Single-cell resolution imaging of retinal ganglion cell apoptosis in vivo using a cell-penetrating caspase-activatable peptide probe

Abstract

Peptide probes for imaging retinal ganglion cell (RGC) apoptosis consist of a cell-penetrating peptide targeting moiety and a fluorophore-quencher pair flanking an effector caspase consensus sequence. Using ex vivo fluorescence imaging, we previously validated the capacity of these probes to identify apoptotic RGCs in cell culture and in an in vivo rat model of N-methyl- D-aspartate (NMDA)-induced neurotoxicity. Herein, using TcapQ488, a new probe designed and synthesized for compatibility with clinically-relevant imaging instruments, and real time imaging of a live rat RGC degeneration model, we fully characterized time- and dose-dependent probe activation, signal-to-noise ratios, and probe safety profiles in vivo. Adult rats received intravitreal injections of four NMDA concentrations followed by varying TcapQ488 doses. Fluorescence fundus imaging was performed sequentially in vivo using a confocal scanning laser ophthalmoscope and individual RGCs displaying activated probe were counted and analyzed. Rats also underwent electroretinography following intravitreal injection of probe. In vivo fluorescence fundus imaging revealed distinct single-cell probe activation as an indicator of RGC apoptosis induced by intravitreal NMDA injection that corresponded to the identical cells observed in retinal flat mounts of the same eye. Peak activation of probe in vivo was detected 12 hours post probe injection. Detectable fluorescent RGCs increased with increasing NMDA concentration; sensitivity of detection generally increased with increasing TcapQ488 dose until saturating at 0.387 nmol. Electroretinography following intravitreal injections of TcapQ488 showed no significant difference compared with control injections. We optimized the signal-to-noise ratio of a caspase-activatable cell penetrating peptide probe for quantitative non-invasive detection of RGC apoptosis in vivo. Full characterization of probe performance in this setting creates an important in vivo imaging standard for functional evaluation of future probe analogues and provides a basis for extending this strategy into glaucoma-specific animal models.

Figure 1. In vivo confocal scanning laser ophthalmoscopic imaging of single retinal ganglion cell (RGC) apoptosis.

(A) Infrared reflection fundus image at 25 hours post-probe injection from a rat eye pretreated with 25 mM NMDA and followed 24 hrs later by injection of 0.387 nmol TcapQ488. The image was focused on the Nerve Fiber Layer using a 55°FOV lens. (B) Fluorescein angiograph (FA) mode fundus image from the same eye at the same time point and focal plane showing distinct fluorescent signals from probe activation. Both images were recorded as an average of 100 aligned frames to obtain a single low-noise, high contrast image. Scale bar in both images, 200 µm.

Figure 2. Correspondence of an in vivo fluorescent fundus image with an ex vivo retinal flat mount.

(A) Fluorescent fundus image obtained in vivo using the CSLO (28 hours post-probe injection) from a rat eye pretreated with NMDA followed by TcapQ488. Strong, punctate fluorescent signals were detected in the retina ganglion cell (RGC) layer. (B) Higher magnification of the boxed area in A in which prominent fluorescent signals are highlighted. (C) Ex vivo flat mount of the same retina showed excellent correspondence with in vivo images in A and B, indicating that real time images reflect single cell resolution of probe activation. Scale bar: A, 200 µm; C, 100 µm.

Figure 3. Representative kinetics of TcapQ488 activation in vivo.

In vivo images were taken in a rat eye pretreated with 12.5 mM NMDA immediately before (A) and at 4 hours (B), 12 hours (C), 25 hours (D), 48 hours (E) and 72 hours (F) post intravitreal injection of 0.313 nmol TcapQ488. Evidence of initial probe activation was noted at 4 hours after TcapQ488 injection (B). Scale bar, 200 µm in all images.

Figure 4. Quantitative plot of probe activation from live images.

(A) Plot of probe activation from images in Figure 3 B–F. (B) Probe activation normalized to maximum counts from an individual eye at each time point was averaged across all probe dose/NMDA concentration combinations. Probe activation increased significantly in the first 12 hours and generally decreased slowly thereafter. Data represent mean ± SEM.

Figure 5. In vivo probe activation as a function of NMDA concentration.

Rat eyes were pretreated with various NMDA concentrations (2.5, 12.5, 25, 40 mM) followed by 0.313 nmol TcapQ488. A control injection consisting of PBS was also performed to determine background probe activation levels in the absence of NMDA. All eyes were imaged at 4, 12, 24, 48 and 72 hours post-probe injection. Probe activation increased with increasing NMDA concentration at all time points examined. n≥3 (5 to 8 eyes) for each time point. Data represent mean ± SEM.

Figure 6. In vivo probe signal as a function of TcapQ488 dose.

Rat eyes were pretreated with 25(0.097, 0.193, 0.313, 0.387 and 0.775 nmol). All eyes were imaged at 4, 12, 24, 48 and 72 hours post-probe injection. Probe activation increased with increasing TcapQ488 doses at all but the earliest time point (4 hours post probe injection). n≥3 (5 to 8 eyes) for each time point. Data represent mean ± SEM.

Figure 7. Probe activation in vivo as a function of both NMDA concentration and probe dose.

Retinal ganglion cell apoptosis was induced in rats by intravitreal injection of NMDA at various concentrations (0 (PBS only), 2.5, 12.5, 25 and 40 mM), followed by intravitreal injection of TcapQ488 (at 0 (PBS only), 0.097, 0.193, 0.313, 0.387 and 0.775 nmol) for a total of 27 NMDA-probe-dose combinations. Data reflect imaging at 12 hours post-probe injection. Probe signal increased as both NMDA concentration and probe dose increased. The plateau in TcapQ488 activation above 0.313 nmol indicates saturation of the dynamic range of probe dosing. n≥3 (5 to 8 eyes) at each combination. Data represent mean ± SEM.

Figure 8. Signal-to-noise ratio across NMDA concentrations and TcapQ488 probe doses.

TcapQ488 activation data from Figure 7 was normalized to PBS pre-treatment (i.e., NMDA = 0) for each probe dose to reflect relative “signal-to-noise.” For lower NMDA concentrations (2.5 and 12.5 mM), the ratio was highest at 0.097 nmol probe, while for higher NMDA concentrations (25 and 40 mM), the ratio was similar from probe doses 0.097 to 0.313 nmol. Data represent mean normalized labeled cells ± Error Propagation.

Figure 9. Dark-adapted B-wave amplitudes as a function of probe dose and time post-treatment.

Treatment indicates PBS or probe only injections. (A–C) Comparison among pre-treatment, 1 week post-treatment, and 2 months post-treatment time points for each probe condition (PBS, 0.193 nmol TcapQ488, 0.387 nmol TcapQ488). (D–F) Comparison of each probe condition at each time point (pre-treatment, 1 week and 2 months post-treatment). There were no significant differences on dark-adapted B-wave amplitudes among probe conditions when tested at each time point. Data represent mean ± SD.