All animal care protocols were approved by Texas A&M University Laboratory Animal Care and Use Committee (AUP 2020-0065, approved May 21, 2020) and complied with NIH Guidelines for Care and Use of Laboratory Animals. Mice were group-housed under controlled standard conditions in an ambient temperature room and 12-hour light/dark cycles, with ad libitum food and water. All testing was performed during the light phase.

Female and male C57BL/6J mice were allowed to mate and deliver pups without intervention. Once pregnant, female mice were checked every morning for parturition. Day of birth was designated as post-natal day (PND) 0. Litters contained pups of both sexes during the neonatal period and no sex-based separation was made until weaning at PND 21. The quantity of mice in this study depended on the reproductive success of our in-house breeding program.

On PND 0, full litters were randomly sorted (Table 1) yielding the following six exposure groups (Figure 1): (A) PFOA control + No injection– pups were supplied with deionized drinking water and did not undergo TMEV injection procedure; (B) PFOA + No injection– pups were supplied with 70 ppt PFOA via drinking water and did not undergo TMEV injection procedure; (C) PFOA control + Sham-injected– pups were supplied with deionized drinking water and were injected with Phosphate Buffer Solution (PBS), as described below in “Virus Inoculation”; (D) PFOA + Sham-injected– pups were supplied with 70 ppt PFOA via drinking water and were injected with PBS; (E) PFOA control + TMEV-infected– pups were supplied with deionized drinking water and were injected with TMEV; (F) PFOA + TMEV-infection– pups were supplied with 70 ppt PFOA via drinking water and were injected with TMEV.

Crystallized PFOA (96% purity; Millipore Sigma Chemical Company, St. Louis, MO, USA) was dissolved by agitation in deionized water at a stock concentration of 0.07 mg/mL. PFOA-dosed drinking water was prepared by diluting the PFOA stock to a final concentration of 70 ng/L (ppt), a human relevant dose and EPA’s standard for PFOA in drinking water [41,96]. PFOA drinking water was prepared fresh bi-weekly and water bottles were refilled weekly. Mice in PFOA exposed groups were given 70 ppt PFOA in drinking water (Groups B, D, and F) or deionized water (PFOA control) (Groups A, C, and E) starting at PND 1 until the end of the study (PND 42). Pups were exposed via milk from the dams until weaning at PND 21. Mice have been shown to exhibit significant developmental effects following PFOA exposure from either drinking water or from milk (3, 43). Water consumption was measured every morning to ensure PFOA intake.

At PND 28, mice in Groups C, D, E, and F were anesthetized by isoflurane inhalation (MWI, Meridian, ID, USA) and intracerebrally (i.c.) injected into the right mid-parietal cortex at a depth of ~1.5 mm with 20 µL of 1 × PBS (Groups C and D) or with 5.0 x 104 plaque-forming units (PFU) of TMEV BeAn strain (Groups E and F; American Type Culture Collection [ATCC] VR 995, Manassas, VA, USA), as previously described (29, 44–46).

All mice were weighed and evaluated for TMEV-induced clinical signs (phenotypes) from PND 28 to 42. Phenotypes such as kyphosis, piloerection, limb clasping, and delay in righting reflex were evaluated and scored as described previously (46). Clinical signs of limb weakness and paralysis were observed and scored on a scale of 0 – 4, with a score of 0 given to mice having normal stride and no signs of weakness, and a score of 4 representing the total loss of limb mobility characterized by lack of grip function and flaccid limb extension (45). Seizures were scored between 0 – 5, based on the Racine scale, with a score of 0 given to mice having no seizures, up to a score of 5 given to mice experiencing generalized tonic-clonic seizures accompanied by falling on back and rearing (29).

At PND 25 – 26, blood was collected from the submandibular vein via puncture with a lancet at a depth of ~1 mm. At the end of study (PND 42), mice were euthanized by intraperitoneal (IP) injection of Beuthansia at 150 mg/kg (Merck & Co., Kenilworth, NJ, USA) as described previously (45, 47, 48). Thereafter, blood was collected from the right atrium and was refrigerated at 4°C for an hour. Blood was then centrifuged at 2,000 × rpm to collect sera and was then stored at –20°C for further analysis. Following blood collection, mice were perfused with a 1 × PBS solution through the left ventricle. Then, thoracic spinal cord and hippocampus were removed and flash-frozen in liquid nitrogen. These tissues were selected for analysis based on their known involvement in TMEV infection (49–51). Tissues were stored at -80°C for RNA extractions, as described below.

Cytokine and chemokine levels were measured in the sera collected at PND 25 – 26 and at PND 42. Sera collected at PND 25 – 26 was considered as the baseline timepoint when determining TMEV related responses and for comparing with levels observed at the end of the study (PND 42; 14 dpi). Immune response proteins were measured with Bio-Plex Pro™ Mouse Cytokine 23-plex Assay kit (Bio-Rad, Hercules, CA, USA) to determine concentration levels of 23 cytokines and chemokines (IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-9, IL-10, IL-12 [p40], IL-12 [p70], IL-13, IL-17α, IFN-γ, CCL11 [Eotaxin], G-CSF, GM-CSF, CXCL1 [KC keratinocyte-derived chemokine], CCL2 [MCP-1 Monocyte Chemotactic Peptide 1]), CCL3 [MIP-1α Macrophage Inflammatory Protein1α], CCL4 [MIP-1β], CCL5 [RANTES], and TNF-α). Data was processed using the Bio-Plex Manager software program (Bio-Rad version 4.1.1, Hercules, CA, USA).

Thoracic spinal cord and hippocampal samples were thawed at room temperature, and RNA extractions were performed using Maxwell-16^® automated equipment with a LEV Simply RNA tissue kit (Promega, Sunnyvale, CA, USA). RNA quality and concentration was verified using an Agilent Tape Station with all samples having a RIN score over 8.0 (Agilent Technologies, Santa Clara, CA, USA).

To detect TMEV RNA in both CNS tissues, 125 ng of total RNA was reverse-transcribed to cDNA in 20 µL reactions using Superscript II (ThermoFisher Scientific, Waltham, MA, USA) with gene-specific primers at a final concentration of 10 µM. The resulting cDNA was diluted (1:5) in nuclease-free water.

Primers were designed using the Oligo 7 Primer Analysis Software version 7.6 (Molecular Biology Insights, Cascade, CO, USA). Primers for TMEV were designed from the BeAn strain sequence (M16020.1): Forward 5’–CTG CAA ACA TGG ATA CCC AGA T–3’; Reverse 5’–GTC CAC ACA AAG AAG GTC CGT A–3’; Standard oligonucleotide length: 131 nucleotides (Integrated DNA Technologies, Coralville, IA, USA). Primer pairs were assessed via a BLAST sequence similarity search to avoid cross-amplification across different species and organ systems or between different mRNA transcripts (52). Real-time quantitative PCR (RTq-PCR) was performed according to standard protocols. The sample order was randomized, and amplifications were repeated in triplicate in a 384-well plate at a total reaction volume of 20 µL, containing 10 µL Power SYBR Green PCR Master Mix (ThermoFisher Scientific, Waltham, MA, USA), 0.6 µL of each primer pair at 10 µM, 6.8 µL nuclease-free water, and 2.0 µL of template cDNA. A standard curve was constructed using the linear range of the 10‐fold serial dilution standards ranging from 1012 to 103 copies/µL for absolute quantification. PCR reactions were run on a CFX-384 Touch Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) and analyzed using Bio-Rad CFX Maestro Software (version 1.1).

Serum samples of 160 µL were thawed for 1 min in a water bath (37°C) and mixed with 640 μL of 50% acetonitrile containing 0.2% formic acid. After centrifugation at 8000 g for 8 min, a 700 μL aliquot of the supernatant was loaded into Strata-X-AW solid phase extraction (SPE) cartridges (Phenomenex, Torrance, CA, USA) that were preconditioned with methanol and water (53). The loaded SPE cartridges were washed with 1 mL of 25 mM ammonium acetate to eliminate interferences. The analytes were eluted using 1 mL of 5% ammonium hydroxide in methanol, dried in a vacuum evaporator, and reconstituted in 5% ammonium hydroxide in 60% acetonitrile/40% methanol containing 0.05 µM of ¹³C₄-PFOA (Wellington Laboratory, Canada).

PFAS was analyzed using a Waters Acquity ultraperformance liquid chromatograph/tandem mass spectrometer (LC/MS-MS) equipped with triple quadrupole, following a previous method (54). An Acquity BEH C18 column (2.1 × 50 mm, 1.7 μm) with a VanGuard pre-column (2.1 × 5 mm, 1.7 μm) at 40°C was used for separation. 20 mM ammonium acetate (eluent A) and acetonitrile (eluent B) were carried out at 0.6 mL/min flow rate. The gradient program for elution was 10% eluent B (initial), 10%-55% (0 to 0.1 min), 55%-99% (0.1 to 4.5 min), 99% (4.5 to 5 min), and 99%-10% (5 to 6.5 min). The injection volume was 50 μL for each analysis. The mass spectrometer was operated in a negative ion mode at 4.5 kV spray voltage and 450°C source temperature. The monitored precursor and product ions (m/z) for PFOA and ¹³C₄-PFOA were 413 to 369, and 417 to 372, respectively, at 20 kV cone voltage. To ensure consistency of the detection methods and linearity of peak concentrations, PFOA standard solutions at concentrations between 0.05 and 50 ng/mL in distilled water were prepared and run with each analysis. The 7-point standard curve was linear (r² > 0.99) with a limit of detection at 0.05 ng/mL and a recovery percentage averaged 114%.

GraphPad Prism version 9.5.0 for Mac (GraphPad Software, San Diego, CA, USA) was used to perform one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test to compare individual phenotypic responses between the six exposure groups. A two-way ANOVA test with Tukey’s multiple comparisons test was used to compare daily weights per exposure group, address sex-specific differences among seizure counts, evaluate TMEV RNA expression in spinal cord tissue, identify sex-specific differences in TMEV RNA expression in CNS tissue, and detect sex-specific differences when comparing 0 ppt and 70 ppt pre-infection cytokine levels. Sidak’s multiple comparison test was implemented to identify sex-specific differences for individual phenotypes, compare seizure counts between the six exposure groups, evaluate TMEV RNA expression in hippocampal tissue, and evaluate serum cytokine levels at PND 42 for PFOA-induced changes and sex-specific differences. To evaluate the influence PFOA on pre-infection cytokine levels among the six exposure cohorts an unpaired t-test was utilized. A Pearson R correlation analysis between serum cytokine levels in Groups E and F and their TMEV RNA expression and seizure frequencies indicated the influence of PFOA on serum cytokine levels in TMEV-exposed groups. All reported p-values were based on two-tailed statistical tests, with a significance level of 0.05.

Serum collected at postnatal day (PND) 26 (before infection) and at PND 42 (end of study) was used to measure PFOA concentration levels. PFOA was measured in mice from Groups A (0 ppt + no injection) and B (70 ppt + no injection), but did not statistically differ among the two. In Group B, the average PFOA concentration on day 42 (1.60 ng/mL) was half of the level on day 26 (3.53 ng/mL), correlating with its half-life in serum (55). Low levels of PFOA were also detected in the double blank, single blank, and serum blank samples after 26 and 42 days, showing background contamination and prevalence of PFOA. While serum levels do not necessarily reflect PFOA accumulation in other organs such as the brain, these measurements confirmed the presence of PFOA in the body. Overall, we determined that a low dose of 70 ppt PFOA via drinking water can be detected in the serum via mass spectrometry (Supplementary Figure 1).

Once the mice were weaned and assigned to their respective exposure cohorts, their weights were measured daily from PND 28 (0 days post-infection; dpi) to 42 (14 dpi, the end of the acute phase of TMEV infection). To allow for direct comparisons of weight changes across the different exposure groups, weights were normalized by computing the ratios between the daily weights to the baseline weight measured at 0 dpi (Supplementary Figure 2A). Overall, B6 mice exposed to 70 ppt PFOA and without TMEV infection (Groups B and D) displayed the greatest weight gain. TMEV-infected mice exposed to 70 ppt PFOA (Groups E and F) lost weight after infection and regained their original weight by 3 dpi. Thereafter, weights of TMEV-infected mice increased slightly (1.2 × from the original weight). Finally, mice exposed to 70 ppt PFOA and no TMEV infection (Groups C and E) gained the least amount of weight, increasing their starting weight by 1.1 × from their original weight. Weights at 14 dpi were analyzed using two-way ANOVA to determine the differences between the exposure groups’ final weight changes (Supplementary Table 1). Weights of mice exposed to 70 ppt PFOA but not infected (Groups B and D) were significantly higher than those of mice exposed to 0 ppt PFOA, regardless of infection status (Groups A, C, and E). These differences and their significance were as follows: 0 ppt + no injection vs. 70 ppt + no injection or 70 ppt + sham-infected, both p < 0.0001; 0 ppt + Sham-infected vs. 70 ppt + no injection or 70 ppt + sham-infected, p < 0.001 and p < 0.0001, respectively; and 0 ppt + TMEV-infected vs. 70 ppt + no injection or 70 ppt + sham-infected, p < 0.05 and p < 0.01, respectively. In addition, the weights of mice exposed to 70 ppt PFOA + sham-infected (Group D) were significantly higher than those of mice exposed to 70 ppt + TMEV-infected (Group F), p < 0.05, suggesting that exposure to 70 ppt PFOA resulted in rapid weight gain. However, exposure to both PFOA and TMEV resulted in slower weight gain; thus, the viral infection curbed the weight gain observed in those solely exposed to 70 ppt PFOA.

To determine whether there were sex differences in weight gain, a two-way ANOVA was performed on the exposure groups. By 14 dpi, all male mice, except for those exposed to 0 ppt PFOA + not injected (Group A), had the greatest weight gain compared with their female counterparts, representing typical weight gain behavior (Supplementary Figure 2B).

A variety of neurological phenotypes resulting from TMEV exposure have been characterized in different mouse strains (29, 45, 47). B6 mice, considered a TMEV-resistant strain, were used in this study to determine whether PFOA exposure resulted in an increased incidence and/or severity in neurological clinical symptoms. The 14 dpi frequency scores were calculated by dividing the number of observations recorded for each phenotype by the total number of phenotyping events during the acute phase, which included the first 14 days post-infection (Figure 2A). All phenotypes were recorded as true observations only when the same phenotypes were not observed in control mice. For example, circling on the grate and ground was observed in multiple treatment groups, regardless of exposure, suggesting circling was a normal behavior for this strain. On the other hand, piloerection, a clinical indicator of sickness, was observed in mice exposed to 70 ppt PFOA and either sham- or TMEV-infected. Piloerection was significantly higher in mice exposed to 70 ppt PFOA + TMEV infection (Group F) than their sham-infected counterparts (Group D; p < 0.05) and mice exposed to 0 ppt PFOA + TMEV-infected (Group E; p < 0.0001). Both TMEV-infected groups (E and F) exhibited seizures at 14 dpi. However, the 70 ppt PFOA-exposed + TMEV-infected Group F mice had a significantly higher seizure frequency than mice exposed to 0 ppt PFOA + TMEV-infected (Group E; p < 0.01). Other common phenotypes observed in at least one mouse per group across the TMEV-infected groups, regardless of PFOA exposure (Groups E and F), included limb weakness and paralysis, backing up, and delayed righting reflex. Though these observations may be biologically relevant, not enough mice were affected to be considered statistically significant. Overall, mice exposed to both 70 ppt PFOA and TMEV infection displayed greater cumulative frequencies of phenotypic responses than those exposed to TMEV infection alone (Figure 2B). No sex-specific differences were found for any clinical signs except for retropulsion (p < 0.05) and delayed righting reflex response (p < 0.05) in mice exposed to 70 ppt PFOA + TMEV infection (Group F; Figure 2A).

TMEV infection resulted in seizures; however, seizures were more predominant in TMEV-infected mice exposed to 70 ppt PFOA (Group F; Figure 2A). Seizures were observed in TMEV-infected mice not exposed to PFOA (Group E) between 4 and 7 dpi; whereas for TMEV-infected mice exposed to 70 ppt PFOA (Group F), seizures started as early as 2 dpi and ended at 8 dpi. The Racine seizure scale was used for clinical scoring purposes, with numerical scores given as follows: 1 - freezing with mouth and facial movement, 2 - freezing with head nodding, 3 - sitting with forelimb clasping, 4 - rearing with forelimb clasping, 5 - rearing and falling (Figure 3). Overall, seizures scoring 1, 3, and 5 were observed more often, with a score of 3 being significantly more common (p < 0.05) in TMEV-infected mice exposed to 70 ppt PFOA (Group F) versus 0 ppt PFOA (Group E). A score of 2 trended higher in TMEV-infected but unexposed mice (Group E), while scores of 4 were similarly noted in both TMEV-infected groups.

Sex-specific differences were also recorded in relation to the types of seizures observed in TMEV infected mice (Figure 3). Male mice exposed to 70 ppt PFOA and infected with TMEV (Group F) had significantly higher counts of seizures with scores 1 and 3 than TMEV-infected male mice with 0 ppt PFOA exposure (Group E; both p < 0.05). Furthermore, male mice exposed to 70 ppt PFOA and infected with TMEV (Group F) had more seizures with scores of 3 than TMEV-infected female mice with 0 ppt PFOA exposure (Group E; p < 0.05). In contrast, female mice exposed to 70 ppt PFOA and infected with TMEV (Group F) had more seizures with scores of 5 than TMEV-infected female mice with 0 ppt PFOA exposure (Group E; p < 0.05). Overall, seizures experienced by mice exposed to both PFOA and TMEV tended to have higher scores than those in mice exposed to TMEV alone.

TMEV RNA expression was measured in the hippocampus and thoracic spinal cord, CNS tissues typically targeted by TMEV (47, 56, 57), for all mice infected with TMEV (Figure 4). Our interest in measuring TMEV RNA expression was to determine whether PFOA would suppress the immune system, resulting in higher viral RNA levels. In the hippocampus at 14 dpi (Figure 4A), TMEV-infected mice not exposed to PFOA (Group E) had, as expected, significantly higher levels of TMEV RNA than those of unexposed mice not injected or sham-infected (Groups A and C, both p < 0.01). While TMEV RNA was present in the hippocampus of TMEV-infected mice exposed to 70 ppt PFOA (Group F), differences in levels of TMEV RNA were not quite significant between not injected or sham-infected mice exposed to 70 ppt PFOA (Groups B and D, both p = 0.053). There were no significant differences in TMEV RNA expression in all TMEV-infected mice, regardless of PFOA exposure (Groups E and F), meaning both TMEV infected groups had similar levels of TMEV RNA, suggesting PFOA did not directly alter mechanisms of viral clearance.

Levels of TMEV RNA expression in the spinal cord resembled the expression found in the hippocampus (Figure 4B), as levels in TMEV-infected, unexposed mice (Group E) were significantly higher compared to those in the not injected and sham-infected unexposed mice (Groups A and C; p < 0.05 and p < 0.01). TMEV RNA expression was detected in the spinal cords of TMEV-infected mice exposed to 70 ppt PFOA (Group F), but expression levels did not significantly differ from those of the 70 ppt PFOA-exposed, not injected or sham-infected mice (Groups B and C). TMEV RNA transcript levels for TMEV-infected mice exposed to 70 ppt PFOA (Group F) were lower than those in TMEV-infected, unexposed mice (Group E), but did not significantly differ between the two groups. As expected, in both CNS tissues, the non-injected and sham-infected groups did not express TMEV RNA.

TMEV RNA expression was compared only in groups infected with TMEV (Groups E and F) between both tissues to determine the overall localization of TMEV RNA at this time point (Figure 4C). As expected, TMEV-infected, unexposed Group E had higher TMEV RNA expression in the hippocampus than in the spinal cord (p < 0.01). Although TMEV-infected, exposed to 70 ppt PFOA Group F tended to have lower TMEV RNA expression in the spinal cord, this was not significantly different from that found in the hippocampus (p = 0.084). No sex-related differences were identified in TMEV RNA expression between the different groups and tissues (Figure 4).

Serum collected on PND 25–26, before i.c. injection, was used to measure baseline cytokine and chemokine levels. Prior to PND 28, mice were only exposed to either 0 ppt or 70 ppt PFOA. For this analysis, samples were grouped according to their PFOA exposure to evaluate the potential adverse effects of 70 ppt PFOA on the immune system by identifying any disruption in cytokine and chemokine levels (Supplementary Figure 3). Ten of the 23 cytokines investigated were significantly altered by exposure to 70 ppt PFOA (Figure 5). These cytokines included Th1 and cytokines IL-1β, IL-3, IL-12 [p40], GM-CSF, IFN-γ, and KC; Th2 and cytokines IL-4, IL-10, and IL-13; and IL-6, a cytokine that regulates cytokine/chemokine products of Th1 and Th2 cells. Interestingly, most of these cytokines and chemokines had lower levels in PFOA-exposed relative to non-PFOA exposed groups, as indicated by their low concentration levels compared to the control groups. PFOA-exposed mice had higher levels of IL-6, IL-10, IL-12 (p40), and KC compared to non-PFOA exposed mice. Overall, observations included a mixed response of Th1 cytokines and a suppression of Th2 cytokines, suggesting a PFOA-induced imbalance of Th1/Th2 cytokines.

Sex-related differences in cytokine and chemokine levels were examined between the two exposure groups (Supplementary Figure 4). Lower levels of IL-13 and GM-CSF were identified in males exposed to 70 ppt compared to their unexposed male counterparts (p < 0.05) and higher levels of IL-10 in both PFOA-exposed females and males compared to PFOA control females and males (p < 0.0001 and p < 0.001, respectively). Based on this study, the sex-specific disruptions may have originated from Th2 and Th17 cytokines.

Serum cytokine and chemokine levels at PND 42 (14 dpi for TMEV-infected groups) were measured to assess the longitudinal effects of PFOA exposure and determine whether PFOA altered immune responses after viral infection (Supplementary Figure 5). Cytokines and chemokines significantly affected by exposure to 70 ppt PFOA were identified by comparing the cytokine/chemokine levels of the 0 and 70 ppt PFOA groups in relation to TMEV infection status (non-infected, sham-infected, and TMEV-infected) (Figure 6).

The greatest disruption in cytokines and chemokines occurred in non-infected Groups A and B (Figure 6A). The cytokines IL-1β, IL-2, IL-12 [p70], GM-CSF, MCP-1, and TNF-α were produced at significantly higher levels than those in not-injected, 0 ppt PFOA Group A, apart from lower IL-1α levels from not-injected, 70 ppt PFOA exposure (Group B). Here, the cytokines IL-5 and IL-13 were produced at higher levels in the 70 ppt PFOA exposure, not injected, Group B. The findings among these two groups represent the sole effects of exposure to 42 consecutive days of 70 ppt PFOA, as these mice did not undergo the intracranial injection procedure.

In the analysis between sham-infected mice, four cytokines and chemokines significantly differed between 0 ppt and 70 ppt PFOA exposure groups (Groups C and D, Figure 6B). The chemokines Eotaxin and G-CSF were produced at lower levels in sham-infected mice exposed to 70 ppt PFOA (Group D). Similarly, IL-6 was produced at lower levels. In contrast, Th1-related IL-12 (p40) was produced at higher levels than in the sham-infected mice exposed to 0 ppt PFOA (Group C). The injection procedure itself can initiate an inflammatory response at 0 dpi (as previously reported (39)); whereas at 14 dpi PBS-injected (sham-infected) mice exposed to 0 or 70 ppt PFOA (Groups C and D) demonstrated differences in levels of cytokines and chemokines which were attributable to PFOA exposure.

Previously, immune responses after TMEV infection have been characterized throughout the acute phase (39). In this study, PFOA exposure and TMEV infection (Group F) resulted in low levels of IL-10 compared to mice exposed to 0 ppt PFOA + TMEV infection (Group E). Interestingly, IL-10 levels were low in all three TMEV exposure groups. Thus, subchronic exposure to PFOA resulted in suppressed IL-10 levels, regardless of viral infection.

Sex-specific differences were evaluated in cytokine and chemokine levels among each group (Supplementary Figure 6). Males exposed to 0 ppt PFOA + sham-infected (Group C) produced higher levels of IL-6 and G-CSF, and lower levels of RANTES, compared to females of Group C. Males also produced higher levels of MIP-1β than females when treated with 0 ppt PFOA + TMEV infection (Group E).

A Pearson R correlation analysis was performed to determine whether cytokine and chemokine levels in mice infected with TMEV and exposed to either 0 ppt or 70 ppt PFOA (Groups E and F) were associated with hippocampal and spinal cord TMEV RNA expression levels (Supplementary Figure 7). As expected, no direct correlations were found. A similar analysis was performed using the cytokine and chemokine levels from both groups to assess their influences on the 14 dpi seizure frequency scores (Supplementary Figure 8). Similarly, the association strength of cytokine and chemokine levels with seizure scores was low for TMEV-infected mice exposed to 0 ppt PFOA (Group E; r = – 0.37 to 0.46). In contrast, seizure frequency scores were associated with cytokine and chemokine levels for TMEV-infected mice exposed to 70 ppt PFOA (Group F): five cytokine levels with low associations (r = – 0.22 to 0.10), six cytokine levels with moderate associations (r = – 0.69 to – 0.53, 0.61), and 12 cytokines with strong negative associations (r = – 0.80 to – 0.71). Of these, IL-12 (p70) had the strongest correlation (r = – 0.80, p < 0.05) with seizure frequency (Table 2). IL-12 (p70) stimulates the growth and function of T cells, production of TNF-α from T cells and NK cells, and IFN-γ by reducing IL-4 mediated suppression of IFN-γ, thus promoting an inflammatory environment contributing to the onset of seizures (58).