Background

In the vertebrate visual system, the retina initiates visual information processing. While retinal neurons use synaptic neurotransmitters, including glutamate, gamma-aminobutyric acid (GABA), and glycine, to encode and transmit visual information, neuromodulators, e.g. dopamine, nitric oxide, orexin, neuropeptide Y, and other neuropeptides, are also pivotal in the retina for visual information processing via modulating neuronal activity and synaptic transmission [1,2,3,4,5]. Dysfunction of neuromodulatory systems in the retina may result in retinal diseases, including diabetic retinopathy and glaucoma [6,7,8,9]. Hence, deciphering the expression and function of neuromodulator-related signals in the retina will promote our understanding of the retinal visual system.

Oxytocin is a well-known neurohormone for its function in labor induction and lactation. Oxytocin also acts as a neuropeptide in modulating diverse brain functions, including social, maternal, and emotional behaviors [10,11,12]. Endogenous oxytocin is synthesized and released by oxytocin neurons, which are mainly distributed in the paraventricular nucleus of the hypothalamus (PVN) and supraoptic nucleus (SON). Oxytocin modulates neuronal activity and synaptic transmission via activating the Gq protein-coupled oxytocin receptors (OxtRs), which are broadly expressed in mouse brain areas, including the visual-related brain regions [13, 14]. In the mouse primary visual cortex, oxytocin signaling mediates the experience-dependent cortical development [15] and also modulates spontaneous activity patterns in the develo** visual cortex [16]. Few studies have investigated the involvement of oxytocin signaling in visual information transmission and processing. In the primate visual pathway, OxtRs have been detected in the superior colliculus, pulvinar, and primary visual cortex to modulate gaze direction and attention [17]. Oxytocin has also been detected in the retina of rats, bovines, and humans, and the retinal oxytocin concentration is synchronized with the day/night cycle [18, 19], but the expression and function of oxytocin signal in the retina are still largely unknown. Recent studies observed the OxtR expression in the retinal pigment epithelium (RPE) of humans and rhesus, and they inferred that oxytocin signaling played an important role in the RPE-photoreceptor communication [20, 21]. However, whether OxtRs are functionally expressed in retinal neurons and the role of retinal oxytocin signaling in visual information processing remains largely unknown.

In the mammalian retina, dopamine (DA), released by DA amacrine cells, has multiple neuromodulatory roles in visual functions related to light and contrast adaptation, visual acuity, and circadian rhythmicity [1]. In addition to integrating light inputs from classic photoreceptors and intrinsically photosensitive retinal ganglion cells [22], DA amacrine cells also express some neuromodulator receptors [23] and can be modulated by some neuropeptides, including orexin and neuropeptide Y [5, 24]. Oxytocin signaling regulates the DA system at different levels, including directly modulating DA neuronal activity and DA release via OxtR in the brain [16, 16]. In the visual cortex, OxtR is found to be expressed in the GABAergic interneurons to modulate the excitatory/inhibitory ratio to refine the develo** circuits [16]. As OxtR is also expressed in the GABAergic interneurons in the retina, endogenous oxytocin signal in the retina may act to refine the retinal circuitry during the development. In humans, the role of endogenous oxytocin signals in social perceptual processes, such as perception of facial emotions and visual processing of infant faces [64], has been reported. Enriched social environment, social touch, parental caring, and parent-infant interaction are reported to increase oxytocin levels [15, 61,62,63,64]. Endogenous oxytocin signal in the retina may be involved in promoting social recognition and parental behavior via elevating visual acuity and detection, which requires to be investigated in the future.

In addition to directly regulating neuronal excitability, oxytocin plays a pivotal role in regulating neuronal synaptic transmission via OxtRs expressed in both presynaptic and postsynaptic membranes [13, 16]. In our study, we compared the differences in retinal oxytocin signals between males and females in the following four aspects: (1) Oxtr fluorescence intensities in the INL and GCL are not significantly different between males and females (Additional file 1: Fig. S1A); (2) the numbers of OxtR-eYFP+ neurons from Oxtr-Cre; Ai3 mice in the INL and GCL did not exhibit sex difference (Additional file 1: Fig. S1E); (3) oxytocin-induced retinal DA release was similar in both male and female mice (Additional file 1: Fig. S5D); and (4) intravitreal injection of oxytocin significantly reduced the amplitude of ERG b-wave in both males and females, and the oxytocin-induced ERG b-wave change did not show significant sex difference (Additional file 1: Fig. S6C-S6D). Together, these results indicate that the expression and function of oxytocin signals in adult mouse retinas largely do not exhibit sex differences.

Conclusions

In summary, our study finds that OxtR is functionally expressed in retinal GABAergic ACs, especially in DA subtypes. The possible functions of neuromodulators synthesized by retinal neurons have been extensively investigated [1, 5, 55, 78], our study indicates that neuropeptides synthesized and released from the brain may also involve in regulating retinal visual perception. Though our results suggest activation of OxtR is sufficient to regulate visual information processing, the physiological functions of endogenous oxytocin signal are worth to be investigated in the future. Previous studies found that endogenous oxytocin directly regulates midbrain dopamine neurons to promote pro-social behavior [Tissue processing, immunohistochemistry, and imaging

Mice were anesthetized with isoflurane, and the eyes were removed quickly and dissected in 0.1 M phosphate-buffered saline (PBS). Isolated eye cups were fixed in 4% PFA (dissolved in 0.1 M PBS) at 4 °C for 6 h then dehydrated with 10% (w/v) sucrose solution (dissolved in 0.1 M PBS) for 2 h, 20% for 2 h, and 30% for 24 h in sequence. The dehydrated eyes were embedded in OCT and cut at a thickness of 15 μm using a cryostat (CM1950, Leica Microsystems). The sections were mounted onto Superfrost Plus slides (Thermo Fisher Scientific, Waltham, MA) and stored at − 80 °C.

For immunostaining to determine neuronal identity, tissues were rinsed with PBS and pretreated in 0.2% Triton-X100 for 1 h at room temperature (RT) then blocked with 0.05% Triton-X100 and 10% bovine serum albumin (BSA) in PBS for 2 h at RT. Tissues were then incubated with primary antibody solution in PBS with 0.2% Triton-X100 and 1% BSA for 2–3 days at 4 °C. After rinsing in PBS three times, the tissues were incubated with secondary antibody solution (goat anti-rabbit 488, 594, 647; goat anti-mouse 647, 1:800; goat anti-guinea pig 647, 1:800; goat anti-sheep 648, 1:800, Life Technologies; donkey anti-goat 594, 1:200, Jackson ImmunoResearch) in PBS for 2 h at RT then dried and covered under glycerol:TBS (3:1) with Hoechst 33,342 (1:1000, Thermo Fisher Scientific). The primary antibodies used in this study include mouse anti-HPC-1 (1:1000, S0664, Sigma), sheep anti-Chx10 (1:1000, ab16141, Abcam), mouse anti-Calbindin (1:1000, CB300, Swant), mouse anti-Brn3a (1:50, MAB1585, Sigma), guinea pig anti-RBPMS (1:500, 43,691, PhosphoSolutions), rabbit anti-Melanopsin (UF008, 1:10,000, AB-N39, ATS), rabbit anti-vasoactive intestinal peptide (VIP, 1:1000, 20,077, Immunostar), rabbit anti-GAD65 + GAD67 (1:1000, ab183999, Abcam), rabbit anti-GABA (1:1000, A2052, Sigma), rabbit anti-GlyT1 (1:1000, AGT-011, Alomone), rabbit anti-PPP1R17 (1:500, HPA047819, the Human Protein Atlas), goat anti-ChAT (1:800, ab144p, Millipore), rabbit anti-tyrosine hydroxylase (TH, 1:1000, ab152, Millipore), mouse anti-TH (1:1000, 22,941, Immunostar), and rabbit anti-CRH (1:100, ab8901, Abcam). The sections were imaged with an Olympus VS120 slide scanning microscope. Confocal images were acquired with a Nikon A1 confocal laser scanning microscope with a × 25 objective. Images were analyzed in ImageJ (FIJI).

Quantitative fluorescence single-molecule in situ hybridization (smFISH)

The retina sections were prepared in the same way as used for immunohistochemistry. Samples were then processed according to the manufacturer’s instructions in the RNAscope Fluorescent Multiplex Assay manual (Advanced Cell Diagnostics, Newark, CA). After finishing smFISH, some samples were further stained with TH or GFP primary antibodies for 24 h at 4 °C then washed and incubated with secondary antibody. Samples were coverslipped with ProLong Gold antifade reagent with DAPI (Molecular Probes). The following probes were used in this study: Oxtr (C1, 406491), Crh (C1, 318931), and EYFP (C3, 312131). Sections were subsequently imaged with a Nikon A1 confocal laser scanning microscope with a × 25 objective lens, with 1 μm between adjacent z-sections. Probe omission or negative probes were carried out as control for every reaction.

smFISH images were analyzed as previously reported [25]. Every four adjacent z-stack images were combined. All channels were thresholded to remove background noise. Cellular regions of interest (ROIs) were defined using the GFP IF channel or TH IF channel to localize cell bodies. Since it is not easy to discriminate the single Oxtr punctum within ROIs, the cell in ROI was considered positive for Oxtr when the fluorescence intensity of Oxtr signal within the soma was more than 200 a.u. (based on the negative probe control). All counting experiments were conducted blinded to the experimental group.

Western blot analysis

The mouse retinas were isolated under the microscope, and the retinal pigment epithelium (RPE) was isolated following a previous study [79]. Tissues were lysed with RIPA lysis buffer containing PMSF protease inhibitor (100:1), and total proteins were extracted and protein concentrations were quantified with a bicinchoninic acid (BCA) assay kit (Beyotime Biotech, China). The protein sample’s final concentration was 2 or 3 μg/μl by diluting with sample loading buffer and ddH2O. A total of 20 μg protein was loaded into the polyacrylamide gel and electrophoretically transferred to the polyvinylidene difluoride (PVDF) membrane. The PVDF membrane was blocked for 1 h in 5% non-fat powdered milk and then incubated with primary antibody (rabbit anti-OxtR, 1:2000, AB181077, Abcam) overnight at 4 °C. Mouse anti-GAPDH antibody (1:80,000, 60,004, Proteintech) was used as the control. After incubating with the primary antibody, the PVDF membrane was then rinsed in TBST three times, and incubated with HRP-goat anti-mouse IgG antibody (1:7000, SA00001-1, Proteintech) or HRP-goat anti-rabbit IgG antibody (1:6500, SA00001-2, Proteintech) at room temperature for 1 h. ECL Prime Western Blotting Detection Reagent was used for fluorescence detection by an Odyssey near-infrared imaging scanner (FluorChem E System, Protein Simple, USA). The analysis of images of blots was performed with the AlphaView SA software (Protein Simple, USA). The images of the original uncropped blots have been provided in Additional file 2.

Electrophysiological recording

The retinas from Oxtr-Cre; Ai3 mice were prepared, and retinal neurons were recorded as previously described [2]. Briefly, mice were dark-adapted for at least 2-h and then anesthetized with 25% urethane (0.2 ml/100 g). The mouse retinas were dissected under dim red light in Ames’ medium (MilliporeSigma) and bubbled with 95% O2 and 5% CO2. The retina was placed in a recording chamber and perfused with oxygenated Ames’ solution at a rate of ~ 3 ml/min. eYFP-labeled neurons in the GCL were visualized using an IR-DIC microscopy. Current-clamp recordings were established with glass pipettes (5–7 MΩ) containing the following (in mM): 120 K-gluconate, 5 NaCl, 4 KCl, 10 HEPES, 2 EGTA, 4 Mg-ATP, 0.3 Na-GTP, and 7 Tris-phosphocreatine (pH was adjusted to 7.3); 30 μM D-AP5, 40 μM DNQX, 50 μM L-AP4, 2 μM ACET, 10 μM bicuculline, 10 μM TPMPA, and 10 μM strychnine are used to block NMDA receptors, AMPA receptors, KA receptors, metabotropic glutamate receptors (mGluRs), GABA(A) receptors, GABA(C) receptors, and glycine receptors during recording. 50 μM L-AP4 will completely activate mGluRs to block the further response induced by presynaptic glutamate release in the retina [2, 22, 31]. Both spontaneous activity and current injection-induced responses were recorded before and during the application of 1 μM oxytocin. Data were obtained using an Axon 700B amplifier, digitized at 10 kHz, filtered at 4 kHz, and collected using the pCLAMP software (Molecular Devices).

Intravitreal injection

Animals were dark-adapted for 24 h and then deeply anesthetized with 0.6% pentobarbital sodium (15 μl/g). One microliter oxytocin (1 mM, 500 μM, 100 μM, 10 μM, and 1 μM) or 1 μl WAY267464 (1 mM), dissolved with saline solution, was injected into the vitreous of one eye by a NanojectIII microinjector (Drummond Scientific Company, USA) at a speed of ~ 5 nl/s, and the other eye was injected with the same volume of saline as control. The 1 μl volume was chosen based on a previous study [2]. Since 1 mM oxytocin application had a reliable and large effect on the reduction of ERG b-wave, 1 mM oxytocin was used for the experiments to investigate the possible mechanisms. To further verify the results of exogenous oxytocin application, the specific and selective OxtR agonist WAY267464 [25, 80] was used, which has been widely used as the non-peptide agonist to investigate the function of oxytocin signal.

Measuring retinal dopamine level with high-performance liquid chromatography (HPLC)

About 1 h after oxytocin and saline injection, the retinas or vitreous bodies (collected by an Eppendorf pipette with a 10 μl pipette tip) were harvested [42]. As described previously [81], each frozen sample was homogenized into 100 μl of ice-cold 0.1 M perchloric acid containing 10 μM ascorbic acid, 0.1 mM EDTA disodium salt, and 0.02 μM 3,4-dihydroxybenzyl-amine. Dopamine, DOPAC (3,4-dihydroxyphenylacetic acid), and HVA (Homovanillic acid) levels were measured with the Agilent 1200 series neurotransmitter analyzer (Agilent Technologies, Santa Clara, CA, USA). Data were collected and analyzed by ChemStation (Agilent Technologies).

cFos immunostaining and electroretinographic (ERG) recording

For the cFos immunostaining experiment, mice were exposed to the environment with light intensity at ~ 1.0 cd/m2 for 1 h, and mice were anesthetized during the whole procedure. Then, mice were sacrificed and the retinas were dissected, fixed, and sliced for cFos immunostaining as described previously. Rabbit anti-cFos (1:1000, 5348, Cell Signaling) and mouse anti-TH (1:1000, 22,941, Immunostar) were used to detect cFos expression in retinal DA neurons.

To assess retinal function, ~ 30 min after the intravitreal injection, ERG was measured as previously reported [81]. The whole procedure was conducted in darkness. Mice were kept anesthetized with 0.6% pentobarbital sodium, and the pupils were dilated by compound tropicamide eye drops (Mydrin-p, Santen Pharmaceutical, Japan). Gold wire ring electrodes (3104RC, Roland, Germany) were placed onto the surface of both corneas, and ERGs were acquired by a pre-amplifier (FZG-81, Jia Long Educational Instruments, China) and band-pass filtered (0.1–100 Hz); 3-ms white light flashes were generated by a LED light source (CQ-LU9079, Qianhan Lighting, China) and presented by a custom-built Ganzfeld dome with 5 different stimulus strengths (0.0016 cd·s/m2, 0.0388 cd·s/m2, 0.31 cd·s/m2, 1.47 cd·s/m2, and 2.65 cd·s/m2). Light stimulation was controlled by a multi-data acquisition card (PCIe 6321, National Instruments, USA) with a LabVIEW-based code. The animals were placed on a thermostatic plate to maintain body temperature during the recording. The amplitudes of ERG a-wave and b-wave were analyzed after recording. OxtR antagonist—L368, 899 (5 mg/kg); vasopressin 1a receptor antagonist—SR49059 (10 mg/kg); dopamine D1 receptor antagonist—SCH23390 (5 mg/kg); or dopamine D2 receptor antagonist—L741, 626 (3 mg/kg)—were given intraperitoneally ~ 30 min prior to the intravitreal injection of oxytocin. The dosage for L368, 899; SR49059; SCH23390; and L741, 626 was determined from previous studies in mice and rats [82,83,84,85].

Quantification and statistical analysis

All image analyses were carried out in ImageJ (FIJI, NIH). The number of neurons and the number of animals used in every experiment are provided in the figure legends. Group data are expressed as the mean ± SEM. Statistical analysis was performed in GraphPad Prism (GraphPad). Normality was evaluated by the Kolmogorov–Smirnov normality test using GraphPad Prism. For two-group comparisons, statistical significance was determined by two-tailed paired or unpaired Student’s t-tests, and Wilcoxon signed-rank test or Mann–Whitney test when assumptions for parametric testing were not satisfied. For multiple group comparisons, two-way and one-way analyses of variance (ANOVA) tests were used for normally distributed data, followed by post hoc analyses. For data that were not normally distributed, non-parametric tests for the appropriate group types were used instead. p < 0.05 was considered statistically significant.