IL-7) and IL-7 Splice Variants Affect Differentiation of Human Neural Progenitor Cells

ORIGINAL ARTICLE Interleukin-7 (IL-7) and IL-7 splice variants affect differentiation of human neural progenitor cells

M Moors1,4, NK Vudattu2,, J Abel1, U Kra¨mer1, L Rane2, N Ulfig3, S Ceccatelli4, V Seyfert-Margolies5, E Fritsche1 and MJ Maeurer2 1Group of Toxicology, Group of Epidemiology, Institut fu¨r Umweltmedizinische Forschung, Du¨sseldorf, Germany; 2Microbiology, Tumor and Cell Biology and Smittskyddsinstitutet, Stockholm, Sweden; 3Department of Anatomy, University of Rostock, Rostock, Germany; 4Division of Neurotoxicology, Department of Neuroscience, Karolinska Institutet, Sweden and 5Department of Medicine, University of California, San Francisco, CA, USA

Alternative splicing of pre-mRNA increases proteomic diversity, a crucial mechanism in defining tissue identity. We demonstrate differentially spliced interleukin (IL)-7 in distinct anatomic areas in the adult, in developing human brains and in normal human neuronal progenitor (NHNP) cells. IL-7c (c, the canonical form spanning all six exons) or its variants IL-7d5, d4 or d4/5 were cloned and expressed as recombinant proteins. IL-7 and splice variants were able to shift the differentiation of NHNP cells as compared with the diluent control (Po0.01) defined by anti-b (III)-tubulin and glial fibrillary acidic protein expression, with different degrees (IL-7c4d4/54IL-7d5); IL-7d4 exhibited a significantly weaker potency. Differentiation was confirmed by transcriptome analysis of IL-7c-stimulated neural NHNP cells, resulting in 58 differentially expressed genes; some of these are involved in neural differentiation, for example, the developmentally regulated transcription factor kru¨ppel-like factor 12, musashi 2, a translational regulator of cell fate or the sonic hedgehog receptor patch 1. This suggests that IL-7 influences neural development at a molecular level by participating in human brain architecture through glia cell formation: a paradigm that alternative splicing in cytokines, for example, for IL-7, has a physiological role in human organ development and progenitor cell differentiation. Genes and Immunity (2010) 11, 11–20; doi:10.1038/gene.2009.77; published online 22 October 2009

Keywords: interleukin-7; neuronal progenitor cells; stem cells; differentiation; differential splicing

Introduction This creates a soluble IL-7R, which is associated with an increased risk of developing multiple sclerosis.5–7 Cytokines have pleiotropic functions, which include The nominal IL-7R ligand, IL-7,8 is encoded by six targets in the central nervous system (CNS). They are exons (1–611 first, 612–718 second, 749–829 third, crucial factors in shaping neural plasticity, neuronal cell 830–961 fourth, 962–1015 fifth and 1016–2078 sixth) and differentiation and memory formation. For instance, forms a 177-amino-acid protein with a 25-amino-acid- impaired interleukin (IL)-1b expression in the hippo- long signal peptide. IL-7 has also been described to be campus is associated with spatial memory loss in mice,1 alternatively spliced.8–11 This is not driven by allelic interleukin (IL)-6 promotes sleep through trans-signaling variants of the human IL-7 gene, located on chromosome to neurons2 and IL-10, which exhibits predominantly 8q12–13. IL-7 splice variants have been identified in anti-inflammatory properties, provides direct trophic different conditions: malignant transformation,9 expo- support to neurons and may be useful for neuroprotec- sure to proinflammatory cytokines and tissue/organ- tive strategies.3 Other cytokines, including IL-1a, may specific expression.11 affect neurogenesis through transdifferentiation of adult Alternative splicing has not been studied for IL-7 in mesenchymal stem cells.4 Previous studies showed that the CNS, although alternative splicing of pre-mRNA is the IL-7 receptor (IL-7R) and IL-7 are involved in increasingly appreciated to increase proteomic diversity. neuronal development and pathophysiology. Three Up to 75% of the human genome is currently estimated independent studies recently showed that allelic differ- to be alternatively spliced.12 Alternative splicing ences of the IL-7R, located on chromosome 5p13.2, contributes to the evolution of protein diversity;13–15 lead to an alternative splicing of exon 6 of the IL-7R. both microRNAs and alternative pre-mRNA splicing are implicated in the development of the neuronal system. Neuron-specific microRNA (miR)-124 contri- Correspondence: Professor MJ Maeurer, Microbiology, Tumor and butes to progenitor cell differentiation and matures Cell Biology, Smittskyddsinstitutet, Nobels Va¨g 18, Stockholm neurons by regulating a complex network of alternative S-17182, Sweden. 16 E-mail: [email protected] splicing associated with neuronal development. Received 26 May 2009; revised 1 September 2009; accepted 3 Alternative pre-mRNA splicing has a pivotal role in the September 2009; published online 22 October 2009 establishment of neuronal identity,17 yet the regulatory IL-7 splice variants in neuronal differentiation M Moors et al 12

Figure 1 Interleukin (IL)-7 isoforms are expressed in adult brain tissue. A cDNA library from human brain, standardized for the expression for human b-actin and tested for mRNA integrity, was analyzed for IL-7 isoform expression. The canonical IL-7 isoform (IL-7c) spans six exons. Expression of different IL-7 mRNA isoforms is indicated with different colors (see figure legend, top right corner: skipped exons are in black, colors of the IL-7 isoforms correspond to IL-7 expression analysis from brain tissue). IL-7d2(À56 bp) represents a shifted reading frame resulting in the addition of seven new aa, the omitted 56 bp corresponds to nucleotides 476–531 of the IL-7c transcript. IL-7c mRNA is exclusively expressed in the cerebellum and pons; IL-7d4 is present in the caudate nucleus; the spinal cord shows IL-7c and IL-7d4/5 expression. The licence to use the brain image for publication was purchased from Fotosearch, LifeART, Waukesha, WI, USA, reference 20410757. The numbers in the boxes for tissues from the caudate nucleus and spinal cord indicate the ratio of differentially spliced IL-7 isoforms.

network that governs neuronal-specific splicing is poorly IL-7 affects neuronal cell differentiation understood. To investigate the impact of IL-7c and splice variants on We report here the analysis of IL-7 splice variants neural development, we used neurospheres prepared identified in the CNS. The six-exon-spanning IL-7 is from normal human neural progenitor (NHNP) cells as crucial for T-cell homeostasis,18,19 yet the biological an in vitro model system. NHNP neurospheres represent functions of IL-7 splice variants are poorly understood. heterogeneous, self-regulating cell populations of human IL-7, lacking exon 5 (IL-7d5), showed superior activity on neural precursor and neural stem cells defined by the CD4 þ and CD8 þ T cells compared with the ‘canonical’ expression of nestin and other early neural markers.22 IL-7.11 As IL-7 has been identified in developing brains RNA expression analysis showed that NHNP neuro- from rats20 and data from animal studies suggested spheres express IL-7c and the splice variants IL-7d4 and that IL-7 represents a neural growth and differentia- d3/4 (Figure 3a); they express the full-length IL-7R and tion factor in vitro,21 we examined adult and developing the IL-7R splice variants IL-7d6 and IL-7Rd5/6 human brains for IL-7 mRNA expression patterns, (Figure 3a). IL-7 protein was detected by staining with followed by cloning and expression of IL-7 splice an anti-IL-7-specific mAb (Figure 3b), yet this does not variants as recombinant proteins, which were tested allow the discrimination of IL-7 splice variants, as for their capacity to drive neural progenitor cell monoclonal and polyclonal reagents recognize several differentiation. IL-7 isoforms.11 Human neurospheres were differentiated for 4 days in the presence or absence of IL-7 or its splice variants (for experimental set up see Figure 4). Immuno- Results cytochemical staining of NHNP cells harvested at day 4 revealed a distribution of approximately 15–30% b-(III)- Mapping of IL-7 isoforms in adult and developing brains tubulin-positive neurons and 70–85% glial fibrillary A cDNA library from human brains showed an exclusive acidic protein (GFAP)-positive glia cells (Figure 5a). expression of IL-7c (c, the canonical form spanning all six Treatment with IL-7c or IL-7 splice variants resulted in a exons) in the cerebellum and pons; the IL-7 exon 4 splice significant increase in the number of developed GFAP- variant (IL-7d4) is exclusively expressed in the caudate positive cells in relation to the number of formed nucleus. Tissue from spinal cord showed expression of neurons (Figures 5a and b) compared with the diluent IL-7c and the IL-7 splice variant lacking exon 4/5 (d4/5) control. The potency of the different IL-7 isoforms in (Figure 1). Human embryonic brains express IL-7c and shifting differentiation toward the glia cell lineage varied splice variants that lack either exon 5 (d5), IL-7d4, IL-7d4/ (Figure 5b): IL-7c4IL-7d4/54IL-7d5 were the most 5 or IL-7 lacking exon 3/4 (d3/4) (Figure 2). Each of the potent isoforms, whereas IL-7d4 exhibited a significantly individual IL-7 splice variants was confirmed by weaker potency. A shift in differentiation was not sequence analysis, cloned and expressed as recombinant detected at earlier time points (that is, days 1–3, data proteins. not shown) defined by immunocytochemical staining.

Genes and Immunity IL-7 splice variants in neuronal differentiation M Moors et al 13

Figure 2 Interleukin (IL)-7 isoforms are expressed in embryonic human brain tissue. Brain tissues of different embryos (from natural miscarriage) were analyzed for a housekeeping gene (GAPDH), as well as for IL-7 Ra expression (cell membrane, M and soluble form, S) and IL-7 mRNA expression. Differential expression of IL-7c and IL-7 splice variants including IL-7d5, d4, d4/5 and d3/4 in embryonic brain tissue. ND, not determined. Colors correspond to the legend in Figure 1 designating different IL-7 isoforms. The days of gestation are indicated as described in the Materials and methods section.

Figure 3 Interleukin (IL)-7 isoforms and the IL-7 receptor are expressed in normal human neuronal progenitor (NHNP) neurospheres. Proliferating neurospheres were analyzed for IL-7 and IL-7 receptor variants. (a) Expression of IL-7c, the splice variants d4 and d3/4, the IL-7 receptor and IL-7 receptor variants lacking exon 6 and exons 5/6. (b) Cryostat sections of proliferating neurospheres stained for IL-7 (green) and nuclei (blue). Scale bar represents 100 mm.

Genes and Immunity IL-7 splice variants in neuronal differentiation M Moors et al 14

Figure 4 Experimental setup. Treatment of normal human neuronal progenitor (NHNP) cells with interleukin (IL)-7c and its isoforms. Neuronal human progenitor cells were cultured in neural progenitor maintenance medium as free-floating proliferating neurospheres at

37 1C with 5% CO2. Differentiation was initiated by plating onto poly-D-lysine/laminin-coated chamber slides under growth factor (GF) withdrawal. Five neurospheres were differentiated for each treatment group in Dulbecco’s modified Eagle’s medium/F12-N2 medium containing 1.25 pg mlÀ1 of the canonical IL-7 protein (IL-7c) or an alternatively spliced IL-7 isoform. IL-7-containing medium was freshly prepared and changed every second day. Controls received the IL-7 solvent phosphate-buffered saline (PBS). (a) Phase contrast image of a proliferating neurosphere (day 0). (b) Phase contrast image of a differentiating neurosphere. Single cells migrate out of the sphere in a radial manner. The distance between the outer limit of the sphere and the most distant (migrated) cells was measured after 2 days to quantify migration (indicated by white bars). (c) Cells are continuously cultured in the presence of IL-7c or its splice variants. Four days after initiation of exposure to IL-7, NHNP cells were fixed and stained with anti-b (III)-tubulin (green) and glial fibrillary acidic protein (red). Nuclei were counterstained with Hoechst 33342 (blue). For quantification, the number of neuronal- (green) and astrocyte (red)-like cells was determined by counting. A representative image of the cellular migration area around a neurosphere is depicted. A full colour version of this figure is available at the Genes and Immunity journal online.

Figure 5 Interleukin (IL)-7c and the IL-7d5 isoform mediate glia cell formation in human neural progenitor cells. (a) Normal human neuronal progenitors (NHNPs) were treated as in Figure 4. Fluorescent images show typical representatives of cells grown with diluent (control), IL-7c, IL-7d5 or in IL-7d4/5 (1.25 pg mgÀ1 for 4 days). Neurons stain positive for b-(III)-tubulin (green) and glia cells for glial fibrillary acidic protein (GFAP) (red) as described.22 (b) The mean increase in the number of GFAP-positive cells relative to the number of neurons as a percentage of four independent experiments. Five neurospheres were used for each treatment group (that is, IL-7c or IL-7 isoforms) in every experiment. The number of neurons and glia cells in three different fields per sphere was determined. Data are shown as percentage increase compared with the diluent control; mean±s.e.m. is provided. A paired t-test was performed and the significance of the result is indicated (*Po0.05; **Po 0.01). (c) Quantification of NHNP cell migration. Neurospheres were incubated with diluent (control), IL-7c or IL-7 isoforms. After 2 days of exposure, NHNP cell migration was measured. Data are expressed as mean±s.e.m. of four independent experiments. One- way analysis of variance in combination with a Bonferroni post hoc test (a ¼ 0.05) was used to test for differences between the experiments. No significant differences were detected. Ctrl, control (medium), c, ‘canonical’, six-exon-spanning IL-7.

As differentiating NHNP cells migrate out of the sphere exposed to IL-7 (48 h) and the migration distance of structure, we performed migration studies in IL-7- differentiating cells was measured. We did not detect stimulated neurospheres to explore whether the shift in significant alterations of cellular migration by IL-7 or its neural differentiation is linked to a change in NHNP cell isoforms (Figure 5c). A standard STAT5 phosphorylation migration. We therefore chose a time point before we assay11 in NHNP cells showed a fast response in IL-7c could detect changes defined by immunohistology, that and IL-7d5 (data not shown) and confirmed that IL-7 is, before day 4 after IL-7 culture. Neurospheres were shows biological activity.

Genes and Immunity IL-7 splice variants in neuronal differentiation M Moors et al 15 IL-7 effects on gene expression in neural cells gamma (gc) chain,31 yet we can not exclude that they To identify molecular targets of IL-7, whole-transcrip- bind to alternate receptors. IL-7 may also exhibit trans- tome expression analysis of (IL-7-stimulated) NHNP signaling, similar to IL-6.2 Not mutually exclusive, IL-7 neurospheres was performed. To detect early changes in and its isoform may also use different signaling path- gene expression that correlated with changes in neural ways in neurons, that is, the phosphoinositol 3-kinase cell differentiation, NHNP cells were stimulated for 24 h and the src family tyrosine kinase p56lck and p59fyn as with a physiological concentration of IL-7, which showed suggested earlier.32 We did not perform IL-7R-blocking previous responses in human T lymphocytes.11 A total of experiments because of the limited availability of NHNP 58 genes were identified to be significantly differentially cells. RNA coding for cell-membrane-associated and expressed in IL-7c-stimulated NHNP neurospheres in soluble IL-7R is expressed in human developing brains control versus cells grown in the diluent control (Figure 6, (Figure 2), as well as in NHNP cells (Figure 3). The Table 1). Expression of 50 out of 58 genes were down- expression of IL-7R seems not to be restricted to neural regulated, whereas 8 out of 58 genes were upregulated. progenitor cells, as Nunnari et al.32 showed the IL-7R Functional cluster analysis revealed that most of the expression in matured human neurons and astrocytes. affected genes are involved in transcription regulation IL-7R engagement leads to the activation of Janus kinases (40% of clustered genes) and protein metabolism (34%), (JAK), which selectively activate STATs.31,33 Previous and 17% of the genes are related to neural differentiation experiments showed that JAK/STAT triggering during (Table 1 and Supplementary Table S1). neural development leads to astrogliogenesis, which is instrumental in injury-induced reactive gliosis34 with inhibition of astrocyte proliferation.35,36 It could very well be that IL-7 or IL-7 variants, Discussion elaborated in situ in CNS inflammation, contribute to IL-7 splice variants in neural differentiation reactive gliosis. The potential effects of IL-7 or IL-7 Alternative splicing has a central role in the enhance- isoforms in the adult brain are a matter of future ment of protein diversity in a development/tissue- investigation. Up to now, exogenous IL-7 has been specific manner and in the development of cellular shown to induce upregulation of proapoptotic genes phenotypes.23 It is currently estimated that 10–30% of associated with apoptosis in matured human neurons.32 alternatively spliced genes have tissue-specific iso- We could not identify expression changes in apoptosis- forms24 and 73% of all genes show differences in related genes in neuronal progenitor cells. IL-7 may expression levels between tissues, on the basis of therefore exert specific effects depending on the matura- microarray expression analysis.25 Alternative splicing in tion stage of neural cells. Not mutually exclusive, cytokine genes and in their nominal receptors has been induction of apoptosis is also required for controlled described. For instance, alternative splicing of the IL-23 tissue and organ differentiation. receptor a chain most likely represents the cytokine receptor with the highest number of potential transcripts; IL-7 splice variants modulate neuronal system-associated 18 individual proteins are predicted.26 Alternatively gene expression spliced IL-4, lacking exon 2, serves as a natural Transcriptome analysis supported the immunohistologi- antagonist for IL-4.27 IL-15 has two isoforms, one form cal finding that IL-7 drives human neural progenitor cell is secreted and acts through the cell-bound IL-15 differentiation. Experiments were performed at a single receptor, the other isoform represents an intracellular time point because of limited access to a sufficient IL-15 protein,28 which acts as an ‘intracrine’ signaling number of human neuronal progenitor cells. It may very factor in mast cells to limit the recruitment of neutrophils well be that other time points may show disparate gene to sites of infection,29 a paradigm that cytokine isoforms expression levels, a situation that is true for complex may show differential functions in different cellular changes in cellular differentiation and organization. IL-7 compartments. stimulation of NHNPs resulted in the differential The alternatively spliced IL-7R is associated with expression of 58 genes (Figure 5, Table 1 and Supple- increased risk to develop multiple sclerosis identified mentary Table S1). One gene cluster, regulated by IL-7, is through a genome-wide single-nucleotide polymorph- linked to neuronal development. Musashi 2 (MSI2) has ism search.5–7 IL-7 did not show allelic differences, an essential role in maintaining the undifferentiated state alternative splicing of IL-7 cannot be predicted on the of neural stem cells;37 it is repressed in IL-7-stimulated basis of genomic analysis.30 RNA expression profiling is cells. The same was found to be true for the mitogen- therefore necessary to show tissue and development- activated protein kinase 1 (MAPK1), which has been associated differences11 in IL-7 splicing. IL-7 mRNA shown to be linked to neurogenesis in vivo. Conditional mapping shows that IL-7 isoforms exhibit a unique inactivation of the murine mapk1 gene resulted in the expression profile in human adult brains. The fact that generation of fewer neurons and more astrocytes in the pons and cerebellum exhibit similar IL-7 splicing developing cortex.38 IL-7-mediated MAPK1 downregula- patterns may stem from the fact that they are a part of tion may therefore support the observation that IL-7 the embryological hindbrain. Some of these IL-7 splice resulted in an increase in GFAP-positive cells (Figures 5a variants (IL-7d4 and d3/4) are also present in human and b). Repression of the jumonji domain containing 2C neuronal stem cells (Figure 3). This pattern is similar to (JMJD2C) is also involved in stem cell differentiation. the IL-7 isoform expression pattern in tumor cells of JMJD2C acts as a positive regulator for the transcription neuronal origin.11 factor Nanog, which is essential for stem cell self- It is most likely that the effects of IL-7 splice variants renewal. JMJD2C depletion in embryonic stem cells has are mediated through the cell-associated heterodimeric been shown to lead to cellular differentiation.39 Reticu- IL-7R composed of IL-7Ra (CD127) and the common localbin 2 (RCN2) is a calcium-binding protein and aids

Genes and Immunity IL-7 splice variants in neuronal differentiation M Moors et al 16

Figure 6 Interleukin (IL)-7 mediates changes in gene transcription in neural cells. Normal human neuronal progenitors neurospheres were stimulated with diluent (phosphate-buffered saline) or IL-7c (1.25 pg mlÀ1) for 24 h. Gene expression analysis was performed with the human HG U133 2.0 GeneChip array from Affymetrix. Genes exhibiting a 41.7-fold differential expression in control versus IL-7c-stimulated spheres, testing significantly different in the t-test and significance analysis of microarrays, were scored (Po0.05). The heat map shows data from control (diluent) and IL-7-stimulated cells (three samples per group). Gene expression levels are color coded, red indicates high expression and green represents a low gene expression. in synaptic refinement. Dysregulation of RCN2 was A different gene cluster, (down)regulated by IL-7, is observed in idiopathic absence epilepsy.40 The Kru¨ ppel- associated with lipid synthesis. Farnesyl diphosphate like factor 12 (KLF12) and patch 1 (PTCH1) are also farnesyl transferase 1 (FDFT1) and HMG-CoA synthase 1 involved in neuronal regulatory and developmental (HMGCS1) have an important role in cholesterol processes.41 biosynthesis, and downregulation of these genes has

Genes and Immunity IL-7 splice variants in neuronal differentiation M Moors et al 17 Table 1 IL-7-induced changes in gene expression

Cluster Gene symbol Direction Log2-fold P-value RefSeq ID a ProbeSetID qRT-PCR regulation by IL-7 (P-value)

Transcription ZMYM2 À 3 580939969 0043 NM_197968 210282_at REL À 2 085587248 0020 NM_002908 206035_at NCOR1 À 2 041050201 0038 NM_006311 200854_at SEPSECS À 1 923882469 0038 NM_016955 235516_at KLF12 À 1 760305708 0016 NM_007249 206965_at 0.1 RFX1 À 1 738069774 0033 NM_002918 226786_at NME7 À 2 157049574 0016 NM_197972 227556_at 0.08 ZKSCAN1 + 2 072735461 0026 NM_003439 1557953_at LOC729433 /// MLLT1 À 2 117113575 0010 NM_005934 224993_at TTC14 À 2 114140611 0046 NM_133462 225178_at JMJD2C À 1 884345267 0048 NM_015061 214861_at 0.044 TADA2 L + 1 693315341 0036 NM_133439 209938_at SOS2 À 1 83166119 0014 NM_178169 212870_at 0.12 TTC5 + 1 74308026 0028 NM_138376 226557_at Neural differentiation MAPK1 À 2 386600352 0017 NM_138957 1552263_at MSI2 À 2 103956194 0011 NM_138962 239232_at 0.093 PTCH1 À 1 857376989 0018 NM_001083602 209816_at PAK2 À 1 713319646 0042 NM_002577 208876_s_at PROS1 À 1 79845991 0036 NM_000313 207808_s_at 0.034 RICTOR À 2 211736218 0012 NM_152756 228248_at Signaling HRAS + 1 698493464 0013 NM_005343 212983_at Cytosceleton CENTB2 À 2 061773712 0044 NM_012287 1552472_a_at Cell cycle E2F1 + 1 775979998 0039 NM_005225 204947_at Protein metabolism AGA À 2 052709304 0003 NM_000027 204332_s_at FDFT1 À 1 812809927 0048 NM_004462 208647_at 0.033 HMGCS1 À 1 724841346 0005 NM_002130 221750_at 0.05 CEPT1 À 1 851932694 0002 NM_006090 219375_at PIGL À 1 923559781 0022 NM_004278 232262_at TCP1 À 2 200095274 0036 NM_030752 222011_s_at 0.09 MDN1 À 2 077203652 0043 NM_014611 1569484_s_at SEC63 + 1 813133307 0015 NM_007214 201914_s_at NPEPL1 À 2 003461797 0018 NM_024663 235033_at GOLT1B À 1 992493475 0045 NM_016072 218193_s_at AP3M1 À 2 216158909 0027 NM_207012 222516_at MRPL52 À 2 64895622 0010 NM_181307 221997_s_at Not clustered RCN2 À 1 94241078 0038 NM_002902 201485_s_at 0.06 ATP7A À 1 858764934 0014 NM_000052 205197_s_at SLC39A6 + 1 958242365 0035 NM_012319 202089_s_at LOC388969 + 2 02247565 0041 NM_001013649 238768_at

Abbreviations: IL, interleukin; qRT-PCR, quantitative real-time-PCR. IL-7-mediated transcriptional changes in neural cells. Differentially expressed genes in control (diluent) versus IL-7-stimulated cells are grouped in functional clusters. Values represent log2-fold regulation by IL-7 (P-values are shown). Gene designations provided in bold and italics were used for validation by qRT-PCR (s, Po0.05; (s), Po0.1). aReference sequence number corresponding to the Affymetrix ProbeSetID. been shown to be associated with neuroprotection.42 sion is associated with CNS injuries, particularly in the NME7 (nm23-H7), a nucleoside diphosphate kinase, dentate gyrus of the hippocampus.46 provides NTPs for nucleic acid synthesis and adenosine In summary, our data suggest that IL-7 may participate triphosphate generation. Nm23-H4, H6 and H7 have in human brain development/plasticity. The observation been shown to be overexpressed in human primary colon that IL-7 is able to shift differentiation in human neural and gastric carcinomas;43,44 their role in neuronal stem progenitor cells represents a paradigm that cytokines cells is yet to be elucidated. nmH23-H1 and -H2 levels may not only have a physiological role in immune are decreased in myeloid and lymphoid differentiation of function, but also in neuronal tissue development and hematopoietic progenitor cells, which may reflect NME7 plasticity. repression in NHNP cells after IL-7 stimulation. Other IL-7-regulated genes that may have a role in CNS-associated diseases are the T-complex polypeptide 1 (TCP-1) and son-of-sevenless 2 (SOS2). TCP-1 is Materials and methods involved in tubulin biogenesis (b-1 tubulin). The ratio Cell culture of TCP-1 and b-1 tubulin is decreased in Alzheimer’s Normal human neural progenitor cells (from a disease and significantly increased in brain tissues from male fetus, between the sixteenth and nineteenth patients with Down’s syndrome.45 SOS2 codes for an week of gestation) were purchased from Cambrex, ubiquitously expressed guanine nucleotide exchange Walkersville, MD, USA and cultured in DFB medium factor crucial for signal transmission. SOS2 overexpres- (Dulbecco’s modified Eagle’s medium and Hams F12

Genes and Immunity IL-7 splice variants in neuronal differentiation M Moors et al 18 (3:1) supplemented with B27 (Invitrogen GmBH, using a primer set covering the entire eight exons: F50- Karlsruhe, Germany), 20 ng mlÀ1 EGF (Biosource, GAATGACAATTCTAGGTACAAC, B50-GTTTTGGTAG Karlsruhe, Germany) and 20 ng mlÀ1 rhFGF (R&D Sys- AAGCTGGAC, followed by a nested PCR using primers tems, Wiesbaden-Nordenstadt, Germany) in suspension that cover exons 4–8: F50-TAATGCACGATGTAGCTTA 0 culture as neurospheres at 37 1C with 5% CO2. CCGC, B5 -CACCCTATGAATCTGGCAGTC. PCR pro- Differentiation of NHNP cells was initiated by plating ducts were run on a gel, bands were extracted and onto poly-D-lysin/laminin-coated chamber slides (BD cloned into the pCR 2.1 TOPO vector (Invitrogen) and Biosciences, Eremnodegem, Belgium) under growth fac- subjected to a sequencing reaction using a Big Dye tor withdrawal in DFN medium (Dulbecco’s modified Terminator V.3.1 Cycle sequencing kit (Applied Biosys- Eagle’s medium and Hams F12 (3:1) supplemented with tems, Foster City, CA, USA); sequences were analyzed N2 (Invitrogen GmBH). For IL-7 stimulation, NHNP cells and confirmed using Chromas Lite 2.01 software were plated for differentiation in DFN medium, without (Technelysium, Tewantin, Australia). growth factors, containing 1.25 pg mlÀ1 of IL-7c or its splice variants. IL-7-containing medium was freshly prepared and changed every 2 days. Controls received Immunocytochemistry the IL-7 solvent phosphate-buffered saline. Normal human neuronal progenitor cells were fixed in 4% paraformaldehyde for 30 min, followed by costaining Recombinant IL-7 protein expression with the neuronal marker anti-b (III)-tubulin (1:100) and Interleukin-7c or its variants IL-7d5, d4ord4/5 were the glia marker anti-GFAP (1:1000, both from Sigma- cloned and expressed using the pIZ/V5-His vector in Aldrich, St Louis, MO, USA) as described.22 Nuclei were HighFive insect cells (Invitrogen, Carlsbad, CA, USA) as stained with Hoechst 33342 (0.1 mgmlÀ1, Sigma-Aldrich). described.11 The integrity of proteins was evaluated Alterations in differentiation were determined by count- using monoclonal or polyclonal mAbs directed against ing the number of b-(III) tubulin and GFAP-positive cells IL-7 (mAb clone B-N18, murine IgG1, Diaclone, Besan- in three different fields per sphere with Soft Imaging con, France, pAb: rabbit anti-IL-7 Lot no. 276–01B, System program (Muenster, Germany) analySIS. For each Biosource, Keystone, CO, USA); only IL-7c, or IL-7d4 treatment group, variance analysis was performed to test and d5, but not d4/5 was recognized by both reagents. whether differentiation in spheres (n ¼ 5) differed between experiments (n ¼ 4) (factor sphere nested in factor IL-7 and IL-7R expression analysis experiment). As no significant differences could be A cDNA library from human brain, standardized for the detected, the mean was calculated from all values expression for human b-actin and tested for mRNA combined (n ¼ 20 spheres) for each treatment group. T- integrity (HumanBrain Rapid-SCAN), was obtained tests were performed for each group to test whether the from OriGene Technology (Rockville, MD, USA); spinal percentage increase in GFAP þ cells was significantly cord tissue was derived from the cervical and thorax different from zero. Values were transformed using area. Tissues from human embryonic brains were arcsin transformation before analysis to ensure variance obtained from the Department of Anatomy, University homogeneity. of Rostock, Germany, from spontaneous abortions (and IL-7 protein staining: Proliferating spheres were fixed reviewed from the local ethics committee). The time in 4% paraformaldehyde for 30 min as described.22 After length of gestation was calculated (time point of the last washing in phosphate-buffered saline, spheres were period and spontaneous abortion) and indicated in incubated overnight in a 10% sucrose solution (w/v) at Figure 2. mRNA from brain tissues was isolated and 4 1C. The neurospheres were then transferred to a cDNA prepared. IL-7 isoform expression was analyzed freezing medium (Tissue Tek, Sakura Finetek Europe as described in detail earlier, using primers that span the B.V., Histolab Products AB, Go¨teborg, Sweden). Cryostat entire six exons (F50-TCCCGCAGACCATGTTCCATG, sections (10 mm) were stained with an anti-IL7-mAb B50-TCAGTGTTCTTTAGTGCC, and a nested primer set (clone BVD10–11C10, Nordic Biosite, Stockholm, F50: GCAGACCATGTTCCATGTTTC, B50: CAGTGTTC Sweden) diluted 1:25 for 1 h at room temperature. A TTTAGTGCCCATCA). A run-off reaction with Fam- secondary antibody, coupled with Alexa 488 (1:500; labeled primers was performed accordingly (F50 Invitrogen), was used to visualize anti-IL-7 mAb bind- GTTGCCAGTAGCATCATCTG, F50-GATGGCAAACAA ing. Nuclei were stained with Hoechst 33342 (0.1 mgmlÀ1, TATGAGAG, B50: CTGTTCCTTTAAA-GATTTAT). The Sigma-Aldrich). Slides were mounted in phenylendia- length of the individual mRNA transcripts was mea- mine (Sigma-Aldrich). Pictures were taken with a sured using an ABI-310 sequencer (GMI Inc, Ramsey, Hamamatsu digital camera (Solna, Sweden, C4742–95). MN, USA) using appropriate length standards. IL-7 mRNA transcripts were confirmed by DNA sequence analysis, the nucleotide sequence in the splice variants Migration analysis was identical to the original IL-7 sequence. IL-7R Neurospheres were stimulated with phosphate-buffered expression was analyzed accordingly using the primers saline (control), IL-7c or IL-7 splice variants in DFN on 0 0 F5 -CTCCAGAGATCAATAATAGCTC/B5 -TTGTCGCT poly-D-lysin/laminin-coated chamber slides (BD Bios- CACGGTAAGTTCA. The soluble and membrane-bound ciences) for 48 h. The distance between the outer limit of IL-7R forms were tested by mRNA fragment analysis the sphere and the migrated cells was measured using the Fam-labeled primer 50-TGCCACTCTCCCT microscopically at four different positions per sphere. GCAGT. IL-7 and IL-7R in human neuronal pro- One-way analysis of variance in combination with a genitor cells were characterized using a nested primer Bonferroni post hoc test for a-correction (a ¼ 0.05) was set for IL-7 F50-GCAGACCATGTTCCATGTTTC, B50- used to test for significant differences between treatment CAGTGTTCTTTAGTGCCCATCA. IL-7R was analyzed groups (factor treatment group).

Genes and Immunity IL-7 splice variants in neuronal differentiation M Moors et al 19 Gene expression analysis CACACATTGCTATTCCCATCC, RP: CTTTTCTCCCC Normal human neuronal progenitor cells were stimu- CGTTAGGTC); b-actin (LP: CCCCAGGCACCAGGG lated with phosphate-buffered saline (control) or IL-7c in CGTGAT, RP: GGTCATCTTCTCGCGGTTGGCCTTGG DFN medium, in the absence of growth factors, on poly- GGT). Product accumulation was measured by real-time D-lysin/laminin-coated 24-well plates for 24 h, followed PCR analysis on the basis of SYBR Green detection with by RNA isolation with the RNeasy Mini Kit (VWR, ABI Prism 7500 (Applied Biosystems Inc, Foster City, CA, Stockholm, Sweden) according to the manufacturer’s USA). Expression levels were normalized to the expres- instructions. Integrity of extracted RNA was confirmed sion of b-actin. (PCR reaction: 1 ml cDNA, 4 mM Primers by NanoDrop (Thermo Scientific, Wilmington, DE, USA). each, SybrGreen PCR Master Mix (Applied Biosystems). For Affymetrix MicroArray analysis, samples were The amplification protocol was as follows: 10 min 95 1C processed at the Karolinska Biomic Center Genomic (AmpliTaq Gold Activiation), 15 s 95 1C, 1 min 60 1C Facility, Stockholm, Sweden, according to the Affymetrix (40 cycles). GeneChip Expression Analysis Manual (Affymetrix, Santa Clara, UK). Double-stranded cDNA was synthesized with 50 ng of total RNA using the SuperScript Conflict of interest Choice system (Invitrogen). A T7-(dT24) oligomer was used for priming the first-strand cDNA synthesis. Using The authors declare no conflict of interest. cDNA as the template, cRNA was synthesized using the MEGAscript High Yield Transcription Kit (Applied Biosystems). cRNA was purified using a sample clean Acknowledgements up Kit (Affymetrix). Double-stranded cDNA was synthesized using the SuperScript Choice system (Invitrogen). This work was supported from Karolinska Institutet and Random primers were used for priming the first-strand by a grant from Cancerfonden, Sweden to MM. We thank cDNA synthesis. cDNA was purified using a sample Dr Markus Frericks for help with gene expression clean up Kit (Affymetrix). Using cDNA as the template, analysis and Ali Moshfegh, Karolinska University cRNA was synthesized using an In Vitro Transcription Hospital (Affymetrix core facility lab). We are indebted (IVT) kit (Affymetrix). IVT reactions were carried out at to Hanni Ho¨hn and Deshun Pan, Department of Med. 37 1C for 16 h. The labeled cRNA was purified using the Microbiology, University of Mainz, Germany for pre- Affymetrix sample clean up Kit. cRNA was fragmented paration of recombinant IL-7 proteins and to Michel in a fragmentation buffer (40 mmol lÀ1 Tris-acetate, pH Morre, Cytheris, Paris, France. 8.1, 100 mmol lÀ1 KOAc, 30 mmol lÀ1 MgOAc) for 35 min at 94 1C. Fragmented cRNA (15 mg per probe array) was hybridized with the human HG U133 2.0 GeneChip array References at 45 1C for 18 h in a hybridization oven with constant rotation (60 r.p.m.). Chips were washed and stained 1 Labrousse VF, Costes L, Aubert A, Darnaudery M, Ferreira G, using Affymetrix fluidics stations. Staining was per- Amedee T et al. Impaired interleukin-1beta and c-Fos expres- formed using streptavidin phycoerythrin conjugate sion in the hippocampus is associated with a spatial memory (Molecular Probes, Eugene, OR, USA), followed by the deficit in P2X(7) receptor-deficient mice. PLoS One 2009; 4: addition of a biotinylated anti-strepavidin antibody e6006. (Vector Laboratories, Burlingame, CA, USA), and finally 2 Benedict C, Scheller J, Rose-John S, Born J, Marshall L. with streptavidin phycoerythrin conjugate. Probe arrays Enhancing influence of intranasal interleukin-6 on slow-wave activity and memory consolidation during sleep. FASEB J were scanned using fluorometric scanners (Affymetrix). 2009; 23: 3629–3636. Raw excel files were processed with RMA-express 3 Zhou Z, Peng X, Insolera R, Fink DJ, Mata M. Interleukin-10 (Bolstad). Statistical data analysis was performed with provides direct trophic support to neurons. J Neurochem 2009; the TIGR-Multi-Experiment-Viewer (MEV) version 4.2. 110: 1617–1627. Only genes showing 41.7-fold differential expression 4 Greco SJ, Rameshwar P. Enhancing effect of IL-1alpha on and testing significantly different in the t-test and neurogenesis from adult human mesenchymal stem cells: significance analysis of microarrays were scored. implication for inflammatory mediators in regenerative medicine. J Immunol 2007; 179: 3342–3350. Quantitative real-time PCR 5 Hafler DA, Compston A, Sawcer S, Lander ES, Daly MJ, De Jager PL et al. Risk alleles for multiple sclerosis identified by a cDNA was prepared using 25 mM of Oligo-dT primer and genomewide study. N Engl J Med 2007; 357: 851–862. 25 ng of random primers according to the oligo-dT 6 Lundmark F, Duvefelt K, Iacobaeus E, Kockum I, Wallstrom E, protocol from the Superscript III first-strand cDNA Khademi M et al. Variation in interleukin 7 receptor alpha synthesis kit (Invitrogen). cDNA was stored at À20 1C chain (IL7R) influences risk of multiple sclerosis. Nat Genet until use. For quantitative real-time-PCR reactions, the 2007; 39: 1108–1113. following primer sequences were used: JMJD2C (LP: 7 Gregory SG, Schmidt S, Seth P, Oksenberg JR, Hart J, Prokop A ATGCAGGTGGAGCAGAATTT, RP: TCGGATGGGTCT et al. Interleukin 7 receptor alpha chain (IL7R) shows allelic GATTTCTC); FDFT1 (LP: GGTCCCGCTGTTACACAA and functional association with multiple sclerosis. Nat Genet CT, RP: AAAACTCTGCCATCCCAATG); HMGCS1 (LP: 2007; 39: 1083–1091. TTGGCTGCCACTCTGTACTC, RP: TGGGTGTCCTCT 8 Goodwin RG, Lupton S, Schmierer A, Hjerrild KJ, Jerzy R, Clevenger W et al. Human interleukin 7: molecular cloning CTGAGCTT); RCN2 (LP: CCCCAAGAGCTGTTAC and growth factor activity on human and murine B-lineage CTTG, RP: CATGGAGCTGTCTGCCATAA); NME7 (LP: cells. Proc Natl Acad Sci USA 1989; 86: 302–306. GAAAAGACTGCTGGGACCTG, RP: TGACAGCATG 9 Korte A, Moricke A, Beyermann B, Kochling J, Taube T, GGGTTTAACA); SOS2 (LP: CCAAGCACTCCTCCTAG Kebelmann-Betzing C et al. Extensive alternative splicing of CAC, RP: TGGGGAGTTTCTGCATTTTC); TCP1b (LP: interleukin-7 in malignant hematopoietic cells: implication of

Genes and Immunity IL-7 splice variants in neuronal differentiation M Moors et al 20 distinct isoforms in modulating IL-7 activity. J Interferon 29 Orinska Z, Maurer M, Mirghomizadeh F, Bulanova E, Metz M, Cytokine Res 1999; 19: 495–503. Nashkevich N et al. IL-15 constrains mast cell-dependent 10 Maeurer MJ, Walter W, Martin D, Zitvogel L, Elder E, Storkus antibacterial defenses by suppressing chymase activities. Nat W et al. Interleukin-7 (IL-7) in colorectal cancer: IL-7 is Med 2007; 13: 927–934. produced by tissues from colorectal cancer and promotes 30 Lundmark F, Duvefelt K, Hillert J. Genetic association analysis preferential expansion of tumour infiltrating lymphocytes. of the interleukin 7 gene (IL7) in multiple sclerosis. Scand J Immunol 1997; 45: 182–192. J Neuroimmunol 2007; 192: 171–173. 11 Vudattu NK, Magalhaes I, Hoehn H, Pan D, Maeurer MJ. 31 Pernis A, Gupta S, Yopp J, Garfein E, Kashleva H, Schindler C Expression analysis and functional activity of interleukin-7 et al. Gamma chain-associated cytokine receptors signal splice variants. Genes Immun 2009; 10: 132–140. through distinct transducing factors. J Biol Chem 1995; 270: 12 Xing Y, Lee C. Relating alternative splicing to proteome 14517–14522. complexity and genome evolution. Adv Exp Med Biol 2007; 623: 32 Nunnari G, Xu Y, Acheampong EA, Fang J, Daniel R, Zhang C 36–49. et al. Exogeneous IL-7 induces Fas-mediated human neuronal 13 Irimia M, Rukov JL, Roy SW, Vinther J, Garcia-Fernandez J. apoptosis: potential effects during human immunodeficiency Quantitative regulation of alternative splicing in evolution virus type 1 infection. J Neurovirol 2005; 11: 319–328. and development. Bioessays 2009; 31: 40–50. 33 De-Fraja C, Conti L, Magrassi L, Govoni S, Cattaneo E. 14 Xing Y. Genomic analysis of RNA alternative splicing in Members of the JAK/STAT proteins are expressed and cancers. Front Biosci 2007; 12: 4034–4041. regulated during development in the mammalian forebrain. 15 Garcia-Blanco MA, Baraniak AP, Lasda EL. Alternative J Neurosci Res 1998; 54: 320–330. splicing in disease and therapy. Nat Biotechnol 2004; 22: 34 He F, Ge W, Martinowich K, Becker-Catania S, Coskun V, Zhu 535–546. W et al. A positive autoregulatory loop of Jak-STAT signaling 16 Makeyev EV, Zhang J, Carrasco MA, Maniatis T. The controls the onset of astrogliogenesis. Nat Neurosci 2005; 8: MicroRNA miR-124 promotes neuronal differentiation by 616–625. triggering brain-specific alternative pre-mRNA splicing. Mol 35 Wishcamper CA, Brooks DM, Douglas Coffin J, Lurie DI. Focal Cell 2007; 27: 435–448. cerebral ischemia upregulates SHP-1 in reactive astrocytes in 17 Smirnova L, Grafe A, Seiler A, Schumacher S, Nitsch R, juvenile mice. Brain Res 2003; 974: 88–98. Wulczyn FG. Regulation of miRNA expression during neural 36 Wishcamper CA, Coffin JD, Lurie DI. Lack of the protein cell specification. Eur J Neurosci 2005; 21: 1469–1477. tyrosine phosphatase SHP-1 results in decreased numbers of 18 Mackall CL, Fry TJ, Bare C, Morgan P, Galbraith A, Gress RE. glia within the motheaten (me/me) mouse brain. J Comp IL-7 increases both thymic-dependent and thymic-indepen- Neurol 2001; 441: 118–133. dent T-cell regeneration after bone marrow transplantation. 37 Okano H, Imai T, Okabe M. Musashi: a translational regulator Blood 2001; 97: 1491–1497. of cell fate. J Cell Sci 2002; 115 (Pt 7): 1355–1359. 19 Schluns KS, Kieper WC, Jameson SC, Lefrancois L. Inter- 38 Samuels IS, Karlo JC, Faruzzi AN, Pickering K, Herrup K, leukin-7 mediates the homeostasis of naive and memory CD8 Sweatt JD et al. Deletion of ERK2 mitogen-activated protein T cells in vivo. Nat Immunol 2000; 1: 426–432. kinase identifies its key roles in cortical neurogenesis and 20 Michaelson MD, Mehler MF, Xu H, Gross RE, Kessler JA. cognitive function. J Neurosci 2008; 28: 6983–6995. Interleukin-7 is trophic for embryonic neurons and is 39 Loh YH, Zhang W, Chen X, George J, Ng HH. Jmjd1a and expressed in developing brain. Dev Biol 1996; 179: 251–263. Jmjd2c histone H3 Lys 9 demethylases regulate self-renewal in 21 Araujo DM, Cotman CW. Trophic effects of interleukin-4, -7 embryonic stem cells. Genes Dev 2007; 21: 2545–2557. and -8 on hippocampal neuronal cultures: potential involve- 40 Helbig I, Matigian NA, Vadlamudi L, Lawrence KM, Bayly ment of glial-derived factors. Brain Res 1993; 600: 49–55. MA, Bain SM et al. Gene expression analysis in absence 22 Moors M, Cline JE, Abel J, Fritsche E. ERK-dependent and epilepsy using a monozygotic twin design. Epilepsia 2008; 49: -independent pathways trigger human neural progenitor cell 1546–1554. migration. Toxicol Appl Pharmacol 2007; 221: 57–67. 41 Wang Y, Ng EL, Tang BL. Rab23: what exactly does it traffic? 23 Rosenfeld MG, Mermod JJ, Amara SG, Swanson LW, Traffic 2006; 7: 746–750. Sawchenko PE, Rivier J et al. Production of a novel neuropep- 42 Yoshida E, Atkinson TG, Chakravarthy B. Neuroprotective tide encoded by the calcitonin gene via tissue-specific RNA gene expression profiles in ischemic cortical cultures pre- processing. Nature 1983; 304: 129–135. conditioned with IGF-1 or bFGF. Brain Res Mol Brain Res 2004; 24 Mortazavi A, Williams BA, McCue K, Schaeffer L, Wold B. 131: 33–50. Mapping and quantifying mammalian transcriptomes by 43 Lacombe ML, Milon L, Munier A, Mehus JG, Lambeth DO. RNA-Seq. Nat Methods 2008; 5: 621–628. The human Nm23/nucleoside diphosphate kinases. J Bioenerg 25 Xu Q, Modrek B, Lee C. Genome-wide detection of tissue- Biomembr 2000; 32: 247–258. specific alternative splicing in the human transcriptome. 44 Seifert M, Welter C, Mehraein Y, Seitz G. Expression of the Nucleic Acids Res 2002; 30: 3754–3766. nm23 homologues nm23-H4, nm23-H6, and nm23-H7 in 26 Kan SH, Mancini G, Gallagher G. Identification and char- human gastric and colon cancer. J Pathol 2005; 205: acterization of multiple splice forms of the human interleukin- 623–632. 23 receptor alpha chain in mitogen-activated leukocytes. Genes 45 Schuller E, Gulesserian T, Seidl R, Cairns N, Lube G. Brain Immun 2008; 9: 631–639. t-complex polypeptide 1 (TCP-1) related to its natural 27 Atamas SP, Choi J, Yurovsky VV, White B. An alternative substrate beta1 tubulin is decreased in Alzheimer’s disease. splice variant of human IL-4, IL-4 delta 2, inhibits IL-4- Life Sci 2001; 69: 263–270. stimulated T cell proliferation. J Immunol 1996; 156: 435–441. 46 Marciano PG, Brettschneider J, Manduchi E, Davis JE, East- 28 Tan X, Lefrancois L. Novel IL-15 isoforms generated by man S, Raghupathi R et al. Neuron-specific mRNA complexity alternative splicing are expressed in the intestinal epithelium. responses during hippocampal apoptosis after traumatic brain Genes Immun 2006; 7: 407–416. injury. J Neurosci 2004; 24: 2866–2876.

Supplementary Information accompanies the paper on Genes and Immunity website (http://www.nature.com/gene)

Genes and Immunity