HB-EGF affects astrocyte morphology, proliferation, differentiation, and the expression of intermediate filament proteins

Authors

  • Till B. Puschmann,

    Corresponding author
    1. Center for Brain Repair and Rehabilitation, Department of Clinical Neuroscience and Rehabilitation, Institute of Neuroscience and Physiology, Sahlgrenska Academy at the University of Gothenburg, Gothenburg, Sweden
    • Address correspondence and reprint requests to Dr Till B. Puschmann and Prof Milos Pekny, Center for Brain Repair and Rehabilitation, Department of Clinical Neuroscience and Rehabilitation, Institute of Neuroscience and Physiology, Sahlgrenska Academy at the University of Gothenburg, Box 440, SE-405 30 Gothenburg, Sweden. E-mails: till.puschmann@neuro.gu.se and milos.pekny@neuro.gu.se

    Search for more papers by this author
  • Carl Zandén,

    1. SMIT Center and Bionano Systems Laboratory, Department of Microtechnology and Nanoscience (MC2), Chalmers University of Technology, Gothenburg, Sweden
    Search for more papers by this author
    • These authors contributed equally to this work.
  • Isabell Lebkuechner,

    1. Center for Brain Repair and Rehabilitation, Department of Clinical Neuroscience and Rehabilitation, Institute of Neuroscience and Physiology, Sahlgrenska Academy at the University of Gothenburg, Gothenburg, Sweden
    Search for more papers by this author
    • These authors contributed equally to this work.
  • Camille Philippot,

    1. Center for Brain Repair and Rehabilitation, Department of Clinical Neuroscience and Rehabilitation, Institute of Neuroscience and Physiology, Sahlgrenska Academy at the University of Gothenburg, Gothenburg, Sweden
    Search for more papers by this author
  • Yolanda de Pablo,

    1. Center for Brain Repair and Rehabilitation, Department of Clinical Neuroscience and Rehabilitation, Institute of Neuroscience and Physiology, Sahlgrenska Academy at the University of Gothenburg, Gothenburg, Sweden
    Search for more papers by this author
  • Johan Liu,

    1. SMIT Center and Bionano Systems Laboratory, Department of Microtechnology and Nanoscience (MC2), Chalmers University of Technology, Gothenburg, Sweden
    Search for more papers by this author
  • Milos Pekny

    Corresponding author
    1. Center for Brain Repair and Rehabilitation, Department of Clinical Neuroscience and Rehabilitation, Institute of Neuroscience and Physiology, Sahlgrenska Academy at the University of Gothenburg, Gothenburg, Sweden
    2. Florey Institute of Neuroscience and Mental Health, Parkville, Victoria, Australia
    • Address correspondence and reprint requests to Dr Till B. Puschmann and Prof Milos Pekny, Center for Brain Repair and Rehabilitation, Department of Clinical Neuroscience and Rehabilitation, Institute of Neuroscience and Physiology, Sahlgrenska Academy at the University of Gothenburg, Box 440, SE-405 30 Gothenburg, Sweden. E-mails: till.puschmann@neuro.gu.se and milos.pekny@neuro.gu.se

    Search for more papers by this author

Abstract

Heparin-binding epidermal growth factor-like growth factor (HB-EGF), a vascular-derived trophic factor, belongs to the epidermal growth factor (EGF) family of neuroprotective, hypoxia-inducible proteins released by astrocytes in CNS injuries. It was suggested that HB–EGF can replace fetal calf serum (FCS) in astrocyte cultures. We previously demonstrated that in contrast to standard 2D cell culture systems, Bioactive3D culture system, when used with FCS, minimizes the baseline activation of astrocytes and preserves their complex morphology. Here, we show that HB-EGF induced EGF receptor (EGFR) activation by Y1068 phosphorylation, Mapk/Erk pathway activation, and led to an increase in cell proliferation, more prominent in Bioactive3D than in 2D cultures. HB-EGF changed morphology of 2D and Bioactive3D cultured astrocytes toward a radial glia-like phenotype and induced the expression of intermediate filament and progenitor cell marker protein nestin. Glial fibrillary acidic protein (GFAP) and vimentin protein expression was unaffected. RT-qPCR analysis demonstrated that HB-EGF affected the expression of Notch signaling pathway genes, implying a role for the Notch signaling in HB-EGF-mediated astrocyte response. HB-EGF can be used as a FCS replacement for astrocyte expansion and in vitro experimentation both in 2D and Bioactive3D culture systems; however, caution should be exercised since it appears to induce partial de-differentiation of astrocytes.

image

HB-EGF (heparin-binding EGF-like growth factor) was previously suggested to replace serum, a common and undefined component in primary astrocyte cultures. We show that both in standard 2D and in our newly developed Bioactive3D culture system, HB-EGF affects astrocyte morphology, proliferation, gene/protein expression and leads to partial de-differentiation of astrocytes. Thus, HB-EGF should only be used with caution as a serum replacement in astrocyte cultures.

Abbreviations used
B3D

bioactive3D

EGF

epidermal growth factor

EGFR, ErbB1

epidermal growth factor receptor

FCS

fetal calf serum

GFAP

glial fibrillary acidic protein

GS

Glul, glutamate synthetase

HB-EGF

heparin-binding EGF-like growth factor

MMP

matrix metalloprotease

RT-qPCR

quantitative reverse transcription real-time PCR

Standard in vitro protocols for expansion, maintenance, and experimentation on CNS cells such as astrocytes, foresee the addition of fetal calf serum (FCS) to the culture medium. FCS constitutes a highly undefined culture medium additive with high batch-to-batch variability in composition, likely to influence astrocyte properties. To obtain more consistent cell cultures, it was proposed that Heparin-binding epidermal growth factor-like growth factor (HB-EGF) can replace FCS in astrocyte cultures and that HB-EGF is necessary and sufficient to support astrocyte cultures in vitro (Foo et al. 2011).

The hypoxia-inducible EGFR ligand HB-EGF (Oyagi et al. 2011) belongs to the epidermal growth factor (EGF) family containing, for example, transforming growth factor alpha, amphiregulin, beta-cellulin, epiregulin, and heregulin (Higashiyama et al. 1991; Dreux et al. 2006). HB-EGF binds and activates two subtypes of ErbB receptors (Elenius et al. 1997), ErbB1 (EGFR) and ErbB4 (Citri and Yarden 2006). The expression pattern of HB-EGF in the developing brain suggests the involvement of HB-EGF in the maturation of neurons and glial cells (Kornblum et al. 1999). Furthermore, more than 60% of HB-EGF deficient mice die within the first post-natal week (Yamazaki et al. 2003). The effects of HB-EGF in vivo include improved neuro-protection after brain ischemia (Jin et al. 2002, 2004), improved wound healing (Shirakata et al. 2005; Tolino et al. 2011) and functional recovery after stroke (Sugiura et al. 2005). Astrocyte migration is increased by HB-EGF in an in vitro scratch wound model only when used with a co-factor, insulin-like growth factor 1 (IGF-1) (Faber-Elman et al. 1996). HB-EGF functions as a potent mitogen for fibroblasts, smooth muscle cells, and keratinocytes (Higashiyama et al. 1993). In zebra-fish retinas, HB-EGF induces de-differentiation of Müller glia into multipotent progenitors (Wan et al. 2012).

The importance of HB-EGF during development and in the adult as well as its potential as a replacement for FCS in astrocyte cultures, make HB-EGF a highly interesting molecule to study in astrocyte cell culture systems. Here, we have investigated the effect of HB-EGF on astrocyte morphology, proliferation, cytoskeletal protein expression, and transcriptional regulation in standard 2D cell cultures and in a newly developed 3D astrocyte culture system, Bioactive3D, in which the undesired astrocyte baseline activation is minimized (Puschmann et al. 2013bb). We report that HB-EGF stimulates cortical astrocyte proliferation, induces morphological changes toward radial glia–like phenotypes as well as changes in cytoskeletal and progenitor cell marker protein expression, and modifies Notch pathway mRNA expression.

Materials and methods

Bioactive3D/nanofiber preparation

The solutions for electrospinning were prepared by mixing 11 wt% biocompatible polyether-based polyurethane resin in a 60 : 40 mixture of tetrahydrofuran and N,N-dimethylformamide. The solution was mixed for 24 h and transferred to a syringe with a metal 21 G cannula for electrospinning as described before (Liang et al. 2007; Puschmann et al. 2013ba,b). The electrospinning process parameters were as follows: feeding rate: 2 mL/h: potential: +18 kV, distance from the nozzle tip to the collector: 18 cm; nanofiber diameter 1350 ± 250 nm (mean ± SD).

Bioactive3D and 2D coating

Nanofibers were sterilized with 70% ethanol, washed in dH2O. Bioactive3D scaffolds and 2D culture plates were incubated with poly-l-ornithine (10 μg/mL in dH2O with 285 μL/cm2 surface) for 2 h followed by three wash steps in dH2O and subsequently with laminin (5 μg/mL in Dulbecco's phosphate buffered saline with 285 μL/cm2 surface) overnight. All incubation steps were conducted in a humid atmosphere at 37°C containing 5% CO2. Bioactive3D-coated nanofibers for astrocyte cultures are now available from 3Dtro AB at www.3Dtro.com.

Astrocyte enriched cultures

All mice were housed and bred in a barrier facility of the University of Gothenburg, and experiments were conducted according to protocols approved by the Ethics Committee of the University of Gothenburg. Where applicable the ARRIVE guidelines for animal experimentations have been followed. Brains from post-natal day three mice (both sexes) of C57Bl6-129Sv-129Ola mixed genetic background were dissected under sterile conditions. The cortex was isolated, meninges removed, and tissue chopped in Dulbecco's Modified Eagle Medium (DMEM) with a scalpel blade. The tissue was incubated for 20 min at 37°C in trypsin solution (1 mg DNase1, 25 mg trypsin in 100 mL DMEM, sterile filtered). After 20 min, DMEM containing 10% FCS was added to stop trypsination and cells were pelleted at 235 g, for 5 min at 21°C. Cell pellets were re-suspended in DMEM supplemented with 10% FCS (Cat# SH30071.03, batch# ATE32026; HyClone, Thermo Scientific, South Logan, UT, USA), penicillin (100 U/mL) and l-glutamine (2 mM) (culture medium). The contamination of these cultures by non-astrocytes is minimal as previously described (Stahlberg et al. 2011). These cultures were expanded in DMEM containing 10% FCS for 8–10 days before being passaged and used in HB-EGF treatment experiments.

HB-EGF treatment of astrocyte cultures

Previously expanded un-passaged primary astrocyte cultures were transferred onto 2D plates or onto Bioactive3D plate inserts (www.3Dtro.com). Astrocytes were incubated with or without 10 ng/mL HB-EGF in neurobasal medium containing 1xB27, penicillin (100 U/mL) and l-glutamine (2 mM). For immunocytochemistry, experimentation cells were seeded with a seeding density of 35 k cells/well for 24-well plates (2D) and 45 k cells/well for 24-well plate inserts (Bioactive3D); for western blot analyses, cells were seeded with a seeding density of 750 k cells/well for 60 mm dishes (2D) and 1000 k cells/well for 60 mm dish inserts (Bioactive3D).

Immunocytochemistry

Astrocytes were fixed with 4% paraformaldehyde followed by permeabilization with 0.1% Triton-X 100 in phosphate buffered saline (PBS). Unspecific binding sites were blocked in PBS containing 2% normal donkey serum and 2% normal goat serum (blocking solution). After blocking, cells were incubated for 30 min with primary antibodies diluted in blocking solution, washed three times in PBS followed by secondary antibody incubation diluted in blocking solution. Primary antibodies used were: rabbit anti-glial fibrillary acidic protein (GFAP) 1 : 2000 (DAKO, Z0334), rabbit anti-vimentin 1 : 500 (ab45939; Abcam, Cambridge, UK), mouse anti-nestin 1 : 1000 (611658; BD Bioscience, San Jose, CA, USA), mouse anti-glutamine synthetase 1 : 250 (MAB302; Millipore Corporation, Bedford, MA, USA). Nuclei were counter-stained with 4',6-Diamidin-2-phenylindol (DAPI) (Sigma, St Louis, MO, USA). Secondary antibodies used were: Alexa Fluor 488-conjugated donkey anti-rabbit (A-21206), Alexa Fluor 594-conjugated goat anti-rabbit (A-11012), and goat anti-mouse [A-11005; all Invitrogen and all 1 : 1000 (Invitrogen, Carlsbad, CA, USA)]. Images were taken with a Leica DMI 600B microscope (Leica Microsystems, Wetzlar, Germany).

EdU incorporation

The uptake of 5-ethynyl-2′-deoxyuridine (EdU) and its incorporation into cell nuclei was detected using the click-iT kit (Invitrogen) according manufacturer's instructions. In brief, after 7 days in neurobasal medium alone or with 10 ng/mL HB-EGF and/or 10% FCS, cell cultures were incubated with 10 μM EdU in the respective neurobasal culture medium for 3 h followed by 20 min cell fixation in 4% paraformaldehyde. Cells were permeabilized as described above and EdU was detected by using the click-iT reaction kit (Invitrogen).

Evaluation of cell morphology

Contours of cells labeled with glutamine synthetase were used for determination of form factor. The form factor was defined as 4pi(area/perimeter2). Thus, perfectly round cells have a form factor of 1, whereas more elongated or stellate cells have lower form factor values (Soll et al. 1988).

Protein extraction

Total protein of astrocyte cultures prepared from mice of C57Bl6-129Sv-129Ola mixed genetic background was obtained by adding protein lysis buffer to cell culture plates for 2D cultures, or by submerging the Bioactive 3D inserts in an Eppendorf-tube containing lysis buffer. The protein lysates were sonicated for 30 s with an amplitude of 14 μm and stored at −80°C until further use. The protein lysis buffer was composed of 20 mM Tris-HCL pH 7.5, 150 mM NaCl, 1 mM EDTA, 0.1% v/v Triton-X 100, 1 tablet protease inhibitor/10 mL lysis buffer (Roche Molecular Biochemicals, Indianapolis, IN, USA), 1 tablet phosphatase inhibitor/10 mL lysis buffer (Roche). The protein concentrations within the protein lysates were measured using Bio-Rad DC protein assay (Bio-Rad Laboratories, Hercules, CA, USA) and compared to bovine serum albumin standards of known concentration.

Western blot analyses

Sodium dodecyl sulfate–polyacrylamide gel electrophoresis was performed using Any-kD gels (Bio-Rad) with Tris-glycine running buffer (Bio-Rad). Blotting was conducted on polyvinylidene difluoride membranes (Immobilon, Millipore) at 100 V for 1.5 h. Transfer buffer composed of Tris-base (2.9 g/L), glycine (14.4 g/L) and MeOH (20% v/v) in dH2O. A total amount of 15 μg of protein lysate per lane was loaded. Unspecific binding was blocked using 3% bovine serum albumin in Tris-buffered saline containing 0.1% TWEEN for 1 h at 21°C. Primary antibody incubation was done overnight at 4°C followed by incubation with secondary horseradish peroxidase (HRP)-linked anti-rabbit, anti-mouse or anti-rat antibodies (Cell Signaling Technology, Beverly, MA, USA) for 45 min at 21°C. Secondary HRP antibodies were detected using chemiluminescence detection kit (Luminogen PS-3; GE Healthcare, Little Chalfont, UK) and a LAS-3000 luminescent image analyzer (Fujifilm, Tokyo, Japan), the signal was transformed into arbitrary units using MulitGauge software (Fujifilm). The following antibodies were used in 1 : 1000 dilution: mouse anti-GFAP (MAB360; Millipore), mouse anti-nestin (611658; BD Bioscience), rabbit anti-vimentin (ab45939; Abcam), rabbit anti-MAPK(Erk1/2) (#9102; Cell Signaling), rabbit anti-phospho-Mapk/Erk1/2 (Thr202/Tyr204;#9101; Cell Signaling), rabbit anti-phospho-EGFR (Y1068;ab5650; Abcam), rat anti-EGFR (ab231; Abcam). HRP-conjugated beta-actin (#5125; Cell Signaling) was used in 1 : 3000 dilution. Western blot intensity was measured using ImageJ 1.47v software (National Institutes of Health, Bethesda, MD, USA). For nestin western blot analysis, the values were normalized to beta actin signal. Values normalized for each experiment are presented in arbitrary units. Data were analyzed by two-way anova and Sidak-Bonferroni post hoc test using GraphPad Prism 6.0 software (GraphPad Software, Inc., La Jolla, CA, USA). Differences were considered significant at p < 0.05.

RNA extraction

Total RNA was extracted and purified with the RNeasy Micro Kit (Qiagen, Valencia, CA, USA) without the DNaseI digestion step, according to the manufacturer's instructions. Extracted RNA was stored at −80°C until use.

Quantitative RT-PCR

CDNA synthesis by reverse transcription was performed using TATAA Grandscript cDNA Synthesis Kit (TATAA Biocenter AB, Gothenburg, Sweden) and CFX 96 Real-Time System instrument (Bio-Rad). Reverse transcription was performed 5 min at 25°C, 30 min at 42°C, followed by 5 min at 85°C. Reaction volume of RT-qPCR was 10 μL, containing 5 μL TATAA SYBR Grandmaster Mix (TATAA Biocenter) and a final concentration of 0.4 μM of each primer (synthesized by Sigma-Aldrich, St. Louis, MO, USA). RT-qPCR were performed with a LightCycler 480 instrument (Roche). The amplification protocol consisted of 3 min initial denaturation at 95°C followed by 45 cycles of 5 s at 95°C, 20 s at 60°C and 20 s at 72°C, followed by a melt curve analysis from 60°C to 95°C to determine correct product sizes of RT-qPCR amplification products. The primer sequences are listed in Table S1. Primers for the reference genes GAPDH and PGK1 were obtained from TATAA Biocenter AB, Gothenburg, Sweden.

Affymetrix data

The data set was generated in the context of a previous study (Puschmann et al. 2013b) and analyzed in this study. Raw Affymetrix data were normalized using the Robust Multichip Average (RMA) method. Normalized data were analyzed using the Multiexperiment viewer (MeV v4.8) software (open source Java tool for genomic data analysis) (Saeed et al., 2006). Significance was set as p < 0.05, Student's t-test, with Welch approximation (assuming unequal group variance).

Statistical analysis

All results are presented as mean ± SEM. Grouped data for cells/mm2, EdU counts, form factor, cell area were analyzed by one-way or two-way anova, followed by Tukey or Sidak-Bonferroni post hoc test, respectively, using GraphPad Prism 6.0 software. Gene expression data were normalized against the stably expressed reference genes GAPDH and PGK1. Gene expression data were analyzed by two-way anova, followed by Sidak-Bonferroni post hoc test, using GraphPad Prism 6.0 software. Differences were considered significant at p < 0.05.

Results

HB-EGF induces stronger proliferative response in astrocytes maintained in Bioactive3D than in standard 2D culture system

First, we assessed the proliferative effects of HB-EGF on cortical 2D and Bioactive3D mouse astrocyte cultures. An increase in cell numbers after 7 days of treatment with HB-EGF was observed in both 2D and Bioactive3D grown astrocytes (Fig. 1a). Interestingly, the increase in cell number after HB-EGF treatment was higher in Bioactive3D grown astrocytes (569 ± 28%) compared to 2D grown astrocytes (238 ± 54%). Next, the proportion of cells positive for the exogenous proliferation marker EdU was determined following a 3 h pulse. Consistent with the total cell number data, after treatment with HB-EGF, a higher percentage of cells were EdU positive in Bioactive3D than in 2D cultures (Fig. 1b). Hence, the addition of HB-EGF to primary cortical mouse astrocytes cultured in 2D and Bioactive3D resulted in an increase in proliferation in both culture systems, with a stronger response to HB-EGF in Bioactive3D. Proliferation of Bioactive3D grown cortical astrocytes in the presence of HB-EGF, FCS or both, was comparable (Fig. 1c).

Figure 1.

Heparin-binding epidermal growth factor-like growth factor (HB-EGF) increases cell proliferation with a stronger effect on Bioactive3D grown astrocytes. (a) Number of cells after 7 days in HB-EGF treated and untreated 2D and Bioactive3D (B3D) cultures. HB-EGF has a mitogenic effect on cortical astrocytes in both culture systems, albeit a stronger one in B3D. (b) HB-EGF increases the number of 5-ethynyl-2′-deoxyuridine (EdU) positive cells in both culture systems with the percentage of EdU positive cells being higher in Bioactive3D than in 2D grown astrocyte cultures. (c) Proliferation of Bioactive3D grown cortical astrocytes in the presence of HB-EGF, fetal calf serum (FCS) or both HB-EGF and FCS, was comparable. Results show mean ± SEM of = 3 experiments for (a), anova F1,8 = 103.2, p < 0.0001; *p < 0.05, **p < 0.01, ***p < 0.001 (Sidak-Bonferroni), = 3 experiments for (b), anova F1,8 = 23.3, p < 0.01; *p < 0.05 (Sidak-Bonferroni), = 4 experiments for (c), anova F3,12=5.4, p < 0.02; *p < 0.05 (Tukey).

HB-EGF activates the Mapk/Erk pathway downstream of the EGF receptor

Since EGF receptor activation by HB-EGF leads to activation of the Mapk/Erk pathway in cultured neurons (Kornblum et al. 1999), we investigated whether this signaling cascade is activated also in cortical mouse astrocytes. Western blot analysis showed that following treatment with HB-EGF, the EGF receptor was phosphorylated at Y1068 (Fig. 2a) and that the EGF receptor activation leads to the activation of the Mapk/Erk pathway (Fig. 2b). This signaling cascade was activated in both 2D and Bioactive3D grown astrocytes.

Figure 2.

Western blot analysis of Heparin-binding epidermal growth factor-like growth factor (HB-EGF) treated 2D and Bioactive3D grown astrocytes. (a) Treatment with HB-EGF for 5 min phosphorylates EGF receptor (EGFR) at Y1068 in both 2D and Bioactive3D astrocyte cultures. (b) Treatment with HB-EGF for 7 h activates the MapK/Erk pathway in both 2D and Bioactive3D astrocyte cultures. Results show mean ± SEM of = 3 experiments for (b), anova F1,8 = 23.9, p < 0.01; *p < 0.05 (Sidak-Bonferroni).

HB-EGF affects astrocyte morphology

Astrocytes grown in Bioactive3D assume a more complex morphology compared to polygonal shaped astrocytes in 2D culture systems (Puschmann et al. 2013b) and Fig. 3a. After treatment with HB-EGF for a minimum of 5 days, distinct morphological changes were observed as visualized by immunocytochemistry for glutamine synthetase (GS), an enzyme distributed within astrocyte cytoplasm (Fig. 3a). Form factor analysis of individual astrocytes revealed an increase in morphological complexity of both 2D and Bioactive3D grown astrocytes after HB-EGF treatment (Fig. 3b). A large proportion of astrocytes in both culture systems exhibited radial glia-like features such as a bipolar shape, retracted lamellopodia and long cellular protrusions (Fig. 4a, b; Fig. 5b). After HB-EGF treatment, the total area covered by individual astrocytes decreased in 2D but not in Bioactive3D grown cells (Fig. 3c).

Figure 3.

Changes in astrocyte morphology after 7 day Heparin-binding epidermal growth factor-like growth factor (HB-EGF) treatment. (a) Images of astrocytes visualized by immunocytochemical staining for glutamine synthetase (GS) reveal more complex cell morphology of HB-EGF treated 2D grown astrocytes. Astrocytes in Bioactive3D (B3D) show a complex morphology already prior to HB-EGF treatment. Scale bar = 30 μm. (b) Form factor analysis reveals increased morphological complexity after HB-EGF treatment of 2D and Bioactive3D grown astrocytes. Already prior to HB-EGF treatment, 3D grown astrocytes show increased morphological complexity compared to 2D grown astrocytes. (c) Cell area measurements demonstrate a decrease in the area covered by HB-EGF treated astrocytes in 2D cultures. Already prior to HB-EGF treatment, 3D grown astrocytes cover a smaller area compared to 2D grown astrocytes. Results show mean ± SEM of n = 12–16 individual cells per culture and treatment for (b) anova F1,52 = 40.5, p < 0.0001; *p < 0.05, **p < 0.01, ***p < 0.001 (Sidak-Bonferroni) and (c) anova F1,52 = 11.1, p < 0.01; *p < 0.05, **p < 0.01, ***p < 0.001 (Sidak-Bonferroni).

Figure 4.

Heparin-binding epidermal growth factor-like growth factor (HB-EGF) treatment induces nestin protein expression and radial glia-like cell morphology. (a) Astrocyte intermediate filament bundles were visualized by immunocytochemical staining for nestin (green) and vimentin (red), and nuclei were counterstained with DAPI (blue). The majority of cells express nestin after HB-EGF treatment, and many astrocytes in 2D and Bioactive3D (B3D) assume radial glia-like morphology (white arrowheads). (b) Astrocyte intermediate filament bundles were visualized by immunocytochemical staining for glial fibrillary acidic protein (GFAP) (green), astrocytes were visualized with antibodies against glutamate synthetase (GS) (red), and nuclei were counterstained with DAPI (blue). Prior to HB-EGF treatment, GS and GFAP immunostaining reveals polygonal shape of 2D grown astrocytes. After HB-EGF treatment, many astrocytes in 2D and Bioactive3D assume radial glia-like morphology; examples are indicated by white arrowheads. Scale bar = 30 μm.

Figure 5.

Western blot analysis and immuno-cytochemical detection of intermediate filament proteins glial fibrillary acidic protein (GFAP), vimentin, and nestin. (a) Western blot analysis demonstrates that nestin, but not GFAP or vimentin, is up-regulated after Heparin-binding epidermal growth factor-like growth factor (HB-EGF) treatment in both 2D and Bioactive3D. (b) Visualization of GFAP (green) and glutamate synthetase (GS) (red) by immunocytochemical staining reveals that after HB-EGF treatment, astrocytes form foci of either GFAP positive or GFAP negative cells both in 2D and Bioactive3D (B3D) cultures. Cell nuclei were counterstained with DAPI (blue). Results show mean ± SEM of = 3 experiments for 2D control and 2D HB-EGF,= 2 experiments for 3D control and 3D HB-EGF in (a) anova F1,6 = 18.6, p < 0.01; *p < 0.05 (Sidak-Bonferroni).

HB-EGF induces the expression of intermediate filament protein nestin but not vimentin or GFAP

The expression of intermediate filament proteins GFAP, vimentin, and nestin is known to be up-regulated under culture conditions in astrocytes as well as after injury in vivo [for review see (Pekny and Nilsson 2005)]. In addition, nestin expression has been reported to be a progenitor cell marker and to be induced in de-differentiating astrocytes (Yang et al. 2009; Shimada et al. 2012). HB-EGF treatment induced the expression of nestin in 2D and Bioactive3D grown astrocytes as shown by immunocytochemistry (Fig. 4a). The vast majority of astrocytes expressed nestin after HB-EGF treatment, whereas vimentin and GFAP were expressed in a subpopulation of cultured astrocytes (Fig. 4a, b). Western blot analysis further confirmed the increased expression of nestin after treatment with HB-EGF in both 2D and Bioactive3D grown astrocytes (Fig. 5a), while the expression levels of GFAP and vimentin were unaffected (Fig. 5a).

The progeny of HB-EGF-induced proliferating cells stay GFAP positive or negative

In HB-EGF treated astrocyte cultures, the immunocytochemical analysis of GFAP showed neighboring astrocytes forming foci that were either strongly GFAP positive or GFAP negative (Fig. 5b). This phenomenon was observed in both 2D and Bioactive3D grown astrocytes and it most likely resulted from a clonal expansion of GFAP positive and GFAP negative cells.

The expression of HB-EGF signaling pathway related genes is affected by Bioactive3D

Data obtained using Affymetrix gene expression arrays in the context of our previous study (Puschmann et al. 2013b) showed that HB-EGF mRNA levels were increased to 138% (p < 0.05) in Bioactive3D grown astrocytes (Fig. 6a). We found no differences in HB-EGF receptor EGFR mRNA levels between 2D and Bioactive3D grown astrocytes. However, ErbB4 mRNA was down-regulated to 45% in Bioactive3D compared to 2D (p < 0.05) (Fig. 6a), which was also confirmed by RT-qPCR, where mRNA was down-regulated to 36% (Table 1). The cleavage of HB-EGF from pro-HB-EGF is performed by matrix metalloproteases (MMPs) and members of the disintegrin and metalloproteinases (ADAMs) family. Interestingly, we found the MMP inhibitor TIMP-1 mRNA to be up-regulated to 450% in Bioactive3D grown astrocytes (p < 0.05). Of the MMPs known to cleave pro-HB-EGF, only MMP-9 mRNA showed an increase by 16% (p < 0.001) in Bioactive3D compared to 2D grown astrocytes.

Figure 6.

Schematic drawing of the Heparin-binding epidermal growth factor-like growth factor (HB-EGF) activation and Notch signaling pathways in 2D and Bioactive3D. (a) Components of the HB-EGF signaling cascade are depicted. Pro-HB-EGF is cleaved by metalloproteases (MMPs) resulting in an extracellular soluble fraction of HB-EGF (sHB-EGF) and an intracellular fraction. Pro-HB-EGF cleavage into sHB-EGF by MMPs is inhibited by TIMP-1. Relative mRNA expression (determined by Affymetrix) of Bioactive3D grown astrocytes is compared to 2D grown astrocytes. In Bioactive3D, mRNA of HB-EGF is up-regulated to 138% and MMP-9 to 116%, while ErbB4 is down-regulated to 45% and TIMP-1 up-regulated to 450%. The up-regulation of HB-EGF pathway inhibitor, TIMP-1, together with the down-regulation of the HB-EGF receptor ErbB4 indicates that the HB-EGF pathway is less active in the Bioactive3D cell culture system. (b) Schematic drawing of the HB-EGF-Notch signaling cascade. Relative mRNA expression (determined by RT-qPCR) of HB-EGF treated versus untreated cultures is shown. Values for 2D cultures are given in green and for Bioactive3D cultures in red. After 7 days of incubation with HB-EGF, EGFR mRNA is up-regulated to 126% in 2D and to 225% in Bioactive3D cultures. The Notch activator Ascl1 mRNA is up-regulated to 277% in 2D and 189% in Bioactive3D, and the Notch pathway activator Dll1 mRNA is up-regulated to 150% and 151% in 2D and Bioactive3D, respectively. Notch pathway suppressor Hes1 mRNA is down-regulated to 37% in Bioactive3D only.

Table 1. Relative mRNA expression in Bioactive3D and 2D grown astrocytes, treated or untreated with HB-EGF
  2D HB-EGF/2D con3D HB-EGF/3D con3D con/2D con3D HB-EGF/2D HB-EGF
  1. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

  2. Each value is an average obtained from astrocyte cultures derived from six individual mice and three replicas per each mouse. Significant values are highlighted in bold. Data were analyzed using two-way anova, p < 0.05; followed by Sidak-Bonferroni post hoc tests.

GFAP 0.48 *** 0.38 **** 0.910.73
Vim 0.780.710.74 0.67 *
Nes 1.111.021.321.22
HB-EGF 0.641.040.701.15
EGFR 1.26 ** 2.25 **** 0.79 ** 1.42 ***
ErbB4 0.841.64 0.36 **** 0.69
Sox2 0.841.04 0.68 ** 0.84
Pax6 0.61 *** 0.77 * 0.60 **** 0.76 *
Ascl1 2.77 **** 1.89 *** 1.69 ** 0.98
Jag1 0.72 0.43 ** 2.061.23
DLL1 1.5 * 1.51 ** 1.54 1.56 *
Notch1 0.840.980.850.98
Hes1 0.75 0.37 *** 1.190.59
Glul 0.72 0.42 * 0.980.57

HB-EGF treatment affects the expression of cytoskeletal and Notch signaling pathway genes

Next, we investigated the effect of HB-EGF on the expression of intermediate filament protein mRNA in both 2D and Bioactive3D cultures. Surprisingly, nestin mRNA expression did not change after HB-EGF activation of the EGFR (Table 1) even though the nestin protein levels were up-regulated (Fig. 5a). GFAP mRNA levels were reduced after the HB-EGF treatment (Table 1), while the GFAP protein expression in astrocyte cultures with and without HB-EGF treatment was comparable. These findings indicate prominent post-transcriptional regulation of nestin and GFAP.

Nestin re-expression occurs in reactive astrocytes after neurotrauma, stroke, or neurodegenerative disease (Pekny and Nilsson 2005) and in cells that de-differentiate in response to injury (Yang et al. 2009, 2011). Müller glia, for example, de-differentiate into multi-potent progenitor cells after treatment with HB-EGF in a process mediated by Notch signaling (Wan et al. 2012). Since nestin protein expression was increased in astrocytes after HB-EGF treatment, we investigated the effects of HB-EGF on mRNA expression levels of genes involved in the Notch signaling pathway. In response to HB-EGF, we detected changes in mRNA level of regeneration-associated genes Ascl1, Dll1, Hes1, and Pax6, members of the Notch signaling pathway, (Fig. 6b, Table 1) and the astrocyte marker glutamine synthetase (Glul). After HB-EGF treatment, Glul expression was down-regulated to 42% in Bioactive3D only. The EGFR mRNA level was increased to 126% and to 225% in 2D and Bioactive3D, respectively, Ascl1 mRNA was increased to 277% and to 189% in 2D and Bioactive3D, respectively. Interestingly, the Notch ligand Dll1 was up-regulated to 150% and 151% in 2D and Bioactive3D, respectively, whereas Hes1, a suppressor of Dll1 and of Notch signaling (Kobayashi et al. 2009; Kobayashi and Kageyama 2010), was down-regulated to 37% in Bioactive3D only. Furthermore, the Notch ligand Jag1 was down-regulated to 43% in Bioactive3D, but not in 2D. Together, these data indicate that the Notch signaling pathway was modulated by HB-EGF in cortical astrocytes in culture and that this activation was more prominent in astrocytes grown in Bioactive3D.

Discussion

Astrocytes undergo complex morphological and functional changes when activated in situations such as neurotrauma, stroke, or neurodegenerative diseases (Wilhelmsson et al. 2006; Sofroniew 2009; Parpura et al. 2012; Verkhratsky and Butt 2013). Characteristic changes of reactive astrocytes are hypertrophy of astrocyte processes, up-regulation of intermediate filament (also known as nanofilament) protein gene expression and the release of soluble factors such as HB-EGF. As the effects of HB-EGF on cortical astrocytes regarding morphology, proliferation, intermediate filament protein expression and activation of specific pathways, are largely unknown, we investigated these cellular responses to HB-EGF utilizing both standard 2D astrocyte cultures and a newly developed nanofiber based 3D cell culture system, Bioactive3D.

HB-EGF regulation of astrocyte proliferation

Cell culture systems utilized to successfully expand primary astrocytes for in vitro experimentation have to provide the environment that maintains viable cells and supports their proliferation and metabolic and other functional activities. In most standard 2D cell culture systems, the addition of FCS to the culture medium is essential for providing growth and survival factors. Recently, we developed a 3D cell culture system that maintains astrocytes in a minimal reactive state with reduced proliferation in the culture medium that contains FCS (Puschmann et al. 2013b). Data presented here show that the replacement of FCS by HB-EGF has proliferation supporting effects on 2D and Bioactive3D primary mouse astrocyte cultures and confirms that it is sufficient for cell survival (Foo et al. 2011). We showed that HB-EGF activation of EGFR at Y1068 induced Mapk/Erk pathway activation and increased cell proliferation. Interestingly, the HB-EGF induced proliferative response of astrocytes was stronger in Bioactive3D grown astrocytes, possibly because of their lower baseline reactivity.

Changes in astrocyte morphology induced by HB-EGF and Bioactive3D

Morphological changes of astrocytes in vitro correlate with changes in metabolism and receptor expression (Zagami et al. 2005; Lau et al. 2011). Therefore, it is desirable to design cell culture systems that provide a cell culture environment supporting the complex in vivo-like morphology of astrocytes. Here, we show that this can be achieved by either utilizing a suitable 3D cell culture system (Puschmann et al. 2013b) or to some extent even by replacing FCS with HB-EGF. However, the treatment with HB-EGF, a growth factor released after injury in vivo (Marikovsky et al. 1993), triggered morphological changes in 2D and Bioactive3D grown astrocytes toward a radial glia-like phenotype. Such changes in astrocyte phenotype were observed in de-differentiating astrocytes in vitro after addition of scratch-wounded astrocyte conditioned medium to the cell cultures (Yang et al. 2011). Therefore, it is conceivable that the HB-EGF-induced morphology reflects a progenitor cell-like phenotype rather than the phenotype of in vivo-like mature astrocytes.

The effect of HB-EGF on intermediate filament (nanofilament) protein expression

The expression of nestin, a marker of reactive astrocytes (Clarke et al. 1994), neural progenitor cells (Shimada et al. 2012), and de-differentiated astrocytes (Yang et al. 2009, 2011), was increased in 2D and Bioactive3D cultures after HB-EGF treatment. However, the expression of reactive astrocyte markers GFAP and vimentin did not increase after a weeklong exposure to HB-EGF. In vivo stab wound experiments showed that nestin is transiently expressed in distinct astrocyte populations at the injury site in the brain for 1–2 weeks after injury (Lin et al. 1995; Krum and Rosenstein 1999; Pekny et al. 1999), which is compatible with a phenotype of transiently de-differentiated astrocytes. Such an effect might be mediated by growth factors such as vascular endothelial growth factor (Krum and Rosenstein 1999). This concept was supported by data suggesting that after injury GFAP positive astrocytes can revert to a progenitor cell type (Buffo et al. 2008).

The effect of Bioactive3D on the HB-EGF signaling pathway

Affymetrix and/or RT-qPCR demonstrated lower expression of HB-EGF receptors (EGFR, ErbB4) and higher expression of HB-EGF pathway inhibitor TIMP-1 in Bioactive3D compared to 2D. The up-regulation and activation of the EGFR induces astrocyte activation in the optic nerve glaucoma model (Liu and Neufeld 2003). Therefore, reduced expression of EGFR mRNA and protein in Bioactive3D further corroborates our findings that cortical astrocytes, when grown in Bioactive3D, are less reactive. Minimized baseline reactivity of astrocytes in Bioactive3D culture is likely to allow to measure responses to stimuli that might not be detected in 2D culture systems. The observed increase in EGFR mRNA after HB-EGF treatment and the higher increase in proliferation in Bioactive3D, compared to 2D, demonstrate that 3D-grown astrocytes are fully responsive to activating stimuli. Therefore, Bioactive3D culture system seems to constitute an ideal platform for in vitro experimentation.

The expression of Notch signaling genes after HB-EGF administration

In both culture systems, mRNA expression of Notch pathway genes was affected, albeit differentially between the two culture systems. In vivo, Notch signaling triggers reactive gliosis and expression of nestin in astrocytes after ischemia (Marumo et al. 2013). Notch signaling from astrocytes to neural stem cells inhibits stem cell differentiation and depends on the intermediate filament system of astrocytes (Breunig et al. 2007; Wilhelmsson et al. 2012). Notch signaling was also suggested to enhance nestin expression in progenitor cells in human gliomas (Shih and Holland 2006). In line with our findings of increased nestin protein expression after HB-EGF treatment, we found expression of genes involved in the Notch signaling cascade to be activated in both HB-EGF treated 2D and Bioactive3D astrocyte cultures. The mRNA expression of the Notch ligand Dll1 was up-regulated in both systems, however the Dll1 suppressor Hes1 was down-regulated only in Bioactive3D. Increased ligand expression concomitant with decreased ligand suppressor expression possibly lead to an amplified signaling cascade in Bioactive3D compared to 2D cultures after HB-EGF treatment. Together with the morphological changes toward the radial glia-like phenotype and increased nestin protein expression, the activation of Notch signaling pathway genes and the down-regulation of the astrocyte marker Glul, indicate de-differentiation of astrocytes following HB-EGF treatment.

In summary, we confirm here that replacement of FCS with HB-EGF in astrocyte culture medium is sufficient to maintain viable astrocytes. The observed effects of HB-EGF were more profound in the Bioactive3D cell culture system, most probably because of minimal baseline reactivity of 3D-grown astrocytes. We demonstrate that HB-EGF has a strong proliferative effect on astrocyte cultures and induces changes in morphology toward a radial glia-like phenotype. The increased expression of nestin after HB-EGF treatment, and modulation of the Notch pathway on a transcriptional level support this concept of HB-EGF induced astrocyte de-differentiation. Therefore, our data suggest that HB-EGF can replace FCS, however it alters astrocyte phenotype and hence should be used as a standard additive to astrocyte cultures only with caution.

Acknowledgements

This work was supported by Swedish Medical Research Council (11548), ALF Gothenburg (11392), AFA Research Foundation, Söderbergs Foundations, Sten A. Olsson Foundation for Research and Culture, Hjärnfonden, Hagströmer's Foundation Millennium, Amlöv's Foundation, E. Jacobson's Donation Fund, VINNOVA Health Program, the Swedish Stroke Foundation, the Swedish Society of Medicine, the Free Mason Foundation, Chalmers University of Technology, NanoNet COST Action (BM1002), EU FP 7 Program EduGlia (237956), EU FP 7 Program TargetBraIn (279017). TP, CZ, JL and MP own stock in 3Dtro, a biotech company that produces systems for 3D culture of astrocytes.

Ancillary