Prolactin protects retinal pigment epithelium by inhibiting sirtuin 2-dependent cell death

Abstract

The identification of pathways necessary for retinal pigment epithelium (RPE) function is fundamental to uncover therapies for blindness. Prolactin (PRL) receptors are expressed in the retina, but nothing is known about the role of PRL in RPE. Using the adult RPE 19 (ARPE-19) human cell line and mouse RPE, we identified the presence of PRL receptors and demonstrated that PRL is necessary for RPE cell survival via anti-apoptotic and antioxidant actions. PRL promotes the antioxidant capacity of ARPE-19 cells by reducing glutathione. It also blocks the hydrogen peroxide-induced increase in deacetylase sirtuin 2 (SIRT2) expression, which inhibits the TRPM2-mediated intracellular Ca(2+) rise associated with reduced survival under oxidant conditions. RPE from PRL receptor-null (prlr(-/-)) mice showed increased levels of oxidative stress, Sirt2 expression and apoptosis, effects that were exacerbated in animals with advancing age. These observations identify PRL as a regulator of RPE homeostasis.

Keywords: Age-related retinal degeneration; Antioxidant; Prolactin; Retinal pigment epithelium; SIRT2; TRPM2 channels.

Copyright © 2016 The Authors. Published by Elsevier B.V. All rights reserved.

Figures

Fig. 1

PRL receptor and PRL expression in the RPE. (a) Representative image of in situ hybridization of transverse sections from rat eyes showed marked expression of the PRL receptor in the RPE, the outer limiting membrane (OLM), the outer (ONL) and inner (INL) nuclear layers, and ganglion cell layer (GCL) with the antisense probe (left panel). No specific staining was observed with the sense probe (right panel). Arrows indicate RPE. (b) Representative confocal stack image of RPE whole mounts from wild-type (prlr+/+) and PRL receptor-null (prlr−/−) mice showing PRL receptor (green) and F-actin (stained by phalloidin, red) immunofluorescence in x-y axis and z-axis projection (0.5 μm step size). No specific signal against PRL receptor was observed in prlr−/− RPE cells. Corresponding separate x-y axis images were included in Supplemental Fig. 2. (c) Representative confocal images of transverse sections of rat retinas showing RPE-65 (red) and PRL (green) immunofluorescence. PRL was present in choroid and retinal cell layers. Co-labeling of RPE and PRL showed strong immunostaining in the cytoplasm of RPE cells (yellow, right panel). Magnification bars were as indicated. In experiments of each panel, images were captured in three different regions of the same retina section (n = 3). Both retinas of four animals (per group in b) were analyzed (N = 6). Two male and two female from each genotype (prlr+/+ and prlr−/−) were analyzed in b.

Fig. 2

Endogenous PRL is a trophic factor for RPE cells. (a) RT-PCR of PRL receptor (PRLR, 200 bp) in ARPE-19 cell extracts. GAPDH was used as a positive control. bp, DNA ladder. RT-PCR was performed in RNA extracted from three independent cell cultures (N = 3). (b) Confocal stack image of ARPE-19 cells showing F-actin (red), nucleus (blue), and PRL receptor (green) immunofluorescence in x-y axis. Corresponding x-y axis images and z-axis projection were included in Supplemental Fig. 3a and b, respectively. PRL receptor was present in a non-uniform, punctuate distribution along the ARPE-19 cell border (left panel). The secondary antibody control was carried out by omitting the anti-PRL receptor antibody (right panel). Images were captured in three different regions of the culture plate (n = 3), three independent cultures were analyzed (N = 3). (c) Western blot analysis of PRL in conditioned media (CM) from 3-day ARPE-19 cells. A standard of 23-kDa PRL (Std) was included, and equal concentrations of total proteins were loaded. Also, a series of standard PRL dilutions was included at indicated concentrations. CM from five 3-day ARPE-19 cell cultures concentrated by a ~ 2-fold factor was analyzed and this same experiment was repeated three times (N = 3). (d) Quantification of PRL in the CM of 3-day ARPE-19 cells by the Nb2-bioassay. 10% FBS medium was included. Note that CM was not concentrated. Three samples from CM from five 3-day ARPE-19 cell cultures were analyzed and this same experiment was repeated three times (N = 3). (e) Effect on survival of ARPE-19 cells treated with 10% FBS, polyclonal α-PRL antibody (α-PRL Ab) or preimmune serum (ctl Ab), pure PRL receptor antagonist Del1-9-G129R-hPRL at 0.1 and 1 μM or untreated (ctl) for 48 h was measured by MTT assay (n = 8; N = 3 independent experiments). BLK, averaged blank. (f) ARPE-19 proliferation after treatment with 100 pM hPRL, 1 μM Del1-9-G129R-hPRL or no treatment was measured by incorporation of [3H]thymidine for 24 h after the treatments. Scintillation signals were normalized to the untreated condition (n = 9; N = 3 independent replicates). All bar plots, mean plus S.E.M.; P values: ANOVA and Bonferroni post-hoc test. Of note, the control condition corresponds to cells that were pre-treated with 1% FBS medium during 12 h followed by a 48-h period in the same medium.

Fig. 3

PRL promotes human RPE cell survival. (a) Dose-response curve showing the inhibitory effect of hydrogen peroxide (H2O2) on ARPE-19 survival. Increasing H2O2 doses diminished ARPE-19 cell survival. (b) Proliferation of ARPE-19 cells was reduced by 82 ± 6% and (c) apoptosis was increased by 51 ± 12% with 100 μM H2O2 present for 24 h. (d) Timeline diagram depicting the design of the preventive experiments e and f. (e) When applied alone, hPRL promoted ARPE-19 cell survival, while application of hPRL 24 h before the 24-h treatment with 100 μM of H2O2, prevented the H2O2-induced reduction of survival. (f) hPRL prevented the H2O2-mediated increase of apoptosis, while it had no effect alone. (g) Timeline diagram depicting the design of the restorative experiments h and i. (h, i) When applied during the last 6 h of a 24-h treatment with 100 μM H2O2, hPRL prevented the H2O2-induced reduction of survival (h) and the H2O2-mediated induction of apoptosis (i), while hPRL alone had no effect. Note that the restorative scheme of hPRL administration showed results similar to those with the preventive scheme of administration. Incorporation of [3H]thymidine was normalized to the untreated condition. (n = 12; N = 3 independent replicates). All bar plots, mean plus S.E.M.; P values: ANOVA and Bonferroni post-hoc test. Of note, the control condition corresponds to cells that were pre-treated with 1% FBS medium during 12 h followed by a 24-h (b, c, h, and i) or 48-h (e, f) period in the same medium.

Fig. 4

PRL boosts antioxidant defenses and prevents the oxidant-induced SIRT2 effects on human RPE cells. (a) Effect of the antioxidant N-acetyl cysteine (NAC, 10 mM) on ARPE-19 survival. NAC was applied 24 h before of the 24-h H2O2 insult. (b) ROS production and (c) antioxidant capacity in control condition, following H2O2 (100 μM) treatment for 24 h, and in the presence of hPRL (100 pM) for 48 h alone or 24 h before the 24-h treatment with H2O2. The same treatment scheme was used in d, e, and f. (d) qRT-PCR analyses of mRNAs for catalase, Mn2+-dependent SOD (Mn2+-SOD), Cu2+/Zn2+-dependent SOD (Cu2+/Zn2+-SOD), and glutathione peroxidase in ARPE-19 cell lysates. qRT-PCR data are normalized to TBP levels (n = 3; N = 3 independent replicates). (e) Reduced glutathione (GSH) levels in lysates from ARPE-19 cells. (f) qRT-PCR analyses of SIRT2 mRNA in ARPE-19 cells. qRT-PCR data were normalized to TBP levels (n = 3–6; N = 3 independent replicates). (g) Effect of the SIRT2 inhibitor AGK2 on ARPE-19 survival. AGK2 (10 μM) was applied 24 h before of the 24-h H2O2 insult. (h) Effect of hPRL on ARPE-19 survival challenged by the SIRT2 activator piceatannol. When hPRL (100 pM) was applied 24 h before of the 24-h piceatannol (10 μM) challenge, it prevented piceatannol-induced reduction of survival. ROS accumulation values were normalized to the untreated conditions. Excluding (d), n = 12; N = 3 independent replicates. All bar plots, mean plus S.E.M.; P values: ANOVA and Bonferroni post-hoc test. Of note, the control condition corresponds to cells that were pre-treated with 1% FBS medium during 12 h followed by a 48-h period in the same medium.

Fig. 5

PRL maintains human RPE cell survival by inhibiting the oxidant-induced SIRT2-dependent induction of TRPM2-mediated intracellular Ca2+ increase. (a) RT-PCR and (b) Western-blotting of TRPM2 in ARPE-19 cell lysates. (a) TBP was used as a positive control. bp, DNA ladder. RT-PCR was performed in RNA extracted from three independent cell cultures (N = 3). (b) ARPE-19 cells were untreated (Ctl), subjected to lipofectamin alone (Lipo.) or transfected with siRNA against TRPM2 (siRNA) or scramble sequence (Scr.). TRPM2 siRNA efficiently reduced TRPM2 expression as observed by immunoblotting using an anti-TRPM2 antibody that labeled a protein at the expected molecular weight for TRPM2 (171 kDa). β-Actin served as loading control. Extracts from three independent ARPE-19 cell cultures in each condition were analyzed (N = 3). (c, d) Measurement of the change in intracellular Ca2+ measured by the change in fluo-8 fluorescence (Δ fluorescence) in ARPE19 cells exposed to (c) H2O2 (100 μM) or (d) piceatannol (10 μM) while TRPM2 was blocked by siRNA against TRPM2 or PRL was applied (hPRL, 100 pmol/l, 15-min pretreatment). Scramble sequence (siRNA ctl) was used as a negative control for siRNA against TRPM2. Treatments with H2O2 or piceatannol began at time = 0 s (n = 170–190 cells; N = 3 independent replicates). (e) Ca2+-dependent fluorescence change in ARPE19 cells 180 s after application of SIRT2 inhibitor AGK2 (10 μM) combined or not with H2O2 (100 μM) (n = 160–190 cells; N = 3 independent replicates). (f) Effect of TRPM2 inhibition on survival of ARPE-19 subjected to a 24-h H2O2 insult (100 μM) by MTT assay. ARPE-19 cells were untreated (Ctl), treated with lipofectamin alone (Lipo.) or transfected with siRNA against TRPM2 or the scramble sequence (Scr.) 24 h prior initiating the MTT assay. In (c–e), signals were normalized by subtracting the Fluo-8 Δ fluorescence to the one in untreated conditions. All bar plots, mean plus S.E.M.; P values: ANOVA and Bonferroni post-hoc test.

Fig. 6

RPE death in wild-type and PRL receptor-null mouse. (a) Representative stack image of TUNEL staining (green) in RPE flat-mounts of 5-month- and 14-month-old wild-type (prlr+/+) and PRL receptor-null (prlr−/−) mice in x-y axis. Rhodamine-phalloidin labeled F-actin (red), and DAPI labeled the nuclei (blue). Arrow indicates positive TUNEL stain and asterisks indicate morphological alterations. Corresponding separate images in x-y and z-axis projection were presented in Supplemental Fig. 4. Six sections per flat mount in three mice were observed. (b) Quantification of TUNEL-positive apoptosis signal in whole mounts of RPE (left upper panel; n = 4–11; N = 3 independent replicates). Quantification of cell number in RPE flat mounts from wild-type and prlr−/− mice. Data are presented as raw numbers (n = 4 sections of 200–200 μm per group from 6 flat mounts from 3 mice were analyzed). Note that no cell had 3, 4 or 6 nuclei in RPE flat mounts. (c) Averaged surface (right upper panel) and perimeter (right lower panel) of individual RPE cells from 5-month- and 14-month-old prlr+/+ and prlr−/− mice. In (a–c), the 5-month-old prlr+/+ and prlr−/− mouse group was composed of four males and three females and of three males and four females, respectively, while the 14-month-old prlr+/+ and prlr−/− mouse group was composed of four males and two females and of three males and five females, respectively. (d) Representative images of superoxide anion production stained with DHE in RPE flat mounts of 5-month-old prlr+/+ and prlr−/− mice, and of 14-month-old prlr+/+ mice. (e) Quantification of fluorescence intensity as mean number of pixels positive for DHE staining normalized to the mean number of pixels positive for DAPI staining (not shown). Six sections per flat mount in three mice per group were observed. The 5-month-old prlr+/+ and prlr−/− mouse group was composed of one male and two females, respectively, while the 14-month-old prlr+/+ and prlr−/− mouse group was composed of two males and one female. (f) qRT-PCR analyses of catalase and SIRT2 mRNA in retina extracts of 5-month- and 14-month-old prlr+/+ and prlr−/− mice. (g) qRT-PCR analyses of PRL receptor and PRL mRNA from retina extracts of 5-month- and 14-month-old prlr+/+ and prlr−/− mice. All qRT-PCR data were normalized to TBP levels (n = 6 per genotype, 3 males and 3 females each; N = 3 independent replicates). Magnification bars as indicated. All bar plots, mean plus S.E.M. except in (b); P values: ANOVA and Bonferroni post-hoc test.

Fig. 7

Schematic depiction of the antiapoptotic conditions including the PRL signaling pathway in young RPE versus the proapoptotic conditions in aging RPE. Young RPE cells express PRL and its receptor (PRLR) that (1) signal for the synthesis of reduced glutathione (GSH) that transforms two molecules of hydrogen peroxide (H2O2) into water and (2) induces the transcription of catalase, which also processes H2O2 into water; both of these processes limit ROS levels. This antioxidant pathway is more prominent than the pro-oxidant pathway comprised by the TRPM2-mediated intracellular Ca2+ rise induced by the metabolite 2′O-acetyl-ADP-ribose (O-Ac-ADPR) that results from the deacetylation (removal of Ac group) of various substrates by SIRT2. The viability of RPE is therefore not jeopardized. In contrast, in aged RPE cells, the PRL signaling pathway is blunted. Therefore, the pro-oxidant pathway predominates over the antioxidant one. High levels of ROS, illustrated by increased H2O2, promote the up-regulation of SIRT2 which, in turn, activates TRPM2 by producing O-Ac-ADPR. The subsequent sustained rise in intracellular Ca2+ levels can depolarize mitochondria (Berridge et al., 2000), which causes apoptosis. RPE from PRLR null (−/−) mice emphasizes these features.