Targeted Replacement (TR) E3 (Sullivan et al., 1997) and E4 (Knouff et al., 1999) mice crossed with NL-G-F mice (Saito et al., 2014) and backcrossed to only contain human APP and human apoE (Holden et al., 2022; Kundu et al., 2022) were used for the current study (n = 44 mice; n = 15 NL-G-F/E3 mice (n = 7 PBS; n = 8 KUNV; n = 5 males; n = 10 females) and n = 29 NL-G-F/E4 mice (n = 14 PBS; n = 15 KUNV; n = 4 males; n = 25 females). Mice were implanted with temperature sensors. TS100 millimeter-scale (7.5 × 7.5 × 4.2 mm) CubiSensTM wireless sensors (CubeWorks, Ann Arbor, MI), packaged in bio-compatible epoxy and coated with parylene, were implanted under the skin for accurate, real-time temperature measurement. The TS100 is capable of transmitting up to 100 m in distance, last up to 2 years in sensing operation, and allows measuring circadian body temperature in group-housed mice, including during the 4-week period following inoculation we are not allowed access to the mice in the ABSL-2 facility. The sensors were sterilized using the Cidex solution (CubeWorks, Ann Arbor, MI). A heating pad and bead sterilizer were used for the surgeries. For the surgery, the mice were anesthetized with Isoflurane (4% for induction of the anesthesia and 1–3% for maintenance of the anesthesia). Lidocaine (6 mg/kg of 0.5%) was injected subcutaneously around the incision site, immediately prior to the aseptic preparation of the abdomen. To close the skim, 9 mm AUTOCLIP stainless steel clips were used. For pain control, Meloxicam (10 mg/kg) orally prior to the induction of anesthesia and every 24 h for two additional days was used. The mice were treated as indicated below. Body temperatures were acquired and analyzed for prior and during the behavioral testing. Due to technical issues, in NL-G-F/E3 mice no data were saved for days 19–22 and 29–30, and in NL-G-F/E4 mice on day 22. Homozygous breeding of the mice was used to generate the experimental mice for this study. Throughout testing, all the mice were group-housed with mice in the same cage receiving the same treatment. Animals were maintained on a 1,200 h light/dark schedule (lights on at 06:00). Laboratory chow (PicoLab Rodent diet 20, # 5053; PMI Nutrition International, St. Louis, MO, USA) and water were provided ad libitum. Behavioral testing took place during the light cycle. All procedures complied with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and with IACUC approval at Oregon Health & Sciences University. Experimenters were blinded to the genotype, sex, and treatment of the mice.

WNV_KUNV, abbreviated KUNV hereaffter) was obtained from BEI Resources, (NIAID, NIH: Kunjin Virus, MRM 16, NR-51653). Viral stocks were grown on C6/36 cells, followed by centrifugation of culture supernatant through a 20% sorbitol cushion at 30,000 rpm for 1.5 h in an SW32 rotor (Beckman). Following centrifugation, virus was resuspended in 1/100 of the original volume in DMEM, and aliquots were stored at −80°C. Viral titers were determined by focus-forming assay on Vero cells. Mice were infected intra-peritoneally (i.p.) with 1,000 focus-forming units (ffu) KUNV (n = 23 mice) or PBS control (n = 21 mice). We have determined that this dose is sub-lethal in most infected mice. The mice were 13.79 ± 0.53 (PBS) and 13.39 ± 0.51 (KUNV) months of age and there was no different in age in the treatment groups in either genotype. The mice underwent behavioral testing, starting 30 days post infection (dpi), as described below.

Mice were behaviorally tested as illustrated in Figure 1. In the mornings of days 30 and 31, mice were tested for measures of activity, measures of anxiety, and spatial habituation in the open field. In the afternoon of day 31, mice were tested for spontaneous alternation in the Y maze. On days 32 and 33, the mice were tested for object recognition. On days 36 and 37, the mice were tested in the spatial Y maze. On day 38, the mice were euthanized by cervical dislocation and blood was collected in EDTA-containing tubes and the hippocampus and cortex from each hemibrain were dissected.

The mice were put in an open field enclosure (16 × 16 inches, Kinder Scientific, Poway, CA) for 10 min on two subsequent days (Figure 2). On day 3, the open field contained two identical objects for a 15-min trial. The objects were placed 10 cm from each other. The next day, one object was replaced with a novel object for a 15-min trial. Images of the actual objects used are illustrated in Figure 3B. Between trial, the arenas and objects were cleaned with 0.5% acetic. Interaction with the object was coded as object exploration (i.e., nose sniffing the objects) by hand scoring videos acquired with Noldus Ethovision software (version 17, Wageningen, The Netherlands). The percent time spent exploring the familiar object and the novel object was analyzed. Different objects were used during the first and second week of testing.

The outcome measures in the open field analyzed were: (1) distance moved in the open field in the absence and presence of objects, an activity measure; (2) the difference in the distance moved in the open field over days, habituation to the open field, a cognitive measure; (3) time spent in the center of the open field, an anxiety measure; and (4) percent time spent exploring the familiar and novel object in the object recognition test, a cognitive measure.

Activity levels and hippocampus-dependent spontaneous alternations were assessed in the Y-shaped maze from O′ Hara & Co., Ltd. (Tokyo, Japan) had raised sides (3.8 cm bottom width, 12.55 cm top width, 12.55 cm height) with plastic, opaque gray arms (37.98 cm length). the Y-maze (O′ Hara & Co., Ltd., Tokyo, Japan) in a 5-min trial. The maze was cleaned with 0.5% acetic acid between trials. Performance was recorded using Noldus Ethovision software and hand scoring used to assess the number of arm entries and the percent spontaneous alternations. The outcome measures in the Y maze were total arm entries, an activity measure, and percent spontaneous alternations, a cognitive measure.

The spatial Y-maze test was conducted using smaller and distinct Y-mazes from those used for assessment of spontaneous alternation (Harvard Apparatus, Panlab, Holliston, MA, United States). This task was conducted over 2 consecutive days. On day 1, one arm was blocked off and mice were allowed to explore the maze for 15 min. Extra-maze spatial cues were taped on all three walls of the biosafety cabinet in which the mice were being tested. On day 2, all arms were accessible, and mice were allowed to explore for a 5-min trial. Performance of the mice was tracked using Ethovision software (version 17). Digital videos of day 2 were later analyzed to measure the number of entries into and the percent time spent in the novel arm (the arm that was blocked off during day 1) out of the entries in all three arms in trial 2. The criterion for an arm entry was when all four limbs were within the arm.

Following euthanasia by quick cervical dislocation without anesthesia, trunk blood was collected in EDTA-coated tubes, and brains were dissected and RNA isolated from cortical and hippocampal tissue of one hemisphere. RNA isolated from hippocampi and cortices was analyzed by qRT-PCR for expression of the inflammatory mediators TNF-α (assay ID: Mm00443258_m1), IFN-γ (Μm01168134_m1), CXCL10 (Mm00445235_m1) and T cell receptor α (Mm01313019_g1) using predesigned primer/ probe sets (Thermofisher). RNA was extracted using TRIzol reagent (Invitrogen) according to manufacturer’s protocol. Relative expression of cytokines was determined by qRT-PCR using gene specific primer-probe sets (ThermoFisher) and normalized to β-actin mRNA expression (Mm00607939) using the ΔΔCt method (Livak and Schmittgen, 2001). Viral genomes were quantified using WNV_KUNV specific primer probes. The primer/probe sequences for KUNV detection and quantitation are:

Forward Primer 5’-AGTGGAGAAGTGGAGCGATGTT-3’

Reverse Primer 5’-CAGGCTGCCACACCAAATG-3’

Probe FAM-CATACTCTGGCAAACGA-MGB

Relative expression levels (RQ) were calculated by the ∆∆C_T method as previously described (Livak and Schmittgen, 2001). Briefly, Ct values were obtained in duplicate for the gene of interest (GOI) and β-actin in each sample. ∆C_T (Avg. GOI C_T − β-actin C_T) and ∆∆C_T (∆C_{T, Sample} − ∆C_{T, reference control sample}) were calculated for each sample. Relative quantitation = 2^-∆∆C_T.

The cortex of the other hemisphere was processed for analyses of soluble and insoluble Aβ. Cortices were processed for analyses of soluble and insoluble Aβ40 and Aβ42 levels using MyBiosource ELISA kits (catalog numbers MBS760432 and MBS268504, respectively; San Diego, MA, United States), according to the recommended guidelines in the production information sheets. To a thawed cortical tissue sample (both hemispheres), 200 μl of buffer A (phosphate-buffered saline containing a protease inhibitor tablet (cOmpleteTM, 11,836,170,001 Roche, Millipore Sigma, Burlington, MA, United States and filtered before use) was added. The tissue was homogenized using a Polytron for 10 s and subsequent a sonicator and centrifuged at 45,000 rpm for 20 min at 4°C. The supernatant was collected as the soluble fraction. The same volume of buffer A was used to loosen the pellet. The sample was centrifuged again at 45,000 rpm for 5 min at 4°C. After removing the supernatant in a separate tube. The pellet was dissolved in Buffer B (containing 6 M Guanidine H-Cl and 50 mM Tris and filtered before use) and incubated at room temperature for 1 h. After this incubation, the sample was sonicated for 20 s and the extracted pellet was centrifuged at 45,000 rpm for 20 min at 4°C. The supernatant was collected as the insoluble fraction. For each ELISA, pilot experiments were performed to determine the optimal sample dilution to assure that the optical density was within the range of the standard curve. For the analyses of insoluble Aβ levels, the tissue samples were diluted 1: 4,000. For analysis of the insoluble Aβ levels, undiluted tissues samples were used. Standard curves were generated with the same buffer dilution as the samples. The ELISAs were read at 450 nm using a SpectraMax iD5 Multi-Mode Microplate Reader (Molecular Devices, VWR 76175–474, San Jose, CA, United States). The standard curves were generated and the levels in the samples determine using GraphPad Prism software, San Diego, CA, United States). Total protein amounts in the samples were determined by BCA protein assay kit (Pierce, Thermo Scientific, catalog #23225, Waltham, MA, United States) and reading the samples at 562 nm using the iD5 Reader.

All behavioral data are reported as mean ± standard error of the mean and were analyzed using SPSS v.22 (IBM, Armonk, NY, USA) or GraphPad v.8 (La Jolla, CA, USA) software. Genotype and treatment were included as factors in analysis of variance (ANOVAs). Sex was used as a covariate in this study, as there were not sufficient mice of each sex to include sex as a factor. In case there were statistical interactions, genotypes were analyzed separately, as indicated. Repeated-measures ANOVAs were used when appropriate. For analysis of object recognition, the percent time spent exploring the familiar and novel object in each group was analyzed. For the circadian data, based on the pattern of the data, the light and dark periods were analyzed as separate analyses, with the mean body temperature in the light or dark period of each day as the repeated measure. Statistical significance was considered as p < 0.05. When sphericity was violated (Mauchly’s test), Greenhouse–Geisser corrections were used. Mice were tested in separate cohorts, each containing mice of all experimental groups. All researchers were blinded to genotype and treatment and the code was only broken after the data were analyzed.

Mice were inoculated with the KUNV subtype of WNV. Although this virus is less lethal than WNV_NY99 or other recent lineage I isolates from the U. S. or Europe, it is capable of neuroinvasion in C57Bl/6 background mice (Donadieu et al., 2013). Mice were inoculated intraperitoneally with 1,000 ffu KUNV. Subsequent experiments were performed as shown in Figure 1.

When the activity levels in the open field were analyzed over the two subsequent days, there was an overall effect of day 1 [F(1,39) = 10.583, p = 0.002] (Figure 2A). PBS-treated E3 mice (t = 3.278, p = 0.0169, paired t-test) and KUNV-infected E3 mice (t = 5.535, p = 0.0009, paired t-test) showed spatial habituation learning and moved less on day 2 than day 1. However, while spatial habituation was significant in PBS-treated E4 mice (t = 5.473, p = 0.0001, paired t-test), in KUNV-infected E4 mice, it was not (t = 1.1617, p = 0.1281, paired t-test). There was also an effect of genotype [F(1,39) = 5.809, p = 0.021], with lower activity levels in NL-G-F/E4 than NL-G-F/E3 mice. There was also a trend towards a day x genotype x treatment interaction [F(1,39) = 10.583, p = 0.056]. As the open field is a novel environment on day 1, performance on day 1 was also analyzed separately. For activity levels on day 1 in the open field, there was an effect of genotype [F(1,44) = 9.102, p = 0.004].

When time in the more anxiety-provoking center of the open field was analyzed, there was an effect of genotype [F(1,39) = 5.582, p = 0.023] (Figure 2B). For time spent in the center of the open field on only day 1, there was also an effect of genotype [F(1,44) = 10.819, p = 0.002].

The objects used in the object recognition test are illustrated in Figure 3B. When activity in the open field in the presence of the two objects was analyzed, there was a day x genotype interaction [F(1,39) = 14.864, p < 0.001] (Figure 3A). In NL-G-F/E3 mice, there was an effect of treatment [F(1,12) = 5.472, p = 0.037], with lower activity in KUNV-infected than PBS-treated NL-G-F/E3 mice. In addition, there was an effect of sex [F(1,12) = 6.000, p = 0.031], with slightly higher activity levels in female (mean distance moved: 2512 cm) than male (mean distance moved: 2451 cm) mice. In NL-G-F/E4 mice, there were no significant effects.

When time spent in the center of the open field containing objects was analyzed, there was a trend towards a day x genotype interaction [F(1,39) = 3.614, p = 0.065] (Supplementary Figure S1).

PBS-treated NL-G-F/E3 mice showed novel object recognition and spent significantly more time exploring the novel than familiar object (t = 2.153, p = 0.0262; Figure 3B). In contrast, KUNV-infected NL-G-F/E3 mice spent significantly more time exploring the familiar than novel object (t = 2.431, p = 0.0291; Figure 3B). PBS-treated NL-G-F/E4 mice also spent significantly more time exploring the familiar than novel object (t = 3.206, p = 0.0036; Figure 3B). However, KUNV-infected NL-G-F/E4 mice did not a preference for exploring either the novel or familiar object (Figure 3B).

When the number of entries was analyzed as activity measure, there were no significant effects (Figure 4A). However, when spontaneous alternation was assessed in the Y maze, there was an effect of treatment [F(1,44) = 4.785, p = 0.035], with less spontaneous alternation in KUNV infected—than PBS-treated mice (Figure 4B). There was also an effect of sex [F(1,44) = 5.487, p = 0.024], with slightly higher spontaneous alternation in females (61%) than males (57%).

When the time spent in the novel arm of the spatial Y maze was analyzed, there were no significant effects (Figure 5A). When the percent entries in the novel arm when analyzed, there were also no significant effects (Figure 5B).

The circadian body temperatures are illustrated in Figures 6A–H. The light and dark periods for each were analyzed as separate analyses, with the mean body temperature in the light or dark period of each day as the repeated measure. For each panel, we first analyzed all groups together. Based on the genotype effects and interactions with genotype, we followed up with analyses of each genotype. The detailed statistical analyses are described in the Supplementary data. The body temperature was higher in NL-G-F/E4 than NL-G-F/E3 mice. The effect of KUNV on body temperature was more profound in NL-G-F/E3 than NL-G-F/E4 mice, with higher body temperatures in PBS- than KUNV-infected NL-G-F/E3 mice. This effect was most pronounced the week following inoculation, D11-D16, with the effect becoming more pronounced over subsequent days.

Following behavioral testing, insoluble and soluble cortical levels of Aβ40, Aβ42, and the Aβ42/40 ratio were analyzed. In NL-G-F/E3 mice, there was a trend towards higher insoluble Aβ40 levels in the cortex following KUNV than PBS treatment (t = 2.093, p = 0.0626; Figure 7A). This was not seen in NL-G-F/E4 mice. For cortical soluble Aβ40 levels, there was an effect of genotype (F = 4.609, p = 0.0431), with higher cortical soluble Aβ40 levels in NL-G-F/E4 than NL-G-F/E3 mice (Figure 7B).

In NL-G-F/E3 mice, insoluble cortical Aβ42 levels were also higher following KUNV than PBS treatment (t = 2.710, p = 0.0219; Figure 7C). This was not seen in NL-G-F/E4 mice. For cortical soluble Aβ42 levels in NL-G-F/E3 mice, there was a trend towards lower levels following KUNV than PBS treatment (t = 2.417, p = 0.0573). This was not seen in NL-G-F/E4 mice.

For the cortical insoluble Aβ40/40 ratio, there was an effect of genotype (F = 6.448, p = 0.0187), with a lower cortical insoluble Aβ42/40 ratio in NL-G-F/E4 than NL-G-F/E3 mice (Figure 7E). For the cortical soluble Aβ42/40 ratio, there also was an effect of genotype (F = 5.140, p = 0.0336), with a lower cortical soluble Aβ42/40 ratio in NL-G-F/E4 than NL-G-F/E3 mice (Figure 7F).

Next, we assessed viral loads in the cortex and hippocampus 7 weeks after KUNV inoculation. In the cortex, viral load was detected in more NL-G-F/E4 than NL-G-F/E3 mice (Figure 8). In the cortex of KUNV-infected NL-G-F/E3 mice, viral load was detected in only 2 out of the 12 mice, while in the cortex of KUNV-infected NL-G-F/E4 mice, viral load was detected in 11 out of the 21 mice (p = 0.034, 2-sided Chi-square test). In the hippocampus of KUNV-infected NL-G-F/E3 mice, viral load was detected in 2 out of the 11 mice, while in the hippocampus of KUNV-infected NL-G-F/E4 mice, viral load was detected in 9 out of the 21 mice but this genotype difference did not reach statistical significance.

Transcripts of four immune measures were analyzed in the cortex: IFN-γ, CXCL10, TCR-α, and TNF-α. First, the transcript levels of these four immune measures in cortex were analyzed in PBS-treated NL-F-G/E3 and NL-G-F/E4 mice and KUNV-infected NL-F-G/E3 and NL-G-F/E4 mice that did not show detectable viral loads at seven weeks post inoculation. No significant effects were seen for IFN-γ (Figure 9A) or CXCL10 (Figure 9B). However, for TCR-α there was an effect of treatment [F(1,37) = 4.115, p = 0.0497] and a genotype x treatment interaction [F(1,37) = 7.304, p = 0.0103] (Figure 9C). TCR-α transcript levels were higher in KUNV-infected NL-F-G/E3 mice that did not show detectable viral loads than PBS-treated NL-G-F/E3 mice (t = 2.573, p = 0.0259). This pattern was not seen in NL-G-F/E4 mice. For TNF-α levels, there was a trend towards a genotype difference [F(1,37) = 3.207, p = 0.0815], with a trend towards higher TNF-α transcript levels in NL-G-F/E4 than NL-G-F/E3 mice (Figure 9D).

Next, we analyzed whether in KUNV-infected NL-G-F/E4 mice the cortical transcript levels of these measures were higher in mice showing a viral load in the cortex than those who did not. Transcripts of IFN-γ (t = 6.531, p < 0.0001, Figure 9E), CXCL10 (t = 4.267, p = 0.0007, Figure 9F), TCR-α (t = 3.848, p = 0.0016, Figure 9G), and TNF-α (t = 3.102, p = 0.0073, Figure 9F) were all higher in KUNV-infected NL-G-F/E4 mice showing a viral load in the cortex than those who did not. As only two KUNV-infected NL-G-F/E3 mice showed viral load in the cortex, we also combined NL-G-F/E3 and NL-G-F/E4 mice to assess whether the cortical transcript levels of these measures were higher in mice showing a viral load in the cortex than those who did not. Transcripts of IFN-γ (t = 5.398, p < 0.0001, Figure 9I), CXCL10 (t = 3.769, p = 0.0010, Figure 8J), TCR-α (t = 3.197, p = 0.0042, Figure 9K), and TNF-α (t = 3.564, p = 0.0017, Figure 9L) were all higher in KUNV-infected mice showing a viral load in the cortex than those who did not.

Relative expression of cytokines was determined by qRT-PCR using gene specific primer-probe sets (ThermoFisher) and normalized to β-actin mRNA expression (Mm00607939) using the ΔΔCt method and expressed as RQ (relative quantification).