an additive effect of the responses of the individual cell lines.

Figure 5   ( Jump back to reading ◀ )

We also observed that hypoxia induced an increase in VEGF and VEGFR-2 protein expression following western blot analysis [Figure 5E and 5F]. Our Western blot data support mRNA data with the low protein expression levels of VEGF and VEGFR-2 in HRPC and HUVEC were cultured under normoxia environment. We have found that exogenous VEGF expression is increased in HRPC and HUVEC cultured alone under hypoxia condition as compared with normoxia. There was a four-fold increase in VEGF expression by HRPC exposed to hypoxia for 24 h, as compared with normoxia. When HRPC were cultured in presence of HUVEC under hypoxia, there was a further two-fold increase in VEGF expression, compared to HRPC cultured alone under hypoxic conditions. In addition, our data indicated that hypoxia was capable of inducing not only VEGF expression, but also its receptor - VEGFR-2 when HRPC were co-cultured with HUVEC for 24 h. Data from RT-PCR and Western blot analysis are consistent with our VEGF ELISA results, and showed an increased in VEGF in all the hypoxia treatments and co-cultures compared with normoxia controls. These results suggest the association of VEGF-VEGFR-2 signaling regulate the retinal neovascularization.

Role of hypoxia on the expression of NFκB and HIF-1α

Cultures of HRPC and HUVEC (alone or as co-cultures), maintained under normoxia (control) and hypoxia, were harvested, then total RNA and protein extracted. These extracts were used to measure the expression of transcription factors NFκB and HIF-1α. Figures 6A shows that either cell line alone constitutively expressed low levels of NFκB mRNA under normoxia conditioned. Hypoxia treatment of either HRPC or HUVEC resulted in a three- and two-fold increase in NFκB mRNA expression compared with normoxia controls, respectively. Additionally, we found that under control (normoxia) conditions, HRPC co-cultured with HUVEC as either contacting or non-contacting cultures, showed significantly higher levels of NFκB compared to single-cell cultures maintained similarly. In contrast, hypoxia treatment resulted in a much higher level of expression for NFκB (5-fold), compared to control co-cultures. There was more than a seven-fold induction of NFκB seen for co-cultures under hypoxia, compared to control co-cultures. To confirm our RT-PCR results, we also measured the protein expression for these transcriptional factors using western blotting. Total cellular proteins were isolated from culture cells maintained under hypoxia and normoxia (controls) for 24 h. Figure 6C shows western blots for NFκB; we found a 2.6-fold increase in NFκB expression by HRPC exposed to hypoxia for 24 h, compared with normoxia controls. A similar observation was found for hypoxia-treated HUVEC; there was a three-fold increase in NFκB expression. Lastly, we found that hypoxia was capable of significantly increasing the levels of NFκB expression by four-fold for co-cultures, whether non-contacting or contacting.

Figure 6   ( Jump back to reading ◀ )

Figure 7 showed the effects of hypoxia treatment on HIF-1α expression for single and co-cultured cells. Similar to the results found for NFκB, HIF-1α mRNA expression showed a two-fold increase in following hypoxia treatment for both HRPC and HUVEC cultured alone (Figure 7A). We also found that co-culture of these two cell lines resulted in a two-fold higher level of HIF-1α mRNA compared to single-cell HRPC cultures maintained similarly. Hypoxia treatment induced a four-fold increase in HIF-1α expression by co-cultures, compared to normoxia control co-cultures. Figure 7C shows that hypoxia resulted in a three-fold increase in HIF-1α protein by either HRPC or HUVEC cultured alone, but increased HIF-1α expression by six-fold for co-cultures. Contacting or non-contacting co-cultures had similar responses.

Figure 7   ( Jump back to reading ◀ )

Effect of HUVEC-CM on the Subcellular Localization of NFκB and HIF-1α

When HRPC single-cells cultures were treated for 24 h with either HUVEC-CM collected under normoxia (control) or hypoxia conditions, we observed significant effects on possible nuclear translocation of NFκB and HIF-1α. Figure 8A and 8B show immunofluorescent photomicrographs of HRPC treated with normoxic or hypoxic HUVEC-CM, respectively. With normoxic HUVEC-CM, there was higher NFκB immunofluorescence observed in the cytoplasm. This finding was subsequently confirmed by measuring the expression of NFκB protein in cytosolic and nuclear extracts using western blotting analyses (Figure 8C). Treatment with hypoxic HUVEC-CM resulted in the activation of NFκB and promoted its translocation to the nuclear compartment (Figure 8B) and resulted in higher expression of NFκB in the nuclear fraction of HRPC and corresponding decrease in cytoplasmic NFκB (Figure 8D).

Figure 8   ( Jump back to reading ◀ )

A similar finding was seen for the subcellular localization of HIF-1α (Figure 9). HRPC treated with the control normoxic HUVEC-CM showed HIF-1α was localized primarily in the cytoplasm (Figure 9A). Subsequently, treatment with hypoxic HUVEC-CM resulted in the translocation of HIF-1α to the nucleus (Figure 9B). Western blotting of proteins extracts from the cytoplasmic and nuclear fraction also showed a similar change. Treatment with hypoxic HUVEC-CM resulted in a higher expression of HIF-1α in the nuclear fraction compared to control cultures treated with normoxic HUVEC-CM (Figure 9C and 9D). Thus, these studies suggest that the hypoxia HUVEC-CM promoted the translocation of both NFκB and HIF-1α to the nuclear compartment of HRPC.

Figure 9   ( Jump back to reading ◀ )