Overexpression of HtrA1 and Exposure to Mainstream Cigarette Smoke Leads to Choroidal Neovascularization and Subretinal Deposits in Aged Mice

Abstract

Purpose.: We determined the function of ARMS2 and HtrA1 in the choroid and retina using transgenic (Tg) mice and evaluated the effects of mainstream cigarette smoke on these mice.

Methods.: The chicken actin promoter (CAG) was used to drive mouse HtrA1, human ARMS2, and ARMS2 (A69S) expression in the entire body of a mouse for one year. Fundus observations were performed with a Spectralis HRA+ optical coherence tomograph (OCT). Eyes were sectioned, stained with hematoxylin and eosin (H&E), and analyzed with immunohistochemistry. Mice were exposed to cigarette smoke for 30 min/d, 5 d/wk for 12 weeks using a mainstream smoking chamber (INH06-CIGR02A, MIPS). After 12 weeks, fundus observations and pathological analyses were performed.

Results.: Approximately 18.2% of 12-month-old HtrA1 Tg mice exhibited choroidal neovascularization (CNV) by OCT and positive immunostaining with anti-CD31 and anti-fibronectin antibodies. Furthermore, elastic van Gieson (EVG) staining showed Bruch's membrane damage in HtrA1 Tg mice. No retinal changes were observed in ARMS2 and ARMS2 (A69S) Tg mice. A total of 12 weeks of exposure to mainstream cigarette smoke led to CNV rates of 7.7% for wild type (Wt) mice and 20% for HtrA1 Tg mice, but had no effect on ARMS2 Tg mice. In addition, abnormal deposits were observed between photoreceptor cells and the RPE in an HtrA1 Tg mouse exposed to mainstream cigarette smoke.

Conclusions.: The HtrA1 overexpression and mainstream cigarette smoke can independently lead to CNV. The HtrA1 gene is a strong risk factor for wet AMD, but not all of the HtrA1 Tg mice developed CNV, suggesting that CNV development depends on multiple risk factors.

Introduction

Age-related macular degeneration (AMD) is the leading cause of severe irreversible central vision loss and blindness in individuals over 65 years of age in developed countries. ^1,2 This condition is separated into two types based on pathological characteristics. In dry AMD, drusen deposits are observed initially between the retinal pigment epithelium (RPE) and Bruch's membrane (BM) and there is a gradual progression to geographic atrophy, whereas in wet AMD, vision loss occurs by choroidal neovascularization (CNV) or polypoidal choroidal vasculopathy (PCV) of the macula. ³ Caucasian AMD patients predominantly exhibit geographic atrophy, while Japanese AMD patients present with CNV or PCV with few or no drusen. ^4,5 Genetic and environmental factors are believed to be involved in the onset of AMD. ⁶ A recent genome-wide association study (GWAS) revealed more than 30 AMD susceptibility gene loci. ^4,7–9 Single nucleotide polymorphisms (SNPs) in complement factor H (CFH), such as rs800292 (I62V) and rs1061170 (T402H) on chromosome 1q32, have been associated with AMD in Caucasians, ^4,7,10 but not with AMD in Chinese or Japanese patients. ^7,11–16 Another locus strongly associated with AMD is on chromosome 10q26, which is the chromosome where ARMS2 (age-related maculopathy susceptibility) and HtrA1 (high-temperature requirement factor A1) also are located. ¹⁰ Our previous GWAS study also showed strong associations between SNPs rs10490924, rs800292, and rs2241394 at the same locus, and wet AMD and PCV in Japanese patients. Within the linkage disequilibrium (LD) of these SNPs, a large insertion/deletion (indel) sequence variation was found downstream of ARMS2 at the promoter region of HTRA1. ^4,17  

It has been controversial whether this indel variant is responsible for the instability of ARMS2 mRNA or for the upregulation of HTRA1 transcription. ^10,18–22 We showed that the indel in the HTRA1 promoter upregulates transcription in a photoreceptor-derived cell line, but not in the RPE (data not shown; Iejima et al., submitted for publication). The presence and localization of ARMS2 also has been controversial. ^19,23,24 A previous study found that ARMS2 localized to the mitochondria, and that ARMS2 and ARMS2 (A69S) were found in extracellular fractions. ²³ Another study showed that ARMS2 was distributed primarily in the cytosol rather than in the outer membrane of mitochondria. ²⁴ In addition, a study suggested that ARMS2 is not synthesized in individuals homozygous for the indel variant; thus, ARMS2 may be the AMD risk-factor within this locus. ²³ Furthermore, SNPs in ARMS2 that change alanine to serine at codon 69 (A69S) may function as surrogate markers for the downstream indel. The ARMS2 gene consists of two exons, which encode a poorly characterized 107 amino acid protein with no similarities to known protein motifs. ²³ In this study, ARMS2 and ARMS2 (A69S) were expressed ubiquitously in mice to observe their effects on eye development and aging. 

Environmental risk factors also have been associated with AMD. Smoking is a major environmental risk factor that increases disease onset approximately 2- to 20-fold. ^7,25–31 Cigarette smoke, which contains chemical toxins, may affect metabolic processes in the RPE by repressing the release of antioxidants and by altering choroidal blood flow. ⁷ A recent report showed that smoking damaged mitochondrial DNA, but not nuclear DNA, and increased degradation processes in the RPE. ³² A number of studies found an independent association between smoking and the risk of advanced AMD with no susceptibility genes acting as modifiers, except ARMS2. ^7,33 In mouse models, exposure to cigarette smoke led to oxidative damage with structural degradation of the RPE and BM. ³⁴ Smoke-related oxidants also led to the formation of subretinal deposits. ³⁵  

Based on the observation that HtrA1 expression was higher in fibroblasts isolated from AMD patients than in fibroblasts from healthy subjects, ¹⁸ we created human ARMS2, human ARMS2 (A69S), and HtrA1 transgenic (Tg) mice. We used the chicken actin promoter (CAG) to ubiquitously express these genes. We characterized pathological changes during aging and compared them to those observed in human AMD patients. Pathological changes were observed for up to 12 months after birth by fundus observation, fluorescein angiography (FA), indocyanine green angiography (IA), and optical coherence tomography (OCT). Because upregulation of susceptible genes and mainstream cigarette smoke are two essential risk factors for AMD, Tg mice also were exposed to mainstream cigarette smoke to evaluate genetic and environmental risk factors for disease progression. 

Methods

Construction of ARMS2, ARMS2 (A69S), and HtrA1 Tg Mice

The ARMS2 cDNA was PCR amplified (PrimeSTAR HS DNA Polymerase; TaKaRa Bio, Otsu, Japan) from a human retina cDNA library (Clontech Laboratories, Inc., Mountain View, CA, USA) and ligated into pCAGGS, an expression vector containing a CAG promoter and an HA-tag at the C-terminus. The HtrA1 cDNA was PCR amplified from a mouse brain cDNA library and ligated into pCAGGS, an expression vector containing a CAG promoter and an HA-tag at the N-terminus. Linearized inserts were injected into pronuclear stage BDF1/C57BL6N embryos to create the Tg mice (PhoenixBio Co., Hiroshima, Japan). 

The ARMS2 and ARMS2 (A69S) Tg mice were screened by PCR (PrimeSTAR HS DNA Polymerase; TaKaRa Bio) using the forward primer 5′-ATGCTGCGCCTATACCCAGG-3′ and the reverse primer 5′-AGTGTCAGGTGGTGCTGAGG-3′ followed by Sanger sequencing (BigDye Terminator v3.1 Cycle Sequencing Kit; Invitrogen, Carlsbad, CA, USA). The HtrA1 Tg mice were screened by PCR using the forward primer 5′-GAGCCTCTGCTAACCATGTTGA-3′ and the reverse primer 5′-CAGCAGTAGCAAAGACAGGAGC-3′ followed by Sanger sequencing. 

Nine Tg lines with 2 to 11 copies of mouse HtrA1 cDNA were generated. The Tg line with 11 copies was selected and bred with C57BL6N for seven generations before using the mice for experiments. All experiments were approved by the National Hospital Organization Experimental Animal Committee and the Hoshi University Experimental Animal Committee. All mice experiments were performed in accordance with the Association for Research in Vision and Ophthalmology (ARVO) Statement for the Use of Animals in Ophthalmic and Vision Research. 

Western Blotting

Proteins from mouse brain and eye tissue were extracted in ice-cold TNE buffer containing protease and phosphatase inhibitors (1697498; Hoffmann-La Roche, Basel, Switzerland). Protein concentrations were determined using the BCA assay kit (Pierce; Thermo Fisher Scientific, Waltham, MA, USA). Equal amounts of protein (25 μg/lane) were separated by 7.5% SDS-PAGE and transferred to polyvinylidine fluoride (PVDF) membranes (Trans-Blot Turbo; Bio-Rad Laboratories, Inc., Hercules, CA, USA) for Western blotting with anti-mouse HtrA1 (1:100 dilution; ab38611; Abcam, Cambridge, UK), anti-HA (1:1000 dilution; 631207; Clontech Laboratories, Inc.), anti-human ARMS2 (1:1000 dilution; ABN160; Millipore Corporation, Billerica, MA, USA), and anti-actin (1:1000 dilution; MAB1501; Millipore Corporation) antibodies. The FluorChem Western Blot Imaging Station (ChemiDoc XRS+; Bio-Rad Laboratories, Inc.) was used to capture the images. The pixel value for each protein band was determined and normalized using image analysis software (Image Lab; Bio-Rad Laboratories, Inc.). 

Quantitative RT-PCR (qRT-PCR)

Total RNA was isolated from the eyes and brains of Tg and wild type (Wt) mice using an RNeasy Mini kit (Qiagen, Venlo, The Netherlands) according to the manufacturer's instructions, and aliquots were reverse-transcribed to generate single-stranded cDNA (High Capacity cDNA Reverse Transcription Kit; Life Technologies, Carlsbad, CA, USA). The qRT-PCR was performed using an ABI STEP-One Real-time PCR system (Life Technologies) and a TaqMan probe (mouse HtrA1) according to the manufacturer's instructions. All reactions were run in triplicate using human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as the internal control and ΔCt normalization. 

RNA Preparation and PCR Amplification in ARMS2 (A69S) Tg Mice

Total RNA was isolated from mice brains and eyeballs using TRIzol Reagent (Life Technologies). Single-stranded oligo (dT)-primed cDNA was generated using Superscript II Reverse Transcriptase (Life Technologies). The ARMS2 was PCR amplified (PrimeSTAR HS DNA Polymerase, TaKaRa Bio) using the forward primer 5′-GCCAACTGGAGCTTCTCATC-3′ and the reverse primer 5′-GATTTCACCAGCTGCACAGA-3′. Mouse GAPDH (450 base pairs [bp]) was used as the control (forward primer, 5′-GGGGAGCCAAAAGGGTCATCATCT-3′; reverse primer, 5′-CGACGCCTGCTTCACCACCTTCTT-3′). 

Angiography Observation and Retinal OCT of CNV

For in vivo imaging, 12-month-old ARMS2, ARMS2 (A69S), and HtrA1 Tg mice were anesthetized by injecting ketamine (Daiichi Sankyo, Chuo, Japan) at 0.002 mL/g body weight into their abdominal cavities. Pupils were dilated with 5 mg/mL Tropicamide (Santen Pharmaceutical Co., Ltd., Osaka, Japan). Fundus examinations were performed using a small animal retinal-imaging microscope (Micron III; Phoenix Research Products, Candler, NC, USA). The FA, IA, and OCT studies were performed using a retinal-imaging device (Spectralis HRA+OCT, Heidelberg Engineering, Heidelberg, Germany). Retinal vasculature imaging by FA and choroidal vasculature imaging by IA were performed simultaneously. Morphological changes, and angiographs of mouse retinas and choroids were visualized by OCT using a 3.2-mm diameter polymethylmethacrylate (PMMA) mouse contact lens (Ocular Instruments, Inc., Bellevue, WA, USA) to adjust the light path. Mice were tail-injected with 100 μL fluorescein sodium (100 mg/mL, 1:10 dilution; Alcon, Fort Worth, TX, USA) immediately before imaging and with 100 μL indocyanine green (25 mg/3 mL, 1:10 dilution; Santen Pharmaceutical Co.) 5 minutes before imaging. 

Histology and Immunohistochemistry

Mice were anesthetized by injecting 1 mL pentobarbital (64.8 mg/mL, Kyoritsu Seiyaku, Tokyo, Japan) into their abdominal cavities and their hearts were cut open to sacrifice the mice. The eyes were removed quickly and immersed in a fixative solution containing 5% formaldehyde overnight at 4°C. The eyes were embedded in paraffin and sectioned at 5-μm thickness. After deparaffinization and rehydration, sections were hematoxylin and eosin (H&E) stained and elastica van Gieson (EVG) stained (Elastic Statin kit HT25A, Sigma-Aldrich Corp., St. Louis, MO, USA). The H&E- and EVG-stained images were collected using a Nikon Eclipse light microscope (Nikon Corporation, Tokyo, Japan). After deparaffinization and rehydration, eye sections were treated with Target Retrieval Solution (DakoCytomation, Glostrup, Denmark) at 120°C for 10 minutes. They then were incubated with blocking solution for 1 hour followed by overnight incubation with primary anti-CD31 antibody (1:50 dilution; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) in PBS containing 2% BSA at 4°C. Sections were washed with PBS three times and then were incubated with Alexa Fluor 488 conjugated rabbit anti-mouse IgG (1:500 dilution; Life Technologies) and 4′6-diamidino-2-phenylindole (DAPI) for nuclear staining (1:1000 dilution; Wako Chemicals GmbH, Neuss, Germany) for 1 hour at room temperature. Slides were mounted with Ultramount Aqueous Permanent Mounting Medium (DakoCytomation) and viewed with a confocal fluorescence laser microscope (LSM 700, Carl Zeiss Meditech, Jena, Germany). 

Transmission Electron Microscopy (TEM)

Mice eyes were dissected and immersed in a fixative containing 2% formaldehyde overnight at 4°C. Tissue samples for TEM were fixed in phosphate-buffered 2% glutaraldehyde (EM grade; Electron Microscopy Science, Hatfield, PA, USA) and postfixed in 2% osmium tetroxide (Crystal; Heraeus Chemicals, Port Elizabeth, South Africa) for 3 hours in an ice bath. Specimens then were dehydrated in graded ethanol (Nakarai Tesque, Inc., Nakagyo-ku, Japan) and embedded in epoxy resin (TAAB Laboratories, Berkshire, UK). Ultrathin sections were cut at 70 to 80 nm by an ultramicrotome with diamond knives (Diatome Ltd., Biel, Switzerland) and mounted on 200 mesh copper grids. Ultrathin sections were stained with uranyl acetate for 10 minutes and with a lead staining solution for 5 minutes; stained sections were submitted for TEM (JEM-2000EX and JEM-1200; JEOL Ltd., Akashima, Japan) to the Hanaichi Ultrastructure Research Institute (Aichi, Japan). 

Exposure to Mainstream Cigarette Smoke

The 12-month-old ARMS2, ARMS2 (A69S), HtrA1 Tg, and Wt mice with normal fundi were exposed to mainstream cigarette smoke for 30 min/d, 5 d/wk for 12 weeks using a mainstream smoking chamber for small animals (INH06-CIGR02A, MIPS, see Fig. 4A). Natural American Spirit cigarettes, which contain 12 mg of tar and 1.5 mg of nicotine per cigarette, were used in this experiment. Mainstream cigarette smoke was diluted 1:7 with compressed air. Mice were placed in the chamber and exposed to smoke at a flow rate of 0.35 L/min. Control mice were placed in the chamber without cigarette smoke to evaluate the effect of stress on the retina. After 12 weeks of exposure to mainstream cigarette smoke, fundus observations, FA, IA, and OCT were performed. Mice then were sacrificed for eye sectioning followed by H&E staining and immunostaining. All experiments were approved by the National Hospital Organization Experimental Animal Committee and the Hoshi University Experimental Animal Committee. All mice experiments were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Statistics

Data of qRT-PCR were analyzed using the StatView 5.0 statistical software package (SAS Institute, Inc., Chicago, IL, USA). Continuous data between two groups were compared using Student's t-test. All pathological data were analyzed using Fisher's exact test. P values < 0.05 were considered statistically significant. 

Results

Ubiquitous Expression of Mouse HtrA1 and Human ARMS2 in Mice

The CAG was used to create Tg mice that overexpress mouse HtrA1, human ARMS2, or ARMS2 (A69S). An HA-tag at the C-terminus of recombinant HtrA1 was used to detect expression of recombinant Htra1 in the brains and eyes of mice by western blotting. 

Levels of HtrA1 in the brains and eyes of mice were studied by Western blotting. The HtrA1 levels were 6.6- and 3.0-fold higher in HtrA1 Tg mice brains and eyes, respectively, than in Wt mice brains and eyes (Fig. 1A). In addition, HtrA1 mRNA levels were higher in HtrA1 Tg mice than in Wt mice. The brains and eyes of HtrA1 Tg mice had 93.2- and 98.2-fold higher HtrA1 mRNA levels, respectively, compared to Wt (Fig. 1B). 

Figure 1

 

Ubiquitous expression of mouse HtrA1, human ARMS2, or ARMS2 (A69S) using the CAG promoter in mice. (A) Western blot analysis of mouse HtrA1 expression in the brains and eyes of mice. HtrA1 expression was detected in brains and eyes using an anti-HtrA1 and anti-HA antibody. The relative intensity of a band for the transgenic eye tissue toward a housekeeper was taken as 1.0. The HtrA1 protein level was significantly higher in the HtrA1 Tg mouse eye than in the Wt mouse eye. (B) Quantitative real-time PCR determination of HtrA1 mRNA levels from total RNA isolated from brains and eyes of Wt and HtrA1 Tg mice. The Wt mouse brain was assigned a value of 1.0. The HtrA1 mRNA level in an HtrA1 Tg mouse was significantly higher than in a Wt mouse. (C) Western blot analysis of human ARMS2 expression in the mouse brain and eye. The ARMS2 protein was detected in transgenic mouse using an anti-ARMS2 antibody. (D) Analysis of ARMS2 mRNA expression in ARMS2 and ARMS2 (A69S) Tg mice by PCR amplification. The ARMS2 mRNA was detected in the brain and eye.

Human ARMS2 was detected in the brains and eyes of ARMS2 Tg mice (Fig. 1C). The ARMS2 and ARMS2 (A69S) mRNAs were detected in the brains and eyes of ARMS2 and ARMS2 (A69S) Tg mice (Fig. 1D). Fundus observations and H&E stains of eye sections showed that ARMS2 and ARMS2 (A69S) Tg mice had no abnormalities in their retinas and choroids up to 12 months after birth (Fig. 2). 

Figure 2

 

No retinal abnormalities in 12-month-old ARMS2 Tg and ARMS2 (A69S) Tg mice by fundus observation. The ARMS2 Tg and ARMS2 (A69S) Tg mice showed no retinal abnormalities at 1 year by fundus observation (scale bars: 200 μm) and H&E staining (×40 magnification; scale bars: 20 μm).

HtrA1 Tg Mice Develop CNV but ARMS2 and ARMS2 (A69S) Tg Mice Remain Normal

Fluorescein diapedesis was observed in HtrA1 Tg mice fundi, suggesting the breakdown of the blood–retinal barrier (BRB) and aggravation of retinal vascular permeability. Lesions with low fluorescence were observed in HtrA1 Tg mice choroids by IA and indicated the presence of networks of abnormally branched vessels (Fig. 3A). Additionally, radial capillary branching from the choroid through the RPE and into the retina was observed by OCT (Fig. 3B). Radial CNV formation was explored further by H&E staining and immunostaining with anti-CD31, a marker for endothelial cells (Fig. 3C). Furthermore, HtrA1 Tg mice showed ruptures and deficiencies in BMs by EVG staining (Fig. 3D). Of the 22 12-month-old HtrA1 Tg mice examined, four developed CNV (18.2%), compared to none of the 40 12-month-old Wt mice. In contrast, none of the 20 12-month old ARMS2 and ARMS2 (A69S) Tg mice had any abnormal retinal changes (Table 1). 

Figure 3

 

The 12-month-old HtrA1 Tg mice developed CNV. (A) The HtrA1 Tg mice showed hyperfluorescent lesions by FA (red asterisks) and networks of abnormally branching vessels by IA (yellow asterisks). (B) The CNV was observed in the choroid and retina by OCT (red arrowheads; scale bar: 200 μm). (C) The H&E staining of the FA and IA lesions on the retinal sections showed radial CNV spreading from the choroid through the RPE into the retina (black asterisks; ×40 magnification; scale bars: 20 μm). The endothelial cell marker CD31 was stained (green) in a radial CNV (white arrowheads; scale bars: 20 μm). Nuclei were stained with DAPI (blue). (D) The EVG staining of BM (red arrowheads). The HtrA1 Tg mice displayed ruptures and deficiencies in the BM (red asterisks; ×100 magnification; scale bars: 10 μm).

Table 1

 

CNV in Wt, HtrA1, ARMS2, and ARMS2 (A69S) Tg Mice

	Wt	HtrA1 Tg	ARMS2 Tg	ARMS2 (A69S) Tg
N of mice	40	22	20	20
N of mice with CNV	0	4	0	0
CNV rate, %	0	18.2	0	0

 Approximately 18.2% of HtrA1 Tg mice displayed CNV. P < 0.005, Fisher's exact test.

Mainstream Cigarette Smoke Leads to CNV in HtrA1 Tg and Wt Mice

The HtrA1 Tg, ARMS2 Tg, and Wt mice were exposed to mainstream cigarette smoke as described in the Materials and Methods section. All mice used for this experiment were 12 months old. Two of 10 HtrA1 Tg mice that were exposed to mainstream cigarette smoke developed CNV (Table 2), compared to none of the control group that was exposed to air (Table 4). The HtrA1 Tg mice exposed to mainstream cigarette smoke displayed hyperfluorescent lesions by FA and lesions with low fluorescence by IA. Networks of abnormally branching vessels were observed in lesions with low fluorescence (Fig. 4B). The OCT images of the FA and IA lesions showed abnormal morphological changes in the choroid and retinal layer (Fig. 4C). Radial CNV spreading from the choroid to the retina was observed by H&E staining and confirmed by immunostaining with anti-CD31 (Figs. 4D, 4E). The HtrA1 Tg mice showed ruptures and deficiencies in BMs, but equivalent damage also was observed in Wt mice exposed to mainstream cigarette smoke (Fig. 4F). Our results showed that exposure to mainstream cigarette smoke triggered CNV, and the formation of CNV was independent of the level of HtrA1 expressed. 

Table 2

 

Effects of Mainstream Cigarette Smoke on CNV and Subretinal Deposits in Wt and HtrA1 Tg Mice

	Wt	HtrA1 Tg	ARMS2 Tg
N of mice exposed to smoke, 30 min/d, 5 d/wk, 12 wks	13	10	9
N of mice with CNV	1	2	0
CNV rate, %	7.7	20	0

 There was no significant difference in the CNV rates of HtrA1 Tg and Wt mice. P > 0.5, Fisher's exact test.

Figure 4

 

Effect of cigarette smoke on CNV in 12-month-old Wt and HtrA1 Tg mice. (A) The 12-month-old Wt and HtrA1 Tg mice were exposed to mainstream cigarette smoke (diluted 1:7 with compressed air) from Natural America Spirit cigarettes (left). Mice were placed into smoke chambers (right) for 30 min/d, 5 d/wk for 12 weeks. The smoke flow rate was 0.35 L/min. (B) After 12 weeks of exposure to cigarette smoke, hyperfluorescent lesions were observed by FA and networks of abnormally branching vessels were observed by IA in Wt and HtrA1 Tg mice (red boxes). The networks of abnormally branching vessels spread further and were more clearly visible in HtrA1 Tg mice than in Wt mice. (C) The OCT of FA and IA lesions in Wt and HtrA1 Tg mice, respectively. Photoreceptor cells were thinner. HtrA1 Tg mice displayed more severe retinal abnormalities than Wt mice (red arrowheads; scale bars: 200 μm). (D) Retinal sections stained with H&E displayed radial CNV formations (black arrowheads; ×40 magnification; scale bars: 20 μm). (E) The endothelial cell marker CD31 was stained (green) within a radial CNV formation (white arrowheads; scale bars: 20 μm). Nuclei were stained with DAPI (blue). (F) The EVG staining showed ruptures or deficiencies in the BMs of Wt mice, but HtrA1 Tg mice showed more BM damage after exposure to mainstream cigarette smoke (red asterisks, ×100 magnification; scale bars: 10 μm).

Accumulation of Subretinal Deposits in HtrA1 Tg Mice Exposed to Mainstream Cigarette Smoke

A single HtrA1 Tg mouse exposed to mainstream cigarette smoke for 12 weeks developed dense deposits in the subretinal space (Fig. 5A). Electron micrographs revealed RPE degeneration with severe pigment loss and abnormal deposits or vacuolization in the subretinal space of one HtrA1 Tg mouse exposed to mainstream cigarette smoke (Fig. 5B). The molecular content of these deposits currently is under investigation. Although only one of 10 HtrA1 Tg mice had subretinal deposits, no ARMS2 Tg mice or Wt mice, with or without exposure to mainstream cigarette smoke, had any subretinal deposits (Table 3). Further, none of the HtrA1 Tg mice exposed to air for 12 weeks had retinal degeneration (Table 4). Our results suggested that overexpression of HtrA1 combined with mainstream cigarette smoke can cause formation of abnormal subretinal deposits. 

Figure 5

 

Effect of cigarette smoke on subretinal deposits in 12-month-old HtrA1 Tg mice. (A) Retinal electron micrograph of an HtrA1 Tg mouse exposed to smoke shows RPE degeneration with severe pigment loss and vacuolization (red arrowheads). Abnormal deposits were observed between photoreceptor cells and RPE cells in an HtrA1 Tg mouse exposed to smoke (red boxes; ×20 magnification; scale bars: 20 μm). (B) Vacuolization (black asterisks) and the fragmented outer segment (red arrowheads) were observed in RPE (lower left). Photoreceptor cells were aligned in a disorderly manner and vacuolization occurred between and within the inner segment (lower middle, black asterisks). Basal infolding with thick subretinal deposits (black arrowheads) and a BM lacking elastic lamina (red asterisks) were observed (lower right photo). BI, basal infolding; EL, elastic lamina. Scale bars: 2 μm.

Table 3

 

Effect of Cigarette Smoke on Subretinal Deposits in HtrA1 Tg Mice

	Wt	HtrA1 Tg	ARMS2 Tg
N of mice exposed to smoke, 30 min/d, 5 d/wk, 12 wks	13	10	9
N of mice with subretinal deposits	0	1	0
Subretinal deposits rate, %	0	10	0

 Of HtrA1 Tg mice, 10% had subretinal deposits (P > 0.05, Fisher's exact test).

Table 4

 

No Retinal Changes Were Found in Wt, HtrA1 Tg, and ARMS2 Tg Mice Exposed to Air

	Wt	HtrA1 Tg	ARMS2 Tg
N of mice exposed to air, 30 min/d, 5 d/wk, 12 wks	10	10	10
N of mice with CNV	0	0	0
N of mice with subretinal deposits	0	0	0

Discussion

The HTRA1 gene is a member of a family of serine proteases and is involved in the degradation of extracellular matrix (ECM) proteins, like fibronectin and aggrecan. ^36,37 Mutations in this gene have been associated with hereditary cerebral small-vessel disease (CSVD). ^6,38 Because HtrA1 was found to be expressed similarly in mouse and human retinas, ^39,40 we generated a mouse line that overexpressed mouse HtrA1 throughout the entire mouse body and observed CNV progression in aged mice. Wet AMD is characterized by CNV or PCV. The CNV results from the growth of new blood vessels from the choroid into the RPE and subretinal spaces, whereas PCV results from inner choroidal vessel abnormalities with intact basement membranes and collagenous fibers in the BM. ^41,42 The BM depends on the adjacent RPE and choroidal cells, and is important for AMD development, particularly for CNV formation in wet AMD. Degradation of BM and upregulation of VEGF are risk factors for CNV. ^43,44 Targeted expression of human HtrA1 in the mouse RPE has led to the development of PCV, but not CNV. ⁴¹ Our results demonstrated, for the first time to our knowledge, that overexpression of mouse HtrA1 in the entire mouse body can induce CNV. No drusen-like deposits at the basal RPE in HtrA1 Tg mice were observed. Of the 22 12-month-old HtrA1 Tg mice that we examined, four displayed CNV. 

In AMD patients with CNV, the promoter region of HTRA1 is highly associated with CNV. Risk variant 18 leads to a 2.7-fold mRNA increase in HTRA1 in the RPE and is estimated to confer a population attributable risk of 49.3%. In our mouse model, an approximately 98.2-fold increase in HtrA1 mRNA in the mouse eye (Fig. 1B) led to an attributed CNV risk of 18.2%. Our results suggested that the level of HtrA1 expression may influence CNV progression and that other unknown environmental risk factors may influence the risk variant. None of 40 12-month-old Wt mice examined displayed CNV, suggesting that the level of HtrA1 expression is more influential than aging. Previous studies have suggested that overexpression of HtrA1 in the RPE can lead to an altered BM with fragmentation of the elastic lamina. ^6,41 Our study also found ruptured or deficient BMs when HtrA1 was overexpressed ubiquitously. 

Cigarette smoke, which contains numerous potential oxidants, including nitric oxide, carbon monoxide, and many other toxic chemical moieties, is considered to be the environmental risk factor most strongly associated with early-stage AMD. ⁴⁵ In human studies, cigarette smoking induces RPE abnormalities, such as geographic atrophy of the RPE and cell death from apoptosis, which also are changes associated with aging and early-stage AMD. ^34,46,47 Mice exposed to chronic cigarette smoke develop evidence of oxidative damage, such as ultrastructural degeneration of the RPE and BM, as well as RPE apoptosis. ³⁴ In our study, exposure to mainstream cigarette smoke enhanced the rates of CNV in HtrA1 Tg mice and Wt mice. Two (20%) of 10 HtrA1 Tg mice examined had radial CNV, compared to one (7.7%) of 13 Wt mice. The number of Tg mice, with and without CNV, at 12 months was extremely limited. Although the number of mice tested was low and although our results were statistically insignificant, we speculated that smoking does not significantly alter the occurrence of CNV. 

In this study, HtrA1 Tg mice exposed to cigarette smoke displayed RPE degeneration, severe pigment loss, vacuolization, and numerous wave-like morphological disorders in the outer nuclear layer of the retina, as well as abnormal deposits at the apical side of the RPE. Although the mechanisms of this retinal degeneration still are unknown, previous reports have shown that treatment with endogenous double-stranded RNAs (dsRNA) led to RPE loss and that activation of a receptor involved in innate immunity, Toll-Like receptor 3 (TLR3), affected retinal morphology. ^48–50 A recent report showed that TLR3 activation led to a wave-like morphology in the outer nuclear retinal layer. This wave-like layer also was found to contain late apoptotic cells. ⁵¹ Of the 10 HtrA1 Tg mice exposed to smoke that we examined, one displayed signs of RPE degeneration, including subretinal deposits similar to the advanced form of dry AMD. No RPE damage was found in Wt mice or in HtrA1 Tg mice that were not exposed to smoke. Our results showed that overexpression of HtrA1 combined with exposure to mainstream cigarette smoke leads to RPE degeneration and morphological changes in the retina. 

A recent report showed that ARMS2 is a constituent of the ECM; this localization suggests that ARMS2 may be necessary for proper matrix function and may protect against drusen formation. ²³ We examined 20 aged ARMS2 Tg mice and 20 aged ARMS2 (A69S) Tg mice, but none displayed any retinal changes. Furthermore, nine of the 20 aged ARMS2 Tg mice that were exposed to mainstream cigarette smoke showed no abnormal retinal changes. Our results demonstrated that ubiquitous overexpression of ARMS2 or ARMS2 (A69S) does not lead to typical AMD phenotypes, such as drusen or CNV formation. We were unable to confirm whether ARMS2 protects against CNV formation induced by mainstream cigarette smoke. 

In summary, aged mice that ubiquitously overexpress HtrA1 develop CNV similar to human AMD patients. Exposure to mainstream cigarette smoke enhanced the CNV rate in HtrA1 Tg and Wt mice. Furthermore, HtrA1 Tg mice exposed to mainstream cigarette smoke developed subretinal deposits and a wave-like retinal morphology. Extended networks of branching vessels covering much of the retina were observed in HtrA1 Tg mice, but not in ARMS2 Tg or Wt mice exposed to mainstream cigarette smoke. Morphological changes also were observed in the photoreceptor layer of the CNV. Our study suggested that HtrA1 overexpression alone is a strong risk factor for wet AMD, which is the predominant form of AMD in the Japanese population. 

Acknowledgments

The authors thank Koichi Yanagisawa of JFC Sales Plan Co., Ltd. for use of the Spectralis HRA+OCT (Heidelberg Engineering) and for his technical assistance. 

Supported by grants to by the Japanese Ministry of Health, Labor and Welfare (10103254 [TI]) and the Japan Society for the Promotion of Science (24592664 [MN], 23890258 [DI], and 22791704 [MA]). 

Disclosure: M. Nakayama, None; D. Iejima, None; M. Akahori, None; J. Kamei, None; A. Goto, None; T. Iwata, None 

Footnotes

 MN and DI are joint first authors.

Footnotes

 MN and DI contributed equally to the work presented here and should therefore be regarded as equivalent authors.