[¹⁸F]DPA-714 PET Imaging Reveals Global Neuroinflammation in Zika Virus-Infected Mice

The Senegal, ZIKV strain DAK AR D 41525 obtained from the World Reference Center for Emerging Viruses and Arboviruses (R. Tesh, University of Texas Medical Branch) was used in the studies. The virus was initially amplified, passaged once each in AP61 and C6/36 cells, and then twice in Vero cells. Subsequently, for infection studies in this report, the virus was passaged once in Vero cells (ATCC, CCL-81) and fidelity was verified by sequence analysis [25].

Five-week-old female C57BL/6 mice (n = 15/group; Jackson Laboratories) were injected intraperitoneally (IP) with a total of 3.0 mg in three doses (2.0 mg first dose, 0.5 mg sustaining doses) of 5A3 (produced by Leinco Technologies, St. Louis, MO) [26, 27] or phosphate-buffered saline (PBS). The mice were injected on days − 1, + 1, and + 4 PI. In order to saturate the IFNAR-1 receptor, it was determined that a large bolus injection was required initially, i.e., 2.0 mg, followed by subsequent injections to cover the antibody half-life of 5.2 days, [26]. On day 0, mice were infected with 6.4 log₁₀ PFU of ZIKV strain DAK AR D 41525 by IP exposure route in a total volume of 200 μl. The animals were evaluated by PET on days 3, 6, and 10 PI. The mean time to death for this mouse model of ZIKV infection is reported to be 9.7 days [24].

Research was conducted under a US Army Medical Research Institute of Infectious Diseases (USAMRIID) IACUC-approved animal research protocol in compliance with the Animal Welfare Act, PHS Policy, and other Federal statutes and regulations relating to animals and experiments involving animals. The facility where this research was conducted is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care, International and adheres to principles stated in the Guide for the Care and Use of Laboratory Animals, National Research Council, 2011.

Tissue and serum samples from mice were processed as previously described [24]. Essentially, mouse serum samples were inactivated using a 3:1 ratio of TRIzol LS Reagent (Thermo Fisher Scientific, Waltham, MA). Tissues were homogenized in 1× minimum essential medium with Earle’s salts and l-glutamine (MEM) with 1 % penicillin/streptomycin and 5 % heat-inactivated fetal bovine serum (FBS-HI) using a gentleMACS dissociator (Miltenyi Biotec, San Diego, CA) followed by centrifugation at 10,000×g for 10 min, and the supernatant was stored at − 80 °C until further evaluation. Supernatant was inactivated using a 3:1 ratio of TRIzol LS. Total nucleic acid from all samples was purified using the EZ1 Virus Mini Kit v 2.0 (Qiagen, Valencia, CA) and the EZ1 Advanced XL robot (Qiagen) according to the manufacturer’s recommendations. Samples were eluted in 60 μl. Viral load was determined using a real-time RT-PCR assay specific to the 5′-untranslated region of ZIKV Standard plaque assays were completed as previously described [24] (details are provided in the Electronic Supplementary Material (ESM)).

[¹⁸F]DPA-714 was obtained from the Imaging Probe Development Center, NIH, Rockville, MD and was produced by employing slight modifications to procedures already reported [28] and using a commercially available GE TRACERLab FX-N Pro synthesizer [29]. Ready-to-inject, > 99 % radiochemically pure [¹⁸F]DPA-714 (formulated in physiological saline containing ~ 10 % ethanol) was obtained with 30–40 % (n = 12) non-decay-corrected yields and the specific activity at the end of the 70-min radiosynthesis ranged from 48 to 200 GBq/μmol.

On days 3, 6, and 10 post-infection, mice (n = 5/group/day) were bled for determination of viremia by qRT-PCR and then PET/CT scanned. PET/CT scanning was performed using an Inveon preclinical PET/CT system (Siemens Medical Solutions, Knoxville, TN) with a spatial resolution of ~ 1.5 mm full width at half maximum at the center of the field of view. Mice were anesthetized with isoflurane (4 % induction, 2 % maintenance) delivered in oxygen. Once anesthetized, animals were administered ~ 8.0 MBq of [¹⁸F]DPA-714 in ~ 150 μl volume via IV tail vein injection and were allowed to wake in their home cages for a 60-min radiotracer uptake phase. Just prior to scanning, mice were reanesthetized with isoflurane and PET imaging was performed (details are provided in the ESM).

All image reconstructions were performed using Siemens’ Inveon Acquisition Workplace v2.0 software package (Siemens Medical Solutions, Knoxville, TN) (details are provided in the ESM). PET images were coregistered to corresponding CT data using VivoQuant v2.5 image processing software (inviCRO, LLC, Boston, MA) and were subsequently coregistered to a 3D mouse brain atlas (included in VivoQuant software package) so that brain [¹⁸F]DPA-714 biodistribution could be quantified. PET imaging data were reported in terms of percent injected dose per gram of tissue (%ID/g), calculated as a ratio of tissue radioactivity concentration (Bq/g) at time of scan to total injected activity (Bq) at time of scan.

Mouse brains were collected at necropsy and fixed by immersion in 10 % neutral buffered formalin for at least 2 days. Brains were subsequently trimmed and processed according to standard protocols [30]. Brain sections were cut at 5 to 6 μM on a rotary microtome, mounted onto glass slides, and stained with hematoxylin and eosin (HE). Histological examination was performed unblinded by a board-certified veterinary pathologist.

In situ hybridization was performed using RNAscope® 2.5 HD RED kit according to the manufacturer’s recommendations (Advanced Cell Diagnostics, Hayward, CA) (details are provided in the ESM). The slides were counter-stained with hematoxylin, air dried, and mounted.

Formalin-fixed, paraffin-embedded mouse brain sections on slides were processed and stained for Iba-1 (details are provided in the ESM). Images were captured on a Zeiss LSM 700 confocal system and processed with Zen 2011 software. A LI-COR-Odyssey CLx (LI-COR Biosciences) scanned sections at 42 μm/pixel resolution. The average intensity of Iba-1 on each slide was obtained from fields-of-interest drawn around each section with the LI-COR-Odyssey analysis software, and two slides per brain were scanned. Negative control staining, for which the primary antibodies were omitted, showed no detectable labeling in immunofluorescence or infrared (IR) imaging.

Statistical analysis was carried out using SAS/STAT® software, version 9.4 (SAS Institute Inc.). All values are means ± SD, with a testing level (α) of 0.05 and adjusted p values as indicated. p < 0.05 signifies statistical significance. One-way ANOVA with post-hoc Dunnett’s one-tailed t tests were used to compare each treatment group by day with historical control data. One-way ANOVA with post-hoc Tukey-Kramer tests were used to compare between treatment groups by day.

Prior to PET scanning at days 3, 6, and 10 post-infection, mice designated for imaging from each treatment group were bled and viral titers in sera were determined by qRT-PCR (Fig. 1a). For the PBS + 5A3 and ZIKV + PBS treatment groups, no viremia was detected in any mouse at any time point tested. In the group susceptible to ZIKV infection (ZIKV + 5A3), average viremia was highest (~ 5.5 log₁₀ PFUe/ml; n = 5) at day 3 PI and decreased at both days 6 (~ 4.2 log₁₀ PFUe/ml; n = 5) and 10 PI (~ 3.0 log₁₀ PFUe/ml; n = 4). These data demonstrate a productive infection in the subjects challenged with ZIKV and treated with 5A3 with the exception of one mouse at day 10 PI.

At the conclusion of each PET scan, mice were euthanized and brains were harvested for viral titer determination (Fig. 1b). As expected, the mice in the PBS + 5A3 treatment group showed no detectable viral titers in brain at days 3, 6, or 10 post-infection. In the ZIKV + PBS treatment group, four out of the five mice tested at each time point also showed no detectable viral titers in brain, however, a single mouse at each time point tested (days 3, 6, and 10 post-infection) had viral brain titers of ~ 3.9, ~ 6.8, and ~ 4.9 log₁₀ PFUe/g, respectively, suggesting perhaps some level of viral CNS penetration and subsequent clearance with no detectable viremia. In contrast with the decreasing viremia that was measured in the susceptible ZIKV + 5A3 animals, brain titers in this group were shown to increase over time as determined by qRT-PCR from brain homogenates. From day 3 PI to day 10 PI, brain viral titers increased from an average of ~ 5.3 log₁₀ PFUe/g at day 3 PI to an average of ~ 6.95 log₁₀ PFUe/g in the remaining living mice (n = 4/5) by day 10 PI. It should be noted, however, that at days 3 and 10 post-infection, a single mouse at each time point, respectively, had brain viral titers below the limit of detection for our assay. To confirm the presence of infectious virus in the brain, qRT-PCR titers were verified by conventional plaque assay. Altogether, these data are consistent with our previous findings [24] and those of others [20‐23] and demonstrate a substantial increase in ZIKA neuroinvasion when IFN-I signaling is blocked.

To evaluate the ability of in vivo PET imaging to detect ZIKV-induced neuroinflammation in mouse brain, we used [¹⁸F]DPA-714, a PET radiotracer that is specific for TSPO [11], a biochemical marker of neuroinflammation that is highly upregulated in activated microglia, CNS macrophages, and reactive astrocytes [12‐19]. By as early as day 3 PI, ZIKV-infected mice treated with MAb-5A3 showed a modest but significant (approximately twofold) increase in mean whole-brain [¹⁸F]DPA-714 binding (2.21 ± 0.14 %ID/g; n = 5) compared with historic PBS controls (Fig. 2a, b). By days 6 and 10 post-infection, the increase in mean whole-brain [¹⁸F]DPA-714 binding in ZIKV + 5A3 mice was approximately four- (4.48 ± 1.35 %ID/g; n = 5) and sixfold (5.70 ± 3.19 %ID/g; n = 3), respectively, compared with historic PBS controls. The changes in whole-brain [¹⁸F]DPA-714 binding was highly correlated with brain viral titer, i.e., r = 0.716. Based on brain subregion analysis, all brain regions evaluated (olfactory bulb, striatum, hippocampus, thalamus, hypothalamus, cortex, and cerebellum) were increased similar to the whole brain (details are provided in Table 1 of the ESM). Of note, one of the three mice from the ZIKV + 5A3 group at day 10 PI had [¹⁸F]DPA-714 binding (2.25 %ID/g) similar to the ZIKV + PBS (2.53 ± 0.41 %ID/g, n = 5) and PBS + 5A3 (1.97 ± 0.37 %ID/g, n = 5) control groups at day 10 PI, which reduced the mean whole-brain [¹⁸F]DPA-714 binding for this group. Moreover, viral RNA was not detected in the brain of this mouse at day 10 PI (Fig. 1b). This subject was not excluded from our analyses because viremia was not assessed in this subject at day 3 and/or day 6 PI allowing for the possibility that this subject was infected and cleared the virus prior to CNS penetration, i.e., it recovered from the neuroinflammatory disease prior to day 10 or it never became infected.

ZIKV alone (ZIKV + PBS) showed a similar modest (2.25 ± 0.11 %ID/g; n = 5) but significant increase in mean whole-brain [¹⁸F]DPA-714 binding at day 3 PI relative to historic PBS controls. Mean whole-brain [¹⁸F]DPA-714 binding in ZIKV + PBS remained at modest levels at day 6 PI (2.51 ± 0.30 %ID/g, n = 5) and returned to the level of historic PBS controls by day 10 PI suggesting that ZIKV infection in the presence of IFN signaling is sufficient to produce a mild yet transient neuroinflammatory effect measureable by [¹⁸F]DPA-714 PET imaging. Importantly, 5A3 treatment in the absence of ZIKV had no detectable effect on mean whole-brain [¹⁸F]DPA-714 uptake and shared a similar tracer uptake profile as the historic PBS control mice at all three time points tested. These results demonstrate for the first time the ability of [¹⁸F]DPA-714 PET imaging to detect and quantify ZIKV-related neuroinflammation disseminated throughout the brains of infected mice.

To confirm the in vivo findings of ZIKV-induced neuroinflammation as determined by [¹⁸F]DPA-714 PET imaging, brain sections from the PET imaged mice were assessed for pathology, in situ hybridization (ISH) to detect viral RNA, and IR and immunofluorescent detection of Iba-1, a microglia/macrophage-specific calcium-binding protein that is highly upregulated during neuroinflammation [31]. Consistent with our imaging results, brains from the susceptible ZIKV + 5A3 treatment group exhibited histopathological lesions consistent with encephalitis with minimal microgliosis beginning as early as day 3 PI (Fig. 3a). By days 6 and 10 PI, microgliosis was more pronounced, with the presence of mononuclear cell perivascular cuffs peaking at day 6 PI and waning by day 10 PI (Fig. 3b, c). Other histological findings consistent with encephalitis include minimal necrotic cellular debris scattered throughout the parenchyma (days 6 and 10 PI) and neuronal degeneration and necrosis observed in the cerebrum as early as day 6 PI and becoming more pronounced and widespread on day 10 PI (cerebrum, hippocampus, and thalamus). Additionally, variable minimal perivascular edema and hemorrhage was present on both days 6 and 10 PI. Moreover, ZIKV RNA was detected by ISH in the cerebral cortex on days 6 and 10 PI and in the hippocampus on day 10 PI in mice administered ZIKV + 5A3 (Fig. 3e, f). No ZIKV RNA was detected on day 3 PI in mice administered ZIKV + 5A3 (Fig. 3d). Mice administered ZIKV + PBS exhibited no signs suggestive of ZIKV infection on day 3 or 6 PI; however, 1/5 animals exhibited minimal lesions consistent with encephalitis (perivascular cuffing, microgliosis) on day 10 PI that may be attributable to ZIKV infection, though no viral RNA was detected in this animal by ISH. As expected, animals administered 5A3 + PBS exhibited no CNS lesions suggestive of ZIKV infection and no viral RNA was detected by ISH at any of the time points tested.

In terms of Iba-1 expression, no statistical difference was observed by IR imaging at days 3 and 6 PI for any of the treatment groups tested, PBS + 5A3, ZIKV + PBS, or ZIKV + 5A3 (Fig. 4a, b, first two rows). At day 6 PI, however, whole-brain Iba-1 expression in the susceptible ZIKV + 5A3 group is modestly, though not significantly, increased. By day 10 PI, and in agreement with the [¹⁸F]DPA-714 findings, Iba-1 expression was significantly increased (approximately fourfold) in the ZIKV + 5A3 group compared with either of the control groups (Fig. 4a, b, bottom row). These findings were confirmed by immunofluorescent labeling of Iba-1 in these brains (Fig. 4c, d). Importantly, the single animal in the ZIKV + 5A3 treatment group at day 10 post-infection that showed no appreciable Iba-1 labeling by IR imaging was the same animal that at day 10 PI showed no detectable virus in the brain (Fig. 1b) and no significant whole-brain [¹⁸F]DPA-714 binding over controls (Fig. 2a, b). These data are consistent with and serve to confirm our in vivo PET imaging results. These results also demonstrate the ability of in vivo [¹⁸F]DPA-714 PET imaging to provide a more sensitive measure of ZIKV-induced neuroinflammation than Iba-1 IR or immunofluorescent imaging.