**3. Results**

### *3.1. Hyaluronan is Elevated in Plasma and RPE*/*Choroid of Patients with AMD*

Age-related macular degeneration (AMD) is usually seen as two main types. "Dry" AMD where deposits called drusen develop in the macular region that ultimately progress to a late stage in which there is atrophy of the macula (geographic atrophy). "Wet" AMD describes AMD in which patients develop abnormal growth and leakage of the choroid vessels beneath and into the retina, termed choroidal neovascularization (CNV). HA contents were measured in plasma from patients with late stage AMD (geographic atrophy or choroidal neovascularization) and from age-matched controls without the disease. ELISA analysis (Figure 1A) indicates that HA contents were significantly increased in the plasma of patients with late-stage AMD (mean ± SEM: 111.8 ± 5.78 ng/mL) compared with plasma of controls without AMD (32.91 ± 5.75).

**Figure 1.** Hyaluronan HA is increased in circulation and in the RPE of age-related macular degeneration (AMD) patients. (**A**) HA was increased in plasma from patients with late-stage AMD (GA or CNV) compared to age-matched, normal controls. Data are presented as mean ± SEM (**B**–**J**) Representative human retina sections stained with HA binding protein (HABP) (**B**) Human retina section stained with streptavidin-AlexaFluor 647 in the absence of HA binding protein serves as a specificity control. (**C**–**J**) Human retina sections stained with streptavidin-AlexaFluor 647 in the presence of HA binding protein. HA is increased in the RPE in patients with dry AMD (**D**,**H**), wet AMD (**E**,**I**), and around drusen (**F**,**J**) compared to the RPE from an aged-match normal control (**C**,**G**). 40× images (**B**–**F**), 63× images (**G**–**J**). Green: HA; red: Bruch's membrane determined by its autofluorescence at 405 nm; blue: DAPI. Asterisks indicate drusen (**F**,**J**). GA—geographic atrophy; CNV—choroidal neovascularization; ONL—outer nuclear layer; RPE—retinal pigment epithelium; Ch—choroid; DAPI—4,6-diamidino-2-phenylindole.

To evaluate the distribution of HA in the retina under physiological and pathological conditions, sections from post-mortem human donor eyes from 3 controls and 6 AMD (4 dry and 2 wet AMD) patients were stained for HA using biotinylated HA binding protein (HABP). HA was found to be localized predominantly in the choroid of normal eyes (Figure 1C) as described previously [39,40]. Increased deposition of HA was seen around the RPE in AMD eyes (both in the dry (Figure 1D,H) and wet AMD specimens (Figure 1E,I)). HA was particularly enhanced in drusen and in areas of atrophy (Figure 1F,J) in AMD specimens. The sections stained with secondary antibody alone serves as a specificity control and shows minimal staining compared with sections stained with HABP (Figure 1B).

### *3.2. Increased Plasma HA and Accumulation of HA in the RPE of SFD Mice*

We utilized two mouse models to study the potential role of TIMP3 in the regulation of HA in the retina: mice lacking TIMP3 [34] and mice carrying the S179C-TIMP3 SFD mutation [33]. Plasma from S179C-TIMP3 and TIMP3-KO mice at 4–6 weeks of age was collected and HA contents were analyzed by ELISA. HA content of plasma was significantly increased in mice lacking TIMP3 as well

as in mice carrying the S179C-TIMP3 mutation (Figure 2A), suggesting that TIMP3 may be important in regulating HA.

**Figure 2.** HA is increased in circulation and in the RPE and choroid in mouse models of Sorsby's fundus dystrophy. (**A**) HA was increased in plasma from S179C-TIMP3 knockin mice (Timp3S179C6/S179C) and TIMP3-KO (Timp3-/-) mice compared to wild-type (WT) littermates. (n ≥ 5). Data are presented as mean ± SEM. (**B**) Mouse retina sections stained with biotinylated HA binding protein (HABP) in the absence (upper panel) or presence (lower panel) of hyaluronidase to detect HA. HABP staining is specific for HA as shown by the absence of staining in sections treated with hyaluronidase (lower panel). Green: HA; blue: DAPI. (**C**–**H**) Representative images of HA staining of mouse sections from wild-type (WT) mice (**C**,**F**), S179C-TIMP3 mutant mice (**D**,**G**) and TIMP3-KO mice (**E**,**H**). HA is increased in the RPE and choroid of S179C-TIMP3 (**D**) and TIMP3-KO (**E**) mice compared to wild-type (WT) littermate controls (**C**). HA (green) is predominantly localized to the basal surface of the RPE (\*\*) (**F**–**H**) and not to the apical surface (\*) as shown by co-staining with ezrin, a marker for the apical microvilli of RPE (red). 40× images (**C**–**E**); 63× images (**F**–**H**). n ≥ 3 for all immunohistochemistry data.

To determine if there was a similar correlation between the plasma HA levels and the accumulation of HA in the RPE as observed in human sections with AMD, we evaluated accrual of HA in the retinas of mice (8 weeks of age) lacking TIMP3 or carrying the SFD mutation. Cryosections of retina from mice of each specific genotype and wild-type littermates were stained for HA content with HABP. To ascertain that HABP binds HA specifically, sections were treated with hyaluronidase prior to staining with HABP. Indeed, pre-treatment with hyaluronidase resulted in absence of staining with HABP (Figure 2B, lower panel). S179C-TIMP3 mice (Figure 2D) and TIMP3-KO mice (Figure 2E) show increased accumulation of HA beneath the RPE and in the choroid compared to that seen in wildtype littermates (Figure 2C). Staining with antibodies to ezrin served as a marker for RPE apical microvilli (Figure 2C–H), and the higher magnification images (Figure 2F–H) confirmed the RPE localization of HA to the basal surface of the cells.

To identify the potential mechanism by which S179C-TIMP3 regulates HA we utilized stable human RPE lines (ARPE-19) expressing S179C-TIMP3 [20]. ARPE-19 cells (expressing S179C-TIMP3, wildtype TIMP3 (WT-TIMP3) and empty vector (Vector)) were cultured for 2–6 weeks in 1% serum on trans-well inserts and stained for HA. Increased accumulation of HA was observed in RPE cells expressing S179C-TIMP3 (Figure 3C,G) compared with cells transfected with empty vector (Figure 3A,G) or expressing wildtype TIMP3 (Figure 3B,G). The accumulation was predominantly intracellular and apical in the RPE (Figure 3D–F). There appear to be multiple layers of S179C-TIMP3 RPE cells on the transwell compared with a single monolayer for WT-TIMP3 and vector cells, which might sugges<sup>t</sup> epithelial-mesenchymal transition.

**Figure 3.** HA is increased in S179C-TIMP3 RPE cells in culture. (**A**–**C**) ARPE-19 cells expressing S179C-TIMP3 grown in culture for at least 2 weeks on trans-well inserts have increased HA (**C**) compared to WT-TIMP3 expressing cells (**B**) or vector only controls (**A**).(**D**–**F**) Z-plane images of HA in RPE monolayers grown on trans-well inserts show increased intracellular HA in S179C-TIMP3 cells. Green: HA; blue: DAPI. (**G**) Fluorescence intensity was quantitated by integrated density measurement (n ≥ 4, for each cell line). Data are presented as mean ± SEM.

### *3.3. FGF-2 Contributes to HA Accumulation in the RPE*

We have recently reported that RPE cells expressing S179C-TIMP3 secrete higher amounts of FGF-2 compared with control cells [28]. Previous studies have suggested that FGF signaling has the propensity to increase HA accumulation [32]. To experimentally test this hypothesis in RPE cells, we evaluated the ability of FGF-2 to induce HA accumulation in primary porcine RPE cells. Cells cultured for 3 weeks on trans-well inserts in 1% serum were treated with 0 ng/mL, 10 ng/mL, 25 ng/mL, or 100 ng/mL of FGF-2 in the presence of 1 μg/mL heparin, a cofactor for FGF receptor signaling. FGF-2 induced HA accumulation in a dose-dependent manner (Figure 4A–I) with maximum HA deposits being observed with a dose of 25 and 100 ng/mL (Figure 4D,H). This was confirmed by quantitation of fluorescence by integrated density measurements (Figure 4I). The FGF-2 induced accumulation of HA was seen predominantly on the apical surface of the RPE with increased basal and peri-cellular accumulation at higher doses (Figure 4E–H).

**Figure 4.** FGF-2 induces HA accumulation in primary RPE cells. (**A**–**D**) FGF-2 induced HA accumulation in primary porcine RPE cells in a dose-dependent manner (**A**) 0 ng/mL, (**B**) 10 ng/mL, (**C**) 25 ng/mL, (**D**) 100 ng/mL. Fluorescence intensity was quantitated by integrated density measurement (**I**) (n ≥ 4, for each cell line). Data are presented as mean ± SEM (**E**–**H**) Z-plane images show that increased concentrations of FGF-2 induce increased apical accumulation of HA in addition to some peri-cellular and basal deposits of HA at high doses of FGF-2. Green: HA, red: phalloidin; blue: DAPI.

To determine if the accumulation of HA seen in RPE cells expressing S179C-TIMP3 was a consequence of increased FGF signaling, cells (RPE cells expressing S179C-TIMP3, wildtype TIMP3 or empty vector) were treated with BGJ-398, an FGF receptor inhibitor. 10 μM BGJ-398 decreased HA content in RPE cells expressing S179C-TIMP3 (Figure 5C,F,H,I) and WT-TIMP3 (Figure 5B,E,I) but not vector only transfected cells (Figure 5A,D,I) when compared with their respective untreated cells that served as controls (Figure 5A: vector, 5B: WT-TIMP3 and 5C,G: S179C-TIMP3). Quantitation of fluorescence by integrated density analysis revealed that 10 μM BGJ-398 decreased HA accumulation in S179C-TIMP3 cells 63.1% (SD = 0.241) and only 43.4% (SD = 0.291) in wildtype cells (Figure 5I). Quantitation confirmed that BGJ-398 had no significant effect on vector only cells (Figure 5I), suggesting an FGF-specific mechanism for HA accumulation in S179C-TIMP3 cells.

**Figure 5.** Inhibition of FGF signaling decreases HA accumulation in S179C-TIMP3 RPE cells. Vector only (**A**), WT-TIMP3 (**B**), and S179C-TIMP3 (**C**) expressing ARPE-19 cells were grown on trans-well inserts for 4 weeks before treatment with FGF receptor inhibitor BGJ-398. Treatment with BGJ-398 decreased HA accumulation in S179C-TIMP3 expressing RPE cells (**F**,**H**,**I**) and to a lesser extent in WT-TIMP3 expressing cells (**E**,**I**) but has no effect on control vector RPE cells (**A**,**D**,**I**). (**I**) Control (Black Bar) indicates respective untreated cells compared with BGJ-398 treated cells (Grey bar). (**G**,**H**) Z-plane images of S179C-TIMP3 cells in the absence (**G**) and presence (**H**) of 10 μM BGJ-398 S179C-TIMP3 cells show an overall reduction in HA accumulation, including intracellular HA, after treatment with BGJ-398. Green: HA; blue: DAPI. (**I**) data are presented as mean ± SEM (n ≥ 6).

### *3.4. Increased HA is Associated with CNV in AMD and SFD*

Sections of post-mortem eyes from a patient with CNV showed significant HA deposition around the RPE (Figure 6C,D) when compared with control eyes (Figure 6A,B). While S179C-TIMP3 mice do not demonstrate a florid SFD phenotype as seen in humans, they do have increased susceptibility to experimental laser-induced CNV [28] as do the TIMP3-KO mice [41]. We have previously shown that FGF-2 from S179C-TIMP3 RPE cells can stimulate angiogenesis [28]. Since FGF-2 contributes to HA accumulation in the RPE and to angiogenesis, we evaluated if HA content and distribution was altered in laser-induced CNV lesions in S179C-TIMP3 mice. As described previously lesions in S179C-TIMP3 mice were larger and leakier compared to controls [28]. Increased HA accumulation was observed in CNV in S179C-TIMP3 mice (Figure 6H–J) when compared to lesions in control mice (Figure 6E–G) which appears to be a consequence of altered distribution to the CNV lesions in S179C-TIMP3 mice.

**Figure 6.** HA is increased in choroidal neovascular lesions in AMD patients and in S179C-TIMP3 mice. (**A**–**D**) HA is increased in the RPE in a patient with CNV (**C**,**D**) and in the CNV lesion compared to the RPE of a normal patient (**A**,**B**). Green: HA; blue: nuclei; red: Bruch's membrane. Arrows indicate RPE, double asterisks indicate CNV lesion. (**E**–**G**) HA is increased in laser-induced CNV lesions in S179C-TIMP3 mice compared to wild-type (WT) littermates 5 days post injury. (**E**,**H**) Arrows indicate CNV lesion. (**F**,**I**) Brightfield image overlaid on fluorescent image shows disruption of RPE and Bruch's membrane. (**G**,**J**) HA accumulation is diffuse and appears predominantly in the borders of the lesion in WT (**E**–**G**) compared to dense mass within the lesions of S179C-TIMP3 mice (**H**–**J**). Green: HA; blue: nuclei.
