Bestrophin 1 is indispensable for volume regulation in human retinal pigment epithelium cells

Abstract

In response to cell swelling, volume-regulated anion channels (VRACs) participate in a process known as regulatory volume decrease (RVD). Only recently, first insight into the molecular identity of mammalian VRACs was obtained by the discovery of the leucine-rich repeats containing 8A (LRRC8A) gene. Here, we show that bestrophin 1 (BEST1) but not LRRC8A is crucial for volume regulation in human retinal pigment epithelium (RPE) cells. Whole-cell patch-clamp recordings in RPE derived from human-induced pluripotent stem cells (hiPSC) exhibit an outwardly rectifying chloride current with characteristic functional properties of VRACs. This current is severely reduced in hiPSC-RPE cells derived from macular dystrophy patients with pathologic BEST1 mutations. Disruption of the orthologous mouse gene (Best1(-/-)) does not result in obvious retinal pathology but leads to a severe subfertility phenotype in agreement with minor endogenous expression of Best1 in murine RPE but highly abundant expression in mouse testis. Sperm from Best1(-/-) mice showed reduced motility and abnormal sperm morphology, indicating an inability in RVD. Together, our data suggest that the molecular identity of VRACs is more complex--that is, instead of a single ubiquitous channel, VRACs could be formed by cell type- or tissue-specific subunit composition. Our findings provide the basis to further examine VRAC diversity in normal and diseased cell physiology, which is key to exploring novel therapeutic approaches in VRAC-associated pathologies.

Keywords: bestrophin 1; induced pluripotent stem cell; mouse sperm; retinal pigment epithelium; volume-regulated anion channel.

Fig. 1.

RNA and protein expression analysis of BEST1. (A) RNA expression of BEST1 in 14 mouse tissues and (B) RPE and testis across five mammalian species by RT-PCR. GAPDH served as a control for RNA integrity. For primer sequences, see SI Appendix, Table S5. (C) Western blot analysis in 14 mouse tissues and (D) RPE and testis from human, pig, and mouse using α-C45 or α-334 antibody, respectively. Costaining was done with cell-specific markers for glia cells (α-Gfap), RPE (α-Rpe65), and retina (α-Rs1h). Anti-beta actin served as the control. (E) Spatiotemporal RNA gene expression in mouse testis development by quantitative RT-PCR at indicated postnatal time points. Testis samples (n = 3) were measured in triplicates and normalized to a fixed amount of extracted total RNA. For each sample, the mean ± SD is given. (F) Relative mRNA expression of indicated chloride channels in mouse RPE (n = 2, prepared from 24 eyes), mouse testis (n = 3), human RPE (n = 2), and human testis (n = 3), normalized to Hprt1. For each sample, the mean ± SD is given. For primer sequences and Roche library probes, see SI Appendix, Table S7.

Fig. 2.

Effects of Best1 deficiency on sperm quality parameters in CD-1 mice. (A) Immunofluorescence staining of spermatozoa from wild-type and Best1^−/− mice and (B) epididymal cryosection with α-C45 (also see SI Appendix, Fig. S1 A–C). (Scale bar, 10 µm and 40 µm, respectively.) (C) Spermatozoa from 12 CD-1 male mice were subjected to cell surface biotinylation. Labeled sperm sample was immunoprecipitated with streptavidin, transferred to nylon membranes, and probed as indicated (n.b., unbound fraction). (D) CD-1 Best1^−/− spermatozoa showing enhanced decapitation and angulated sperm tails. Percentages of straight tails and (E) stained (dead) and colorless (live) spermatozoa from the cauda epididymis after eosin–nigrosin staining from 100 spermatozoa of wild-type (n = 10) and CD-1 Best1^−/− (n = 5) mice. (F) CD-1 Best1^−/− mice showing acrosome-reacted spermatozoa using PNA-488 staining. Percentages of acrosome intact cells (uniform apple green fluorescence) to acrosome-reacted sperm (perforated or absent acrosome cap) from 150 spermatozoa of wild-type (n = 16) and Best1^−/− mice (n = 10) of spermatozoa from caput, corpus, and cauda epididymis. (G) Box and Whisker plots of sperm velocity parameters including VSL, VCL, ALH, and VAP. Each box plot shows the interquartile range of 382 spermatozoa from CD-1 wild-type (n = 3) and 654 spermatozoa from Best1^−/− mice (n = 7). For further details, see SI Appendix, Table S2. Spermatozoa were released into TYH290 media. Values are given as mean + SD. Two-sided Student’s t test or Wilcoxon signed rank test: * P < 0.05; ** P < 0.01; *** P < 0.001.

Fig. 3.

Consequences of Best1 deficiency on calcium signaling and osmotic tolerance in mouse sperm. (A–F) Ca²⁺ imaging in the sperm head from spermatozoa of corpus epididymis from CD-1 wild-type and CD-1 Best1^−/− mice (n = 3–6) released into TYH420 media. (A) Representative images of cytosolic Ca²⁺ levels [Ca²⁺]_i in mouse sperm unstimulated (Left) or upon ionomycin application (Right). The pseudocolor image shows the [Ca²⁺]_i increase in sperm head. (B) Bar graphs showing basal [Ca²⁺]_i levels. (C) Averaged graphs of [Ca²⁺]_i increase upon hypotonic stimulation. (D) Summary of basal [Ca²⁺]_i levels, ATP-induced, and swelling-activated increases in [Ca²⁺]_i obtained from experiments shown in B, C, and F. Ionomycin was used for determination of maximum [Ca²⁺]_i. Black lines indicate the averaged responses of at least 30 spermatozoa; gray area indicates SEM. Also see SI Appendix, Fig. S4. (E) Bar graphs showing recovery rates (slope·s⁻¹) from swelling-induced [Ca²⁺] maximum peak to initial levels. (F) Averaged graphs of [Ca²⁺]_i increase in response to 100 µM ATP.·(G) Percentages of total sperm motility from caudal spermatozoa of three CD-1 wild-type and seven Best1^−/− mice exposed to defined osmotic conditions. Recordings were taken at indicated time points after initial exposure to TYH290, TYH350, or TYH420 media. (H) Summery of G. Mean values were pooled for the three indicated osmotic conditions. For further details, see SI Appendix, Table S2. Values are given as mean ± SEM. Two-sided unpaired Student’s t test: * P < 0.05; ** P < 0.01; *** P < 0.005. n, number of spermatozoa.

Fig. 4.

Rescue of swelling-induced membrane rupture of AQP1-expressing X. laevis oocytes by coexpression of hBEST1 or mBest1. (A) Photomicrograph showing two X. laevis oocytes under hypotonic condition injected with water as the control (Left) or AQP1 (Right). (B) Cumulative plot of ruptured X. laevis oocytes, coexpressing AQP1 and the indicated BEST1 constructs in a time course of 0–8 min after exposure to hypotonic media. Data were plotted as colored circles and fitted by double-exponential regression. (C) Bar graph summarizing B. T50, time to burst with 50% of the total number of oocytes indicated. (D–F) Voltage-clamp recordings of X. laevis oocytes, expressing AQP1 (black) or AQP1/mBest1 (blue). (D) Kinetic of swelling-induced anion currents (I_swell) at –80 and 60 mV over time. (E) Current traces in response to voltage steps (20 mV intervals from –0 to 60 mV) in isotonic solution (290 mmol·kg⁻¹) and after 5 min in hypotonic solution (200 mmol·kg⁻¹). (F) Current–voltage relationship under hypo-osmotic conditions. (G) Summary of I_swell from oocytes, coexpressing AQP1 and the indicated BEST1 constructs at 60 mV. AQP1-expressing oocytes served as control. Values are given as mean ± SEM; n, number of measured cells. An asterisk indicates significant differences compared with AQP1-expressing oocytes (two-sided unpaired Student’s t test, * P < 0.05). Defolliculated oocytes were carefully monitored for endogenous I_swell before experiments.

Fig. 5.

Expression and localization of normal and macular dystrophy-associated mutant BEST1 in hiRPE cells. (A) Images of hiRPE cells from two BD patients (+/A243V or +/Q238R) after 2 mo grown on 12-well–transwell filters (Left). Bright-field microscopy of pigmented hiRPE cells (Right). (Scale bar, 100 µm.) (B) Flat mount immunofluorescence imaging of BEST1 and ZO-1 proteins. Shown are x–y and x–z projections of confocal image stacks. (Scale bars, 20 µm.) (C) Quantification of BEST1 RNA expression by quantitative RT-PCR of total RNA extracted from hiRPE transwell filters. Samples were performed in triplicates and normalized to HPRT1. (D) Western blot analysis of hiRPE cell lysates using BEST1 antibody α-334 for detection. Beta-actin (ACTB) served as the loading control. (E) Relative mRNA expression of indicated genes in hiRPE cells from a healthy donor (BEST1^+/+) (n = 3) normalized to HPRT1. For results from patient-derived RPE cells, see SI Appendix, Fig. S5G. Primer sequences and Roche library probes are given in SI Appendix, Table S7.

Fig. 6.

Swelling-induced anion currents in hiRPE cells are greatly reduced in patients with macular dystrophy. (A–D) Whole-cell voltage-clamp recordings of hiRPE cells from a healthy donor (^+/+) and two macular dystrophy patients (+/A243V or +/Q238R) under indicated conditions. (A) Activation of endogenous anion currents upon hypotonic challenge (260 mmol·kg⁻¹) over time. Data were extracted from recordings of voltage ramps at –100 and 100 mV. (B) Averaged currents recorded at –100 to 100 mV upon hypotonic challenge for 5 min. (C) IV plot of selected recordings from A. Also see SI Appendix, Fig. S6 C and D. (D) Statistical analysis of I_swell obtained from A–C. Data points are corrected for baseline currents under isotonic conditions (290 mmol·kg⁻¹). Values are given as mean ± SEM. Two-sided paired Student’s t test: * P < 0.05.

Fig. 7.

Swelling-induced anion currents in hiRPE cells display time- and voltage-dependent inactivation. (A–D) Voltage-dependent inactivation and recovery from inactivation of I_swell during a two voltage step experiment. After inactivation of I_swell during the first voltage step to 120 mV for 1 s, the cell was held at –60 mV with increasing time intervals and then stepped back to 120 mV. (A) Representative recording from the control (BEST1^+/+). Asterisks indicate time points of current amplitudes analyzed in B and C. (B) Summary of experiments from hiRPE cells of control and BD patients. Also see SI Appendix, Fig. S6 E–H. (C) Kinetics of I_swell (at –100 mV and 100 mV) in hiRPE cells (BEST1^+/+) under the Ca²⁺ conditions indicated. A 1-s depolarizing prepulse led to voltage-dependent inactivation. A second pulse with varying delay revealed time-dependent recovery from inactivation. (D) Summary of voltage-dependent inactivation and time-dependent recovery from inactivation of I_swell in two-pulse experiments in hiRPE cells (BEST1^+/+) under the Ca²⁺ conditions indicated. The currents are corrected for baseline. Values are given as mean ± SEM.

Fig. 8.

Stable knockdown of LRRC8A in hiRPE cells has no effect on I_swell. (A) Relative mRNA expression of hLRRC8A in hiRPE^+/+ cells virally transducted with scrambled (hiRPE_scramble) and LRRC8A shRNA (hiRPE_shLRRC8A), normalized to Hprt1 expression (n = 2, pooled data from analysis after 1 and 4 wk of lentiviral transduction; mean ± SD). (B) Western blot analysis of hiRPE_scramble and hiRPE_shLRRC8A lysates using α-LRRC8A. Per lane, three 12-well–transwell filters were pooled. Anti-beta actin served as the control. (C) Representative graph showing swelling-induced whole-cell currents of a hiRPE_scramble cell using a standardized protocol with current activation at 260 mmol·kg⁻¹ for 300 s and inactivation at 290 mmol·kg⁻¹ for at least 200 s Data were extracted from recordings of voltage ramps at –100 and 100 mV. (D) Bar graph depicting the fraction of responders versus nonresponders of analyzed hiRPE_scramble (n = 63) and hiRPE_shLRRC8A (n = 50) cells.