Forkhead box protein p1 is a transcriptional ... - Oxford Academic

6 downloads 2997 Views 548KB Size Report
Apr 6, 2012 - and a manually curated Huntington's disease (HD)-signaling pathway. Similar results were found when the gene expression data set was ...
Human Molecular Genetics, 2012, Vol. 21, No. 14 doi:10.1093/hmg/dds132 Advance Access published on April 6, 2012

3097–3111

Forkhead box protein p1 is a transcriptional repressor of immune signaling in the CNS: implications for transcriptional dysregulation in Huntington disease Bin Tang1, Kristina Becanovic2, Paula A. Desplats3, Brian Spencer3, Austin M. Hill2, Colum Connolly2, Eliezer Masliah3,4, Blair R. Leavitt2 and Elizabeth A. Thomas1,∗ 1

Department of Molecular Biology, The Scripps Research Institute, 10550 North Torrey Pines Rd., La Jolla, CA, USA, Centre for Molecular Medicine and Therapeutics, Department of Medical Genetics, University of British Columbia, Vancouver, Canada V5Z 4H4, 3Department of Neuroscience and 4Department of Pathology, University of California, San Diego, La Jolla, CA, USA 2

Received December 16, 2011; Revised March 6, 2012; Accepted April 2, 2012

Forkhead box protein p1 (Foxp1), a transcription factor showing highly enriched expression in the striatum, has been implicated in central nervous system (CNS) development, but its role in the mature brain is unknown. In order to ascertain functional roles for Foxp1 in the CNS, we have identified gene targets for Foxp1 both in vitro and in vivo using genome-wide expression microarrays and chromatin-immunoprecipitation followed by high-throughput sequencing (ChIP-seq) assays. We found that mouse Foxp1 overexpression in striatal cells elicited expression changes of genes related to immune signaling, transcriptional regulation and a manually curated Huntington’s disease (HD)-signaling pathway. Similar results were found when the gene expression data set was integrated with Foxp1-binding data determined from ChIP-seq analysis. In vivo lentiviral-mediated overexpression of human FOXP1 in the context of mutant huntingtin (Htt) protein resulted in a robust downregulation of glial cell-associated, immune genes, including those encoding a variety of cytokines and chemokines. Furthermore, Foxp1-induced expression changes were significantly negatively correlated with those changes elicited by mutant Htt protein in several different HD mouse models, and most significantly in post-mortem caudate from human HD subjects. We finally show that Foxp1 interacts with mutant Htt protein in mouse brain and is present in nuclear Htt aggregates in the striatum of R6/1 transgenic mice. These findings implicate Foxp1 as a key repressor of immune signaling in the CNS and suggest that the loss of Foxp1-mediated gene regulation in HD contributes to the immune dysfunction in this disease. We further suggest that Foxp1-regulated pathways might be important mediators of neuronal-glial cell communication.

INTRODUCTION The forkhead box (Fox) proteins constitute a large family of transcription factors with diverse functions, from development and organogenesis to regulation of metabolism and immune function (1,2). Fox transcription factors are characterized by a 100-amino-acid winged helix/forkhead DNA-binding domain. Additionally, the subfamily of Foxp proteins,

including Foxp1-p4, contain a zinc finger domain and a leucine zipper motif and can act as transcriptional repressors by forming homo- or hetero-dimers with other family members. While the function of Foxp1 has been widely studied in blood, lung, heart and immune cells (3 –6), the function of Foxp1 in neuronal processes remains unclear. Previous studies have demonstrated a role for this transcription factor in central nervous system (CNS) development, whereby it has



To whom correspondence should be addressed at: Department of Molecular Biology, The Scripps Research Institute, 10550 North Torrey Pines Rd., La Jolla, CA 92037, USA. Fax: +1 8587842212; Email: [email protected]

# The Author 2012. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]

3098

Human Molecular Genetics, 2012, Vol. 21, No. 14

been shown to be an important accessory factor in Hox transcriptional output, thus regulating motor neuron diversification and connectivity to target muscles (7,8). A link between Foxp1 and development has also been suggested from genetic studies on humans, which have shown that the FOXP1 gene, like its related family member FOXP2 (9 – 11), may be involved in developmental conditions that are associated with language and speech deficits (12,13). In addition to showing abundant expression during developmental stages, Foxp1 also exhibits high levels of expression in the adult striatum (14– 17), suggesting that this transcription factor plays an important role in gene expression regulation of mature medium spiny neurons, although Foxp1 target genes have not yet been identified. The goal of this study was to identify genome-wide gene targets for Foxp1 in the CNS, in order to surmise the potential functional roles of Foxp1 in the mature striatum under normal and diseased states. Huntington’s disease (HD) is one of the most notable of striatal disorders, whereby expression of a polyglutamineexpanded mutant huntingtin (Htt) protein results in predominant loss of medium spiny striatal neurons in the brain (18). HD is associated with a range of transcriptional abnormalities, and several specific transcription factors and co-factors have been proposed as mediators of mutant Htt toxicity; however, which transcription factors are most important to pathogenesis/pathophysiology is not known (19,20). In this study, we used genome-wide transcriptome and chromatin-immunoprecipitation (ChIP) binding assays to identify Foxp1 target genes. Our results indicate that Foxp1 is strongly associated with repression of immune-related genes in striatal cells, both in vitro and in vivo, and that a majority of genes downregulated by Foxp1 are glia derived. We further found a strong link between Foxp1 and pathways related to HD, and that Foxp1 can interact with mutant Htt protein, suggesting that Foxp1 is an essential player in the transcriptional dysregulation observed in HD, especially that involving immune signaling.

RESULTS Foxp1 target genes are related to HD signaling pathways in vitro In order to identify candidate target genes for Foxp1 expression regulation in vitro, we performed whole transcriptome analysis of wild-type (WT) immortalized striatal cells [STHdhQ7 striatal cells (21)] overexpressing Foxp1. In total, 1473 genes (740 downregulated and 733 upregulated) were significantly altered by Foxp1 (Supplementary Material, Table S1). We picked eight genes showing the most statistically significant expression differences, including both up- and down-regulated genes, to validate by real-time quantitative polymerase chain reaction (RT-qPCR) analysis. Six of these showed validated expression differences by this method: RAB18, member RAS oncogene family (Rab18), 3hydroxymethyl-3-methylglutaryl-Coenzyme A lyase-like 1 (Hmgcll1), Sulfotransferase family 1E, member 1 (Sult1e1), Activating transcription factor 3 (Atf3), Follistatin (Fst) and FBJ osteosarcoma oncogene (Fos), while the changes in two

others (Mdh2 and Etfdh) did not reach statistical significance (Fig. 1A). As a positive control, we also detected a robust increase in Foxp1 mRNA expression in these cells (Supplementary Material, Fig. S1). We further tested the effects of RNA knock-down of Foxp1 in these cells using small-interfering RNA methodology. We found that several of the genes upregulated by Foxp1 overexpression, including Atf3, Fst and Fos, were decreased in expression when Foxp1 mRNA levels were repressed (Supplementary Material, Fig. S1). To assess potential functional consequences of Foxp1 gene regulation, pathway analysis using the Ingenuity Pathways Analysis software was implemented. Foxp1 target genes were associated with known canonical pathways related to immune signaling, cancer and neuronal development (Supplementary Material, Table S2). Notably, 40 Foxp1 target genes were associated with a manually curated ‘Huntington’s disease pathway’, which includes brain-derived neurotrophic factor signaling, caspase activity, calcium and glutamate signaling and several transcription factors (Table 1). Integrated expression microarray and ChIP-seq analysis of Foxp1 target genes In order to identify binding patterns of Foxp1, we performed ChIP with an anti-Foxp1 antibody in Foxp1-transfected striatal cells followed by massively parallel sequencing of the co-immunoprecipitated genomic DNA fragments (ChIP-Seq). We applied MACS software to identify Foxp1-enrichedbinding regions and mapped these regions to genes using the UCSC Genome Browser. We identified 1000 Foxp1-binding regions showing fold-enrichments between 3 and 95, ranging in size from 44 to 633 bp, and located throughout all chromosomal regions (Fig. 1B and C). Of the top ChIP-seq hits, only 20% was within the 5′ proximal promoter region, with a majority being within intronic regions (Supplementary Material, Table S3). Interestingly, a 242 bp intronic region of Foxp1 was among the top binding regions for Foxp1. Given that many regulatory elements reside in introns, or up- and downstream of the transcription unit (22), this suggests that Foxp1 can regulate its own expression. Foxp1 was also among the expression targets for itself; however, we could not distinguish endogenous mRNA levels from transgene sources in mouse cells. We integrated all gene hits showing Foxp1 binding with the list of genes whose expression levels were altered by Foxp1, in order to identify a stringent group of direct Foxp1 targets in vitro. We identified 75 top target genes for Foxp1 whose functional annotations revealed associations to neuronal development, genetic disorder, inflammatory disease and transcriptional regulatory pathways, similar to what was detected from the gene expression data (Fig. 1D). Strikingly, 33 of the 75 direct Foxp1 targets were significantly associated with inflammatory disease, highlighting a role of Foxp1 in this process (Fig. 1D). Foxp1 target genes relate to immune function in vivo We have previously demonstrated that Foxp1 expression was decreased in R6/1 transgenic mice and in human post-mortem caudate from HD patients (16). In addition, microarray data

Human Molecular Genetics, 2012, Vol. 21, No. 14

3099

Figure 1. Real-time qPCR validation of microarray data and ChIP-seq data annotation. (A) Real-time qPCR validation of genes regulated by Foxp1 in vitro. Data are shown as mean expression levels + SEM of the indicated genes in Foxp1-transfected striatal cells versus mock-transfected cells from three separate experiments. Expression levels were normalized against Hprt. Genes are indicated by their official UniGene ID. Open bars represent control-transfected striatal cells, and closed bars represent Foxp1-transfected striatal cells. Asterisks denote significant differences in gene expression as determined by Student’s t-test; ∗ P , 0.05; ∗∗ P , 0.01; ∗∗∗ P , 0.001. (B) Fold-enrichment of ChIP-seq Foxp1-binding tags. The top genes showing Foxp1-binding sites are shown according to P-value for enrichment. Genes are indicated by their official UniGene ID. (C) Chromosomal distribution of ChIP-seq hits. (D) Functional analysis of the integrated gene target list from the ChIP-seq and expression microarray analyses. Data reflect Gene Ontology categories identified by DAVID. Left y-axis shows the number of gene in each category (bar graphs), and right y-axis show – log(P-value) for the representation of the given category (data points).

from several HD mouse models, as well as from human caudate, also show significantly reduced expression levels of Foxp1 (23– 25). These studies are in line with the strong link of Foxp1 to HD signaling pathways that we found in vitro striatal cells, suggesting that Foxp1 might be an important mediator of the transcriptional dysregulation observed in HD. Therefore, we analyzed gene expression profiles resulting from Foxp1 overexpression in vivo in the presence of the fulllength mutant Htt protein. We performed bilateral overexpression of human Foxp1, which shows 98.6% identity to mouse Foxp1 at the protein level, or a lentiviral construct containing green fluorescent protein (GFP), in the striatum of 3-monthold YAC128 HD transgenic mice, which express a full-length human mutant Htt protein driven by the endogenous human HTT promoter (26). Importantly, at this age, YAC128 mice do not show a significant difference in Foxp1 expression when compared with age-matched WT control mice (Supplementary Material, Fig. S2). Microarray analysis revealed 2486 genes showing altered expression in response to

Foxp1 overexpression, with 73.6% of these genes being downregulated in expression (Supplementary Material, Table S4). Similar to what was observed in striatal cells in vitro, pathway analysis of the differentially expressed genes revealed overrepresentation of genes associated with known canonical pathways related to immune signaling and immune function (Table 2). In this experimental setting, however, the magnitude and significance of the immune-related expression changes detected in vivo were striking, with 100 immune-related genes showing .5-fold reductions in expression (Supplementary Material, Table S4), suggesting that Foxp1 is an immune repressor in the CNS. Several immune-related pathways, including nuclear factor-kappaB (NF-kB), T-cell receptor– B-cell receptor-, triggering receptor expressed on myeloid cells 1 (TREM1)-, toll-like receptor- and IL-10-signaling, were significantly associated with Foxp1 overexpression at P-values , 2.5E210 (Table 2). The NF-kB signaling pathway was the most highly significant CNS pathway encompassing

3100

Human Molecular Genetics, 2012, Vol. 21, No. 14

Table 1. Genes comprising the manually curated HD signalling pathway (from Ingenuity Systems analysis) Gene ID

Gene description

Location

Protein type

APAF1 BDNF CAPN1 CAPN5 CASP6 CDK5 CREB5 EGFR EP300 GNAQ GNB1 GNG7 GOSR1 HAP1 HDAC4 HDAC5 HDAC10 HSPA5 IGF1 JUN MAPK1 MAPK3 MAPK9 MTOR NAPB NEUROD1 NGF PACSIN1 PIK3C2A PIK3R3 PLCB3 PRKCD PRKCZ SHC1 SNCA STX1A TCERG1 UBB UBC UBD

Apoptotic peptidase activating factor 1 Brain-derived neurotrophic factor Calpain 1, (mu/I) large subunit Calpain 5 Caspase 6, apoptosis-related cysteine peptidase Cyclin-dependent kinase 5 cAMP responsive element binding protein 5 Epidermal growth factor receptor E1A-binding protein p300 Guanine nucleotide binding protein, q polypeptide Guanine nucleotide binding protein, beta polypeptide 1 Guanine nucleotide binding protein, gamma 7 golgi SNAP receptor complex member 1 Huntingtin-associated protein 1 Histone deacetylase 4 Histone deacetylase 5 Histone deacetylase 10 Heat shock 70 kDa protein 5 Insulin-like growth factor 1 jun proto-oncogene Mitogen-activated protein kinase 1 Mitogen-activated protein kinase 3 Mitogen-activated protein kinase 9 Mechanistic target of rapamycin N-ethylmaleimide-sensitive factor attachment protein, beta Neurogenic differentiation 1 Nerve growth factor (beta polypeptide) Protein kinase C and casein kinase substrate in neurons 1 Phosphoinositide-3-kinase, class 2, alpha polypeptide Phosphoinositide-3-kinase, regulatory subunit 3 Phospholipase C, beta 3 Protein kinase C, delta Protein kinase C, zeta SHC (Src homology 2 domain containing) transforming protein 1 Synuclein, alpha Syntaxin 1A (brain) Transcription elongation regulator 1 Ubiquitin B Ubiquitin C Ubiquitin D

Cytoplasm Extracellular Space Cytoplasm Cytoplasm Cytoplasm Nucleus Nucleus Plasma Membrane Nucleus Plasma Membrane Plasma Membrane Plasma Membrane Cytoplasm Cytoplasm Nucleus Nucleus Nucleus Cytoplasm Extracellular Space Nucleus Cytoplasm Cytoplasm Cytoplasm Nucleus Cytoplasm Nucleus Extracellular Space Cytoplasm Cytoplasm Cytoplasm Cytoplasm Cytoplasm Cytoplasm Cytoplasm Cytoplasm Cytoplasm Nucleus Cytoplasm Cytoplasm Nucleus

Other Growth factor Peptidase Peptidase Peptidase Kinase Transcription regulator Kinase Transcription regulator Enzyme Enzyme Enzyme Transporter Other Transcription regulator Transcription regulator Transcription regulator Other Growth factor Transcription regulator Kinase Kinase Kinase Kinase Transporter Transcription regulator Growth factor Kinase Kinase Other Enzyme Kinase Kinase Other Other Transporter Transcription regulator Other Other Other

Foxp1 target genes (P ¼ 1.9E219), with 59 genes in this pathway being altered by Foxp1 and nearly all being downregulated in expression (Fig. 2; Table 2). Because a majority of the top Foxp1-elicited expression changes were decreased in expression, we ran a separate analysis using only those genes showing elevated expression. Genes upregulated in response to Foxp1 overexpression were significantly related to cAMP-mediated signal transduction, G protein-coupled receptor signaling, glutamate receptor function and gamma-aminobutyric acid receptor signaling, although far fewer members of these pathways were altered by Foxp1, compared with the immune-related pathways, as shown in Table 2 (Supplementary Material, Table S5). Lentiviral delivery of human FOXP1 into mouse brain allowed us to measure its effect on the expression of endogenous Foxp1. Our microarray data analysis showed that overexpression of human Foxp1 resulted in a 1.53-fold elevation of mouse Foxp1 mRNA (P ¼ 1.5E24). This finding was validated by RT-qPCR (Fig. 3), providing additional support that Foxp1 regulates its own expression, as observed in our in vitro ChIP-seq analysis. We also validated the expression elevation of Kv channel-interacting protein 2 (Kcnip2), a

known striatal-enriched gene (27), and the robust expression decrease in glial fibrillary acidic protein (Gfap), a classical marker of mature astrocytes in the CNS (Fig. 3). Because our expression data were generated from whole striatal extracts containing a heterogeneous population of cell types, we compared the list of Foxp1-induced expression changes to lists of CNS cell-type-specific molecular markers generated from microarray data performed on purified populations of astrocytes, neurons and oligodendrocytes from mouse forebrain (28) and from isolated neurons, astrocytes and microglial cells derived from primary human cortical cultures (29). Thirty-five percent of the mRNAs altered by Foxp1 were associated with a particular CNS cell type: n ¼ 382 were astrocyte-enriched, n ¼ 208 were oligodendrocyteenriched and n ¼ 290 were neuron-enriched. Moreover, considering the direction of the expression changes, we found that genes downregulated by Foxp1 are significantly overrepresented among glial expressed genes (Fisher’s exact test, two-tailed; P ¼ 3.64E25), while neuronally expressed genes are mostly upregulated by Foxp1 overexpression (Fisher’s exact test, two-tailed; P ¼ 3.05E29). Similar results were found when comparing to the data sets from primary human

Human Molecular Genetics, 2012, Vol. 21, No. 14

3101

Table 2. Pathways associated with Foxp1 striatal overexpression in vivo Ingenuity canonical pathway

2log(P-value)

# genes altered

% of genes in pathway

% down-regulated

Altered T/B cell signaling in rheumatoid arthritis NF-kB signaling PI3K signaling in B lymphocytes TREM1 signaling T helper cell differentiation Type I diabetes mellitus signaling Acute phase response signaling CD28 signaling in T helper cells T cell receptor signaling B cell receptor signaling Toll-like receptor signaling G-protein-coupled receptor signaling Regulation of IL-2 expression in T lymphocytes IL-10 signaling NFAT regulation of the immune response Dendritic cell maturation Molecular mechanisms of cancer TNFR1 signaling IL-15 signaling Apoptosis signaling IL-6 signaling Phospholipase C signaling Innate and adaptive immune cell communication TNFR2 signaling IL-12 signaling and production in macrophages B cell activating factor signaling Primary immunodeficiency signaling LXR/RXR activation Renin–angiotensin signaling Endothelin-1 signaling Colorectal cancer metastasis signaling Induction of apoptosis by HIV1 Prolactin signaling Interferon signaling Glucocorticoid receptor signaling Activation of IRF by pattern recognition receptors IL-17 signaling IL-8 signaling Lymphotoxin beta receptor signaling Thrombin signaling Role of PKR in interferon induction and antiviral response Death receptor signaling

16.60 15.70 15.30 12.80 12.70 12.20 11.60 11.50 11.20 10.60 10.30 9.96 9.80 9.65 9.55 9.25 8.89 8.16 8.12 7.86 7.82 7.82 7.58 7.32 7.27 7.14 7.05 7.03 6.92 6.84 6.81 6.80 6.79 6.75 6.64 6.59 6.53 6.48 6.41 6.38 6.34 6.27

41 59 51 29 32 41 53 42 37 46 24 107 31 28 50 47 76 21 25 29 31 56 30 15 34 18 21 26 32 44 55 22 25 16 57 23 25 43 21 46 20 22

45 33.5 34.9 42 44.4 33.9 29 31.1 33.6 29.3 43.6 20.2 34.4 35.9 24.6 24.6 20.3 40.4 35.7 32.2 31 21.5 27.5 45.5 24.8 40 33.3 28 25.8 22.9 21.2 33.8 31.2 44.4 20.1 31.1 32.5 22.3 32.8 22.1 37 32.8

100.0 93.2 90.0 100.0 100.0 92.6 92.5 96.2 97.2 91.3 100.0 70.0 93.0 700.0 88.0 95.0 78.0 95.2 96.0 89.0 90.0 85.0 100.0 100.0 91.0 94.0 100.0 92.0 75.0 82.0 80.0 90.9 84.0 100.0 80.0 95.6 88.0 88.0 95.2 76.0 100.0 100.0

cultures (29). Additionally, in this latter study by Loerch et al. (29), purified microglia cell populations were examined; of the 66 identified microglial-enriched genes, 23 were found to be downregulated by Foxp1, compared with only 2 genes whose expression levels were upregulated by Foxp1 (Fisher’s exact test, two-tailed; P ¼ 2.68E26). In light of these findings, we investigated which CNS cell types express Foxp1 using co-immunofluorescence. In WT mice, we found that the Foxp1 protein was co-localized with the neuronal marker NeuN, but not with GFAP, a marker for astrocytes, or CD11b, a marker for microglia (Fig. 4A). As a more sensitive means to detect Foxp1 expression, we measured Foxp1 mRNA using RT-qPCR analysis in cultured microglia cells generated from whole brain at post-natal day 1. RT-qPCR analysis of primary microglial cultures from WT, YAC18 and YAC128 mice revealed that Foxp1 mRNA is expressed in microglial cultures, although at much lower levels than that found in medium spiny neurons (.4-fold lower; Fig. 4B). We further found that Foxp1 mRNA levels were significantly up-regulated upon microglia activation by

pro-inflammatory stimulation with endotoxin and interferon gamma (Fig. 4B). Foxp1 target genes versus HD transcriptional signatures To further test our hypothesis that Foxp1 may be an important mediator of the transcriptional dysregulation observed in HD, we compared the Foxp1-induced transcriptome profiles in vivo to gene expression abnormalities detected in human HD caudate and several different HD mouse models. Even though the microarray data from these models were performed on different platforms and analyzed using different statistical methods, we primarily aimed for qualitative comparisons, noting the direction of the expression changes and significance (P , 0.05) among models. Human homologs of the top 500 Foxp1-regulated mouse genes in the striatum (the top 250 downregulated genes and the top 250 upregulated genes based on rank) were retrieved and screened for expression differences in the human HD caudate data set provided in Hodges et al. (23). We identified 348 unique Foxp1 target genes,

3102

Human Molecular Genetics, 2012, Vol. 21, No. 14

Figure 2. Schematic depiction of NF-kB signaling pathway highlighting FOXP1 target genes. Canonical pathways analysis was performed using Ingenuity Pathways Analysis. Foxp1 target genes were significantly associated with the NF-kB signaling pathway (Fisher’s exact test; P ¼ 1.9E216). Green indicates downregulated genes and red indicates upregulated genes in response to FOXP1 overexpression in the striatum. Dark green indicates a greater magnitude of expression change compared to light green. With permission from Ingenuity Systems Inc. # 2000– 2012. All rights reserved.

whose human homolog was identified on the human microarray chip. Strikingly, the expression of 78% of these genes was significantly altered in the human caudate data set, although largely in the opposite direction to that caused by Foxp1. Only nine genes were significantly concordantly

downregulated, and 11 concordantly upregulated, in both FOXP1-injected mouse striatum and human HD caudate (Fig. 5A), while 101 of the Foxp1-induced downregulated genes were significantly elevated in human HD caudate and 147 of the Foxp1-induced upregulated genes were

Human Molecular Genetics, 2012, Vol. 21, No. 14

3103

Figure 3. Real-time qPCR analysis of gene expression differences elicited by Foxp1 in vivo and of Foxp1 itself in cultured cells. Validation of microarray expression differences elicited by Foxp1 in vivo. Expression changes of Kcnip2, Gfap, mouse Foxp1 and human FOXP1 elicited by lentiviral-mediated human FOXP1 overexpression in the striatum of 3-month-old YAC128 mice. Tissue was harvested 1 month after lentiviral injections. Data represent the mean expression levels + SEM from eight GFP-injected control mice and nine FOXP1-injected mice normalized by a normalization factor (NF4) as described in Materials and Methods.

downregulated in human HD caudate (Fig. 5A). The log2 ratios of the gene expression changes from human caudate showed a significant negative correlation with those due to Foxp1 overexpression (Pearson correlation value ¼20.621; P , 0.0001; Fig. 5A and B). Examining the list of mapped genes that were significantly altered in both Foxp1-injected mouse striatum and human HD caudate using the Ingenuity Pathways Analysis software, we find that 52.5% (n ¼ 115) was significantly associated with inflammatory/immunological disease (Fisher’s exact test, P ¼ 3.43E212). We next compared the striatal expression levels of the same top 500 genes elicited by Foxp1 overexpression to gene expression abnormalities detected in several different HD mouse models. We included three microarray data sets performed previously by our group and on the same Illumina microarray platform, R6/2-Q150 (6 weeks old) (30), R6/2-Q300 (5 months old) (30) and DE5 transgenic mice (14 months old) (31) and three performed on the Affymetrix platform, the R6/2-Q150 (12 weeks old) (24), YAC128 (12 months old) (25) and YAC128 (24 months old) (25) [all freely available in the NCBI gene expression omnibus (GEO) database]. Similar to what was observed when compared with the human HD data set, Foxp1-elicited expression changes were significantly negatively correlated with the profiles of the HD mouse models tested, with the exception of the YAC128 mice at 24 months of age (Fig. 5B). In contrast, the expression profiles of each mouse model were positively correlated with one another and to the human HD data set (Fig. 5B). In addition, more concordant gene expression changes were observed between Foxp1 overexpression and the different HD mouse models than between Foxp1 overexpression and HD human brains. This is visualized in the scatterplots in Figure 5A and C, whereby expression

change concordance, reflected by the log2 ratios, is shown in the upper right (upregulated–upregulated) and lower left (downregulated–downregulated) quadrants. Foxp1 interacts with mutant Htt in mouse brain Taken in whole, our results suggest the involvement of Foxp1 in Htt-elicited transcriptional dysregulation. To further explore the potential role of Foxp1 in HD pathology, we investigated whether Foxp1 could directly interact with the Htt protein in the brains of R6/1 HD trangenic mice. Co-immunolocalization experiments revealed that .98% of Foxp1-positive neurons in the striatum of symptomatic, 6-month-old R6/1 transgenic mice also contained large neuronal nuclear Htt aggregates (Supplementary Material, Fig. S3). To test for a direct interaction between Foxp1 and Htt, we performed co-immunoprecipitation experiments in the brains of R6/1 transgenic mice. Immunoprecipitation of Htt in homogenates from the cortex of R6/1 transgenic mice and WT control mice using an antibody directed against the N-terminal region of Htt resulted in the co-precipitation of Foxp1 (Fig. 6A). Conversely, immunoprecipitation with the Foxp1 antibody resulted in the co-precipitation of the Htt protein (Fig. 6B).

DISCUSSION In this study, we used genome-wide methodologies to identify Foxp1 target genes both in vitro (Foxp1-transfected striatal cells) and in vivo (FOXP1- injected striatum). Our results reveal that Foxp1-regulated genes are strongly associated

3104

Human Molecular Genetics, 2012, Vol. 21, No. 14

Figure 4. Expression of Foxp1 in CNS cells. (A) Foxp1 expression is localized to neurons in WT mouse brain (3 months of age). Top row represents localization of Foxp1 and the neuronal marker NeuN. Foxp1 protein expression did not localize with the astrocyte marker GFAP (middle row) or microglial marker Cd11b (bottom row). Immunofluorescent images of 25 mm coronal sections were taken at ×100 magnification. Scale bars ¼ 10 mm. (B) Primary cultures of microglia from WT, YAC18 and YAC128 mice express Foxp1 under basal conditions, with expression in medium spiny neurons (MSN) shown for a reference. Pro-inflammatory stimulation of microglia with interferon gamma and endotoxin resulted in significant upregulation of Foxp1 mRNA levels assessed by a two-way ANOVA (WT fold-change ¼ 1.28; YAC18 fold-change ¼ 2.61; treatment effect: F ¼ 5.507, P ¼ 0.027. Data are from four to seven independent experiments per condition.

with immune-related functions in the CNS and that Foxp1 is linked to HD-related pathological pathways, especially those associated with glial cell activation. Prior to this study, the functional significance for Foxp1 in the adult CNS was not known. However, Foxp1 has previously been shown to play diverse roles in the functioning of the immune system, where it is expressed in lymphocytes, monocytes and macrophages (32,33). Foxp1 was originally cloned from a mouse B-cell leukemia cell line and was subsequently shown to be an essential transcriptional regulator of B lymphopoiesis (5) and B-cell activation (34). Foxp1 can also act as a transcriptional repressor regulating monocyte differentiation (6) and is essential for the generation of quiescent naive T cells during thymocyte development (35). In immune cells, Foxp1 has been shown to form heterodimers with Foxp3, another Foxp family member known to be essential to immune cell regulation (36– 38). In this study, we show that Foxp1 is fundamentally associated with repression of immune signaling in the CNS, similar to its role in immune cells. In cultured striatal cells, Foxp1 elicited both up- and downregulated changes in the expression of immune-related genes, such as those encoding different types of cytokines. However,

Foxp1 overexpression in vivo resulted in primarily downregulated expression changes of immune-related genes. Further, we found that the immune-related genes downregulated by Foxp1 in vivo were significantly associated with glial cells, especially astrocytes and microglia. One pathway showing statistically significant inhibition by Foxp1 is the NF-kB signaling pathway. As a transcriptional repressor, Foxp1 might be expected to directly repress the activity of genes in the NF-kB pathway; however, it is also possible that the expression changes of these genes are secondary. NF-kB is widely known to play important roles in inflammation and immune responses, as well as cell division regulation, apoptosis and glial cell activation in the CNS (39). In general, stimuli, such as TNFa, IL-1 and IL-6 trigger NF-kB activation both in astrocytes and microglia (Fig. 2) resulting in the production of proinflammatory mediators, including chemokines and cytokines (40). The NF-kB-signaling pathway has been shown to be upregulated in HD transgenic mice (41), and specific increases in proinflammatory cytokines, including IL-6, IL-8 and TNFa, have been demonstrated in the post-mortem striatum from HD patients (42,43). This is consistent with the astrocyte and microglia activation known to occur in the human HD brain (44– 48). Genes encoding many of these same proinflammatory mediators and their receptors were found to be decreased in expression by Foxp1 in the striatum, suggesting that Foxp1 could be working to counteract the effects of glia activation in the CNS. Although glia activation in the human HD brain is substantial, it is not clear whether activated glial cells in the brain represent a secondary effect to neuronal cell death, a primary contributor, or represent a neuroprotective mechanism. Data are emerging to implicate immune activation as playing an important role in the pathogenic mechanism of cell death. In particular, studies using positron emission tomography detected activated microglia in preclinical HD brains that correlated with striatal neuronal dysfunction, suggesting that microglial activation is an early event in the pathogenic processes of HD, and may be associated with subclinical progression of disease (49,50). It is also possible that glial activation plays a neuroprotective role, in which case Foxp1-mediated suppression of immune signaling could be detrimental, further contributing to neuronal cell death. For example, IL-10, which is upregulated in several HD mouse models (43) and down-regulated by Foxp1 (Supplementary Material, Table S4), is a cytokine with potent anti-inflammatory properties, repressing the expression of inflammatory cytokines, such as TNFa, IL-6 and IL-1b (51). Its primary recognized biological function relates to the limitation and termination of inflammatory responses and the regulation of differentiation and proliferation of several immune cells. Although IL-10 is considered a potent anti-inflammatory cytokine, studies have also suggested that IL-10 mediates immunostimulatory effects that help eliminate infectious agents (51–53). Hence, a complex interplay of cytokine signaling likely exists in the CNS, especially during disease states, although the exact role of Foxp1 in this requires further investigation. Glial activation in the HD brain is also reflected in microarray gene expression studies performed on human HD caudate. The expression of genes associated with gliosis and neuroinflammatory processes, such as GFAP, and those encoding tumor necrosis factor receptor superfamily members and

Human Molecular Genetics, 2012, Vol. 21, No. 14

3105

Figure 5. Correlation of the log2 ratios (M) of the top 500 FOXP1-elicited gene expression changes in the striatum across different HD mouse models and human HD caudate. (A) Scatterplots of the log2 ratios of FOXP1-elicited expression changes compared with those from human HD caudate (23). (B) Heatmap depiction of the linear correlation coefficients of M for all pair-wise comparisons between the HD models indicated. (C) Scatter plots of the comparisons of M between FOXP1-injected striatum against different HD mouse models as indicated: R6/2-Q150 (6 weeks) (30), R6/2-Q300 (30), DE5 transgenic mice (31), R6/2-Q150 (12 weeks) (24), YAC128 (12 months) (25) and YAC128 (24 months) (25).

receptors, interleukins and complement components, were found to be up-regulated in human HD caudate (23). Given the evidence for a role of Foxp1 as an immune repressor in the CNS, including the robust downregulation of GFAP expression that we validated by RT-qPCR analysis, the highly significant negative correlation in expression changes observed when comparing the human expression profiles to those elicited by Foxp1 overexpression is not surprising. Further, the fact that 52% of the genes that were significantly altered in both Foxp1-injected mouse striatum and in human HD caudate were significantly associated with inflammatory/ immunological disease (P ¼ 3.43E212), implicates Foxp1 as a mediator of these glial effects. The significant glia molecular signature of Foxp1 target genes begs the question of which cell types express Foxp1 in the adult mouse brain. In the striatum, we found that

Foxp1 is predominantly expressed by neurons, based on co-localization with the neuronal marker NeuN. In contrast, Foxp1 was not found to be co-localized with the astrocyte marker, GFAP or the microglial marker, Cd11b; however, RT-qPCR analysis revealed the presence of Foxp1 mRNA in purified microglia cultures, where it was upregulated upon activation in WT and YAC18 microglia cultures. Interestingly, Foxp1 levels were not upregulated in microglia from YAC128 mice. One explanation is that the presence of mutant Htt, which is now known to be expressed in glial cells, alters the ability of these cells to be activated, or alternatively, glial cells from YAC128 mice are already in a partially activated state, such that further activation by interferon gamma and endotoxin results in less-pronounced effects. Importantly, lentiviral-mediated delivery of Foxp1 induces expression in a similar manner, preferentially transducing neurons,

3106

Human Molecular Genetics, 2012, Vol. 21, No. 14

Figure 6. Interaction between Foxp1 and mutant Htt protein. (A) Co-immunoprecipitation of Foxp1 and mutant Htt protein in cortical homogenates from R6/1 transgenic mice (6 months of age). The immunoprecipitation (IP) was performed using an anti-N-terminal Htt antibody (EM48) followed by western blotting using an anti-Foxp1 antibody. (B) Reverse co-immunoprecipitation of mutant Htt protein by Foxp1, where the IP was performed using the anti-Foxp1 antibody followed by western blotting using the anti-Htt EM48 antibody.

with only a small effect on glial cells (54,55). Previous studies have shown that, after cytomegalovirus (CMV) promoterdriven, lentiviral-mediated expression of GFP in the striatum, 82% of transduced cells showed co-localization with the neuronal marker, NeuN, compared with only 8% showing co-localization with GFAP (54). Only a few scattered cells showed double labeling with Ng2 (an oligodendrocytic marker), and microglial markers were not tested in that study (54). Hence, the strong glial signature associated with the Foxp1 target genes in vivo could result from overexpression of Foxp1 in a low percentage of glia cells or overexpression of Foxp1 in neurons that express immune-related genes. For example, many immune-related genes that are typically associated with astrocyte or microglia are also expressed by neurons, including many cytokines (TNFa, IL-1 and IL-6) (56–58) and chemokines (CCL2-CCL5, CXCL1 and CXCL10) (59). Additionally, the expression of glial-expressed genes could represent a secondary response to Foxp1-elicited expression of target genes involved in neuron-glia cross-talk in the brain. Accumulating evidence suggests that neuron-glia communication and reciprocal signaling play important roles in neurodegenerative disorders; however, the detailed mechanisms are not completely characterized. Several genes upregulated in response to Foxp1 overexpression could represent important players in cell-to-cell communication in the brain. For example, in vivo overexpression of Foxp1 resulted in an increase in the expression of suppressor of cytokine signaling 5 (Socs5; see Supplementary Material, Table S4). This protein family of suppressors of cytokine signaling (SOCS) is known to be negative regulators of cytokine signaling under diverse conditions (60). Nine members (SOCS1–SOCS9) have been identified to date, although the specific functions of each have not yet been determined. Such an upregulation of Socs5 by Foxp1 in neurons could elicit a secondary effect on glia cells resulting in a decrease in expression of cytokine genes, such as those we detected in this study. Another upregulated gene, Gdnf, encodes glial cell line-derived neurotrophic factor. This factor has previously been shown to protect dopaminergic neurons by suppressing microglia activation and subsequent excessive nitric oxide production (61). Our previous studies have demonstrated decreased Foxp1 mRNA expression in R6/1 transgenic mice and in human postmortem caudate from HD patients (62) (although mouse Foxp1

mRNA levels were not altered in 3-month-old YAC128, which were used for the human FOXP1 striatal injections). In the current study, we found that overexpression of human FOXP1 resulted in a 1.53-fold elevation of mouse Foxp1 mRNA (P ¼ 0.00015), a finding that was validated by RT-qPCR, providing additional evidence to our in vitro ChIP-seq findings that Foxp1 regulates its own expression and also suggesting that part of the increased mouse Foxp1 expression we detected in striatal cells was from endogenous sources. Importantly, we demonstrated an interaction between mutant Htt protein and Foxp1 by co-immunoprecipitation assays and found Foxp1 to be localized in Htt aggregates in the striatum of R6/1 transgenic mice. We suggest that the decreased Foxp1 expression levels observed in the striata of HD mouse models and human subjects is due to Htt-mediated sequestration, and subsequent loss of function of Foxp1 self-regulation. Lower Foxp1 levels, in turn, compounded by Htt-mediated sequestration, lead to a loss in the repression of immune-related genes, resulting in elevated levels of damaging cytokines and chemokines, which are observed in HD brain. Therefore, we suggest that Foxp1 plays a neuroprotective role under normal physiological conditions, whereby it counteracts the effects of glia activation elicited in response to various pathological stimuli. Further, our findings implicate Foxp1-mediated pathways in neuron-glia communication, providing mechanistic insight into the connections between glial cell activation and neuronal dysfunction.

MATERIALS AND METHODS Striatal cell culture and RNA preparation Conditionally immortalized WT STHdhQ7striatal neuronal progenitor cells were a kind donation from Dr Marcy MacDonald (21). The striatal cells were grown at 338C in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 1% non-essential amino acids, 2 mM L-glutamine and 400 mg/ml G418 (Sigma-Aldrich, St Louis, MO, USA). Cells were plated at 3 × 105 cells per well in six-well tissue culture plates and the following day the cells were transfected with an expression plasmid (pcDNA3.1) containing the cDNA of mouse Foxp1 (nts 312 – 2429 of accession # NM_053202) using Polyjet transfection reagent (SignaGen, Ijamsville, MD, USA) according to the manufacturer’s instructions. Empty expression

Human Molecular Genetics, 2012, Vol. 21, No. 14

vector, pcDNA3.1, was used to ensure that all wells received the same amount of total DNA. Transfection efficiency was assessed in duplicate sets of transfected cells by quantifying the percentage of GFP-positive cells using fluorescence microscopy and by measuring Foxp1 mRNA expression with qPCR. We did not observe any notable changes in cell growth or morphology in striatal cells overexpressing Foxp1. Two days after transfection, cells were harvested, and RNA was extracted using RNAeasy mini kit (Qiagen) with DNase I treatment to eliminate genomic DNA contamination. Lentiviral injections in mice The human FOXP1 cDNA was subcloned from pcDNA3.1 into the third-generation self-inactivating lentivirus vector (pLV-Bobi) with the human CMV promoter driving expression. The resulting vector plasmids were packaged with the three packaging plasmids by transient transfection in 293T cells as previously described (63) to generate LV-hFoxP1. Lentiviruses expressing GFP, LV-GFP or empty control, LV-Control, have been previously described (55). Bilateral striatal injections with the FOXP1 lentiviral vector were performed in anesthetized YAC128 mice (age 2 – 4 months) (n ¼ 9), which express a full-length human mutant Htt protein containing 128 CAG repeats driven by the endogenous human HTT promoter (26). Littermate YAC128 mice were injected with lenti-GFP vector as controls (n ¼ 8). In total, 5 ml of viral preparation (5 × 108 TDU) was delivered per site at a rate of 0.5 ml per min. The needle was kept at the injection site 2 min prior to injection and 5 min post-injection. Striatal tissue was collected 4 weeks after lentiviral delivery. Each striatal hemisphere was collected separately and snap-frozen in liquid nitrogen immediately upon dissection. RNA was extracted from each striatal hemisphere using the PurelinkTM RNA Mini kit (Invitrogen, Carlsbad, CA, USA), including the on-column DNase treatment (Invitrogen). Reverse transcription was performed using the Superscript VILOTM cDNA synthesis kit (Invitrogen) according to the manufacturer’s instructions. Microarray experiments Total RNA was quantified using NanoDrop (ND-1000, manufacturer) and quality was checked with the Agilent 2100 Bioanalyzer using the RNA 6000 Nano LabChip. One microgram of total RNA was taken through Ambion’s Illumina Total Prep RNA Amplification System (protocol available at http://ambion.com/ techlib/prot/fm_IL1791.pdf). Post-amplification RNA (750 ng cRNA) product was hybridized onto the Illumina Sentrix BeadChip Array MouseRef-8 v2 for 18 h at 588C (protocol available at http://www.illumina.com/products/mouseref-8_expression_bea dchip_kits_v2.ilmn). After hybridization, the BeadChip Arrays were washed and stained as per protocol requirements. BeadChip Arrays were scanned using the Illumina BeadArray Reader with default settings. Data normalization was performed using GenomeStudio Gene Expression Module v1.0 with quantile normalization (64,65). The sample clustering was done using BRBArrayTools (http://linus.nci.nih.gov/BRB-ArrayTools.html), with centered correlation and average linkage. The Limma package in the R software was used to find transcripts showing

3107

differential expression in response to Foxp1 overexpression (66). Microarray data are available at GEO accession # GSE31560. Chromatin-immunoprecipitation sequencing (ChIP-seq) ChIP assays on cultured striatal cells were performed as described in previous studies (67). Briefly, cellular homogenates were cross-linked by incubating with 1% of formaldehyde for 15 min at room temperature, and then homogenizing to isolate nuclei. DNA was sonicated in 100 ml of buffer [1% sodium dodecyl sulfate (SDS), 10 mM ethylenediaminetetraacetic acid (EDTA), 50 mM Tris–HCl, pH 8.0, 1× protease inhibitors cocktail] per 15 mg of brain to achieve 0.2–1.0 kb sized DNA fragments. An aliquot (100 ml) of precleared homogenates was diluted with dilution buffer (1% Triton X-100, 2 mM EDTA, 20 mM Tris–HCl, pH 8.0, 150 mM NaCl and 1× protease inhibitors cocktail), and incubated with 3 mg of rabbit anti-Foxp1 antibody (Aviva Systems Biology) or 3 mg of control IgG (BD Biosciences) overnight at 48C, followed by hybridization to 60 ml of protein-sepharose A beads for 2 h at 48C. The protein–DNA complexes were washed with a low salt buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris–HCl, pH 8.0, 150 mM NaCl) three times, followed by a high salt buffer wash (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris–HCl, pH 8.0, 500 mM NaCl). Elution buffer (1% SDS and 0.1 M NaHCO3) was then added and incubated at 658C for 10 min, and then incubated with proteinase K for 30 min at 378C. The cross-linking reaction was reversed by incubating at 658C overnight. DNA was purified with DNA purification kit (ZYMO research Corp, Irvine, CA, USA). The recovered DNA, as well as input DNA, was used for massively parallel sequencing. The sequencing libraries were prepared using 10 ng doublestranded DNA prepared by ChIP (pull-down and input). ChIP DNA samples ends were repaired using 15 units T4 DNA polymerase, 5 units Klenow Large Fragment and 50 units T4 polynucleotide kinase at 208C for 30 min in 50 mM Tris–HCl (pH 7.5), 10 mM MgCl2, 1 mM adenosine 5′ -triphosphate (ATP), 10 mM dithiothreitol and 400 mM dNTP mixture. DNA products were purified again using DNA Clean & ConcentratorTM -5 Kit (Zymo Research). Next, DNA ends were A-tailed with 15 units Klenow (3′ 5′ exo) at 378C for 30 min in 10 mM Tris–HCl (pH 7.9), 50 mM NaCl, 10 mM MgCl2, 1 mM dithiothreitol and 0.2 mM dATP. DNA products were again purified using the DNA Clean&ConcentratorTM -5 Kit. Illumina Paired End-adapter oligonucleotides (0.33 mM) were then ligated to the A-tailed cDNA ends with 3000 units T4 DNA ligase at 208C for 15 min in 66 mM Tris–HCl (pH 7.6), 10 mM MgCl2, 1 mM dithiothreitol, 1 mM ATP and 1 mM PEG in a 30 ml reaction volume. DNA products were purified using DNA Clean&ConcentratorTM -5 Kit. The DNA library products were separated on an Invitrogen 2% Size-Select agarose gel and products corresponding to a size of 200–250 bases were removed from the gel and cleaned using the Agencourt SPRI system. The DNA material was PCR amplified with 1 unit of PhusionTM Polymerase in standard 1X PhusionTM HF buffer with 0.2 mM dNTPs and 0.6 mM PCR primers PE 1.0 and PE 2.0 for 15 cycles. The amplified DNA products were further purified on 2% NuSieve GTGw agarose gel, excised and

3108

Human Molecular Genetics, 2012, Vol. 21, No. 14

isolated again using ZymocleanTM Gel DNA recovery kit. The purified DNA library was quantitated using the Qubit quantitation platform (Invitrogen) and sized using the 2100 Bioanalyzer. DNA products were then denatured in 0.1N NaOH and diluted to a final concentration of 10 pM before being loaded onto the Illumina paired-end flow cell for massively parallel sequencing by synthesis on the Illumina GAIIx. The Genome Analyzer Pipeline Software v1.7 was used to perform the early data analysis of a sequencing run, including the image analysis, base calling and alignment. Alignment was performed with Efficient Large-Scale Alignment of Nucleotide Databases (ELAND2). The uniquely aligned reads containing less than three errors to the mouse genome are used as input to the Model-based Analysis for ChIP-Seq (MACS) program, a publicly available open source ChIP-Seq analysis (http://liulab.dfci.harvard.edu/MACS/) (68). Peaks found by MACS (v1.4alpha2) were mapped to within 10 kb of a refSeq transcript of the mouse version mm9 database. The peak coordinates were then positioned outside the 5′ or 3′ transcript coordinates or within exons/introns coordinates for the annotated genes, by comparing to the transcript and exon start/end positions from the UCSC Genome Browser (http://genome.ucsc. edu/). The peak sequences were obtained from queries to the UCSC Genome database using the chromosome coordinates. Real-time qPCR analysis For in vitro experiments, real-time qPCR analysis was performed on cDNA prepared from striatal cells/homogenates using primers designed in the exonic regions of selected genes (Supplementary Material, Table S6). Primers were designed to generate amplicons of 80– 150 nucleotides with similar melting temperatures (648C) using Invitrogen’s Primer Designer. The real-time PCR thermal cycling was performed using the ABI PRISMs 7900HT Sequence Detection System (Applied Biosysterms, Foster City, CA, USA) as described previously (16). The amount of cDNA in each sample was calculated using SDS2.1 software by the comparative threshold cycle (Ct) method and expressed as 2exp(Ct). Hypoxanthine guanine phosphoribosyl transferase (Hprt) was used as an internal control. For in vivo gene expression experiments, real-time qPCR analysis was performed on cDNA generated from striatal RNA from nine FOXP1 lentiviral-treated YAC128 mice and eight GFP lentiviral-treated YAC128 mice. Amplification of cDNA was performed using the StepOneTM Real-Time PCR System (Applied Biosystems) and analyses were performed using the StepOne software v2.1 (Applied Biosystems). Primers were constructed over exon/exon boundaries using the Primer Express software v3.0 (Applied Biosystems). Primer sequences are available as Supplementary Material, Table S6. Relative quantification of mRNA levels was calculated using the standard curve method. The relative amount of mRNA in each well was calculated as the ratio between the target mRNA and a normalization factor (NF). GeNorm software (http://medgen.ugent.be/~jvdesomp/genorm/) was used for calculation of the most accurate NF, which was based on geometric averaging of multiple reference genes and analyses were based on the average expression stability and pairwise variation. An accurate normalization factor (NF4) was

calculated based on Pak1ip1, Cyc1, Hprt and Gapdh for the in vivo lentiviral-experiments; and based on Pak1ip1, Rpl13a and Rplp0 (NF3) for the microglia experiments. Statistical differences in gene expression were determined using Student’s t-test (unpaired, two-tailed) (Graph Pad Prism version 5.0). Co-immunofluorescence Animals were perfused by cardiac puncture with 4% paraformaldehyde, post-fixed for 24 h and then transferred to 30% sucrose overnight. Mounted sections (25 mm) were permeabilized in 0.1% Triton X-100 in PBS, blocked using 10% normal donkey serum and incubated at RT overnight in primary antibody. The primary antibodies used were: neuronal marker NeuN (Millipore; Temecula, CA, USA; 1:1000), astrocyte marker GFAP (Sigma-Aldrich; 1:1000), microglial marker CD11b (AbD serotec; Oxford, UK; 1:1000) and Foxp1 (Cell Signaling Technology, Danvers, MA, USA; 1:400). Immunofluorescent detection was achieved by an 1 hour incubation at room temperature with secondary antibodies appropriate to recognize primary antibodies. The secondary antibodies used were: donkey anti-rabbit, donkey anti-mouse IgG antibodies complexed to either Alexa 488 or Alexa 594 dyes (Molecular Probes, Invitrogen; 1:1000). All sections were then incubated for 2 min in 4′ ,6-diamidino-2-phenylindole (DAPI; 1:10 000; Sigma) at RT. Sections were cover slipped using DEPEX mounting media (Electron Microscopy Sciences; Hatfield, PA, USA) and stored at RT. Co-immunoprecipitation experiments Co-immunoprecipitation of Htt and Foxp1 proteins was performed on the cortex from WT and R6/1 transgenic mice (3 animals/condition, 6 months old) as described previously (62). For the immunoprecipitation, aliquots containing 650 mg of protein were incubated for 3 h with either 1 mg of mouse anti-htt/exon1 antibody, EM48 (Chemicon International, Temecula, CA), 1 mg of anti-Foxp1 antibody (Bethyl Laboratories, Montgomery, TX, USA) or 1 mg of a normal mouse IgG (Santa Cruz Biotechnology, Santa Cruz, CA, USA) as control, in a rocker at 48C. Although EM48 preferentially binds mutant aggregated Htt, it can also recognize the WT form of the protein (69,70). Protein A-sepharose/PBS 50:50 v/v was added to the homogenates and incubated overnight on a rocker at 48C. Collection of the conjugates was done by spinning the samples at 12 000g and thoroughly washing the pellets. Protein-sepharose A beads were resuspended in 50 mM TRIS/HCl, pH 6.5; 1× Nupage loading buffer (Invitrogen) and 40 mM DTT and heated at 1008C for 5 min before separation by SDS–polyacrylamide gel electrophoresis and subsequent western blotting. Western blotting was performed using anti-Foxp1 (1:500 dilution) or mouse anti-Htt/exon1 antibody EM48 (1:500 dilution). Equal loading of western blots was verified by Ponceau S stain. Comparison of FOXP1-elicited gene expression changes to Htt-elicited changes Human homologs of the top 500 FOXP1-induced mouse genes in the striatum (the top 250 downregulated genes and the top

Human Molecular Genetics, 2012, Vol. 21, No. 14

250 upregulated genes based on rank) were mapped and their expression differences screened in the human HD caudate dataset provided in Hodges et al. (23). We also compared these top 500 genes to several different HD mouse models, including R6/2-Q150 (6 weeks) (30), R6/2-Q300 (30), DE5 transgenic mice (31), R6/2-Q150 (12 weeks) (24), YAC128 (12 months) (25) and YAC128 (24 months) (25) (all freely available in the NCBI GEO database). Where multiple mouse probe sets representing a single gene were identified, we selected the mouse probe set reporting the most significant change. The corresponding M values (log2-fold changes) for each of the top genes in each model were subjected to Pearson product-moment correlation analyses to calculate the pair-wise distances between every pair of HD models (based on Pearson’s correlation coefficient).

Microglia cultures Whole brains were obtained from post-natal day 1 WT, YAC18 transgenic and YAC128 transgenic mouse pups and placed in Hanks Balanced Salt Solution on ice. Meninges were removed carefully and the remaining brain tissue was placed into the growth medium [DMEM, 10% FBS (PAA), 1% L-glutamine, 1% penicillin/streptomycin (Invitrogen)] and homogenized using a 5 ml pipette. Cells from each brain were pelleted, resuspended in the growth medium and transferred into a T150 flask that was cultured at 378C at 5% CO2. The growth medium was replaced with fresh media after 24 h and then every 7 days. After 18–21 days in culture, loosely attached microglia were harvested and seeded at 6 × 105 cells/ml with their own conditioned media into six-well PRIMARIATM tissue culture plates (BD Falcon). Cells were plated on glass cover slips for verification of the purity of the established cultures. The media were replaced by 1% DMEM medium (DMEM, 1% FBS, 0.1% L-glutamine) after 24 h. IFNg (R&D Systems) (final concentration 10 ng/ml) was added to prime the cells 1 h before stimulation with control standard endotoxin (CSE) (Associates of Cape Cod) (final concentration 2.5 ng/ml). Stimulated microglia cultures were incubated at 378C at 5% CO2 for 24 h prior to harvesting.

SUPPLEMENTARY MATERIAL Supplementary Material is available at HMG online.

ACKNOWLEDGEMENTS We would like to thanks Lana Schaffer and the Scripps DNA Array Core facility for performing ChIP-seq assays. Conflict of Interest statement. None declared.

FUNDING These studies were funded by NIH grant NS44169 to E.A.T. and NIH grants AG022074, AG18440, AG10435 and NS050041 to E.M.

3109

REFERENCES 1. Carlsson, P. and Mahlapuu, M. (2002) Forkhead transcription factors: key players in development and metabolism. Dev. Biol., 250, 1 –23. 2. Jackson, B.C., Carpenter, C., Nebert, D.W. and Vasiliou, V. (2010) Update of human and mouse forkhead box (FOX) gene families. Hum. Genomics, 4, 345–352. 3. Shu, W., Yang, H., Zhang, L., Lu, M.M. and Morrisey, E.E. (2001) Characterization of a new subfamily of winged-helix/forkhead (Fox) genes that are expressed in the lung and act as transcriptional repressors. J. Biol. Chem., 276, 27488– 27497. 4. Wang, B., Weidenfeld, J., Lu, M.M., Maika, S., Kuziel, W.A., Morrisey, E.E. and Tucker, P.W. (2004) Foxp1 regulates cardiac outflow tract, endocardial cushion morphogenesis and myocyte proliferation and maturation. Development, 131, 4477– 4487. 5. Hu, H., Wang, B., Borde, M., Nardone, J., Maika, S., Allred, L., Tucker, P.W. and Rao, A. (2006) Foxp1 is an essential transcriptional regulator of B cell development. Nat. Immunol., 7, 819–826. 6. Shi, C., Zhang, X., Chen, Z., Sulaiman, K., Feinberg, M.W., Ballantyne, C.M., Jain, M.K. and Simon, D.I. (2004) Integrin engagement regulates monocyte differentiation through the forkhead transcription factor Foxp1. J. Clin. Invest., 114, 408–418. 7. Dasen, J.S., De Camilli, A., Wang, B., Tucker, P.W. and Jessell, T.M. (2008) Hox repertoires for motor neuron diversity and connectivity gated by a single accessory factor, FoxP1. Cell, 134, 304–316. 8. Rousso, D.L., Gaber, Z.B., Wellik, D., Morrisey, E.E. and Novitch, B.G. (2008) Coordinated actions of the forkhead protein Foxp1 and Hox proteins in the columnar organization of spinal motor neurons. Neuron, 59, 226– 240. 9. Lai, C.S., Fisher, S.E., Hurst, J.A., Vargha-Khadem, F. and Monaco, A.P. (2001) A forkhead-domain gene is mutated in a severe speech and language disorder. Nature, 413, 519–523. 10. MacDermot, K.D., Bonora, E., Sykes, N., Coupe, A.M., Lai, C.S., Vernes, S.C., Vargha-Khadem, F., McKenzie, F., Smith, R.L., Monaco, A.P. and Fisher, S.E. (2005) Identification of FOXP2 truncation as a novel cause of developmental speech and language deficits. Am. J. Hum. Genet., 76, 1074–1080. 11. Vernes, S.C., Nicod, J., Elahi, F.M., Coventry, J.A., Kenny, N., Coupe, A.M., Bird, L.E., Davies, K.E. and Fisher, S.E. (2006) Functional genetic analysis of mutations implicated in a human speech and language disorder. Hum. Mol. Genet., 15, 3154–3167. 12. Hamdan, F.F., Daoud, H., Rochefort, D., Piton, A., Gauthier, J., Langlois, M., Foomani, G., Dobrzeniecka, S., Krebs, M.O., Joober, R. et al. (2010) De novo mutations in FOXP1 in cases with intellectual disability, autism, and language impairment. Am. J. Hum. Genet., 87, 671– 678. 13. Horn, D., Kapeller, J., Rivera-Brugues, N., Moog, U., Lorenz-Depiereux, B., Eck, S., Hempel, M., Wagenstaller, J., Gawthrope, A., Monaco, A.P. et al. (2010) Identification of FOXP1 deletions in three unrelated patients with mental retardation and significant speech and language deficits. Hum. Mutat., 31, E1851– E1860. 14. Ferland, R.J., Cherry, T.J., Preware, P.O., Morrisey, E.E. and Walsh, C.A. (2003) Characterization of Foxp2 and Foxp1 mRNA and protein in the developing and mature brain. J. Comp. Neurol., 460, 266– 279. 15. Tamura, S., Morikawa, Y., Iwanishi, H., Hisaoka, T. and Senba, E. (2003) Expression pattern of the winged-helix/forkhead transcription factor Foxp1 in the developing central nervous system. Gene Expr. Patterns, 3, 193– 197. 16. Desplats, P.A., Kass, K.E., Gilmartin, T., Stanwood, G.D., Woodward, E.L., Head, S.R., Sutcliffe, J.G. and Thomas, E.A. (2006) Selective deficits in the expression of striatal-enriched mRNAs in Huntington’s disease. J. Neurochem., 96, 743 –757. 17. Takahashi, K., Liu, F.C., Hirokawa, K. and Takahashi, H. (2003) Expression of Foxp2, a gene involved in speech and language, in the developing and adult striatum. J. Neurosci. Res., 73, 61–72. 18. The Huntington’s Disease Collaborative Research Group. (1993) A novel gene containing a trinucleotide repeat that is expanded and unstable on Huntington’s disease chromosomes. The Huntington’s Disease Collaborative Research Group. Cell, 72, 971– 983. 19. Sugars, K.L. and Rubinsztein, D.C. (2003) Transcriptional abnormalities in Huntington disease. Trends Genet., 19, 233–238. 20. Thomas, E.A. (2006) Striatal specificity of gene expression dysregulation in Huntington’s disease. J. Neurosci. Res., 84, 1151–1164.

3110

Human Molecular Genetics, 2012, Vol. 21, No. 14

21. Trettel, F., Rigamonti, D., Hilditch-Maguire, P., Wheeler, V.C., Sharp, A.H., Persichetti, F., Cattaneo, E. and MacDonald, M.E. (2000) Dominant phenotypes produced by the HD mutation in STHdh(Q111) striatal cells. Hum. Mol. Genet., 9, 2799– 2809. 22. Kleinjan, D.A. and van Heyningen, V. (2005) Long-range control of gene expression: emerging mechanisms and disruption in disease. Am. J. Hum. Genet., 76, 8 –32. 23. Hodges, A., Strand, A.D., Aragaki, A.K., Kuhn, A., Sengstag, T., Hughes, G., Elliston, L.A., Hartog, C., Goldstein, D.R., Thu, D. et al. (2006) Regional and cellular gene expression changes in human Huntington’s disease brain. Hum. Mol. Genet., 15, 965– 977. 24. Kuhn, A., Goldstein, D.R., Hodges, A., Strand, A.D., Sengstag, T., Kooperberg, C., Becanovic, K., Pouladi, M.A., Sathasivam, K., Cha, J.H. et al. (2007) Mutant huntingtin’s effects on striatal gene expression in mice recapitulate changes observed in human Huntington’s disease brain and do not differ with mutant huntingtin length or wild-type huntingtin dosage. Hum. Mol. Genet., 16, 1845– 1861. 25. Becanovic, K., Pouladi, M.A., Lim, R.S., Kuhn, A., Pavlidis, P., Luthi-Carter, R., Hayden, M.R. and Leavitt, B.R. (2011) Transcriptional changes in Huntington disease identified using genome-wide expression profiling and cross-platform analysis. Hum. Mol. Genet., 19, 1438–1452. 26. Hodgson, J.G., Agopyan, N., Gutekunst, C.A., Leavitt, B.R., LePiane, F., Singaraja, R., Smith, D.J., Bissada, N., McCutcheon, K., Nasir, J. et al. (1999) A YAC mouse model for Huntington’s disease with full-length mutant huntingtin, cytoplasmic toxicity, and selective striatal neurodegeneration. Neuron, 23, 181–192. 27. de Chaldee, M., Gaillard, M.C., Bizat, N., Buhler, J.M., Manzoni, O., Bockaert, J., Hantraye, P., Brouillet, E. and Elalouf, J.M. (2003) Quantitative assessment of transcriptome differences between brain territories. Genome Res., 13, 1646–1653. 28. Cahoy, J.D., Emery, B., Kaushal, A., Foo, L.C., Zamanian, J.L., Christopherson, K.S., Xing, Y., Lubischer, J.L., Krieg, P.A., Krupenko, S.A., Thompson, W.J. and Barres, B.A. (2008) A transcriptome database for astrocytes, neurons, and oligodendrocytes: a new resource for understanding brain development and function. J. Neurosci., 28, 264– 278. 29. Loerch, P.M., Lu, T., Dakin, K.A., Vann, J.M., Isaacs, A., Geula, C., Wang, J., Pan, Y., Gabuzda, D.H., Li, C., Prolla, T.A. and Yankner, B.A. (2008) Evolution of the aging brain transcriptome and synaptic regulation. PLoS ONE 3, e3329. 30. Tang, B., Seredenina, T., Coppola, G., Kuhn, A., Geschwind, D.H., Luthi-Carter, R. and Thomas, E.A. (2011) Gene expression profiling of R6/2 transgenic mice with different CAG repeat lengths reveals genes associated with disease onset and progression in Huntington’s disease. Neurobiol Dis., 42, 459–467. 31. Thomas, E.A., Coppola, G., Tang, B., Kuhn, A., Kim, S., Geschwind, D.H., Brown, T.B., Luthi-Carter, R. and Ehrlich, M.E. (2011) In vivo cell-autonomous transcriptional abnormalities revealed in mice expressing mutant huntingtin in striatal but not cortical neurons. Hum. Mol. Genet., 20, 1049– 1060. 32. Li, B., Samanta, A., Song, X., Iacono, K.T., Brennan, P., Chatila, T.A., Roncador, G., Banham, A.H., Riley, J.L., Wang, Q. et al. (2007) FOXP3 is a homo-oligomer and a component of a supramolecular regulatory complex disabled in the human XLAAD/IPEX autoimmune disease. Int. Immunol., 19, 825–835. 33. Banham, A.H., Beasley, N., Campo, E., Fernandez, P.L., Fidler, C., Gatter, K., Jones, M., Mason, D.Y., Prime, J.E., Trougouboff, P., Wood, K. and Cordell, J.L. (2001) The FOXP1 winged helix transcription factor is a novel candidate tumor suppressor gene on chromosome 3p. Cancer Res., 61, 8820– 8829. 34. Shaffer, A.L., Rosenwald, A. and Staudt, L.M. (2002) Lymphoid malignancies: the dark side of B-cell differentiation. Nat. Rev. Immunol., 2, 920–932. 35. Feng, X., Ippolito, G.C., Tian, L., Wiehagen, K., Oh, S., Sambandam, A., Willen, J., Bunte, R.M., Maika, S.D., Harriss, J.V. et al. (2010) Foxp1 is an essential transcriptional regulator for the generation of quiescent naive T cells during thymocyte development. Blood, 115, 510 –518. 36. Chung, H.S., Lee, J.H., Kim, H., Lee, H.J., Kim, S.H. and Kwon, H.K.ImS.H.BaeH. (2010) Foxp3 is a novel repressor of microglia activation. Glia, 58, 1247–1256.

37. Li, B. and Greene, M.I. (2007) FOXP3 actively represses transcription by recruiting the HAT/HDAC complex. Cell Cycle, 6, 1432–1436. 38. Schubert, L.A., Jeffery, E., Zhang, Y., Ramsdell, F. and Ziegler, S.F. (2001) Scurfin (FOXP3) acts as a repressor of transcription and regulates T cell activation. J. Biol. Chem., 276, 37672–37679. 39. Memet, S. (2006) NF-kappaB functions in the nervous system: from development to disease. Biochem. Pharmacol., 72, 1180– 1195. 40. Mattson, M.P. and Camandola, S. (2001) NF-kappaB in neuronal plasticity and neurodegenerative disorders. J. Clin. Invest., 107, 247–254. 41. Khoshnan, A., Ko, J., Watkin, E.E., Paige, L.A., Reinhart, P.H. and Patterson, P.H. (2004) Activation of the IkappaB kinase complex and nuclear factor-kappaB contributes to mutant huntingtin neurotoxicity. J. Neurosci., 24, 7999–8008. 42. Dalrymple, A., Wild, E.J., Joubert, R., Sathasivam, K., Bjorkqvist, M., Petersen, A., Jackson, G.S., Isaacs, J.D., Kristiansen, M., Bates, G.P. et al. (2007) Proteomic profiling of plasma in Huntington’s disease reveals neuroinflammatory activation and biomarker candidates. J. Proteome Res., 6, 2833–2840. 43. Bjorkqvist, M., Wild, E.J., Thiele, J., Silvestroni, A., Andre, R., Lahiri, N., Raibon, E., Lee, R.V., Benn, C.L., Soulet, D. et al. (2008) A novel pathogenic pathway of immune activation detectable before clinical onset in Huntington’s disease. J. Exp. Med., 205, 1869–1877. 44. Pavese, N., Gerhard, A., Tai, Y.F., Ho, A.K., Turkheimer, F., Barker, R.A., Brooks, D.J. and Piccini, P. (2006) Microglial activation correlates with severity in Huntington disease: a clinical and PET study. Neurology, 66, 1638–1643. 45. Sapp, E., Kegel, K.B., Aronin, N., Hashikawa, T., Uchiyama, Y., Tohyama, K., Bhide, P.G., Vonsattel, J.P. and DiFiglia, M. (2001) Early and progressive accumulation of reactive microglia in the Huntington disease brain. J. Neuropathol. Exp. Neurol., 60, 161– 172. 46. Shin, J.Y., Fang, Z.H., Yu, Z.X., Wang, C.E., Li, S.H. and Li, X.J. (2005) Expression of mutant huntingtin in glial cells contributes to neuronal excitotoxicity. J. Cell Biol., 171, 1001–1012. 47. Chou, S.Y., Weng, J.Y., Lai, H.L., Liao, F., Sun, S.H., Tu, P.H., Dickson, D.W. and Chern, Y. (2008) Expanded-polyglutamine huntingtin protein suppresses the secretion and production of a chemokine (CCL5/RANTES) by astrocytes. J. Neurosci., 28, 3277–3290. 48. Vonsattel, J.P., Myers, R.H., Stevens, T.J., Ferrante, R.J., Bird, E.D. and Richardson, E.P.J. (1985) Neuropathological classification of Huntington’s disease. J. Neuropathol. Exp. Neurol., 44, 559– 577. 49. Tai, Y.F., Pavese, N., Gerhard, A., Tabrizi, S.J., Barker, R.A., Brooks, D.J. and Piccini, P. (2007) Microglial activation in presymptomatic Huntington’s disease gene carriers. Brain, 130, 1759– 1766. 50. Tai, Y.F., Pavese, N., Gerhard, A., Tabrizi, S.J., Barker, R.A., Brooks, D.J. and Piccini, P. (2007) Imaging microglial activation in Huntington’s disease. Brain Res. Bull, 72, 148– 151. 51. Ouyang, W., Rutz, S., Crellin, N.K., Valdez, P.A. and Hymowitz, S.G. (2011) Regulation and functions of the IL-10 family of cytokines in inflammation and disease. Annu Rev. Immunol., 29, 71–109. 52. Asadullah, K., Sterry, W. and Volk, H.D. (2003) Interleukin-10 therapy– review of a new approach. Pharmacol. Rev., 55, 241– 269. 53. Lauw, F.N., Pajkrt, D., Hack, C.E., Kurimoto, M., van Deventer, S.J. and van der Poll, T. (2000) Proinflammatory effects of IL-10 during human endotoxemia. J. Immunol., 165, 2783– 2789. 54. Jakobsson, J., Ericson, C., Jansson, M., Bjork, E. and Lundberg, C. (2003) Targeted transgene expression in rat brain using lentiviral vectors. J. Neurosci. Res., 73, 876 –885. 55. Spencer, B., Potkar, R., Trejo, M., Rockenstein, E., Patrick, C., Gindi, R., Adame, A., Wyss-Coray, T. and Masliah, E. (2009) Beclin 1 gene transfer activates autophagy and ameliorates the neurodegenerative pathology in alpha-synuclein models of Parkinson’s and Lewy body diseases. J. Neurosci., 29, 13578– 13588. 56. Fattori, E., Lazzaro, D., Musiani, P., Modesti, A., Alonzi, T. and Ciliberto, G. (1995) IL-6 expression in neurons of transgenic mice causes reactive astrocytosis and increase in ramified microglial cells but no neuronal damage. Eur. J. Neurosci., 7, 2441–2449. 57. Liu, L., Li, Y., Van Eldik, L.J., Griffin, W.S. and Barger, S.W. (2005) S100B-induced microglial and neuronal IL-1 expression is mediated by cell type-specific transcription factors. J. Neurochem., 92, 546– 553.

Human Molecular Genetics, 2012, Vol. 21, No. 14

58. Gahring, L.C., Carlson, N.G., Kulmar, R.A. and Rogers, S.W. (1996) Neuronal expression of tumor necrosis factor alpha in the murine brain. Neuroimmunomodulation, 3, 289–303. 59. de Haas, A.H., van Weering, H.R., de Jong, E.K., Boddeke, H.W. and Biber, K.P. (2007) Neuronal chemokines: versatile messengers in central nervous system cell interaction. Mol. Neurobiol., 36, 137–151. 60. Larsen, L. and Ropke, C. (2002) Suppressors of cytokine signalling: SOCS. APMIS, 110, 833–844. 61. Xing, B., Xin, T., Zhao, L., Hunter, R.L., Chen, Y. and Bing, G. (2011) Glial cell line-derived neurotrophic factor protects midbrain dopaminergic neurons against lipopolysaccharide neurotoxicity. J. Neuroimmunol., 225, 43– 51. 62. Desplats, P.A., Lambert, J.R. and Thomas, E.A. (2008) Functional roles for the striatal-enriched transcription factor, Bcl11b, in the control of striatal gene expression and transcriptional dysregulation in Huntington’s disease. Neurobiol. Dis., 31, 298– 308. 63. Tiscornia, G., Singer, O. and Verma, I.M. (2006) Design and cloning of lentiviral vectors expressing small interfering RNAs. Nat. Protoc., 1, 234–240. 64. Bolstad, B.M., Irizarry, R.A., Astrand, M. and Speed, T.P. (2003) A comparison of normalization methods for high density

65. 66. 67. 68.

69.

70.

3111

oligonucleotide array data based on variance and bias. Bioinformatics, 19, 185– 193. Irizarry, R.A., Bolstad, B.M., Collin, F., Cope, L.M., Hobbs, B. and Speed, T.P. (2003) Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res., 31, e15. Smyth, G.K. (2004) Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol. Biol., 3, Article 3. Luo, R.X., Postigo, A.A. and Dean, D.C. (1998) Rb interacts with histone deacetylase to repress transcription. Cell, 92, 463–473. Zhang, Y., Liu, T., Meyer, C.A., Eeckhoute, J., Johnson, D.S., Bernstein, B.E., Nusbaum, C., Myers, R.M., Brown, M., Li, W. and Liu, X.S. (2008) Model-based analysis of ChIP-Seq (MACS). Genome Biol. 9, R137. Bradford, J., Shin, J.Y., Roberts, M., Wang, C.E., Li, X.J. and Li, S. (2009) Expression of mutant huntingtin in mouse brain astrocytes causes age-dependent neurological symptoms. Proc. Natl Acad. Sci. USA, 106, 22480– 22485. Nguyen, T., Hamby, A. and Massa, S.M. (2005) Clioquinol down-regulates mutant huntingtin expression in vitro and mitigates pathology in a Huntington’s disease mouse model. Proc. Natl Acad. Sci. USA, 102, 11840– 11845.