Neonatal IL-4 Over-Exposure Elevated IL-4 Levels and Reduced the Pro-Inflammatory Cytokine Levels in Serum before P35

Firstly, we observed the effect of IL-4 over-exposure in neonatal rats on the levels of IL-4 and three classic pro-inflammatory cytokines in serum from the neonatal period to adulthood (P70). Specifically, the levels of IL-4, IL-1β, IL-6 and TNF on P7, P14, P21, P28, P35, P42, P49, P56, P63 and P70 in the serum of IL-4 group and CON group were measured.

Data in Fig. 1 present the levels of cytokines in the serum from the neonatal period to adulthood. The levels of IL-4 in serum of IL-4 group were significantly higher than CON group from P7 to P35, with no significant differences between the two groups from P42 to P70 (Paired samples t-test with Bonferroni-Holm correction; P7, p < 0.001, Bonferroni–Holm α = 0.005; P14, p < 0.001, Bonferroni–Holm α = 0.0071; P21, p < 0.001, Bonferroni–Holm α = 0.0056; P28, p < 0.001, Bonferroni–Holm α = 0.0063; P35, p < 0.001, Bonferroni–Holm α = 0.0083; n = 15; all p values > Bonferroni–Holm α values; n = 15; Fig. 1A).

The levels of IL-1β in serum of IL-4 group showed a downward trend before P28 compared to CON group with statistically significant differences on P21 and P28, with no significant differences at the other ages (Paired samples t-test with Bonferroni-Holm correction; P21, p = 0.005, Bonferroni–Holm α = 0.0056; P28, p = 0.004, Bonferroni–Holm α = 0.0050; all p values for the other time points > Bonferroni–Holm α values; n = 15; Fig. 1B).

The levels of IL-6 in serum of IL-4 group showed a downward trend before P28 compared to CON group with statistically significant differences on P14, P21 and P28, with no significant differences at the other ages (Paired samples t-test with Bonferroni-Holm correction; P14, p = 0.002, Bonferroni–Holm α = 0.0063; P21, p < 0.001, Bonferroni–Holm α = 0.0050; P28, p = 0.002, Bonferroni–Holm α = 0.0056; all p values for the other time points > Bonferroni–Holm α values; n = 15; Fig. 1C).

The levels of TNF showed mild decreases than CON group on P14 and P21, with no significant alterations at other ages (Paired samples t-test with Bonferroni-Holm correction; P14, p = 0.004, Bonferroni–Holm α = 0.0056; P21, p = 0.002, Bonferroni–Holm α = 0.0050; all p values for the other time points > Bonferroni–Holm α values; n = 15; Fig. 1D).

To summarize, the results shown above indicate that the decreased levels of pro-inflammatory cytokines is accompanied by elevation of the peripheral IL-4 levels.

Neonatal IL-4 Over-Exposure Induced an Instant Anti-Inflammatory Effect and a Subsequent Pro-Inflammatory Effect in the Hippocampus

Having observed the effect of neonatal IL-4 over-exposure on the levels of IL-4, IL-1β, IL-6, and TNF in the serum, we next measured the levels of these cytokines in the hippocampus at every same age point.

Data in Fig. 2 present the levels of cytokines in the hippocampus from the neonatal period to adulthood. The IL-4 levels in hippocampus of IL-4 group were significantly higher than CON group from P7 to P28(Paired samples t-test with Bonferroni-Holm correction; P7, p < 0.001, Bonferroni–Holm α = 0.063; P14, p < 0.001, Bonferroni–Holm α = 0.0056; P21, p < 0.001, Bonferroni–Holm α = 0.0050; P28, p < 0.001, Bonferroni–Holm α = 0.0071; n = 15; Fig. 2A), with no significant differences between the two groups from P35 to P70 (Paired samples t-test with Bonferroni-Holm correction, all p values > Bonferroni–Holm α values; n = 15; Fig. 2A).

Fig. 2

Neonatal IL-4 over-exposure induced an instant anti-inflammatory effect and a subsequent pro-inflammatory effect in the hippocampus. (A–D) The dots represent the average levels of IL-4 (A), IL-1β (B), IL-6 (C) and TNF (D) in the hippocampus. The data represent the means ± SD. *p value < Bonferroni–Holm α value; n = 15/group; Paired samples t-test was used. HIP hippocampus; ud undetectable

Both IL-6 and IL-1β levels showed a downward trend in IL-4 group from P7 to P28, with statistically significant decreases compared to CON group but showed a period of rebounding increases from P42 to P63 compared to CON group, with no significant differences between the two groups on P35 and P70 (IL-1β: Paired samples t-test with Bonferroni-Holm correction; P7, p < 0.001, Bonferroni–Holm α = 0.0125; P14, p < 0.001, Bonferroni–Holm α = 0.0071; P21, p < 0.001, Bonferroni–Holm α = 0.0083; P28, p < 0.001, Bonferroni–Holm α = 0.0167; P35, p = 0.071, Bonferroni–Holm α = 0.0250; P42, p < 0.001, Bonferroni–Holm α = 0.0063; P49, p < 0.001, Bonferroni–Holm α = 0.0050; P56, p < 0.001, Bonferroni–Holm α = 0.0100; P63, p < 0.001, Bonferroni–Holm α = 0.0056; P70, p = 0.390, Bonferroni–Holm α = 0.0500; n = 15; Fig. 2B) (IL-6: Paired samples t-test with Bonferroni-Holm correction; P7, p < 0.001, Bonferroni–Holm α = 0.0083; P14, p < 0.001, Bonferroni–Holm α = 0.0071; P21, p < 0.001, Bonferroni–Holm α = 0.0067; P28, p < 0.001, Bonferroni–Holm α = 0.0100; P35, p = 0.278, Bonferroni–Holm α = 0.500; P42, p < 0.001, Bonferroni–Holm α = 0.0050; P49, p < 0.001, Bonferroni–Holm α = 0.0056; P56, p < 0.001, Bonferroni–Holm α = 0.0063; P63, p < 0.001, Bonferroni–Holm α = 0.0125; P70, p = 0.087, Bonferroni–Holm α = 0.0250; n = 15; Fig. 2C). From P14 to P70, the TNF level showed a tendency in similar to the IL-6 and IL-1β levels in IL-4 group, but only two statistically significant differences were found on P21 and P42, with no significant alterations at other ages (Paired samples t-test with Bonferroni-Holm correction; P21, p < 0.001, Bonferroni–Holm α = 0.0056; P42, p = 0.004, Bonferroni–Holm α = 0.0063; all p values for the other time points > Bonferroni–Holm α values; n = 15; Fig. 2D).

All of these suggested that neonatal IL-4 over-exposure induced an instant anti-inflammatory effect in the preceding 28 days and a subsequent pro-inflammatory effect that lasted from P42 to P63 in the hippocampus of IL-4 group, which consists with the previous study (Wang et al. 2018).

Neonatal IL-4 Over-Exposure Induced an Instant Anti-Inflammatory Effect and a Subsequent Pro-Inflammatory Effect in the CSF, Which is Phenomenally Similar to Hippocampus, but the Pro-Inflammatory Cytokine Response Appeared Relatively Earlier

Neuroinflammation is constantly accompanied by changes in CSF cytokines. Therefore, we hypothesized that changes in CSF cytokines could reflect the progress of neuroinflammation earlier. To verify this hypothesis, we then measured levels of IL-4, IL-1β, IL-6 and TNF in the CSF of neonatal IL-4 over-exposure rats to investigate the characteristics of CSF cytokine changes and the relation between cytokine changes of CSF and hippocampus.

Data in Fig. 3 present the levels of cytokines in the CSF from the neonatal period to adulthood. The IL-4 levels were significantly higher in the CSF of IL-4 group compared to that of CON group from P7 to P21 (Paired samples t-test with Bonferroni-Holm correction. P7, p < 0.001, Bonferroni–Holm α = 0.063; P14, p < 0.001, Bonferroni–Holm α = 0.0050; P21, p < 0.001, Bonferroni–Holm α = 0.0056; all the other p values > Bonferroni–Holm α values; n = 15; Fig. 3A), with no significant differences from P28 to P70 (Paired samples t-test with Bonferroni-Holm correction. all the p values > Bonferroni–Holm α values; n = 15; Fig. 3A).

Fig. 3

Neonatal IL-4 over-exposure induced an instant anti-inflammatory effect and a subsequent pro-inflammatory effect in the CSF. (A–D) The dots represent the average levels of IL-4 (A), IL-1β (B), IL-6 (C) and TNF (D) in the cerebrospinal fluid. The data represent the means ± SD. *p value < Bonferroni–Holm α value; n = 15/group; Paired samples t-test was used. CSF cerebrospinal fluid; ud undetectable

Just like in the hippocampus, the levels of IL-1β, IL-6 and TNF in the CSF showed a anti-inflammatory profile in IL-4 group from P7 to P21 and a pro-inflammatory profile from P35, with a eventually returning to the normal levels at P70. (IL-1β: Paired samples t-test with Bonferroni-Holm correction; P7, p < 0.001, Bonferroni–Holm α = 0.0056; P14, p < 0.001, Bonferroni–Holm α = 0.0050; P21, p < 0.001, Bonferroni–Holm α = 0.0125; P28, p = 0.0572, Bonferroni–Holm α = 0.0050; P35, p = 0.071, Bonferroni–Holm α = 0.0083; P42, p < 0.001, Bonferroni–Holm α = 0.0071; P49, p < 0.001, Bonferroni–Holm α = 0.0063; P56, p < 0.001, Bonferroni–Holm α = 0.0100; P63, p = 0.0300, Bonferroni–Holm α = 0.0167; P70, p = 0.476, Bonferroni–Holm α = 0.0250; n = 15; Fig. 3B).

(IL-6: Paired samples t-test with Bonferroni-Holm correction; P7, p < 0.001, Bonferroni–Holm α = 0.0063; P14, p < 0.001, Bonferroni–Holm α = 0.0050; P21, p < 0.001, Bonferroni–Holm α = 0.0083; P28, p = 0.0700, Bonferroni–Holm α = 0.0167; P35, p = 0.071, Bonferroni–Holm α = 0.0125; P42, p < 0.001, Bonferroni–Holm α = 0.0071; P49, p < 0.001, Bonferroni–Holm α = 0.0056; P56, p < 0.001, Bonferroni–Holm α = 0.0100; P63, p = 0.3680, Bonferroni–Holm α = 0.0250; P70, p = 0.8760, Bonferroni–Holm α = 0.0500; n = 15; Fig. 3C).

(TNF: Paired samples t-test with Bonferroni-Holm correction; P7, p < 0.001, Bonferroni–Holm α = 0.0050; P14, p < 0.001, Bonferroni–Holm α = 0.0056; P21, p < 0.001, Bonferroni–Holm α = 0.0071; P28, p = 0.2005, Bonferroni–Holm α = 0.0167; P35, p = 0.1486, Bonferroni–Holm α = 0.0100; P42, p < 0.001, Bonferroni–Holm α = 0.0063; P49, p = 0.0083, Bonferroni–Holm α = 0.0083; P56, p = 0.1663, Bonferroni–Holm α = 0.0125; P63, p = 0.5381, Bonferroni–Holm α = 0.0500; P70, p = 0.5330, Bonferroni–Holm α = 0.0250; n = 15; Fig. 3D).

These results hinted that the neonatal IL-4 over-exposure induced an anti-inflammatory effect in the preceding 21 days and a subsequent pro-inflammatory effect that lasted from P35 to P56 in the CSF in IL-4 group. Moreover, the pro-inflammatory cytokine response in CSF appeared relatively earlier than in the hippocampus.

The Levels of Pro-Inflammatory Cytokines in Hippocampus Positively Correlated to Those in CSF at the Individual Level on P42

All the above results suggest that delayed neuroinflammation occurred both in the CSF and hippocampus of IL-4 group at all or some of the age points after P35 or P42, indicated by IL-1β, IL-6 and TNF altered significantly at the two sites. To further explore the relationship of the alterations of four cytokines in the CSF to that in the hippocampus, correlation analyses at the individual level were then performed between the hippocampus level and the CSF level of each of the four cytokines we measured on P42. Both CSF and hippocampus are in the pro-inflammatory cytokine response on P42.

IL-4 could spread to the brain through the choroid plexus, promoting the production of proinflammatory cytokines by microglia and affects inflammation status there.Therefore, we also correlated the cytokine levels in the hippocampus with those in blood at P42.

The results showed that levels of the three pro-inflammatory cytokines in CSF were positively correlated with those in hippocampus on P42 (Pearson’s correlation coefficient; IL-1β: r2 = 0.3073, p < 0.0320; IL-6: r2 = 0.7360, p < 0.001; TNF: r2 = 0.7245, p < 0.001; n = 15; Fig. 4B–D). The levels of IL-4 in CSF and hippocampus also have a positive correlation (Pearson’s correlation coefficient; IL-4: r2 = 0.6473, p < 0.01; n = 15; Fig. 4A).

Fig. 4

The positive correlation of the levels of IL-4, IL-1β, IL-6 and TNF in CSF with the levels of IL-4, IL-1β, IL-6 and TNF in hippocampus in IL-4 group. A Correlation analyses between the CSF IL-4 level and the hippocampal levels of IL-4. B Correlation analyses between the CSF IL-1β level and the hippocampal levels of IL-1β. C Correlation analyses between the CSF IL-6 level and the hippocampal levels of IL-6. D Correlation analyses between the CSF TNF level and the hippocampal levels of TNF. n = 15/group; Pearson’s correlation analysis. CSF cerebrospinal fluid; HIP hippocampus

As for the correlation analyses between the cytokine levels in serum with those in hippocampus, our results showed that there was no significant correlation for any cytokines between serum and hippocampal levels (Pearson’s correlation coefficient; IL-4: r2 = 0.0844, p = 0.2934; IL-1β: r2 = 0.0216, p = 0.6010; IL-6: r2 < 0.0001, p = 0.9934; TNF: r2 = 0.1368, p = 0.1748; n = 15; Fig. 5).

Fig. 5

The positive correlation of the levels of IL-4 in serum with those in hippocampus in IL-4 group. A Correlation analyses between the serum IL-4 level and the hippocampal levels of IL-4. B Correlation analyses between the serum IL-1β level and the hippocampal levels of IL-1β. C Correlation analyses between the serum IL-6 level and the hippocampal levels of IL-6. D Correlation analyses between the serum TNF level and the hippocampal levels of TNF. n = 15/group; Pearson’s correlation analysis. HIP hippocampus

Neonatal IL-4 Over-Exposure Induced Macrophage Accumulation in Dura Mater after Instant Anti-Inflammatory Cytokine Response in CSF

It has been reported that macrophages in dura mater are an important source of cytokines in cerebrospinal fluid, and their number and activation state are changed by the regulation of cytokines in cerebrospinal fluid (Wieseler-Frank et al. 2007; Reuter et al. 2001). Therefore, macrophages in dura mater were observed at P42 when pro-inflammatory effect appeared in the CSF and hippocampus in IL-4 group. NF-κB signaling is the major pathway by which macrophage polarize towards M1 phenotype and F4/80+/NF-κB+ cells represent pro-inflammatory activation profile of macrophages (Wu et al. 2022a; Wang et al. 2014; Zhou et al. 2023). So we also observed the expression of NF-κB in dural macrophages by double labeling using immunofluorescence staining.

There were more F4/80+ and F4/80+/NF-κB+ cells in dural mater of IL-4 over-exposed rats (Paired samples t-test; p < 0.001; n = 10; Fig. 6) than in CON group (Paired samples t-test; p > 0.05; n = 10; Fig. 6). These data indicated more total macrophages as well as more pro-inflammatory activated macrophages in dura mater IL-4 over-exposed rats at P42.

Fig. 6

More F4/80+ and F4/80+/NF-κB+ cells in dural mater of IL-4 group. A, B Images showing the whole dural mater and the selected scopes shown as C and D in higher magnification. C, D Images showing the selected scopes containing the medial parts of both transverse sinuses and the selected scopes shown as E and F in higher magnification. E, F Images showing the selected scopes containing the parts connecting both transverse sinuses where exist some F4/80+/NF-κB+ co-labeled cells (green: F4/80+; red: NF-κB; blue: nuclei). Scale bars, 500 μm in (A, B), 100 μm in (C, D), 50 μm in (E, F). (G, H) Dots represent the numbers of F4/80+ and F4/80+/NF-κB+ cells in each group of rats. n = 10 rats/group. ***: p < 0.001. Paired samples t-test. Data are presented as the mean ± SD

Neonatal IL-4 Over-Exposure Increased Significantly the Number of Microglia in the Hippocampus and Induced their Activation on P49

The numbers and activation of microglia in the hippocampus have also been observed on P49 when the pro-inflammatory cytokines had a largest extent of increase as indicated by the data shown in Fig. 2. As shown in Fig. 7, there were significantly more microglia (Paired samples t-test., t = 6.25; p < 0.001; n = 8 rats/group) in the hippocampus of the rats in IL-4 group than the CON group. The activation of microglia were confirmed through their morphological changes of amoeboid that were more obvious and of a larger number in IL-4 group.

Fig.7

Neonatal IL-4 over-exposure increased significantly the number of microglia in the hippocampus and induced their activation on P49. A, B Representative micrographs showing the microglia (green signal: Iba-1+) in the hippocampus in CON rats (A) and IL-4 rats (B) at the age of P49. Scale bar, 100 μm. C Data represent the numbers of microglia in the hippocampus of each group of rats. The data in (C) represent the means ± SD. *** p < 0.001; n = 8 rats/group; Paired samples t-test. HIP hippocampus

Neonatal IL-4 Over-Exposure Increased M-CSF Slightly in the Hippocampus on P35 and P49.

To further evaluate hippocampal inflammation induced by neonatal IL-4 over-exposure, another main pro-inflammatory cytokines, including IFN-γ, IL-2, IL-8, IL-17 and M-CSF in the hippocampus, were also detected on P35, P49, P63. These three time points were chosen because they spread the time span from just before the emerging of the hippocampal inflammation, indicated by hippocampal levels of IL-1β, IL-6 and TNF (Fig. 2), and throughout its lasting period.

The data showed that only M-CSF has a slight increase (about 1.1 fold compared with CON mice) on P35 and P49 (Paired samples t-test with Bonferroni-Holm correction; P35, p < 0.001, Bonferroni–Holm α = 0.0167; P49, p < 0.001, Bonferroni–Holm α = 0.0250; P63, p = 0.1220, Bonferroni–Holm α = 0.0500; n = 15; Fig. 8A). The other cytokines shown no differences between groups (Paired samples t-test with Bonferroni-Holm correction, all p values > 0.05 and all p values > Bonferroni–Holm α values; n = 15; Fig. 8D–E).

Fig. 8

Neonatal IL-4 over-exposure increased M-CSF slightly in the hippocampus on P35 and P49. (A–E) The dots represent the average levels of each of the cytokines in the hippocampus. The data represent the means ± SD. *p value < Bonferroni–Holm α value; n = 15/group; Paired samples t-test was used. HIP hippocampus

These results suggested that the increased numbers of microglia in the hippocampus on P49 (Fig. 7) may be associated with the higher level of M-CSF in the hippocampus.