CD44 signaling in Müller cells impacts photoreceptor function and survival in healthy and diseased retinas

Abstract

Retinitis pigmentosa (RP), an inherited retinal disease, affects 1,5 million people worldwide. The initial mutation-driven photoreceptor degeneration leads to chronic inflammation, characterized by Müller cell activation and upregulation of CD44. CD44 is a cell surface transmembrane glycoprotein and the primary receptor for hyaluronic acid. It is involved in many pathological processes, but little is known about CD44's retinal functions. CD44 expression is also increased in Müller cells from our Pde6b^STOP/STOP RP mouse model. To gain a more detailed understanding of CD44's role in healthy and diseased retinas, we analyzed Cd44^-/- and Cd44^-/-Pde6b^STOP/STOP mice, respectively. The loss of CD44 led to enhanced photoreceptor degeneration, reduced retinal function, and increased inflammatory response. To understand the underlying mechanism, we performed proteomic analysis on isolated Müller cells from Cd44^-/- and Cd44^-/-Pde6b^STOP/STOP retinas and identified a significant downregulation of glutamate transporter 1 (SLC1A2). This downregulation was accompanied by higher glutamate levels, suggesting impaired glutamate homeostasis. These novel findings indicate that CD44 stimulates glutamate uptake via SLC1A2 in Müller cells, which in turn, supports photoreceptor survival and function.

Keywords: CD44; Gliosis; Glutamate; Inflammation; Müller cells; Retinitis pigmentosa; SLC1A2.

Fig. 1

Progressive photoreceptor degeneration results in upregulation of CD44 expression in Müller cells. (A-D) Retinas from Pde6b^STOP/WT (ST/WT) and Pde6b^STOP/STOP (ST/ST) mice were analyzed at 4 and 8 weeks of age. Expression of CD44 increases with disease progression. (A) Quantitative analysis of Cd44 mRNA by qRT-PCR. N = 3 per group. (B) Representative images of retinal sections immunostained for glutamine synthetase (GLUL) to visualize Müller cells and for CD44. CD44 is mainly expressed in the apical microvilli of Müller cells. Scale bar, 30 μm. (C) Representative CD44 immunoblot of retinal lysates. β-Actin was used as a loading control. (D) Quantitative analysis of CD44 immunoblots. N = 4 per group. (A, D) Data, presented as mean ± SEM, were compared by ANOVA. *** P ≤ .001. (E-G) Retinas from 3 retinal degeneration mouse models were analyzed: Pde6b^H620Q/H620Q (week 4), Rho^P23H/+ (week 9) , and Cnga3^−/−, Rho^−/−, Opn4^−/− triple-knockout mice (week 5). (Upper panel) Representative images of retinal sections immunostained for CD44. Scale bar, 30 μm. (Lower panel) Representative CD44 immunoblots of retinal lysates. β-Actin was used as a loading control. ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer

Fig. 2

Absence of CD44 does not affect GFAP expression levels. Retinas from Pde6b^STOP/WT (ST/WT), Cd44^−/−Pde6b^STOP/WT(CD44^−/− ST/WT), Pde6b^STOP/STOP (ST/ST), and Cd44^−/−Pde6b^STOP/STOP (CD44^−/− ST/ST) mice were analyzed at 8 weeks of age. (A) Quantitative analysis of Cd44 mRNA by qRT-PCR. (B) Representative images of retinal sections immunostained for CD44. Scale bar, 30 μm. (C) Representative CD44 immunoblot of retinal lysates. β-Actin was used as a loading control. (D) Quantitative analysis of Gfap mRNA by qRT-PCR. (A, D) N = 4 per group. Data, presented as mean ± SEM, were compared by ANOVA. * P ≤ .05; *** P ≤ .001. (E) Representative images of retinal sections immunostained for GLUL and GFAP. Scale bar, 25 μm (F) Representative GFAP immunoblot of retinal lysates. β-Actin was used as a loading control. ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer

Fig. 3

Progression of photoreceptor degeneration and inner retinal remodeling in CD44-deficient mice. Retinas from ST/WT, CD44^−/− ST/WT, ST/ST, and CD44^−/− ST/ST mice were analyzed. (A) Representative images of retinal sections from 12-week-old mice immunostained for ARR3 (cones). Scale bar, 25 μm. (B) ONL thickness and cone inner and outer segment length at 3, 4, 8, and 12 weeks of age. pw 3, N = 5 per group. pw 4, N = 6 per group. pw 8, N = 5 for ST/WT, and ST/ST, N = 7 for CD44^−/− ST/WT, N = 6 for CD44^−/− ST/ST. pw 12, N = 8 for ST/WT, N = 6 for CD44^/− ST/WT, and CD44^−/− ST/ST, N = 5 for ST/ST. (C, E, G) Representative images of retinal sections from 12-week-old mice immunostained for PKC–α (rod bipolar cells) (C), SCGN (cone bipolar cells) (E), and calbindin (horizontal cells) (G). Arrowheads indicate shorter and less-branched bipolar dendrites. Scale bars, 25 μm. (D, F, H) Quantification of dendritic area from rod bipolar cells (D), cone bipolar cells (F), and horizontal cells (H). pw 12, N = 6 for ST/WT, CD44^−/− ST/WT, and CD44^−/− ST/ST, N = 7 for ST/ST. Data, presented as mean ± SEM, were compared by ANOVA. * P ≤ .05; ** P ≤ .01; *** P ≤ .001. ONL, outer nuclear layer; INL, inner nuclear layer; IS, inner segment; OS, outer segment; OPL, outer plexiform layer; IPL, inner plexiform layer

Fig. 4

Decreased scotopic and mesopic retinal function in CD44-deficient mice. (A-E) ERGs were recorded from 8-week-old ST/WT, CD44^−/− ST/WT, ST/ST, and CD44^−/− ST/ST mice. (A) Scotopic and mesopic a-wave amplitudes. A-wave amplitudes of CD44^−/− ST/WT and CD44^−/− ST/ST at 0.5 and 1.0 log (cd*s/m²) were significantly decreased compared to ST/WT and ST/ST, respectively. (B) Scotopic and mesopic b-wave amplitudes. At the highest measured light intensity of 1.0 log (cd*s/m²), the b-wave amplitude of CD44^−/− ST/WT mice was significantly decreased compared to ST/WT mice. (C) Representative ERG responses in the scotopic and mesopic range. (D) Photopic b-wave amplitudes. B-wave amplitudes were similar between all groups. (A, B, D) Data, presented as mean ± SEM, were compared by ANOVA, * P < .05; ** P < .01. N = 7 per group. (E) Representative ERG responses in the photopic range

Fig. 5

Proteomic profiling of Müller cells and neurons. MACS enriched Müller cells and neuron cell fractions were subjected to label-free liquid chromatography-tandem mass spectrometry (LC-MS/MS)-based proteomics. ST/WT N = 4, CD44^−/− ST/WT, CD44^−/− ST/ST N = 5, ST/ST N = 6. (A) Representative image of Müller cell fraction immunostained for glutamine synthetase (GLUL) and counterstained with Hoechst 33,342. (B) Representative image of neuron fraction immunostained for recoverin and counterstained with Hoechst 33,342. Scale bar, 50 μm. (C-F) Quantitative analysis of GLUL (Müller cell marker) (C), rhodopsin (RHO) (rod photoreceptor marker) (D), PDE6B (E) and CD44 (F). (G) Volcano plot showing differentially expressed proteins between ST/ST and CD44^−/− ST/ST Müller cells. Red dot, solute carrier family 1 member 2 (SLC1A2) with fold change > 5, and FDR < 0.1. (H) Heat map representation of Müller cell-specific proteins involved in metabolic processes. (I, J) Quantitative analysis of SLC1A2 (I) and MAPK3 (J). Data, presented as mean ± SEM, were compared by ANOVA. * P ≤ .05; ** P ≤ .01; *** P ≤ .001

Fig. 6

Validation of decreased SLC1A2 expression in CD44-deficient mice. (A-D) Retinas from ST/WT, CD44^−/− ST/WT, ST/ST, and CD44^−/− ST/ST were analyzed at 8 weeks of age. (A) Representative SLC1A2 immunoblot of retinal lysates. β-Actin was used as a loading control. (B) Quantitative analysis of SLC1A2 immunoblots. Data are presented as mean ± SEM. N = 5 per group. (C) Representative images of retinal sections immunostained for the glutamate transporter SLC1A2. SLC1A2 is expressed in Müller cells and cones. Scale bar, 35 μm. (D) Fluorometric glutamate assay revealed a significantly higher glutamate concentration in CD44^−/− ST/WT retinas compared to ST/WT. Data, presented as mean ± SEM, were compared by ANOVA, * P ≤ .05; ST/WT, CD44^−/− ST/WT, CD44^−/− ST/ST N = 3, ST/ST N = 4. ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer

Fig. 7

Increased neuroinflammation in CD44^−/−ST/ST mice. (A) Representative images of retinal sections from 12-week-old mice immunostained for Iba1 and CD68 to visualize reactive microglia/ macrophages. Scale bar, 35 μm. (B) Quantification of reactive microglia at 4, 8, and 12 weeks of age. pw 4, N = 5 for ST/WT, N = 6 for CD44^−/− ST/WT, ST/ST, and CD44^−/− ST/ST. pw 8, N = 4 for ST/WT, N = 5 for CD44^−/− ST/WT, ST/ST, and CD44^−/− ST/ST. pw 12, N = 8 for ST/WT, N = 6 for CD44^−/− ST/WT, and CD44^−/− ST/ST, N = 5 for ST/ST. (C-D) Quantitative mRNA analysis of TNFα (C), and NFκB (D) by qRT-PCR at 8 weeks of age. N = 4 per group. (E-F) Quantitative analysis of NFκB2 (E) and IL-16 (F) in isolated Müller cells. ST/WT N = 4, CD44^−/− ST/WT, CD44^−/− ST/ST N = 5, ST/ST N = 6. (B-F) Data, presented as mean ± SEM, were compared by ANOVA. ** P ≤ .01; *** P < .001. (G) Heat map representation of Müller cell-specific proteins involved in immune system processes. ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer