Abstract
Background: Chronic infection with Toxoplasma gondii (TOXO) results in microcysts in the brain that are controlled by inflammatory activation and subsequent changes in the kynurenine pathway. TOXO seropositivity is associated with a heightened risk of schizophrenia (SCZ) and with cognitive impairments. Latency of the acoustic startle response, a putative index of neural processing speed, is slower in SCZ. SCZ subjects who are TOXO seropositive have slower latency than SCZ subjects who are TOXO seronegative. We assessed the relationship between kynurenine pathway metabolites and startle latency as a potential route by which chronic TOXO infection can lead to cognitive slowing in SCZ. Methods: Fourty-seven SCZ subjects and 30 controls (CON) were tested on a standard acoustic startle paradigm. Kynurenine pathway metabolites were measured using liquid chromatography-tandem mass spectrometry were kynurenine (KYN), tryptophan (TRYP), 3-hydroxyanthranilic acid (3-OHAA), anthranilic acid (AA), and kynurenic acid (KYNA). TOXO status was determined by IgG ELISA. Results: In univariate ANCOVAs on onset and peak latency with age and log transformed startle magnitude as covariates, both onset latency [F((1,61)) = 5.76; p = 0.019] and peak latency [F((1,61)) = 4.34; p = 0.041] were slower in SCZ than CON subjects. In stepwise backward linear regressions after stratification by Diagnosis, slower onset latency in SCZ subjects was predicted by higher TRYP (B = 0.42; p = 0.008) and 3-OHAA:AA (B = 3.68; p = 0.007), and lower KYN:TRYP (B = -185.42; p = 0.034). In regressions with peak latency as the dependent variable, slower peak latency was predicted by higher TRYP (B = 0.47; p = 0.013) and 3-OHAA:AA ratio (B = 4.35; p = 0.010), and by lower KYNA (B = -6.67; p = 0.036). In CON subjects neither onset nor peak latency was predicted by any KYN metabolites. In regressions stratified by TOXO status, in TOXO positive subjects, slower peak latency was predicted by lower concentrations of KYN (B = -8.08; p = 0.008), KYNA (B = -10.64; p = 0.003), and lower KYN:TRYP ratios (B = -347.01; p = 0.03). In TOXO negative subjects neither onset nor peak latency was predicted by any KYN metabolites. Conclusions: KYN pathway markers predict slowing of startle latency in SCZ subjects and in those with chronic TOXO infection, but this is not seen in CON subjects nor TOXO seronegative subjects. These findings coupled with prior work indicating a relationship of slower latency with SCZ and TOXO infection suggest that alterations in KYN pathway markers may be a mechanism by which neural processing speed, as indexed by startle latency, is affected in these subjects.