In a first step, we immunocytochemically characterized the hNSC population generated from immature pluripotent embryonic stem cells. After neural pre-differentiation, almost all cells expressed nestin and most cells (>80%) expressed both, nestin and Sox2 (Figure 1A). After withdrawal of bFGF, NSCs terminally differentiated into βIII-tubulin⁺ neurons or GFAP⁺ astrocytes (Figure 1A). As the expression of membrane-bound IFNγ receptors (2 IFNγ receptor-1 chains and 2 IFNγ receptor-2 chains) is necessary to transduce the IFNγ signal, we performed immunocytochemical labelings against both receptor chains and were indeed able to demonstrate their expression (Figure 1B). We further investigated the expression levels of both receptor chains on mRNA level by means of real-time quantitative PCR. We compared these data with those generated on a murine background and found, that murine and human NSCs did not show significant differences (Figure 1C). Data were generated with the Wicell H9 line. We also confirmed these findings by using the HUES 6 line (data not shown).

One major characteristic of the IFNγ-induced mNSC dysregulation is the coexpression of neuronal and glial markers under the influence of growth factors that normally hold NSCs in an immature and proliferating state (Figure 2A). This phenomenon is visible after a 3-days treatment with 1000 U/ml IFNγ and leads to a portion of around 39% of all cells that co-express GFAP and βIII-tubulin. The detailed characterization of this phenomenon is published elsewhere (Walter et al., 2011). However, when immature and proliferating hNSC populations under the influence of growth factors were treated with human recombinant IFNγ in identical concentrations compared to the murine situation, we were not able to detect this phenomenon (Figure 2A). These results were also confirmed on mRNA level (Figure 2B). Data were generated with the Wicell H9 line. We also confirmed these findings by using the HUES 6 line (data not shown).

As previously reported, the genes for inducible nitric oxid synthase (i-NOS) (Komatsu et al., 1996), IRF-9 (Ousman et al., 2005), c-Myc (Ramana et al., 2000), major histocompatibility complex 1 (MHC 1) (Johansson et al., 2008) and Stat 1 (Lehtonen et al., 1997) are significantly regulated in murine cells after IFNγ exposure. We now investigated the expression of these down-stream signaling genes after IFNγ exposure in hNSCs by means of real-time quantitative PCR. We found all of these genes to be significantly up-regulated after IFNγ exposure also in hNSCs (Figure 3A). As previously reported, we also found SHH to be significantly up-regulated in mNSCs after IFNγ treatment. Interestingly, we could not detect that phenomenon in hNSCs. Another IFNγ-induced phenomenon on a murine background was a strong reduction of the population size of undifferentiated mNSCs. However, we were not able to detect a significant decrease in the population size of hNSCs after IFNγ exposure (Figure 3B). Data were generated with the Wicell H9 line. We also confirmed these findings by using the HUES 6 line (data not shown).

In contrast to the extensively described up-regulation of IFNγ in animal models of human central nervous system diseases, there is only very limited information about amounts of IFNγ in human individuals suffering from neurological diseases. Therefore, we measured the concentration of IFNγ in cerebrospinal fluid (CSF) specimens, taken after lumbar punction from patients suffering from multiple sclerosis, peripheral nervous plexus affections, Alzheimer dementia, viral myelitis, or stroke. We found, that patients suffering from multiple sclerosis during an acute relapse showed a significant increase of IFNγ (three control samples vs. three MS samples). Due to the limitation of patient material, we could only screen single patients with other diseases (one patient per other indicated disease). But also here we found a clear trend toward an up-regulation of IFNγ in the CSF (Figure 3C).

The human NSCs used in this study were either derived from the human embryonic stem cell lines HUES 6 (hESC facility; Harvard University; Cambridge, MA, USA) or purchased in a neuroectodermal, predifferentiated form of the human embryonic stem cell line H9 (WiCell Research Institute, Madison, WI, USA/Invitrogen, Karlsruhe, Germany). The detailed neuronal differentiation protocol is described elsewhere (Dhara et al., 2008). Use of hESCs for derivation of neural progenitors was approved by the regulatory authorities at the Robert Koch Institute, Berlin, Germany, and derivation of neural progenitors was performed as follows: the hESCs were cultivated in knockout serum replacement (KSR) media (Cowan et al., 2004) on gelatin coated plates and a feeder layer of γ-irradiated mouse embryonic fibroblasts. Neural differentiation was performed by a protocol previously described by Lappalainen et al. (2010). Briefly, hESCs were dislodged and separated from the feeder cells and afterwards dissected into clusters that contained in between 1000 and 3500 cells. These clusters were placed in neural differentiation/proliferation media that consisted of 1:1 DMEM/F12 and Neurobasal media supplemented with 2 mM GlutaMax, 1XB27, 1XN2, 0.25% BSA, 0.1 mM non-essential amino acids, 25 U/ml penicillin/streptomycin (all products Invitrogen, Karlsruhe, Germany) and 20 ng/ml bFGF (Tebu-bio, Le Perray en Yvelines Cedex, France). Cells were propagated for 8–10 weeks with a medium change every 2–3 days, the formed clusters were dissected once per week. After this time period, the vast majority of cells expressed NSC markers like nestin and Sox 2. Cells were then frozen. For experiments, human NSC were thawn and single cells were seeded on culture surfaces that were coated with poly-L-ornithine (PLO; 0.001%; Sigma-Aldrich, Munich, Germany) and fibronectin (20 μg/ml; BioPur, Bubendorf, Switzerland). In parallel, we used the neuroectodermally pre-differentiated hNSC line (hNSC H9) for comparison. HNSC H9 cells were cultivated and propagated in the same way and in the same media as described above. Human NSCs derived from the cell line H9 were cultivated as free floating clusters and were dissected to single cells for experiments. The attached cells were either treated with human recombinant IFNγ (1000 U/ml; Millipore, Billerica, USA) or kept under control conditions for 3 days. This amount is equal to 100 ng/ml recombinant protein (murine and human). For differentiation experiments, cells had a day of recovery after seeding, then bFGF was withdrawn from the cells and differentiation was performed up to 3 months.

For immunocytochemistry, cells were seeded on coated cover slips (VWR International, Darmstadt, Germany). After 3 days under the influence bFGF (20 ng/ml both Tebu-bio) and under different treatments as indicated, the cells were fixed with 4% PFA (Roti-Histofix, Carl Roth, Karlsruhe, Germany) for 15 min at room temperature. Cells were blocked for 30 min at room temperature with 1 fold Roti-Immuno-Block containing 0.25% Triton X-100 for cell wall permeabilization (Carl Roth, Karlsruhe, Germany). Primary antibodies used at 4°C overnight were against βIII-tubulin (Tuj1; 1:500; R&D Systems, Minneapolis, USA), glial fibrillary acid protein (GFAP) (1:500; Dako, Hamburg, Germany), Sox2 (1:50; R&D Systems, Minneapolis, USA), IFNγ-R1 (1:500 Santa Cruz Biotechnology, Heidelberg, Germany), IFNγ-R2 (1:500; Santa Cruz Biotechnology, Heidelberg, Germany) and nestin (1:200; Covance, Munich, Germany). For detection of primary antibodies, fluoresceine-isothiocyanate (FITC; 1:500; Millipore, Billerica, USA) or indocarbocyanine (Cy3; 1:800; or Cy5; 1:200; Millipore, Billerica, USA) coupled secondary antibodies were used. The first and secondary antibodies were diluted in 1-fold Roti-Immuno-Block without Triton X-100 (Carl Roth, Karlsruhe, Germany). For visualization of cell nuclei cells were co-stained with DAPI (Invitrogen, Karlsruhe, Germany). For negative controls, primary antibodies were omitted in each experiment.

To analyze the population size of hNSCs, the optical density, indicative of conversion of 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT; Sigma-Aldrich) into formazan crystals which takes place in live cells only, was determined after IFNγ treatment at indicated concentrations. An OD value of 0.5 represents approximately 50,000 and an OD value of 1.0 represents approximately 100,000 live NSCs. The population size was measured after the indicated time intervals. Therefore, MTT (Invitrogen, Karlsruhe, Germany) was added to the cell culture media at a final concentration of 0.5 mg/ml after 5 h of incubation media was skipped and DMSO (Invitrogen, Karlsruhe, Germany) was added for the solubilization of formed crystals. We seeded the same amount of cells at day 0 and performed an MTT-Assay on day 3, 5, and 7. At every time point control cells and cells treated with human recombinant IFNγ were tested. We used the same dose of human recombinant IFNγ, which was used in the murine study (1000 U/ml).

RNeasy Kit (Qiagen, Hilden, Germany) was used for RNA isolation of cultured NSCs. Then a reverse transcription into cDNA (ABI, Darmstadt, Germany) was performed. Real-time quantitative PCR was carried out by the usage of the 7500 fast or 7500 real-time quantitative PCR cycler (ABI, Darmstadt, Germany). SYBR green master mix (Qiagen, Hilden, Germany) or equivalent chemistry from another supplier (Quantace, London, UK) was used. The specific primers for genes of interest or the housekeeping gene were either purchased (QuantiTect primer assays, Qiagen) or self-designed (BioTEZ, Berlin, Germany). The genes of interest (target gene) in IFNγ-treated groups or control groups (PBS-treated) were analyzed in at least 3 independent cultures in triplicate each. Every experiment in IFNγ-treated or control (PBS-treated) groups provided delta CT values (ΔCT: gene of interest minus reference gene).

Experiments were repeated with independent cultures at least three times in triplicate each. The resulting data sets were statistically analyzed und illustrated using the GraphPad Prism 4 (GraphPad Software Inc., San Diego, CA, USA, 2003) software. For approval of statistical significance between groups, a two-tailed unpaired t-test was performed. P-values <0.05 were considered to indicate significant differences.

Lumbar CSF was obtained from patients admitted to our department for diagnostic purposes. Patients suffered from multiple sclerosis (three samples), peripheral nervous plexus affections (one samples), Alzheimer's disease (one sample), viral myelitis (one sample), or stroke (one sample), were compared to control CSF (three samples). All CSF specimens were immediately centrifuged, aliquoted and stored at –35°C. Ethics approval for the use of human CSF was obtained from the institutional ethics committee. For IFNγ detection in the CSF samples an IFNγ ELISA by R&D systems (R&D systems, Wiesbaden-Nordenstadt, Germany) was used. The experimental protocol followed the manufacturer's instructions. For statistical testing 3 control patients were compared with 3 MS patients with an unpaired two-tailed t-test.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.