[alpha]-Chemokines Regulate Proliferation ... - Wiley Online Library

EMBRYONIC STEM CELLS ␣-Chemokines Regulate Proliferation, Neurogenesis, and Dopaminergic Differentiation of Ventral Midbrain Precursors and Neurospheres LINDA C. EDMAN,a HELENA MIRA,a ALEJANDRO ERICES,a,b SETH MALMERSJO¨ ,a EMMA ANDERSSON,a PER UHL E´ N,a ERNEST ARENASa a

Laboratory of Molecular Neurobiology, Department of Medical Biochemistry and Biophysics, Karolinska Institute, Stockholm, Sweden; bDepartamento de Biologı´a Celular y Molecular, Facultad de Ciencias Biolo´gicas, Pontificia Universidad Cato´lica de Chile, Santiago, Chile Key Words. Stem cells • Chemokines • Parkinson’s disease • Dopamine

ABSTRACT Increasing evidence suggests that ␣-chemokines serve several important functions in the nervous system, including regulation of neuroimmune responses, neurotransmission, neuronal survival, and central nervous system development. In this study, we first examined the function of two ␣-chemokines, chemokine ligand (CXCL) 6 and CXCL8, and their receptors, CXCR1 and CXCR2, in the developing rat ventral midbrain (VM). We found that CXCR2 and CXCL6 are regulated during VM development and that CXCL6 promotes the differentiation of nurr77-related receptor (Nurr1)ⴙ precursors into dopaminergic (DA) neurons in vitro. Intriguingly, CXCL8, a ligand expressed only in Homo sapiens, enhanced progenitor cell division, neurogenesis, and tyrosine hydroxylase-

positive (THⴙ) cell number in rodent precursor and neurosphere cultures. CXCL1, the murine ortholog of CXCL8, was developmentally regulated in the VM and exhibited activities similar but not identical to those of CXCL8. THⴙ cells derived from chemokine-treated VM neurospheres coexpressed Nurr1 and VMAT and were functionally active, as shown by calcium (Ca2ⴙ) fluxes in response to AMPA. In conclusion, our data demonstrate that CXCL1, CXCL6, and CXCL8 increase the number of DA neurons in VM precursor and neurosphere cultures by diverse mechanisms. Thus, ␣-chemokines may find an application in the preparation of cells for drug development or Parkinson’s disease cell replacement therapy. STEM CELLS 2008;26:1891–1900

Disclosure of potential conflicts of interest is found at the end of this article.

INTRODUCTION Chemokines (chemotactic cytokines) constitute a family of small protein ligands that were initially recognized for playing a role in leukocyte migration and communication [1]. Chemokines are classified into four major groups (based on the location and organization of their cysteine residues), referred to as C, CC, CXC, and CX3C. The CXC members, also called the ␣-chemokine subfamily, have a single amino acid residue between the first two conserved cysteine residues [1– 4]. In addition, on the basis of the presence or absence of a glutamic acid-leucine-arginine motif (ELR) immediately preceding the first cysteine, the CXC chemokines are further divided in two subclasses, ELR and non-ELR. Both chemokine ligand (CXCL) 6 and CXCL8 are members of the ELR subclass and interact with the CXCR1 and CXCR2 receptors [5]. Interestingly, chemokines and their receptors are expressed by all major cell types in the central nervous system (CNS), and

a growing body of evidence suggests that chemokines and their receptors also mediate cellular communication in the CNS [3]. For instance, CXCR2 has been shown to enhance the survival of hippocampal neurons [6, 7], and together with CXCL1, it is involved in patterning the spinal cord by controlling the positioning of oligodendrocyte precursors [8]. Furthermore, both the CXC ligand CXCL12 and its cognate receptor CXCR4 are required for normal development of the hippocampal dental gyrus [9, 10] and cerebellar granule cell layers [11, 12], where they regulate precursor proliferation in cooperation with Sonic hedgehog [13]. Thus, substantial evidence suggests that CXC chemokines and CXC receptors regulate diverse mechanisms and functions, such as proliferation, maturation, and migration of several subtypes of neural precursors during CNS development. However, little is known about the roles of chemokines and chemokine receptors in the development of the midbrain. To date, only one chemokine receptor, CXCR4b—a zebrafish homolog to human CXCR4 —and one ligand, CXCL15/WECHE, have been found expressed in the ventral midbrain (VM) [14,

Author contributions: L.C.E.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing; H.M.: conception and design, collection and/or assembly of data; A.E. and S.M.: collection and/or assembly of data, data analysis and interpretation; E. Andersson: collection and/or assembly of data; P.U.: data analysis and interpretation; E. Arenas: conception and design, financial support, manuscript writing, final approval of manuscript. Correspondence: Ernest Arenas, M.D., Ph.D., Laboratory of Molecular Neurobiology, Department of Medical Biochemistry and Biophysics, Karolinska Institute, Scheeles va¨g 1, 17177 Stockholm, Sweden. Telephone: 46-8-52487663; Fax: 46-8-341960; e-mail: Ernest.Arenas@ ki.se Received September 9, 2007; accepted for publication April 8, 2008; first published online in STEM CELLS EXPRESS April 24, 2008. ©AlphaMed Press 1066-5099/2008/$30.00/0 doi: 10.1634/stemcells.2007-0753

STEM CELLS 2008;26:1891–1900 www.StemCells.com

Chemokines in Dopaminergic Development

1892

Figure 1. CXCR2 and CXCL6 are expressed in the developing rat ventral midbrain (VM) and increase the number of dopaminergic (DA) neurons in rat precursor cultures by different mechanisms. cDNAs from VM were analyzed by quantitative polymerase chain reaction. (A, B): CXCR2 and CXCL6 were first detected in the VM at embryonic day (E) 11.5, prior to DA neurogenesis in the rat. The expression of CXCR2 peaked at E13.5. By the end of DA neurogenesis, at E15.5, the expression of CXCR2 decreased. Instead, CXCL6 was expressed at a relatively constant level throughout the neurogenic period and until P1. (C): CXCL8 treatment increased the number of Nurr1⫹ cells by threefold. (D): Both CXCL6 and CXCL8 increased the number of TH⫹ cells in a dose-dependent manner after 3 days in vitro. (E–I): Double immunocytochemistry revealed that CXCL6, but not CXCL8, increased the differentiation of Nurr1⫹ precursors that acquired TH expression in rat E14.5 cultures. (E–G) show cultures treated with 0, 2 and 0.5 ␮g/ml CXCL6 and CXCL8, respectively. Each value represents the mean ⫾ SEM expression of CXCR2 and CXCL6 mRNA in at least three embryos. ⴱ, p ⬍ .05; ⴱⴱ, p ⬍ .01. One-way analysis of variance with Bonferroni post test was used (n ⱖ 3). Scale bar ⫽ 37.5 ␮m. Abbreviations: Ctrl, control; CXCL, chemokine ligand; TH, tyrosine hydroxylase.

15]. Moreover, the function of chemokines in dopaminergic (DA) neuron development is unknown, and only cytokines such as interleukin-1b and interleukin-11 have been reported to regulate and promote DA neuron development [16 –19]. In this study, we have investigated whether two ␣-chemokine receptors (CXCR1 and CXCR2) and their common ligands (CXCL1, CXCL6, and CXCL8) regulate midbrain development. Our results describe, for the first time, the involvement of ␣-chemokines in this process. In particular, our data point to the two endogenous ligands, CXCL1 and CXCL6, as key regulators of progenitor proliferation and DA precursor differentiation, respectively.

MATERIALS

AND

METHODS

Animals Male and female wild-type CD-1 mice (25–35 g; Charles River Laboratories, Sulzfeld, Germany, http://www.criver.com) were housed, bred, and treated according to the guidelines of the European Communities Council (directive 86/609/EEC) and the Society for Neuroscience (available at http://www.sfn.org), and all experiments were approved by the local ethical committee. Sprague-Dawley rats were time-mated (Scanbur, Stockholm, Sweden, http://www.scanbur.eu/). Ethical approval for animal experimentation was granted by Stockholm Norra Djurfo¨rso¨ksetisks Na¨mnd.

Edman, Mira, Erices et al.

1893

Quantitative Polymerase Chain Reaction and Quantification of Gene Expression Total RNA was isolated from rat VM on embryonic day (E) 10.5, E11.5, E13.5, E14.5, E15.5, and postnatal day (P) 1 with the RNeasy Mini extraction kit (Qiagen, Hilden, Germany, http:// www1.qiagen.com). Reverse transcription reactions and real-time quantitative polymerase chain reaction (Q-PCR) were carried out as described previously [20]. The following polymerase chain reaction (PCR) program was used for SYBR Green detection on the ABI Prism 5700 Detection System (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com): 94°C for 2 minutes; 35– 40 cycles of 94°C for 30 seconds, 59°C for 30 seconds, and 72°C for 15 seconds; and 80°C for 5 seconds. Quantum RNA classic 18S internal standard was purchased from Ambion (Austin, TX, http://www.ambion.com), and PCR primers were purchased from DNA Technology a/S (Aarhus, Denmark, http://www.dnatechnology.dk). Statistical analysis of the results was performed using a two-tailed Wilcoxon signed rank test. Significance for all tests was assumed at the level of p ⬍ .05 (ⴱ, p ⬍ .05; ⴱⴱ, p ⬍ .01; ⴱⴱⴱ, p ⬍ .001).

Precursor Cultures and Treatments VM from E14.5 rat embryos were dissected, mechanically dissociated, and plated at a final density of 1 ⫻ 105 cells per cm2 on poly-D-lysine-coated 24- or 48-well plates in N2, consisting of a 1:1 mixture of Ham’s F-12 medium and minimal essential medium, containing HEPES and glutamine, N2 supplement (Gibco, Grand Island, NY, http://www.invitrogen.com), and 1 mg/ml bovine serum albumin (BSA). All factors, 0.2 ␮g/ml CXCL1, 2 ␮g/ml CXCL6, and 0.5 ␮g/ml CXCL8 (R&D Systems Inc., Minneapolis, http:// www.rndsystems.com), were added at once at the initiation of culture, and bromodeoxyuridine (BrdU) (10 ␮M) was added 6 hours prior to fixation. To test specificity, a CXCL8 blocking antibody (MAB208; R&D Systems) was added at the beginning of the experiment simultaneously with or prior to CXCL8, at a concentration of 16 ␮g/ml.

Neurosphere Cultures and Treatments VM from E11.5 embryos obtained from wild-type CD-1 mice were dissected, mechanically dissociated, and plated at a final density of 1 ⫻ 105 cells per milliliter in 35-mm dishes in N2 supplemented with 1.5 mg/ml Albumax (Sigma-Aldrich, St. Louis, http://www. sigmaaldrich.com). Basic fibroblast growth factor (bFGF) and epidermal growth factor (EGF) were added at a concentration of 20 ng/ml each with or without CXCL6 and CXCL8. After 4 days in vitro (DIV), neurospheres were collected, centrifuged, resuspended in N2, triturated using a 200-␮l pipette, and plated as described above. For differentiation, neurospheres were plated at a final density of approximately 50 spheres per poly-D-lysine-coated well, and cells were analyzed 3 days later. BrdU (10 ␮M) was added 2 hours prior to fixation.

Measurement of Intracellular Ca2ⴙ Signaling VM precursor cultures untreated or treated with CXCL1, CXCL6, or CXCL8 were loaded with the Ca2⫹-sensitive fluorescence indicator Fluo-3/AM (Molecular Probes, Eugene, OR, http://probes. invitrogen.com) at a concentration of 5 ␮M in Krebs buffer at 37°C for 30 minutes. Ca2⫹ measurements were carried out at 37°C in a heat-controlled chamber (QE-1; Warner Instruments, Hamden, CT, http://www.warneronline.com) with a cooled EMCCD camera (Cascade II; Photometrics, Tucson, AZ, http://www.photomet.com) mounted on an inverted microscope equipped with a ⫻25, 0.8 numerical aperture water objective (Axiovert 100; Carl Zeiss, Jena, Germany, http://www.zeiss.com). Excitation at 495 nm was performed with a filter wheel (Lambda 10-3 with SmartShutter; Sutter Instrument, Novato, CA, http://www.sutter.com). Sampling frequency was set to 0.2 Hz. The computer software MetaFluor (Molecular Devices, Downington, PA, http://www.moleculardevices. com) was used to control all devices and to analyze acquired images. Experiments were performed in Krebs buffer, and drugs were bath-applied. Cells were considered responsive when the in-

www.StemCells.com

Figure 2. CXCL8 increases the number of proliferating cells in rat embryonic day 14.5 ventral midbrain precursor cultures. (A–D): Neither CXCL6 (2 ␮g/ml) nor CXCL8 (0.5 ␮g/ml) increased cell death after 3 days in vitro. (E–H): CXCL8, but not CXCL6, increased the percentage of BrdU⫹ cells by threefold after a 6-hour pulse on day 3. (E): 100% of cells per well in Ctrl were (mean ⫾ SD) 14.4 ⫾ 3.7 cells per field. ⴱ, p ⬍ .01. Ten fields per well were counted. One-way analysis of variance with Bonferroni post test was used (n ⱖ 3). Scale bar ⫽ 37.5 ␮m. Abbreviations: BrdU, bromodeoxyuridine; Ctrl, control; CXCL, chemokine ligand.

crease in Ca2⫹ exceeded 1.25 of the baseline. Four independent experiments were analyzed, and a total of 400 cells were recorded (100 cells per condition). The recorded cells were examined by immunocytochemistry, as described below.

Immunocytochemical Analysis Cells were fixed for immunocytochemistry in ice-cold 4% paraformaldehyde for 15–20 minutes, washed with phosphate-buffered saline (PBS), blocked with serum for 1 hour, and incubated overnight in dilution buffer (1⫻ PBS, 1% BSA, and 0.3% Triton X-100) with one of the following antibodies: mouse ⬀Tyrosine hydroxylase (TH) (1:250 dilution; Diasorin, Saluggia, Italy, http://www. diasorin.com), rabbit ⬀TH (1:1,000; Pel-Freez, Rogers, AK, http:// www.invitrogen.com), rabbit ⬀nurr77-related receptor (Nurr1) (1: 150; Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www. scbt.com), rabbit ⬀active caspase 3 (1:100; Cell Signaling Technology, Beverly, MA, http://www.cellsignal.com), mouse ⬀BrdU (1:100; Dako, Glostrup, Denmark, http://www.dako.com), rat ⬀BrdU (1:250; Abcam, Cambridge, MA, http://www.abcam. com), mouse ⬀␤ tubulin (1:500; Promega, Madison, WI, http:// www.promega.com), rabbit ⬀glial fibrillary acidic protein (GFAP) (1:250; Dako), rabbit ⬀Nestin (1:150; HybridomaBank), rat ⬀myelin basic protein (MBP) (1:250; Chemicon, Temecula, CA, http:// www.chemicon.com), rabbit ⬀Ki67 (1:800; NeoMarkers), or rabbit ⬀ VMAT (1:200; Chemicon). After washing, cultures were incubated for 1 hour in dilution buffer with the appropriate secondary antibody (CY2, CY3, or rhodamine IgG; 1:500; all from Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com). Double staining was performed by sequential single staining, as described above. At the end of all staining procedures, cultures were incubated with Hoechst 33258 reagent or propidium iodide for 10 minutes. BrdU immunocytochemistry included incubation for 30

1894


Figure 3. CXCL8 increases progenitor cell division and neurogenesis in embryonic day 11.5 mouse ventral midbrain (VM) neurospheres. (A): Treatment with CXCL8, but not CXCL6 increased the number of BrdU⫹ cells, after a 2 hour BrdU pulse in mouse VM neurosphere cultures at passage 2. (B): No change in the number of Ki67⫹ cells after administration of CXCL6 or CXCL8 was detected. (C): CXCL8 increased the proportion of BrdU⫹ cells out of the total number of Ki67⫹ cells. (D): Neither CXCL6 (2 ␮g/ml) nor CXCL8 (0.5 ␮g/ml) changed the number of nestin⫹ cells in the cultures. CXCL8 increased the number of Nurr1⫹ cells (E) and Tuj1⫹ cells (F) in the neurosphere cultures. (G): Quantitative polymerase chain reaction showed a similar upregulation in Nurr1 and Tuj1 mRNA expression. (H, I): Confocal image showing TuJ1⫹ cells in a Ctrl and a CXCL8-treated neurosphere. ⴱ, p ⬍ .05; ⴱⴱ, p ⬍ .01. One-way analysis of variance with Bonferroni post test was used (n ⫽ 3–5). Abbreviations: BrdU, bromodeoxyuridine; Ctrl, control; CXCL, chemokine ligand; PI, propidium iodide.

minutes with 2 M HCl prior to primary antibody addition. Images were acquired (from stained cells in PBS at room temperature) with a Zeiss Axioplan 100M microscope (LD Achrostigmat, ⫻20, 0.3 PHII ⬀0.2; LD Achroplan, ⫻40, 0.60 Korr PHII ⬀0.2) and collected with a Hamamatsu C4742.95 camera (Hamamatsu Corp., Bridgewater, NJ, http://www.hamamatsu.com) (with OpenLab imaging software, Improvision, Coventry, England, http://www.improvision. com).

Statistical Analysis Quantitative immunocytochemical data represent the means ⫾ SEM of counts from 10 nonoverlapping fields in three wells per condition from three separate experiments. Neurospheres were considered positive for stainings when one or more cells were clearly labeled. For the reverse transcription-PCR experiments, three separate experiments were analyzed. Statistical analysis was performed using Prism 4 (GraphPad Software, Inc., San Diego, http://www. graphpad.com) as described in the figure legends, with significance for all tests assumed at the level of p ⬍ .05 (ⴱ, p ⬍ .05; ⴱⴱ, p ⬍ .01; ⴱⴱⴱ, p ⬍ .001). Statistical tests were chosen according to the distribution of the sample population.

RESULTS CXCR2 and CXCL6 Are Expressed in the VM To study the expression of CXCR1 and CXCR2 and their ligands during the course of VM development, we performed Q-PCR using rat VM tissue at different developmental stages (E10.5, E11.5, E13.5, E14.5, E15.5, and P1). We found that CXCR2 mRNA, but not CXCR1 mRNA, was expressed in the developing rat VM (Fig. 1A; data not shown). The expression of CXCR2 was first detected at the onset of DA neurogenesis, E11.5, and it peaked at E13.5 with the birth of DA neurons. We next examined the expression of the CXCR1 and CXCR2 ligand CXCL6. The expression of CXCL6 mRNA was first detected at E11.5, and CXCL6 was expressed at relatively constant levels through neurogenesis and until P1 (Fig. 1B). Thus, our results showing expression of CXCR2 and CXCL6 in the developing VM during DA neurogenesis suggested a possible function of ␣-chemokines in DA neuron development.


1895

Figure 4. CXCL8 increases the number of TH⫹ ventral midbrain neurospheres after 3 days of differentiation. (A–D): Spheres differentiated for 3 days in vitro gave rise to the three major lineages of the central nervous system; neurons (Tuj1⫹), oligodendrocytes (MBP⫹), and astrocytes (GFAP⫹). (E–G): CXCL6, but not CXCL8, increased the number of MBP⫹ cells per sphere. (H): Neither CXCL6 nor CXCL8 increased the number of GFAP⫹ cells in the spheres. (I–M): CXCL6 and CXCL8 increased both the number of TH⫹ cells per neurosphere (I) and the percentage of TH⫹ neurospheres (J). The spheres were differentiated for 3 days in vitro after passage 2. ⴱ, p ⬍ .05; ⴱⴱ, p ⬍ .01; ⴱⴱⴱ, p ⬍ .001. One-way analysis of variance with Bonferroni post test was used (n ⫽ 3– 6). Abbreviations: Ctrl, control; CXCL, chemokine ligand; GFAP, glial fibrillary acidic protein; MBP, myelin basic protein; TH, tyrosine hydroxylase.

CXCL6 and CXCL8 Increase the Number of THⴙ Cells, but Only CXCL6 Promotes the Differentiation of Nurr1ⴙ/THⴚ Precursors into Nurr1ⴙ/THⴙ Neurons To examine the function of CXCL6/CXCR2 in DA neuron development, we treated rat E14.5 VM primary precursor cultures with CXCL6. We also included a control treatment with CXCL8, another ␣-chemokine that signals through the same receptors as CXCL6 but is present only in Homo sapiens. Thereafter, we examined the number of cells expressing Nurr1 or TH, two markers that identify DA precursors (Nurr1⫹/TH⫺) and DA neurons (Nurr1⫹/TH⫹). Surprisingly, dose-response analysis demonstrated that both CXCL6 and CXCL8 increased the number of Nurr1⫹ cells in a dose-dependent manner, reaching maximal effects at 0.5 ␮g/ml CXCL8 and 2 ␮g/ml CXCL6 www.StemCells.com

(Fig. 1C; control value [mean ⫾ SD], 4,475 ⫾ 63 Nurr1⫹ cells per cm2). A similar dose response was observed for TH⫹ neurons, which showed 3.5- and 4.5-fold increases after treatment with CXCL8 and CXCL6, respectively (Fig. 1D–1G; control value [mean ⫾ SD], 37.85 ⫾ 11 TH⫹ cells per cm2). To examine the mechanism by which the number of TH⫹ neurons increased in the culture, we examined whether CXCL6 and CXCL8 enhanced the differentiation of DA precursors as assessed by the conversion of Nurr1⫹/TH⫺ cells into Nurr1⫹/ TH⫹ cells. Interestingly, the endogenous ␣-chemokine CXCL6, but not CXCL8, increased the number of TH⫹ neurons in the Nurr1 population, suggesting that CXCL6 has a predominant effect on the conversion of DA precursors into DA neurons (Fig. 1H, 1I). Thus, our results indicate that CXCL6 promotes DA differentiation in VM precursor cultures, whereas CXCL8 does not.

1896


Figure 5. The CXCL8 blocking antibody MAB208 decreases the CXCL8-induced, and basal, number of TH⫹ cells in rat embryonic day 14.5 ventral midbrain precursor cultures. (A–C): Administration of MAB208 decreased the CXCL8-induced number of TH⫹ cells by 60%. The number of TH⫹ cells in basal conditions also decreased after administration of CXCL8 antibody, suggesting that the CXCL8-blocking antibody cross-reacted with an endogenous ␣-chemokine similar to CXCL8. (D–F): Colocalization of TH and VMAT in precursor cultures treated with CXCL6 and CXCL8. ⴱⴱ, p ⬍ .01; ⴱⴱⴱ, p ⬍ .001. One-way analysis of variance with Bonferroni post test was used (n ⫽ 3). Scale bar ⫽ 37.5 ␮m. Abbreviations: Ab, MAB208; C, control; Ctrl, control; CXCL, chemokine ligand; L8, chemokine ligand 8; TH, tyrosine hydroxylase.

CXCL8 Promotes Neurogenic Divisions in VM Precursor Cultures To complete the characterization of the mechanism by which CXCL6 and CXCL8 increase the number of DA neurons, we examined survival in the cultures by determining the number of active caspase 3⫹ cells. Proliferation was also assessed by means of BrdU incorporation. No difference in active caspase 3 could be seen in response to CXCL6 or CXCL8 treatment (Fig. 2A–2D), indicating that neither of these chemokines exerted a survival-promoting effect. However, we found that CXCL8, but not CXCL6, increased the number of proliferating cells by threefold (Fig. 2E–2H). Since CXCL8 induced a threefold increase in both BrdU⫹ and TH⫹ cells, we hypothesized that such a simultaneous increase could correspond to an induction of neurogenesis. In agreement with this possibility, BrdU⫹/ nestin⫹ cells, found during the 1st day of culture, were not detected after a 3-day differentiation period. Thus, our results suggest that CXCL8 promotes the neurogenic division of progenitors and that CXCL6 promotes the differentiation of Nurr⫹ precursors into DA neurons.

CXCL8 Promotes Neurogenesis in Proliferating VM Neurospheres Next we set out to determine whether these two ligands also influence VM development at a proliferative progenitor level. We therefore investigated whether CXCL6 or CXCL8 could be used to promote cell division and/or differentiation of progenitors in VM neurospheres. Mouse VM E11.5 neurospheres were grown in the presence of bFGF and EGF. Similar to our results in primary precursor cultures, incorporation of BrdU in the presence of mitogens was increased threefold by CXCL8, but not by CXCL6 (Fig. 3A), indicating that CXCL8 indeed promotes cell division in VM neurosphere cultures. Interestingly, the number of Ki67⫹ cells did not change (Fig. 3B), but when the labeling index (percentage BrdU⫹ and ki67⫹/total Ki67⫹) was examined, we found that CXCL8 clearly increased the proportion of Ki67⫹ cells that went through DNA synthesis (Fig. 3C). Since the pool of Ki67⫹ progenitors did not increase, our results suggest that the increase in BrdU by CXCL8 does not reflect an increase in proliferation. We then examined the number of nestin-positive cells in the neurospheres (Fig. 3D) and the number of spheres (data not shown). However, these parameters did not change, indicating that the effect of CXCL8 was not on

self-renewal of VM progenitors. We thus examined whether CXCL8 increased the number of Nurr1⫹ precursors and neurons. CXCL8 increased the number of postmitotic cells, including Nurr1⫹ cells (Fig. 3E) and TuJ1⫹ cells (Fig. 3F, 3H, 3I). These results were also confirmed by Q-PCR, which showed a significant increase in the levels of TuJ1 and Nurr1 transcripts (Fig. 3G). Thus, our results indicate that the increase in BrdU incorporation by CXCL8 in proliferating progenitors does not result in self-renewal or further proliferation but rather enhances neurogenesis and the number of Nurr1⫹ precursors. However, no TH⫹ neurons were detected. We therefore examined the effects of chemokines in neurosphere cultures after removal of bFGF and EGF.

CXCL6 and CXCL8 Enhance the Number of DA Neurons in VM Neurospheres In the absence of bFGF and EGF, most VM neurospheres gave rise to cells belonging to the three different neural lineages: neurons (TuJ1⫹ cells), oligodendrocytes (MBP⫹ cells), and astrocytes (GFAP⫹ cells) (Fig. 4A– 4D). As in proliferating neurospheres, CXCL8 increased the number of neurons but did not affect the number of MBP⫹ cells or GFAP⫹ cells (Fig. 4E– 4H, and data not shown). Instead, CXCL6 induced a modest increase in MBP⫹ cells but not in GFAP⫹ cells (Fig. 4E, 4H). Importantly, we found that in the absence of bFGF and EGF, both CXCL6 and CXCL8 increased the number of TH⫹ cells per sphere by 50% and the percentage of spheres that contained TH⫹ cells after 3 DIV (Fig. 4I– 4M). These findings support the idea that CXCL6 and CXCL8 also increase the number of DA neurons in neural stem cell preparations grown as neurospheres.

A Blocking CXCL8-Antibody Decreases the Basal and CXCL8-Induced Number of DA Neurons in Precursor Cultures Since CXCL8 is not detected in mouse or rat and since chemokines are known for their promiscuity, we decided to examine the specificity of the CXCL8 effects. We therefore examined the effects of a CXCL8 blocking antibody on rat E14.5 primary cultures treated with CXCL8 and found an almost 60% decrease in the number of TH⫹-expressing cells. Interestingly, the antibody also decreased the number of TH⫹ cells in control well without CXCL8 (Fig. 5). Since CXCL8 is not expressed in mice or rats, our results suggested that the antibody cross-reacts with an endogenous ␣-chemokine exhibiting structure/function sim-


1897

Figure 6. CXCL1 is expressed in the ventral midbrain (VM) and induces dopaminergic neurogenesis. (A): Developmental regulation of CXCL1 expression in the developing VM. (B): CXCL1 (0.2 ␮g/ml) increased BrdU incorporation and the number of TH⫹ cells in embryonic day 14.5 rat VM precursor cultures. (C): A CXCL8blocking antibody decreased the CXCL1induced increase in TH⫹ neurons. (D–F): In proliferating neurospheres, CXCL1 did not change the labeling index (D) or the number of Tuj1⫹ (E) or Nurr1⫹ cells per sphere (F). (G–H): CXCL1 did not increase the number of MBP or GFAP⫹ cells. (I–K): CXCL1 increased the number of TH⫹ cell per sphere, but only in the absence of basic fibroblast growth factor and epidermal growth factor. ⴱ, p ⬍ .05; ⴱⴱ, p ⬍ .01; ⴱⴱⴱ, p ⬍ .001. One-way analysis of variance with Bonferroni post test was used (n ⫽ 3–5). Abbreviations: Ab, antibody; BrdU, bromodeoxyuridine; Ctrl, control; CXCL, chemokine ligand; GFAP, glial fibrillary acidic protein; MBP, myelin basic protein; TH, tyrosine hydroxylase.

ilar to those of CXCL8. This effect was specific, since it was not detected in CXCL6-treated cultures (not shown). Moreover, the effect of the antibody was not related to a different degree of maturation of TH⫹ cells, as both CXCL6- and CXCL8-treated cultures showed similar degrees of morphological differentiation and coexpressed TH and VMAT (Fig. 5D–5F). We next examined whether the decrease in TH⫹ cells in control conditions was due to nonspecific cell death induced by the antibody rather than a blocking effect on endogenous chemokines. Active caspase 3 immunostainings revealed that treatment with the blocking antibody did not increase the number of active caspase 3⫹ cells (data not shown). Thus, our results suggest that the blocking antibody does not induce cell death but rather prevents the generation of TH⫹ DA neurons by CXCL8 treatment or an unknown ␣-chemokine. www.StemCells.com

CXCL1, an Endogenously Expressed ␣-Chemokine, Promotes DA Neurogenesis in the Absence of Mitogens The activity of the human CXCL8 in murine cells prompted us to study CXCL1, a member of the ␣-chemokine subfamily that binds both CXCR1 and CXCR2 and is the murine ortholog to the human CXCL8 [21]. Interestingly, Q-PCR analysis confirmed that CXCL1 is highly expressed in the VM between E11.5 and E14.5 in mouse (Fig. 6A). Next we examined whether CXCL1 could also regulate VM cell division in the same manner as CXCL8. Analysis of BrdU incorporation in precursor cultures revealed that CXCL1 increases the number of cells incorporating BrdU (Fig. 6B). Moreover, we found that CXCL1 does increase the number of TH⫹ neurons in VM precursor cultures without affecting cell


1898

death and that these effects can also be blocked by the CXCL8-blocking antibody (Fig. 6C). These results indicated that CXCL1 and CXCL8 regulate similar functions in precursor cultures. However, CXCL1 did not increase the labeling index (Fig. 6D) or Ki67⫹ cells in proliferating neurospheres (with bFGF and EGF), suggesting that CXCL1 does not increase the pool of proliferating progenitors. In such cultures, we detected a nonsignificant increase in the number of neurons and Nurr1⫹ cells (Fig. 6E, 6F). Similarly, we did not detect any increase in the number of astrocytes (GFAP⫹ cells) or oligodendrocytes (MBP⫹ cells) (Fig. 6G, 6H). However, we detected an increase in the number of TH⫹ cells in the absence of mitogens (Fig. 6I– 6K), suggesting that the increase in BrdU results in increased numbers of TH⫹ cells by a neurogenic mechanism. Thus, CXCL1 exerts selective effects on TH⫹ neurons, since it did not increase the overall number of neurons.

␣-Chemokine-Treated Precursor Cultures Differentiate into DA Neurons That Activate Ca2ⴙ Signaling in Response to AMPA Receptor Activation To further validate that CXCL1, CXCL6, and CXCL8 enhanced the number of functional DA neurons, we examined whether the TH⫹ cells born in vitro were capable of responding to physiological stimuli, such as glutamatergic afferents. We therefore carried out intracellular Ca2⫹ flux experiments. The glutamate receptor agonist AMPA is known to evoke a robust intracellular Ca2⫹ increase in neurons, and it is well documented that the AMPA receptor is expressed in DA neurons [22]. Cells were loaded with the Ca2⫹-sensitive probe Fluo-3/AM, and Ca2⫹ signaling was monitored during agonist exposure. After Ca2⫹ imaging, the DA neurons were immunostained for TH and back-traced/correlated to their respective Ca2⫹ response to AMPA. When control cells were exposed to 10 ␮M AMPA and immunostained for TH, prominent Ca2⫹ peaks were recorded in TH⫹ cells (Fig. 7A). A similar transient Ca2⫹ increase in response to AMPA was observed in TH⫹ cells induced by CXCL1, CXCL6, or CXCL8 treatments (Fig. 7B–7D). A statistical analysis of total number of AMPA-responding cells showed no significant difference among control, CXCL1, CXCL6, and CXCL8 cells (Fig. 7E). The numbers of control, CXCL1, CXCL6, and CXCL8 cells that responded to AMPA were 90.0%, 88.9%, 88.6%, and 87.7%, respectively. These results indicate that the CXCL1-, CXCL6-, and CXCL8-treated VM precursors produced functional DA neurons capable of responding to afferent stimuli and activating cell signaling mechanisms.

DISCUSSION In recent years, chemokines have been suggested to play a major role in the development of the CNS [3]. Until recently, CXCR4b and CXCL15/WECHE were the only two chemokines described in the VM [14, 15]. However, there is no previous functional information available on the role of chemokines in the developing midbrain. Our results show that chemokines regulate VM DA neuron development. In particular, we found that three ␣-chemokines, CXCL1, CXCL6, and CXCL8, regulate three crucial aspects of DA neuron development: the proliferation of DA progenitors, neurogenesis, and precursor differentiation into VM DA neurons. Although all of them increase the number of TH⫹ cells in the absence of mitogens, they do so by regulating different mechanisms. Both CXCL8 and CXCL1 promoted progenitor cell divisions. However, although CXCL8 increased neurogenesis and the numbers of Nurr1⫹ DA neurons and

Figure 7. Analysis of the functionality of dopaminergic (DA) neurons in chemokine-treated precursor cultures. Shown is the Ca2⫹ response evoked by AMPA in TH⫹ cells in ventral midbrain (VM) precursor cultures treated with CXCL1, CXCL6, or CXCL8. (A–D): TH-positive (red) cells and Fluo-3/AM (green) loaded cells (left panels) treated with AMPA 10 ␮M showed a strong Ca2⫹ increase (right panels). Cells from different culture conditions showed a similar AMPA-triggered Ca2⫹ peak. Traces are representative single-cell recordings from the first TH-positive cell to the left in each of the pictures. (E): The mean number of dopamine neurons responding to AMPA with a Ca2⫹ increase exceeding 1.25 of the baseline did not change and was not different from embryonic day 11.5 mouse VM primary cultures (not shown). These results indicated that TH⫹ cells generated in vitro, in the presence of chemokines, are as functional and capable of responding to glutamatergic inputs via AMPA receptors as endogenous DA neurons. (F): Schematic representation of the function and the steps regulated by CXCL1, CXCL6, and CXCL8 during DA neuron development. We found that CXCL1 and CXCL8 regulate neurogenesis in proliferating VM progenitors, whereas CXCL6 promoted the differentiation of Nurr1⫹ DA precursors. Both mechanisms led to an increase in the number of TH⫹ DA neurons. Please note that CXCL8 also induced non-DA neurogenesis. Abbreviations: AMPA, alpha-amino-3-hydroxy5-methyl-4-isoxazolepropionic acid; BrdU, bromodeoxyuridine; CTRL, control; CXCL, chemokine ligand; sec, seconds; TH, tyrosine hydroxylase.

non-DA neurons, CXCL1 selectively enhanced DA neurogenesis in VM neurospheres. On the other hand, CXCL6 induced

Edman, Mira, Erices et al. differentiation of committed Nurr1⫹ postmitotic precursors into DA neurons. Our data also indicate that to generate TH⫹ cells from progenitors, it is necessary to remove mitogens from the media. In the absence of mitogens, CXCL8 increased the number of TH⫹ cells by 1.5-fold in progenitors expanded as neurospheres. Moreover, VM precursor cultures (that had not been exposed to mitogens) were more efficient in generating TH⫹ cells by differentiation than neurosphere cultures (threefold increase). However, since the expansion of neurospheres also contributes to the increased number of progenitors, the overall production of neurons was similar. Since bFGF has been previously used to expand and differentiate progenitors [23], our results suggest that the factor that limits DA differentiation in proliferating neurosphere cultures could be EGF or the combination of bFGF and EGF. Future work will aim at improving the expansion and differentiation conditions with chemokines. Our finding that 90% of the DA neurons produced by CXCL1, CXCL6, or CXCL8 treatment can respond to physiological afferent stimuli and activate Ca2⫹ signaling in response to AMPA suggests that DA neurons generated in vitro are fully functional. Thus, our results suggest a possible application of ␣-chemokines to expand DA progenitors, promote neurogenesis, enhance the number and differentiation of Nurr1⫹ DA precursors, and increase the number of VM TH⫹ DA neurons. Interestingly, our study identified CXCL8, a ligand not expressed in the rodent VM, as capable of promoting an increase in DA precursors. That led us to examine CXCL1, the murine ortholog of human CXCL8 [21]. Similarly to CXCL8, CXCL1 increased BrdU incorporation and the number of TH⫹ cells. This effect could be blocked by an anti-CXCL8 antibody, suggesting that they indeed have structural and functional similarities. Thus, our results suggest that CXCL1 is the endogenous ligand in the rodent VM that exerts the CXCL8-like activity blocked by the CXCL8 antibody. Interestingly, a more detailed analysis of the function of CXCL1 revealed that its effects were less pronounced, but more restricted, than those of CXCL8. Importantly, CXCL1 increased BrdU incorporation and the number of TH⫹ cells but not Ki67⫹ cells, the labeling index, or the number of TuJ1⫹ neurons, suggesting that CXCL1 does not affect overall neurogenesis but rather has a specific effect on DA neurogenesis. The finding that three chemokines, CXCL1, CXCL6, and CXCL8, that share the CXCR2 receptor, exert different biological activities was unexpected. These results, together with the stage-specific effects of CXCL6 on Nurr1⫹ precursors and of CXCL1 and CXCL8 on proliferative progenitors, suggest that CXCR2 might be coupled to different signaling pathways at different developmental stages. Alternatively, the high degree of receptor promiscuity exhibited by chemokines may allow them to bind and activate other chemokine receptors. However, despite the apparent promiscuity, our results suggest that ␣-chemokines exhibit a remarkable and highly specific biological activity profile on midbrain DA neurons. Another possibility is that CXCL6 and/or CXCL8 recruit additional coreceptors. One example of this is another ␣-chemokine, CXCL12, which has been recently reported to bind to CXCR4 and RDC1, a novel coreceptor [24]. Future studies will aim at defining whether receptors other than CXCR2 contribute to the distinct biological

1899

activities and whether other ␣-chemokines regulate DA neuron development. Previous studies have shown that CXCL1 signaling through its receptor, CXCR2, induces oligodendrocyte precursor cell (OPC) proliferation [25, 26] and inhibits OPC migration in the spinal cord [8]. However, we did not detect an effect by CXCL1 on MBP⫹ cells in the VM, but rather an effect by CXCL6. These studies show that ␣-chemokines exert remarkably specific and unsuspected functions in the developing brain. Our results also indicate that ␣-chemokines regulate multiple aspects of DA neuron development. Other cytokines used by the immune system, such as interleukin (IL)-1 and IL-11, have previously been shown to play a role in the conversion of mesencephalic-derived progenitors to DA neurons [17]. IL-1 also exerts trophic effects on DA and other catecholaminergic neurons both in vitro and in vivo [27, 28]. Thus, combined, emerging evidence indicates that signaling molecules typically used postnatally by the immune system, such as interleukins and chemokines, play an essential role during embryonic development of the CNS.

CONCLUSION Our study supports the idea that chemokines regulate ventral midbrain development. In this context, our report identifies CXCL1, CXCL6, and CXCL8 as novel regulators of distinct aspects of DA development, including proliferation, neurogenesis, and differentiation. Moreover, our data show that VM progenitors and DA precursors respond to different ␣-chemokines in a very specific manner (Fig. 7F). We therefore suggest that chemokines could be useful in the development of novel protocols to enhance the expansion and differentiation of stem/ progenitor cells into DA neurons. Moreover, cells produced by this method may find a therapeutic application as tools for drug development or cell replacement therapies for neurodegenerative diseases affecting DA neurons, such as Parkinson’s disease.

ACKNOWLEDGMENTS We thank Paola Sacchetti and Ruani Fernando for critical reading of the manuscript, Lottie Jansson-Sjo¨strand and Johnny So¨derman for technical assistance, Annika Ka¨ller and Alessandra Nanni for secretarial assistance, and members of the Arenas laboratory for fruitful discussions. Financial support was obtained from the Swedish Foundation for Strategic Research, the Swedish Royal Academy of Sciences, the Knut and Alice Wallenberg Foundation, the European Commission (Eurostemcell) and the Karolinska Institutet. H.M. and A.E. contributed equally to this work. H.M. is currently affiliated with Instituto de Salud Carlos III, Majadahonda, Madrid, Spain.

DISCLOSURE

2 Yoshimura T, Matsushima K, Tanaka S et al. Purification of human monocyte-derived neutrophil chemotactic factor that has peptide se-

www.StemCells.com

CONFLICTS

All authors have served as consultants, officers, or members of the Board, and own stock in Neuro Therapeutics AB.

REFERENCES 1

OF POTENTIAL OF INTEREST

3

quence similarity to other host defence cytokines. Proc Natl Acad Sci U S A 1987;84:9233–9237. Murphy PM, Baggiolini M, Charo IF et al. International union of pharmacology: XXII. Nomenclature for chemokine receptors. Pharmacol Rev 2000;52:145–176. Bajetto A, Bonavia R, Barbero S et al. Chemokines and their recep-


1900

4 5 6 7 8 9 10 11 12 13 14 15 16 17

tors in the central nervous system. Front Neuroendocrinol 2001;22: 147–184. Gangur V, Birmingham NP, Thanesvorakul S. Chemokines in health and disease. Vet Immunol Immunopathol 2002;86:127–136. Zlotnik A, Yoshie O. Chemokines: A new classification system and their role in immunity. Immunity 2000;12:121–127. Araujo DM, Cotman CW. Trophic effects of interleukin-4, -7 and -8 on hippocampal neuronal cultures: Potential involvement of glial-derived factors. Brain Res 1993;600:49 –55. Horuk R, Martin AW, Wang Z et al. Expression of chemokine receptors by subsets of neurons in the central nervous system. J Immunol 1997; 158:2882–2890. Tsai HH, Frost E, To V et al. The chemokine receptor CXCR2 controls positioning of oligodendrocyte precursors in developing spinal cord by arresting their migration. Cell 2002;110:373–383. Lu M, Grove EA, Miller RJ. Abnormal development of the hippocampal dentate gyrus in mice lacking the CXCR4 chemokine receptor. Proc Natl Acad Sci U S A 2002;99:7090 –7095. Bagri A, Gurney T, He X et al. The chemokine SDF1 regulates the migration of dentate granule cells. Development 2002;129:4249 – 4260. Ma Q, Jones D, Borghesani PR et al. Impaired B-lymphopoiesis, myelopoiesis, and derailed cerebellar neuron migration in CXCR4- and SDF-1-deficient mice. Proc Natl Acad Sci U S A 1998;95:9448 –9453. Zhu Y, Yu T, Zang XC et al. Role of the chemokine SDF-1 as the meningeal attractant for embryonic cerebellar neurons. Nat Neurosci 2002;5:719 –720. Klein RS, Rubin JB, Gibson HD et al. SDF-1 alpha induces signalchemotaxis and enhances Sonic hedgehog-induced proliferation of cerebellar granule cells. Development 2001;128:1971–1981. Chong SW, Emelyanov A, Gong Z et al. Expression pattern of two zebrafish genes, cxcr4a and cxcr4b. Mech Dev 2001;109:347–354. Ohneda O, Ohneda K, Nomiyama H et al. WECHE: A novel hematopoietic regulatory factor. Immunity 2000;12:141–150. Ling ZD, Potter ED, Lipton JW et al. Differentiation of mesencephalic progenitor cells into dopaminergic neurons by cytokines. Exp Neurol 1998;149:411– 423. Potter ED, Ling ZD, Carvey PM. Cytokine-induced conversion of mes-

18 19 20 21

22 23 24 25 26 27 28

encephalic-derived progenitor cells into dopaminergic neurons. Cell Tissue Res 1999;296:235–246. Storch A, Paul G, Csete M et al. Long-term proliferation and dopaminergic differentiation of human mesencephalic neural precursor cells. Exp Neurol 2001;170:317–325. Rolletschek A, Chang H, Guan K et al. Differentiation of embryonic stem cell-derived dopaminergic neurons is enhanced by survival-promoting factors. Mech Dev 2001;105:93–104. Rawal N, Castelo-Branco G, Sousa KM et al. Dynamic temporal and cell type-specific expression of Wnt signaling components in the developing midbrain. Exp Cell Res 2006;312:1626 –1636. Srinivasan S, Hatley ME, Reilly KB et al. Modulation of PPARalpha expression and inflammatory interleukin-6 production by chronic glucose increases monocyte/endothelial adhesion. Arterioscler Thromb Vasc Biol 2004;24:851– 857. Liu QS, Pu L, Poo MM. Repeated cocaine exposure in vivo facilitates LTP induction in midbrain dopamine neurons. Nature 2005;437: 1027–1031. Studer L, Tabar V, McKay RD. Transplantation of expanded mesencephalic precursors leads to recovery in parkinsonian rats. Nat Neurosci 1998;1:290 –295. Balabanian K, Lagane B, Infantino S et al. The chemokine SDF-1/ CXCL12 binds to and signals through the orphan receptor RDC1 in T-lymphocytes. J Biol Chem 2005;280:35760 –35766. Robinson S, Tani M, Strieter RM et al. The chemokine growth-regulated oncogene-alpha promotes spinal cord oligodendrocyte precursor proliferation. J Neurosci 1998;18:10457–10463. Wu Q, Miller RH, Ransohoff RM et al. Elevated levels of the chemokine GRO-1 correlate with elevated oligodendrocyte progenitor proliferation in the jimpy mutant. J Neurosci 2000;20:2609 –2617. Akaneya Y, Takahashi M, Hatanaka H. Interleukin-1 beta enhances survival and interleukin-6 protects against MPP⫹ neurotoxicity in cultures of fetal rat dopaminergic neurons. Exp Neurol 1995;136:44 –52. Nakao N, Itakura T, Uematsu Y et al. Pretreatment with interleukin-1 enhances survival of sympathetic ganglionic neuron grafts. Neurol Med Chir (Tokyo) 1994;34:407– 411.