Bradykinin type 2 receptor deficiency alters vascular endothelial growth factor and atrial natriuretic peptide levels in early-stage diabetic retinopathy

Experimental animals

The study involved 2–6 months old male diabetic C57BL/6-Ins2^Akita/J (Akita, n = 18; RRID: IMSR_JAX: 003548), diabetic B2R knock-out B6.Cg-Ins2^Akita/Bdkrb2^tm1Jfh/SmiJ (Akita/B2R-KO, n = 18; RRID: IMSR_JAX: 006860), and their non-diabetic controls: wild type C57BL/6J (WT, n = 18; RRID: IMSR_JAX: 000664) and B2R knock-out B6; 129S7-Bdkrb2^tm1Jfh/J (B2R-KO, n = 18; RRID:IMSR_JAX: 002641 backcrossed to C57BL/6J) mice. Mice were procured from The Jackson Laboratory (Bar Harbor, USA), bred at the Croatian Institute for Brain Research, University of Zagreb School of Medicine and backcrossed for over 15 generations. The mice were housed under standardized conditions (22 °C ± 2 °C, 12-h light–dark cycle) with free access to food and water. Genotypes were confirmed with PCR upon separation by sex at 1.5 months of age.

Assessment of glycemic status

At 2 and 6 months of age, tail blood samples were collected to measure glucose concentration, HbA1c levels, and conduct intraperitoneal glucose tolerance test (IPGTT), as previously described. ³⁶ Blood glucose was measured before and after a 6-h daytime fasting period using the Rightest glucose meter (GM550; Bionime, Taiwan), while HbA1c levels were determined using the HbA1CNow^® Self Check (PTS Diagnostics, USA). Additionally, WT and B2R-KO mice underwent IPGTT with 1 g/kg body weight of 20% glucose solution (Braun, Germany) administered via intraperitoneal injection (i.p.), followed by blood glucose measurements at 15, 30, 60, and 120 min. Akita and Akita/B2R-KO mice were exempted from the IPGTT due to risk of inducing diabetic coma.

Non-invasive measurement of blood and intraocular pressure

Blood pressure and intraocular pressure (IOP) were measured in anesthetized mice at 2 and 6 months of age, following established protocols. ³⁶ Anesthesia was induced with 4% isoflurane (Isoflurane; Piramal Critical Care, USA) in a 70% N₂/30% O₂ gas mixture, then maintained at 2% isoflurane. Normothermia (38.5 °C ± 0.5 °C) was ensured using a feedback-controlled heating pad (medres—medical research GmbH, Germany). Respiratory rate (90–100 breaths/min) was monitored with an optical probe (medres—medical research GmbH, Germany). IOP was measured with a Tonolab tonometer (Icare, Finland), obtaining six measurements per eye within 5 min post-anesthesia. After the measurements, ophthalmic lubricant (Recugel^®; Bausch and Lomb, Canada) was applied to prevent corneal desiccation. Blood pressure was measured using the CODA Surgical Monitor system (Kent Scientific, USA), with 20 measurement cycles, 5-s intervals between cycles, and a 20-s occlusion cuff relaxation period.

Fluorescein angiography

Fluorescein angiography was performed at 2 and 6 months of age, preceding MRI sessions, using a smartphone ophthalmoscopy setup. ³⁶ Mice were anesthetized in a heated chamber with 4% isoflurane in a 70% N₂/30% O₂ gas mixture, then maintained at 2% isoflurane. For optimal visualization, vibrissae were taped and pupils were dilated with 1% tropicamide eye drops (Mydriacyl; Alcon, Switzerland). Body temperature was continuously monitored using a CODA rectal temperature probe. Artificial tears (0.4%, Vizol S; JGL d.d., Croatia) were applied, followed by a subcutaneous injection of 10 µl of 10% sodium fluorescein (NAF; Alcon, Switzerland). Images were captured with a smartphone (Samsung, South Korea) mounted onto an Olympus SZ-STU2 dissecting microscope (Olympus Corp., Japan), used in combination with a handheld 90D optical lens (Volk Optical, Inc, USA). An excitation/emission filter pair for fluorescein angiography was also mounted onto the microscope via custom adjustable holders, the excitation filter (483 nm, Edmund Optics, USA) positioned in front of the light source, and the emission filter (535 nm, Edmund Optics, USA) placed in front of the smartphone camera.

Semi-automated measurements of retinal vessel diameter, tortuosity, and capillary perfusion area were performed on fluorescein angiograms, using an open-source machine-learning plugin in FIJI/ImageJ v.1.53c. Vessel diameters were measured 0.5–1.0 disc diameters from the optic disc margin and normalized to optic disc diameter. The capillary area was analyzed in a predefined region of interest (ROI) and expressed as arbitrary units.

Magnetic resonance imaging

MRI was conducted using a 7T system (BioSpec 70/20 USR with Paravision 6.0.1; Bruker Biospin, Germany) configured in a Tx/Rx setup with an 86 mm transmit volume coil and a two-element mouse brain surface receive coil, following protocols for anatomical and angiographic imaging, data processing, and analysis as previously described.^36,37 Anesthesia was induced with 4% isoflurane (Piramal Critical Care, Germany) in a 30% O₂/70% N₂ mixture and maintained at 1.5%–2% isoflurane. Respiratory rate was kept at 80–100 breaths/min, monitored with an optical probe (medres—medical research GmbH, Germany), and body temperature was maintained at 37 °C ± 0.5 °C with an MR-compatible rectal temperature probe and a feedback-controlled heating pump (medres—medical research GmbH, Germany). Imaging sessions followed fluorescein angiography. T2-weighted eye images were acquired using a TurboRARE sequence (TE/TR = 28/2500 ms; resolution 0.06 × 0.06 mm; slice thickness 0.4 mm; interslice distance 0.1 mm; nine slices), with the central slice bisecting the eye’s center and optic nerve sagittally to minimize partial-volume effects due to retinal curvature. T2-maps were generated using an MSME sequence (TE/TR = 9/2500 ms; resolution 0.08 × 0.08 mm; slice thickness 0.4 mm; interslice distance 0.1 mm; 25 slices), and T2 values were calculated with Paravision 6.0.1 software. Additionally, angiographic images were obtained using a 3D FLASH (3D-TOF) sequence (TE/TR = 2.1/12.0 ms, flip angle 25°, resolution 0.08 × 0.08 × 0.2 mm³). Each imaging session lasted ~60 min/mouse.

Images were analyzed semi-automatically using FIJI/ImageJ1.53i software by investigators blinded to the experimental groups. Retinal thickness was measured using a single T2-weighted MRI slice intersecting the eye’s center and optic nerve. Measurements were taken at predefined points 30° above and below the optic nerve head. Precise segmentation was performed on subsampled T2-weighted images with sub-pixel accuracy.

Quantitative measurements of T2 relaxation time (spin-spin relaxation) in individual voxels of the retina using a built-in macro in ImageJ calculated the average signal intensity across the entire retina volume using T2-weighted images and T2 maps. The T2-weighted image defined an ROI marking the eye’s edges. The displacement between T2-weighted images and T2 maps was calculated. The macro used ROIs for measuring retinal thickness at three points above and below the optic nerve, adjusted the T2 map resolution to match the T2-weighted image, and moved the ROIs according to the calculated displacement. Average T2 values were measured with integrated subsampling (factor 3) in individual ROIs and within retinal surfaces outlined by two polygonal shapes generated from the individual ROIs below and above the optic nerve.

Angiograms from 3D-TOF scans were analyzed via maximum intensity projections (MIP) to visualize and assess vasculature with high blood flow velocity. Vessel volume for each ophthalmic artery was assessed using a standardized cuboidal 3D region of interest (ROI). Analysis was performed using an open-source machine learning plugin within FIJI/ImageJ-v.1.53c, which included thresholding and image binarization, as previously described.^37,38 Results were expressed as vessel volume (mm³).

Blood–retinal barrier permeability assessment using Evans blue and sodium fluorescein

Evans blue (EB; Merck, Germany) and sodium fluorescein (NAF; Alcon, Switzerland) were used to assess blood-retinal barrier permeability at 6 months of age, as described previously. ³⁶ Permeability was indirectly assessed by measuring retinal tissue fluorescence intensity ex vivo with the IVIS^® Spectrum system (Perkin Elmer, USA). Two hours before imaging, mice received a tail vein injection of 2% EB in saline (50 mg/kg). Ten minutes before imaging, 10 µl of 10% NAF was administered subcutaneously. Upon completion of the circulation period, mice were anesthetized with 2.5% tribromomethanol (Avertin; Sigma–Aldrich) at 250 mg/kg i.p., and perfused transcardially with 30 ml of phosphate-buffered saline (PBS). Retinas were isolated from enucleated eyes under subdued lighting and placed in a Petri dish with 10 µl PBS for imaging. EB fluorescence was captured at 640 nm excitation and 720 nm emission, while NAF fluorescence was recorded at 465 nm excitation and 520 nm emission.

To measure the concentration of applied dyes, EB and NAF were extracted from collected retinal samples. Retinas were placed in tubes with 100 µl of PBS at room temperature and homogenized using an ultrasonic homogenizer (Q55 Sonicator^®; Qsonica, USA) with short pulses. Equal volumes of 100% trichloroacetic acid (TCA; Merck, Germany) were added to the homogenized tissue. The homogenization process was repeated, and samples were incubated for 1 h at 37 °C in an HB-96D Witeg tube heater (Witeg Labortechnik GmbH, Germany). After incubation, samples were centrifuged at 12,000g for 20 min at 25 °C. For each retina, 50 µl of the supernatant was pipetted into three wells of a 96-well plate (Falcon, USA). Fluorescence of EB was measured at excitation/emission wavelengths of 640/720 nm, and NAF at 465/520 nm. EB and NAF concentrations were calculated using standard fluorescence curves in 50% TCA.

Retinal tissue isolation for ELISA, Western blot, and qPCR

Mice were anesthetized with an i.p. of 2.5% tribromomethanol (Avertin; Sigma–Aldrich, USA) at a dose of 250 mg/kg, and transcardially perfused with PBS. Eyes were enucleated, cornea and lens removed, and retinas isolated for further processing using a modification of a previously described method. ³⁹ Details are provided in Supplemental Methods.

ELISA

Bradykinin, VEGF, and ANP levels in retinal lysates (diluted 4×) were quantified using ELISA kits (ab100662 and ab136936; Abcam, Cambridge, UK; EIA-ANP; RayBiotech, Norcross, GA, USA). All standards and samples were assayed in duplicate, with absorbances read at 450 nm using a GloMax^® plate reader (Promega, Madison, WI, USA). Concentrations were calculated according to the manufacturer’s instructions and expressed (pg/ml).

Western blot

Relative protein expression of B2R, bradykinin type 1 receptor (B1R), and GC-A was analyzed by Western blot, as previously described.^40,41 See Supplemental Methods for details.

qPCR

Relative expression of genes, including genes for B1R (Bdkrb1), B2R (Bdkrb2), ANP (Nppa), GC-A/NPRA (Npr1), VEGF (Vegfa), and VEGFR-2 (Kdr), was determined using qPCR, as previously described.^40,41 Details are provided in Supplemental Methods.

Immunohistochemistry and data analysis

Animals were anesthetized with 2.5% tribromoethanol, transcardially perfused, and fixed. The eyes were enucleated, processed, and analyzed as described previously.^36,37 Staining for B2R, B1R, VEGF, VEGFR-2, ANP, and GC-A was performed on cryosections from four animals per group. For colocalization, second primary antibodies were used against cluster of differentiation 31 (CD31) as an endothelial cell marker, cellular retinaldehyde-binding protein (CRALBP) as a Müller cell marker, and neuronal nuclei (NeuN) as a ganglion cell marker. Nuclei were stained with DAPI. Four images per section were captured at 40× magnification using the Olympus FLUOVIEW FV3000 confocal microscope with the accompanying software FV31-SW Fluoview (Olympus, Japan). Image analysis was performed with ImageJ 1.53i. Three regions of interest (ROIs) per image were analyzed for protein expression. Colocalization was imaged at 60× magnification. Retinal ganglion cell counts were performed in ROIs enclosing the retinal ganglion cell layer. Details are provided in Supplemental Methods and Table 1.

Statistical analysis and sample size

Based on preliminary data, a power analysis using G*Power software (Heinrich-Heine-Universität Düsseldorf, Germany) determined that six animals per genotype were needed to detect significant differences in retinal thickness using a two-sided Mann–Whitney U test (power = 0.8, α = 0.05). ⁴² Data were analyzed using GraphPad Prism 9.4.0 (GraphPad Software, USA). Normality was assessed with the Shapiro–Wilk test. A mixed-model two-way analysis of variance (ANOVA) was used to assess baseline glucose, fasting glucose, IPGTT, HbA1c, body weight, IOP, systolic, diastolic, and mean arterial pressure, heart rate, arteriole and venule diameter and tortuosity, capillary area, ophthalmic artery volume, retinal thickness, and average T2 value, with age and genotype as variables. Post hoc comparisons between genotypes were made using Tukey’s test, and within-genotype using Šidak’s post hoc test. The expression of Bdkrb2 gene, B2R band signal intensity, and B2R+-fluorescent signal area were assessed using Student’s t-test. One-way ANOVA with multiple comparison corrections, controlling for false discovery rate at α = 0.05, was used to evaluate EB and NAF extravasation and concentration, expression of Bdkrb1, Nppa, Npr1, Vegfa, and Kdr genes, bradykinin, ANP and VEGF concentration, B2R, B1R and GC-A band signal intensity, B1R+, ANP+, GC-A+, VEGF+, and VEGFR-2+ fluorescent signals, and NeuN+ ganglion cell count. Statistical significance was set as p < 0.05.

B2R deficiency modulates basic metabolic parameters and intraocular pressure in diabetes

To assess the effects of diabetes and B2R deficiency on glycemic status, body mass, IOP, and blood pressure, we examined mice at 2 and 6 months of age (Figure 1). At 2 months old, B2R-KO and WT mice were normoglycemic, while Akita, and Akita/B2R-KO mice exhibited hyperglycemia, which persisted at 6 months (p < 0.0001; Figure 1(a)). Akita and Akita/B2R-KO mice had elevated HbA1c levels (p < 0.0001), with Akita/B2R-KO mice showing worse long-term glycemic status (p < 0.01; Figure 1(b)). IPGTT results indicated altered glucose metabolism in young B2R-KO mice (Figure 1(c)). Diabetic mice had significantly lower body mass at 2 months (p < 0.01). All groups gained weight by 6 months (p < 0.0001). B2R-KO mice gained more body mass than WT (p < 0.0001), and Akita/B2R-KO more than Akita mice (p < 0.001; Figure 1(d)). IOP was higher in Akita mice compared to Akita/B2R-KO mice (p < 0.01), with no significant differences between non-diabetic groups (Figure 1(e)). Blood pressure measurements showed no differences in systolic, diastolic, and mean arterial pressures at 2 months (Figure 1(f)–(h)). At 6 months, Akita mice had lower systolic, diastolic, and mean arterial pressures compared to younger Akita and WT animals (p < 0.05 and p < 0.001, respectively), with no differences among other groups. In summary, B2R deficiency worsened long-term glycemic control in diabetic mice and promoted weight gain across genotypes. It also mitigated diabetes-induced elevation in intraocular pressure, without affecting systemic blood pressure. These results indicate that B2R modulates metabolic and IOP homeostasis in a context-dependent manner.

OPEN IN VIEWER

Figure 1.

Diabetes and B2R deficiency impact glycemic status, body weight, and intraocular pressure. (a) Elevated glucose concentrations at 2 and 6 months and (b) higher percentage of HbA1c in diabetic mice, further elevated by B2R deficiency. (c) B2R deficiency alters glucose uptake, as shown by the intraperitoneal glucose tolerance test in young non-diabetic B2R mice. (d) Diabetic mice exhibit reduced body weight, while B2R deficiency significantly increases body weight in older mice. (e) B2R deficiency lowers intraocular pressure in diabetic mice. (f–h) Diabetes lowers systolic, diastolic, and mean arterial pressure in older animals. Values are presented as median and interquartile range or mean ± standard error. Statistically significant differences among genotypes, using mixed model ANOVA with Tukey’s post hoc test: *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. Differences within genotypes, using mixed model ANOVA with Sidak post hoc test: ^#p < 0.05. ^##p < 0.01. ^####p < 0.0001. ns: nonsignificant.

Both B2R deficiency and diabetes impair retinal microvascular perfusion

Fluorescein angiography showed no obvious leakage in any group (Figure 2(a)), suggesting no neovascularization or locally visible BRB disruption, though limitations of resolution and signal-to-noise ratio should be considered. The effects of diabetes and B2R deficiency were quantitatively evaluated on retinal vessel diameter and tortuosity, as well as capillary area. Capillary area analysis revealed that WT mice had a larger capillary network perfusion area than Akita and B2R-KO (p < 0.0001) mice (Figure 2(b)). Aging decreased the perfused capillary area in WT mice, but it remained larger than in Akita (p < 0.0001) and B2R-KO (p < 0.001) mice. B2R deficiency did not further affect the perfused capillary area in diabetic animals. At 2 months, WT mice had significantly wider arterioles compared to Akita mice (p < 0.05), but aging reduced arteriolar diameter in WT mice, nullifying group differences (Figure 2(c)). Arteriolar tortuosity was similar across 2 month olds but was significantly lower in 6-month-old Akita/B2R-KO mice compared to Akita (p < 0.01) and B2R-KO (p < 0.001) mice (Figure 2(d)). Venular measurements at 2 months showed wider venules in WT mice compared to Akita (p < 0.01) and B2R-KO (p < 0.05) mice, with Akita/B2R-KO mice having wider venules than Akita mice (p < 0.05). This difference disappeared with age (Figure 2(e)). Venular tortuosity showed no significant differences at either age (Figure 2(f)). In short, both diabetes and B2R deficiency were associated with a reduction in retinal capillary perfusion area and venular diameter, indicating impaired microvascular function.

OPEN IN VIEWER

Figure 2.

Diabetes and B2R deficiency reduce retinal venular diameter and capillary network perfusion area. (a) Representative fundus angiograms of 2- and 6-month-old animals. (b) Diabetes and B2R deficiency reduce the retinal capillary perfusion network area. (c) B2R deficiency does not affect, but diabetes reduces the diameter of retinal arterioles. (d) B2R deficiency affects retinal arteriolar tortuosity in diabetic mice with age. (e) Diabetes and B2R deficiency impact retinal venular diameter. (f) Diabetes and B2R deficiency do not affect retinal venular tortuosity. Values are presented as median and interquartile range. Statistically significant differences among genotypes, using mixed model ANOVA with Tukey’s post hoc test: *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. Differences within genotypes, using mixed model ANOVA with Sidak post hoc test: ^#p < 0.05. ns: nonsignificant.

B2R deficiency exacerbates vascular permeability but attenuates neurodegeneration in diabetes

The impact of diabetes and B2R deficiency on retinal morphology, edema, and ophthalmic artery perfusion was assessed using high-resolution MRI imaging (Figure 3(a)). At 6 months, mean retinal thickness was significantly lower in the Akita and non-diabetic B2R-KO groups compared to WT (p < 0.05), but higher in Akita/B2R-KO group than in Akita (p < 0.01) and non-diabetic B2R-KO groups (p < 0.01; Figure 3(b)). T2 values were lower in Akita/B2R-KO compared to Akita and non-diabetic B2R-KO groups (p < 0.05; Figure 3(c)). 3D MIP reconstructions showed significantly reduced ophthalmic artery volume in B2R-KO mice compared to WT (p < 0.05), but this was not observed in diabetic Akita/B2R-KO mice compared to Akita (Figure 3(d)).

OPEN IN VIEWER

Figure 3.

B2R deficiency exacerbates vascular permeability but ameliorates ganglion cell loss and retinal thinning caused by diabetes. (a) Representative T2-weighted images of the eye. (b) Diabetes and B2R deficiency significantly reduce retinal thickness, while B2R deficiency in diabetic animals cancels out this reduction. (c) B2R deficiency significantly reduces mean T2 values in diabetic retinas. (d) B2R deficiency reduces ophthalmic artery volume in non-diabetic mice. (e) B2R deficiency significantly increases high-molecular-weight EB extravasation, independent of the presence of diabetes. (f) B2R deficiency significantly increases retinal EB concentrations, independent of diabetes. (g) Diabetes and B2R deficiency significantly increase low-molecular-weight NAF extravasation. (h) Diabetes and B2R deficiency significantly increase retinal NAF concentrations. (i) Diabetes and B2R deficiency significantly reduce ganglion cell numbers, while B2R deficiency in diabetes mitigates this reduction. Values are presented as median and interquartile range. Statistically significant differences among genotypes, using mixed model ANOVA with Tukey’s post hoc test or with multiple comparison correction controlling the false discovery rate (α = 0.05): *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. ns: nonsignificant.

BRB permeability was assessed by EB and NAF fluorescence. Increased EB fluorescence, indicating albumin extravasation, was shown in Akita/B2R-KO and B2R-KO animals compared to controls (p < 0.05; Figure 3(e)), consistent with higher EB concentration in these groups (p < 0.01, p < 0.05; Figure 3(f)). NAF fluorescence indicated more low-molecular-weight molecule leakage in Akita/B2R-KO compared to Akita (p < 0.001) and in Akita compared to WT (p < 0.01; Figure 3(g)). B2R deficiency alone also increased NAF fluorescence (p < 0.01) but less so than in Akita/B2R-KO mice (p < 0.001), consistent with NAF concentration findings (Figure 3(h)).

Diabetes and B2R deficiency significantly reduced ganglion cell numbers (p < 0.0001, p < 0.01; Figure 3(i)), while B2R deficiency in diabetes mitigated this reduction (p < 0.05). To summarize, B2R deficiency increased blood-retinal barrier permeability in both non-diabetic and diabetic mice, as shown by enhanced leakage of both albumin-bound and small molecules. However, in diabetic mice, B2R deficiency partially protected against retinal thinning and ganglion cell loss, suggesting a complex, differential role of B2R signaling under diabetic versus non-diabetic conditions.

B2R deficiency upregulates B1R expression

To assess the impact of diabetes and B2R deficiency on retinal bradykinin concentrations quantitative ELISA was performed, revealing no differences among the groups (Figure 4(a)). qPCR, immunohistochemistry, and Western blot were used to evaluate the impact of diabetes on Bdkrb2 gene expression and B2R protein levels in Akita and WT mice, as Bdkrb2 transcription was absent in the Akita/B2R-KO and B2R-KO groups. Bdkrb2 mRNA expression was higher in Akita mice compared to WT (p < 0.05; Figure 4(b)). B2R protein expression, indicated by the B2R+ immunofluorescent signal area, was significantly elevated in Akita mice (p < 0.001; Figure 4(c) and (e)), corroborated by Western blot results (p < 0.05; Figure 4(d)). Bdkrb1 gene expression and B1R protein levels were evaluated in all study groups. Bdkrb1 mRNA expression was unchanged in Akita but higher in B2R-KO mice compared to WT (p < 0.0001) and in Akita/B2R-KO mice compared to Akita (p < 0.001; Figure 5(a)). B1R protein expression, indicated by B1R+ immunofluorescent signal area, was significantly elevated in both Akita (p < 0.001) and B2R-KO mice compared to WT mice (p < 0.01; Figure 5(b) and (d)), corroborated by Western blot results (p < 0.0001 and p < 0.05; Figure 5(c)), and further elevated in Akita/B2R-KO mice compared to both Akita (p < 0.0001 and p < 0.05) and B2R-KO mice (p < 0.0001 and p < 0.0001; Figure 5(b)–(d)). In brief, while bradykinin levels remained unchanged, diabetes increased both B2R and B1R expression, while B2R deficiency led to a strong rise in B1R expression, especially in diabetic mice, pointing to a compensatory response between the two receptors.

OPEN IN VIEWER

Figure 4.

Diabetes increases B2R expression. (a) Neither diabetes nor B2R deficiency influences bradykinin concentration in the retina. (b) Diabetes increases retinal B2R gene (Bdkrb2) expression (n = 6). (c) Diabetes significantly increases the B2R+ fluorescent signal intensity in the retina (n = 4). (d) Diabetes significantly increases retinal B2R expression (n = 4). (e) Representative double immunofluorescence labeled retinal tissue sections confirm the expression of B2R (green) on endothelial cells (CD31, red). DAPI-positive nuclei (blue). Scale bar: image 50 μm, insert 10 μm. Values are presented as median and interquartile range. Statistically significant differences among all genotypes were determined using one-way ANOVA with multiple comparison corrections controlling the false discovery rate (α = 0.05), and differences between Akita and WT were determined using Student's t-test: *p < 0.05. ***p < 0.001. ns: nonsignificant.

OPEN IN VIEWER

Figure 5.

Diabetes and B2R deficiency increase B1R expression. (a) B2R deficiency increases retinal B1R gene (Bdkrb1) expression (n = 6). (b) Diabetes and B2R deficiency significantly increase the B1R+ fluorescent signal intensity in the retina (n = 4). (c) Diabetes and B2R deficiency significantly increase retinal B1R expression (n = 4). (d) Representative double immunofluorescence labeled retinal tissue sections confirm the expression of B1R (green) on endothelial cells (CD31, red). DAPI-positive nuclei (blue). Scale bar: image 50 μm, insert 10 μm. Values are presented as median and interquartile range. Statistically significant differences among all genotypes were determined using one-way ANOVA with multiple comparison corrections controlling the false discovery rate (α = 0.05): *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. ns: nonsignificant.

B2R deficiency downregulates VEGF and VEGFR-2 protein synthesis

To assess the effects of diabetes and B2R deficiency on retinal VEGF concentrations, quantitative ELISA was performed. The analysis showed significantly reduced VEGF levels in Akita/B2R-KO compared to Akita mice (p < 0.001), and in B2R-KO compared to WT mice (p < 0.01; Figure 6(a)). Interestingly, while Vegfa gene expression was significantly higher in diabetic Akita mice compared to WT controls (p < 0.001), it was even higher in B2R-KO mice compared to WT controls (p < 0.001) and in Akita/B2R-KO compared to Akita mice (p < 0.05; Figure 6(b)). VEGF protein expression, indicated by VEGF+ fluorescent signal areas, was lower in Akita/B2R-KO and B2R-KO mice compared to their controls (p < 0.0001), suggesting reduced protein synthesis (Figure 6(c) and (d)).

OPEN IN VIEWER

Figure 6.

B2R deficiency decreases VEGF and VEGFR-2 protein expression. (a) B2R deficiency decreases VEGF concentration in the retina. (b) B2R deficiency and diabetes increase retinal VEGF gene (Vegfa) expression (n = 6). (c) B2R deficiency decreases VEGF+ fluorescent signal intensity in the retina (n = 4). (d) Representative double immunofluorescence labeled retinal tissue sections confirm the expression of VEGF (green) on endothelial cells (CD31, red). (e) Diabetes and B2R deficiency increase retinal VEGFR-2 gene (Kdr) expression (n = 6). (f) B2R deficiency significantly decreases VEGFR-2+ fluorescent signal intensity in the retina (n = 4). (g) Representative double immunofluorescence labeled retinal tissue sections confirm the expression of VEGFR-2 (green) on endothelial cells (CD31, red). Scale bar: image 50 μm, insert 10 μm. Values are presented as median and interquartile range. Statistically significant differences among all genotypes were determined using one-way ANOVA with multiple comparison corrections controlling the false discovery rate (α = 0.05): *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. ns: nonsignificant.

Concerning Kdr, VEGFR-2 gene expression, diabetes, and B2R deficiency increased mRNA levels, with higher expression in Akita/B2R-KO mice compared to Akita mice (p < 0.01), and in B2R-KO mice compared to WT mice (p < 0.0001). Interestingly, Akita/B2R-KO mice had lower Kdr expression than B2R-KO mice (p < 0.001; Figure 6(e)). VEGFR-2 protein expression was, however, reduced in Akita/B2R-KO and B2R-KO groups compared to their controls (p < 0.05; Figure 6(f) and (g)). To summarize, despite higher Vegfa and Kdr gene expression, B2R deficiency led to lower VEGF and VEGFR-2 levels in both non-diabetic and diabetic mice, pointing to a regulatory effect on protein synthesis.

B2R deficiency upregulates ANP and GC-A expression

ELISA analysis revealed that diabetes increased ANP levels in Akita compared to WT (p < 0.05) and in Akita/B2R-KO compared to B2R-KO mice (p < 0.001). B2R deficiency alone also elevated ANP in the B2R-KO group (p < 0.05; Figure 7(a)). NppA, ANP gene expression was higher in Akita compared to WT (p < 0.05) and in both diabetic and non-diabetic B2R-deficient animals compared to Akita (p < 0.01) and WT (p < 0.001) controls, respectively (Figure 7(b)). Fluorescent ANP+ signal area (Figure 7(c) and (d)) confirmed the ANP concentration analysis (Akita vs WT p < 0.05; Akita/B2R-KO vs Akita p < 0.001 and vs B2R-KO p < 0.05; B2R-KO vs WT p < 0.01). The increased ANP expression mainly colocalized with Müller cells (see Supplemental Figure 1), and less with ganglion (see Supplemental Figure 2) and endothelial cells (Figure 7(d)).

OPEN IN VIEWER

Figure 7.

B2R deficiency increases ANP and GC-A expression, and diabetes further modulates it. (a) Diabetes and B2R deficiency increase ANP concentration in the retina. (b) Diabetes and B2R deficiency increase retinal ANP gene (Nppa) expression (n = 6). (c) Diabetes and B2R deficiency significantly increase ANP+ fluorescent signal intensity in the retina (n = 4). (d) Representative double immunofluorescence labeled retinal tissue sections confirm the expression of ANP (green) on endothelial cells (CD31, red). (e) Diabetes and B2R deficiency increase retinal guanylyl cyclase/NPR-A gene (Npr1) expression (n = 6). (f) Diabetes and B2R deficiency increase GC-A+ fluorescent signal intensity in the retina (n = 4). (g) Diabetes and B2R deficiency significantly increase retinal GC-A expression (n = 4). (h) Representative double immunofluorescence labeled retinal tissue sections confirm the expression of GC-A (green) on endothelial cells (CD31, red). Scale bar: image 50 μm, insert 10 μm. Values are presented as median and interquartile range. Statistically significant differences among all genotypes were determined using one-way ANOVA with multiple comparison corrections controlling the false discovery rate (α = 0.05): *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. ns: nonsignificant.

Similar trends were found for Npr1, GC-A gene expression. Akita mice had higher expression than WT (p < 0.05), Akita/B2R-KO mice significantly higher than both Akita (p < 0.001) and B2R-KO mice (p < 0.05), and B2R-KO compared to WT mice (p < 0.01; Figure 7(e)). The GC-A+ signal area was also greater in Akita mice compared to WT (p < 0.05), Akita/B2R-KO mice compared to both Akita (p < 0.0001) and B2R-KO (p < 0.0001), and B2R-KO compared to WT mice (p < 0.05; Figure 7(f)). Western blot analysis confirmed these immunofluorescent findings (Akita vs WT: p < 0.05; Akita/B2R-KO vs Akita: p < 0.0001 and vs B2R-KO: p < 0.01; B2R-KO vs WT: p < 0.001; Figure 7(g)). In conclusion, B2R deficiency significantly upregulated ANP and GC-A expression, likely as a compensatory or modulatory response, both in non-diabetic and diabetic animals, but more prominently in the latter. This increased expression was especially seen in Müller cells.