Introduction

Pax5 (Paired box5) is expressed in brain, B-lymphocytes, lymph nodes and spleen (Adams et al. 1992; Nutt et al. 2001; Souabni et al. 2002; Torlakovic et al. 2002; Delogu et al. 2006; Bousquet et al. 2007; Bharti and Mishra 2015). It has been implicated in lymphoma, neuroendocrine tumours, neuroblastoma and astrocytomas (Johri et al. 2016; Song et al. 2010; Baumann Kubetzko et al. 2004; Stuart et al. 1995). In B-cells, four isoforms (Pax5a or BSAP, Pax5b, Pax5d and Pax5e) of Pax5 in mouse (Zwollo et al. 1997; Lowen and Zwollo 2001) and 11 isoforms of PAX5 in human (Robichaud et al. 2004; Arseneau et al. 2009) have been observed. The Pax5 has been shown associated with phenotypic traits of ascitic cells causing Dalton’s lymphoma (Bharti and Mishra 2011) and genes of immune functions (CD19, CD21, CD40, Bach2, Aiolos (Ikzf3), IRF-4, IRF-8, H2-Ob, Ifi30 and C2ta), receptor-mediated signalling, adhesion (Bst1, Cd44, Sdc4, Tnfrsf19, Cd97 and Cd55) and migration (Capn2, Eps8, Gsn) of cells (Delogu et al. 2006; Schebesta et al. 2007; Pridans et al. 2008). It affects regulators of the NF-κB, genes of actin-remodelling (Schebesta et al. 2007), and immune pathway in the progression of Glioblastoma multiforme (Li et al. 2016). It also functions as metabolic gatekeeper for malignant transformation (Chan et al. 2017).

The microglia has been considered immune cells of brain. They regulate development of the nervous system, functions of various brain structures and age-related mental pathology (Korzhevskii and Kirik 2016). Data-mining indicates multiple binding sites for Pax5 in promoter of Ionized calcium-binding adapter molecule 1 (Iba1) found in microglia. The Iba1 (Utans et al. 1995) regulates actin-bundling, membrane ruffling, cell migration and phagocytosis in activated microglia (Ohsawa et al. 2004; Ito et al. 1998, 2001). It has also been predicted to interact with proteins essential in activation of leukocytes, inflammatory response, neuronal and glial differentiation. However, the role of Pax5 in brain immunity is not yet elucidated. It is presumed that neuronal Pax5, microglia and Iba1 may have different microenvironments in immune-privileged brain. Therefore, co-localization and interactions of brain-specific Pax5 with Iba1 in brain of mice were analysed through Immuno-fluorescence assay, Chromatin Immuno-precipitation (ChIP) and Co-Immunoprecipitation (Co-IP). The Pax5- and Iba1-positive cells were observed co-localized in almost all regions in microglia and distinctly compartmentalized in some regions of brain. The Pax5a/b and Iba1 interact specifically and indicate association of Pax5 and Iba1 in brain of mice.

Materials and Methods

Animal Model

The adult male albino mice of AKR strain were used for the experiments. Mice were maintained at 25 ± 2 °C as per the guideline of Institutional Animal Ethical Committee in the animal house of the department. They were fed upon standard mice feed in pellet form and tap water. The life span of both male and female mice is 75 ± 5 weeks. The experiments were carried out three times using male mice (n = 5) per experiment.

In Silico Analysis of Pax5 Interacting Proteins and Promoter Sequence Elements of Iba1

The Pax5 interacting proteins in mice were analysed using http://string-db.org/ and evaluated by interaction scores shown in the string database. Annotation of Pax5 biological functions and signalling pathway were based on the interacting proteins. The promoter sequences of Mus musculus Iba1 (Gene ID: 43916) were retrieved from Transcriptional Regulatory Elements database (https://cb.utdallas.edu/cgi-bin/TRED/tred.cgi?process=home). The promoter sequences for Iba1 with promoter ID 61002, 61001, 134800 were annotated from the Genbank Nucleotide database. The promoter sequences were used as an input sequence in TFBIND software for searching transcription factor binding sites on DNA (Tsunoda and Takagi 1999) using http://tfbind.hgc.jp.

Chromatin-Immunoprecipitation (ChIP) with Anti-Pax5 in Brain of Mice

The Chromatin Immunoprecipitation was performed as described earlier (Maurya and Mishra 2017). Briefly, the lysates of the adult brain mice were prepared. Cross linking and chromatin preparation from lysate was done by 1% formaldehyde. 125 mM glycine was used to stop the cross linking reaction. Nuclear extract was collected by centrifugation at 10,000×g for 10 min. Nuclear lysis was performed in ChIP-lysis buffer followed by sonication. Typically four rounds, 30 s pulse with 1-min rest in between rounds at output 5.0 (LABSONIC L, B. Braun Biotech International GmbH, Germany). The desired DNA fragment was in between 0.5 kb and 1 kb in length. The supernatant after sonication containing chromatin was incubated for 4 h at 4 °C for immunoprecipitation with anti-Pax5 antibody (anti-mouse, sc-13146, Santa-Cruz Biotech, USA). In control, anti-human IgG (HPO-1, Merk, India) was used. After centrifugation at 10,000xg, reverse linking was performed by adding 120 mM NaCl and incubation at 65 °C for 1 h. DNA obtained through immunoprecipitation was purified through phenol: chloroform purification method. The pulled DNA was checked on 1% agarose gel. Input DNA was obtained by reverse linking the chromatin prepared without antibody. The fold enrichment of Pax5 and Iba1 in Pax5 pulled DNA as compared to control were analysed by qPCR using gene-specific primers for Pax5 and Iba1. The primers sequence used were Pax5F, 5′ AATGACACCGTGCCTAGCGT 3′, Pax5R, 5′ TCAGCGGGGGTGGG 3′; Iba1F, 5′ ACTGCCAGCCTAAGACAACC 3′, Iba1R, 5′ GACCAGTTGGCCTGTTGTGT 3′.

Analysis of Expression and Co-Localization of Pax5 with Iba1

The antigen retrieval of cryo-sections of adult brain mice was done in 0.1% trypsin + 0.1% CaCl2 for 10 min. Sections after antigen retrieval were blocked with 1% BSA for 1 h. For double labelling, anti-Pax5 (anti-mouse, sc-13146, Santa-Cruz Biotech, USA, 1:500 dilution) and anti-Iba1 (anti-goat, sc-28528, Santa-Cruz Biotech, USA, 1:200 dilution) antibodies were used at 4 °C for overnight. The sections were washed with PBS and probed with TRITC (red)-conjugated goat anti-mouse IgG secondary antibody, FITC (green)-conjugated rabbit anti-goat IgG secondary antibody (1:2000 dilution) (Merk, India) for 2 h for detecting Pax5 and Iba1 immunoreactivity, respectively. In negative control, slides were stained with TRITC (red)-conjugated goat anti-mouse IgG secondary antibody, FITC (green)-conjugated rabbit anti-goat IgG secondary antibody (1:2000 dilution) without incubating with primary antibody. The slides were washed with PBS with Tween 20 (0.02%) and counter stained with DAPI (Molecular Probe) for nuclear staining as previously described (Tripathi and Mishra 2010). Images were scanned with a fluorescence microscope (Evos FLc) and confocal microscope (Zeiss LSM 780). Image analysis was performed by Zen software.

Analysis of Interaction of Pax5 and Iba1 by Co-Immunoprecipitation (Co-IP) in Brain of Mice

For Co-Immunoprecipitation, 50 µl of Protein-A bead was taken into spin column and washed twice with 1 ml 1X IP buffer by centrifugation at 3000 rpm for 20 s. 2 µg of anti-Pax5 (anti-mouse, sc-13146, Santa-Cruz Biotech, USA) and anti-Iba1 (anti-goat, sc-28528, Santa-Cruz Biotech, USA) diluted to 200 µl in buffer were added to each column containing resin, respectively. The columns were incubated at 4 °C for 30 min. After 30 min, the resin was washed with 1 ml cold 1X IP buffer thrice and 150 µl of adult mice brain tissue lysate was added to each column, respectively. Then columns were incubated for 1 h at 4 °C and washed with cold 1X IP buffer thrice. In negative control, the beads were incubated with antibody and lack brain tissue lysate. After washing, 100 µl of sample loading buffer was added on the resin, mixed well and the suspension was transferred into the 1.5-ml microfuge tube. Samples were heat denatured for 5 min, centrifuged for 1 min at 3000 rpm, and 50 µl of the samples was resolved through 12% SDS-PAGE and analysed by Western blotting for Pax5 and Iba1-reactive peptide band, respectively.

Statistical Analysis

Each experiment was repeated three times (n = 5 mice per experiment). For Chromatin Immunoprecipitation (ChIP), fold enrichment was calculated using ΔΔCt value as compared to IgG control and plotted as histogram with the mean ± SEM of three values calculated from three independent experiments. ‘*’ denotes significant differences (p ≤ 0.05) as compared to control (independent-samples t-test).

Results

Pax5 Binds to DNA-Sequence Elements of Iba1, Shows Co-Localization and Physical Interaction with Iba1 in Brain of Mice

In silico analysis predicts association of Pax5 with genes and proteins of neurons and glia, T-cell, leukocyte, inflammatory responses and cytokines. The Pax5 shows highest scores with Myc, Lef1, Rag2, Ebf1, Tnfsf11, Gdnf, Csfr1, Tcf12, Sox11, Foxo1, Mapk9, Mapk10, Ikzf1, Irf4, Ptprc, Gata3, Fgf8 and Crebbp (Fig. 1a). The Pax5 binds to the promoter sequence elements of Iba1 was also predicted. Out of the three promoter sequences of Iba1 retrieved from a transcriptional regulatory element database, the Pax5 binding sites have been observed on 8 sequences at (+) strand and 7 sequences at (−) strand of Iba1 Promoter ID (61002), 6 sequences at (+) strand and 6 sequences at strand of Iba1 Promoter ID (61001) and 7 sequences at (+) strand and 11 sequences at (−) strand of Iba1 Promoter ID (134800) summarized in Table 1. Apart from Pax5, binding sites for major transcription factors were also observed on promoter sequences of Iba1 (Fig. 1b).

Fig. 1
figure 1

a In silico analysis of Pax5 interacting proteins through neural network indicates that Pax5 could be a central molecule in the regulation of brain homeostasis by interacting with proteins responsible from cell cycle to cell death. b Schematic diagram representing the binding of major transcription factors on the Iba1 promoter region analysed by TFBIND software

Full size image
Table 1 In silico analysis for Pax5 binding to the promoter sequence elements of Iba1 using TFBIND
Full size table

The qPCR with gene-specific primers for Pax5 and Iba1 from DNA obtained by Chromatin-Immunoprecipitation (ChIP) using anti-Pax5 revealed Pax5-binding to the sequence elements of the its own Pax5 gene and microglia-specific gene (Iba1) in brain of mice (Fig. 2). The enrichment of Pax5 and Iba1 were observed 200- and 300-fold, respectively, in anti-Pax5 pulled down DNA as compared to anti-human IgG pulled down DNA.

Fig. 2
figure 2

Analysis of genetic sequence elements of the brain obtained by Chromatin Immunoprecipitation (ChIP) with anti-Pax5 in brain of mice. The qPCR shows the fold enrichment of Pax5 (200 fold) and Iba1 (300 fold) genes in Pax5 pulled DNA as compared to IgG pulled DNA (control). Data are represented mean ± SEM and ‘*’ denotes significant differences (p ≤ 0.05) as compared to control (independent-samplest-test)

Full size image

The Pax5 and Iba1 show co-localization in different forms of microglia like ramified (Rm), amoeboid (Am) and round (rm) in cerebral cortex (Cc) (Fig. 3a; A–E), in amoeboid form of microglia (Am) of white matter (Whm), cell body (Cb) of ramified microglia (Rm) in grey matter (Gm) and molecular layer of cerebellum (C) (Fig. 3a; F–J), in amoeboid and round phenotype of olfactory glomerular layer (OlGl) of olfactory bulb (OB) (Fig. 3b; A–E) and in microglia of inner molecular layer (IML), Granular cell layer (GCL) and sub-granular zone (SGZ) of hippocampus (H) (Fig. 3b; F–J). The co-localization was detected in choroid plexus (ChP) and cerebrospinal fluid (CSF) of third ventricle (TV) (Fig. 4a; A–E), in clusters of blood vessels (BV) of lateral ventricles (LV) (Fig. 4a; F–J) and choroid plexus (ChP) (Fig. 4b; A–E). In glomerular layers (OlGl) of olfactory bulb, peri-glomerular cells (PGC) were positive for Pax5 only, whereas round- to amoeboid-shaped microglia phenotypes were positive for both Pax5 and Iba1 (Fig. 3b; A–E). In hippocampus, elongated processes of ramified microglia and round-shaped activated microglia showed co-localization of Pax5 and Iba1 and in CA1 region of dentate gyrus of adult mice brain (Fig. 3b; F–J). Apart from co-localization of Pax5 and Iba1 in microglia cells in inner molecular layer (IML) and granular cell layer (GCL) of CA1 region of dentate gyrus, Pax5 positive cells were observed in adult born granule cells (AGC) in sub-granular zone (SGZ) of hippocampus (Fig. 3b; F–J). Western blot analysis with anti-Pax5 and anti-Iba1 from Immunoprecipitated samples with anti-Pax5 and anti-Iba1 shows physical interaction of Pax5a/b and Iba1 in the brain of mice where tissue lysate was taken as positive control. In tissue lysate, two isoforms of Pax5 i.e. Pax5a/b and Pax5d/e were detected whereas in anti-Pax5 and anti-Iba1 pulled down proteins, only Pax5a/b was detected (Fig. 5).

Fig. 3
figure 3

a Photomicrographs of Pax5- and Iba1-positive cells in Cc, cerebral cortex (A–E); C, cerebellum (F–J) in brain of mice. In cerebral cortex and cerebellum, Iba1-positive cells (green) are microglia which appear ramified (Rm), amoeboid (Am) and round (rm). They also express Pax5 and show co-localization with Pax5-positive cells (red) in cerebral cortex. In cerebellum, Iba1-positive cells (green) show ramified (Rm) and amoeboid (Am) microglia phenotype where Pax5-positive cells (red) were observed localized only in the cell body (Cb) of ramified microglia in grey matter (Gm) and molecular layer (ML) and in amoeboid microglia phenotype of white matter (Whm), complete co-localization of Pax5 and Iba1 were observed. b Photomicrographs of Pax5- and Iba1-positive cells in OB, olfactory bulb (A–E); H, Hippocampus (F–J) in brain of mice. In olfactory bulb, Iba1-positive cells (green) indicate amoeboid (Am) to round (rm) phenotype of microglia in olfactory glomerular layer (OlGl). They were positive for Pax5 (red) whereas peri-glomerular cells (PGC) of olfactory glomerular layer were observed only positive for Pax5 (red). The merged image showed co-localization of Pax5- and Iba1-positive cells in the round and amoeboid phenotype of microglia but not in peri-glomerular cells of glomerular layer. In hippocampus, Iba1-positive cells represent ramified (Rm) elongated processes of microglia and amoeboid to round phenotypes (Am and rm) in inner molecular layer (IML), granular cell layer (GCL) and round phenotype in sub-granular zone (SGZ) where Pax5-positive cells were also observed in the adult granule cells (AGC) of sub-granular zone (SGZ). The merged image of Pax5- and Iba1-positive cells showed co-localization of Pax5 and Iba1 in ramified elongated processes of microglia in hippocampus of brain in mice

Full size image
Fig. 4
figure 4

a Photomicrographs of Pax5- and Iba1-positive cells in TV, Third Ventricles (A–E), LV, Lateral Ventricles (F–J) in brain of mice. The co-localization of Pax5 and Iba1 were observed in the cerebrospinal fluid (CSF), choroid plexus (ChP) of third ventricle (A–E) and in clusters of blood vessels chain (BV) in lateral ventricles (LV). b Photomicrographs of Pax5- and Iba1-positive cells in ChP, Choroid plexus (A–E) in brain of mice. The merged image of Pax5-positive cells and Iba1-positive cells indicates co-localization of Pax5 and Iba1 in ChP

Full size image
Fig. 5
figure 5

Analysis of interaction of Pax5 with Iba1 in brain of mice by Co-Immunoprecipitation. It shows Western blot analysis with anti-Pax5 and anti-Iba1 from samples Co-Immunoprecipitated with anti-Pax5 and anti-Iba1 from brain of mice. Crude protein lysate was taken as input

Full size image

Discussion

The co-localization of Pax5 and Iba1 in microglia and interaction of Pax5 with Iba1 both at the DNA- and protein levels indicates brain-specific association of Pax5 and Iba1. It was also evident that only full length Pax5 isoform (Pax5a/b) interacts with Iba1 in brain of mice. The analysis of Pax5-interacting proteins from databases indicate its association with regulation of differentiation and activation of neurons and glia, inflammatory responses and production of cytokines, signalling pathways of Wnt, MAPK, Foxo, Hedgehog, NF-κB, PI3KT-AKT, Ras, TGF-beta, JAT-STAT, Neurotrophin, Toll-like receptor, TNF, Insulin and Notch. The similarity index for Pax5-binding sites and analysis of promoter sequences suggested binding of Pax5 to Iba1 to more than one site on (+) strand and (−) strand of a particular promoter sequence. The binding of Pax5 to sequence element of Iba1 indicates Pax5-mediated transcriptional regulation of Iba1 in brain of mice. Chromatin Immunoprecipitation also revealed that Pax5 binds to its own regulatory sequences and may directly regulate its own expression.

The co-localization of Pax5 and Iba1 in microglia of cerebral cortex, cerebellum, olfactory lobe, hippocampus and ventricles of brain shows association of Pax5 with Iba1 in almost all regions of brain of adult mice. The region-specific ramified and radiated, resting or activated round Iba1-positive microglia cells support their impact on status of activation of microglia (Torres-Platas et al. 2014). Their co-localization in cerebellum accords (Perez-Pouchoulen et al. 2015) their affects on regulation of population of microglia. Their co-localization in hippocampus may be related to pro-inflamed brain or neuronal plasticity as suggested earlier (Sandhir et al. 2008). Since adult granule cells (AGC) in sub-granular zone (SGZ) region of hippocampus were positive for Pax5 but not for Iba1, Pax5 appears to be involved in adult neurogenesis in hippocampus. It is likely to be involved in memory and learning, cognition and encoding of novel information (Christian et al. 2014; Danielson et al. 2016) as suggested. The presence of Pax5 and Iba1 in microglia of olfactory bulb may be associated with immunological surveillance (Herbert et al. 2012) because this region is considered potential route for the entry of microorganisms and pathogens into the central nervous system. The Pax5 were also observed localized in the peri-glomerular cells (PGC), gatekeeper of olfactory system which not only receive information from olfactory sensory neurons, but also control mitral cells spiking (Ye et al. 2010; Arruda et al. 2013). Choroid plexus serves as a port for microglia to enter first the CSF and then brain at the ventricular surface (Lun et al. 2015), co-localization in blood vessels of sub-ventricles, lateral ventricles and in CSF of third ventricles signifies their impact in transport of immune cells and microglia within the brain. It is in agreement with reports (Lun et al. 2015; Devorak et al. 2015) that choroid plexus of the brain serves as a communicating network. The microglial cells have been considered long lived and not normally replaced by bone marrow-derived cells. They may have distinctive genetic signature from neurons, astrocytes, oligodendrocytes and peripheral immune cells including other tissue macrophages (Prinz et al. 2017). The Iba1 has also been described more suitable markers than trans-membrane- and plasma-membrane-specific markers for studies on the complex organization and structural analysis of microglia (Korzhevskii and Kirik 2016). The co-localization of Pax5 and Iba1 appears compartmentalized because all Pax5-positive or Iba1-positive cells do not show co-localization as peri-glomerular cells (PGC) of olfactory bulb (OB) and adult granule cells (AGC) in sub-granular zone (SGZ) region of hippocampus in brain were only observed positive for Pax5. The observation also supports views of novel source of microglia (Sakuma et al. 2016) and microenvironment of Pax5- and Iba1-positive cells in immune-privileged brain. Immunoprecipitation revealed Pax5a/b but not Pax5d/e physical interact with Iba1 in brain which indicates brain-specific activities of Pax5a/b as reported for B-cells (Lowen and Zwollo, 1991; Zwollo et al. 1997).

In summary, brain-specific co-localization and interaction of Pax5a/b with Iba1 indicate impact of Pax5 on microglia-mediated immunity in brain.