Inhibition of microRNA-155 Protects Retinal Function Through Attenuation of Inflammation in Retinal Degeneration

Abstract

Although extensively investigated in inflammatory conditions, the role of pro-inflammatory microRNAs (miRNAs), miR-155 and miR-146a, has not been well-studied in retinal degenerative diseases. We therefore aimed to explore the role and regulation of these miRNA in the degenerating retina, with a focus on miR-155. C57BL/6J mice were subjected to photo-oxidative damage for up to 5 days to induce focal retinal degeneration. MiR-155 expression was quantified by qRT-PCR in whole retina, serum, and small-medium extracellular vesicles (s-mEVs), and a PrimeFlow™ assay was used to identify localisation of miR-155 in retinal cells. Constitutive miR-155 knockout (KO) mice and miR-155 and miR-146a inhibitors were utilised to determine the role of these miRNA in the degenerating retina. Electroretinography was employed as a measure of retinal function, while histological quantification of TUNEL⁺ and IBA1⁺ positive cells was used to quantify photoreceptor cell death and infiltrating immune cells, respectively. Upregulation of miR-155 was detected in retinal tissue, serum and s-mEVs in response to photo-oxidative damage, localising to the nucleus of a subset of retinal ganglion cells and glial cells and in the cytoplasm of photoreceptors. Inhibition of miR-155 showed increased function from negative controls and a less pathological pattern of IBA1⁺ cell localisation and morphology at 5 days photo-oxidative damage. While neither dim-reared nor damaged miR-155 KO animals showed retinal histological difference from controls, following photo-oxidative damage, miR-155 KO mice showed increased a-wave relative to controls. We therefore consider miR-155 to be associated with the inflammatory response of the retina in response to photoreceptor-specific degeneration.

Keywords: Inflammation; Macrophage; Microglia; Photo-oxidative damage; Retina; miR-155; miRNA.

Fig. 1

Inhibition of miR-155 was protective against photo-oxidative damage (PD)–induced degeneration. a–b miR-155 inhibitor–injected retinas demonstrated a significantly greater function, as measured by ERG, with a higher (a) a-wave and (b) b-wave amplitude compared with the negative control (p < 0.05). c–e Quantification of photoreceptor cell death by TUNEL labelling demonstrated no significant difference in TUNEL⁺ cell counts between groups. d–e Representative images displaying TUNEL immunolabelling of retinal cryosections. f ONL thickness measured across the length of the retina showed a significantly thicker ONL in the miR-155 inhibitor mice (p < 0.05). The box indicates the region of focal cell death. g–l IBA1⁺ cells were quantified in miR-155 inhibitor– and negative control–injected retinas which showed (g) no significant difference in total number of IBA1⁺ cells between groups. h–i Representative images of IBA1⁺ cells infiltrating the ONL. j Within the outer retina, miR-155 inhibitor–injected retinas showed a significantly greater percentage of ramified IBA1⁺ cells (p < 0.05). k–m Significantly, more IBA1⁺ cells were found within both the IPL and OPL of the retina in miR-155 inhibitor–injected mice compared with that in negative controls (p < 0.05). l–m Representative images show IBA1⁺ cells in the IPL and OPL of the peripheral retina. n Gene expression of Cfh, C3, Il-10, Tnfα and Socs1, and predicted targets of miR-155 (Bdnf, Il6st and Antxr2) were analysed by qRT-PCR. No significant difference was seen in the gene expression of Cfh, C3 Socs1, Bdnf, Il6st or Antxr2 (p > 0.05); however, the expression of Il-10 and Tnfα was significantly decreased in miR-155 inhibitor–injected retinas (p < 0.05). Statistical significance was determined by Student’s t test and two-way ANOVA with post hoc multiple comparison (n = 10 animals per group, *represents p < 0.05). ONL, outer nuclear layer; INL, inner nuclear layer. For all images, scale bars represent 20 μm (Hi, Ii: scale bar represents 10 μm)

Fig. 2

DR miR-155 constitutive knockout (KO) retinas demonstrated no difference compared with that of WT controls. a–d Representative images of (a–b) H&E-stained whole retinas (c–d) and a region of superior retina approximately 600 μm from the optic nerve. e Thickness of retinal layers represented as a ratio of each layer to total retinal thickness indicated no significant difference in thickness of retinal layers between WT and miR-155 KO DR mice. Furthermore, no difference was observed in (f) photoreceptor row counts in the superior retina of control and miR-155 KO mice (p > 0.05). g The number of IBA1⁺ cells in the IPL, OPL and the ONL/SRS was quantified in control and miR-155 KO retinas and showed no significant difference between groups (p > 0.05). h–i ERG recordings of DR miR-155 KO and control mice recorded over a range of flash intensities showed no significant difference between groups for either the (h) a-wave or the (i) b-wave response (p > 0.05). Statistical significance was determined by Student’s t test and two-way ANOVA with post hoc multiple comparison (n = 6 animals per group, *represents p < 0.05). GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; IS, photoreceptor inner segments; OS, photoreceptor outer segments; SRS, sub-retinal space. For all images, scale bars represent 20 μm

Fig. 3

miR-155 KO mouse retinas demonstrated some protection against photo-oxidative damage (PD). a–b Retinal function of miR-155 KO and WT retinas at 5 days PD (a) demonstrated a significantly higher a-wave amplitude in the miR-155 KO (b) and no significant difference in b-wave amplitude between the miR-155 KO and control (p < 0.05). c–e Quantification of photoreceptor cell death by TUNEL immunolabelling of retinal cryosections showed no significant difference in TUNEL⁺ cells between groups. f ONL thickness across the length of the retina was significantly thicker in miR-155 KO mice compared with that in WT controls (p < 0.05). The box indicates the region of focal cell death. g IBA1⁺ cells were quantified and compared between miR-155 KO and control mice showing no significant difference in total counts at 5 days PD. h–i Representative images of IBA1⁺ cells in the ONL. j–m There was no difference in the percentage of ramified cells in (j) the outer retina or (l–m) plexiform layers between miR-155 KO and WT mice (p > 0.05). n The gene expression of complement (Cfh, C3), general cytokines (Il-10, Tnfα and Socs1) and predicted targets of miR-155 (Bdnf, Il6st and Antxr2) was measured by qRT-PCR. There was no change in the expression of Cfh, Tnfα, Il-10 and Socs1;however, there was a significant decrease in the downstream complement component C3 (p < 0.05). Additionally, Bdnf, Il6st and Antxr2 showed a significant increase in expression in miR-155 KOs compared with that in controls (p < 0.05). Statistical significance was determined by Student’s t test and two-way ANOVA with post hoc multiple comparison (n = 10 animals per group, *represents p < 0.05). ONL, outer nuclear layer; INL, inner nuclear layer; IPL, inner plexiform layer, OPL; outer plexiform layer. For all images, scale bars represent 20 μm (Hi, Ii: scale bar represents 10 μm)

Fig. 4

miR-155 expression was predominantly detected in glia of DR WT retinas. a–g Retinal cell suspensions were stained for lineage markers using fluorophore-conjugated antibodies and miR-155 expression using a PrimeFlow assay. a–c Single, DAPI-positive retinal cells were gated on by flow cytometry. d–g Cell populations were identified using lineage markers, rhodopsin (photoreceptors), vimentin (glia), CD11b (microglia) and Thy1.2 (CD90.2; ganglion cells). d Flow cytometric analysis revealed approximately 6% of Rhodopsin⁺ cells to be positively stained for miR-155. ImageStream visualisation shows miR-155 staining to be primarily located within the cytoplasm of Rhodopsin-miR-155 double-positive cells (arrow). e Approximately 62% of vimentin⁺, CD11b⁻ cells were positive for miR-155. The brightest miR-155 staining occurred in vimentin⁺ cells. f Approximately 70% of CD11b⁺ cells were miR-155-positive. g Approximately 35% of Thy1.2⁺ cells showed miR-155 expression. miR-155 fluorescence was detected in the nucleus and cytoplasm of miR-155 positive, vimentin⁺, CD11b⁺ and Thy1.2⁺ cells. n = 3 samples per group with 4 pooled retinas per sample. For all images, scale bars represent 7 μm

Fig. 5

miR-155 in primary and immortalised retinal cells following stimulation. a–d Flow cytometric analysis of miR-155 expression demonstrated no difference in (a) rhodopsin⁺ cells (b) vimentin⁺ cells or (c) thy1.2⁺ cells before and after PD. (d) There was an increase in miR-155 mean fluorescence intensity (MFI) in CD11b⁺ cells following PD. e–f Representative images of miR-155 labelling by AF-647 in (e) DR and (f) 5-day PD CD11b⁺ cells. g–i Immortalised cell lines, 661W (photoreceptors), MIO-M1 (Müller glia) and C8-B4 (brain-derived microglia) were stimulated to mimic a degenerative response. g There was no change in miR-155 expression in 661W photoreceptor-like cells following PD compared with dim controls (p > 0.05); however, a significantly greater expression of miR-155 was seen in both (h) MIO-M1 and (i) C8-B4 cultures following stimulation with a cytokine cocktail compared with unstimulated controls (p < 0.05). j Quantification of miR-155 in rhodopsin⁺ photoreceptors showed more cells with no detectable cytoplasmic miR-155 in DR compared with PD and consequently, more rhodopsin⁺ cells with a higher miR-155 spot count following PD. k Representative image indicating miR-155 located in the cytoplasmic space of rhodopsin⁺ photoreceptors. l miR-155 expression, measured by qPT-PCR, was significantly higher in s-mEV isolated from PD retinas compared with that in DR controls (p < 0.05). m MiR-155 expression was also significantly increased in the serum of mice following PD (p < 0.05). Statistical significance was determined by Student’s t test (n = 3–4 samples per group with 4 pooled retinas per sample, *represents p < 0.05). For all images, scale bars represent 7 μm