HBV infection increases the risk of macular degeneration: the roles of HBx-mediated sensitization of retinal pigment epithelial cells to UV and blue light irradiation

The National Health Research Institutes (NHRI) built a large database, the National Health Insurance Research Database (NHIRD), which includes the claims data from the Taiwan National Health Insurance (NHI) program. The Taiwan NHI is a single-payer, compulsory health insurance program for Taiwan citizens. The data for the present study were derived from the Longitudinal Health Insurance Database (LHID), which contains the claims data of 1 million insured people within the NHIRD, including beneficiary registration, inpatient and outpatient files, drug use, and other medical services. One million insured people were randomly selected between 1 January 1996, and 31 December 2000, and followed up in the LHID. Notably, the disease record system in the Taiwan NHI was established according to the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). To protect the privacy of those insured, the NHRI replaced the original identification numbers with anonymous numbers before releasing the database publicly.

To investigate the association between MD risk and HBV infection, we designed a retrospective population-based cohort study and established an HBV-infected cohort and a comparison cohort. The HBV-infected cohort comprised patients with newly onset HBV (ICD-9-CM 070.20, 070.22, 070.30, 070.32, and V02.61) who were aged ≥ 20 years; the patients were recruited from 1 January 2000, to 31 December 2011, and the index date was ordered as the day of the HBV diagnosis. The comparison cohort was formed by selecting individuals in the LHID without a history of HBV infection and frequency matching them against the patients at a ratio of 1:4 (HBV-infected cohort vs. comparison cohort); the matching criteria included age. The index dates of the comparisons were randomly assigned a month and a day, as well as the same index year as the matched cases. We excluded patients with a history of HCV infection (ICD-9-CM 070.41, 070.44, 070.51, 070.54, and V02.62) or MD (ICD-9-CM 362.5) before the index date. The observation outcome of interest was the occurrence of an MD event. We observed these two cohorts at the index date and stopped the follow-up process when the patients were removed from the Taiwan NHI, developed MD, or on December 31, 2011, whichever occurred first. Age, sex, and comorbidity are the common confounding factors in NHIRD research; we therefore collected the patients’ comorbidity histories prior to the index date. Specifically, the comorbidities examined in this study comprised hypertension (ICD-9-CM 401–405), hyperlipidemia (ICD-9-CM 272), alcohol-related illness (ICD-9-CM 291, 303, 305, 571.0, 571.1, 571.2, 571.3, 790.3, A215, and V11.3), diabetes (ICD-9-CM 250), asthma (ICD-9-CM 493), cirrhosis (ICD-9-CM 571.2, 571.5, and 571.6), anxiety (ICD-9-CM 300.00), and coronary artery disease (ICD-9-CM 410–414).

ARPE19 cells, which form a human retinal pigment epithelial cell line (ATCC number: CRL-2302), were used in this study. The cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM)/Nutrient Mixture F-12 medium (Hyclone), supplemented with 10% fetal bovine serum (Hyclone), 100 U/mL penicillin, and 100 µg/mL streptomycin, in a humidified incubator at 37 °C and with 5% CO₂. The cells were subcultured every 2–3 days to maintain exponential growth.

The UV crosslinker, CL-1000L model (Ultra-Violet Products, LCC, CA, USA), was used as the source of UV light. The UV light source consists of five tubes of 8 W UV dual bipin discharge type (115 V/60 Hz/0.7 A), and the wavelength of the UV tube is 365 nm (UV-A). Two operational settings are available in the UV crosslinker: (1) preset UV energy exposure and (2) preset UV time exposure. The energy needed in the experiments can be set on the touch pad. It takes 4 min for the UV light exposure to 1 J/cm². The exposure time of the UV light depends on the energy used in our experiments. The blue light LED lamp, MIC-209 model (60 W, blue light), was used as the source of visible blue light. The wavelength and the chromaticity diagram were measured by using the spectrometer USB2000+ (Ocean Optics, FL, USA) and the luminance colorimeter BM-7A (Topcon Tech. Co., Tokyo, Japan), respectively. The irradiance of the blue LED lamp was directly measured by a solar power meter SPM1116SD (Lutron Electronic Enterprise Co., Ltd., Taiwan) at the distance of 6.5 cm below the blue light source, where the cultured cells were exposed in the experiments.

Cell viability was analyzed and measured using the 3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyltetrazolium bromide (MTT) method. Briefly, the cells were seeded in a 35-mm dish at a density of 2 × 10⁵ cells/cm². After overnight culturing, the cultured medium was discarded and the cells were exposed to the indicated dose of UV radiation, followed by a replacement of fresh medium at 37 °C. After incubation for a different period, MTT (Sigma, St. Louis, MO, USA) was added to a final concentration of 0.5 mg/mL and incubated in a CO₂ incubator for an additional 4 h. Subsequently, the medium was aspirated, and 500 μL of dimethyl sulfoxide (Sigma, St. Louis, MO, USA) was added to the dish to dissolve formazan crystals. The absorbance was then obtained using a Synergy 2 microplate reader (BioTek Instruments, VT, USA) at a test wavelength of 490 nm with a reference wavelength of 630 nm. Finally, cell viability was determined by the relative absorbance of the experimental treatment compared with that of the control treatment.

Two thousand cells were seeded into a 60-mm dish. After 24 h, the cells were exposed to the indicated dose of UV irradiation and cultured with fresh medium for 2 weeks. Subsequently, the cells were fixed with a 4% paraformaldehyde solution and stained with 0.1% crystal violet for 30 min. After washing, the crystal violet was dissolved with 10% acetic acid and the absorbance was measured at 590 nm. The relative colony number was calculated according to the relative absorbance of the experimental treatment in comparison with that of the control treatment.

Following treatment, the total RNAs of each group of cells were extracted using the Trizol reagent (Invitrogen, Carlsbad, CA, USA). The RNA yields and purity were checked by OD260/OD280 (> 1.8) and OD260/OD230 (> 1.6) using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Additionally, we used the human oligonucleotide DNA microarray (Human Whole Genome OneArray^®v6, Phalanx Biotech Group, Taiwan), which contains 32,679 DNA oligonucleotide probes. Of these probes, 31,741 correspond to the annotated genes in the RefSeq v51 and Ensembl v65 databases. To control the experiment quality, the remaining 938 control probes were also included. Detailed descriptions of the gene array list are available from http://​www.​phalanx.​com.​tw/​Products/​HOA_​Probe.​php.

For the in vitro studies, the experiments were performed, at a minimum, in triplicate. In each experiment, the mean value of the repetitions was calculated and then used in the statistical analysis. All of the data were normalized to control the values of each assay and are presented as the mean ± SD. Additionally, the data were analyzed using one-way ANOVA, and significance was again set at P < 0.05. The array data were analyzed using the Rosetta Resolver System (Rosetta Biosoftware), which was applied to correct the data by removing both systematic and random errors. Signals that passed the criteria were then normalized through a 50% median scaling normalization method. The technical repeat data were analyzed by calculating the Pearson correlation coefficient to review the reproducibility (R value ≥ 0.975), and then the normalized data were transformed to gene expression log2 ratios between the mock and HBx groups. Signals with a log2 ratio of ≥ 1 or log2 ratio of ≤ −1 and a P value of < 0.05 were selected and defined as differentially expressed (DE) genes for further analysis. Scatter plots were made to visually assess the variation between chips. In addition, volcano plots (Fig. 4a) and hierarchical clustering (Fig. 4c) were performed to visually demonstrate distinguishable gene expression profiles among samples.

From the LHID, 39,796 HBV-infected patients, along with 4-fold comparison patients, were enrolled in this study (Table 1). Their mean age was nearly 44 years and 58% were male. The most common comorbidities in the HBV-infected cohort were hypertension (21.7%), hyperlipidemia (19.0%), and coronary artery disease (10.2%); moreover, the percentages of all comorbidities were significantly greater in the HBV-infected cohort than in the comparison cohort (all P < 0.001).

MD occurred in 450 and 1441 patients in the HBV-infected and comparison cohorts, respectively (Table 2). Notably, the incidence of MD was nearly 1.29-fold higher among the HBV-infected patients than among the non-HBV-infected patients (1.90 vs. 1.47/1000 person-years). Figure 1 illustrates the incidence curve of MD (log-rank test P < 0.001). After adjustment for age, sex, and the comorbidities, the HBV-infected patients had a 1.31-fold increased risk of developing MD compared with the non-HBV-infected patients (HR = 1.31, 95% CI = 1.17–1.46).

Table 3 presents the risk of MD development in the HBV-infected patients compared with the non-HBV-infected patients, according to age, sex, and the presence of certain comorbidities. The risk of MD was nearly 1.24-fold higher in the HBV-infected cohort than that in the comparison cohort for the patients aged ≥ 50 years (HR = 1.24, 95% CI = 1.10–1.40); by contrast, there was no statistical significance regarding the risk of MD between the HBV-infected cohort and the comparison cohort among patients aged ≤ 35 (HR = 0.99, 95% CI = 0.58–1.71), nor in those aged 35–49 years (HR = 1.28, 95% CI = 0.98–1.71). Additionally, the risks of developing MD were both nearly 1.3-fold greater in the male HBV-infected patients (HR = 1.30, 95% CI = 1.12–1.51) and female HBV-infected patients (HR = 1.31, 95% CI = 1.12–1.54), compared with the comparison cohort. Among the entire study population, the HRs of developing MD were 1.47 (95% CI = 1.20–1.80) and 1.23 (95% CI = 1.08–1.39) when there was no comorbidity and at least one comorbidity, respectively.

To determine the functional effects of the HBx protein in retinal epithelial cells, ARPE19 cells were stably transfected with either a mock plasmid or HBx-containing plasmid. As depicted in Fig. 2a, the HBx protein was stably expressed in the HBx-transfected ARPE19 cells, after being cultured in 0.5 mg/mL G418 for several weeks. Previous studies have reported that the expression of HBx induces spontaneous apoptotic cell death in mouse fibroblasts [23], human HepG2 cells [24, 25], and the liver of HBx transgenic mice [26]. However, we found that the expression of HBx at tolerant levels was not deleterious toward ARPE19 cells. In addition, the morphology (observed under a phase-contrast microscope; Fig. 2b) and cell growth rate (measured by the MTT assay; Fig. 2c) of the HBx- and mock-transfected cells were similar.

To clarify the cytotoxic effects of the HBx protein in retinal epithelial cells following UV irradiation, cell viability and clonogenic survival were measured using the MTT assay and colony formation assay, respectively. The results revealed that the expression of the HBx protein significantly decreased the cell viability of ARPE19 cells in dose- (from 1 to 5 J/cm²) and time- (from 1 to 3 days) dependent manner (Fig. 3). Similarly, the expression of the HBx protein markedly reduced the clonogenic survival of the ARPE19 cells in response to 1–5 J/cm² UV irradiation (Fig. 3c, d).

The global gene expression profiles of ARPE19 cells with or without the HBx protein were examined pre- and at 4 h post-UV irradiation. Standard selection criteria with log2 ratios of ≥ 1 (upregulated at least 2-fold) or log2 ratios of ≤ − 1.0 (downregulated at least 2-fold) and P < 0.05 were used to identify DE genes (Fig. 4a). Comparing the gene expression profiles from the HBx- and mock-transfected ARPE19 cells revealed that a total of 497 genes (185 upregulated and 312 downregulated) and 3569 genes (1560 upregulated and 2009 downregulated), respectively, were significantly DE pre- and post-UV irradiation (Fig. 4b). Moreover, these DE genes were hierarchically clustered, demonstrating a clear pattern of differential transcriptional regulation in cells with and without HBx pre- and post-UV irradiation (Fig. 4c).

To further functionally classify the DE genes in the HBx- and mock-transfected ARPE19 cells, an enrichment analysis was performed according to the multiple cellular process categories of the GO database. The results from the canonical pathway analysis of the HBx-transfected cells, compared with those of the mock-transfected cells, indicated that a significant number of the DE transcripts was assigned to functional categories of metabolic processes prior to UV irradiation: prostaglandin metabolism, the p53 signaling pathway, TNF signaling pathway, cytokine–cytokine receptor interaction, extracellular matrix–receptor interaction, steroidogenesis, various cancer-related pathways, and the PI3K-Akt signaling pathway. At 4 h post-UV irradiation, a significant number of the DE transcripts was reclassified into ten functional categories, namely BER, NER, MMR, HR, DNA replication, cell cycle regulation, p53 signaling pathway, various cancer-related pathways, circadian rhythms, and the TNF signaling pathway (Table 4 and Fig. 5a), in which most of them were down-regulated (Fig. 5b). Furthermore, according to the GO enrichment analysis, these DE genes were categorized into three groups predominantly involved in molecular functions (Table 5A), biological processes (Table 5B), and cellular components (Table 5C).

To further clarify the effects of HBV infection on the sensitivity of retinal epithelial cells to short-wavelength visible light, the cytotoxic effects of the HBx protein in ARPE19 cells following blue light exposure were also determined. The wavelength of the blue light LED lamp peaked around 470 nm (Additional file 1: Figure S1A), the chromaticity diagram indicated the light belonged to blue light area (Additional file 1: Figure S1B), and the irradiance of the blue light was approximate 26 W/m² at the distance of 6.5 cm below the light source, where the cells were exposed (Additional file 1: Figure S1C). The results of cell viability and clonogenic survival revealed that HBx-transfected ARPE19 cells were more sensitive to blue light than the mock-transfected ARPE19 cells in a dose-dependent manner (Fig. 6). In short, the HBx protein can sensitize retinal epithelial cells to UV irradiation and blue light exposure, suggesting that the HBx protein of HBV may be (at least partially) the critical factor that increases the incidence of MD in HBV-infected patients.