Mesodermal cell types induce neurogenesis from adult human hippocampal progenitor cells


Address correspondence and reprint requests to Alexander Storch, Technical University of Dresden; Department of Neurology, Fetscherstrasse 74, 01307 Dresden, Germany.


Neurogenesis in the adult human brain occurs within two principle neurogenic regions, the hippocampus and the subventricular zone (SVZ) of the lateral ventricles. Recent reports demonstrated the isolation of human neuroprogenitor cells (NPCs) from these regions, but due to limited tissue availability the knowledge of their phenotype and differentiation behavior is restricted. Here we characterize the phenotype and differentiation capacity of human adult hippocampal NPCs (hNPCs), derived from patients who underwent epilepsy surgery, on various feeder cells including fetal mixed cortical cultures, mouse embryonic fibroblasts (MEFs) and PA6 stromal cells. Isolated hNPCs were cultured in clonal density by transferring the cells to serum-free media supplemented with FGF-2 and EGF in 3% atmospheric oxygen. These hNPCs showed neurosphere formation, expressed high levels of early neuroectodermal markers, such as the proneural genes NeuroD1 and Olig2, the NSC markers Nestin and Musashi1, the proliferation marker Ki67 and significant activity of telomerase. The phenotype was CD15low/–, CD34, CD45 and CD133. After removal of mitogens and plating them on poly d-lysine, they spontaneously differentiated into a neuronal (MAP2ab+), astroglial (GFAP+), and oligodendroglial (GalC+) phenotype. Differentiated hNPCs showed functional properties of neurons, such as sodium channels, action potentials and production of the neurotransmitters glutamate and GABA. Co-culture of hNPCs with fetal cortical cultures, MEFs and PA6 cells increased neurogenesis of hNPCs in vitro, while only MEFs and PA6 cells also led to a morphological and functional neurogenic maturation. Together we provide a first detailed characterization of the phenotype and differentiation potential of human adult hNPCs in vitro. Our findings reinforce the emerging view that the differentiation capacity of adult hNPCs is critically influenced by non-neuronal mesodermal feeder cells.

Abbreviations used

hippocampal neural progenitor cells


mouse embryonic fibroblasts, NPCs, neural progenitor cells



Neural precursor cells (NPCs) have been the focus of increased attention over recent years because of their potential in cell replacement and gene therapy in the adult brain of patients with neurological diseases, for example stroke and Parkinson's disease. Recent in vitro studies indicated that multipotent, self-renewing NPCs with the capacity to differentiate into glial and neuronal cell types can be isolated from several mammalian adult brain regions namely the hippocampus and the subventricular zone (SVZ) of the lateral ventricles (Reynolds and Weiss 1992; Johansson et al. 1999b; Rietze et al. 2001; Doetsch 2003; Hack et al. 2004) and in lower amounts from non-neurogenic areas, such as neocortex and the spinal cord (Magavi et al. 2000; Shihabuddin et al. 2000; Wachs et al. 2003). Indeed, a few recent studies demonstrated isolation of NPCs from different regions of the adult human brain including cortex, amygdala, hippocampus and SVZ (Kirschenbaum et al. 1994; Johansson et al. 1999a; Kukekov et al. 1999; Arsenijevic et al. 2001; Nunes et al. 2003; Westerlund et al. 2003; Moe et al. 2005), but only three previous studies investigated human adult NPCs from the hippocampus region in vitro (Johansson et al. 1999a; Kukekov et al. 1999; Roy et al. 2000). However, there is no detailed analysis yet of the expression of early neuroectodermal or neural stem cell markers in proliferating adult human hippocampal NPCs (hNPCs). These progenitor cells are reported to be multipotent cells differentiating into both glial and neuronal cells with some functional properties of neurons, such as the expression of sodium and potassium channels (Roy et al. 2000). In contrast to extensive analyses of the differentiation capacity of adult non-human mammalian hNPCs in the absence and presence of various feeder layer cell types, such as astrocytes and fibroblasts (Song et al. 2002), there is no systematic evaluation of the influence of feeder cell layers on the differentiation behavior of adult NPCs from human origin.

There is a growing body of evidence that cell-to-cell contacts are important for neuroectodermal specification of fetal and adult NPCs (Kawasaki et al. 2000; Storch et al. 2001; Song et al. 2002). Various cell types derived from neuroectodermal tissue including cortical, striatal and mesencephalic astrocytes support not only the neuronal differentiation, but also the neuronal subtype specification of fetal and adult mammalian brain-derived NPCs (Ling et al. 1998; Storch et al. 2001; Song et al. 2002; Nakayama et al. 2004; Hermann et al. 2005). Additionally, several mesodermal-derived cell types such as fibroblasts and PA6 stromal cells are reported to significantly alter the differentiation capacity of various embryonic, fetal and adult neuroprogenitor cell types (Kawasaki et al. 2000; Perrier et al. 2004; Hermann et al. 2005; Kitajima et al. 2005). Adult and embryonic fibroblasts promote the differentiation of various NPC types including embryonic stem (ES) cell-derived and adult NPCs into all neural cell types including astroglial and neuronal cells. PA6 cells drive the differentiation of embryonal stem (ES) cell-derived NPCs and adult mesencephalic NPCs into nerve cells with a subset of cells showing a functional dopaminergic phenotype. In most studies the conditioned media were not as effective as the corresponding cocultured feeder cell layer (Kawasaki et al. 2000; Song et al. 2002) suggesting that direct cell-to-cell contacts promote the neuroectodermal differentiation of NPCs.

Although there are a few reports on isolation and short-term propagation of adult hippocampal NPCs from human origin (Johansson et al. 1999a; Kukekov et al. 1999; Roy et al. 2000), the phenotype of long-term expanded adult human hippocamal NPCs as well as their neuroectodermal differentiation capacity is largely unknown. Here we describe the isolation, long-term expansion and differentiation potential of multipotent adult hippocampal NPCs derived from epileptic-surgery procedures. We further analyzed the expression of early neuroectodermal and NPC marker genes as well as the telomerase activity during the proliferation and their differentiation capacity on various coculture systems including mouse cortical astrocytes, mouse embryonic fibroblasts (MEFs) and PA6 stromal cells.

Materials and methods

Cell culture

Adult human hippocampal tissue was obtained from routine epilepsy surgery procedures (selective amygdale-hippocampectomy or anterior temporal lobectomy; 14 samples; age 18–55 years) following informed consent of the patients. All procedures were in accordance with the Helsinki convention and approved by the Ethical Committee of the University of Ulm. The tissue was taken from the removed hippocampi and stored in ice-cold Hank's balanced salt solution (HBSS) supplemented with 11 mm glucose and 1% penicillin/streptomycin for transporting the tissue from the operating theatre. All patients underwent high-resolution magnetic resonance imaging excluding tumors and were screened for the presence of infectious disease. In all cases, the hippocampus and the neuropathological examination did not reveal evidences for tumor formation. For expansion of neurospheres consisting of hNPCs, tissue samples were cut into small pieces with a scalpel, incubated in 0,1% trypsin (Sigma, St Louis, MO, USA) for 30 min at RT, incubated in DNase (40 mg/mL; Sigma) for 10 min at RT and homogenized to a quasi single cell suspension by gentle triturating (Storch et al. 2001). The cells were added to 25 cm2 flasks (2 × 3 106 viable cells per flask) in Knock-Out DMEM (Gibco BRL, Life Technologies; Tulsa, OK, USA), supplemented with 10% serum replacement (Gibco); 0.5 mm glutamine; 1% penicillin/streptomycin and 20 ng/mL of both EGF and FGF-2 (both from Sigma) at 5% CO2; 92% N2 and 3% O2 using an incubator equipped with an O2-sensitive electrode system (Haereus, Germany). After 10–20 days neurosphere formation was observed and these spheres were expanded for additional 5–8 weeks (in total 7–12 weeks, 5–12 passages) before differentiation was initiated. The medium was changed once a week, while the growth factors were added twice a week. For BrdU labeling, cells were incubated for 30 h with 10 µm BrdU.

Differentiation conditions

Induction of neural differentiation was initiated by plating the cells on poly-d-lysine (PDL) coated glass cover-slips in Knock-out-DMEM, 10% serum replacement, 0.5 mm glutamine, 1% penicillin/streptomycin, 10 ng/mL rh-BDNF (Promega, Madison, WI, USA), 100 µm di-butyryl-(db) cAMP and 0.5 µm retinoic acid (both from Sigma). Cells were differentiated for 14 days. For coculture analyses, various feeder cells (mouse embryonic fibroblasts [MEFs]), PA6 stromal cells as well as primary mixed cortical cultures (containing astrocytes and neurons) from E17 mouse embryos were cultured on gelatine-coated cover-slips as described earlier (Kawasaki et al. 2000; Storch et al. 2001). In coculture experiments, hNPCs were pre-stimulated with expansion media supplemented with 500 ng/mL Shh and 100 ng/mL FGF-8 for 48 h prior to seeding the cells. When coculture was started, confluent feeder layers were washed twice with PBS. The hNPCs were seeded at a concentration of 1.5–2.0 × 105 cells cm−2 in D-MEM high glucose/F12 (50 : 50) containing 5% fetal calf serum, 5% horse serum; 1% Penicillin/Streptomycin. For conditioned media experiments confluent PA6 cell layers were washed three times with the medium described above for coculture experiments and incubated further with the same medium. The media, conditioned by exposure to PA6 cells for 72 h, was then removed, centrifuged at 1800 × g for 15 min, and the supernatant was used immediately or stored at − 20°C until for further use. For conditioned media experiments, hNPCs were seeded on PDL at a concentration of 1.5–2.0 × 105 cells cm−2 similar to coculture experiments, but the media were supplemented with PA6 conditioned medium (50%). Medium change was performed on day 4 and every other day following that.

Flow cytometry

Human hNPCs were treated with Accumax® (Gibco) and washed with PBS. Dead cells were excluded from analysis by forward scatter gating. Samples were processed using a FACSCalibur flow cytometer and analyses were performed with the Cellquest software (both from Becton Dickinson, Franklin Lakes NJ, USA). Antibodies were used as follows: CD15-FITC 1 : 10, CD34-FITC 1 : 10, CD45-PE 1 : 10, CD133-PE 1 : 10 (all from Miltenyi Biotech, Bergisch Gladbach, Germany). A minimum of 10 000–12 000 events were acquired for each sample.


Cell cultures were fixed in 4% paraformaldehyde in PBS or with 4% paraformaldehyde/PBS followed by ice-cold acidic ethanol and 2 N HCl for BrdU staining. Immunocytochemistry was carried out using standard protocols. Cell nuclei were counter stained with 4,6-diamidino-2-phenylindole (DAPI). The following primary antibodies were used: Mouse anti-human nuclei 1 : 200; mouse anti-Nestin 1 : 500, rabbit anti-GFAP 1 : 1000, mouse anti-GalC 1 : 750, rabbit anti-Neurogenin2 1 : 500, rabbit anti-Musashi1 1 : 500, rabbit anti-NeuroD1 1 : 500 (all from Chemicon International, Temecula, USA); rabbit anti-Tuj1 1: 2000, mouse anti-Tuj1 1 : 500 (both from Covance, Richmond, CA); rabbit anti-Ki67 1 : 500 (Novocastra, Newcastle, UK), rabbit anti-Olig2 1: 2000 (kindly provided by Dr H. Takebayashi) and secondary antibodies conjugated to Alexa 488, 568 or 647 1 : 500 (all from Invitrogen-Molecular Probes, Carlsbad, California), and rat anti-BrdU 1 : 40 with fluorescence labeled secondary antibody (both from Abcam, Cambridge, UK). Images were captured using a fluorescence microscope (Axiovert 135, Zeiss, Oberkochen, Germany) or a Leica TCS/NT confocal microscope equipped with krypton, krypton/argon and helium lasers.

Telomerase activity

A highly sensitive in vitro assay known as the quantitative real-time telomeric repeat amplification protocol (Emrich et al. 2002; Saldanha et al. 2003) has been used for detecting telomerase activity (OTD kit; Allied Biotech, Ijamsville, MD). The telomerase activity in the cell or tissue extract is determined through its ability to synthesize telomeric repeats onto an oligonucleotide substrate in vitro. Telomerase from the cell extract or tissue adds telomeric repeats onto a substrate oligonucleotide and the resultant extended product are subsequently amplified by PCR. The PCR products are then visualized using DNA fluorochromes SYBR Green as described above. Mouse ES cells (D3 ES cell line) were used as positive control and cultured as described previously (Chung et al. 2002), whereas adult mouse cortical tissue was used as negative control (Milosevic et al. 2005).

RNA extraction, and quantitative real-time RT-PCR analysis

Total cellular RNA was extracted from hNPCs during expansion using RNAeasy total RNA purification kit followed by treatment with RNase-free DNase (Qiagen, Hilden, Germany). Quantitative real-time one step RT-PCR was carried out using the LightCycler® System (Roche, Mannheim, Germany), and amplification was monitored and analyzed by measuring the binding of the fluorescence dye SYBR Green I to double-stranded DNA. 1 µL (50 ng) of total RNA was reverse transcribed and subsequently amplified using QuantiTect SYBR Green RT-PCR Master mix (Qiagen) and 0.5 µmol l−1 of both sense and antisense primers. Tenfold dilutions of total RNA were used as external standards. Standards and samples were simultaneously amplified. After amplification, melting curves of the RT-PCR products were acquired to demonstrate product specificity. The results are expressed relative to the housekeeping gene HMBS (hydroxymethylbilane synthase). Primer sequences, lengths of the amplified products and melting point analyses are summarized in Table 1.

Table 1.   Primers for quantitative real-time RT-PCR
Gene (Protein)Sequence (forward; reverse)Accession Number
(frizzled homologue 1)
(hydroxymethylbilane synthase)
(musashi 1)
(neurogenic differentiation 1)
(neurogenin 2)
(Notch homolog 1 translocation associated)
(neurotrophic tyrosine kinase, receptor, type 1)
(orthodenticle homolog 1)
(orthodenticle homolog 2)
P75 (NTR)
(nerve growth factor receptor)
(sex determining region Y-box 1)
SOX2 (sex determining region Y-box 2)5′-CAG GAG AAC CCC AAG ATG C-3′
(sex determining region Y-box 10)


Cells were investigated for membrane currents 3–10 days after initiation of differentiation on PDL using the standard whole cell patch clamp technique with an EPC-7 amplifier (List Electronics, Heidelberg, Germany) and pClamp data acquisition (Axon Instruments, Union City, CA) essentially as described previously (Storch et al. 2003; Hermann et al. 2004). For better sealing we added 5% FCS 5 h prior to the patch clamp experiments. The extracellular solution contained (in mmol l−1): 142 NaCl, 8.1 KCl, 1CaCl2, 6 MgCl2, 10 HEPES, 10 D-Glucose, pH value 7.4. The pipette solution contained (in mmol L−1): 153KCl, 1 MgCl2, 5 EGTA, 10 HEPES, pH value 7.3. Using these solutions, borosilicate pipettes had resistances of 6–10 MΩ. Seal resistances in the whole cell mode were between 0.5 and 1 GΩ. Data were analysed using pClamp 8.0, Microsoft Excel 97 and Origin 5.0 software. Resting membrane potentials (RMP) were determined in the current clamp mode immediately after establishing the whole cell configuration. Action potentials (APs) were elicited by applying increasing depolarizing current pulses (300 ms, 10 pA current steps).

Determination of GABA, glutamate, dopamine and serotonin production

For determination of dopamine production, media were supplemented with 100 µmol l−1 tetrahydrobiopterin and 200 µmol l−1 ascorbate 2 days prior to medium harvest. Dopamine levels were determined in medium stabilized with EGTA/glutathione solution as reported previously (Storch et al. 2001), and stored at − 80°C until analysis. Dopamine was assayed by reverse-phase HPLC with an electrochemical detector as previously described (Gerlach et al. 1996). GABA, glutamate and serotonin were assayed by HPLC with fluorescence detection (Gerlach et al. 1996), employing precolumn derivatisation with ortho-phtaldialdehyde and an automatic HPLC system (Kontron Instruments, Neufahrn, Germany). The excitation and emission wave-length of the fluorescence detector were set at 330 and 450 nm, respectively.

Cell counting and statistics

For quantification of the percentage of cells producing a given marker, in any given experiments the number of positive cells of the whole well surface was determined relative to the total number of DAPI-labeled nuclei. In a typical experiment, a total of 100–500 cells were counted per marker. In coculture experiments, only cells positive for the antihuman nuclei antibody were determined in confocal microscopic images to avoid the counting of overlaying cells. Human nuclei/GFAP+ and human nuclei/Tuj1+ cells, respectively, were counted in sister cultures. Statistical comparisons were made by Dunnett's t-test. If data were not normally distributed, a non-parametric test (Mann–Whitney U-test) was used for comparisons of results. Data presented are pooled from experiments performed on cells obtained from tissue of all 14 patients. All data are presented as mean ± SEM.


Isolation and long-term expansion of neurosphere-forming cells from the adult human hippocampus

After 10–20 days in vitro, the cells formed small spheres with all typical morphological properties of neurospheres (Fig. 1a). FACS analysis revealed that the phenotype of these neurosphere-forming NPCs was CD15/Lex1low/–, CD34, CD45, CD133 (Fig. 1b). Extensive analyses of the expression of early and late neuroectodermal marker genes using quantitative real-time PCR and immunocytochemistry (Figs 1a,c, for complete names of genes and encoded proteins refer to Table 1) revealed that hNPCs expressed high levels the early neuroectodermal or CNS precursor cell markers, such as Nestin and Musashi1 (Cattaneo and McKay 1990; Reynolds et al. 1992; Kaneko et al. 2000) on both the mRNA and protein level. Furthermore, adult hNPCs showed detectable expression levels of the proneural genes NeuroD1 and Olig2 on the mRNA and/or the protein level (Figs 1a,c). Consistent with the FACS analysis data, only a few cells were weakly positive for CD15 protein (Fig. 1a). Quantitative real-time PCR additionally showed the expression of several early neuroectodermal marker genes, such as NTRK1, OTX1/2 and SOX1,2,10 as well as surface receptors important for neural development [NOTCH1, NTRK1, FZD1, p75(NTR); Fig. 1(c)]. This phenotype is similar to that of NSCs derived from adult mouse SVZ or hippocampus (Rietze et al. 2001; Doetsch et al. 2002; Hack et al. 2004). hNPCs could be passaged and cultured as secondary and tertiary neurospheres without changing morphology and phenotype for up to 12 weeks (≈ 8–10 passages).

Figure 1.

 Characteristics of adult human hippocampal NPCs during in vitro expansion. (a) Morphology and marker expression of adult hNPCs during expansion in the presence of EGF and FGF-2. Spheres (phase contrast) were cultured for 2–3 h to allow attachment and then stained for Nestin, Ki-67, Musashi1, Nestin and NeuroD1 or Olig2, MAP2ab, GFAP and Lex1/CD15. Nuclei were counterstained with DAPI. Scale bars, 100 µm (all except Musashi1 and NeuroD1 stainings, respectively) or 50 µm (Musashi1 and NeuroD1 stainings). (b) Flow cytometry of adult hNPCs cultured for 4–12 weeks (5–10 passages). Cells were labeled with fluorescence-coupled antibodies against CD15, CD34, CD45, CD133 or immunoglobulin isotype control antibodies. Cells were analyzed using a FACSCalibur flow cytometer. Black line, control immunoglobulin; red line, specific antibody. (c) Quantitative transcription profile of hNPCs. Results of real-time RT-PCR analysis of the NSC markers Nestin and Musashi1 (NES, MSl1), proneural genes (NeuroD1, NGN2), genes early expressed in the development (Frizzled 1 (FZD1)), Notch1, OTX1, OTX2, SOX1, SOX2 and SOX10) and tyrosin-receptor kinases like p75NGF. and NTRK1 are displayed. Expression levels are expressed relative to the housekeeping gene HMBS. For primers complete names of the genes and accession numbers see Table 1. Results are mean values ± s.e.m. from at least three independent experiments. (d) Telomerase activity in human NPCs after 3 to 8 weeks as well as adult cortical tissue and mouse ES cells during expansion measured by the quantitative real-time telometric repeat amplification protocol. Telomerase activity was normalized to protein content. Results are mean values ± s.e.m. from three independent experiments. * indicates P < 0.05, ** represents P < 0.01 when compared to adult cortical tissue; + indicates P < 0.05 when compared to ES cells.

To confirm the proliferation potential of adult hNPCs, we used the proliferation marker Ki67 as well as BrdU incorporation to identify DNA-synthesizing cells (Fig. 1a). Ki67 is detected in the nucleus of proliferating cells in all active phases of the cell cycle from the late G1 phase through the M-phase but is absent in non-proliferating and early G1-phase cells and cells undergoing DNA repair (Gerdes et al. 1983; Gerdes et al. 1984; Key et al. 1994). As expected, total Ki-67-staining (representing late G1- through M-phase) was higher at each time point (67 ± 8% of cells were Ki67+n = 3) than total BrdU staining (11 ± 1% BrdU+ cells; n =3), which only marks cells within the S-phase of the cell cycle. After 4–7 weeks, an average number of 7 ± 6% of hNPCs were found as spontaneously apoptotic (chromatin condensation and margination, apoptotic bodies in some cells) by DAPI staining. These cells were negative for Ki67 (data not shown) as well as for most other markers tested such as Nestin and Olig2 (Fig. 1a).

Telomerase is inactive in most somatic cells, but present in various stem cell populations (Meyerson et al. 1997; Ostenfeld et al. 2000; Mattson and Klapper 2001). Quantification of telomerase activity in hNPCs in comparison to adult human differentiated tissue using the telomeric repeat amplification protocol (Emrich and Karl 2002; Saldanha et al. 2003) showed significant telomerase activity in hNPCs compared to human cortical differentiated tissue as negative control without changes over the three months expansion period (Fig. 1d). Together, investigations of the proliferation characteristics revealed the typical pattern of slowly dividing cells with a small amount of spontaneous apoptosis within the hNPC neurospheres.

Adult human hippocampal NPCs differentiate into functional neuronal cell types

Differentiation of adult hNPCs was initiated by removal of mitogens, plating the cells onto poly d-lysine and addition of BDNF, db-cAMP and retinoic acid. After 14 days, 36 ± 8% acquired morphologic and phenotypic characteristics of astrocytes (GFAP+), 31 ± 13% that of oligodendrocytes (GalC+), 29 ± 7% that of young neurons (Tuj1+) and 11 ± 8% that of mature neurons (MAP2ab+n = 7; Figs2a,b). As we did not find MAP2ab+ cells during expansion of hNPCs within the neurospheres (Fig. 1a), all MAP2ab+ neurons must be derived from progenitor cells not expressing this marker for mature neurons. GFAP and Tuj1 were never found in the same cell (Fig. 2a). Very few cells expressed markers for dopaminergic (tyrosin hydroxylase) or serotoninergic (5-hydroxytyramine, 5-HT) cells (data not shown). Pre-stimulation of hNPCs with expansion media supplemented with Shh (500 ng/mL) and FGF-8 (100 ng/mL) for 48 h prior to seeding the cells did not lead to significant changes in their differentiation behavior (compare Fig. 2b with 3b).

Figure 2.

In vitro differentiation capacity of adult human hippocampal NPCs. (a) hNPCs differentiated by plating onto poly-d-lysine (PDL), mitogen withdrawal and incubation with BDNF (10 ng/mL), db-cAMP (100 µm) and retinoic acid (0.5 µm) for 14 days. The cells were stained against markers for astrocytes (GFAP), oligodendrocytes (GalC), and/or neurons (Tuj1; MAP2ab), respectively. In the right panel, arrows mark GFAP+ astrocytes and arrowheads mark Tuj1+ neurons. Nuclei are counterstained with DAPI (blue). Scale bar, 30 µm. (b) Quantification of differentiation capacity of 14-day cultures of hNPCs shown in (a). Data shown are mean values ± SEM from three to seven independent experiments. (c–e) Electrophysiological recordings on hNPCs differentiated without feeder layers as described under (a). For voltage-clamp measurements, cells were held at − 70 mV and hyper- or depolarized in 10 mV steps to + 100 mV for 50 ms. (c) Example of a fast inward current and a sustained outward current (inset: only currents for depolarizing steps to − 60, − 50, 0, 20 and 60 mV are shown). (d) Current-voltage relationships of the sustained outward currents (upper diagram) and the inward currents (lower diagram) of the cell displayed in (c). (e) Representative voltage traces of current-clamp recordings of differentiated hNPCs in response to depolarizing current pulses of increasing amplitude. The resting membrane potential was ∼− 65 mV and the action potential peaked at ∼ + 30 mV. (f–g) Neurotransmitter production of differentiated hNPCs. Glutamate (f) and GABA (g) production were measured in hNPCs differentiated on poly-d-lysine (PDL), MEFs or PA6 cells for 14 days (see Material and methods) using a HPLC-based assay. Quantification of the neurotransmitters was performed in medium conditioned for 2 days. Feeder cell layer did not produce detectable amounts of glutamate or GABA. *p < 0.05; **p < 0.01 when compared to cells differentiated on PDL.

Figure 3.

 Differentiation capacity of adult hippocampal hNPCs on various coculture systems including mouse fetal mixed cortical cultures, mouse embryonic fibroblasts (MEFs), and PA6 stromal cells, as well as by PA6 cell conditioned media (CM). Differentiation was initiated after an expansion phase of 3–10 weeks as described in the Material and methods. (a) Confocal microphotographs of triple immunostainings of hNPCs for markers of astroglia (GFAP) or neurons (Tuj1) as well as human cells (human nuclei). Nuclei were counterstained with DAPI (blue). Scale bar, 75 µm. (b) Quantification of GFAP+ or Tuj1+ derived from hNPCs (counterstained with human nuclei antibody) by differentiation on PDL, mixed cortical feeder cell layers, MEFs, PA6 cells, or on PDL with PA6 cell conditioned media (50%). (c) Length of the neurites of Tuj1+/human nuclei+ neurons derived from hNPCs under coculture conditions. Results are mean values ± SEM from at least three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 in comparison to PDL and PA6 CM, respectively; †p < 0.05, ††p < 0.01 in comparison to cortical cocultures; ‡p < 0.05 in comparison to MEFs.

To analyze the functional properties of these differentiated hNPCs, we performed both whole-cell voltage-clamping to record voltage-gated sodium and potassium currents and current-clamping to search for the capacity to generate action potentials after differentiation for 3–7 days. The resting potential of the cells, determined in the current clamp mode immediately after establishing the whole cell conformation, was − 41 ± 12 mV (n = 8). The majority of differentiated adult hNPCs (5 out of 8 recorded cells) expressed a sustained outward current ranging from a few hundred pA to up to 4 nA (Fig. 2c). These currents showed a voltage dependence and kinetics characteristic for delayed rectifier K+ channels (Fig. 2d). In a few cells, we could identify inward currents with voltage dependence and kinetics typical for voltage-activated Na+ channels and were able to record action potentials (Fig. 2c–e). Neurotransmitter (glutamate, GABA, dopamine and serotonin) production was studied on hNPCs differentiated for 14 days using a HPLC-based method. We found significant production of glutamate and GABA by differentiated hNPCs (Fig. 2f–g), but no generation of dopamine and serotonin (data not shown).

PA6 stromal cells and mouse embryonic fibroblasts induce neurogenesis from adult human hippocampal NPCs

We next tested the neuronal differentiation potential of adult hNPCs by plating the progenitor cells on feeder layers of several cell types known to promote neuronal differentiation of ES cells and/or NPCs including fetal mixed cortical cultures, mouse embryonic fibroblasts (MEFs) and PA6 stromal cells (Kawasaki et al. 2000; Perrier et al. 2004; Hermann et al. 2005; Kitajima et al. 2005). The human origin of immunostained cells was confirmed by costaining the cultures with a human-specific antibody and investigating the cells using a confocal microscope. As shown in Fig. 3a and b, coculture of hNPCs with primary cortical cells, MEFs or PA6 stromal cells for 14 days resulted in a significant increase of Tuj1+ neurons, while only astrocytes and MEFs produced more GFAP+ glial cells compared to control conditions (PDL-coated culture surface). The potential of different feeder cells to induce neurogenesis from adult hNPCs could be ordered as follows: PA6 cells > MEFs > cortical feeder cells > PDL, whereas the ranking order for the promotion of an astroglial phenotype was: Cortical feeder cells > MEFs > PA6 cells = PDL. Consistently, morphological analysis of the Tuj1+ neurons derived from adult hNPCs by measuring the length of neurites revealed that coculturing the hNPCs with MEFs or PA6 cells promoted neurite outgrowth of differentiating hNPCs compared to cultures on PDL or mixed cortical cultures (Fig. 3c). Media conditioned by PA6 stromal cells for 72 h did not show any relevant effects on glial or neuronal differentiation of hNPCs (no significant changes of GFAP+ or Tuj1+ cell counts compared to hNPCs differentiated in normal media; Figs 3a,b). Furthermore, hNPCs on MEFs or PA6 cells produced significant higher amounts of the neurotransmitters glutamate and GABA compared to cells cultured on PDL or mixed cortical cultures (Fig. 2f–g). We did not find any dopamine or serotonin production under all culture conditions.


Here we describe the isolation, long-term in vitro expansion and the differentiation potential of multipotent adult human hippocampal NPCs derived from epilepsy surgery. Noteworthy, we used diseased hippocampal tissue to isolate hNPCs that most likely contains mesangial sclerosis reported to include an increased number of neural progenitor cells in vivo (Blumcke et al. 2001; Crespel et al. 2005). These brain-derived NPCs were expanded in neurosphere cultures similar to fetal human NPCs (Svendsen et al. 1998; Svendsen et al. 1999; Caldwell et al. 2001; Storch et al. 2001) and expressed several markers for early neuroectodermal cells or NPCs as well as significant telomerase activity. The adult NPCs differentiated into all major cell types of the brain, namely astroglia, oligodendroglia and neurons. Functional analyses of the differentiated NPCs revealed obligate functional properties of neurons including the expression of voltage-gated sodium and potassium channels, the generation of action potentials, and the production of neurotransmitters such as glutamate and GABA. The central finding of our study is that non-neuroectodermal cells derived from the mesoderm (mouse embryonic fibroblasts [MEFs] and PA6 stromal cells) promote neurogenesis and maturation of neurons from adult human hippocampal NPCs.

Since there are only a few reports on isolation and short-term propagation of adult hippocampal NPCs from human origin (Johansson et al. 1999a; Kukekov et al. 1999; Roy et al. 2000), detailed data on their morphological properties and marker expression during expansion as well as their differentiation capacity are missing. NPCs are often defined in vitro by the presence of the fetal and adult CNS stem cells marker Nestin (Cattaneo and McKay 1990; Lendahl et al. 1990; Dahlstrand et al. 1995), and the recently recognized NPC marker Musashi1 (Kaneko et al. 2000). Consistently, Nestin and Musashi1 are highly expressed in adult human hippocampal NPCs on both mRNA and protein levels. Another family of genes expressed in early neuroectodermal cells are the proneural genes encoding transcription factors of the basic helix-loop-helix (bHLH) class such as NeuroD1, Neurogenin2 and Olig2 (Franklin et al. 2001; Nieto et al. 2001; Simmons et al. 2001; Hack et al. 2004). Adult hippocampal hNPCs expressed these proneural genes, which are necessary and sufficient to promote the generation (proliferation and self-renewal) of progenitors that are committed to neural differentiation (Bertrand et al. 2002; Bylund et al. 2003; Graham et al. 2003). The early neurogenic markers OTX1/2, several SOXB1 transcription factors (SOX1,2,10) and surface receptors important for neural development (NOTCH1, NTRK1, FZD1, p75(NTR)) were also found on the mRNA level in adult hippocampal NPCs. FACS analysis revealed that the hNPCs are CD15low/–, CD34, CD45 and CD133. This phenotype is similar to that found in fetal human and adult mouse NPCs (Capela and Temple 2002; Vogel et al. 2003). The stem cell marker CD133 (or prominin-1) is expressed on several primitive cells such as hematopoietic stem or progenitor cells, embryonic neuroepithelial stem cells and endothelial stem cells (Weigmann et al. 1997; Uchida et al. 2000; Bhatia 2001), but there is no systematic report on CD133 expression in adult NPCs. However, in neuroepithelial stem cells, CD133 is mainly localized at the apical cell membrane facing the ventricular wall (Weigmann et al. 1997; Corbeil et al. 1999; Corbeil et al. 2000). Consistently, the adult hNPCs isolated from the hippocampus region without ventricular contacts were CD133. Together, human adult hippocampal NPCs expressed a typical pattern of early neuroectodermal and/or NPC marker genes, including major factors for proliferation/self-renewal and neurogenic differentiation. Due to a very slow growing behavior of the adult hNPCs (approx. 8–14 days doubling time), we failed to perform single cell (clonal) analysis of their differentiation capacity. However, we did not found MAP2ab+ cells during expansion of hNPCs within the neurospheres demonstrating that all MAP2ab+ neurons after differentiation must be de novo generated from MAP2ab-negative progenitor cells. Finally, showing proliferation by Ki67 and BrdU immunostaining as well as the expression of telomerase (Mattson and Klapper 2001), these adult NPCs fulfill the major characteristics of NPCs during in vitro expansion.

Differentiation of adult hNPCs was initiated by removal of mitogens and plating the cells onto poly-lysine and incubation for 14 days with BDNF, db-cAMP and retinoic acid with or without pre-stimulation with Shh and FGF-8 for 48 h. This mixture of differentiation factors was reported to induce neurogenic effects in adult hippocampal and tegmental NPCs from mouse without affecting the feeder cell layers (Hermann et al. 2005). In the present study, the pre-incubation of adult hNPCs with Shh and FGF-8 did not induce significant changes of the neuro-astroglia differentiation behavior of the adult hNPCs. Overall, hNPCs differentiated into all major cell types of the CNS, namely mature (MAP2ab+) neurons with long neurites (≈ 350 µm after 14 days of differentiation), oligodendrocytes and astrocytes. The majority of these cells expressed sustained outward currents with the typical properties of delayed rectifier potassium channels, but no detectable sodium channels. A few cells showed properties of young neurons, such as expression of sodium channels and generation of immature action potentials with a long duration and no after- hyperpolarization. This in vitro differentiation capacity is similar to that previously described for adult human SVZ and hippocampal NPCs (Johansson et al. 1999a; Kukekov et al. 1999; Roy et al. 2000; Westerlund et al. 2003; Moe et al. 2005). Roy and coworkers described potassium and sodium channels, but no action potentials in neuronal cells derived from adult hippocampal hNPCs expanded for only a short period of time (≈ 2 weeks) (Roy et al. 2000). In contrast, Westerlund et al. 2003 and Moe et al. 2005 reported detailed electrophysiological analysis of adult SVZ NPCs showing not only sodium channels and action potentials in differentiated NPCs, but also functional neuronal networks (synaptic connections) from single differentiated hNPCs (Moe et al. 2005). However, these cells were derived from the ventricular wall and not from hippocampal hNPCs and it seems to be important to distinguish well between NPCs from different brain regions as shown for mice counterparts (Liu et al. 1999; Doetsch 2003; Kempermann et al. 2004). In addition to these previous results, we were able to demonstrate that differentiated hNPCs produce various neurotransmitters, such as glutamate and GABA.

There is a growing body of evidence that cell-to-cell contacts and/or soluble trophic factors produced by feeder cells are important for neuroectodermal specification of fetal and adult NPCs (Kawasaki et al. 2000; Storch et al. 2001; Song et al. 2002). Various cell types derived from neuroectodermal tissue including cortical, striatal and mesencephalic astrocytes support not only neuronal differentiation, but also neuronal subtype specification of fetal and adult brain-derived NPCs (Kawasaki et al. 2000; Perrier et al. 2004; Hermann et al. 2005; Kitajima et al. 2005). Additionally, several mesoderm-derived cell types, such as fibroblasts and PA6 stromal cells, are reported to significantly alter the differentiation capacity of ES cells as well as various fetal and adult NPC types (Kawasaki et al. 2000; Perrier et al. 2004; Hermann et al. 2005; Kitajima et al. 2005). In contrast to the data from Song and coworkers showing that adult astrocytes derived from neurogenic regions (hippocampus) promote neurogenesis from adult mouse hNPCs (Song et al. 2002), mixed cortical cultures from E17 mouse embryos containing mainly astroglial did not alter neurogenesis from adult human hNPCs in the present study. This discrepancy most likely resulted from differences of the developmental stage and the brain region from which the primary cultures were isolated. In contrast, both mesoderm-derived cell types (MEFs and PA6 cells) induced neurogenesis of adult human hNPCs with over 80% Tuj1+ neurons in cocultures of hNPCs together with PA6 cells. Moreover, MEFs and PA6 cells promote morphological and functional maturation of adult hNPC-derived neurons. These neurons showed longer neurites with more branches and produced higher amounts of neurotransmitters. On the other hand, trophic support provided by PA6 conditioned media did not lead to significant changes of glial and neuronal differentiation of hNPCs. This is in agreement with studies on mouse and human ES cell-derived NPCs showing that fibroblasts and PA6 cells promote neurogenesis from these NPCs with coculturing being much more effective than PA6 conditioned media (Kawasaki et al. 2000; Zeng et al. 2004). Interestingly, PA6 feeder cells did not influence the glial potential of hNPCs, whereas MEFs and particularly mixed cortical cultures increased the number of astrocytes differentiated from hNPCs. Together, our results demonstrate a neurogenesis-inducing activity of mesoderm-derived cell types not only in ES cells as reported previously (Kawasaki et al. 2000; Zeng et al. 2004), but also in adult human hNPCs. The mechanisms or factors of this neurogenic activity are largely unknown, but most likely include cell-to-cell contacts and/or cell (membrane)-bound factors, such as the Wnt ligands (Kawasaki et al. 2000; Storch et al. 2001; Song et al. 2002; Schmidt and Patel 2005). However, due to the different neuronal-glia induction potential of the investigated feeder cell types, it seems to be most likely a combination of different trophic factors and cell surface interactions which influence the neuro-glia fate determination. We used rodent feeder cell layers to analyse the neural differentiation capacity of adult hNPCs and therefore it is unclear whether human feeder might have different trophic effects.

In summary, we provide a detailed characterization of long-term expanded human adult hippocampal neuroprogenitor cells derived from tissue obtained from epilepsy surgery procedures. The described technique for long-term expansion of these adult human NPCs allows the generation of high yields of cells for further characterizations by molecular biology or protein biochemistry, such as gene array analysis or proteomics. This study provides therefore a framework for standardized comparative analysis of brain-derived adult NPCs with adult NPCs derived from other human cell sources, such as bone marrow stromal or skin cells (Toma et al. 2001; Hermann et al. 2004). However, future studies are warranted to further define the functional (electrophysiological) properties including synaptic neurotransmitter release of differentiated adult human hNPCs and compare these data with data obtained on primary human astrocytes and neurons. Furthermore, we demonstrated that mesoderm-derived cell types, such as MEFs and PA6 stromal cells, not only induce neurogenesis from adult human NPCs, but also promote morphological and functional maturation of neurons derived from these progenitor cells. The analysis of the underlying molecular mechanisms in the future might help to better understand neurogenesis and neuronal maturation in the adult human brain.


The authors would like to thank S. Kanzler, T. Lenk, N. Meyer and R. Rainer for technical assistance. This work was supported in part by the Interdisziplinäres Zentrum für Klinische Forschung (IZKF) Ulm (Project D6) to AS, the Bundesministerium für Bildung und Forschung (Verbundvorhaben ‘Tissue Engineering’; AZ 0312126) to AS and JS, and the Landesstiftung Baden-Württemberg (Förderprogramm ‘Adulte Stammzellen’; AZ 37610) to AS. AH was supported by an IZKF fellowship as member of the graduate college GRK460, Ulm. HL is a Heisenberg fellow of the Deutsche Forschungsgemeinschaft.