Role of Type I Interferon Signaling and Microglia in the Abnormal Long-term Potentiation and Object Place Recognition Deficits of Male Mice With a Mutation of the Tuberous Sclerosis 2 Gene

Background Tuberous sclerosis complex is a genetic disorder associated with high rates of intellectual disability and autism. Mice with a heterozygous null mutation of the Tsc2 gene (Tsc2+/−) show deficits in hippocampal-dependent tasks and abnormal long-term potentiation (LTP) in the hippocampal CA1 region. Although previous studies focused on the role of neuronal deficits in the memory phenotypes of rodent models of tuberous sclerosis complex, the results presented here demonstrate a role for microglia in these deficits. Methods To test the possible role of microglia and type I interferon in abnormal hippocampal-dependent memory and LTP of Tsc2+/− mice, we used field recordings in CA1 and the object place recognition (OPR) task. We used the colony stimulating factor 1 receptor inhibitor PLX5622 to deplete microglia in Tsc2+/− mice and interferon alpha/beta receptor alpha chain null mutation (Ifnar1−/−) to manipulate a signaling pathway known to modulate microglia function. Results Unexpectedly, we demonstrate that male, but not female, Tsc2+/− mice show OPR deficits. These deficits can be rescued by depletion of microglia and by the Ifnar1−/− mutation. In addition to rescuing OPR deficits, depletion of microglia also reversed abnormal LTP of the Tsc2+/− mice. Altogether, our results suggest that altered IFNAR1 signaling in microglia causes the abnormal LTP and OPR deficits of male Tsc2+/− mice. Conclusions Microglia and IFNAR1 signaling have a key role in the hippocampal-dependent memory deficits and abnormal hippocampal LTP of Tsc2+/− male mice.

The object place recognition (OPR) task is designed to assess spatial memory and discrimination. This test is based on a rodent's tendency to spend more time investigating a familiar object that has been moved to a new location (16,17). Previous studies showed that lesions (16,18), pharmacological antagonists (19), and genetic manipulations (20) directed at the hippocampus resulted in impairments in OPR, demonstrating a role for the hippocampus in this task (16,(18)(19)(20)(21). Recently, we demonstrated that microglia and type I interferon (IFN) play a critical role in the social memory deficits of Tsc2 1/2 mice triggered by early postnatal immune activation (22). Here, we used the OPR task to test the hypothesis that microglia and type I IFN signaling play a role in hippocampal-dependent memory deficits and abnormal hippocampal LTP of Tsc2 1/2 mice.

METHODS AND MATERIALS
The Chancellor's Animal Research Committee at the University of California Los Angeles approved the research protocols used here.

Experimental Design and Subject Details
Tsc2 1/2 Mice. We first crossed male Tsc2 1/2 mice (23) with female wild-type (WT) mice. Tsc2 1/2 male breeders were on a C57BL/6Ncrl genetic background (Charles River Laboratories, Cat. # 027). We used C57BL/6J females (JAX, Cat. # 000664) to generate experimental mice. Pregnant females were single housed and left undisturbed except for weekly cage changes. Pregnancy was determined by checking for abdominal distension. For all of the mice studied here, pregnant females were checked every day to determine the exact day when the pups were born (designated postnatal day [P] 0). Tail biopsies for genotyping were taken around P40.
To test the role of type I IFN alpha and beta in the phenotype of male Tsc2 1/2 mice, we generated mice that included both Tsc2 1/2 and Ifnar1 2/2 null mutations by crossing Tsc2 1/2 male breeders with Ifnar1 2/2 females (bred on a C57BL/6J background) followed by intercrosses. Tail biopsies for genotyping were taken around P40.
Cx3cr1 Cre -Tsc2 Flox Mice. To confirm whether microglia play a critical role in the hippocampal-dependent memory deficits that male Tsc2 1/2 mice show, we crossed Cx3cr1 Cre mice (JAX stock #025524) with Tsc2 Flox mice (JAX stock #027458) to generate mice that carry the Tsc2 mutation only in microglia. Tail biopsies for genotyping were taken around P40.

Method Details
OPR Test. Before the test, mice were handled for 8 minutes daily for 6 consecutive days. Then, in the next 2 days, they were habituated in the OPR open field (41.5 3 41.5 3 40.5 cm) for 12 minutes each day. During the training session, mice were placed back in the OPR open field, this time with 2 identical objects, and they were allowed to explore freely for 7 minutes. After either a 1-hour or 24-hour interval, mice were tested for OPR with the previously presented objects, one of them in a new location. The object in the new location during the test was counterbalanced between trials. The open field was cleaned after each session using 70% ethanol. Mice were trained or tested once per day. Sessions were video recorded and scored offline by either 1 or 2 observers, as described above. The variation between observers was normally ,2 seconds in the 7-minute exploration sessions. Exploration was counted only when the test mouse touched one of the objects with its nose. All experiments and scoring were carried out blinded to genotype and treatment condition. The objects included in this study were small bottles or containers of different shapes.
Novel Object Recognition Test. The novel object recognition test was carried out as described previously (24), 24 hours before or after OPR (to counterbalance possible order effects). During the training session, mice were placed in the novel object recognition open field (same as the OPR open field) with 2 identical objects, and they were allowed to explore freely for 7 minutes. After a 24-hour interval, mice were tested for object memory with the previously presented object and a new object. Location of the new object during the test was counterbalanced between trials. Mice were trained or tested once per day. The open field was cleaned after each session using 70% ethanol. Sessions were video recorded and scored offline as described above. The variation between observers was normally ,2 seconds in the 7-minute exploration sessions. Exploration was counted only when the test mouse touched the objects with its nose. All experiments and scoring were carried out blinded to genotype and treatment condition. The objects included in this study were small bottles or containers of different shapes.
Rapamycin Treatment. Rapamycin (5 mg/kg; LC Laboratories, Cat. # R-5000) was freshly dissolved in vehicle solution (100% DMSO; Sigma-Aldrich, Cat. # D5879-500ML) before use. Adult (4-6 months old) Tsc2 1/2 mice were treated with a single intraperitoneal injection of rapamycin (5 mg/kg) or vehicle (DMSO) daily for 5 days prior to the OPR test. Mice were tested for OPR 18 hours after the last injection of rapamycin or DMSO as well as 2 months later. Because 100% DMSO could have an impact, we monitored that possibility in our studies. We observed that the administration of 100% DMSO did not lead to deficits in social interaction in the Tsc2 1/2 mice. In addition, we did not observe any deficits in locomotion or exploration in the Tsc2 mice injected with DMSO compared with mice not injected.
PLX5622 Treatment. PLX5622 (PLX) and control rodent diet was provided by Plexxikon Inc. and formulated in AIN-76A standard chow by Research Diets. Adult (4-6 months old) Tsc2 1/2 mice were fed with PLX (1200 mg/kg; Cat. # D11100404i) or control chow (Cat. # D10001i) for 21 days (25). At day 21 of PLX treatment, mice were tested for OPR. Mice were tested again in the OPR test 2 months after the last day of PLX treatment.
Immunohistochemistry. Immunohistochemistry was performed as described earlier (26). Briefly, mice were perfused transcardially with a fixative containing 4% paraformaldehyde and cryoprotected with 30% sucrose. Coronal and sagittal brain sections 60 mm thick were incubated overnight at 4 C with monoclonal rabbit anti-Iba1 (1:1000 dilution; Wako Chemicals, Cat. # 019-19741). The sections were then incubated for 90 minutes with Alexa Fluor 488 goat anti-rabbit IgG (immunoglobulin G) (1:500 dilution; Invitrogen, Cat. # A11011). After that, the sections were incubated in DAPI (1:1000 dilution) for 15 minutes and then in phosphate-buffered saline for 14 minutes. Immunofluorescence labeling was detected with a confocal microscope. Field excitatory postsynaptic potentials (fEPSPs) in the Schaffer collateral-CA1 in the dorsal hippocampus were evoked with an FHC bipolar platinum microelectrode (9). The input-output curve was constructed by varying stimuli intensity (w 10-100 mA) while measuring the corresponding presynaptic volley and the initial fEPSP slope. LTP of Schaffer collateral-CA1 synapses was induced by a 100-Hz (1 second) tetanus.
All stimulation pulses were 100 ms in duration and at approximately one third to one half of the intensity that induces a maximal fEPSP response. Data were obtained using Multi-Clamp 700B amplifier equipped with a Digidata 1440A interface and pCLAMP 10.0 software (Axon Instruments). Digitalized signals were sampled at 10 kHz and filtered at 2 kHz. Initial fEPSP slopes after tetanic stimulation were normalized to the average baseline fEPSP slope. For statistical analyses, two-way repeated-measures analysis of variance and Sidak's multiple comparisons test were used on the average of the first 5 minutes and the last 20 minutes of recording after LTP induction.

Statistical Analysis
Mouse behavioral data are presented as mean 6 SEM and as individual data. For behavioral experiments, statistics presented in the figures were based on Student t tests. Data were also analyzed using two-way analysis of variance plus Holm-Sidak post hoc analyses with identical results. p , .05 was considered significant. p . .05 was denoted as nonsignificant. GraphPad Prism 7 software was used to perform statistical analyses and to generate graphical representations of the data.  Outline of treatment with rapamycin or DMSO and behavior approach. (B) Tsc2 1/2 male mice before rapamycin treatment (n = 13; p = .11, t = 1.64) show OPR deficits tested 24 hours after training. Tsc2 1/2 mice during rapamycin treatment (n = 12; p , .0001, t = 8.90) show normal OPR. In contrast, Tsc2 1/2 mice 2 months after rapamycin treatment (n = 9; p = .58, t = 0.55) again show OPR deficits. Data represent mean 6 SEM as well as individual data points. ****p , .0001. n.s., not significant; OPR, object place recognition; Rap, rapamycin.

Role of Microglia and IFN1 in OPR and LTP of TSC Mice
Biological Psychiatry: Global Open Science July 2023; 3:451-459 www.sobp.org/GOS

Male but Not Female Tsc2 1/2 Mice Show Deficits in OPR
Male and female Tsc2 1/2 mice and their WT littermates (WT mice) were tested as adults (4-6 months old) in the OPR task ( Figure 1A). Male Tsc2 1/2 mice, but not female Tsc2 1/2 or WT mice, showed OPR deficits (no preference for the object in the novel location versus the object in the same location as during training) ( Figure 1B). All groups tested ( Figure S1A) showed no deficits and equivalent performance in the object recognition memory task (novel object recognition) (24) (Figure S1B), indicating that the OPR deficits of male Tsc2 1/2 mice are not due to deficits in object recognition. The Tsc2 gene is known to regulate mTOR (mechanistic target of rapamycin) signaling (27)(28)(29)(30). Accordingly, reductions of TSC2 cause upregulation of mTOR signaling in rodents and humans (23,31). Previously, our laboratory demonstrated that Tsc2 1/2 mice (same genotype/background that we used in this study) have upregulated mTOR signaling (9). The germline mutation that these mice have should affect all cells including the microglia. Previous results showed that an mTOR inhibitor (rapamycin) was capable of reversing multiple phenotypes observed in Tsc2 1/2 mice (9,32,33). Moreover, our laboratory demonstrated that rapamycin treatment effectively reduces hippocampal p-S6 (on Ser235 and Ser236) in both WT and Tsc2 1/2 mice (9). To test whether the inhibition of mTOR signaling in adult male Tsc2 1/2 mice can reverse their OPR deficits, we treated those mice with rapamycin (5 mg/kg) or vehicle (DMSO) as adults (Figure 2A) for 5 days. Before treatment with rapamycin, Tsc2 1/2 mice showed deficits in OPR ( Figure 2B), while Tsc2 1/2 mice treated with rapamycin show normal OPR (preference for the object in the novel location) ( Figure 2B). However, 2 months after rapamycin treatment, Tsc2 1/2 mice again show OPR deficits ( Figure 2B). These results demonstrate the critical role of mTOR signaling in OPR deficits of adult Tsc2 1/2 male mice and show that a 5-day treatment of rapamycin can only temporarily rescue OPR deficits of Tsc2 1/2 male mice.

Adult Depletion of Microglia Results in the Permanent Rescue of OPR Deficits of Male Tsc2 1/2 Mice
Previous results from our laboratory (22) showed that microglia play a critical role in the social memory deficits of Tsc2 1/2 mice given early postnatal immune activation. Therefore, we tested whether depletion of microglia could also reverse OPR deficits of Tsc2 1/2 male mice. Inhibition of CSF1R leads to the depletion of approximately 99% of all microglia brain-wide (34) ( Figure 3B). Thus, we treated male Tsc2 1/2 mice with a CSF1R (B) IBA1 immunostaining of Tsc2 1/2 control mice, PLX mice, and mice 2 months after PLX. Treatment with PLX led to elimination of microglia in the whole brain (hippocampus shown as example) compared with the control group (hippocampus shown as example). Two months after PLX, the microglia had repopulated the brain (hippocampus shown as example). (C) Tsc2 1/2 /PLX mice (n = 8; p , .0001, t = 6.50), but not Tsc2 1/2 /control mice (n = 10; p = .73, t = 0.33), show normal OPR memory tested 24 hours after training. Role of Microglia and IFN1 in OPR and LTP of TSC Mice inhibitor (PLX chow) or vehicle control chow ( Figure 3A) for 21 days and then tested them for OPR. Treatment with PLX rescued the OPR deficits of male Tsc2 1/2 mice ( Figure 3C). Remarkably, we also found that this treatment results in a permanent rescue of OPR in male Tsc2 1/2 mice, because the mice showed normal OPR when they were tested 2 months after microglia depletion ( Figure 3D). Histological studies demonstrated that, at this time, microglia had already repopulated the brain of these mice ( Figure 3B). These results demonstrate a role for microglia in OPR deficits of male Tsc2 1/2 mice.
As we mentioned above, our laboratory has previously shown that early postnatal immune activation reveals social memory deficits in Tsc2 1/2 mice (Tsc2 1/2 Ep) (22). Thus, we determined whether the depletion of microglia is still capable of rescuing OPR deficits of Tsc2 1/2 Ep adult mice (4-6 months old). Remarkably, treatment with PLX also rescued OPR deficits of male Tsc2 1/2 Ep mice ( Figure S2B), and this rescue persisted for at least 2 months ( Figure S2C).

The Tsc2 1/2 Mutation Restricted to Microglia During Development, but Not in Adults, Results in OPR Deficits
To test whether the Tsc2 1/2 mutation restricted to the microglia (35) is sufficient to cause OPR deficits, we used a mouse model with a floxed Tsc2 allele specifically deleted in microglia by a Cre recombinase expressed from a microglia specific promoter (Cx3cr1; Cx3cr1 Cre -Tsc2 Flox ) (35). We tested those mice as adults (4-6 months old) for OPR ( Figure 4A). Similar to germline Tsc2 1/2 male mice, male Cx3cr1 Cre -Tsc2 Flox mice, but not control mice, showed OPR deficits ( Figure 4B).
The results presented above ( Figure 3D; Figure S2C), showed that depletion of microglia permanently rescue OPR deficits of male Tsc2 1/2 mice. Therefore, once microglia are depleted in adult mice, the Tsc2 1/2 mutant microglia that repopulate the adult brain of these mice (after PLX treatment in adults) are not able to trigger OPR deficits. This suggests that the Tsc2 mutation, specifically during development, is critical for OPR deficits of male Tsc2 1/2 mice.

Microglia Depletion Reverses Abnormal LTP of Male
Previous studies suggested that abnormal hippocampal CA1 LTP may underlie the spatial memory deficits of Tsc2 1/2 mice (9,13). The results presented above demonstrated that microglia play a critical role in the OPR deficits of male Tsc2 1/2 mice ( Figure 3C, D; Figure S2B, C). We recently demonstrated that poly(I:C) given early postnatally triggers social memory deficits (also hippocampal dependent) in Tsc2 1/2 mice (36). To test if microglia are also responsible for the abnormal LTP of Tsc2 1/2 mice, we treated adult male Tsc2 1/2 Ep mice with PLX or control chow, and 4 months later, we tested their LTP ( Figure 5A). We tested Tsc2 1/2 Ep mice, and not Tsc2 1/2 mice, because the hippocampal memory deficits of Tsc2 1/2 Ep mice (which also include social memory deficits) are more severe than those of Tsc2 1/2 mice (36). Remarkably, the results show that depletion of microglia rescued the abnormal LTP of male Tsc2 1/2 Ep mice ( Figure 5B-D). These results demonstrate that microglia have a critical role in the abnormal LTP of Tsc2 1/2 mice and suggest that this may contribute to their spatial learning and memory deficits.

Normal Short-term Memory for OPR in Tsc2 1/2 Male Mice
Prior results (9), as well as the LTP experiments described above ( Figure 5B-D), revealed that Tsc2 1/2 mice (with or without early postnatal immune activation) have normal LTP tested 60 minutes after induction, whereas LTP tested 3 hours after induction was abnormal. In the experiments described above (Figure 1), OPR memory was tested 24 hours after training. In addition, TSC2 affects mTOR signaling (23,31), and this signaling pathway has been implicated in memory consolidation (37). Thus, we next tested whether adult Tsc2 1/2 male mice have deficits in OPR at a time both before memory consolidation (60 minutes after OPR training) and when their LTP is normal ( Figure 5B-D). Consistent with the idea that abnormal LTP may contribute to OPR memory deficits of Tsc2 1/2 male mice, we found that these mutant mice show normal OPR when tested 60 minutes after training ( Figure 6B). When tested 24 hours after training, Tsc2 1/2 male mice once again revealed OPR deficits ( Figure 6B). These results support the hypothesis that the abnormal LTP of Tsc2 1/2 mice contributes to their OPR memory deficits. In our previous work, we determined that type I IFN plays a critical role in the social memory deficits of male Tsc2 1/2 mice (22). Here, we determined whether type I IFN signaling also has a role in OPR deficits of male Tsc2 1/2 mice ( Figure 7A). Remarkably, although male Tsc2 1/2 mice showed robust OPR deficits ( Figure 7B; as we observed in Figure 1B), male Tsc2 1/2 mice with a null mutation for the Ifnar1 gene (Tsc2 1/2 /Ifnar1 2/2 ) showed normal OPR ( Figure 7B). These results demonstrate the role of type I IFN signaling in the OPR deficits of male Tsc2 1/2 mice.

DISCUSSION
Although previous studies focused on the role of neuronal deficits in the memory phenotypes of rodent models of TSC (14,15), the results presented here demonstrate a role for microglia in these memory deficits. In addition to rescuing their OPR deficits, we show that depletion of microglia also reversed the abnormal LTP of Tsc2 1/2 mice.
Role of Microglia and IFN1 in OPR and LTP of TSC Mice targeting glial mechanisms could be a viable therapeutic approach for addressing their cognitive deficits.
Pathology studies of tubers dissected from epileptic patients with TSC revealed evidence of microglial activation (41,42). However, seizures are common in patients with TSC and may cause microglia activation themselves (43), thus making it difficult to determine whether microglia activation is a consequence of other pathological causes in TSC or a contributing cause of TSC phenotypes. For example, previous studies showed that mice with an astrocyte-specific Tsc1 mutation showed evidence of microglia activation, as well as increased microglia size and number (44). The results presented here make a strong case that abnormalities in microglia may not simply be a consequence of changes in other cell types involved in the pathophysiology of TSC, such as astrocytes and neurons, because we showed that the Tsc2 1/2 mutation restricted to microglia recreated the hippocampal memory deficits associated with animal models of TSC (7,9,36,45,46). In addition, we also showed that the hippocampal CA1 LTP abnormalities and OPR deficits of Tsc2 1/2 mice could be reversed by a manipulation that depleted most microglia from the brain of these mice. We demonstrated that a null mutation (Ifnar1 2/2 ) of a key signaling pathway in microglia (47) also prevented the hippocampal-dependent memory deficits of these mice. Previous RNA sequencing studies in patients with autism spectrum disorder (48) pointed to a lasting upregulation of microglia and IFN response pathways, suggesting that our findings with an animal model of TSC may be more generally significant.
We previously proposed (22) that Tsc2 haploinsufficiency (Tsc2 1/2 ) leads to abnormally elevated mTOR signaling (23,31) and, consequently, to increased production of type I IFN. Elevated IFN levels result in overactivation of IFNAR1, which in turn further activates mTOR signaling and triggers even more IFN production. We propose that this cycle is perpetuated in microglia, and it is responsible for the OPR and LTP deficits of male Tsc2 1/2 mice.
Our results suggest that the Tsc2 1/2 mutation during development is critical for OPR deficits: first, we showed that rapamycin treatment during early postnatal development (at Tsc2 1/2 mice tested 60 minutes after training (Tsc2 1/2 /60 min; n = 10; p , .0001, t = 6.97) show normal OPR (they explore the object in the novel location significantly more than the object in the familiar location), but Tsc2 1/2 mice tested 24 hours after training (Tsc2 1/2 /24h; n = 10; p = .73, t = 0.33) show OPR deficits (show no preference for the object in the novel location). Data represent mean 6 SEM as well as individual data. ****p , .0001. Cont. control; F, familiar; N, novel; n.s., not significant; OPR, object place recognition. Role of Microglia and IFN1 in OPR and LTP of TSC Mice Biological Psychiatry: Global Open Science July 2023; 3:451-459 www.sobp.org/GOS P3, P7, and P14) can prevent OPR deficits of Tsc2 1/2 mice (tested w4 months after this treatment) ( Figure S3), which suggests that this developmental period is critical for OPR deficits of male Tsc2 1/2 mice. Second, rapamycin treatment specifically in adult Tsc2 1/2 mice rescues their OPR deficits ( Figure 2B). Resident microglia are long lived, with a median lifetime of well over 15 months (49). Thus, approximately half of these cells survive the entire mouse life span. Therefore, changes in these cells during development could contribute to OPR deficits of Tsc2 1/2 mice. Finally, depletion of microglia in adult Tsc2 1/2 mice rescues OPR deficits ( Figure 3C), and the rescue persists after new microglia repopulate the brain in adults ( Figure 3D), which indicates that new mutant microglia do not trigger OPR deficits in Tsc2 1/2 mice, suggesting that changes in development are critical.
Altogether, the results presented here demonstrate that microglia and type I IFN signaling have a key role in the electrophysiological and memory phenotypes of an animal model of TSC, a result consistent with the hypothesis that abnormal type I IFN signaling in microglia is responsible for abnormal LTP and for hippocampal memory deficits in Tsc2 1/2 mice. These phenotypes can be reversed by depletion of microglia, a result that suggests that therapeutic strategies targeting this cell type may be a viable strategy to address cognitive impairments in TSC.