p38α‐MAPK‐deficient myeloid cells ameliorate symptoms and pathology of APP‐transgenic Alzheimer's disease mice

Abstract Alzheimer's disease (AD), the most common cause of dementia in the elderly, is pathologically characterized by extracellular deposition of amyloid‐β peptides (Aβ) and microglia‐dominated inflammatory activation in the brain. p38α‐MAPK is activated in both neurons and microglia. How p38α‐MAPK in microglia contributes to AD pathogenesis remains unclear. In this study, we conditionally knocked out p38α‐MAPK in all myeloid cells or specifically in microglia of APP‐transgenic mice, and examined animals for AD‐associated pathologies (i.e., cognitive deficits, Aβ pathology, and neuroinflammation) and individual microglia for their inflammatory activation and Aβ internalization at different disease stages (e.g., at 4 and 9 months of age). Our experiments showed that p38α‐MAPK‐deficient myeloid cells were more effective than p38α‐MAPK‐deficient microglia in reducing cerebral Aβ and neuronal impairment in APP‐transgenic mice. Deficiency of p38α‐MAPK in myeloid cells inhibited inflammatory activation of individual microglia at 4 months but enhanced it at 9 months. Inflammatory activation promoted microglial internalization of Aβ. Interestingly, p38α‐MAPK‐deficient myeloid cells reduced IL‐17a‐expressing CD4‐positive lymphocytes in 9 but not 4‐month‐old APP‐transgenic mice. By cross‐breeding APP‐transgenic mice with Il‐17a‐knockout mice, we observed that IL‐17a deficiency potentially activated microglia and reduced Aβ deposition in the brain as shown in 9‐month‐old myeloid p38α‐MAPK‐deficient AD mice. Thus, p38α‐MAPK deficiency in all myeloid cells, but not only in microglia, prevents AD progression. IL‐17a‐expressing lymphocytes may partially mediate the pathogenic role of p38α‐MAPK in peripheral myeloid cells. Our study supports p38α‐MAPK as a therapeutic target for AD patients.


| INTRODUC TI ON
Alzheimer's disease (AD), the major cause of dementia in the elderly, is pathologically characterized by three components: (i) extracellular deposits of amyloidβ peptide (Aβ), (ii) intracellular neurofibrillary tangles (NFT) that is composed of hyper-phosphorylated tau protein (p-tau), and (iii) microglia-dominated inflammatory activation in the brain parenchyma . Interactions between Aβ, p-tau and inflammatory activation are primarily responsible for the progressive neurodegeneration in AD. However, many clinical trials to reduce Aβ accumulation or p-tau aggregation or inflammatory activation (Pleen & Townley, 2022) have failed to produce AD therapies that modify the disease progression. A simple explanation for these failures is that the study population may have already reached a disease stage too late for effective intervention. However, it is important to recognize that AD is a heterogeneous disease. For example, the pathological and biochemical features of Aβ deposits or molecular structure of Aβ aggregates in the brain (Thal et al., 2015) varies among AD patients. Variations in Aβ structure affect how microglia respond to the Aβ deposits, which, in turn, affects inflammatory activation and Aβ internalization (Parvathy et al., 2009). A growing number of subtypes of activated microglia have recently been identified in AD brains (Chen & Colonna, 2021). Moreover, pathological examination of postmortem brain tissues and imaging studies show different distributions of tau-related pathology and patterns of brain atrophy in AD patients (Ferreira et al., 2020). Therefore, targeting multiple pathogenic pathways might be more effective as a therapeutic intervention than focusing on a single step in AD disease progression. p38α mitogen-activated protein kinase (p38α-MAPK) is a protein kinase present in a variety of cells that respond to external stress stimuli (Kumar et al., 2003). p38α-MAPK is activated in both neurons and microglia in brains of AD patients (Hensley et al., 1999).
Our recent study indicates that p38α-MAPK deficiency in neurons reduces both Aβ and p-tau levels in the brain of AD mice (Schnöder et al., 2016(Schnöder et al., , 2020(Schnöder et al., , 2021. A systemic administration of chemical p38α-MAPK inhibitor has been observed to reduce inflammatory activation in the brain of APP-or tau-transgenic mice (Bachstetter et al., 2012;Maphis et al., 2016). Thus, p38α-MAPK inhibition might simultaneously target Aβ, p-tau and inflammation in AD. A recent phase 2 clinical trial showed that a 24-week treatment with p38α-MAPK inhibitor decreased tau proteins in the cerebral spinal fluid of mild AD patients; although it did not improve the cognitive function (Prins et al., 2021). We believe that the therapeutic protocol can be optimized, if the pathogenic mechanisms of p38α-MAPK are better understood. Pharmacological treatments with p38α-MAPK inhibitors affect both microglial p38α-MAPK and neuronal p38α-MAPK, without the ability to distinguish their effects. The inhibition of inflammatory activation in the brain might come from neuronal p38α-MAPK inhibition-mediated attenuation of Aβ and p-tau generation, or even from neuronal protection (Schnöder et al., 2020). In this study, we investigated specific effects of p38α-MAPK in microglia or myeloid cells on AD pathogenesis.
The pathogenic role of microglia in AD is extremely heterogeneous. For example, the rare variants in the triggering receptor expressed on myeloid cells-2 (TREM2) gene increase the risk of developing AD. One group reported that TREM2 deficiency in APPtransgenic mice increases hippocampal Aβ burden and accelerates neuron loss (Wang et al., 2015); however, another group showed that TREM2 deletion reduces cerebral Aβ accumulation (Jay et al., 2015). Subsequent work suggested that the effect of TREM2 deficiency on cerebral Aβ accumulation depends on the stage of disease (Jay et al., 2017). Consistent with this conclusion is the observation from a longitudinal imaging study of human subjects with mild cognitive impairment that several peaks of microglial activation appear over the disease trajectory (Fan et al., 2017). These studies underscore the effects of the changing cellular environment and reinforce the idea that the pathogenic role of microglial activation should be dynamically investigated during disease progression.
In this study, we conditionally knocked out Mapk14 gene (encoding p38α-MAPK) in the myeloid cell lineage or specifically in microglia in amyloid precursor protein (APP)-transgenic mice and investigated the AD pathology and microglial activation in early and late disease stages. We observed that deletion of p38α-MAPK attenuated Aβ load and neuronal deficits of AD mice; however, the pathogenic mechanism of p38α-MAPK is evolving during the disease progresses, which potentially involves peripheral interleukin (IL)-17a-expressing T lymphocytes. Mapk14 transcription nor p38-MAPK protein in CD11b+ brain cells, but decreased Mapk14 transcription by 88% in CD11b+ blood cells (Figure S1b-e).
We further constructed APP-transgenic green fluorescence protein (GFP)-expressing LysM-Cre reporter mice (APP tg ROSA mT/mG LysM-Cre +/− ; Muzumdar et al., 2007). GFP was mainly expressed in microglia associated with Aβ deposits ( Figure S1f); Aβ deposits were also surrounded by microglia without expression of GFP, indicating heterogeneity of Aβ plaques. GFP was rarely expressed in neurons ( Figure S1g). APP tg p38 fl/fl LysM-Cre +/− mice were also mated to CCR2-RFP reporter mice expressing red fluorescent protein (RFP) under the control of Ccr2 gene promoter (Saederup et al., 2010). Both histological and flow cytometric analysis showed that p38α-MAPK deficiency does not affect the recruitment of peripheral myeloid cells into the brain of 9-month-old APP-transgenic mice ( Figure S2).

| Deficiency of p38α-MAPK in myeloid cells improved the cognitive function of APP-transgenic mice
We used the Morris water maze test to examine cognitive function of 9-month-old APP tg and their non-APP-transgenic (APP wt ) littermate mice. During the acquisition phase, APP wt mice with or without deletion of p38α-MAPK in myeloid cells showed no significant differences in either swimming time or swimming distance before climbing onto the escape platform The swimming velocity did not differ between p38α-MAPKdeficient and wildtype APP-transgenic mice or for the same mice on different training dates ( Figure 1c).
Twenty-four hours after the end of training phase, the escape platform was removed and a 5-min probe trial was performed to test the memory of mice. Compared to APP wt p38α fl/fl LysM- We further used Western blot analysis to quantify the levels of four synaptic proteins: Munc18-1, synaptophysin, SNAP-25, and PSD-95 in the brain homogenate of 9-month-old APP tg and APP wt littermate mice. As shown in Figure 1f-k, protein levels of

| Deficiency of p38α-MAPK in myeloid cells reduces Aβ load in the brain of APPtransgenic mice
As Aβ is the key molecule leading to neurodegeneration in AD , we analyzed the effects of myeloid p38α-MAPK on Aβ pathology in the APP-transgenic mice. Using immunohistological and stereological Cavalieri methods, we observed that the volume of immunoreactive Aβ load in 9-month-old APP tg p38α fl/fl LysM-Cre +/− mice was significantly lower than that in  Figure S4).

F I G U R E 1
Deficiency of p38α-MAPK in myeloid cells improves cognitive function and attenuates AD-associated loss of synaptic proteins in APP-transgenic mice. (a-c) Nine-month-old APP-transgenic (APPtg) and non-APP-transgenic (APPwt) littermate mice with (p38α ko) and without (p38α wt) deletion of p38α-MAPK in myeloid cells were assessed for cognitive function using the water maze test. In the training phase, deficiency of p38α-MAPK decreases swimming distance (a) and latency (b) to reach the escape platform in APPtg but not in APPwt mice. Deficiency of p38α-MAPK does not affect the traveling velocity in APPtg mice (c). Two-way ANOVA from day 3 to day 6 followed by Bonferroni's post hoc test, n is shown in the figure. The latency of first visit to the region where the escape platform was previously located (d) and the frequency, with which mice crossed the platform region (e), were recorded in the 5-min probe trial. One-way ANOVA followed by Bonferroni's post hoc test. (f-k) The amount of synaptic proteins, Munc18-1, SNAP25, synaptophysin, and PSD-95 in the brain homogenate of 9-month-old APPtg and APPwt mice was determined using Western blotting. One-way ANOVA followed by Bonferroni's post hoc test, n ≥ 11 per group for APPtg mice and n ≥ 6 per group for APPwt mice. Here, representative Western blot images from five independent experiments are shown. Munc18-1 and SNAP15, PSD-95 and synaptophysin, and their corresponding α-tubulin immunoblots were performed on the same membrane. Data was represented as mean ± SEM

| Deficiency of myeloid p38α-MAPK differently regulates microglial inflammatory activation in early and late disease stages of APPtransgenic mice
Inflammatory activation of microglia is another pathogenic factor in AD . After immunofluorescent staining of Iba-1, we used the stereological method, Optical Fractionator probe, to count microglia in the hippocampus and cortex. Deficiency of p38α-MAPK in myeloid cells significantly decreased Iba-1-immunoreactive microglia in both 4 and 9-month-old APP-transgenic, but not in 9-month-old non-APP-transgenic mice (Figure 3a,b,x). We also observed that deficiency of p38α-MAPK decreased the number of P2RY12-immunoreactive microglia in the hippocampus ( Figure S5).
It has been reported that P2RY12 is a more specific protein marker for endogenous microglia (McKinsey et al., 2020).

| Deficiency of myeloid p38α-MAPK increases microglial clearance of Aβ in APP-transgenic mice at the late disease stage
Microglia play like a double-edged sword. Their uptake of Aβ is an important mechanism of Aβ clearance in AD brain . We asked whether deficiency of p38α-MAPK facilitates microglial internalization of Aβ in AD mice. After observing that there were more microglia surrounding Aβ deposits in 9-month-old In 4-month-old APP-transgenic mice, we repeated all experiments for 9-month-old mice. We observed that p38α-MAPK deficiency neither altered the intracellular Aβ in microglia, nor affected the transcription of Aβ internalization-associated receptors, including SR-A, CD36, and RAGE ( Figure S1d and Figure S8).
In order to verify our in vivo observation that p38α-MAPK deficiency enhances Aβ internalization in microglia, we cultured p38α-MAPK-deficient and wildtype bone marrow-derived macrophages Co-treatment with fucoidan, an antagonist of SR-A, abolished p38α-MAPK deficiency-enhanced Aβ internalization ( Figure S9g,h).

| Deficiency of p38α-MAPK specifically in microglia reduces AD-associated pathologies in the brain of APP-transgenic mice, but with low efficiency
After observing that deficiency of p38α-MAPK in whole myeloid cells prevented AD progression, we asked whether p38α-MAPK deficiency specifically in microglia served the same beneficial effects. A second AD mouse model was constructed by cross-breeding APP tg mice with p38 fl/fl mice and Cx3Cr1-CreERT2 mice as we did in a recent study (Quan et al., 2021). Six or nine-month-old and Mrc1 genes in microglia of APP-transgenic mice (Figure 5j-o).
However, as the same as in 9-month-old APP tg p38 fl/fl LysM-Cre +/− mice, p38α-MAPK deficiency specifically in microglia also promoted the accumulation of Iba-1-positive microglia around Aβ deposits in both cortex and hippocampus of 12-month-old APP tg p38 fl/fl Cx3Cr1-Cre +/− mice compared with p38α-MAPK-wildtype AD mice (Figure 5p,q). We also observed that deletion of p38α-MAPK in microglia at 6 to 12 months of age significantly decreased Aβ deposition in the hippocampus but not in the cortex (Figure 5r-t). Concentrations of both Aβ40 and Aβ42 in RIPA-soluble brain homogenates quantitated by ELISA did not differ between APP tg p38 fl/fl Cx3Cr1-Cre +/− and APP tg p38 fl/fl Cx3Cr1-Cre −/− mice (Figure 5u,v).
As APP tg p38 fl/fl Cx3Cr1-Cre +/− mice were haploinsufficient for Cx3cr1 gene, additional experiments were performed to examine whether Cx3Cr1 haploinsufficiency affects AD pathogenesis.
Our recent study showed that Cx3Cr1 haploinsufficiency does not change Aβ deposition and inflammation in the brain of APPtransgenic mice (Quan et al., 2021). Our current study indicated that haploinsufficiency of Cx3Cr1 altered neither the recruitment of microglia toward Aβ deposits, nor the transcription of inflammatory genes, Tnfα, Il-1β, Ccl-2 and Il-10 in individual microglia ( Figure S10).
In further experiments, we observed that microglial deficiency of p38α-MAPK attenuated the cognitive deficits of 12-month-old APPtransgenic mice in Morris water maze test; however, p38α-MAPK deficiency did not prevent the loss of synaptophysin, Munc18-1, PSD-95, and SNAP25 in APP-transgenic AD mice ( Figure S11). F I G U R E 2 Deficiency of p38α-MAPK in myeloid cells reduces Aβ load in the brain of APP-transgenic mice. Four and 9-month-old APPtransgenic mice with (p38α ko) and without (p38α wt) deletion of p38α-MAPK in myeloid cells were analyzed with stereological Cavalieri methods for cerebral Aβ volumes (adjusted by the volume of analyzed tissues) after immunohistochemical (a, b), Congo red (c, d) and immunofluorescent (h, i) staining. T test, n ≥ 7 per each group. (e, f) Aβ was also detected in brain homogenates of 9-month-old p38α-MAPKko and -wt APP-transgenic mice with Western blot. Here, representative Western blot images from two independent experiments are shown. Aβ and β-Actin immunoblots were performed on the same membrane. T test, n ≥ 4 per each group. (g) APP-transgenic mouse brains were further serially homogenized in TBS-, TBS-T-, and guanidine-soluble fractions, in which monomeric, oligomeric and high-molecularweight Aβ aggregates were enriched, respectively. Aβ40 and Aβ42 were measured by ELISA and normalized to the amount of homogenate protein. T test, n ≥ 10 per group. Data was represented as mean ± SEM F I G U R E 3 Deficiency of p38α-MAPK in myeloid cells differently regulates microglial inflammatory activation in the brain of APP-transgenic mice at early and late disease stages. (a, b, x) microglia stained with red fluorescent Iba-1 antibody were counted with the optical fractionator stereological probe in brains of 4 and 9-month-old APP-transgenic (APPtg) and non-transgenic (APPwt) mice with (p38α-ko) and without (p38α-wt) deletion of p38α-MAPK in myeloid cells.

|
The cell number was adjusted by the volume of analyzed tissues. One-way ANOVA followed by Bonferroni's post hoc test for 9-month-old mice, n ≥ 7 per group for APPtg mice and ≥3 per group for APPwt mice; t test for 4-month-old mice, n ≥ 7 per group. (a) Images show the immunofluorescent staining of 9-monthold mouse brains. (c-j) The inflammatory gene transcripts in brains of 9-month-old APPtg mice were measured with real-time PCR. T test, n ≥ 8 per group. (k, l, y, z) four and nine-month-old APP tg p38 fl/fl LysCre +/− and APP tg p38 fl/fl LysM-Cre −/− mice were further analyzed with Western blot for the levels of phosphorylated (Tyr705; p-) and total (t-) Stat3 in the brain. The same membrane was serially blotted with antibodies against p-Stat3, t-Stat3 and β-Actin. The activity of Stat3 is shown in the ratio of p-/t-Stat3. T test, n ≥ 7 per group for 9-month-old mice and n = 5 per group for 4-month-old mice. (y) To avoid overexposure of the film, the membrane for t-Stat3 blotting was additionally washed for 17 h after the first exposure to the film after 1 h of washing. Here, representative Western blot images from two independent experiments are shown. (m-w, aa-ad) in following experiments, microglia were selected from brains of 4-and 9-month-old p38α-wt and ko APPtg mice. The transcriptional level of inflammatory genes and other DAMassociated genes was determined by real-time PCR. T test, n ≥ 6 and 4 per group for 9 and 4-month-old mice, respectively.

Data was represented as mean ± SEM
We then isolated CD4+ spleen cells from 4-and 9-month-old APP tg p38 fl/fl LysM-Cre +/− and APP tg p38 fl/fl LysM-Cre −/− littermate mice, and observed that p38α-MAPK deficiency significantly reduced the transcription of Il-17a, but not Ifnγ, Il-4 and Il-10 in AD mice at the age of 9, but not 4 months (Figure 6a-d). Thus, we hypothesized that IL-17a-expressing cells might be involved in cerebral F I G U R E 4 Deficiency of p38α-MAPK in myeloid cells promotes microglial internalization of Aβ in the brain of 9-month-old APPtransgenic mice. (a, b) Brain sections from 9-month-old p38α-MAPK-deficient (ko) and wildtype (wt) mice were stained for microglia with green fluorescent Iba-1 antibodies and for Aβ deposits with Congo red. Under the red channel, total 105 Aβ deposits in p38α-MAPK-ko mice and 76 Aβ deposits in p38α-MAPK-wt mice were randomly chosen. Microglia with clear DAPI-stained nuclei and with contact to Aβ deposits were counted. The number of microglia was adjusted by the area of Congo red-positive Aβ deposits. T test, n = 5 and 4 for p38α-MAPK-ko and wt APP-transgenic mice, respectively. (c, d) Adult microglia were also isolated from 9-month-old p38α-MAPK-ko and wt APP-transgenic mice, and quantified for intracellular Aβ by Western blot using human Aβ and β-Actin antibodies. As a control, no Aβ was detected in the microglia isolated from APP-wildtype mice (c). Here, representative Western blot images from three independent experiments are shown. Aβ and β-Actin immunoblots were performed on the same membrane. The overall picture of the Aβ-immunoblot is shown in Figure S3c. We cross-bred APP tg mice with IL-17a −/− mice (Nakae et al., 2002) and observed that the extent of immunoreactive Aβ deposits in both the cortex and hippocampus of 6-month-old APP tg IL17a −/− (IL-17a knockout) mice was significantly less than that in APP tg IL17a +/+ (IL-17a wildtype) littermates (Figure 6e-g).
To investigate whether IL-17a deficiency models p38α-MAPKdeficient myeloid cells in regulating microglial activation, we analyzed and compared the morphology of microglia surrounding Aβ deposits.
Deletion of p38α-MAPK in myeloid cells decreased the total number and end points of branches of microglial processes in 9-but not 4-month-old APP-transgenic mice (Figure 6h-p). In Sholl analysis, microglial branches crossed concentric circles significantly less in 9-month-old APP tg p38 fl/fl LysM-Cre +/− than in APP tg p38 fl/fl LysM-Cre −/− littermates, especially at 40-70 μm from the soma (Figure 6m).

| DISCUSS ION
We constructed two AD mouse models with deletion of p38α- receptor (Paresce et al., 1996), in microglia. We also observed that p38α-MAPK deficiency in myeloid cells inhibited IL-10-Stat3 signaling in the brain of APP-transgenic mice. It has been reported that deficiency of IL-10 or Stat3 facilitates microglial clearance of Aβ in AD mice (Guillot-Sestier et al., 2015;Reichenbach et al., 2019). In APP tg p38 fl/fl Cx3Cr1-Cre +/− mice, p38α-MAPK deficiency inhibited inflammatory activation in microglia, which may prevent p38α-MAPK deficiency from enhancing Aβ internalization. Indeed, mild inflammatory activation has the potential to increase Aβ clearance in the brain. Systemic injection of TLR4 or TLR9 agonists induces both pro-and anti-inflammatory activation and decreases Aβ in the brain of APP-transgenic mice (Michaud, Halle, et al., 2013;Scholtzova et al., 2014). TREM2 antibody administration also decreases Aβ load in the presence of increased expression of inflammatory cytokines and chemokines in the brain of APP-transgenic mice (Price et al., 2020). However, the mechanisms of inflammatory regulation of microglial Aβ clearance remain unclear. The gene transcription in microglia from our APP tg p38 fl/fl LysM-Cre +/− mice showed partial DAM signatures (e.g., induction of proinflammatory genes); however, transcription of homeostatic genes (e.g., Cx3cr1) was also upregulated and transcription of Trem2 gene was reduced.

MAPK in all myeloid cells (APP
Our study was not yet able to answer the question of how peripheral p38α-MAPK-deficient myeloid cells reduced cerebral Aβ in 4-month-old APP tg p38 fl/fl LysM-Cre +/− mice. The decrease in F I G U R E 5 Deficiency of p38α-MAPK specifically in microglia inhibits inflammatory activation and decreases Aβ load in the brain of APP-transgenic mice. (a) APP tg p38 fl/fl Cx3Cr1-Cre +/− and APP tg p38 fl/fl Cx3Cr1-Cre −/− littermate mice were injected with tamoxifen at 6 or 9 months of age and analyzed at 12 months. (b, c) Microglia in the hippocampus of microglial p38α-MAPK-deficient (μp38α-MAPK-ko) and wildtype (μp38α-MAPK-wt) APP-transgenic (APPtg) mice were stained with green fluorescence-conjugated Iba-1 antibodies and counted with the optical fractionator probe. The number of microglia was adjusted by the volume of analyzed brain tissue. T test, n ≥ 9 per group. (d-i) Inflammatory gene transcripts in brain tissues were measured with real-time PCR. As a control, non-APP-transgenic (APPwt) littermates were treated with tamoxifen, as in APPtg mice, to induce deletion of p38α-MAPK in microglia. One-way ANOVA followed by Bonferroni's post hoc test, n ≥ 11 and 4 per group for APPtg and APPwt mice, respectively. (j-o) CD11b+ cells were further selected from brains of 12-month-old APPtg mice. Transcripts of various inflammatory genes in microglia were quantified with real-time RT-PCR. T test, n ≥ 5 per group. (p, q) in following experiments, brain sections of 12-month-old μp38α-wt and -ko APPtg mice were stained with red fluorescent Iba-1 antibodies for microglia and with methoxy-XO4 (in green) for Aβ deposits. Microglia around Aβ deposits were counted and the number of microglia was adjusted by the area of Aβ deposits. T test, n ≥ 5 per group. (r-t) Finally, the coverage of Aβ deposits in the brain as stained by human Aβ antibodies was estimated with Cavalieri method and adjusted by the area of analyzed brain tissue. T test, n ≥ 10 per group. (u, v) Aβ40 and Aβ42 in RIPA-soluble brain homogenates of 12-month-old APPtg mice with and without deletion of p38α-MAPK in microglia were measured with ELISA. T test, n ≥ 6. Data was represented as mean ± SEM  (Erny et al., 2021), and may also indicate active Aβ internalization (Huang et al., 2021). Very interestingly, the morphological pattern of microglia in p38α-MAPK-deficient mice can be generated in IL-17a-deficient APP-transgenic mice. It is known that p38α-MAPK signaling in dendritic cells drives differentiation of T helper 17 (Th17) cells and sustains autoimmune inflammation (Huang et al., 2012). We have observed that APP is expressed in myenteric neurons of the gut (Semar et al., 2013) and is able to increase IL-17a expression in CD4-positive gut lymphocytes ( Figure S12). During disease progression, AD pathology in the gut is sufficient to induce differentiation and activation of T lymphocytes and myeloid p38α-MAPK has the opportunity to alter the immune response. Our study suggests that IL-17a may at least partially mediate the pathogenic role of myeloid p38α-MAPK in AD pathogenesis. It has been reported that the number of Th17 cells increases in the blood of AD patients (Oberstein et al., 2018). IL-17a-expressing T lymphocytes accumulate in the meanings and brain of triple-transgenic AD mice (3× Tg-AD; Brigas et al., 2021). It is worthwhile to reanalyze the pathogenic effects of myeloid p38α-MAPK in AD mice on the basis of IL-17a deficiency in our future studies.
It should be noted that deficiency of p38α-MAPK promotes the recruitment of microglia around Aβ deposits in both APP tg p38 fl/fl LysM-Cre +/− and APP tg p38 fl/fl Cx3Cr1-Cre +/− mice, possibly favoring Aβ clearance (Hao et al., 2011;Quan et al., 2021). It has also been suggested that microglia clustered around Aβ deposits protect local neurites from damage by forming a physical barrier and condensing Aβ into dense plaques (Condello et al., 2015). Indeed, we observed p38α-MAPK deficiency in all myeloid cells as well as specifically in microglia protecting neurons in APP-transgenic mice, albeit with varying efficiency. The mechanism that drives microglia to migrate to Aβ deposits needs to be further identified.
In summary, deficiency of p38α-MAPK in all myeloid cells, not just microglia, triggers efficient Aβ clearance in the brain and improves cognitive function of APP-transgenic mice. Together with our previous observations that neuronal deficiency of p38α-MAPK reduces Aβ and phosphorylated tau proteins in the brains of AD mice (Schnöder et al., 2016(Schnöder et al., , 2020(Schnöder et al., , 2021, our serial studies support that inhibition of p38α-MAPK is a novel therapeutic option targeting multiple pathogenic processes in AD. As a potential anti-AD mechanism, F I G U R E 6 Deficiency of p38α-MAPK in myeloid cells and knockout of IL-17a similarly modify microglial morphology and reduce Aβ deposits in APP-transgenic mice. (a-d) CD4-positive spleen cells were selected from 4-and 9-month-old APP-transgenic mice with (ko) and without (wt) deletion of p38α-MAPK in myeloid cells. Real-time PCR was used to quantify transcripts of marker genes for Th17 (IL-17a), Th1 (Ifnγ), Th2 (IL-4) and regulatory T (Treg) lymphocytes (Il-10). T test, n ≥ 3 per group. (e-g) To investigate the pathogenic role of IL-17a in AD, APP-transgenic mice were mated to IL-17a knockout mice. Brains of 6-month-old APP-transgenic mice with (ko) and without (wt) knockout of IL-17a were stained with antibodies against human Aβ (e). The volume of immunoreactive Aβ-positive staining was estimated with stereological Cavalieri method and adjusted by the volume of analyzed brain tissue. T test, n ≥ 9 per group. (h-p) In following experiments, the morphology of microglia in contact with Aβ deposits was analyzed in 4 and 9-month-old p38α-MAPK-ko and -wt APP-transgenic mice after immunofluorescent staining of Iba-1 and Aβ (h, i, images from 9-month-old APP-transgenic mice). The number of branches (j, n), endpoints of branches (k, o) and length of branches (l, p) of microglia were calculated and adjusted by the number of microglia. Deficiency of p38α-MAPK in myeloid cells significantly decreases the number of branches, but not the length of branches of microglia in 9-month-old APP-transgenic mice (j, k). T test, n ≥ 3 per group. (m) The Sholl analysis further shows that p38α-MAPK deficiency decreases the number of microglial processes in 9-month-old APP-transgenic mice. Total 14 microglia from five p38α-MAPK-ko mice and 12 microglia from three p38α-MAPKwt mice were analyzed. Two-way ANOVA testing the difference between p38α-MAPK-ko and -wt mice. (n-p) Deficiency of p38α-MAPK in myeloid cells does not change the morphology of microglia in 4-month-old APP-transgenic mice. T test, p > 0.05, n ≥ 3 per group. (q, r) Similarly, the morphology of microglia in contact with Aβ deposits was analyzed in 6-month-old IL-17a-ko and -wt APP-transgenic mice after immunofluorescent staining of Iba-1 and Aβ. (s-u) IL-17a deficiency significantly reduces the number and total length of branches of microglia in APP-transgenic mice compared with IL-17a-wt controls. T test, n ≥ 6 per group. (v) for the Sholl analysis, 10 microglia from 2 IL-17a ko mice and 14 microglia from 2 IL-17a wt mice were analyzed. Two-way ANOVA was performed to test the difference between IL-17a ko and wt mice. Data was represented as mean ± SEM  (Clausen et al., 1999) were bought from The Jackson Laboratory, Bar Harbor, ME (stock number 004781) and were back-crossed to C57BL/6J mice for >6 To delete IL-17a in AD mice, APP-transgenic mice were crossbred with Il-17a knockout mice (Nakae et al., 2002), which were kindly provided by Y. Iwakura, Tokyo University of Science, Japan.
Moreover, to investigate the location of LysM-Cre-expressing cells in the brain, APP-transgenic mice were cross-bred with ROSA mT/mG Cre reporter mice (Muzumdar et al., 2007) and LysM-Cre +/− mice to obtain APP tg ROSA mT/mG LysM-Cre +/− of genotype, which express enhanced green fluorescence protein (eGFP) in LysM-Cre-expressing cells. To examine whether peripheral myeloid cells migrate into the brain parenchyma, APP tg p38 fl/fl LysM-Cre +/− were mated to CCR2-RFP reporter mice (The Jackson Laboratory; stock number 017586), in which the chemokine (C-C motif) receptor 2 (CCR2) -coding sequence has been replaced with monomeric RFP-encoding sequence (Saederup et al., 2010). To track IL-17a-expressing cells in APP-transgenic mice, APP tg mice were cross-bred with IL-17a-eGFP reporter mice (kindly provided by R. Flavell, Yale University, USA), which express eGFP under the control of mouse Il-17a gene promoter (Esplugues et al., 2011).
Animal breeding, experimental procedure and methods of killing were conducted in accordance with national rules and ARRIVE guidelines, and were authorized by Landesamt für Verbraucherschutz, Saarland, Germany (registration numbers: 40/2014, 12/2018 and 34/2019).

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflicts of interest with the contents of this article.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analyzed during this study are included in this published article. Raw data are available upon reasonable request.