Neural Origins of Human Sickness in Interoceptive Responses to Inflammation

Background Inflammation is associated with psychological, emotional, and behavioral disturbance, known as sickness behavior. Inflammatory cytokines are implicated in coordinating this central motivational reorientation accompanying peripheral immunologic responses to pathogens. Studies in rodents suggest an afferent interoceptive neural mechanism, although comparable data in humans are lacking. Methods In a double-blind, randomized crossover study, 16 healthy male volunteers received typhoid vaccination or saline (placebo) injection in two experimental sessions. Profile of Mood State questionnaires were completed at baseline and at 2 and 3 hours. Two hours after injection, participants performed a high-demand color word Stroop task during functional magnetic resonance imaging. Blood samples were performed at baseline and immediately after scanning. Results Typhoid but not placebo injection produced a robust inflammatory response indexed by increased circulating interleukin-6 accompanied by a significant increase in fatigue, confusion, and impaired concentration at 3 hours. Performance of the Stroop task under inflammation activated brain regions encoding representations of internal bodily state. Spatial and temporal characteristics of this response are consistent with interoceptive information flow via afferent autonomic fibers. During performance of this task, activity within interoceptive brain regions also predicted individual differences in inflammation-associated but not placebo-associated fatigue and confusion. Maintenance of cognitive performance, despite inflammation-associated fatigue, led to recruitment of additional prefrontal cortical regions. Conclusions These findings suggest that peripheral infection selectively influences central nervous system function to generate core symptoms of sickness and reorient basic motivational states.

blood samples and administered the injections of saline and Typhoid, was unblinded though played no role in the above procedures to ensure blinding of procedures while maintaining subject safety. Salmonella typhi vaccination is a standard vaccination for travel to regions with poor sanitation; other than mild sickness symptoms, local soreness and erythema, serious reactions are rare. Informed consent was obtained in accordance with the Declaration of Helsinki (Helsinki 1991) and the procedures were approved by the joint University College London (UCL/University College London Hospitals (UCLH) Ethics Committee). Participants were recruited by advertisement on the UCL campus and given a small financial reimbursement for involvement in the study. Of note these are the same group of subjects that were reported on in our previous study (1) and in (2).

Generation and measurement of inflammatory response
Participants were injected with either Salmonella typhi capsular polysaccharide vaccine 0.025 mg (Typhim Vi, Aventis Pasteur MSD, Berkshire, UK) or 0.9% NaCl into the non-dominant deltoid muscle, receiving the other injection on their return visit.
Venesection was performed at baseline and 3 hours post vaccination immediately after subjects completed MRI scanning. Participants were scanned from two to three hours after vaccination, which was designed to coincide with the previously observed peak cytokine response to Typhoid vaccination (3). Blood (10 ml) was drawn into vacutainer tubes containing EDTA as anti-coagulant, and then centrifuged immediately at 1250xg for ten minutes at room temperature. Plasma was removed, aliquoted and frozen at -70°C prior to analysis. Plasma interleukin-6 (IL-6) and tumor necrosis factor alpha (TNF-α) were assessed using high sensitivity twosite ELISAs (R&D Systems, Oxford, UK). The limit of detection of the IL-6 assay was 0.09 pg/ml, with intra-and inter-assay coefficients of variation (CVs) of 5.3 and 9.2%.
The TNF-α assay had a detection limit of 0.10 pg/ml with intra-and inter-assay CVs of 6.9 and 8.4% respectively. Plasma IL-1Ra concentrations were determined by a commercial ELISA from R&D Systems (Oxford, UK). This assay had a limit of detection of 15 pg/ml and inter-and intra-assay CVs of less than 10%. Salivary cortisol was collected using cotton dental rolls at baseline, 2 and 3 hours (Salivettes, Sarstedt, Leicester, UK) and analyzed using a time resolved immunoassay with fluorescence detection. Intra-and inter-assay variability was <10% and 12% respectively. Body temperature, heart rate and resting blood pressure were measured at baseline, 2 and 3 hours using sub-lingual digital thermometer and an electronic sphygmomanometer (A&D UA779, Tokyo, Japan) respectively.

Behavioral and mood ratings
Mood and other psychological symptoms were assessed using a modified version of the Profile of Mood States (POMS) (4). This consisted of 5-6 items from each of five scales (vigor, fatigue, depression, tension-anxiety, and mental confusion), together with four somatic symptom items. Each was rated from 0-4, and scores were computed by summing ratings on individual items. Paired t-test was used to compare responses to vaccine and placebo conditions.

Color word Stroop task
The target color word was presented with the four possible response words (red yellow green blue) below. Target words and the order of response words were displayed randomly. Subjects were instructed to respond as rapidly as possible to the color of the target word using a 4 button response pad corresponding to the response words below. In the incongruent condition all words were printed in a color incongruent with the target word; in the congruent condition the font color of all words matched the read target word. Both targets and possible response words were displayed for 3000 ms preceded by a central fixation cross presented for 2000 ms.
Feedback was not given. Trials were presented in 5 blocks of 36 events with 10000 ms breaks between blocks. Percentage of incongruent trials was varied between blocks and ranged from 14% (2 blocks), 28% (1 block) to 42% (2 blocks), with 17% null events per block.

Functional imaging and imaging data analysis
Functional MRI data were acquired on a 1.5T Siemens Sonata magnetic resonance scanner equipped with a standard head coil. Mild external head restraint was used to minimize head movement during scanning. Heart rate was continuously recorded using a pulse oximeter (Nonin 8600, Nonin Medical Inc., Plymouth, MN), pulse probe on the left index finger. Visual stimuli were projected onto a screen visible via a mirror on the head coil. Functional brain imaging data were acquired using T2*-weighted echoplanar imaging, sensitive to BOLD contrast. For the Stroop task, data were acquired with whole brain coverage using a sequence with 44 contiguous slices, 2mm slice thick, 1mm inter-slice gap, tilted -30˚ from intercommisural plane, TE 40 ms, TR 3.96 s per volume (5). 248 volumes were acquired for each participant in a single session of 16.4 minutes.

Functional
MRI datasets were analyzed using SPM5 (http://www.fil.ion.ucl.ac.uk/spm). The first five volumes were discarded to allow for T1 equilibration effects. Individual scans were realigned and unwarped, timecorrected, normalized and spatially smoothed with an 8-mm FWHM Gaussian kernel using standard SPM methods. A high-pass frequency filter (cut off 120 s) and corrections for auto-correlation between scans (AR1) were applied to the time series.
In analysis of the Stroop task, each event was modeled by a standard synthetic hemodynamic response function at each voxel across the whole brain.
Congruent and incongruent trials and errors of commission (response within 3 seconds) and omission (response after 3 seconds) errors were modeled as separate regressors in first-level multiple regression analysis. In the experimental design (see above) null events were included to facilitate identification of differential hemodynamic responses to stochastically-ordered stimuli.
The first level individualized design matrices for each participant were were used as co-regressors in both the analysis of the main effect of inflammation and the between subject correlations. All co-ordinates relate to MNI space.

Change in inflammatory cytokines following vaccine and placebo
Vaccine Placebo   Results reported for clusters of ten or more contiguous voxels after whole brain false detection rate (FDR) correction at p < 0.05.