Endotoxin-Induced Physiological and Psychological Sickness Responses in Healthy Humans: Insights into the Post-Acute Phase

Inflammation is an innate protective response of the body to infection and tissue injury. One key component of the inflammatory response is the release of soluble mediators such as cytokines and prostaglandins from activated immune cells. The main function of these mediators is to coordinate the local and systemic immune responses to microbial antigens or endogenous danger signals [1], but their action is not limited to the immune system. Cytokines and prostaglandins also affect the brain, inducing a wide spectrum of autonomic, endocrine, and behavioral/psychological changes [2‒5]. The inflammation-induced behavioral symptoms are collectively termed “sickness behavior” and include both bodily and psychological symptoms such as loss of appetite, sleepiness, deterioration of mood, lethargy, cognitive disturbances, social withdrawal, and increased pain sensitivity [6, 7]. Despite its negative impact on subjective well-being, sickness behavior is considered an adaptive response that saves energy for the energy-demanding fever and immune responses [8]. However, while protective in the first line, excessive or persistent inflammation can lead to an exacerbation of sickness behavior, promoting the development of mood disorders including major depressive disorder (MDD) [4, 9, 10].

One translational model that has largely contributed to our current understanding of sickness behavior and inflammation-associated depression is the administration of bacterial endotoxin (lipopolysaccharide [LPS]) in animals and humans [11‒13]. LPS, a major component of the outer membrane of Gram-negative bacteria, is a prototypical pathogen-associated molecular pattern that activates the innate immune system through a Toll-like receptor (TLR) 4-dependent pathway. Intravenous administration of LPS in healthy humans rapidly elicits peripheral and central inflammatory responses as well as activation of the sympathetic nervous system and the hypothalamic-pituitary-adrenal (HPA) axis [14‒18]. The LPS-induced inflammatory response is accompanied by behavioral and negative affective changes (e.g., decrease in positive mood, increase in negative mood and anxiety) that resemble core symptoms of depression [17‒19]. The depression-like symptoms typically emerge within 2 h after LPS administration, persist during the acute phase of inflammation for about 4 h, and correlate with cytokine levels in the blood and cerebrospinal fluid [18‒23]. In inflammation-associated depression, mood symptoms are present for longer periods of time and correlate with elevated serum levels of C-reactive protein (CRP), a clinical marker of inflammation [24‒26]. In the human endotoxin model, however, CRP levels in the systemic circulation do not begin to rise before 6 h after LPS administration, i.e., when mood symptoms already have improved [18, 21]. Thus, it remains open whether there is a second wave of CRP-driven affective symptoms that occurs at a later time point after endotoxin exposure or if endotoxin-induced symptoms are restricted to the acute phase of inflammation.

The aim of this randomized, double-blind, placebo-controlled crossover study in healthy human volunteers was to look beyond the already well-characterized LPS-induced acute inflammatory phase to obtain insights into potential post-acute physiological and psychological consequences of endotoxin administration that may either persist or newly emerge between 24 and 72 h after injection. In addition to inflammatory and psychological measures, we also assessed diurnal cortisol profiles on the days following the LPS challenge as HPA axis dysregulation is another key finding in inflammation-associated depression [27, 28].

The study was part of a larger trial on the effects of endotoxemia on B- and T-cell dynamics [29, 30]. The data reported here are previously unpublished data from a subgroup of trial participants who provided a complete data set for all psychological and physiological measures reported. Healthy male volunteers, aged 18–40 years, were recruited by public advertisement. After an initial telephone interview, prospective study participants were invited for an extensive medical screening consisting of a structured personal interview, a physical examination, and the measurement of blood and clinical chemistry parameters (i.e., complete blood cell count, CRP, coagulation factors, liver enzymes, and renal parameters). General exclusion criteria were body mass index <19 or ≥29 kg/m², any known preexisting or current mental or physical illness, regular medication use, smoking, regular alcohol use, sleep disturbances, and anxiety and/or depression scores exceeding published cutoffs of the Hospital Anxiety and Depression Scale (HADS) [31]. Participants were instructed to refrain from excessive exercise for 48 h before testing and to maintain regular night sleep before and during study blocks.

The study used a randomized, double-blind, placebo-controlled crossover design. Each participant received an injection of low-dose endotoxin (LPS) and placebo, respectively, on two separate but otherwise identical study days. The order of the treatment conditions was randomized and counterbalanced, with a minimum interval of 7 days in between study days.

On treatment days, participants first underwent a medical checkup and were then prepared for the study. An intravenous catheter was inserted into a forearm vein for repeated blood collection and LPS or placebo administration. After a rest period of 30 min, a baseline blood sample was obtained. Thirty minutes later, subjects received an injection of either LPS (0.8 ng/kg of body weight; reference standard endotoxin from Escherichia coli O113:H10, lot H0K354; United States Pharmacopeia, Rockville, MD, USA) or placebo (sterile, isotonic NaCl solution, B Braun Melsungen, Melsungen, Germany). The LPS had been subjected to a microbial safety testing routine by the German Federal Agency for Sera and Vaccines (Paul-Ehrlich Institute, Langen, Germany) and was stored in endotoxin-free borosilicate tubes (Pyroquant Diagnostik, Mörfelden-Waldorf, Germany) at −20°C until use. To minimize diurnal effects, all injections were performed between 12:00 and 14:00 h for each participant at identical times on both treatment days. Additional blood samples were collected during acute (1, 2, 3, 4, 6 h) and post-acute (24, 48, 72 h) phases after endotoxin and placebo injection, respectively (see Table 1). After collection of the 6-h sample, the catheter was removed, a physical exam was performed, and participants were discharged home. For post-acute samples, participants had to return to the study site, and blood was collected by venipuncture. In addition to blood, saliva was collected for cortisol analyses, and participants were required to complete questionnaires (for details, see below).

Blood samples were drawn into collection tubes (S-Monovette, Sarstedt, Nümbrecht, Germany) containing either an anticoagulant (EDTA) or clot activator. Plasma and serum, respectively, were obtained by centrifugation (2,000 g, 10 min, 4°C) and were stored at −80°C until analysis. Plasma concentrations of the pro- and anti-inflammatory candidate cytokines IL-6 and IL-1 receptor antagonist (ra) were quantified using Human Quantikine ELISAs (R&D Systems, Minneapolis, MN, USA), and serum concentration of CRP was measured using Human CRP ELISA (IBL International, Hamburg, Germany) according to the manufacturer’s instructions. The sensitivity of the assays was 0.70 pg/mL for IL-6, 6.3 pg/mL for IL-1ra, and 0.23 mg/L for CRP. Mean inter- and intra-assay coefficients of variation (CV) were <10%.

Saliva samples were collected at the same times as blood samples. Additional saliva samples were obtained in the morning and evening of day 1, day 2, and day 3 after endotoxin or placebo administration to analyze potential effects on diurnal cortisol profile (for exact times, refer to Table 1). Saliva was self-collected by participants with commercial collection devices (Salivette Cortisol; Sarstedt, Nümbrecht, Germany) and stored at −20°C until analysis. Salivary cortisol levels were measured by enzyme-linked immunosorbent assay (Cortisol Saliva ELISA, IBL International, Hamburg, Germany) according to the manufacturer’s instructions. Cross-reactivity of the anti-cortisol antibody with other relevant steroids was 7% (11-desoxycortisol), 4.2% (cortisone), 1.4% (corticosterone), 0.4% (progesterone), and <0.01% (testosterone). Sensitivity of the assay was 0.082 nmol/L, and mean inter- and intra-assay CVs were <10%.

Physical sickness symptoms, calmness, alertness, sleepiness, positive and negative moods, and state anxiety were assessed before (baseline) as well as at different occasions during the acute and post-acute phases after endotoxin or placebo administration using a comprehensive battery of validated questionnaires (for exact time points, see Table 1). Physical sickness symptoms such as nausea, fatigue, headache, muscle pain, and reduced appetite among others were measured with an adaptation of the General Assessment of Side Effects (GASE) questionnaire as previously described [21]. Calmness and alertness were quantified with two 4-item subscales of the Multidimensional Mood Questionnaire (MDBF) [32]. Sleepiness was assessed with the 7-point Stanford Sleepiness Scale (SSS) [33]. Positive mood (euthymia) and negative mood (dysthymia) were measured with two 5-item subscales of the state version of the State-Trait Anxiety Depression Inventory (STADI) [34]. State anxiety was assessed with the 20-item state version of State-Trait Anxiety Inventory (STAI) [35].

Statistical analyses were performed using IBM SPSS Statistics 29 (IBM Corporation, Armonk, NY, USA). Normal distribution of the variables was examined using the Shapiro-Wilk test, and data were log-transformed prior to analysis when necessary. Differences between endotoxin and placebo conditions were analyzed by two-way repeated measures ANOVA with “treatment” and “time” as within-subject factors. Greenhouse-Geisser correction was applied when the assumption of sphericity was violated. In case of significant treatment × time interaction, paired t tests were computed to compare endotoxin and placebo conditions at the different sampling times. All results are reported as mean ± SEM, and the level of significance was set at p < 0.05.

Eighteen healthy male volunteers with a mean age of 25.9 ± 1.0 years and a mean body mass index of 24.0 ± 0.5 kg/m² were included in the study. HADS anxiety (2.94 ± 0.55) and depression (1.53 ± 0.43) scores were within normal ranges, and baseline serum CRP levels were in all subjects below the clinical cutoff (5 mg/L). Descriptive baseline data for subjects receiving placebo as a first or second treatment are provided as online supplementary Table S1 (for all online suppl. material, see https://doi.org/10.1159/000534444).

LPS administration induced a systemic inflammatory response, which was reflected in a subfebrile rise in body temperature (F[8, 136] = 29.9, p < 0.001; peak: 37.7 ± 0.1°C; data not shown) as well as significant increases in circulating concentrations of IL-6 (F(8, 136) = 36.3, p < 0.001; Fig. 1a), IL-1ra (F(8, 136) = 72.1, p < 0.001; Fig. 1b), and CRP (F(5, 75) = 173.2, p < 0.001; Fig. 1c) compared to placebo treatment (all ANOVA time × treatment interaction effects; see Fig. 1 for results of post hoc tests). The LPS-induced increase in plasma IL-6 was short-lived, peaking at 2 h after injection, and rapidly returning to control levels by 6 h after injection. In contrast, plasma levels of anti-inflammatory IL-1ra were highest between 3 and 4 h after LPS treatment and declined slowly thereafter. Twenty-four hours after LPS administration, IL-1ra concentration was still significantly elevated compared to placebo. Interestingly, 72 h after LPS injection, IL-1ra levels even decreased below those of placebo treatment. The increase in serum CRP was delayed relative to IL-6 and IL-1ra, peaking at 24 h and persisting until 72 h post injection. Correlation analyses showed that peak levels (2 h) of IL-6 were significantly positively correlated with peak levels (3 h) of IL-1ra (Pearson’s r = 0.65, p = 0.003) and peak levels (24 h) of CRP (Pearson’s r = 0.64, p = 0.004), respectively. No further correlations were found.

The inflammatory response was accompanied by a wide range of bodily and psychological sickness symptoms. Endotoxin administration induced a significant increase in physical symptoms (GASE score: F(8, 136) = 15.5, p < 0.001; Fig. 2a) and a significant reduction in calmness (MDBF score: F(5, 75) = 7.9, p < 0.001; Fig. 2b) and alertness (MDBF score: F(5, 75) = 10.9, p < 0.001; Fig. 2c) compared to placebo treatment (all ANOVA time × treatment interaction effects; see Fig. 2 for results of post hoc tests). In addition, sleepiness was significantly increased (SSS score: F(8, 136) = 7.1, p < 0.001; Fig. 2d). These bodily sickness symptoms were accompanied by a worsening of mood, as evidenced by a significant decrease in positive mood (STADI euthymia score: F(8, 136) = 4.6, p < 0.001; Fig. 2e) and a significant increase in negative mood (STADI dysthymia score: F(8, 136) = 5.3, p < 0.001; Fig. 2f). Administration of the inflammatory stimulus also induced a significant increase in state anxiety (STAI-S score: F(8, 136) = 11.1, p < 0.001; Fig. 2g). All of these effects were exclusively observed during the acute phase after LPS injection, whereas none of the symptoms persisted or recurred during the post-acute phase.

Administration of the inflammatory stimulus also elicited a pronounced cortisol response (Fig. 3). Compared to the placebo condition, salivary cortisol levels were significantly increased during the acute phase after LPS treatment, reaching peak levels at 3 h and remaining elevated until 6 h post injection (F(14, 196) = 28.9, p < 0.001; ANOVA time × treatment interaction effect; see Fig. 3 for results of post hoc tests). Additional analyses in saliva samples collected in the morning, noon, and evening of the subsequent 3 days revealed pronounced circadian changes but no significant differences in diurnal cortisol profiles between placebo and endotoxin conditions.

Given the increasingly recognized role of inflammation in the etiology and pathophysiology of mood disorders, research on the mechanisms underlying inflammation-induced psychological symptoms has recently received a lot of attention. In this context, administration of bacterial endotoxin in healthy volunteers has emerged as a translational tool to investigate temporal associations between changes in inflammatory markers, sickness behavior, and affective symptoms [13]. However, the focus of previous studies has been predominately on the acute inflammatory phase, covering the early hours after LPS injection. In the present study, we investigated potential post-acute changes that either persist beyond the early phase or develop later, i.e., with some delay to the onset of the inflammatory insult. We paid particular attention to changes in serum CRP, which was found to be increased in about one-third of MDD patients [24‒26] and its possible association with negative affective symptoms. In addition, we aimed to determine whether LPS-induced inflammation not only triggers acute changes in cortisol secretion but also has effects on the diurnal cortisol profile in the following days after the challenge as HPA dysregulation is another key finding in MDD [27, 28].

Consistent with previous work, our data show a short-lived LPS-induced cytokine response evidenced by a transient rise in plasma IL-6 levels that was followed by a more prolonged increase of anti-inflammatory IL-1ra in the circulation [17, 18, 36]. While the IL-6 response had completely vanished by 6 h, systemic IL-1ra levels remained significantly elevated until 24 h after LPS administration. At 72 h, IL-1ra levels in the LPS condition were even lower than those in the placebo condition, indicating a potential overcompensation of the anti-inflammatory response during the post-acute phase. However, this needs to be confirmed in future studies by determining other important anti-inflammatory cytokines such as IL-4 and IL-10. At the symptom level, the acute cytokine response was accompanied by an increase in physical and psychological sickness symptoms, including a worsening of mood and an increase in anxiety, again confirming previous findings from the human LPS model [17‒19, 23].

In addition to the transient rise of pro- and anti-inflammatory cytokines during the acute phase, endotoxin administration also resulted in a prolonged increase of serum CRP that was delayed compared to IL-6 and IL-1ra, peaking at 24 h and gradually decreasing until 72 h after injection, in line with its half-life of about 19 h [37]. CRP is a routine clinical marker of inflammation that is not only increased in patients with infectious or inflammatory diseases but also in a substantial subset of patients with MDD [24, 25]. Elevated levels of high sensitivity CRP (>3 mg/L) in MDD patients have been found to correlate with the severity of depressive symptoms [26] and predict future development of depression [38, 39] as well as resistance to standard antidepressant therapies [40]. Moreover, clinical trials testing the efficacy of anti-cytokine therapies in inflammation-associated depression typically include only those patients with elevated blood CRP [41]. In contrast to the findings in MDD patients, we show here that the comparatively high serum levels of CRP (20.4 ± 0.9 mg/L) found in the post-acute phase after endotoxin administration were not associated by either physical sickness symptoms or negative affective changes, clearly indicating that CRP per se does not appear to cause depressive symptomatology. The strong positive correlation between serum CRP and the preceding plasma IL-6 response rather suggests that the previously reported associations between CRP and depressive symptoms in MDD may reflect the presence of elevated IL-6 in these patients since hepatic CRP synthesis is mainly dependent on IL-6 [37]. Thus, systemic CRP levels in MDD should be considered an indirect measure of increased pro-inflammatory cytokines, which are often difficult to detect because of their low blood concentrations in these patients. Nevertheless, given its much longer half-life compared to cytokines, CRP should be regarded as a sensitive marker to detect low-grade inflammation in patients with inflammation-associated depression.

Our data clearly show that the depression-like symptomatology in the human endotoxin model occurs only during the acute inflammatory phase and concomitantly with sickness behavior, whereas LPS-induced sickness and depression-like behaviors in animals exhibit different temporal dynamics. Peripheral LPS administration in rodents induces sickness behavior (based on locomotor activity) that peaks between 2 and 6 h and then gradually subsides. Depression-like behavior emerges on this background with a peak around 24 h after LPS [4, 42‒45]. However, it must be emphasized that it is difficult to capture human emotional states in animals and that the tools used to assess depression-like symptoms differ greatly in humans and animals for obvious reasons. In animals, depression-like behavior is usually captured indirectly based on increased immobility in the forced-swim or tail-suspension tests and decreased preference for sweet taste, whereas in humans, depressive symptoms are assessed directly by clinical interview or self-reports. Thus, the differences in findings might be, at least in part, due to the difficulties of comparing subjective ratings in humans with the results of behavioral tests in animals. Besides these translational issues in the assessment of depression-like symptoms across species, there are others factors that could also account for the different temporal dynamics of the behavioral responses between animals and humans. For example, the LPS doses used in animals are much higher compared to humans, which can lead to impairment of the blood-brain barrier and consequently a prolonged central inflammatory response [43, 46, 47]. In turn, this then could result in an exaggerated sickness response with a later emergence of depression-like behaviors. Such endotoxin-induced impairment of the blood-brain barrier integrity does not occur in the human LPS model, at least for the LPS dose used herein as we have recently shown [18].

In addition to behavioral and psychological changes, systemic inflammation also triggers profound autonomic and neuroendocrine responses, including activation of the sympathetic nervous system and HPA axis. These allostatic physiological responses are not only important for adapting the body to the altered metabolic demands of infection or tissue injury but are also part of a centrally mediated counter-regulatory loop that, together with anti-inflammatory cytokines, provides negative feedback control of the inflammatory process [3]. Consistent with this, we show here that LPS administration elicited a pronounced salivary cortisol response that persisted throughout the acute inflammatory phase. Twenty-four hours later, however, cortisol secretion had completely normalized. Additional analyses of saliva samples collected in the morning, noon, and evening of the following 3 days showed the well-documented circadian changes with high cortisol levels in the morning and low levels in the evening [48]. Importantly, we found no differences in diurnal cortisol profiles between placebo and endotoxin conditions during this post-acute phase. This suggests that in healthy individuals, a single inflammatory episode does not lead to disturbances in cortisol secretion that extend beyond the acute inflammatory phase. Therefore, altered diurnal cortisol profiles in inflammation-associated depression more likely either result from persistent/chronic low-grade inflammation or, alternatively, reflect stress-associated dysregulation of the HPA axis [48, 49].

In summary, our results clearly demonstrate that the administration of low-dose endotoxin in healthy humans does not elicit a delayed second wave of sickness symptoms or negative affective changes that appears concomitantly with the rise in systemic CRP levels. In addition, we found no evidence for persistent changes in inflammatory markers or diurnal cortisol profile in the aftermath of the LPS challenge. These findings confirm that experimental endotoxemia in humans is a robust and safe model of systemic inflammation that does not cause long-term changes in healthy individuals, which is important not only for participant safety but also sometimes required for multiday study designs in experimental research. The model provides a highly standardized approach to eliciting a transient inflammatory response along with depression-like symptoms. This allows not only to study the afferent communication pathways and brain networks underlying inflammation-induced behavioral and affective symptoms but may also help to identify potential new treatment targets for inflammation-associated depression. However, our results also highlight the limitations of the model with respect to symptom chronicity which is an important clinical feature of MDD. This translational limitation could be at least partially addressed in future studies by attempting to prolong the LPS-induced symptomatology by continuous infusion of endotoxin.

We thank Larissa Lueg and Eva Stemmler for their support in data collection, Alexandra Kornowski for technical assistance, and Dr. Ingo Spreitzer (Paul-Ehrlich-Institute, Department of Microbial Safety, Langen, Germany) for endotoxin safety testing.

The experiments were conducted in accordance with the Declaration of Helsinki, and the study protocol was reviewed and approved by the Ethics Review Board of the Medical Faculty of the University of Duisburg-Essen (Approval No. 15-6533-BO). All volunteers provided written informed consent and received financial compensation for their participation in the study.

The authors have no conflicts of interest to declare.

This work was partly funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – Project-ID 316803389 – SFB 1280 (subproject A12) and Project-ID 422744262 – TRR 289 (subprojects A10 and A11). The funding organization was not involved in study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the article for publication.

H.E., M.S., and S.B. conceived and designed the study; H.E., A.B., and S.B. carried out data collection; H.E., A.B., and S.B. analyzed the data; B.W., A.K., H.R., O.W., and M.S. contributed to the interpretation of the data. H.E. and S.B. wrote the manuscript. All the authors critically revised the manuscript for important intellectual content. All the authors approved the final manuscript.

The data from this study are not publicly available due to ethical reasons. Further inquiries can be directed to the corresponding author.