The results presented here provide extensive new information concerning TRP channel expression in the retina, and suggest new directions for uncovering their functions. It is somewhat surprising that mRNA from every TRP channel gene is detectable in the mouse retina. Of course, the technique used, RT-PCR is extremely sensitive, and the presence of a small amount of mRNA does not guarantee that a functionally important amount of protein is present. Moreover, the RT-PCR product intensities cannot be reliably correlated with the levels of message present in the retina. Thus, the results imply that it will be worthwhile to check knockouts of every TRP channel gene for effects on retinal function and health, and considering them as candidate genes for hereditary retinal disease.
At the mRNA level, the TRP family members fall into three main groups. One has the highest levels of mRNA as indicated by northern blots, and includes TRPC1, TRPC3, TRPC4, TRPM1, TRPM3, TRPML1, TRPV2, and TRPV4, as well as the TRP-associated protein PKD1. A second group has weaker but still robust expression, and includes TRPC4, TRPC5, TRPM7, and TRPP2. Expression of the other 16 TRP channel genes was much weaker. The high expression levels motivate pursuing further the potential functions of the 12 most abundant species. However, the less abundant mRNA species could be functionally important.
TRPC6 and TRPC7 mRNAs seem to be present at very low levels, but they have been proposed as candidates for the transduction channel of melanopsin-expressing retinal ganglion cells (Sekaran et al., 2007
). In addition, a recent publication identified TRPC6 within cells of the INL and the GCL of the retina (Wang, Teng, Li, Ge, Laties & Zhang, 2011
). We did not detect any TRPC6 mRNA in the retina by ISH, and northern blotting indicates that TRPC6 mRNA levels are low in the retina. The differences between our study and those of Wang et al. may be due in part to probe differences, but may also result from differences in protocol for ISH. The high stringency methods we used may have decreased possible signal from low-abundance TRPC6 mRNA. We tested for TRPC6 protein by immunoblotting with two anti-TRPC6 antibodies and they were able to detect TRPC6 protein only in the brain, but not in the retina. At longer exposures, some signal at a size corresponding to TRPC6 signal also appeared in the retina lane, but at those exposures the antibody was cross-reactive with multiple proteins at sizes not expected for TRPC6. Immunostaining of the retina with these antibodies yielded results consistent with the work by Wang et al. Although the antibodies gave weak signal-to-background levels, slightly stronger signal was detected in the lower row of cells in the INL as well as in the GCL.
TRPV1 is barely detectable in the retina using a probe that reliably detects TRPV1 message in dorsal root ganglion, yet TRPV1 has been reported to play an important role in Ca2+
entry leading to apoptosis in retinal ganglion cells under pressure (Sappington et al., 2009
). Sappington et al. used an mRNA probe corresponding to an N-terminal portion of the expressed protein, whereas we used two probes, corresponding to a C-terminal region of the protein as well as the 5′ UTR of the TRPV1 mRNA. Both probes failed to show significant signal by ISH of retina cryosections. Paraffin section ISH, used by Sappington et al, has inherent differences compared with our ISH method and may have provided for a more stable signal for low-abundance TRPV1 mRNA. We utilized an anti-TRPV1 antibody and tested it against yeast expressed rTRPV1. While our antibody detected the expressed protein, it failed to detect significant levels of protein in brain or retina, and failed to provide signal by immunostaining of the retina. The antibody we utilized, though specific for TRPV1, may not have been sensitive enough to replicate results from the previous report.
The most striking pattern of localization of both mRNA and immunoreactivity was observed for TRPM1, which is now well established as essential for light responses of ON bipolar cells (Shen et al., 2009
). As seen previously (Kim et al., 2008a
, Koike, Obara, Uriu, Numata, Sanuki, Miyata, Koyasu, Ueno, Funabiki, Tani, Ueda, Kondo, Mori, Tachibana & Furukawa, 2010
, Morgans et al., 2009
, Nakajima, Moriyama, Hattori, Minato & Nakanishi, 2009
), the staining and mRNA localization in the retina was consistent with expression in these neurons, but most of the staining was in the cell body, not the dendritic tips where mGluR6 is located. A new finding is that TRPM1 is also expressed at even higher levels in the ciliary body, and that the distribution of splice variants there is distinctly different from that in the retina. Complimentary observations were made for the closest relative of TRPM1, TRPM3, which is also expressed strongly in the INL, with weaker mRNA signal in the outer nuclear layer and ganglion cell layer, but very strong expression in the retinal pigmented epithelium. Prior to its implication in ON-bipolar cell signaling, interest in TRPM1 centered on its role in pigmented cells and melanoma (Duncan, Deeds, Hunter, Shao, Holmgren, Woolf, Tepper & Shyjan, 1998
, Fang & Setaluri, 2000
, Miller, Du, Rowan, Hershey, Widlund & Fisher, 2004
, Zhiqi, Soltani, Bhat, Sangha, Fang, Hunter & Setaluri, 2004
), so it is notable that the pigmented ciliary body contains high levels, suggesting a functional role related to pigmentation. A relationship between TRPM3 function and pigmentation is suggested by our observation that albino animals show altered levels of expression of some splice variants as compared to pigmented mice.
Decreased expression of shorter TRPM1 variants has been correlated with an increase in metastatic melanomas in human pigmented melanocytes (Duncan, Deeds, Cronin, Donovan, Sober, Kauffman & McCarthy, 2001
, Fang & Setaluri, 2000
). Our discovery that variants of TRPM1 are differentially expressed at high levels in either the ciliary body or in the INL of the retina, suggest the interesting possibility that TRPM1 could play a role in some forms of uveal melanoma and/or forms of melanoma-associated retinopathy (MAR) that show selective reduction of b-wave amplitudes in electroretinogram (ERG) waveforms (Keltner, Thirkill & Yip, 2001
, Kim, Alexander & Fishman, 2008b
Another striking expression pattern is that of TRPM7, which two antibodies localize to cone outer segments. TRPM7 has been suggested to be important in homeostasis of Mg2+
or other metal ions, and it may play such a role in cone outer segments. TRPM7 is a bifunctional channel that could affect cone photoreceptor physiology through its channel activity or its C-terminal alpha kinase activity. Little is known about the endogenous activity and targets of TRPM7 kinase activity, but this activity may be sensitive to changes in the concentrations of intracellular divalent cations (Penner & Fleig, 2007
). Because TRPM7 is outwardly rectifying, it exhibits high divalent selectivity with little voltage-dependence over the physiological voltage range for cone photoreceptors. While TRPM7 current is typically small under physiological concentrations of extracellular cations, it is possible that TRPM7 modulates the intracellular concentration of Zn2+
, and Co2+
as the intracellular concentration of Ca2+
changes with light-dependent hyperpolarization. It is also possible that TRPM7 may have some function in synaptic vesicle release in cone photoreceptors (Krapivinsky, Mochida, Krapivinsky, Cibulsky & Clapham, 2006
TRPP2 immunostaining is also enriched in cone outer segments, which are modified cilia, consistent with its previous localization to cilia in the kidney. While TRPP2 is often localized to primary cilia, it is not exclusively expressed in ciliary membranes (Hoffmeister, Babinger, Gurster, Cedzich, Meese, Schadendorf, Osten, de Vries, Rascle & Witzgall, 2011
), and was not restricted to the connecting cilia of dissociated outer segments. TRPP2 is also expressed in other neurons throughout the retina, especially in the INL, and may play a role in the cilia of those cells, or in some unknown function. The mRNA for PKD1, a known binding partner for TRPP2, is also detected in all nuclear layers, consistent with their forming a complex as they do in the kidney. Hydrostatic pressure and fluid flow are known to be important for retinal health and it may be that this complex plays a sensing role similar to its role in the kidney. While TRPP2 has not been previously reported in the retina (please note that TRPP2 as used here is identical with the product of the PKD2
gene or polycystin-2, and is not to be confused with TRPP3/PKD2L, which some authors refer to as TRPP2, while referring to TRPP2 as “TRPP1”) mutations in TRPP2 may lead to retinal damage (Feng, Wang, Stock, Pfister, Tanimoto, Seeliger, Hillebrands, Hoffmann, Wolburg, Gretz & Hammes, 2009
). A recent publication has demonstrated that TRPP2 can interact with retinitis pigmentosa
protein RP2 suggesting a possible link between TRPP2 and retinal ciliopathies (Hurd, Zhou, Jenkins, Liu, Swaroop, Khanna, Martens, Hildebrandt & Margolis, 2010
TRPC1 has been reported to interact with TRPP2 (Bai, Giamarchi, Rodat-Despoix, Padilla, Downs, Tsiokas & Delmas, 2008
, Tsiokas, Arnould, Zhu, Kim, Walz & Sukhatme, 1999
), and its expression levels and distribution levels are such that it could easily do so in the retina. However, its levels as indicated by both mRNA and immunostaining appear much higher relative to brain and kidney than do the levels of TRPP2, suggesting that most of the TRPC1 likely has a function independent of that complex. One such function of TRPC1 could include modulating the Ca2+
concentration associated with tonic synaptic activity in photoreceptors (Szikra, Cusato, Thoreson, Barabas, Bartoletti & Krizaj, 2008
). TRPC1, TRPC3, TRPC4 and TRPC5 are all expressed robustly in the retina, and, except for TRPC5, are found in multiple cell types, as they are in other tissues, where many functions have been proposed, but not yet unambiguously demonstrated. A common theme may be coupling to activation of PLC-linked GPCR signaling cascades, which are numerous in retinal neurons. The TRPC5 staining of cells whose location suggests they may be amacrine cells, could reflect a function specific for those cells, whose identity will be important to test. TRPC5 is found in hippocampal neurons and has been shown to interact physically with TRPC1 in the brain (Goel, Sinkins & Schilling, 2002
, Strubing, Krapivinsky, Krapivinsky & Clapham, 2001
). Though little is known regarding the function of TRPC5 in neurons, the function of TRPC5 in the retina may be analogous to its function in the brain.
TRPV2 also has a striking pattern of localization. It has previously been reported that TRPV2 is localized to the plexiform layers of the retina (Leonelli, Martins, Kihara & Britto, 2009
, Yazulla & Studholme, 2004
), and our data also demonstrate strong outer plexiform staining using an anti-TRPV2 antibody in 21 day old mice. This staining appears to be localized to the photoreceptor axons, based on its proximity to pre-synaptic ribeye staining. However, we did not detect TRPV2 staining in the IPL as has been reported previously. Both previous studies utilized the same commercial TRPV2 antibody, but neither validated the specificity of the antibody for TRPV2. The antibody we use was confirmed to recognize authentic rTRPV2 expressed in yeast membranes, whose size and identity were further confirmed via an epitope tag expressed at the rTRPV2 terminus. Leonelli et al. (2009)
observed sparse immunostaining in the IPL and GCL in the rat retina at P60, but not at P15, so that even if the signal was really due to TRPV2, it may be that P21 mouse retina does not contain detectable levels of TRPV2 in those cell layers. We did not detect TRPV2 immunoreactivity in mouse RGC neurons as was previously reported, but we cannot rule out the posssibility that at the age examined the levels of TRPV2 is present in those cells but at levels below our detection limit.. In addition to the retina, TRPV2 is localized to the RPE, consistent with our detection of TRPV2 mRNA in the eyecup, and may play a role in thermal regulation of RPE cells (Cordeiro, Seyler, Stindl, Milenkovic & Strauss, 2010
). Although it has been demonstrated that TRPV2 is heat-sensitive (Caterina, Rosen, Tominaga, Brake & Julius, 1999
) additional functional roles have been proposed, and it will be interesting to explore retinal phenotypes of TRPV2 knockout mice.
Determining the levels and localization of TRP channel gene expression is only a first step toward elucidating their roles in retinal function, health and disease. However, having the results of such a survey should prove a valuable tool for guiding future work on TRP channel function in the retina as well in less accessible parts of the nervous system.