Ultrastructural Localization and Expression of TRPM1 in the Human Retina

Abstract

Purpose.: Transient receptor potential subfamily melastatin (TRPM)1 cation channels of retinal ON-bipolar cells are modulated via a mGluR6 (GMR6) signaling cascade. While light-microscopy shows these channels are located on the tips of ON-bipolar cells dendrites, near rod and cone synaptic ribbons, TRPM1 localization at the electron-microscope level is currently not described. The authors report here the ultrastructural localization of TRPM1 in the human retina.

Methods.: TRPM1 was localized in postmortem human retinas by immunohistochemistry at both the light and electron microscope levels. Additionally, TRPM1 expression was studied using in situ hybridization, laser dissection microscopy, and PCR techniques.

Results.: TRPM1-immunoreactivity was located on the dendrites and soma of ON-bipolar cells at the light microscope level. At the electron microscope level TRPM1-immunoreactivity was located on the tips of ON-bipolar cell dendrites that were invaginating cone pedicles and rod spherules. In addition, TRPM1-immunoreactivity was occasionally found on the rod spherules ribbons, suggesting that at least a proportion of rods may also express TRPM1. In situ hybridization showed TRPM1 encoding RNA in inner nuclear layer somata and in some photoreceptors. The presence of TRPM1-RNA in photoreceptors was confirmed by PCR in pure photoreceptor material obtained with a laser dissection microscope.

Conclusions.: In the human retina TRPM1 is expressed on ON-bipolar cell dendrites that invaginate photoreceptor terminals. TRPM1 is also expressed on the synaptic ribbons of a subclass of rods, suggesting a dual function for TRPM1 in the ON-pathway.

The retina sends information about the visual environment to the brain via a number of parallel pathways. ¹ Photoreceptor signals are split into ON- and OFF- channels at the photoreceptor synapse. Photoreceptors stimulated by light hyperpolarize and reduce their glutamate release. The reduction in glutamate causes OFF-bipolar cells to hyperpolarize and ON-bipolar cells to depolarize. The OFF- and ON-bipolar cells react differently to changes in glutamate levels as they express distinct types of glutamate receptors. OFF-bipolar cells express ionotropic glutamate receptors, ¹ which are cation channels that open when stimulated by glutamate. ON-bipolar cells express the metabotropic glutamate receptor mGluR6 ^2,3 that when activated by glutamate leads to the closure of the cation channel: transient receptor potential subfamily melastatin (TRPM)1. ^4,5  

In both mice and human, the lack of TRPM1 impairs retinal functioning at the bipolar cell level evidenced by reduced electroretinogram (ERG) b-wave while the a-wave remains undisturbed. ^{5 –7} A number of other proteins such as the G protein G_o and Nyctalopin (NYX) seems to be essential in the mGluR6/TRPM1 cascade. ^2,3,8 As for TRPM1, mutations in mGluR6 or NYX ^{7,9 –11} also led to a reduced ERG b-wave while leaving the a-wave undisturbed, suggesting that these three proteins belong to the same functional cascade. ⁸ Membership to the same functional cascade is further supported by previous work on the ultrastructural localization of mGluR6 and NYX, which indicates these proteins are localized close together on ON-bipolar cell dendrites near photoreceptor synaptic ribbons. ⁸ As yet, however, the ultrastructural localization of TRPM1 has not been performed. 

Here we show the ultrastructural localization of TRPM1 in the human retina. TRPM1-immunoreactivity (IR) was found on dendrites of ON-bipolar cells invaginating the synaptic terminals of rods and cones near the synaptic ribbon. This localization puts TRPM1 at the same location as mGluR6 and NYX, further strengthening the argument that they form a functional cascade. Surprisingly, TRPM1-IR was also found on the synaptic ribbon of some rods. Both findings were confirmed by in situ hybridization (ISH) and polymerase chain reaction (PCR). These results suggest that TRPM1 has both a pre- and postsynaptic function in the human retina. 

Methods

This study was performed in agreement with the declaration of Helsinki on the use of human material for research. Postmortem human donor eyes were obtained from the Cornea Bank Amsterdam. In accordance with The Netherlands' law, the Cornea Bank Amsterdam ensured donors had consented to their eyes being used for scientific purpose. 

Retinas were isolated by peeling away the sclera and corpus vitreous and fixed in 4% paraformaldehyde buffered in 0.1 M phosphate buffer (PB) pH 7.4 for 30 to 60 minutes. Retinas were cryoprotected in 12.5% sucrose in 0.1 M phosphate buffer for 30 minutes, then 1 hour in 25% sucrose in 0.1 M phosphate buffer before being frozen in embedding compound (TissueTek; Sakura Finetek Holland BC, Alphen aan de Rijn, The Netherlands). 

For light microscopical (LM) purposes 10-μm thick sections were made and stored at −20 °C. Sections were first preincubated in 2% normal goat serum (NGS) for 30 minutes, then incubated with primary antibodies for 24 to 48 hours followed by 35 minutes of incubation with secondary antibodies at 37 °C. Optical examination (optical section thickness is 0.25 μm) was performed with a confocal laser scanning microscope (CLSM; Meta confocal miscroscope; Zeiss, Stuttgart, Germany). 

The primary antibodies used were: TRPM1 (Sigma Aldrich Chemi BV, Zwijndrecht, The Netherlands) 1:200; calbindin (Swant, Marly, Switzerland) 1:500; mGluR6 (gift of Noga Vardi) 1:5000; G_o (Chemicon, Hampshire, United Kingdom) 1:5000; Ribeye (BD Transduction Laboratories, Breda, The Netherlands) 1:1000; Bassoon (Stress Gen, Brussels, Belgium) 1:5000; PKC_α (Sigma Aldrich Chemi BV) 1:200; and PKC_β (Sigma Aldrich Chemi BV) 1:200; diluted in 0.1 M phosphate buffered saline (PBS) containing 5% normal goat serum and 0.05% Triton. Secondary antibodies were visualized by means of goat anti-rabbit Cy3, 1:500, and goat anti-mouse Alexa 488, 1:500. 

For electron microscopical (EM) purposes 40-μm thick sections were incubated with TRPM1 (1:200) in phosphate buffer for 48 hours, then rinsed before being incubated with rabbit peroxidase antiperoxidase (PAP) for 2 hours, rinsed, then developed in a 2,2′-diaminobenzidine (DAB) solution containing 0,03% H₂O₂ for 4 minutes. Afterward the gold substitute silver peroxidase ¹² method was performed; sections were fixed in sodium cacodylate buffer (pH 7.4) containing 1% osmium tetra oxide and 1.5% potassium ferricyanide. Sections were then dehydrated and embedded in epoxy resin, ultrathin sections made, and examined with an electron microscope (FEI Technai 12; Fei Company, Eindhoven, The Netherlands). 

For Western blot analysis, human retinas were homogenized with a nonstick pestle in ice-cold phosphate buffered saline containing one tablet of protease-inhibitor cocktail (Boehringer Mannheim GmbH, Mannheim, Ingelheim, Germany) per 25 mL. Proteins fractions were isolated by centrifuge (14,600g), the supernatants and deposits were collected, sample buffer (Nu Page LDS; Invitrogen, Breda, The Netherlands) was added and then run on a gel (Nu Page 4–12% Bis Tris Cell; Invitrogen). Protein standards (Bio-rad Laboratories, BV, Veenendaal, The Netherlands) were run in adjacent lanes. Gels were electroblotted on blot membrane (PolyVinylideneDiFluoride; Millipore, Amsterdam, The Netherlands) overnight at 80 mA. Membranes were rinsed in a Tris buffer (0.5 M) containing NaCl (1.5 M) and 5% Tween-20, blocked in the same buffer containing 2% dry milk powder for 1 hour, then incubated in the primary antibodies against TRPM1 for 1 hour, washed in Tris-buffer, and incubated in goat anti-rabbit (IRDye 800 CW; 1:3000). Blots were examined on an Odyssey infrared detecting system. Western blot analysis showed a strong band between 150 kDa and 200 kDa (Fig 1). The predicted molecular weight of TRPM1 is 182 kDa, suggesting that the band between 150 kDa and 200 kDa represents TRPM1. Two additional bands were seen; a strong band between 100 kDa and 150 kDa and a weaker band between 75 kDa and 100 kDa, which are most likely degradation products or alternative splicing variations of TRPM1. 

Figure 1.

 

Western blot analysis of TRPM1 antibody shows a band between 150 kDa and 250 kDa. The expected weight for the TRPM1 antigen is 182 kDa. There is an additional band between 100 and 150 kDa, and a very weak band between 75 and 100 kDa. These bands most likely represent degradation products of TRPM1.

For in situ hybridization (ISH) sections were stored at −80 °C. Hybridization was performed using TRPM1 specific, 5′-fluorescein-labeled 19mer antisense oligonucleotides containing locked nucleic acid (LNA) and 2′-O-methyl (2OME)-RNA moieties; TRPM1 (5′-TuuCccAaaGacTugTuuC-3′) and mGluR6 (5′-TguTccTgcGguTguTcuC- 3′), where locked nucleic acid residues are given in capital letters and 2′-O-methyl in lowercase. Sense probes were used as controls (TRPM1 5′-GaaAcaAguCuuTggGaaA-3′, mGluR6 5′-GagAacAacCgcAggAacA 3′). Probes were synthesized by Ribotask ApS, Odense, Denmark. Hybridization signals were detected by incubating the sections in blocking buffer containing anti-fluorescein-alkaline phosphatase (AP) Fab fragments (1:1000; Roche, Lewes, UK) for 1 hour at room temperature. AP signal was detected by using a substrate kit (Vector Blue AP Substrate Kit III; Vector, Burlingame, CA). 

For details of sampling, RNA isolation, amplification, and PCR conditions see van Soest et al. ¹³ PCR amplifications were carried out as follows: mGluR6, 0.5 μL DNA per 50 μL, annealing temperature 52°C, 30 to 35 cycles; Ceacam, 1 μL DNA per 25 μL annealing temperature 52°C, 35 cycles; TRPM1, 1 μL DNA per 25 μL, annealing temperature 55°C, 40 cycles. All PCR experiments were amplified with β-actin as standard. 

Results

TRPM1-IR was present in the outer plexiform layer (OPL), and the inner nuclear layer (INL) of the human retina. TRPM1-IR formed two distinct patterns in the OPL: a punctuated labeling (Fig. 2A, arrowhead) and a band-like labeling (Fig. 2A, arrow). In the INL, TRPM1-IR labeled proximally situated somata with protrusions running into the OPL (Fig 2A). The identity of cells expressing TRPM1 was determined by double labeling experiments with cell type-specific markers. 

Figure 2.

 

TRPM1 localization in the human retina; confocal immunofluorescence[b]. TRPM1 is labeled red. (A) Double labeling of TRPM1 and the horizontal cell marker, calbindin (green). TRPM1-IR band-like (arrow) and punctated-like labeling (arrowhead) was observed in the OPL. Somata in the inner nuclear layer (INL) also showed TRPM1-IR. Colocalization was not found in the OPL or the INL, indicating that horizontal cells do not express TRPM1. A calbindin-IR horizontal cell is indicated by an asterisk. (B) TRPM1 double-labeled with the ON-bipolar cell marker, PKC_α. Both band-like (small arrowhead) and punctuated-like (small arrow) TRPM1 labeling colocalized with PKC_α. Note that somatic TRPM1-IR and PKC_α-IR is also present (large arrowhead). No colocalizatio between PKC_α and TRPM1 in IPL (large arrow). (C) TRPM1 double-labeled with the bipolar cell marker, PKC_β. Colocalization of TRPM1-IR and PKC_β-IR was clearly found in both the band-like (arrows) and punctated-like structures (arrowheads). Somatic colocalization was also observed (large arrowhead). TRPM1 does not colocalize with PKC_β labeling of bipolar cells synaptic terminals in the IPL (large arrows). (D) Double-labeling of TRPM1 and the G protein, G_o. Colocalization of TRPM1-IR and G_o-IR occurred in the OPL (arrows). Note that in the somata of bipolar cells (asterisk) TRPM1-IR (arrow) was separated from G_o-IR (arrowhead). (E) Double-labeling of the metabotropic glutamate receptor, mGluR6, and G_o. Colocalization of mGluR6 and G_o was readily seen in the OPL (arrows). (F) TRPM1 double labeled with the synaptic ribbon marker, Bassoon. TRPM1-IR and Bassoon-IR were strongly associated and sometimes colocalized (arrowhead) in the OPL. Note that no colocalization of TRPM1 and Bassoon was seen in the IPL. (Inset): higher magnification showing that TRPM1-IR and Bassoon-IR are strongly associated. (G) Double-labeling of TRPM1 and the synaptic ribbon marker, Ribeye. TRPM1-IR and Ribeye-IR are strongly associated and sometimes colocalized. Scale bars, 5 μm.

In the OPL, TRPM1-IR colocalization was not evident with the horizontal cell marker, calbindin (Fig. 2A, asterisk), whereas it was with bipolar cell markers. TRPM1-IR colocalized with the bipolar cell markers PKC_α and PKC_β (Figs. 2B and 2C). PKC_α labels ON-bipolar cells whereas PKC_β labels both ON and OFF cone bipolar cell according to Kolb et al., ¹⁴ whereas Haverkamp et al. ¹⁵ suggests that the PKC_β labeling is restricted to OFF types. Both PKCs showed punctated- and band-like labeling that both colocalized with TRPM1-IR (Figs. 2B and 2C; band-like labeling: arrow; punctated-like labeling: arrowhead). The TRPM1 somatic labeling in the INL also colocalized with both PKCs (Figs. 2B, 2C, large arrow). The labeled somata for PKC_α and PKC_β were localized in the same layer directly below the horizontal cells in the distal part of the INL. This suggests that both PKC_α and PKC_β label ON-bipolar cells in the human retina. No colocalization of TRPM1-IR and PKC_α-IR and PKC_β-IR was found in the IPL (Figs. 2B and 2C, large arrow). This suggests that ON-bipolar cells axons do not express TRPM1. 

Combined, all the results outlined above suggest that the somata and dendrites of ON-bipolar cells express TRPM1. The band-like labeling could represent the bipolar cell invaginating the cone terminals; the punctuated labeling could represent the bipolar cell dendrites invaginating the rod spherule. To evaluate this further, we studied the colocalization of TRPM1 with dendritically located components of the ON-bipolar cell signaling cascade that have a close proximity to photoreceptor synaptic ribbons: G_o and mGluR6. We also investigated the relation between TRPM1 and the pre-synaptic complex, using two synaptic markers that label the synaptic ribbon, Ribeye and Bassoon. ^16,17  

Double labeling experiments revealed TRPM1 and G_o are colocalized (Fig. 2D) and that both Ribeye-IR and Bassoon-IR are strongly associated with TRPM1 (Figs. 2F and 2G). Double labeling experiments could not be performed with TRPM1 and mGluR6 as both antibodies were raised in rabbit. Instead we indirectly assessed the colocalization of TRPM1 and mGluR6 by double labeling mGluR6 and G_o. As mGluR6 and G_o were found to colocalize (Fig. 2E) and as G_o also colocalizes with TRPM1 (Fig. 2D) then it can be reasonably assumed that TRPM1 and mGluR6 colocalize as well. 

So far, our results suggest several aspects regarding the location of TRPM1 in human retina. Firstly, TRPM1 is associated with both rod and cone ON-bipolar cells and is localized to both the soma and dendrites. Secondly, dendritic TRPM1 colocalizes with G_o and presumably with mGluR6 as well. Thirdly, given the close proximity of G_o and mGluR6 to photoreceptor synaptic ribbons and the strong association with Ribeye and Bassoon IR, TRPM1 is in close proximity to photoreceptor synaptic ribbons. Combined this suggests that in human retina TRPM1 channels are located on the tips of ON-bipolar cell dendrites, close to the synaptic ribbons of rods and cones. This possibility was explored further at the ultrastructural level. 

Experiments performed at the ultrastructural level focused on the OPL. Ultrastructurally photoreceptor synapses are triads with synaptic ribbons. ^{18 –20} Rod spherules have one ribbon and so can be easily distinguished from cone pedicles, which have several ribbons. At the electron microscopic level, TRPM1-IR was found on the triad's central elements, corresponding to bipolar cell dendrites, and was located close to the synaptic ribbon in both cone pedicles and rod spherules (Fig. 3). TRPM1-IR was never seen on horizontal cell dendrites (H), the triad elements lateral to the ribbons (Fig. 3). Surprisingly, ribbons in the rod spherules occasionally showed TRPM1-IR (Fig. 3C) suggesting that TRPM1 is located both post- and pre-synaptically in the rod triad. 

Figure 3.

 

Ultrastructural localization of TRPM1. (A) TRPM1-IR is located at the central position in cone pedicle. (B, C) TRPM1-IR is localized at the bipolar cell dendrites ending near the synaptic ribbon (R). Horizontal cell dendrites (H) present in the synaptic triads of cone pedicles or rod spherules did not show TRPM1-IR. Note that in (C) TRPM1-IR is visible at the position of the synaptic ribbon (R). Scale bars, 0.5 μm.

To confirm that TRPM1 can be synthesized by bipolar cells and photoreceptors, ISH experiments were performed. TRPM1-mRNA expression occurred in somata of the INL (Fig. 4A) in a similar pattern observed for mGluR6-mRNA in the INL (Fig. 4C) indicating the TRPM1-mRNA expressing somata were ON-bipolar cells. In photoreceptor nuclei mGluR6-mRNA expression was never seen whereas TRPM1-mRNA could be clearly detected in some photoreceptor nuclei (Figs. 4A and 4B, arrowheads). Some ganglion cells showed an mGluR6-mRNA signal, which is consistent with the work of Tehrani et al. ²¹ This signal was not studied further in this study. 

Figure 4.

 

In situ hybridization of antisense TRPM1 RNA. Riboprobes of TRPM1 (A, B) and mGluR6 (C). In (A, C) TRPM1 and mGluR6 label cells at the same position in the INL suggesting bipolar cells express both mGluR6 and TRPM1. In (A), the TRPM1 riboprobe was also detected in the somata of photoreceptors (arrowhead). In (B), where the outer nuclear layer (ONL) is shown at higher magnification, TRPM1 riboprobe labeling of several photoreceptor somata is apparent (arrowheads). In (C), the mGluR6 riboprobe is seen to detect ganglion cells somata (asterisk) in addition to bipolar cells. Scale bars, 10 μm.

To confirm photoreceptors indeed express TRPM1, additional PCR experiments were performed. Pure photoreceptor samples were harvested with a laser dissection microscope. They were used to study TRPM1 expression in photoreceptors and compared with TRPM1 expression total neural retina samples. The samples were taken from two retinal eccentricities, centrally (c) just outside the macula and peripheral to the large arteries (p). PCR detected TRPM1-RNA in both the photoreceptor and total retina samples (Fig. 5A). The positive TRPM1 signal in the ONL sample was not caused by bipolar cell contamination as mGluR6-RNA was present only in the total neural retina (NR) sample and not in the photoreceptor sample (Fig. 5B). To exclude that the positive TRPM1 signal in the photoreceptors was due to leaky expression, we determined the presence of Caecam3-RNA. Caecam3 is abundantly present in blood cells (Fig. 5B) but is not expressed in the retina. ²² We found strong expression of Caecam3-RNA in blood samples (Fig. 5C, lane B) whereas there were no detectable signals in photoreceptor samples (Fig 5C, lanes c and p), thus demonstrating our detection threshold settings prevented the spurious detection of leaky mRNA expression. 

Figure 5.

 

TRPM1 detected by PCR in photoreceptors[b]. (A) TRPM1 PCR product was detected in photoreceptor (photo) and total neural retina samples (retina). (B) mGluR6 PCR product was detected in the neural retina (NR), but not in the photoreceptor, samples. (C) Ceacam3 PCR product was detected in blood cells (B), but not in the photoreceptor samples. c, central; p, peripheral.

Discussion

The present study ascertained the localization and expression pattern of TRPM1 in human retina using immunohistochemistry, ISH, and PCR. In photoreceptor synapses TRPM1 is located postsynaptically, on the dendrite tips of ON-bipolar cells near synaptic ribbons of both cone pedicles and rod spherules, and pre-synaptically on ribbon of some rod spherules. 

All proteins known to be involved in the initiation of the ON-bipolar cell transduction cascade have now been shown in the synaptic triads of cone pedicles and rod spherules. In primate, mouse, and now human retina mGluR6, ³ NYX ⁸ and Go, ²³ and TRPM1, respectively, are found on the central element of the triads close to the synaptic ribbons. This suggests that all these proteins are situated in a single complex. 

In addition to the dendritic labeling, we also observed that the ribbon of some rod spherules had TRPM1-IR. This indicates the TRPM1 protein is indeed present on (some) synaptic ribbons. As this was a most unexpected finding we confirmed the result with PCR and ISH experiments. However, despite the novelty of this result there is precedence for TRP-channel localization on synaptic ribbons. Transient receptor potential subfamily Vanniloid-1 (TRPV1) channels have been found on photoreceptor ribbons in goldfish and zebrafish. ²⁴ These authors suggest the TRPV1 channels are involved in docking the synaptic vesicles to the synaptic ribbon or in the transport of synaptic vesicles along the synaptic ribbon. However, it is clear that further experiments are required to resolve the role of TRP channels in the pre-synaptic complex. 

On average, the human retina contains 92 million rods ²⁵ and it is generally assumed that all rods are equal. However, this may not be entirely true as we found some rods express TRPM1 whereas others do not. In fish retina different types of rods have been described based on their connectivity patterns. ²⁶ Most rods contact both horizontal and bipolar cells but some rods only contact horizontal cells. Together these results suggest that subclasses of rods exist in both the human and the fish retina. 

In humans, mutations in TRPM1 leads to congenital stationary night blindness (CSNB) type 1. ^{7,9 –11} In these patients the ERG b-wave is absent while the a-wave remains intact. The absence of b-wave indicates a deficiency of ON-bipolar cell responses, which is consistent with a disturbed mGluR6-TRPM1 cascade function. Our present results suggest that additional pre-synaptic modifications may also be present. Such pre-synaptic modifications might not be visible in the ERG of congenital stationary night blindness type 1 patients because their remaining ERG signal consists only of the a-wave, which is generated by the light response of the photoreceptor's outer segment and not by light-induced vesicle release. 

Footnotes

 Supported by European Commission FP7 Grant RETICIRC HEALTH-F2-2009-223156 (coordinator, MK), and ZonMW-NWO (MK).

Footnotes

 Disclosure: J. Klooster, None; J. Blokker, None; J.B. ten Brink, None; U. Unmehopa, None; K. Fluiter, None; A.A.B. Bergen, None; M. Kamermans, None

The authors thank Ton Put for preparing the photographs, Marcus Howlett for his critical comments on the manuscript and correcting the English, and the Corneabank Amsterdam for supplying the human retinas (Director, Tamás Csikós).