Divergent Effects of HSP70 Overexpression in Photoreceptors During Inherited Retinal Degeneration

Purpose Disruption of proteostasis is a key event in many neurodegenerative diseases. Heat shock proteins (HSPs) participate in multiple functions associated with intracellular transport and proteostasis. We evaluated the effect of augmented HSP70 expression in mutant photoreceptors of mouse retinal degeneration models to test the hypothesis that failure to sustain HSP70 expression contributes to photoreceptor cell death. Methods We examined HSP70 expression in retinas of wild-type and mutant mice by RNA and protein analysis. A transgenic mouse line, TgCrx-Hspa1a-Flag, was generated to express FLAG-tagged full-length HSP70 protein under control of a 2.3 kb mouse Crx promoter. This line was crossed to three distinct retinal degeneration mouse models. Retinal structure and function were evaluated by histology, immunohistochemistry, and electroretinography. Results In seven different mouse models of retinal degeneration, we detected transient elevation of endogenous HSP70 expression at early stages, followed by a dramatic reduction as cell death ensues, suggesting an initial adaptive response to cellular stress. Augmented expression of HSP70 in RHOT17M mice, in which mutant rhodopsin is misfolded, marginally improved photoreceptor survival, whereas elevated HSP70 led to more severe retinal degeneration in rd10 mutants that produce a partially functional PDE6B. In Rpgrip1−/− mice that display a ciliary defect, higher HSP70 had no impact on photoreceptor survival or function. Conclusions HSP70 overexpression has divergent effects in photoreceptors determined, at least in part, by the nature of the mutant protein each model carries. Additional investigations on HSP pathways and associated chaperone networks in photoreceptors are needed before designing therapeutic strategies targeting proteostasis.

Molecular chaperones, especially the heat shock protein (HSP) family, act to support cellular proteostasis by restoring protein function under physiological conditions or by triggering programed cell death under stress. 6,7 The HSP superfamily is classified by apparent molecular mass into subfamilies. Eight HSP70 proteins identified in humans, belonging to the 70-kDa subfamily, share 52% to 99% amino acid sequence homology and exhibit similar functions yet somewhat distinct characteristics. 8 Hereafter we use the term "HSP70" to collectively refer to two proteins, HSP70-1a and HSP70-1b, which only differ from each other by two amino acid residues and are encoded in mice by Hspa1a and Hspa1b genes, respectively. 8,9 HSP70 is one of the most studied chaperones, especially in the stressed retina. In addition to the stress-inducible HSP70, a cognate 70-kDa HSP HSC70 (Hspa8) 10 is constitutively expressed. Under physiological conditions, HSPs protect and improve folding of newly synthesized proteins, participate in quality control, guide their translocation into distinct organelle destinations, and assist in the assembly/disassembly of multiprotein complexes. [11][12][13] When under stress, transcription regulator heat shock factor 1 (HSF1) activates expression of various HSPs, including HSP70. 14 HSP70 binds to misfolded proteins and facilitates their recovery with assistance of other HSPs, such as HSP40 and HSP90, and co-chaperones such as Hip and Hop. 15,16 Alternatively, when the damage is not repairable, HSP70 facilitates the degradation of aggregated proteins by delivering them to the ubiquitin proteasome system with assistance of co-chaperone CHIP. 17,18 The nexus between disruption in cellular proteostasis and progression of neurodegenerative disease has uncovered an important role for chaperone proteins in regulating the survival or death of stressed neuronal cells. [19][20][21] Disease-causing proteins, such as β-amyloid and tau in Alzheimer disease, huntingtin in Huntington disease, αsynuclein in Parkinson disease, and mutant rhodopsin in retinitis pigmentosa, all interact with the HSP chaperone machinery. [20][21][22] Therefore HSPs are potential therapeutic targets for neurodegenerative diseases.
Although HSP70 deficient mice were generated over a decade ago, the role of HSP70 in retinal development has not been investigated. 23 In the adult retina, HSP70 maintains proteostasis and plays an essential role in the maturation of phototransduction proteins. 24,25 In cooperation with HSP90 and a photoreceptor specific co-chaperone AIPL1, HSP70 is implicated in facilitating correct folding and assembly of retinal cGMP phosphodiesterase (PDE6) holoenzyme. [26][27][28] Elevated HSP70 expression was observed in photoreceptors following induced retinal detachment and hypothesized to play a protective role. 29 HSP70 upregulation was also reported to delay N-methyl-N-nitrosourea-induced photoreceptor death. 30 Furthermore, inhibition of HSP90, a negative regulator of HSF1 activation and of HSP70 expression, leads to improved photoreceptor survival and visual function in a rat model with rhodopsin P23H mutation. 31 Exogenous expression of HSP70 and HSF1 are also suggested to significantly improve the survival of injured retinal ganglion cells. 32,33 Along these lines, HSP70 is suggested to play a protective role in the retina, especially in stressed photoreceptors, in response to either exogenous stress stimuli or inherited mutations. [29][30][31][32]34 Here we tested the hypotheses that compromised HSP70 expression precedes photoreceptor death in IRDs and that expression of exogenous HSP70 could extend the survival of mutant photoreceptors. We first assessed the expression of HSP70 transcripts in RNA-seq data from rods and S-cone-like photoreceptors 35 compared with whole retina 36 and then examined HSP70 protein levels in several retinal degeneration mouse models. We augmented HSP70 expression in the photoreceptors of RHOT17M, Rpgrip1 −/− , and rd10 mice by crossing them with a transgenic mouse line, TgCrx-Hspa1a-Flag, that overexpresses HSP70. Our studies demonstrate divergent effects of HSP70 expression in degenerating photoreceptors carrying different mutations and highlight the importance of distinct approaches when targeting cellular proteostasis for therapy.

Generation of TgCrxp-Hspa1a-Flag Mice
A 2.3 kb mouse cone-rod homeobox (Crx) promoter (from −2286 to +72) and the Hspa1a-coding region with an additional Kozak sequence and FLAG-tag were cloned into a modified promoter-less pCl vector (Promega, Madison, WI, USA). The Crx:Hspa1a-Flag fragment was purified and injected into fertilized FVB/N mouse oocytes. Transgenic founder mice and their progeny were identified by PCR. The founders were backcrossed to C57BL/6J mice to establish a stable colony. TgCrxp-Hspa1a-Flag mice are referred to as TgHspa1a+, whereas littermate controls without the transgene are called TgHspa1a-.

Histology of Retinal Cross-Sections
Eyes were enucleated and marked on dorsal side. Eyeballs were fixed in 4% glutaraldehyde followed by 4% paraformaldehyde, embedded in methacrylate and sectioned at 1-μm thickness along the dorsal-ventral axis. Pictures of hematoxylin and eosin (H&E)-stained retinal sections were taken at 20X and 40X magnifications. Outer nuclear layer (ONL) thickness was determined from sections displaying optic nerve head. Only the dorsal half was measured by taking nine measurements at equal spacing from the optic nerve head to the periphery on each retinal section.

Immunoblotting and Immunofluorescence Microscopy
Whole retina was homogenized in radioimmunoprecipitation assay buffer. Retinal proteins were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes. Membranes were blocked with StartingBlock For immunofluorescence, eyes were enucleated and fixed in 4% paraformaldehyde for 30 minutes, cryo-preserved through a series of sucrose gradients, cryo-sectioned at 10-μm thickness, and processed as previously described. 41 Primary antibodies were used at the following dilutions:

Electroretinography (ERG)
ERG recordings were performed with a computer-based system (E2, Diagnosys LLC, Lowell, MA, USA), using flashes produced with LEDs or Xenon bulbs. Mice were darkadapted overnight and anesthetized with an intraperitoneal injection of ketamine (80 mg/Kg) and xylazine (8 mg/Kg). Pupils were dilated with topical phenylephrine (2.5%) and tropicamide (0.5%). Corneal ERGs were recorded from both eyes using gold wire loop electrodes with a drop of 2.5% hypromellose ophthalmic demulcent solution. Reference was taken with a gold wire loop placed in the mouth. For dark-adapted ERG, mice were stimulated with flashes of increasing light intensity (from 0.0001-1000 cd.s.m −2 ). Responses were computer-averaged and recorded at 3 to 60 second intervals depending on the stimulus intensity. For light-adapted ERG, mice were light-adapted for 2 minutes and were stimulated with flashes (from 0.3-100 cd.s.m −2 ) in the presence of a white 32 cd.m −2 rod-suppressing background.

HSP70 Expression in the Developing and Degenerating Retina
We previously reported developmental transcriptome profiles of rods (flow sorted from Nrl-GFP mice) and Scone-like photoreceptors (flow sorted from Nrl −/− ;Nrl-GFP mice) and of whole retinas. 35,36,42 Focusing on chaperone genes, we noticed that expression of the two HSP70 transcripts, Hspa1a and Hspa1b, in rods progressively increased from postnatal day 2 (P2) to P28 (Fig. 1A). Compared with an even distribution of the transcription factor Hsf1, expression of the cognate Hspa8 was somewhat enriched in rods at P28 (Fig. 1B). Notably, Hspa1a and Hspa1b showed dramatically higher expression in rods at P28 compared with S-cone-like photoreceptors and the whole retina (Fig. 1B). Accordingly, HSP70 protein increased in temporal correlation with maturation of the retina (Fig. 1C). Immunoblotting of retinal extracts from seven mouse models (rd1, rd16, rd10, Rpgrip1 −/− , rds, Nrl −/− , and RHOT17M) revealed transient expression of HSP70 at early/mid stage of degeneration, except in the RHOT17M retina (Fig. 1C). The timing of HSP70 expression correlated with the severity of disease. In mouse models with a relatively fast degeneration (rd1, rd16, and rd10), HSP70 expression was observed at P14, whereas relatively slow mouse models (Rpgrip1 −/− , rds, and Nrl −/− ) revealed later expression from P18 to P28. We also noted that HSP70 expression dramatically decreased during the progression of photoreceptor degeneration in the earlier mentioned six models. Distinctively, HSP70 expression was evident at P14 in a slow degeneration model RHOT17M and persisted as the degeneration progressed. In contrast, we observed a stable temporal expression of the constitutively expressed cognate HSC70 and HSP90, a member of the 90-kDa subfamily, in degenerating retinas (data not shown).

Generation and Characterization of TgCrx-Hspa1a-Flag Mice
To test the potential neuroprotective effect of HSP70, we generated a transgenic mouse line with augmented HSP70 expression in both rod and cone photoreceptors, controlled by a 2.3 kb mouse Crx promoter. The TgCrx-Hspa1a-Flag (referred to as TgHspa1a+) mouse expresses a FLAG-tagged full-length HSP70 protein. Immunofluorescent labeling of FLAG on TgHspa1a+ retina revealed transgene expression in photoreceptors with strong signals in the inner segment (IS) and synaptic terminals ( Fig. 2A). Weaker staining of FLAG was observed in inner plexiform layer, possibly resulting from Crx promoter activity in bipolar cells 43 (Fig. 2A). Robust expression of HSP70 protein driven by the transgene was detected by immunoblotting in P14 TgHspa1a+ retinas, and the expression gradually increased at P28 and P60 (Fig. 2B). Eight-month old TgHspa1a+ mice displayed normal retinal morphology (Fig. 2C) with no difference in ONL thickness between TgHspa1a+ (n = 4) and TgHspa1a-(n = 5) littermates (Fig. 2D). ERG response (a-and b-wave amplitude) showed no change between TgHspa1a+ (n = 3) and TgHspa1a-(n = 6) mice in dark-or light-adapted conditions (Fig. 2E). Thus augmented expression of HSP70 does not alter the development, survival, or function of wild-type photoreceptors.

Improved Photoreceptor Survival but not ERG Response by Augmented HSP70 Expression in RHOT17M Retina
TgHspa1a+ mice were crossed to RHOT17M mice, which express a class II mutant rhodopsin displaying defects in stability and subcellular localization. 37,44 RHOT17M mice exhibit slow degeneration with complete photoreceptor loss by 8 months of age. 37 Immunoblotting confirmed robust expression of HSP70 from the transgene in 3-monthold RHOT17M;TgHspa1a+ retinas ( Fig. 3A). A marginal but significant increase in ONL thickness was observed from center to periphery in cross-sections of 5-month-old RHOT17M;TgHspa1a+ (n = 8) retina compared with littermate controls (n = 6) (Figs. 3B, 3C). An increase in inner nuclear layer and inner plexiform layer thickness was also evident. However, transgene expression in mutant retina did not translate into a significant improvement of scotopic or photopic ERG amplitudes (Fig. 3D).

Accelerated Photoreceptor Degeneration by Augmented HSP70 Expression in rd10 Retina
TgHspa1a+ mice were crossed to rd10 mice, which carry a missense mutation in the Pde6b gene and display relatively fast degeneration with complete photoreceptor loss at 2 months. 45 Robust expression of HSP70 from the transgene was detected by immunoblotting in P20 rd10;TgHspa1a+ retinas (Fig. 5A). More severe degeneration was observed in the peripheral retina of rd10;TgHspa1a+ (n = 5) mice at P30 compared with rd10;TgHspa1a-(n = 5) littermates (Figs. 5B, 5C). At P30, and not earlier, we observed significantly reduced scotopic b-wave amplitude and dramatically reduced photopic a-and b-wave amplitudes in mice expressing the transgene (Fig. 5D).

Altered Rhodopsin and PDE6B Expression with Augmented HSP70 Expression in RHOT17M and rd10 Retina
Immunolabeling revealed robust staining of rhodopsin in the OS of 3-month-old RHOT17M;TgHspa1a-and RHOT17M;TgHspa1a+ retinas (Fig. 6). We observed weaker staining of rhodopsin in the IS area (marked by calreticulin) in RHOT17M;TgHspa1a+ retina ( Fig. 6 insert box), suggesting reduced retention of mutant rhodopsin. High expression of HSP70 in 2-month-old retina under a nondegeneration background (C57BL/6) does not affect expression or localization of PDE6B protein (Fig. 7). In the rd10 retinas, PDE6B expression level at P20 is comparable with or without augmented HSP70 expression, whereas reduced staining of PDE6B was detected in P30 rd10;TgHspa1a+ mice compared with controls (Fig. 7).

DISCUSSION
High turnover of proteins makes the photoreceptors especially vulnerable to disruptions of proteostasis, a feature commonly observed in neurodegenerative diseases. Several retinal disease-causing mutations produce misfolded or mistargeted proteins, accumulation of which can trigger cellular responses that eventually lead to demise of the photoreceptors. 46,47 Thus manipulation of chaperone system and UPR pathways offers an attractive, potential therapeutic strategy to treat retinal degeneration irrespective of the original mutation. The HSP family of proteins control many aspects of cellular proteostasis including nascent protein folding, trafficking to organelles, repair of damaged proteins, and assembly of multiprotein complexes. In doing so, they promote cellular survival and longevity against proteotoxic stress implying a likely role in neuroprotection; yet there is still no detailed investigation of their therapeutic effects in retinal degenerative conditions.
In this study, we observed transient high expression of HSP70 protein as a common feature in early stages of degenerating retinas. Unexpectedly, augmented HSP70 expression through a transgenic approach manifested divergent effects in the three retinal degeneration mouse models we tested. A protective effect on augmentation of HSP70 expression was only attained in the RHOT17M mice. Rhodopsin mutations can be classified into several subgroups based on the mutant protein's structure, stability, function, and intracellular localization. 48,49 T17M rhodopsin at low concentration displays altered post-translational modification and reduced stability; but the majority can fold well enough to exit endoplasmic reticulum (ER) and translocate to the OS. 49 In the RHOT17M mice and in transiently transfected cell cultures, T17M rhodopsin manifests reduced thermal stabil- ity, resulting in partially misfolded proteins and moderate ER retention, lacking full ability of regeneration with the chromophore 11-cis-retinal, thereby failing to integrate in the plasma membrane. 37,49,50 Induction of ER stress was reported in RHOT17M retinas, 51 and manipulating ER stressrelated UPR pathway could improve photoreceptor survival in RHOT17M mice. 52,53 Previous studies examined different methods to improve chaperone activity in treating retinal degeneration related to rhodopsin aggregation. 49,54 In this study, improvement in ONL thickness indicated protective effects in the RHOT17M retina with augmented HSP70 expression. Reduced retention of mutant rhodopsin in the IS suggests that HSP70 facilitates folding and translocation of mutant rhodopsin to the OS. Further studies are however needed to elucidate the precise mechanism of HSP70's protective effect, with quantitative analysis of the changes in mutant opsin localization to the OS versus retention at ER or degradation at proteosome.
In Rpgrip1 −/− and rd10 mice, augmented HSP70 expression did not demonstrate a protective effect; instead, we observed an accelerated degeneration in the rd10 retina and no apparent impact in the Rpgrip1 −/− retina. RPGRIP1 functions as a docking site for ciliary proteins at the base of the connecting cilium and likely facilitates trafficking of nascent proteins from the IS into the OS. 38,[55][56][57] With loss of RPGRIP1, RPGR fails to anchor at the base of the connecting cilium and the proper transport and localization of rod and cone opsin are affected. 38 Enhancing the chaperone system does not rescue the phenotype, and no protective effect was observed. The rd10 mice carry a missense mutation in the gene encoding β-subunit of cGMP phosphodiesterase 6 (Pde6b). 45 The rd10 allele produces a hypomor- phic PDE6B that is partially defective in assembling into the PDE6αβγ 2 holoenzyme with decreased transport to the OS, leading to higher free cGMP, increased channel opening and Ca 2+ influx, and rapid photoreceptor death. [58][59][60] We were intrigued by the accelerated photoreceptor degeneration in rd10 retina by HSP70 overexpression. The mutant PDE6B is expressed at low levels in rd10 retina, exhibits residual catalytic activity, and is partially mislocalized, consistent with a misfolding defect. 58 Increased expression of HSP70 would enhance the quality control mechanisms, which target misfolded proteins, 61 as evident by decreased staining of PDE6B in rd10;TgHspa1a+ mice (Fig. 7). Photoreceptor specific chaperone AIPL1 27,28 directly interacts with HSP70 and HSP90, 25,62 forming a chaperone heterocomplex that regulates biosynthesis, stability, and translocation of PDE6 holoenzyme. 26 HSP70 and HSP90 may compete for binding to the client protein PDE6B. We suggest that overexpression of one arm of the chaperone heterocomplex may lead to imbalance of the system, enhancing PDE6B degradation and accelerating photoreceptor degeneration. Thus augmenting chaperone activity might have limited or no effect when dealing with protein-null defects such as Rpgrip1 −/− , and overstimulating the chaperone machinery may accelerate degeneration in specific cases such as rd10 mice.
It is unclear why improved rod survival in RHOT17M mice was not accompanied by a corresponding improvement in dark-adapted ERG amplitudes. HSP70 may assist mutant rhodopsin folding and/or trafficking to the correct subcellular compartment (i.e., the OS). Although likely relieving ER stress and thereby promoting cell survival, increased mutant rhodopsin in the OS would interfere with phototransduction. Misfolded mutant rhodopsin that fails to fully regenerate with the chromophore can stimulate phototransduction via transducin activation in the absence of light and reduce rods sensitivity because full dark adaptation is not achievable, as if they are constantly exposed to light. 49 We hypothesize that increased trafficking of mutant rhodopsin to the OS may at least in part account for the lack of ERG improvement in this model. Surprisingly, we observed drastically reduced photopic a-and b-wave amplitude in rd10 mice at P30 and significantly reduced photopic b-wave amplitude in Rpgrip1 −/− mice at P40 with augmented HSP70 expression, suggesting compromised cone function or survival. As the TgHspa1a transgene has no deleterious effect on cone function and survival in wild-type background, a likely scenario in rd10 retina might be that accelerated rod decline secondarily impacted cone viability.

CONCLUSIONS
Members of HSP70 family usually work together with HSP40 (DNAJ), HSP90, and HSP110 family co-chaperones, forming a large and complex chaperone network in maintaining cellular proteostasis. DNAJs, which regulate HSP70's function as co-chaperones, play an important role in suppressing disease-causing protein aggregations. 63 DNAJB6b and DNAJB8 are shown to be superior suppressors than HSP70 in treating aggregates associated with polyglutamine expansion. 64 Thus HSP70 may not be the only or best candidate for therapies targeting the chaperone system. We had previously demonstrated retinal expression of the small HSP α-crystallin, traditionally believed to localize in the ocular lens. 65 Expression of α-crystallin is elevated in diseased retina where it may play a protective role. 66,67 In some cases, augmentation of HSP function can be a double-edged sword, and augmentation of different HSPs might result in unintended consequences. A member of the 90 KDa family, HSP90α, predominantly expressed in the retina and brain, 68 has also been targeted for treating retinal degeneration. 69 Although inhibition of HSP90 protected dying photoreceptors in vivo, 31,70 prolonged HSP90 inhibition led to photoreceptor death in beagle dogs, 71 and HSP90α-deficient mice displayed retinal degeneration. 68 Given that HSP70 augmentation in our studies showed distinct impact on degenerating photoreceptors from different mutants, we suggest that a better understanding of the chaperone machinery and especially HSPs in degenerating retinas is required and that various chaperone candidates should be considered when developing treatment paradigms focused on photoreceptor protein homeostasis.