MPC-n (IgG) improves long-term cognitive impairment in the mouse model of repetitive mild traumatic brain injury

IgG was purchased from Solarbio. N-(3-Aminopropyl) methacrylamide hydro-chloride (APM) was purchased from Macklin. 2-Methacryloyloxyethyl phosphorylcholine (MPC), ethylene glycol dimethyl acrylate (EGDMA), fluorescein isothiocyanate (FITC), and 2-(1E, 3E, 5E)-5-(3-(6-(2, 5-dioxopyrrolidin-1-yl) oxy)-6-oxohexyl)-1 (Cy5.5) were purchased from Sigma-Aldrich. Ammonium persulfate (APS), N, N, N′, N′-tetramethylethylenediamine (TEMED) were obtained from Alfa Aesar. All reagents were used without purification.

The preparation principle of nano-microcapsules is to carry out in situ free radical polymerization on the surface of IgG. The specific process is as follows: dissolving IgG in 1 mL phosphate buffer (PBS, pH 7.4), adding APM, and MPC to mix evenly, and then adding cross-linking agent EGDMA, molar ratios of free IgG to APM to MPC to EGDMA = 1: 300: 7000: 600. After mixing for 10 min, the catalyst TEMED and initiator APS (APS: IgG = 300, n/n; TEMED: APS = 2:1, w/w) [16, 17]. The reaction will last 4 h. At the end of the reaction, the nanocapsules with MPC as monomers were obtained. After that, we use dialysis bags (MWCO: 8000–14,000, Solarbio) to remove the free monomer. It was used for dialysis in phosphate buffer saline (10 mM PBS, pH = 7.4). The fresh PBS solution was replaced every 6 h, and the nanocapsules solution was obtained after 48 h of dialysis. And then, to remove unencapsulated proteins, we passed the solution through a hydrophobic interaction column (Phenyl-Sepharose CL-4B, Solarbio, laboratory reagent). The purified nanocapsules were stored at 4 °C.

The particle size of the nanoparticle solution after 1 mL (1 mg/mL) dialysis was measured by Brookhaven's BI-90 Plus Zeta PALS analyzer [18]. The morphology of nanocapsules was characterized by using the JEM-2100F field transmission electron microscope (TEM) with an acceleration voltage of 200 kV. The solution of nanocapsules (0.01 mg/mL) was dripped onto the copper mesh and it was stained with 2% (w/v) phosphotungstic acid solution, washed with deionized water, fully dried, and then observed under TEM. The chemical groups on the surface of nano-microcapsules were analyzed by the Fourier transform infrared (FT-IR). The freeze-dried sample was mixed with potassium bromide and fully ground, and IgG, MPC-n (IgG) and gel without IgG were prepared for scanning [19]. The scanning range is 400–4000 cm⁻¹. The freeze-dried sample MPC, IgG, and MPC-n (IgG) were dissolved in 0.6 mL D₂O, and proton nuclear magnetic resonance (¹H NMR) spectra were measured using AVANCE IIITM HD 400 MHz NanoBAY (Bruker). We can also scan them with a UV–vis spectrophotometer (AOE instruments, A360 spectrophotometer). The range is 200⁻500 nm. The freeze-dried MPC-n (IgG) was added into the solution of PBS (pH = 7.4) to form 1 mg/mL, and the same in H₂O, DMEM, and serum at 37 °C for 48 h. The stability of MPC-n (IgG) was quantified using the scattering light intensity ratio I/I0 by dynamic light scattering (DLS) [20]. Because the crosslinker has an ester group with pH response, we set up two groups of nanocapsules under different pH (pH = 6.5, pH = 7.4) and measured the light scattering intensity I₀ of the two groups of solutions respectively. After the determination, both groups of samples were incubated at 37 °C for 60 min. In the process of incubation, the DLS intensity I was measured by the solution of nanocapsules at different time points, and the enzymatic degradation kinetics of nanocapsules were detected by I/I₀.

Adult male C57BL/6 J mice (aged 8–10 weeks old, weighing 20–25 g) purchased from Hfk Bioscience company (Beijing, China) were housed at the Experimental Animal Laboratories of Tianjin Neurological Institute. They were randomly fed food and water with a 12-h light/dark cycle. All experimental operations were allowed by the Animal Ethics Committee of Tianjin Medical University.

The mice were anesthetized with 4.6% isoflurane and fixed to an acrylic mold. A concave metal disc was placed caudal to bregma on the shaved head of mice, and the impounder tip of controlled cortical impact (electronic CCI model 6.3, American Instruments, Richmond, VA, USA) was positioned at the center of the disc surface, which is discharged at 5 m/s with a head displacement of 5 mm. Mice were divided into four groups: sham, rmTBI, rmTBI + IgG, and rmTBI + MPC-n (IgG) group. Injured mice were impacted 4 times with a 48-h interval. The sham group underwent the same operating procedures without any impact. After the last impact, the rmTBI mice were respectively administrated IgG (600 mg/kg) and MPC-n (IgG) (120 mg/kg) through the tail veil.

The distribution of Cy5.5-labeled MPC-n (IgG) and IgG in mice was observed using the in vivo imaging system (IVIS Lumina II, PerkinElmer, USA) 2, 6, and 24 h post-injection. Mouse brain tissues of all groups were collected 24 h post-injection. Fluorescence intensity values were acquired and analyzed using the Living Image software version 3.1 (Caliper Life Sciences) [17].

Mouse brains were harvested following cardiac perfusion and fixed in 4% PFA for 24 h. The dehydrated tissues were frozen rapidly in liquid nitrogen and sliced into 6-μm-thick sections with the freezing microtome (Leica CM 1950). Sections were blocked with 0.3% TritonX and 5% Albumin Bovine V for 1 h. Sections were incubated with primary antibodies overnight at 4 °C (Iba1, 1:500, Abcam; Phospho-Tau, 1:200, Cell Signaling Technology; Phospho-p38 MAPK, 1:200, Cell Signaling Technology). After being washed in PBS, sections were incubated with Alexa Fluor conjugated secondary antibodies (AlexaFluor 488/555, 1:500, Life Technologies) and stained with 4′,6-diamidino-2-phenylindole (DAPI, Sigma).

Brain tissues were obtained at 28 DPI and total protein was lysed with RIPA lysis buffer (Solarbio), PMSF (Solarbio), and phosphatase inhibitor (Sigma). Equally loaded proteins were electrophoretically separated on 10% and 15% SDS-PAGE gels and then transferred to PVDF membranes (Millipore). After being blocked by silk milk, membranes were incubated with primary antibodies for Phospho-Tau, Iba1, Phospho-p38 MAPK (1:1000, Cell Signaling Technology), and GAPDH (1:1000, Abcam). Horseradish peroxidase-conjugated secondary antibodies were used as a chromogenic reagent (1:5000, Zhongshanxinqiao).

The cortex and hippocampus of mice were fixed in 2.5% glutaraldehyde for 24 h and in 1% osmium tetroxide for 1.5 h at 4 °C. After dehydrated, the tissues were cut into ultra-thin slices and observed under transmission electron microscopy as previously described (TEM, HIT CHI-HT7700) [21].

Mice were anesthetized with isoflurane in a 9.4 T small-bore animal scanner (Bruker bio spec 94/30 USR) at 28 DPI and 42 DPI. The body temperature of mice was monitored and maintained at 37 ± 1° C. T2-weighted images were captured according to the following parameters: repetition time = 2500 ms, echo time = 33 ms, rare factor = 8, the field of vision = 20 × 20 mm, matrix size = 256 × 256, slice thickness = 0.5 mm, scanning time: 2 min 40 s [22].

Morris water maze test was used to evaluate the learning ability and spatial memory of mice. The mice at 28 DPI were placed into the pool from four quadrants to search for the underwater platform in the 90 s. After 90 s, mice that had not found the platform were guided to the platform and stayed for 20 s. Four trials per day for six consecutive days later, the platform was removed on the testing day, and the computer recorded the swimming track, dwelling time, and path length of mice [23].

Mouse brains of all groups were used for cytokine profiling. The relative levels of different cytokines were assessed by Proteome Profiler Mouse Cytokine Array Panel A (R&D Systems). The densitometry of each spot was measured using the ChemiDo XRS + imaging system (Bio-Rad, CA, USA), and the pixel density was evaluated by ImageJ Software.

RNA samples were sent to Shanghai Bohao Biotechnology Co., Ltd., China for cRNA library preparation and RNA sequencing. Total RNA from the samples was extracted by the Animal Total RNA Extraction Kit (Magnetic Bead method, MJYHIVD). Among them, RNA quantity and quality were assessed using Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, US) and RNAClean XP Kit (Cat A63987, Beckman Coulter, Inc. Kraemer Boulevard Brea, CA, USA) and RNase-Free DNase Set (Cat#79,254, QIAGEN, GmBH, Germany) were used for purification. All samples used Illumina NovaSeq6000 sequencer, model: PE150.

FASTQ data were processed using Seqtk (https://​github.​com/​lh3/​seqtk) to extract qualified raw read sub-data (the amount of data is about 6G/sample, and the ratio of base quality greater than 20 (Q20) in each direction is no less than 90%). We used the spliced mapping algorithm of Hisat2 (version: 2.0.4) [24] to perform genome mapping on the preprocessed reads, where the mapped genome is “GRCm38.p4 (mm10)”(ftp://​ftp.​ensembl.​org/​pub/​release83/​fasta/​mus_​musculus/​dna/​Mus_​musculus.​GRCm38.​dna.​primary_​assembly.​fa.​gz), the parameter defaults. Then “Stringtie” (version: 1.3.0) was used to count the “Fragments” of the mapped genes [25, 26], and after normalization using the TMM method [27], the FPKM value of each gene was calculated using the perl script.

All downstream analyses were performed in R version 3.6.3. “edgeR” was used to screen for differential genes between samples. Genes with log2|FC|> 0.2 and p < 0.05 identified were identified as differentially expressed genes (DEGs) [28]. Using “umsp” (version 0.2.7.0) to observe heterogeneity between data samples, DEG volcano plots and heatmaps were visualized using the “ggplot2” and “ComplexHeatmap” packages, respectively.

Gene Ontology (GO) [29] and KEGG enrichment analysis [30] was conducted to detect those DEGs’ function. The “GOplot” package and “cluster profiler” are used to visualize the enrichment results.

Nanocapsules are IgG encapsulated by MPC monomer, and then MPC-n (IgG) is formed by adding pH-responsive cross-linking agent EGDMA under the principle of in situ radical polymerization (Fig. 1A). The particle size of the nanocapsules is about 18 nm (± 2 nm). Compared with the particle size of IgG, the particle size of nanocapsules is about 18.5 nm, which increases obviously. The reason for the increase in particle size is that a polymer shell coated with protein is formed on the surface of IgG by in situ free radical reaction, which makes the whole spherical particles larger, and the experimental results are consistent with the analysis, which proves the successful preparation of nanocapsules. This is consistent with the results shown in the TEM photos that the nanocapsules are more regularly spherical (Fig. 1B, C). The surface morphology of the dialyzed nano-microcapsules was analyzed by TEM. From the TEM photos of nanocapsules, we can see that the morphology of MPC-n (IgG) is mostly spherical particles. Compared with the scale, we can see that the particle size is about 20 nm, which is consistent with the particle size measured by our particle size meter. The formation of spherical particles is mainly due to the negative charge of protein in phosphate buffer, while the positively charged monomer APM is a kind of acrylic hydrochloride. Through the interaction of static electricity and hydrogen bond, APM, MPC monomer, and cross-linking agent carry out in situ free radical polymerization on the protein surface under the action of the initiator to form a three-dimensional network structure. The research group proved earlier that if the monomer does not form a shell structure on the protein surface, only the polymerization between monomers will not form a spherical structure, which further illustrates the formation of spherical nanocapsules.

From the FT-IR (Fig. 1D), the absorption peaks identified by 1726 cm⁻¹, 1240 cm⁻¹, 1084 cm⁻¹, and 960 cm⁻¹ belong to the ester absorption peak of-COOCH₂-, the absorption peak of P = O, P-O-C, and the antisymmetric stretching vibration peak of C-N. Because the polymer material on the surface of MPC-n (IgG) is mainly PMPC [31], the appearance of these four peaks verifies its existence, while the composition of other monomers and cross-linking agents is difficult to see, mainly due to the small proportion added.

In the ¹H NMR spectra (Additional file 1: Fig. S1A), the main proton peaks of MPC-n (IgG) were located at 1.8, 3.1, and 4.1–4.3 ppm, which corresponded to the characteristic peaks of poly 2-methacryloyloxy ethyl phosphorylcholine (PMPC). Because the protein has a characteristic peak in 280 nm, in the ultraviolet spectrum (Additional file 1: Fig. S1B), IgG and MPC-n (IgG) will have a characteristic peak in 280 nm, while gel without IgG will not have a characteristic peak, which proves that nanocapsules do contain IgG.

We further confirmed the degradation behavior of MPC-n (IgG) by DLS (Fig. 1E). Compared with pH = 7.4, in the environment of pH = 6.5, with the extension of culture time, I/I₀ of MPC-n (IgG) gradually decreased, indicating that the particle size of MPC-n (IgG) was decreasing, which further confirmed that it could be degraded.

To investigate the stability of MPC-n (IgG), it was dissolved in H₂O, PBS, DMEM, and serum at 37 °C [17], and the size of MPC-n (IgG) was measured by DLS for 48 h. As shown in Additional file 1: Fig. S1C, the MPC-n (IgG) remained stable in H₂O, PBS, DMEM, and serum at 37 °C for 48 h. These findings suggested that the encapsulation of IgG enabled controlled release and enhanced its stability. The modeling process and disease progression of rmTBI are shown in Fig. 1F.

Given the good stability and BBB targeting of microcapsules, we first assessed the delivery capacity of MPC-n (IgG) in rmTBI mice. IgG and MPC-n (IgG) coupled with Cy5.5 fluorescence were intravenously injected into rmTBI mice after the last impact. In a vivo imaging system, it exhibited a higher Cy5.5 signal in the mice brains of the MPC-n (IgG) group compared to the IgG group at 2 h, 6 h, and 24 h post-injection (Fig. 2A). In addition, we obtained the mouse brain tissue at 24 h for fluorescence detection, which displayed a consistent result. To further investigate the spatial differences in IgG distribution, the distribution of Cy5.5-coupled IgG and MPC-n (IgG) 24 h post-injection was observed under a confocal microscope, which showed that MPC-n (IgG) was more significantly distributed in the cortex and hippocampus compared to the free IgG (Fig. 2B). These results illustrated that MPC-n (IgG) significantly enhanced BBB permeability of IgG and increased the specific accumulation in the cortex and hippocampus of rmTBI mice.

The main pathological manifestation of rmTBI is long-term cognitive dysfunction. Thus, Morris water maze was performed to determine spatial cognition and memory function in rmTBI mice at 28 DPI [32]. On the testing day, compared with the IgG group, the MPC-n (IgG) group had longer dwelling time in the target quadrant (Fig. 3A), shorter latency to first target-site (Fig. 3B), and increasing number of platform-site crossovers (Fig. 3C). These results suggested that MPC-n (IgG) can significantly improve long-term cognitive and memory function more than free IgG.

Due to the sensibility of MRI in anatomical detail, it was performed to further measure the hippocampal volume changes. Although there was no significant statistical difference among the groups at 28 DPI (Fig. 3D), MPC-n (IgG) showed a therapeutic effect in hippocampal atrophy at 42 DPI (Fig. 3E), while IgG had no distinct benefit at each time point. These results indicated that the hippocampal lesions were still progressing after rmTBI in the long term, and MPC-n (IgG) could ameliorate cognitive and memory dysfunction through mitigating hippocampal atrophy.

Tau protein is an axonal protein that can stabilize and bundle microtubules [33]. However, excessive phosphorylation of Tau protein results in its shedding from the axon and formation of neurofilament tangles, which was considered a key factor in rmTBI-related dementia [34, 35]. The expression of p-Tau protein was observed by immunofluorescence staining, which suggested that p-Tau protein was mainly deposited in the hippocampus (Fig. 4A). Meanwhile, MPC-n (IgG) markedly reduced the deposition of p-Tau in the hippocampus of mice (Fig. 3A, B), which was in line with the result of Western Blot (Fig. 3C, D). In mTBI, shearing and strain forces can directly contribute to multifocal axonal injury [36]. Myelin surrounding the axons of neurons acts as an insulator and increases the conduction speed of nerve impulses, as well as protecting the axons. TEM was used to observe the changes in the ultrastructure of the myelin sheath. As the TEM displayed, the lamellar structure of the myelin sheath was destroyed and the cytoplasm of the neuron was dissolved in both cortex and hippocampus of rmTBI, while MPC-n (IgG) obviously amended the myelin injury compared with free IgG (Fig. 4E).

Taken together, MPC-n (IgG) can more effectively remove p-Tau deposition and remit myelin degeneration than free IgG.

Continuous chronic neuro-inflammation often leads to secondary injury and neurodegeneration such as mild cognitive impairment [37, 38]. Microglia acts as the first line of defense in the immune response in the central nervous system, and its activation can last for several months [39]. To evaluate the state of microglia, immunofluorescence was used to detect the expression of activated microglia marker Iba1 in the cortex and hippocampus at 28 DPI. The results exhibited that MPC-n (IgG) was more effective in restraining microglia activation in both the cortex (Fig. 5A, B) and hippocampus (Fig. 5C, D). In addition, cytokines were quantified by Array at 42 DPI, which demonstrated the descent of pro-inflammatory factors under the treatment of MPC-n (IgG), including M-CSF, CXCL12, CCL2/MCP-1, IFN-γ, CCL12, and TNF-α (Fig. 5E, F).

All these results proved that MPC-n (IgG) showed a higher capacity of anti-inflammatory compared with free IgG.

In the differential analysis, the results showed the differential genes between the rmTBI group and the sham group (Fig. 6A, C), including 3,574 differential genes, 1794 upregulated genes, and 1780 downregulated genes. Compared with the MPC-n (IgG) group and the rmTBI group, a total of 1625 differential genes were found, including 811 upregulated genes and 814 downregulated genes (Fig. 6B, D). The completed differential gene profiles can be found in Supplementary Tables S1 and S2.

Functional enrichment analysis demonstrated that the changed pathways in the rmTBI group were mainly MAPK signaling pathway, PI3K-Akt pathway, and cell matrix-related pathways compared with the sham group (Fig. 6E). Interestingly, MAPK signaling pathway was also enriched in the MPC-n (IgG) and rmTBI groups (Fig. 6F). Previous studies have shown that MAPK pathway could play an important role in rmTBI [40, 41]; thus, we marked the differential genes in MAPK pathway. As displayed in Additional file 2: Fig. S2A and B, p-p38 MAPK pathway was the most critical gene which upregulated in the rmTBI group compared with the sham group and was lessened in the MPC-n (IgG) group, suggesting that p-p38 MAPK was an underlying treatment target of MPC-n (IgG).

It was reported that the p38-MAPK pathway contributed to the phosphorylation of Tau protein and synaptic damage [42, 43], and a clinical study showed that the p38α-MAPK inhibitor could improve memory function in AD patients [44]. In addition, the phosphorylation of MAPK in neurons may be a kind of stress response, which turns to promote the activation of microglia and the release of pro-inflammatory factors [45].

According to the sequencing results, we further verified the expression of MAPK pathway in neurons. Western Blot showed that MPC-n (IgG) could effectively reduce the expression of p-p38 MAPK in the hippocampus of rmTBI mice (Fig. 7A, B), and immunofluorescence illustrated that p-p38 MAPK co-localized with neurons in the hippocampus (Fig. 7C), suggesting MPC-n (IgG) can act directly on neurons to exert therapeutic effects (Fig. 8).