A Novel Mouse Model for Polysynaptic Retrograde Tracing and Rabies Pathological Research

Retrograde tracing is an important method for dissecting neuronal connections and mapping neural circuits. Over the past decades, several virus-based retrograde tracers have been developed and have contributed to display multiple neural circuits in the brain. However, most of the previously widely used viral tools have focused on mono-transsynaptic neural tracing within the central nervous system, with very limited options for achieving polysynaptic tracing between the central and peripheral nervous systems. In this study, we generated a novel mouse line, GT mice, in which both glycoprotein (G) and ASLV-A receptor (TVA) were expressed throughout the body. Using this mouse model, in combination with the well-developed rabies virus tools (RABV-EnvA-ΔG) for monosynaptic retrograde tracing, polysynaptic retrograde tracing can be achieved. This allows functional forward mapping and long-term tracing. Furthermore, since the G-deleted rabies virus can travel upstream against the nervous system as the original strain, this mouse model can also be used for rabies pathological studies. Schematic illustrations about the application principles of GT mice in polysynaptic retrograde tracing and rabies pathological research Schematic illustrations about the application principles of GT mice in polysynaptic retrograde tracing and rabies pathological research


Introduction
in the brain are organized to engage in behavioral functions (Hannah and Aron 2021;Tang et al. 2021). Viral tools, with their replication-competent, neurotropism, and transmission properties, are widely used to trace neural circuits (Oh et al. 2014;Pi et al. 2013;Tye et al. 2011). To date, many viral monosynaptic tracers have been developed to explore direct neural projections in the central nervous system (Chatterjee et al. 2018;Kim et al. 2016;Tervo et al. 2016;Zeng et al. 2017;Zhu et al. 2020), but few polysynaptic tracers have been used for mapping indirect projections between the central and peripheral nervous systems.
Rabies virus (RABV) strictly travels upstream against the neural circuit after infection in mammals (Kelly and Strick 2000;Ugolini 1995). However, given its infectious nature and lethality, the original RABV is not ordinarily used in its original form as a polysynaptic retrograde tracer; instead, it is usually modified into a safe monosynaptic retrograde tracer (RABV-EnvA-ΔG) by deleting the glycoprotein (G) and adding the envelope protein of subgroup A (EnvA) (Wickersham et al. 2007). After expressing TVA (ASLV subgroups A, the EnvA receptor) and G in the target neurons with viral vectors, RABV-EnvA-ΔG can selectively infect the target neurons and retrogradely transmit to their firstorder presynaptic neurons. Many direct neural projections in the brain have been precisely mapped using this method (Beier et al. 2017;Rossi et al. 2020;Yang et al. 2020;Yetman et al. 2019). Moreover, researchers can easily obtain RABV-EnvA-ΔG from viral companies or institutes and use it without virulence.
In this study, we generated a novel transgenic mouse line (GT mice, Rosa26-pCAG-tdTomato-P2A-G-IRES -TVA) expressing both G and TVA throughout the body, and with GT mice, RABV-EnvA-ΔG can infect and retrogradely transmit through the nervous system as the original RABV. With combination of GT mice and RABV-EnvA-ΔG of an attenuated strain (SAD-B19), polysynaptic tracing was achieved with remarkably high labeling efficiency and low animal toxicity. Furthermore, after injection with RABV-EnvA-ΔG of a challenge virus standard strain (CVS-N2c), GT mice exhibited the same course of rabies. Therefore, GT mice are a safe model for both polysynaptic tracing and rabies pathological research.

Mice
Mice were bred and reared under the conditions in accordance with institutional guidelines and the Animal Care and Use Committee of the animal core facility at Huazhong University of Science and Technology, Wuhan, China, and housed in groups of 3-5 mice per cage under a 12 h light-dark cycle, with lights on at 8:00 a.m., at a consistent ambient temperature (21 ± 1 °C) and humidity (50 ± 5%). Male mice at 90 ± 2 days old were used in the following experiments.
To express both glycoprotein (G) and TVA receptor in a single mouse, we created a LSL-GT transgenic mouse line by inserting a pCAG-LoxP-stop-LoxP-tdTomato-P2A-G-IRES-TVA-WPRE-pA targeting vector (Supplementary sequence) into the Rosa26 locus of C57BL/6-derived embryonic stem (ES) cells and injecting the targeted ES cells into C57BL/6 blastocysts. Then the LSL-GT mice were sequentially crossed with Dppa3-IRES-Cre mice (Shanghai Model Organisms Center, Inc. Shanghai, China) and wildtype (WT) mice, GT mice (Rosa26-pCAG-tdTomato-P2A-G-IRES-TVA) were obtained. For genotyping, DNA was produced from tail snips and progeny were genotyped by PCR analyses. All primers are listed (Supplementary Table 1). The body fluorescence of GT mice was checked with a stereo-fluorescence microscope (V16, Zeiss).

Virus
The virus, either SAD-B19-EnvA-ΔG-GFP or CVS-N2c-EnvA-ΔG-GFP, was purchased from BrainVTA (Wuhan) Co., Ltd. Wuhan, China, and injected into the GT mice. After injection, mice were housed 1 mouse/cage. All experiments were approved by the Institutional Biosafety Committee and performed in the biosafety level 2 laboratory or animal biosafety level 2 facility following the standard procedures and biosafety guidelines.

Western Blots
Whole brain tissues from GT mice, LSL-GT mice, and their respective controls were quickly isolated after decapitation and stored at -80 °C. Samples were homogenized in protein lysis buffer consisting of the protease inhibitor cocktail (Roche) and the protein concentration was measured by using a BCA kit (Pierce, Rockford, IL). The proteins were separated by 12% SDS-PAGE and transferred onto nitrocellulose membranes, which were then incubated with antibodies against tdTomato (1:500, Takara, 632,496), glycoprotein (G) (1:1000, Abclonal, a customer service), and TVA receptor (1:1000, Abcam, ab271292) overnight. Membranes were washed with PBST buffer and incubated with the appropriate secondary antibodies (1:1000, Invitrogen), and then the protein signals were imaged with an infrared imaging system (Odyssey, LI-COR).

Imaging
Mice were anesthetized with intraperitoneal injection of chloral hydrate and transcardially perfused with cold saline (0.9% w/v NaCl), followed with 4% paraformaldehyde (PFA). Brains were extracted and postfixed in 4% PFA at 4 °C for 24 h, and then cryoprotected with 15% and 30% sucrose in 4% PFA. Coronal brain sections (30 μm) were sliced (Leica Microsystems, Wetzlar, Germany) and washed with 0.1% Triton X-100 in PBS. Then, the brain sections were incubated with DAPI or NeuroTrace (1:200, Thermo Fisher, N21480), followed by rinsing, drying, and coverslipping with antifade mounting medium. The brain sections were imaged on a confocal microscope (LSM800, Zeiss). All the brain regions mentioned in this study were referenced to the Paxinos and Franklin's The Mouse Brain in Stereotaxic Coordinates (the 5th Edition), and the abbreviations for each brain region are summarized (Supplementary Table 2).

Behavioral Tests
Behavioral tests including the open field, elevated plus maze, rotarod, and Morris water maze ( Supplementary  Fig. 3) were performed as previously described (Li et al. 2018(Li et al. , 2021a(Li et al. , b, 2022. All tests were started when the mice were 3 months old and were allowed to acclimatize in the test room for 1 h before test commencement. The test instruments were cleaned with 75% ethanol and each animal was ventilated for 3 min after testing. The behavioral tests of open field, elevated plus maze, and Morris water maze were recorded by a digital camera and analyzed with the video analysis system (TopScan, Noldus, Holland).
Open field: the spontaneous movement was measured of 10 min in a transparent box (60 cm × 60 cm square, 50 cm height). Mice were placed in the center of the open field chamber and allowed to freely move. The traveled distance was recorded to evaluate the locomotor activity, and the times entering the center and the time spent in the center were used to assess the anxiety-like behavior.
Elevated plus maze test: the apparatus was comprised of two closed arms and two open arms elevated of 55 cm off the ground, each arm was 30 cm long and 5 cm wide, and the closed arms were enclosed by 20 cm high walls. Mice were individually placed in the intersection of the four arms, facing one open arm, and allowed to explore for 5 min. The times of entering the open arms and the time spent in the open arms were assessed to assess anxiety-like behavior.
Rotarod tests: mice were trained to perform the task for 6 consecutive days. After the training, mice were subjected to accelerated rotarod testing (uniformly accelerated to 40 rpm within 5 min) three times. The latency time before falling off the expediting rotarod was recorded.
Morris water maze: a 1.5 m-diameter swimming pool was filled with water and nontoxic white ink with the temperature of 25 °C. Each mouse was given four trials (up to 90 s for each trial) a day for 6 consecutive days. The time spent to find the platform (latency) was recorded. One day after the training session, mice were given a probe trial. The platform was removed during the probe trial and mice could swim freely for 60 s in the pool. The time spent in the target quadrant was recorded.
For CVS-N2c-EnvA-ΔG-GFP infected group, the GT mice were adapted to an open field chamber (60 cm × 60 cm square, 50 cm height) 20 min per day for 5 consecutive days before the virus injection. The locomotor activity and body weight were measured every 12 h (8:00 a.m. and 8:00 p.m.) after the viral injection. Mice were placed in the center of the open field chamber and allowed to freely move for 20 min. The total movement distance was measured as locomotor activity.

Statistical Analysis
All values in this study are represented as the mean ± SEM. Unpaired t-test and Bonferroni's post hoc test following two-way ANOVA were used when assumptions of normality and equal variance (F test) were met. Power calculations were performed using GraphPad Prism 8.3.0. Significance was accepted for *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Generation of GT Mice
To express G and TVA throughout the mouse body, we first generated a Cre-dependent G and TVA-expressing mouse line (Fig. 1a), LSL-GT mice, Rosa26-pCAG-LoxP-stop-LoxP-tdTomato-P2A-G-IRES-TVA). After crossing with Dppa3-IRES-Cre mice (Cre recombinase expressed in germ cells under the control of the Dppa3 promoter) to delete the LoxP-stop-LoxP (LSL) cassette and backcrossing with wild-type (WT) mice, we obtained GT mice in which both G and TVA were unconditionally expressed throughout the body (Fig. 1a-c). Due to the existence of the red fluorescent protein tdTomato (tdT), we could check GT mice with body fluorescence instead of genotyping (Fig. 1d).
To explore the growth and development of this new transgenic mouse, we crossed male or female heterozygous GT mice with WT mice during the breeding process and found that approximately 50% of the offspring were GT

GT Mice in Polysynaptic Retrograde Tracing
We chose an attenuated RABV strain of SAD-B19-EnvA-ΔG-GFP to verify the polysynaptic retrograde tracing of GT mice because this strain was created as a vaccine to animals and exhibited low animal toxicity (Vos et al. 1999;Wickersham et al. 2007). We injected 0.2 µl of SAD-B19-EnvA-ΔG-GFP (2.0 × 10 8 pfu/ml) into the left rectus femoris of GT mice, and the mice were sacrificed every 3 days after the injection (Fig. 2a). GFP-labeled cells in different brain regions were detected and quantified at the different time point (Fig. 2b and Supplementary Fig. 4). We first detected GFP cells in the brain 9th days after injection, and the GFP cells were sporadically distributed in the MDV and Gi of the brainstem (Fig. 2b and Supplementary Fig. 5). Over time, the number of brain regions containing GFP-labeled cells gradually increased, and we found green cells throughout mice, consistent with Mendel's law (Fig. 1e). And either male or female GT mice showed no significant differences in body size (Fig. 1f) and weight (Fig. 1g) compared to same-sex and same-month-old WT mice. Moreover, GT mice exhibited normal organ development, including brain, spinal cord, heart, lung, liver, kidney, and spleen (Supplementary Fig. 1). With Nissl staining, we also found there is no difference in the neuronal distribution pattern between GT mice and WT mice ( Supplementary Fig. 2). In addition, we performed several behavior tests to examine the motor and cognitive levels of GT mice. The data suggested that GT mice performed normally in locomotor activity, emotion, and learning and memory ( Supplementary Fig. 3).  In this study, we injected 92 GT mice with SAD-B19-EnvA-ΔG-GFP; 73 mice were infected, GFP-labeled cells were observed, and no death occurred before being sacrificed (6 to 30 days after injection). This 100% survival rate conforms to the property of SAD-B19 as an attenuated strain. Moreover, as a control, no GFP-labeled cells were detected in either the peripheral or central nervous system in WT mice with the same injection methods of SAD-B19-EnvA-ΔG-GFP. Overall, the results of the series of experiments provide strong evidence that with the monosynaptic RABV tracer, GT mice are an ideal model for polysynaptic retrograde tracing.

GT Mice in Rabies Pathological Research
We predicted that GT mice could also be a model for rabies pathological research since RABV-EnvA-ΔG-GFP can infect GT mice and transmit against the nervous system as the original RABV. To test our hypothesis, we selected another well-developed monosynaptic retrograde tracer of a challenge virus standard strain (CVS-N2c). CVS-N2c infection showed low neuronal toxicity but significant animal toxicity compared to SAD-B19 and causes typical rabies symptoms and death. This toxicity variation between neuronal and animal is speculated due to immune escape (Morimoto et al. 1998;Reardon et al. 2016). With 20 GT mice per group, 10 4 or 10 5 virions of CVS-N2c-EnvA-ΔG-GFP (0.1 or 1.0 µl, 1.0 × 10 8 pfu/ml) were injected into the biceps brachii or rectus femoris to mimic the natural infection of rabies virus in the extremities, and 10 5 virions were injected into the biceps brachii of WT mice as a control the whole central nervous system 24 days after injection, including the spinal cord, brainstem, midbrain, thalamus, hippocampus, cerebral cortex, and olfactory bulb (Fig. 2c). Among the regions, GFP-labeled cells were mainly distributed in the motor cortex, sensory cortex, CPU, VPM, and RSG, and more than 200 GFP-labeled cells were detected in these regions (Fig. 2b-c). Moreover, the neurites of the GFP-labeled neurons were clearly visible during single-cell imaging (Fig. 2d). To further confirm the tracing specificity of GT mice, we injected SAD-B19-EnvA-ΔG-GFP into the left knee joint cavity of GT mice. We evaluated the tracing by quantifying the labeled as mentioned above, and a distinct tracing pattern was demonstrated compared with the intramuscular injections ( Supplementary Fig. 6). These experiments validate that GT mice combined with a monotranssynaptic retrograde tracer can achieve specific polysynaptic retrograde tracing from the peripheral system to the central nervous system.
Likewise, there are peripheral nervous system neurons that innervate central system neurons. Here, we also examined whether GT mice could be used to achieve retrograde tracing from the central nervous system to the peripheral system. SAD-B19-EnvA-ΔG-GFP was injected into the NTS (nucleus tractus solitarii), the first site of the brain for the termination and integration of sensory information (Xie et al. 2022). 21 days after injection, GFP-labeled neurons were detected in the sensory-related nucleus of the peripheral nervous system, such as the spinal posterior horn, the nodose ganglion (NG), and the dorsal root ganglion (DRG), and we also detected sensory nerve fibers in the colon and muscle (Fig. 3). and Supplementary Fig. 7, Supplementary Fig. 8a-b, Supplementary Video 1-4). Consistent with the clinical rabies course, the latency period (from injection to morbidity) varied among each group, whereas the onset period (from morbidity to death) was remarkably consistent (Fig. 4e-f, and Supplementary Fig. 8c-d). Indeed, the infective patterns showed that the GFP-labeled cells increased from morbidity to death (Fig. 4g-i), and the infection caused the immune response only in the brain but not in the injection site and spleen . Moreover, consistent with the former research (Reardon et al. 2016), CVS-N2c infection showed low neuronal toxicity by counting the neuronal number at different time points (Supplementary group (Fig. 4a). Then, the relevant physiological parameters (body weight, locomotor activity, and survival) of each mouse were recorded every 12 h for 90 days after the infection, as the CVS-N2c strain causes paralytic-like syndrome in mice (Morimoto et al. 1998). Consistent with our expectations, there were mice that died in each GT group but no deaths in the WT group (Fig. 4b). High lethality, such as high mortality and short survival time, occurred in the injections with more virions and the upper limb (Fig. 4b-c).
After analyzing the dead mice, we found that all the mice exhibited weight loss and locomotor disability before death. Morbidity and paralysis were, respectively defined as locomotor activity impaired to 90% and 10% (Fig. 4d, The p value was calculated by the log-rank test (Mantel-Cox, ****p < 0.0001). (c) The survival days of dead mice in the four groups. n = 13 in the Biceps-10 4 group, n = 17 in the Biceps-10 5 group, n = 11 in the Rectus-10 4 group, and n = 17 in the Rectus-10 5 group (mean ± SEM, two-way ANOVA, **p < 0.01, ****p < 0.0001). (d) The variation trends of body weight and locomotor activity of all dead mice. Each grid of the x-axis is 12 h and the variation trends were shown as 12 and 24 h around every event of the rabies course (mean ± SEM, the data were represented as normalized to the baseline record before viral injection). (e-f) The latency period (e, from injection to morbidity) and the onset period (f, from morbidity to death) of the four groups (mean ± SEM, two-way ANOVA, *p < 0.05, ****p < 0.0001).
(g-i) The infective patterns of CVS-N2c-EnvA-ΔG-GFP at the morbidity phase (g), the paralysis phase (h), and the death phase (i) Overall, we provide a novel mouse model (GT mice) in which glycoprotein (G) and TVA were expressed throughout the body, and by using RABV-EnvA-ΔG of different strain, both polysynaptic regrade tracing and rabies pathological research can be achieved with safety and convenience.

Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1007/s10571-023-01384-y. Fig. 11). Therefore, GT mice are also a mouse model for rabies pathological research.

Discussion
RABV-EnvA-ΔG-GFP, without infectious effects and lethality to people, can be easily obtained and has been used only in monosynaptic retrograde tracing (Reardon et al. 2016;Wickersham et al. 2007). In this study, we successfully generated GT mice in which G and TVA were conditionally expressed in the body. RABV-EnvA-ΔG-GFP can infect GT mice and transmit as the original RABV; hence, researchers can safely and conveniently utilize or explore the properties of RABV, such as polysynaptic regrade tracing and rabies pathological research.
To date, the most commonly used polysynaptic regrade tracers are derived from the pseudorabies virus (PRV). As a type of herpesvirus, PRV inevitably exhibits high toxicity to rodents, and mice injected with PRV usually die from the virus before being sacrificed (Brittle et al. 2004;Pomeranz et al. 2005). Another defect is its high speed; PRV-labeled neurons can distribute throughout the brain only tens of hours after peripheral injection, resulting in a tracing pattern that cannot be precisely mapped (He et al. 2018;Prasad and Chudasama 2013;Wee et al. 2019;Xu et al. 2022;Yao et al. 2018). In this study, when GT mice and SAD-B19-EnvA-ΔG-GFP were combined, no death occurred before being sacrificed, and GFP-labeled neurons were first detected in the brain 9 days after peripheral injection. We can precisely map the tracing patterns by sacrificing the mice within a few days, rather than having to observe immediately within hours. Therefore, this novel mouse line, GT mice, is a good mouse model for retrograde tracing.
It is essential to explore effective therapies for rabies, as tens of thousands of people lose their lives annually because of rabies yet we still have no clue about the mechanism of its lethality. (Fisher et al. 2018). GT mice showed the relevant rabies-like syndromes of the CVS-N2c strain after infection with CVS-N2c-EnvA-ΔG-GFP. Since glycoprotein (G) applies to the transmission of all strains of RABV (Dietzschold et al. 2005), GT mice could be an ideal model for rabies pathological research of different RABV strains accompanied by no infection or lethality to researchers.
Because GT mice showed normal physiological indices and there was no report about the toxicology of G and TVA in vivo (Wirblich and Schnell 2011), the strategy used in GT mice should also be applied to other animals, such as rats, pigs, and even nonhuman primates. Furthermore, we are confident that neural circuits and properties of rabies in organisms closer to human beings can be demonstrated by using the GT models of more animals.