Inducible p27 inactivation results in enhanced intermediate filament GFAP expression, glial cell-cycle entry, and glial nuclear migration. (A and B) Immunostaining for p27 and the marker of glial reactivity GFAP. Micrographs of tamoxifen control (p27^L-/L+) and experimental (p27^L-/L-) retinas six weeks after the start of tamoxifen injections. Asterisks indicate blood vessels. (C and D) Quantitation of Müller glia becoming reactive (GFAP⁺) and entering cell cycle (PCNA⁺) as well as the total number of Müller glial nuclei (SOX9⁺) and stalks (GLUL⁺). Curves were generated using the Smooth marked scatter option in Excel™. (E-F) Bromodeoxyuridine (BrdU) injections followed by a six week chase to reveal the fate of dividing cells. (E-F insets) Higher magnification showing colocalization with glial marker GLUL. Data are expressed as the mean ± SD. Abbreviations: PCNA, Proliferating cell nuclear antigen.

Normal electrophysiology in retinas with reactive p27^L-/L- Müller glia. Electroretinographic analysis was performed in experimental p27^L-/L- and control mice during the peak of GFAP upregulation and during chronic GFAP expression (6 and 16 weeks post-inactivation). During these time periods the Müller glia nuclei migrated from their normal position (mid-inner retinal layers) toward the photoreceptor layer. Four to five mice were analyzed for each condition. (A and B) Representative scotopic and photopic electroretinogram (ERG) responses evoked by 10 decibel (dB) light flashes were recorded after dark adaptation and light adaptation, respectively. The fast components of the a- and b-waves reflect the functional integrity of photoreceptor and bipolar cell signaling, respectively. The slow components of the b-wave reflect Müller glial conductance. (C and D) Wave amplitudes were averaged from 4-5 mice per condition and shown as function of stimulating light intensity. Data are expressed as the mean ± SD.

Summary and model. (A) Comparison of retinal glial and neuronal responses to p27 inactivation. Data values were standardized for cross comparison. (B) Model of p27 inactivation effect on glia and neuronal function. Conditional p27 deficiency results in cell cycle entry and soma displacement, as well as in enhanced intermediate filament expression (e.g., GFAP and VIM). Surprisingly, these reactive changes are compatible with homeostatic neuronal metabolism, electrophysiology and function. The benign impact of inducible proliferative reactive gliosis in p27^L-/L- mice raise the hypothesis that the transition from supportive to deleterious is determined by external factors signaled from either neuronal cell stress, microglial reactivity, inflammatory response, blood barrier damage, or neuronal cell death.

Inducible model of p27^Kip1 inactivation in adult mice. (A) The p27 coding region was conditionally targeted by generating mice harboring LoxP sites at the p27 locus (p27^L+) and expressing a tamoxifen-regulated Cre recombinase under the control of the chimeric chicken beta-actin promoter, CAGG::CreER™. (B) Experimental design: Tamoxifen was administered by single daily intraperitoneal injections (i.p.) for 5 consecutive days in 2 month-old mice. For proliferation analysis, BrdU was injected once daily for three consecutive days starting 8 days after the first tamoxifen injection. All timepoints were calibrated to the start of tamoxifen exposure. (C) Western blots for p27 and the marker of glial reactivity GFAP. Each lane contains 15 μg of protein from individual retinas. The blot was cut at ~35 kDa: portions with the lower and higher molecular weight bands were stained for p27 and GFAP, respectively. (D and E) Immunostaining against p27 and the Müller glial marker Glutamine Synthetase (GLUL). Micrographs of control (p27^L-/L+) and experimental animals (p27^L-/L-) six weeks after the start of tamoxifen injections. Asterisks indicate blood vessels. Arrowheads denote Müller glia that failed to undergo p27 inactivation. (F) Inactivation efficiency was measured by comparing the number of p27⁺ cells colocalizing with the glial nuclear marker SOX9. Data are expressed as the mean ± SD. Abbreviations: ONL, outer nuclear layer; INL, inner nuclear layer; RGC, retinal ganglion cell layer; p27, cyclin-dependent kinase inhibitor CDKN1B; GFAP, glial fibrillary acidic protein; GLUL, glutamine synthetase; SOX9, SRY-box containing gene 9; and DAPI, 4', 6-diamidino-2-phenylindole.

Reactive p27^L-/L- Müller glia maintain neuronal metabolic support. (A) Schematic of neuron-glia metabolic interaction via the glutamate-glutamine cycle. (B-E) Differential metabolite content of cells in p27^L-/L- retina visualized and quantified using computational molecular phenotyping (CMP). Plasticized tissue was serially sectioned at 200 nm, probed with specific anti-hapten IgGs, and visualized with silver-intensification of 1.4 nm gold granules coupled to species-specific secondary IgGs. Each metabolic map was captured as a high-resolution, monochrome image. (F-G) Metabolic mapping of tamoxifen control and experimental retina at the peak of GFAP expression (6 weeks after tamoxifen injections). Red-green-blue channels represents taurine-glutamate-glutamine mapping highlighted with GLUL as an alpha-channel. RGB images with alpha channels were prepared with Adobe Photoshop CS3. The yellow background is the distinctive taurine-glutamine signature of Müller glia. The pink and red compartments in the photoreceptor outer segments (OS) and inner segments (IS) contain distinct taurine-glutamate-glutamine mixtures. While various blue-to-azure cells in the interneuron layer (INL) and ganglion cell layer (GCL) are neurons with distinctive glutamate-glutamine mixtures. (H) Metabolite pixel value in Müller glia was extracted using the GLUL signal. Data are expressed as the mean ± SD. There was no statistically significant difference between tamoxifen control animals (n = 3) and experimental p27^L-/L- animals (n = 3). Asterisks, arrows and arrowheads indicate blood vessels, endfeet and somas, respectively. Additional abbreviations: CRALBP, cellular retinaldehyde-binding protein; GSH, glutathione.

Normal visual acuity in p27^L-/L- mice using the optomotor response. (A) Known visual pathways interrogated by the optomotor response test. (B) Optomotor tracking response to moving gratings was measured using a virtual optomotor system. Spatial frequency thresholds that elicited an optomotor response between tamoxifen control and experimental animals during the peaks of Müller glial proliferation and GFAP upregulation, and during chronic GFAP expression (1.5, 6 and 16 weeks post-inactivation, respectively). A sine wave grating was drawn on a virtual cylinder projected on monitors and the cylinder was rotated to test each. The value for each animal represents the average of both eyes. Data are expressed as the mean ± SD.