Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 34 (10), 3793-806

Macroglia-microglia Interactions via TSPO Signaling Regulates Microglial Activation in the Mouse Retina

Affiliations

Macroglia-microglia Interactions via TSPO Signaling Regulates Microglial Activation in the Mouse Retina

Minhua Wang et al. J Neurosci.

Abstract

Chronic retinal inflammation in the form of activated microglia and macrophages are implicated in the etiology of neurodegenerative diseases of the retina, including age-related macular degeneration, diabetic retinopathy, and glaucoma. However, molecular biomarkers and targeted therapies for immune cell activation in these disorders are currently lacking. To address this, we investigated the involvement and role of translocator protein (TSPO), a biomarker of microglial and astrocyte gliosis in brain degeneration, in the context of retinal inflammation. Here, we find that TSPO is acutely and specifically upregulated in retinal microglia in separate mouse models of retinal inflammation and injury. Concomitantly, its endogenous ligand, diazepam-binding inhibitor (DBI), is upregulated in the macroglia of the mouse retina such as astrocytes and Müller cells. In addition, we discover that TSPO-mediated signaling in microglia via DBI-derived ligands negatively regulates features of microglial activation, including reactive oxygen species production, TNF-α expression and secretion, and microglial proliferation. The inducibility and effects of DBI-TSPO signaling in the retina reveal a mechanism of coordinated macroglia-microglia interactions, the function of which is to limit the magnitude of inflammatory responses after their initiation, facilitating a return to baseline quiescence. Our results indicate that TSPO is a promising molecular marker for imaging inflammatory cell activation in the retina and highlight DBI-TSPO signaling as a potential target for immodulatory therapies.

Keywords: DBI; Müller cells; TSPO; gliosis; microglia; retina.

Figures

Figure 1.
Figure 1.
Expression of TSPO in the developing and adult mouse retina. A, Immunohistochemical analyses in retinal sections showed TSPO expression in microglia and inner retinal blood vessels. In the developing P0 retina of CX3CR1GFP/+ mice, TSPO immunopositivity was found exclusively in rounded, deramified microglia marked by GFP expression (arrowheads). In the P3 and P7 retina, GFP+ microglia, which migrate into the outer retina, maintained immunopositivity for TSPO. In the P14 retina, retinal microglia acquired ramified morphologies and decreased in TSPO immunopositivity (arrowheads), whereas inner retinal vessels became immunopositive (arrow). In the P28 adult retina, TSPO immunopositivity was absent in microglia and was found only in retinal vessels. B, qRT-PCR analyses showed varying levels of TSPO mRNA expression in the retina of C57BL6 mice during postnatal development. Expression levels, normalized to that in P0 retina, demonstrated a relative downregulation in the first postnatal week and then increased slightly in the young (P28) adult. C, Protein expression levels of TSPO in the retina, as demonstrated by quantitative Western blot analyses, were low at P0 and then gradually increased. *Comparisons with P0 of p < 0.05, 1-way ANOVA with Dunnett's multiple-comparisons test, n ≥ 4 animals per group. Scale bar indicates 50 μm.
Figure 2.
Figure 2.
Expression of DBI in the developing and adult mouse retina. A, Immunohistochemical analyses in the retinal sections from developing wild-type C57BL6 mice showed immunopositivity for DBI in the astrocytic layer of the inner retina at P0. At P3, P7, and P14, DBI immunopositivity was found additionally in vertically oriented processes extending through the thickness of the retina and in somata in the inner nuclear layer. B, Immunopositivity for DBI (red) in the adult (P28) retina of CX3CR1GFP/+ mice colocalized with the Müller cell marker, glutamine synthetase (GS, green, middle), but not with GFP-labeled microglia (green, right). C, Immunohistochemical analyses in retinal flat mounts showed that GFAP-expressing retinal astrocytes (blue) in the inner retina were weakly positive for DBI (red), whereas GS+ Müller cell end-foot processes (green) were strongly positive. D, qRT-PCR analyses showing relative levels of DBI mRNA expression in the retina of postnatal and young adult wild-type C57BL6 mice. E, Protein expression levels of DBI as demonstrated by quantitative Western blot analyses showed an increasing trend from P0 to adulthood. *Comparisons with P0 of p < 0.05, 1-way ANOVA with Dunnett's multiple-comparisons test, n ≥ 4 animals per group. Scale bar indicates 50 μm.
Figure 3.
Figure 3.
TSPO expression is upregulated in microglia after LPS-mediated activation in vitro and in vivo. A, qRT-PCR analysis in retinal microglia demonstrated increased TSPO mRNA expression after LPS treatment. B, Top, Western blot analysis demonstrated increased TSPO protein expression in both retinal microglia and BV2 microglia 24 h after LPS treatment. Bottom, Quantitative Western blot analyses in BV2 microglia confirmed increased TSPO protein expression (graphs in A and B are normalized relative to control). *p < 0.05, unpaired t test with Welch's correction, n = 7–13. C, CD11b+ retinal microglia (green) demonstrated TSPO immunopositivity (red) in a perinuclear distribution 6 h after LPS exposure. Scale bar indicates 50 μm. D, qRT-PCR analyses of retinal TSPO mRNA levels after intravitreal LPS injection (black bars) relative to PBS-injected control eyes (white bars), showed a transient increase in TSPO mRNA 1 d after LPS injection. *p < 0.05, unpaired t test with Welch's correction, n = 3–15 eyes. E, Quantitative Western blot analysis of TSPO protein demonstrated similar dynamics. *p < 0.05, paired t test, n = 4–8 eyes. FI) Retinal flat mounts from adult CX3CR1GFP/+ mice that had been injected intravitreally with LPS or with a PBS control at various time-points. After control (PBS) injections (F), GFP+ retinal microglia, which maintained their “resting” activation state, as evidenced by their ramified morphology (left, green) and the absence of F4/80 immunopositivity (right, blue), remained negative for TSPO (middle, red). One to 3 d after LPS injections (G, H), retinal microglia demonstrated increased activation (appearance of amoeboid morphologies and F4/80 immunopositivity). TSPO immunopositivity (G, H, middle, red) in the retina emerged in GFP+ microglia (green) and amoeboid F4/80+ (blue) activated microglia and macrophages 1 d after LPS injection and increased in intensity at 3 d after injection (inset shows TSPO colocalization with GFP and F4/80 labeling at high magnification). TSPO expression developed in microglia found at various layers of the retina (K). GCL, Ganglion cell layer; IPL, inner plexiform layer; OPL, outer plexiform layer. At day 28 after LPS injection (I), GFP+ microglia reverted to a ramified “resting” morphology, with TSPO immunopositivity decreasing to a weak residual punctate pattern (inset). Scale bar indicates 100 μm. J, K, In eyes injected with intravitreal LPS 3 d prior, immunohistochemical analyses of retinal flat mounts demonstrated that TSPO immunoreactivity colocalized with microglia but not in GFAP+ astrocytes (J), whereas those performed in retinal sections demonstrated that TSPO immunoreactivity was absent in GS+ Müller cells (K). Scale bar indicates 50 μm.
Figure 4.
Figure 4.
DBI expression in the retina is increased in Müller cells and astrocytes after LPS-mediated activation. A, qRT-PCR analyses of DBI mRNA levels in the retina at various time points after intravitreal LPS injection (black bars) relative to PBS injected controls (white bars). DBI mRNA levels were increased 1 d after LPS injection and returned to control levels by 3 d. *Comparison with control for which p < 0.05, Mann–Whitney test, n = 6–10 eyes. Under control conditions, Müller cell somata and processes, as well as retinal astrocytes, were immunopositive for DBI as observed in retinal sections (B) and retinal flat mounts imaged at the plane of the inner limiting membrane (ILM, indicated by arrow in B, D) at the inner surface of the retina, which revealed Müller cell end-foot processes and astrocytes (C). Three days after intravitreal LPS injection, increased DBI immunopositivity was observed in both radial Müller cell processes (D) and astrocytes (E, arrow). The cellular localization of increased DBI expression was confirmed by immunolocalization with GS for Müller cells (F) and GFAP for astrocytes (G). DBI expression was not observed in F4/80+ activated microglia (H). Scale bar indicates 100 μm.
Figure 5.
Figure 5.
TSPO expression is induced in retinal microglia in models of retinal injury. AD, TSPO expression was evaluated in a model of optic nerve crush injury in adult CX3CR1GFP/+ mice. Animals were subjected to optic nerve crush and their retinas harvested and examined in cryosections 7 d after injury. In sham-manipulated control mice (A), the ganglion cell marker Brn3 (white) showed prominent labeling in the ganglion cell layer (GCL) and TSPO immunolabeling (red) was present only in retinal vessels and absent in GFP+ retinal microglia (B, inset, green). In treated animals (C), Brn3+ ganglion cells were depleted, punctate, extravascular TSPO-labeling developed in the inner retina (C, D, red) that localized to the processes and perinuclear areas of retinal microglia (D, inset). E, F, Excitotoxic retinal injury was induced in CX3CR1GFP/+ mice with an intravitreal injection of NMDA (1 μl of 40 mm) and their retinas analyzed in cryosections 1 d later. GFP+ microglia in the inner retina (E, green) demonstrated activated morphologies with shortened processes and amoeboid morphologies. Punctate, extravascular TSPO immunostaining also developed in the inner retina (F, red) that colocalized with GFP+ cells (F, arrows, insets). G, H, Outer retinal injury was induced in a model of subretinal hemorrhage in adult C57BL6 mice with a subretinal injection of autologous blood (1.5 μl) and retinas analyzed in vibratome sections 2 d later. Activated, CD68+ microglia with amoeboid morphologies were found aggregated in the outer retina in the region of subretinal hemorrhage (G, green). Punctate, extravascular TSPO immunostaining (H, red) was observed throughout the retina, colocalizing with CD68+ microglial processes (H, inset 1), and microglia somata (H, inset 2). Scale bar indicates 50 μm.
Figure 6.
Figure 6.
Effects of TTN, an endogenous DBI-derived TSPO ligand, on microglia function in vitro. A, B, The effect of TTN on ROS production in cultured retinal microglia and BV2 microglia was evaluated using the H2DCFDA assay. In cultured microglia (A), ROS production was significantly increased from control conditions (white bar) by LPS treatment (0.5 μg/ml for 6 h, black bar). Pretreatment with TTN at 0.1 and 10 μm for 2 h significantly reduced ROS production in microglia in the absence and presence of LPS. In BV2 microglia (B), TTN pretreatment (10 μm) similarly significantly suppressed ROS production in LPS-treated cells. n = 14–21 replicates per condition. C, Production of mitochondrial superoxide in cultured retinal microglia, as monitored with a MitoSOX indicator. Minimal fluorescence was detected under control conditions but was increased after LPS application. Pretreatment with 10 μm TTN reduced the number of Mito SOX+ cells produced by LPS treatment. D, E, Effect of TTN on TNF-α mRNA expression in cultured retinal microglia (D) and BV2 microglia (E). Although TTN pretreatment did not influence TNF-α mRNA expression in microglia in the absence of LPS treatment (white and unshaded bars), it significantly suppressed it in LPS-treated cells (black and shaded bars). n = 6–12 replicates. F, ELISA analyses demonstrate that although TTN pretreatment did not affect TNF-α secretion in BV2 cells in the absence of LPS treatment (white and unshaded bars), it significantly suppressed the TNF-α secretion in LPS-treated cells (black and shaded bars). n = 8–16 replicates. G, Effect of TTN on BV2 microglia proliferation was measured by the incorporation of EdU. TTN pretreatment (10 μm) significantly suppressed microglial proliferation in both control and LPS-stimulated culture conditions. *Comparison with non-TTN-treated controls for which p < 0.05, Mann–Whitney test, n = 38–50 replicates. Data in A, B, D, E, and F were analyzed with the Kruskal-Wallis one-way ANOVA test with Dunn's multiple-comparisons test. *Comparison with non-TTN-treated controls for which p < 0.05.
Figure 7.
Figure 7.
Effects of the synthetic TSPO ligands PK 11195 and Ro5–4864 on microglia function in vitro. A, B, The effect of synthetic ligands in cultured retinal microglia on ROS production was evaluated using the H2DCFDA assay. Pretreatment with either PK 11195 (A) or Ro5–4864 (B) significantly reduced ROS production in microglia in the absence and presence of LPS. C, D, The effect of synthetic ligand pretreatment on TNF-α mRNA expression was evaluated using qRT-PCR. Decrease in TNF-α expression in the presence of LPS was nonsignificant with PK 11195 (C) but significant with Ro5-4864 (D). Data were analyzed with the Kruskal–Wallis one-way ANOVA test with Dunn's multiple-comparisons test. *Comparison with non-ligand-treated controls for which p < 0.05, n = 4–18 replicates per condition.
Figure 8.
Figure 8.
Effects of knock-down of TSPO on microglia function in vitro. A, Knock-down of TSPO protein expression in BV2 microglia. BV2 microglia were transfected with either an shRNA vector targeting TSPO or a control/scramble vector. In BV2 microglia cultured under normal conditions (white bar) or stimulated with LPS (0.5 μg/ml) for 6 h (black bar), shRNA transfection was able to significantly reduce the protein expression of TSPO by 30–40% relative to control. *Comparison with controls for which p < 0.05, Mann–Whitney test, n = 6–13 replicates for each condition. B, Effect of TSPO knock-down on ROS production was evaluated using the H2DCFDA assay. TSPO knock-down in BV2 cells significantly increased ROS production induced by LPS treatment. *Comparison with control for which p < 0.05, Mann–Whitney test, n = 28–33 replicates per condition. C, TSPO knock-down significantly increased TNF-α mRNA expression in BV2 cells treated with LPS (black bars). *Comparison with control for which p < 0.05, Mann–Whitney test, n = 8 replicates for each condition. D, Effect of TSPO knock-down on TNF-α protein production was evaluated in conditioned media using ELISA and normalized to protein concentration of cell lysates. TSPO knock-down significantly increased TNF-α protein production in BV2 cells treated with LPS (black bars). *p < 0.05, unpaired t test, n = 8–16 replicates for each condition. E, Effect of TSPO knock-down on BV2 microglia proliferation was measured by the incorporation of EdU. Although TSPO knock-down did not increase BV2 microglia proliferation under normal conditions, it significantly increased BV2 microglia proliferation in LPS-treated cells. *p < 0.05, Mann–Whitney test, n = 24–27 replicates per condition. Scale bar indicates 50 μm.
Figure 9.
Figure 9.
Effects of TTN in vivo. A, ROS produced in retina was evaluated by measurement of lipid peroxidation using a TBARS assay. Two days after intravitreal LPS injection (1 mg/ml, 1 μl), significantly increased lipid peroxidation in the retina was observed. Although the intravitreal injection of TTN did not affect the lipid peroxidation alone, it effectively suppressed the increase of lipid peroxidation induced by LPS. *p < 0.05, paired t test, n = 8–12 replicates for each condition. B, Two days after intravitreal LPS injection (1 mg/ml, 1 μl), the protein level of TNF-α significantly increased (black bar). The intravitreal injection of TTN significantly suppressed the increase of LPS-induced TNF-α in the retina. *p < 0.05, paired t test, n = 3–7 replicates for each condition.

Similar articles

See all similar articles

Cited by 61 PubMed Central articles

See all "Cited by" articles

Publication types

LinkOut - more resources

Feedback