Injury-independent induction of reactive gliosis in retina by loss of function of the LIM homeodomain transcription factor Lhx2 - PubMed (original) (raw)

Injury-independent induction of reactive gliosis in retina by loss of function of the LIM homeodomain transcription factor Lhx2

Jimmy de Melo et al. Proc Natl Acad Sci U S A. 2012.

Abstract

Müller glia are the primary glial subtype in the retina and perform a wide range of physiological tasks in support of retinal function, but little is known about the transcriptional network that maintains these cells in their differentiated state. We report that selective deletion of the LIM homeodomain transcription factor Lhx2 from mature Müller glia leads to the induction of reactive retinal gliosis in the absence of injury. Furthermore, Lhx2 expression is also down-regulated in Prph2(Rd2/Rd2) animals immediately before the onset of reactive gliosis. Analysis of conditional Lhx2 knockouts showed that gliosis was hypertrophic but not proliferative. Aging of experimental animals demonstrated that constitutive reactive gliosis induced by deletion of Lhx2 reduced rates of ongoing apoptosis and compromised both rod and cone photoreceptor function. Additionally, these animals showed a dramatically reduced ability to induce expression of secreted neuroprotective factors and displayed enhanced rates of apoptosis in light-damage assays. We provide in vivo evidence that Lhx2 actively maintains mature Müller glia in a nonreactive state, with loss of function initiating a specific program of nonproliferative hypertrophic gliosis.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.

Fig. 1.

Expression of Lhx2 in the adult (7-wk-old) mouse retina. (A_–_C) Lhx2 is expressed in the medial INL in the adult retina. A small subpopulation of Lhx2-positive cells is localized to the inner INL (C, arrows). (D_–_F) Cells in the inner INL coexpressed the pan-amacrine cell marker syntaxin (D, arrows). (G_–_O) The overwhelming majority of Lhx2-positive cells in the medial INL coexpressed the Müller glia markers Cralbp, P27Kip1, and GS. (P_–_R) These cells did not express PKCα, a marker of bipolar interneurons. (Scale bars: 50 μm.)

Fig. 2.

Fig. 2.

GLAST-CreERT2 specifically and efficiently excises Lhx2 in Müller glia after induction with 4-OHT. (A) Schema for generation of Lhx2 ΔcKO mice. (B_–_D) Prospective labeling with ROSA26stopYFP;GLAST-CreERT2 confirms the Müller glial specificity of the GLAST-CreERT2 transgene. (E_–_J) Lhx2 expression is eliminated in GLAST-CreERT2;Lhx2lox/lox (Lhx2 ΔcKO) but not GLAST-CreERT2;Lhx2lox/+ (Lhx2 ΔcKO/+) animals. A small population of cells continues to express Lhx2 (J, arrows). (K) Lhx2 ΔKO animals display a significant reduction in Lhx2 expression at P26 and P55 (79%, P < 0.005, n = 3; 83%, P < 0.005, n = 3, respectively). There was no significant reduction of cells expressing the Müller glial marker P27Kip1. (L) The majority of remaining Lhx2-positive cells are Müller glia (P26: 89.3 ± 3.5% P27Kip1-positive, 93.1 ± 1.2% Cralbp-positive; P55: 91.8 ± 1.7% P27Kip1-positive, 92.1 ± 0.9% Cralbp-positive; all n = 3). (M_–_P) Expression of the Müller glial markers P27Kip1 and Cralbp are not qualitatively altered in Lhx2 ΔcKO animals. (Scale bars: 50 μm.)

Fig. 3.

Fig. 3.

Lhx2 ΔcKO mice display progressive and ubiquitous hypertrophic reactive gliosis of Müller glia. (A_–_H) Expression of the neurofilament elements Gfap and vimentin are up-regulated in Lhx2 ΔcKO mice by P55. (A and B) Vimentin expression is unchanged at P26 immediately after 4-OHT treatment. (C and D) By P55, expression of vimentin is up-regulated in Lhx2 ΔcKO mice. (E and F) Gfap expression is unchanged at P26 immediately after 4-OHT treatment. (G and H) By P55, expression of Gfap is up-regulated in Lhx2 ΔcKO mice. (I_–_L) Expression of GS is down-regulated in Lhx2 ΔcKO mice at P26 and P55. (M and N) Lhx2 ΔcKO Müller glia display nuclear hypertrophy. (O) Nuclei of Lhx2 ΔcKO cells are 43% larger than Lhx2-positive nuclei (66.98 μm2 vs. 37.94889 μm2, P < 0.0001, _n_ = 27). (_P_–_S_) Generation of mosaic retinas of Lhx2-positive and Lhx2-negative Müller glia demonstrates reduced reactive gliosis associated with Lhx2-positive Müller glia clusters (_P_, between white dashed lines). (_T_–_Y_) Further mosaics were generated by electroporation of Cre recombinase (pCAG-Cre) expression constructs into P0.5 Lhx2lox/lox and Lhx2lox/+ mice. By P14, induction of reactive gliosis was specifically restricted to the electroporated regions in Lhx2lox/lox (_T_–_V_; _T_, dashed line) but not Lhx2lox/+ (_W_–_Y_) mice. (_Z_) Down-regulation of Lhx2 at P14 precedes induction of reactive gliosis in Prph2Rd2/Rd2 mice. (_Z_′) Relative density of Lhx2 at P14, P21, and P35 normalized to Gapdh demonstrates significant 52% reduction of Lhx2 at P14 (_P_ < 0.02, _n_ = 3). There is no significant difference in Lhx2 relative density at P21 or P35 (_P_ > 0.05, n = 3). (Scale bars: A_–_L, P, T, and W, 50 μm; M and N, 10 μm.)

Fig. 4.

Fig. 4.

Lhx2 ΔcKO Müller glia display hallmarks of reactive gliosis at 1 y after 4-OHT treatment. (A_–_E) Lhx2 ΔcKO retinas display ubiquitous up-regulation of Gfap and β-catenin expression in Müller glia at 1 y after 4-OHT treatment. (C_–_E) Outer limiting membrane breaks are evident in these animals because of Müller glia hypertrophy (arrows). (F) Retinas with reactive gliosis generated by Lhx2 ΔcKO feature significantly fewer TUNEL-positive apoptotic cells than age matched controls do at P55 (P < 0.05, n = 3), 3 mo (P < 0.05, n = 3), and 1 y (P < 0.05, n = 3) after 4-OHT treatment. (Scale bars: 50 μm.)

Fig. 5.

Fig. 5.

Lhx2 ΔcKO mice demonstrate a decreased capacity to protect photoreceptors in light-damage assays. (A) Plot of Lhx2 ΔcKO and Lhx2 ΔcKO/+ retinal and ONL thickness after 6,000-lx light exposure. (B) Statistical comparison of overall, dorsal, and ventral retinal and ONL thickness by using generalized estimating equations (G.E.E.). Plots display the G.E.E. coefficient and 95% confidence intervals. (A and B) Overall retinal thickness was not significantly different between Lhx2 ΔcKO and Lhx2 ΔcKO/+ mice (P > 0.05, n = 3) because of the occasional formation of ventral dysplasia in the inner retina of Lhx2 ΔcKO mice. Retinal thickness was significantly different in the dorsal retina (P < 0.005, n = 3). Overall, dorsal, and ventral ONL thickness was significantly reduced in the Lhx2 ΔcKO mice (P < 0.005, n = 3 each of overall, dorsal, and ventral). (C and D) Lhx2 ΔcKO and Lhx2 ΔcKO/+ retinas do not display cell proliferation after light damage. (E_–_P) Lhx2 ΔcKO retinas feature increased TUNEL staining after light exposure in both the dorsal (E_–_G and K_–_M) and ventral (H_–_J and N_–_P) retina compared with Lhx2 ΔcKO/+ mice. (Q and R) Relative target gene expression in real-time PCR analysis of P55 Lhx2 ΔcKO vs. Lhx2 ΔcKO/+ mice (Q) and P55 Lhx2 ΔcKO vs. Lhx2 ΔcKO/+ mice subjected to light damage (R). For all targets, Gapdh was used as an endogenous control gene. (Scale bars: 50 μm.)

Similar articles

Cited by

References

    1. Bringmann A, et al. Müller cells in the healthy and diseased retina. Prog Retin Eye Res. 2006;25:397–424. - PubMed
    1. Bringmann A, et al. Role of retinal glial cells in neurotransmitter uptake and metabolism. Neurochem Int. 2009;54(3-4):143–160. - PubMed
    1. Parker R, Wang JS, Kefalov VJ, Crouch RK. Interphotoreceptor retinoid-binding protein as the physiologically relevant carrier of 11-cis-retinol in the cone visual cycle. J Neurosci. 2011;31:4714–4719. - PMC - PubMed
    1. Wang JS, Kefalov VJ. An alternative pathway mediates the mouse and human cone visual cycle. Curr Biol. 2009;19:1665–1669. - PMC - PubMed
    1. Furukawa T, Mukherjee S, Bao ZZ, Morrow EM, Cepko CL. rax, Hes1, and notch1 promote the formation of Müller glia by postnatal retinal progenitor cells. Neuron. 2000;26:383–394. - PubMed

Publication types

MeSH terms

Substances

LinkOut - more resources