Human NGF "Painless" Ocular Delivery for Retinitis Pigmentosa: An In Vivo Study

Abstract

Retinitis pigmentosa (RP) is a family of genetically heterogeneous diseases still without a cure. Despite the causative genetic mutation typically not expressed in cone photoreceptors, these cells inevitably degenerate following the primary death of rods, causing blindness. The reasons for the "bystander" degeneration of cones are presently unknown but decrement of survival factors, oxidative stress, and inflammation all play a role. Targeting these generalized biological processes represents a strategy to develop mutation-agnostic therapies for saving vision in large populations of RP individuals. A classical method to support neuronal survival is by employing neurotrophic factors, such as NGF. This study uses painless human NGF (hNGFp), a TrkA receptor-biased variant of the native molecule with lower affinity for nociceptors and limited activity as a pain inducer; the molecule has identical neurotrophic power of the native form but a reduced affinity for the p75NTR receptors, known to trigger apoptosis. hNGFp has a recognized activity on brain microglial cells, which are induced to a phenotype switch from a highly activated to a more homeostatic configuration. hNGFp was administered to RP-like mice in vivo with the aim of decreasing retinal inflammation and also providing retinal neuroprotection. However, the ability of this treatment to counteract the bystander degeneration of cones remained limited.

Keywords: blindness; cell survival; inherited retinal degeneration; microglia; neurotrophic factors; photoreceptors.

Figure 1.

rd10 mutant mice retinal phenotype. Vertical sections of the outer retina of a typical C57Bl6 mouse (A), compared with that from a rd10 mutant (B) near the peak of photoreceptor degeneration (50 d of age). The nicely elongated, regularly spaced cones visible in A after cone arrestin antibody staining (in red) have largely degenerated and remodeled into hypertrophic cells with short outer segments (arrows) in B. Nuclei are stained blue with DAPI. onl and inl, outer and inner nuclear layers, respectively.

Figure 2.

Intravitreal delivery of hNGFp has a moderate effect on cone survival in RP mice. A, Experimental design used for the single intravitreal injection. The animals were divided into two groups as for the previous experiments, control (ctr) and hNGFp-treated (treat) and once intravitreal injected with PBS (ctr) or hNGFp (treat) at age P50; hNGFp dosage, 0.54 ng/gr. B, Whole-mount retinas of control (left) and treated mice (right) stained with cone arrestin (red signal). The dotted lines define the degenerated area. C, qPCR quantification of genes involved in photoreceptors survival and microglial activation. Gnat2 (ctr = 8, treat = 9; p = 0.050) and Arr3 (ctr = 5, treat = 6; p = 0.184) are genes expressed in rods and cones, ntrk1 (ctr = 5, treat = 6; val = 0.003) is the gene for trkA expression while ngfr (ctr = 8, treat = 9; p = 0.327) is the one for p75NTR receptor expression. Gfap (ctr = 5, treat = 6; p = 0.749) refers to macroglial activation while tmem119 (ctr = 5, treat = 6; p = 0.007) represent the activation of microglia. Arg1 (ctr = 8, treat = 9; p = 0.552) and nos2 (ctr = 8, treat = 9; p = 0.916) are, respectively, genes expressed by anti-/inflammatory and pro-/inflammatory microglia. Gene expression is reported as the ratio between the ctr and treat animals. Normalization was done on actin expression for each sample. D, Degenerated area quantification expressed as mm2 (ctr = 3, treat = 3; p = 0.153) and percentage (ctr = 3, treat = 3; p = 0.146) with respect to the total retinal area. E, Quantification of cone arrestin signal. Absolute cone number per retina (ctr = 7, treat = 8; p = 0.835); cone density (ctr = 7, treat = 8; p = 0.723); and retinal area (ctr = 7, treat = 8; p = 0.729). Error bars are ±SEM. t test was used to compare the ctr versus treat mean for each experiment.

Figure 3.

Retinal thickness is not rescued by intravitreal delivery of hNGFp in RP mice. A, B, Representative images of the ctr (left) and treated (right) retinal section used to measure the thickness of the ONL. C, D, zoom of retinal sections. E, Statistical analysis of ONL measurements shows no changes in thickness upon hNGFp administration (ctr = 4, treat = 5: p = 0.495). Error bars are ±SEM.

Figure 4.

Intravitreal delivery of hNGFp shows partial immunomodulation of retinal inflammation in RP mice. A, Whole-mount retinas of the control (left) and treated mice (right) stained with CD11b (green signal). The dotted lines define the degenerated area. A and B are zoom images of retinal microglia in the peripheral and central area (B) qPCR quantification of genes involved in inflammatory responses. Myd88 (ctr = 8, treat = 9; p = 0.019) and csf1 (ctr = 5, treat = 6; p = 0.782) are transcriptional factors involved in the activation of pro- and anti-inflammatory response. Lamp2 (ctr = 5, treat = 6; p = 0.0005) is a protein involved in the autophagy process. Nr3c1 (ctr = 5, treat = 6; p = 0.002) codifies for glucocorticoid receptor expression. Ccl12 ctr = 5, treat = 6; p = 0.250), ccl2 (ctr = 5, treat = 6; p = 0.928) and cxcl5 (ctr = 5, treat = 6; p = 0.861) are chemokines and cytokines.

Figure 5.

Increased dosage of hNGFp intravitreal delivered has no beneficial effects on cone survival and retinal inflammation of RP mice. A, Whole-mount retinas of control (left) and treated mice (right) stained with cone arrestin (red signal). The dotted lines define the degenerated area. B, qPCR quantification of genes involved in photoreceptors survival and microglial activation. Gnat2 (ctr = 7, treat = 8; p = 0.042) and Arr3 (ctr = 7, treat = 8; p = 0.486) are genes expressed in rods and cones, ntrk1 (ctr = 7, treat = 8; p = 0.050) is the gene for trkA expression while ngfr (ctr = 7, treat = 8; p = 0.092) is the one for p75NTR receptor expression. Gfap (ctr = 7, treat = 8; p = 0.500) refers to macroglial activation while tmem119 (ctr = 7, treat = 8; p = 0.270) represent the activation of microglia. Arg1 (ctr = 7, treat = 8; p = 0.831) and nos2 (ctr = 7, treat = 8; p = 0.415) are genes expressed by microglial cells. C, Degenerated area quantification expressed as mm2 (ctr = 6, treat = 7; p = 0.698) and percentage (ctr = 6, treat = 7; p = 0.858) with respect to the total retinal area. D, Quantification of cone arrestin–positive cells. Absolute cone number per retina (ctr = 7, treat = 8; p = 0.221); cone density (ctr = 7, treat = 8; p = 0.316); and retinal area (ctr = 7, treat = 8; p = 0.519). Error bars are ±SEM. t test was used to compare the ctr versus treat mean for each experiment.

Figure 6.

wtNGF and hNGFp treatments show no differences in cone survival. Cone counts in retinal preparations from rd10 mice injected intravitreally with wtNGF (n = 4), hNGFp (n = 6), and control solution (n = 6). Error bars are ±SEM. One-way ANOVA test. p = 0.7087.

Figure 7.

Intranasal delivery of hNGFp has no effect on cone rescue in RP mice. A, Graphical representation of the experimental protocol adopted for the treatment. The animals were divided into two groups, control (ctr) and hNGFp treated (treat), and manipulated for 1 week before the start of the experimental protocol. Finally, they were intranasally treated with drops of PBS (ctr) or hNGFp (treat) 3 d/week from age P25 to P45 (left panel) or P60 (right panel). hNGFp dosage, 0.54 ng/gr. B, C, Volcano plots of qPCR array show no changes of gene expression related to photoreceptor survival and the inflammatory pathway. The dots represent each tested gene with its mean fold change and the statistical value (-log10pval). The red and green dots are, respectively, the downregulated and upregulated genes in the treated animals with respect to the control mice. The dotted line represents the significance threshold [-log10(0.05) = 1.33.]. Volcano plot in B, N:(ctr = 7, treat = 7), and in C, N:(ctr = 6, treat = 6). D, Whole-mount image (left side) of a rd10 retina (aged P45) stained with cone arrestin (red signal) showing how the acquisition for cone counting was done. The 16 dotted squares represent the investigated retinal areas for each animal while the white squares in A and B show zoom images of central and peripheral cone arrestin signals. E, Quantification of cone arrestin signal. Absolute cone number per retina (ctr = 4, treat = 4; p = 0.399); cone density (ctr = 4, treat = 4; p = 0.582); and retinal area (ctr = 4, treat = 4; p = 0.781). Each dot represents a different animal. Error bars are ±SEM.

Figure 8.

RPE analysis after intranasal delivery of hNGFp does not demonstrate structural amelioration. The administration failed to elicit visible effects on the morphology of the RPE, here shown upon immunolabeling with antibodies against ZO-1, a protein of the RPE tight junctional complexes. The number of discontinuities in the ZO-1 array (arrows), which is high in rd10 mutants as a secondary effect of photoreceptor loss (A), was not rescued by hNGFp administration (B), as demonstrated by a quantitative analysis shown in C, where the number of intersections between ZO-1 profiles and a reference grid has been assessed systematically and compared statistically (t test; p = 0.62). Error bars are ±SEM. t test was used to compare the ctr versus treat mean for each experiment.

Figure 9.

Biological action of the hNGFp used in this study. A, Basal forebrain coronal sections. ChAT immunofluorescence staining. The number of ChAT-positive profiles is visibly more numerous in the hNGFp-treated section (right), compared with the control. B, Cell counts on multiple sections show an approximate one-third increment in the number of ChAT-positive cells in hNGFp-treated samples (error bars are ±SEM. t test was used to compare the ctr vs treat mean for each experiment; **p ≤ 0.005).