Developmental and behavioural tests
The acquisition of developmental milestones, reflex ontology and behavioural development of mouse pups were assessed using procedures derived from Fox’s battery of tests and Wahlsten’s adaption of Fox’s tests (Fox 1965; Wahlsten 1974), as described in (Heyser 2004; Feather-Schussler and Ferguson 2016). Pups were tested individually, preferably in the morning (9 a.m.-1 p.m). To limit stress due to maternal separation, the time spent by the pup away from the mother and home cage was limited to the duration of each test. At the end of each test the pup was immediately put back in its home cage.
Ear development: The day of the opening of the ear canal, defined as a fully detached outer ear membrane, was recorded in pups aged P3 to P5. A fixed score of 0, 1 or 2, according to the number of ears developed per animal was assigned.
Eyelid opening: The day of the eyelid opening, defined as any visible break in the membrane covering the eye, was recorded in pups aged P12 to P16. A fixed score of 0, 1 or 2, according to the number of eyes opened per animal was assigned.
Righting reflex: Pups were examined for the righting reflex every two days between P2 and P6. Each pup was placed on its back on a flat, hard surface and kept immobile for 5 s. The pup was then released and the time taken to return to the upright position was recorded. Animals unable to perform the righting after 1 min were assigned a score of 60 s. Each pup was only tested once per day.
Acoustic startle reflex: Pups aged P10 to P14 were examined every day and the day of reflex acquisition was determined. A cell counter was use to generate an auditive stimulus and the animal’s startle response was recorded. A fixed score of either 0 or 1 was assigned. A score of 1 was given to pups which displayed a startle reaction; and a score of 0, to pups which did not exhibit a startle response.
Ambulation test: The ambulation ability of pups was assessed every day between P6 and P12 to monitor the acquisition of walking proficiency. Each pup was placed on a flat, hard surface and the walking pattern was assessed over 1 min, according to previous scoring criteria (Feather-Schussler and Ferguson 2016): 0 = no movement, 1 = crawling with asymmetric limb movement, 2 = slow crawling but symmetric limb movement, and 3 = fast crawling/walking.
Serum cytokine measurements
We used the the V-PLEX® Mouse TNF Kit or the V-PLEX® Mouse Cytokine 29-Plex Kit electrochemiluminescence-based assays from (Meso Scale Diagnostics, Rockville, USA). Assays were performed according to the manufacturer’s instructions. The reagents were stored at 4°C until used, when they were allowed to reach RT immediately prior to usage. All buffers used were provided with the kit, with the exception of the Wash Buffer, which was produced in-house using sterile PBS 1x and 0.05% Tween-20. Lyophilized standards (containing cytokines IFN-γ, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-12p70, KC/GRO, TNF, IL-9, MCP-1, IL-33, IL-27p28/IL-30, IL-15, IL-17A/F, MIP-1α, IP-10, MIP-2, MIP-3α, IL-22, IL-23, IL-17C; IL-31, IL-21, IL-16, IL-17A, IL-17E/IL-25) were reconstituted to master stock solutions. Eight concentrations of the standard were made by fourfold serial dilutions of the master stock. The majority of serum samples were diluted 1:4, with the exception of several samples with low remaining serum volume, for which dilutions between 1:6 to 1:11 were used. All samples were diluted on ice on pre-plates. The background level was determined using a buffer blank. Eight serial dilutions of standards and buffer only (in duplicates) were run together with samples run in singlicate using the Sector Imager 2400 plate reader (Meso Scale Diagnostics, Rockville, USA). Concentrations of cytokines in each sample were interpolated from standard curves generated with a five-parameter logistic regression equation in Discovery Workbench 3.0 software (Meso Scale Diagnostics, Rockville, USA). The software allows for a graphic visualization of the placement of all samples on the standard curve for each cytokine. IL-2, IL-4, IL-12p70, IL-9, IL-17A/F, IL-22, IL-23, IL-17C, IL-31, IL-21, IL-17E, IL-25 were below the Lower Limit of Detection (LLOD) in more than 20% of the samples and were therefore not retained for downstream analyses. For the other cytokines, cytokine levels below the lower limit of detection (LLOD), we imputed a value equal to half the LLOD value indicated by the manufacturer.
Perinatal TNF injections from P1 to P5 yield increased serum levels of TNF at P5
To determine the optimal dose of TNF to use in the study, we performed a pilot experiment in which mouse pups were injected intraperitoneally daily from P1 to P5 with various doses of TNF ranging from 0.25 to 20 μg/Kg. All pups survived, even when injected with the highest TNF dose, and there was no sign of inflammation (redness, swelling) on the injected flank. There was no significant effect of TNF treatment on body weight in all groups, suggesting that TNF treatment did not impair early growth (Fig. 1A).
Compared to saline-injected mice, pups injected with low doses of TNF (0.25, 1 μg/Kg) did not exhibit changes in circulating levels of TNF at P5. For pups injected with higher doses, 5 and 20 μg/Kg, there was a dose-dependent increase in serum TNF levels at P6, with 2-fold and 11-fold increase, respectively (Fig. 1B). Based on these results, we chose to use the highest dose of TNF in further experiments, which yielded high levels of TNF levels while preserving general growth.
TNF-injected pups acquire sensorimotor reflexes more rapidly
We monitored the growth, cytokine circulating levels at P16, the acquisition of developmental milestones (ear eversion, eyelid opening) and sensorimotor reflexes (righting reflex, acoustic startle reflex) in pups injected from P1 to P5 with 20 μg/Kg of TNF and in control pups (Fig. 2A). For each variable under scrutiny, we used GEE to test the effects of treatment considering a repeated measures dependency on the measurements, the effect of time and of having pups from two distinct cohorts. Regarding body weight gain from P2 to P16, there were significant effects of time and cohort variables, but no significant impact of TNF treatment (Fig. 2B, C), suggesting that TNF did not impact general growth.
We assessed serum samples at P16 for the levels of TNF and other cytokines after repeated injections of TNF at 20 μg/Kg from P1 to P5. Vehicle- and TNF-injected pups exhibited similar levels of TNF as well as all other tested cytokines including IL-1β, IL-6, IFN-γ, IL-5 and CXCL1 (Supplementary Fig. 1). Of note, positive correlations between the levels of IL-5 and IL-6, as well as CXCL1 and IL-6 were observed (Supplementary Fig. 2). This suggested that perinatal injections of TNF from P1 to P5 is not sufficient to induce a sustained systemic inflammation by P16.
Furthermore, TNF treatment did not modify the timing of developmental milestones such as ear aversion (Fig. 2D, E) or eyelid opening (Fig. 2F, G), while there were significant effects of time and cohort variables. This suggested that TNF did not induce gross developmental changes.
However, TNF treatment had a significant effect on the acquisitions of early reflexes (Fig. 2H-K). Both the righting reflex (Fig. 2H, I) and the acoustic startle reflex (Fig. 2J, K) occurred at earlier stages in TNF-injected pups compared to control animals, as shown by the significant effect of time and treatment variables. At P2, TNF-injected pups took almost twice less time to right up than control pups (Fig. 2H). By P14, all the TNF-injected mice had acquired the acoustic startle reflex, while only 50% of the control pups had acquired it (Fig. 2J).
TNF-injected pups exhibit enhanced exploratory behaviour
We further characterized the impact of perinatal TNF treatment on: olfactory orientation at P9, walking proficiency (ambulation, from P6 to P12) and locomotor activity at P13 (Fig. 3A). In the olfactory orientation test, TNF- and saline-injected mice performed equally well as measured by the latency or time to first reach the maternal bedding (Fig. 3B) and the equal time spent the three zones of the set-up (Fig. 3C). This indicated that sensorimotor processing was likely similar in both groups for this task.
When assessing walking proficiency in TNF-injected and control pups, there was also a significant effect of the variables time and cohort (Fig. 3D, E). However, there was no effect of TNF treatment, as TNF- and saline-injected pups exhibited similar development overtime of walking proficiency, with the acquisition of a mature walking pattern by P12 in both groups (Fig. 3D, E).
We then monitored the exploratory behaviour of TNF-injected and control pups when individually placed in a novel environment (i.e. an unfamiliar clean cage without bedding) for 10 min. TNF-injected pups spent more time mobile (Fig. 2F, G) and travelled a longer distance during a 10 min session (Fig. 3H), when compared to vehicle-injected pups, but there was no effect of time or cohort variables on these parameters. Given that by P12 both TNF-injected and control pups had acquired a mature walking pattern (Fig. 3D, E), these data can directly be interpreted as an increased exploratory behaviour in TNF-injected pups, as shown by their increase in the exploratory index compared to control pups (Fig. 3I).
Table of contents :
Table of contents
List of abbreviations
1. Interactions between the immune system and the brain during immune activation
1.2. Cytokines and cytokine receptors
1.3. Pathogen recognition
1.4. Pathogen Associated Molecular Patterns
1.5. The peripheral immune system and the brain
1.6. Cytokines and behaviour
2. How cytokines shape neurodevelopment
2.1. The essential role of cytokines during embryonic development
2.2. Microglia: neuroimmune interactions in shaping neuronal circuitry
2.3. Neonatal immune system vs. adult immune system
2.4. The essential role of cytokines during postnatal development
2.5. Specific role of cytokines in neurodevelopment: the example of TNF
2.6. Human studies supporting a role of cytokines in neurodevelopment
3. Cytokines can interfere with neurodevelopment and contribute to neurodevelopmental disorders, including autism spectrum disorders
3.1. Neurodevelopmental disorders – overview
3.2. Autism spectrum disorders
3.3. Human studies: the case of ASD
3.4. Mouse models of neurodevelopmental defects triggered by immune activation
4. Scientific hypotheses and objectives
2. Materials and Methods
5. Conclusions & future directions
2. Materials and Methods
1. How does TNF impact neurodevelopment?
2. What is the physiological relevance of our experimental model in which pups are injected with recombinant TNF?
3. Vulnerability vs. resilience to psychological stress
4. What else could we do to further investigate the impact of TNF on neurodevelopment?
5. What is the impact of poly(I:C) injection on pregnant dams?
6. Why do pups born to poly(I:C)-injected mothers gain weight more rapidly than control pups?
7. Why do pups born to poly(I:C)-injected mothers exhibit communication impairment?
8. Why do pups born to poly(I:C)-injected mothers exhibit decreased locomotor activity?
9. Why did we analyse our data using a multivariable statistical approach, and which conclusions could we draw?
10. What is known about the role of TNF, CXCL10, IL-5 and IL-15 in neurodevelopment and brain function?
Take home message