Visual recovery following optic nerve crush in male and female wild-type and TRIF-deficient mice

2.1 Animals

All animal-related procedures in this study were performed in accordance with Emory University. Division of Animal Resources guidelines for the use of experimental animals were approved by the Institutional Animal Care and Use Committee (Emory University protocol DAR-2003137-063018GN); and conformed to the National Institutes of Health guidelines and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.

A total of 64 adult female and male C57BL/6 mice and C57BL/6J-Ticam1^Lps2/J (TRIF-knockout mice) 8–10 weeks old were used in the experiment. The TKO mice were bred in the laboratory and were the third generation of mice originally obtained from the Jackson Laboratory, Bar Harbor, ME, USA. Mice were housed on a 12-hour light/dark schedule with water and food available in an Emory University DAR animal facility.

2.2 Experimental design

Four experimental groups –TKO ONC (12 males and 8 females), WT ONC (10 males and 8 females), TKO sham (9 males and 5 females), and WT sham (7 males and 5 females) –were used in the experiment. Power analysis was used to determine the group sizes required to reject the null hypothesis of no effect of TKO with a p-value 0.05 at a power of 0.8. From each ONC experimental group, 10 TKO and 8 WT mice were taken for immunohistochemistry, 5 TKO and 5 WT mice for western blot, and 5 TKO and 5 WT for functional tests; from each sham experimental group, 6 TKO and 6 WT mice were taken for immunohistochemistry, 4 TKO and 3 WT animals for western blot, and 4 TKO and 3 WT animals for functional/behavioral tests. Electrodes were implanted over the primary visual cortex of mice in the functional group and the mice allowed to recover for 5 days. ONC surgery or sham operations were then performed on all mice. Visual evoked potentials (VEPs) and optomotor responses (explained below) were assessed on days 14, 30, and 80 after ONC injury. Mice were euthanized and their retinas, optic nerves and brains were collected for immunohistochemistry and western blot of inflammatory markers on day 7 after ONC, and immunohistochemistry of growth-associated protein 43 (GAP-43) and RGCs on day 30 after ONC. The timeline of the experiment is shown in Fig. 1.

OPEN IN VIEWER

Fig. 1

Experimental Timeline.

Seventeen mice were anesthetized with 80 mg/kg ketamine and 16 mg/kg xylazine i.p. and fixed on a stereotaxic frame with their body temperature maintained at 37 °C with a heating pad. Chlorhexidine solution was used to disinfect the skin after the head was shaved. After the application of the topical anesthetic proparacaine 1% to the scalp, a midline scalp incision was made. The exposed skull was cleaned with sterile saline and dried. Stainless steel screws (Fine Science Tools, Heidelberg, Germany) were inserted over the V1b cortical surface of each hemisphere through small burr holes (V1b: 0.0 mm lambda anteroposterior, ±3.0 mm mediolateral) in the skull above the visual cortex. Cyanoacrylate and dental cement were used to stabilize the screws.

2.4 Unilateral ONC

ONC was induced in 20 WT mice and 18 TKO mice. All mice were anesthetized with 80 mg/kg ketamine and 16 mg/kg xylazine i.p. A lateral incision of the conjunctiva was performed, followed by the exposure of the optic nerve by blunt dissection, leaving the dura and blood supply intact. Dumont self-closing forceps N7 (Fine Science Tools, Foster City, CA, USA) with a calibrating micrometer screw (Emory University Physics Machine Shop) was used to apply pressure on the optic nerve for 5 seconds at a distance of 1 mm from the eye, with the forceps jaws 0.04 mm apart. All procedures were performed aseptically on the left eye. Topical antibiotic eye ointment (Bausch and Lomb, USA) was applied to the injured site. No surgical procedures were performed on the right eye. Animals in the sham group were given the anesthesia as well as the exposure of the optic nerve with optic nerves left intact.

2.5 Fixation and sectioning procedures

At 7 and 30 days after ONC, 15 mice from each time point were deeply anesthetized with 80 mg/kg ketamine and 16 mg/kg xylazine i.p. They were then perfused intracardially with 0.1 M phosphate buffered saline (PBS) and then with 10% buffered formalin. Their left retinas and left optic nerves were removed, post-fixed overnight, and immersed in 15% sucrose solution followed by 30% sucrose solution for cryoprotection. Frozen retinas were then sectioned into 14μm cross-sections and the optic nerves into 10-μm longitudinal sections.

2.6 Immunohistochemistry procedures

The retinal cross-sections were rehydrated in 0.1 M TBS and permeabilized in TBS-Tween 0.3% solution. After protein blocking with serum-free blocking solution (X0909 Dako, Carpinteria, CA, USA) in the humid chamber for 30 min at room temperature, slides were incubated with primary antibody at 4°C overnight. Primary antibody raised in rabbit against Iba-1 (Wako, WDJ 3047) diluted 1:1000 was used as a marker for microglial cells. Primary antibody raised in rat against glial fibrillary acidic protein (GFAP) (ab7260, 1:500; Abcam, Cambridge, MA, USA) was used to mark astrocytes. Primary antibody raised in Guinea pig against RMPBS (retina binding protein with multiple splicing; Santa Cruz Biotechnology, Santa Cruz, CA, USA, sc-2930, 1:400) was used to visualize RGCs and primary antibody raised in mouse against GAP43 (sc-17790, 1:500, Santa Cruz) was used to label regenerated axons within the optic nerve. Slides were washed with TBS-Tween 0.3% three times, 10 minutes each time, prior to incubation with fluorescent secondary antibodies for 2 hours at room temperature. Secondary antibody raised in goat against rabbit (5220-0336, 1:200, SeraCare Life Sciences, Milford, MA, USA), goat against rat (#14-16-06, 1:200, Kirkegaard & Perry Laboratories, Gaithersburg, MD, USA), goat against guinea pig (#14-17-06, 1:200, KPL), and goat against mouse (#5220-0341, 1:200; SeraCare) were used as secondary antibodies to visualize the respective primary antibodies. Sections incubated with only secondary antibodies served as negative controls. Afterwards, slides were mounted with Vectashield mounting medium with 4’, 6-diamidino-2-phenylindole (DAPI, 0.5μg/ ml, Vector Laboratories, Burlingame, CA, USA) cover-slipped and then were scanned in a fluorescent microscope (Ziess Axioskop, Zeiss, Jena, Germany).

2.7 Western blot

Isoflurane was used to euthanize the animals before decapitation. Retinas were quickly removed and dissected under a binocular microscope. Retinas were then homogenized in T-PER buffer (78510, Thermo Fisher Scientific, Waltham, MA, USA) containing a protease inhibitor cocktail (P8340, Sigma, St. Louis, MO, USA). The lysate was separated by centrifugation at 11 rpm at 4 Celsius for 10 min. Protein concentration was measured using Bradford protein assay.

Twenty-μg samples were loaded into 20% sodium dodecyl sulfate poly acrylamide gradient gel (SDS-PAGE) and transferred onto polyvinylidene fluoride transfer membranes using a 110 V for 45 minutes. The blots were washed with TBS-Tween 0.3% and blocked with 5% milk for 2 h at room temperature. The membranes were incubated overnight at 4° with primary anti-NF-kB p65 antibody raised in rabbit (1:1000, #8242, Cell Signaling, Danvers, MA, USA), diluted in 5% milk in TBS-Tween 0.3%. Three washes with TBS-Tween 0.3% were followed by incubation with peroxidase-conjugated secondary antibody for 90 minutes at 25°. Blots were visualized with enhanced chemiluminescence reagent and developed with HyBlot CL film in the darkroom by an enhanced chemiluminescence system (E2400, Denville Scientific, South Plainfield, NJ, USA). The same membranes were incubated with anti-β-actin for loading control (1:1000, A5316, Sigma). Blots were quantified using Image J software (NIH) and density of NF-kB normalized to density of β-actin in the same gel.

2.8 Axon quantification in the optic nerve

GAP-43 was used as a marker of re-growing axons. The number of GAP43 positive axons were counted 0.50 mm from the crush site in nine stained optic nerves; we counted four nerve sections per each nerve. Based on the nerve radius and the thickness of the nerve slices, the total number of GAP-43 positive axons was quantified by ∑ad  =  πr2 ×[average axons/mm]/t, where total number of axons 0.50 mm from the crush site in a nerve having a radius of r and a thickness of t was estimated.

2.9 VEP recording and analysis

VEPs were recorded in the left and right primary visual cortices of all mice at 14, 30, and 80 days after ONC. Prior to testing, mice were anesthetized with 80 mg/kg ketamine and 16 mg/kg xylazine, i.p., and placed on a heating pad in front of a computer monitor. Eyes were lubricated with 0.9% saline drops. A bandpass filter set at 0.3–500 Hz from the UTAS-E3000 system and software (LKC, Gaithersburg, MD, USA) was used to acquire and amplify the signals.

Visual stimuli consisting of sinusoidal gratings of 100% contrast (0.05 cycles/deg, reversal cycle 1 s) were presented independently to the left and right eyes in random order at 300 stimuli per block and two blocks per eye (Sergeeva et al., 2018). To prevent its stimulation, the opposite eye was covered with a black plastic monocular patch. Gratings in vertical orientation were used as stimuli. Each single evoked response was 500 msec long and averaged for the 300-stimulus block. The peak-to-peak amplitude of the averaged VEP was calculated in the ipsilateral and contralateral visual cortex of the stimulated eye.

2.10 Measurement of optomotor response

To assess the visual spatial frequency threshold, a virtual optomotor system was used (OptoMotry system; Cerebral Mechanics, Lethbridge, AB, Canada) (Prusky et al., 2004). Stimuli consisted of vertical gratings (12°/s) of 100% contrast with an initial spatial frequency of 0.003 cycles per degree. The ability of the 17 mice to detect and react to stimuli was indicated by their head movements in the corresponding moving direction of the stimuli. Both eyes were stimulated to measure their visual acuity. Spatial frequency increased stepwise until the response threshold was crossed three times, and the highest spatial frequency to induce head movement in accordance with the stimulus direction was considered the spatial frequency threshold for the specific eye.

2.11 Statistical analysis

IBM SPSS Statistics 23.0 for Windows was used to evaluate all data. Repeated and factorial measures ANOVA were used to compare differences in VEP and optical mark recognition (OMR) over time. Immunofluorescent and western blot densities of each marker were measured in the retina of the lesion side with NIH ImageJ FIJI software. Five sections of each retina were selected and analyzed by density. Individual group differences were evaluated by factorial ANOVA and then with a post-hoc Tukey test. Standard errors were used because of unequal group numbers in some of the individual comparisons. Data and probabilities are reported according to the latest guidelines proposed by the American Statistical Association (Wasserstein et al., 2019).

3.1 Reduced inflammatory response in the retina of TKO mice 7 days after ONC

According to previous studies, it is thought that 7 days is when inflammatory responses in the optic nerve reach their peak, followed by a decrease in inflammation around 14 days (MacNair et al., 2016). Using these findings as guidelines for this report, we observed that, 7 days after ONC, our TKO mice showed a reduced inflammatory response in the retina compared to WT mice (Fig. 2). There were fewer microglial cells (as shown by less Iba-1 staining, p = 0.03) and reduced astrocyte and muller glia activation (as shown by less GFAP staining, p = 0.01) in retinas of TKO mice. The inflammatory protein NF-kB was also decreased in male TKO mice 7 days after ONC (Fig. 2C). The sham surgery groups of TKO and WT mice were comparable on all inflammatory measures.

OPEN IN VIEWER

Fig. 2

TKO mice had less inflammatory response compared to WT mice 7 days after optic nerve crush. (A) Fewer microglia were present (indicated by reduced IBa-1 positive area, p = 0.03, df = 14, F = 4.2). NF-KB normalized to β-action density. Left panels show results from 9 mice per crush conditions and 6 mice for sham conditions. (B) Fewer astrocytes were present (indicated by reduced GFAP-positive area, p = 0.01, df = 14, F = 6.5). Left panels show results from 9 mice per crush conditions and 6 mice for sham conditions in total of both TKO and WT groups. (C) Levels of the transcription factor NF-kB were lower in TKO males compared to WT males (p = 0.03, df = 10, F = 7.0). Data were evaluated by one-way ANOVA with post-hoc for between-group comparisons; vertical bars indicate standard errors. Right panels of (A) and (B) show immunofluorescent images of the relevant marker; scale bar is 50μm. The bottom panel of (C) shows samples of western blot.

Previous research from other groups has shown that by 30 days, axon elongation is largely ended. Thirty days is also about the time at which the growth and formation of synaptic connections between the re-growing axons and tectorial neurons (Li et al., 2017) is slowing if not completely finished. Again, we used these data to see if the relationship between the RGC loss and axonal regeneration is by-and-large completed or at least substantial by 30 days post-injury.

Accordingly, in our study, we were able to report less RGC loss and greater axonal regeneration in TKO mice 30 days after ONC. RGCs in the retina were visualized 30 days after ONC (Fig. 3A, bottom panel). TKO mice demonstrated less RGC loss than did WT mice as indicated by higher numbers of RNA-binding protein with multiple splicing (RBPMS) positive cells (p = 0.027)

OPEN IN VIEWER

Fig. 3

TKO mice had better RGC survival and better axon regeneration 30 days after ONC. (A) More RGCs were present in the retina of TKO (measured as RBPMS immunoreactive cells, p = 0.027, df = 14, F = 4.5). Bottom panel shows sample retinal cells of TKO (left) and WT (right) mice. (B) More axons were present in the optic nerve (measured as GAP-43 immunoreactive cells, p = 0.01, df = 14, F = 5.9) of TKO mice. No GAP-43 is manifested under the sham condition. Left panels show results from 9 mice per crush condition and 6 mice per sham condition in both TKO and WT groups; data were evaluated by one-way ANOVA with post-hoc test for between-group comparisons; vertical bars indicate standard errors. Bottom panel shows sample axon GAP-43 count in TKO (left) and WT (right) mice at 0.5 mm from the crush site (white line); scale bar is 500μm.

Compared to WT mice, TKO mice exhibited considerably higher numbers of GAP-43-stained axons 30 days after ONC, indicating greater axonal growth after nerve injury (p = 0.01). The bottom panel of Fig. 3B shows the site of GAP-43 measurement at a distance of 0.50 mm from the crush site in the optic nerves of TKO and WT mice.

Interestingly, despite evidence of the presence of regenerating axons at 30 days post-ONC as detected with GAP-43 staining, no difference in OMR between TKO and WT mice was seen 14 and 30 days after the injury. In our study, the higher spatial frequency threshold in TKO mice compared to WT mice only appeared 80 days after ONC (Fig. 4A; p = 0.05).

OPEN IN VIEWER

Fig. 4

TKO mice may have slightly improved visual function 80 days after ONC. No between-group differences in measures of visual function were found at 14 or 30 days after ONC. At 80 days, (A) optomotor response spatial frequency threshold was slightly higher in TKO compared to WT mice (p = 0.05, df = 16, F = 14.4). (B) We found no comparable differences (p = 0.054, df = 16, F = 3.9) in visual evoked potentials of TKO mice in the primary visual cortices contralateral to the damaged nerve in ONC mice, (C) but visual evoked potentials of TKO mice were greater on the side ipsilateral to the intact nerve compared to WT controls (p = 0.01, df = 16, F = 10.2) when evoked from intact eye. Data were derived from 10 mice per crush and 7 mice per sham condition in both TKO and WT groups and were evaluated by one-way ANOVA with post-hoc test for between-group comparisons; vertical bars indicate standard errors.

The amplitude of VEPs in the contralateral primary cortex of the injured left eye was not different for WT and TKO mice 80 days after ONC (Fig. 4B). In contrast, the pattern of VEPs in the ipsilateral primary visual cortex evoked by the stimulation of the intact eye in WT and TKO mice exhibited differences at 80 days. TKO mice showed an increased response in the right visual cortex compared to that of WT mice when the intact eye was stimulated (Fig. 4C; p = 0.047). Visual functional recovery requires some sort of triggering events –in this case, at least regenerating axons support the projection to the appropriate area of the visual cortex, thus helping to form an accurate map of the visual space. Other researchers have also shown that functional recovery is not seen until ten weeks after ONC (De Lima, 2012). Please see Discussion for more details.

TKO female and male mice showed different levels of NF-kB after nerve injury. Specifically, TKO males demonstrated lower NF-kB levels compared to TKO females at 7 days after ONC (p = 0.01) (Fig. 5A). Sex differences in recovery from ONC were also manifested in the differences in RGC numbers between TKO females and males. As summarized in Fig. 5B, TKO males had less RGC loss compared to TKO female mice 30 days after ONC (shown by RBPMS staining, p = 0.048). Furthermore, 80 days after ONC, there is a difference between the visual acuities of TKO female and TKO male mice (p = 0.028), with TKO males having a higher visual acuity than females (Fig. 5C).

OPEN IN VIEWER

Fig. 5

Female TKO mice had a greater inflammatory response 7 days and greater RGC loss than male TKO mice 30 days after ONC, and lower visual acuity 80 days after ONC. (A) Female TKO mice had more NF-kB than males (p = 0.01, df = 8, F = 16.4); NF-KB normalized to β-action density. (B) Female TKO mice had fewer RGCs than males (p = 0.048, df = 8, F = 5.7). Data came from 3 males and 2 females in the crush condition and 2 males and 2 females in the sham condition; data were evaluated by one-way ANOVA with post-hoc for between-group comparisons; vertical bars indicate standard errors. (C) 80 days after ONC, male TKO mice have a higher visual acuity than those of female TKO mice.

4.1 The TRIF adaptor in inflammatory pathways of the CNS

So far, only a few studies have addressed the involvement of the TRIF pathway in a rodent ONC model, and to the best of our knowledge, only ours has looked specifically at the TRIF pathway in recovery of visual function. We assessed visual responses using optomotor tests and VEPs and then measured injury-associated biomarkers in normal and TKO mice with optic nerve damage. The functional visual tests were conducted over a period of 80 days, allowing for evaluation of regeneration of the optic nerve. The longer observation of recovery used in our study led to greater observable differences between TKO and WT mice in the restitution of visual function. As assessed by RGC survival and visual acuity, the recovery pattern from ONC differed between males and females. Overall, we observed sex differences in two measures of inflammatory processes and in neuronal cell loss and visual acuity. These differences are discussed in section 4.5.

We then evaluated changes in the markers of inflammatory responses (GFAP, Iba-1), RGCs (RMPBS), and regenerating axons (GAP-43) and noted that the recovery of visual function is associated with the reduced inflammatory response with TRIF deficiency. Other researchers have demonstrated the significance of the TRIF adaptor in disease-associated inflammatory responses. For example, it has been shown that TRIF signaling is required for the caspase 11-dependent immune response and thus is involved in the lethality of endotoxemia and sepsis (Tang et al., 2018). TRIF deficiency also reduces the number of natural killer T lymphocytes and CD8-T cells infiltrating into the spinal cord of mice with amyotrophic lateral sclerosis (Komine et al., 2018). Other studies show that TRIF protein is essential in activating the common MyD88-dependent signaling pathway as well as a MyD88-independent pathway. The dual signaling of MyD88 and TRIF is thought to be crucial for dendritic maturation (Yamamoto et al., 2003). In line with the work of Lin et al. (2012), our study showed that NF-kB, a major pro-inflammatory signaling pathway protein, is reduced when TRIF is deficient. In addition, after injury in the RGCs, expression of the downstream proteins TANK binding kinase 1 (TBK1) and IκB kinase ɛ (IKKɛ) in the TRIF pathway also decreased in TKO mice (Lin et al., 2012). Lin et al. further showed that other inflammatory factors such as nitric oxide synthase, tumor necrosis factor-a, interferon-β, and interleukins 1β, 6 and 17 were decreased in the RGCs of TKO animals. These findings suggest to us that deletion of the TRIF gene can result in decreased activation of the inflammatory NF-kB pathway and thus in decreased recruitment of neurotoxic factors in the injured retina.

We also observed that expression of both GFAP and Iba-1 markers was reduced in TKO mice 7 days after ONC, indicating reduced astrocyte and microglial activation in response to the injury. In the retina, TRIF is expressed only in microglia (Lin et al., 2012). Since astrocytes also compose the microenvironment of the unmyelinated part of the RGC axon, lower levels of TRIF may lead to interactions that indirectly reduce astrocyte activation (Laha et al., 2017). This reduction is triggered by the reduction of miRNA-21 activation in astrocytes and by decreased cytokine expression in the microenvironment. Under these conditions, there is decreased astrocytosis (Li et al., 2016) resulting in more circumscribed optic nerve injury.

Since the primary focus of our experiment was to compare post-ONC inflammatory responses in TKO versus WT mice, the measure of acute inflammatory responses was performed only at 7 days after the injury. Future studies could analyze the progression of astrocyte and microglial cell changes over the time course of the inflammatory response to help clarify the association between post-injury development of the two types of glial cells. The phenotype of the microglial cells after ONC could also be examined to evaluate cellular activation, since the morphology of retinal microglial cells changes from a resting dendritic-like shape to an ameboid form upon activation after injury and other factors (Wang et al., 2016; Espinosa-Garcia et al., 2017). It would then need to be determined what role these phenotypic changes play in the chronic recovery processes and optic nerve regeneration, if any.

4.2 The TRIF inflammatory pathway and RGC death

In our study we found that 30 days after ONC, RGC death and inflammatory markers in the retina of TKO mice were less evident than those seen in WT mice. In this context, reactive gliosis is generally considered pro-inflammatory, and anti-inflammatory p38 kinase inhibitors prevent the degradation of axons in the optic nerve (Dapper et al., 2013). The removal of components of astrocytosis, such as GFAP and vimentin, is also associated with adult neurogenesis and greater neuronal survival following injury (Pekny & Pekna, 2014).

Our finding is in line with the study by Liddelow et al. showing that ONC injury in mice induced robust A1-phenotype, neurotoxic, astrocyte generation accompanied by RGC death. The reactive A1 astrocytosis was induced by activated microglia secreting pro-inflammatory factors such as Il-1α, TNFand C1q; blocking of those factors either by antibodies or genetic knockout greatly improved RGC survival (Liddelow et al., 2017).

We observed that RGC survival and levels of GAP-43 protein, a marker of re-growing axons, were higher in both male and female mice with reduction in TRIF-induced inflammation. Down-regulation of TRIF-induced inflammatory pathways and reduced glial cell proliferation correlated with better RGC survival and axonal regeneration in male TKO mice. RGC survival and axonal regeneration are also up-regulated by the inhibition of miRNA-21, which reduces Muller cell gliosis (Li et al., 2019). Down-regulation of both glial cell expression and inflammatory pathways can be regarded as factors contributing to repair of the injured optic nerve (Laha et al., 2017).

It is important to note that recent studies on other parameters of injury have shown that reactive gliosis and inflammation can initiate and/or mediate processes that promote the regeneration of axons (Calkins et al., 2017; Yin et al., 2009). For example, microglia are known to have beneficial roles after injury, including debris clearing and synapse maintenance (Herzog et al., 2019). However, Hilla et al. showed that RGC degeneration after ONC was not affected by microglia depletion, while additional depletion of macrophages slightly compromised RGC regeneration (Hilla et al., 2017). In a study by Anderson et al. (2016), the prevention of astrocyte scar formation resulted in failed regrowth of transected corticospinal axons after severe spinal cord injury lesions in mice. Others have reported that additional recruitment of microglia as a result of inflammation caused more neural tissue regeneration in the central and peripheral nervous systems (Shechter & Schwartz, 2012; Niemi et al., 2013).

It has been known for quite some time that some inflammation is necessary to initiate axon regeneration (Yin et al., 2009; Benowitz & Popovich, 2011), but persistent activation of the inflammatory responses typically leads to further damage to the optic nerve (Schwartz, 2004). It has also already been established that reduced microglial activation is associated with reduced retinal and optic nerve degeneration in mice with glaucoma (Bosco et al., 2012). Interestingly, an increase in TLR4 and other TLRs in retinal glaucoma leads to greater microglial activation and worsening of glaucoma (Luo et al., 2010). Overexpression of TLR receptors and the consequent increase in TRIF protein is associated with greater inflammatory responses leading to detrimental effects to the optic nerve, including the apoptosis of RGCs and interruption of nerve regeneration (Ullah et al., 2016; Mohan et al., 2018; Vijay, 2018).

4.3 RGC survival and axonal regeneration

Since GAP-43 is associated with more RGC survival 30 days after injury in TKO mice, RGC survival and regeneration could be governed by different mechanisms (Calkins et al., 2017). Intraocular inflammation leads to increased survival of RGCs, but increases in brain-derived neurotrophic factor eliminate the effect of intraocular inflammation on axonal regeneration (Pernet & Di Polo, 2006). Similarly, deletion of dual leucine zipper kinase (DLK) in RGCs has been shown to increase survival of RGCs but hampers regeneration (Watkins et al., 2013). In general, there seems to be a link between pro-survival and pro-regeneration effects of molecular factors and pathways as demonstrated in our analyses of the TRIF protein signaling system. Similar linkages were also observed in the effects of mTOR upregulation and cAMP and PTEN deletion on recovery from ONC, and these have been associated with both survival of RGCs and regeneration of the optic nerve (Park et al., 2008; Morgan-Warren et al., 2013). More work is needed to clarify the many mechanisms underlying TRIF’s involvement in the prevention of cell apoptosis and up-regulation of GAP-43 and its pro-survival and pro-regeneration effects.

4.4 Optic nerve regeneration and recovery of visual function

An increased number of GAP-43-positive axons is indicative of better nerve regeneration in TKO mice at 30 days after injury (Kaneda et al., 2008) and supports the notion that this process is involved in functional recovery. Although recovery of visual function can be taken to indicate axon regeneration and protection against neuronal degeneration, to the best of our knowledge only a few studies have looked at the recovery of visual function in TKO mice. Damaged optic nerve results in loss of visual acuity, contrast sensitivity, and electrical responses in the visual cortex. Mice with TRIF deficiency demonstrate a higher spatial frequency threshold in the optomotor test, reflecting the recovery of retinofugal optic nerve conduction (Prusky et al., 2004).

Using VEP as a direct measurement of electrical responses in the visual cortex in mice with severe visual impairment (Tokashiki, 2018), we noticed that the amplitudes of the VEPs are higher in TKO mice in both sides of the visual cortex when the intact eye was stimulated; this can be taken to suggest that collateral sprouting and/or optic nerve regeneration may be implicated in the process (Calkins et al., 2017). One process is not exclusive of the other but the time it takes to reach minimal reorganization of function based on these parameters still remains to be determined. Indeed, point-to-point exact regeneration in the adult CNS is not as likely as injury-induced collateral sprouting to play a role in mediating functional recovery. Furthermore, although not studied in our experiments, injury-induced collateral growth of cerebral and optic nerve blood vessels could also play a role in helping to increase functional responses over time (Nishijima et al., 2015). It has been known for a very long time that collateral sprouting can play a role in visual system functional repair and the time for this to occur can vary depending on the extent of the injury, the species studied, environmental experience (Schneider, 2011) and experimental therapeutics (Sabel & Schneider, 1988).

Many studies of functional recovery from ONC typically examine visual behaviors only up to 21 or 38 days after the injury (Prilloff et al., 2010). Although axons continue to extend over time, no axonal regeneration is seen two weeks after ONC and no functional recovery is seen until ten weeks after ONC (De Lima, 2012). We observed a difference in improvement of visual function between WT and TKO mice at 80 days after ONC. It is not surprising that a longer duration of recovery time increases the chances for better visual function by allowing for more compensatory behaviors to occur as well as permitting the damaged optic nerve to recover more extensively. Under these circumstances, shorter-term studies may not reveal these important changes in visual function.

4.5 Sex differences in recovery from ONC

Although the situation is getting better, sex differences in traumatic brain injury (TBI), including optic nerve injury, are often overlooked (Wright et al., 2014). To provide new information on this subject we measured both physiological and functional differences in the recovery pattern of male and female mice after ONC injury. Mounting evidence shows that signs, symptoms and complications associated with TBI are often expressed differently in males and females (Kraus et al., 2009), although this notion is now being challenged (Shansky, 2019). Reports of sex differences in brain injury outcomes are controversial and likely depend on a number of factors that demand more attention in pre-clinical studies (Arambula et al., 2019; Gupte et al., 2019). For example, females on average experience a longer recovery time from TBI compared to males (Styrke et al., 2013). A similar effect is seen in our work, as male TKO mice show better recovery of OMR compared to female mice 80 days after optic nerve injury. Although sex differences are manifested in the morphological changes in the brain post-TBI (Geddes et al., 2016) and in some cortical functions including visual discrimination (Clark et al., 1989), we may be among the first to look at the sex differences in functional recovery from ONC. Interestingly, in addition to sex differences in recovery of visual function, we also observed greater protection against RGC loss and less inflammation in TKO male mice. This finding is contrary to the idea that females are said to do better than males in recovery from CNS injury, but it strengthens the correlation between inflammation and neuronal degeneration and also directs attention to possible gonadal hormonal effects on neuronal degeneration and inflammation that could account for some of the sex differences in recovery of vision.

4.6 Clinical significance of the TRIF pathway

Optic nerve regeneration has been associated with many transcription factors, neurotrophic factors, cell-intrinsic suppressors, and intraocular inflammation (Benowitz et al., 2017). Inhibition of the TRIF pathway, which can suppress the immune response following neural injury, has yet to be thoroughly investigated in the wound area of an optic nerve injury. Our findings could offer a potential and specific therapeutic target for visual recovery after ONC. The elimination of the TRIF gene helps to suppress the inflammatory response in the retina, leading to greater protection against RGC death and enhanced axonal regeneration. The regenerated projections of the axons are also refined and targeted in that they can restore visual function and VEPs to a certain extent. Targeting of the TRIF gene may not only be an approach to developing a treatment for damage to the optic nerve itself, but also could be relevant to the treatment of glaucoma and other CNS neuropathies involving TLRs.