Abstract
People with Down syndrome (DS) are at high risk of developing pathology similar to Alzheimer’s disease (AD). Modeling of this pathology in vitro may be useful for studying this phenomenon. In this study, we analyzed three different cultures of neural cells carrying trisomy of chromosome 21, which were generated by directed differentiation from induced pluripotent stem cells (iPS cells). We report here that in vitro generated DS neural cells have abnormal metabolism of amyloid-β (Aβ) manifested by increased secretion and accumulation of Aβ granules of Aβ42 pathological isoform with upregulated expression of the APP gene. Additionally, we found increased expression levels of genes that are considered to be associated with AD (BACE2, RCAN1, ETS2, TMED10), as compared to healthy controls. Thus, the neural cells generated from induced pluripotent stem cells with DS reproduce initial cellular signs of AD-type pathology and can be useful tools for modeling and studying this variant of AD in vitro.
INTRODUCTION
There are currently a lot of data showing that Down syndrome (DS) and Alzheimer’s disease (AD) are closely connected [1–3]. This relationship is revealed by abnormally high incidence of cases of AD-type pathology among people with DS by age 40 years (up to 80%) [4] and the presence of AD pathological processes like accumulation of amyloid-β (Aβ) plaques in brain tissue of DS patients beginning from early ages [5, 6]. Thus, it can be assumed that the neural cells from DS patients may be used as an in vitro cellular model of AD-type pathology along with neural cells derived from the patients with genetic forms of AD that are usually linked to mutations in APP or PSEN1 genes [7, 8]. Derivation of induced pluripotent stem cells (iPS cells) with subsequent neural differentiation allows one to obtain patient’s neural cells rather easily. Such an approach was reported previously [9–11]. However, the authors usually compared only one iPS cell line from a DS donor with the iPS cell line from a healthy donor. At the same time, it is known that stochastic rewriting of epigenetic profile during reprogramming significantly affects characteristics of individual iPS cell lines [12]. It should be kept in mind that pre-existing mutations may also contribute to heterogeneity of iPS cell cultures [13, 14]. As a result, there is a risk of deriving iPS cell lines with abnormal unique properties including the impact on Aβ metabolism. To overcome this problem, we analyzed cultures of functional neural cells derived from three iPS cell lines with DS in comparison with three neural lines with normal karyotype. Quantitatively evaluating concentrations of secreted and intracellular Aβ, we have shown that neural cells from DS donors have abnormal metabolism of Aβ compared to healthy controls. Using RT- PCR, we have demonstrated that neural cells from different iPS cell lines with DS differ in the expression levels of genes related to AD (APP, BACE2, RCAN1, ETS2, TMED10) as compared to the healthy controls.
MATERIALS AND METHODS
iPS cell cultures
Synthesis of lentivirus constructions carrying Oct4, Sox2, c-Myc, and KLF4 was performed at Evrogen (Moscow, Russia, http://www.evrogen.com). Amniotic fluid stem cells (AFS cells) were cultured in Chang medium supplemented with 10% fetal bovine serum. One day before transfection, cells were passed at Ø 6 cm Petri dishes, 105 cells per dish. Then co-transfection of four constructs with Oct4, Sox2, c-Myc, and KLF4 genes was performed. The day after, the viral stock was removed and substituted for AFS cells growth medium with 20 ng/ml bFGF. The medium was changed every three days. After 7 days, we observed the growth of colonies with the ES morphology, and the medium was substituted for mTeSR1 (Stem Cell Technologies). From this moment, medium was changed every day. About 14 days after transfection, colonies with ES-like morphology were mechanically transferred to individual wells of a 24-well plate and cultured as individual lines in mTeSR1. In each step, plastic surfaces were coated with Matrigel solution (1/40 in DMEM/F12) (BD Bioscience). For iPS cell passaging, we used 1 μg/ml dispase solution in DMEM/F12 (Stem Cell Technologies). Cells were cultured at 37°C in a CO2-incubator with 5% CO2, 5% O2, 100% humidity.
Neural differentiation of iPS cells
One day before the induction of neural differentiation, iPS cells were seeded into Ø 6 cm dishes at a concentration of 106 cells per dish in mTeSR1 medium. On the next day, the medium was changed to Neural Induction Medium (Stem Cell Technologies). From this moment, the medium was changed every other day. After 7–10 days, we observed the appearance of specific neural rosettes. Neural rosettes were selectively picked up using Neural Rosette Selection Reagent (Stem Cell Technologies) and reseeded into separate Ø 6 cm Petri dishes, coated with BD Matrigel (1/40 in DMEM/F12) (BD Bioscience). Then neural stem cells (NSCs) were cultured in a Neural Proliferation Medium (Stem Cell Technologies) for 1–3 passages. One day before the terminal neuronal differentiation, neural stem cells were reseeded at density of 2*104 cells/cm2 onto plastic culture dishes and plates coated with BD Matrigel. On the next day, the medium was changed to the Neural Differentiation Medium (Stem Cell Technologies) and cells were cultured further (28–42 days). Medium was changed every other day.
Immunocytochemistry
Before staining, the cells were washed with PBS and then fixed for 15 min in 4% paraformaldehyde at room temperature (22°C–24°C). Then the cultures were washed three times in PBS (5 min at room temperature), incubated with primary antibodies in blocking solution (PBS with 10% FBS and 0.1% Tryton-X-100) for 1 h at 37°C, washed again and incubated with the appropriate Alexa Fluor 488/546-conjugated secondary antibodies (1:800, Molecular Probes) for 1 h at 37°C. After that, nuclei were counterstained with 1 mg/ml 4′, 6-Diamidino-2-phenylindole dihydrochloride (DAPI). Images were obtained using fluorescence microscope (CKX41; Olympus, Japan). Primary antibodies are given in Supplementary Table 1.
Specific neuronal activity measurement
Electrophysiological recordings of cell cultures were performed in the recording chamber under an upright microscope (BX-51 WI; Olympus, Japan). Individual cells were identified at 20× magnification using infrared-differential interference contrast (IR-DIC) microscopy. Signals were recorded using a Multiclamp 700B amplifier (Molecular Devices, USA), filtered at110 kHz and digitized at 10 kHz, using a Digidata 1550 interface (Molecular Devices, USA) and Clampex acquisition software. During recordings, the culture was maintained at room temperature (22°C–24°C) in Neural Differentiation Medium (Stem Cell Technologies). Whole-cell current recordings were performed from neurons using pipettes with resistance of 5–7 MOhm when filled with (in mM): 120 K gluconate, 30 KCl, 4 Mg-ATP, 10 phosphocreatine, 0.3 GTP, and 10 HEPES (pH 7.3, KOH, 293 mOsm). After establishing the whole-cell mode, depolarizing and hyperpolarizing current pulses of different amplitudes were applied to a patched neuron.
Quantitative RT-PCR
Total RNA was isolated using RNeasy Mini Kit (Qiagen, # 74106) according to manufacturer’s instructions. Reverse transcription was performed using MMLV RT Kit (Evrogen, # SK021) and oligo (dT)-15 primers. Real-time PCR was performed using qPCRmix-HS-SYBR + Low Rox ready-to-use reaction mixture (Evrogen, # PK156L) and 7500 Real-time PCR System (Applied Biosystems). Primer’s sequences are given in Supplementary Table 2. PCR results were normalized by the method described by Vandesompele et al. [15], using primers for GAPDH, ACTB, and UBC.
Western blot
Western blotting was used for the quantitative evaluation of the intracellular Aβ40 and Aβ42 isoforms content. Cells were lysed in RIPA buffer (150 mM NaCl, 50 mM Tris, 1% TritonX-100). The resulting solutions of total protein were mixed with Laemmli buffer (Bio-Rad, # 1610737) (1:1) and heated for 5 min at 98°C. We used ready-to-use polyacrylamide gels (Bio-Rad, # 4568123) for electrophoresis. Electrophoresis was performed for 1 h at 140 V. The transfer to PVDF-membrane (Bio-Rad, # 1,704,156) was performed for 30 min using a Trans-Blot Turbo Blotting System (Bio-Rad) at 25 V. Then, membranes were washed in TBST buffer (10 mM Tris, 140 mM NaCl, 0,1% Tween-20, pH 8,0) three times for 5 min, blocked in TBST containing 5% skim milk (Bio-Rad, # 1706404) for 1 h, and incubated with primary antibodies for 1 h at room temperature. After washing in TBST, membrane was incubated with secondary antibodies from the ECL Western Blotting System Kit (Amersham, # RPN2108) for 1 h at room temperature. Membrane imaging was carried out using ELC Western Blotting System Kit (Amersham, # RPN2108) and Chemi-Doc MP Imaging System (Bio-Rad). Quantitative analysis was performed using Image Lab Software 5.2.1 (Bio-Rad) according to the manufacturer’s instructions with the normalization of the band intensity to the total amount of protein in the sample.
ELISA
ELISA was used to quantify a secreted isoform Aβ42 in the conditioned cell media. For the ELISA analysis, we used commercial kit (Cusabio, # CSB-E08299H) according to the manufacturer’s instructions. Culture media for ELISA analysis (Neural Differentiation Medium (Stem Cell Technologies)) was incubated with cells for three days (72 h). For the measurement of one point, we used medium from three wells of a 24-well plate (3*100 μl), each measured in three independent wells of the ELISA plate. After collecting the conditioned medium, cells were harvested from plastic and counted using a TC20 counter (Bio-Rad). ELISA results were normalized to the total number of cells.
Karyotype analysis
Metaphase spreads were prepared from exponentially growing cells according to standard techniques using a 0.75-M KCl solution. Giemsa-banding (G-banding) of metaphases, chromosomes imaging, and karyotype analysis were carried out by Perinatal Medical Center CG “Mat’ i Ditya” (Moscow, Russia, http://www.mamadeti.ru). For each cell line, at least 15 metaphases were recorded and two metaphases were fully karyotyped.
Statistical analysis
In all experiments with quantitative analysis of the results, we conducted multiple independent measurements (with at least three independent replications). The data obtained in these independent experiments were averaged. In each experiment, we used at least three technical repeats, which were averaged with the calculation of the standard deviation. Error bars on all graphs represent standard deviation. To evaluate the significance of difference between the experimental data from each other, we used the Student’s t-test with the p value calculation.
RESULTS
Reprogramming cells from donors with DS to a pluripotent state
To derive neural cells from donors with DS and healthy controls with normal karyotype, we reprogrammed six cultures of amniotic human fluid cells (AFS-1, AFS-2, AFS-3 from healthy donors and AFS-DS1, AFS-DS2, AFS-DS3 from donors with DS) to a pluripotent state (Fig. 1). We chose the culture of amniotic fluid cells (Fig. 2a) as initial cells, because of the high efficiency of reprogramming [16] and biomaterial availability since these cells are routinely used for karyotype analysis if DS is suspected. Primary cultures of amniotic fluid cell of both types were obtained from healthy women 35–50 years of age. Reprogramming was performed according to the classical scheme [17] using lentiviral constructs with “Yamanaka cocktail” (Oct4, Sox2, c-Myc, and KLF4) as reprogramming agents in 5% O2 and in the presence of 1 μM valproic acid (Fig. 1) which increases reprogramming efficacy [18]. In general, we isolated 3 to 10 clonal lines of iPS cells from each culture of amniotic fluid cells. Subsequently, we chose one line from each culture carrying mostly pronounced iPSCs characteristics (morphology and immunostaining for markers of pluripotent stem cells).
As a result, we obtained three lines of iPS cells with DS (IPS-AFS-DS1, IPS-AFS-DS2, IPS-AFS-DS3) and three lines of iPS cells with normal karyotype (IPS-AFS1, IPS-AFS2, IPS-AFS3). All obtained cell cultures showed morphological characteristics of pluripotent stem cells (Fig. 2b) with high nuclear-cytoplasmic ratio growing in dense colonies. These cells were positive for pluripotent stem cells markers: Oct4 (Fig. 2c, Supplementary Figure 1), Sox2 (Fig. 2d, Supplementary Figure 1), SSEA3 (Fig. 2e, Supplementary Figure 2), SSEA4 (Fig. 2f, Supplementary Figure 2), Tra-1-60 (Fig. 2g, Supplementary Figure 3), and Tra-1-81 (Fig. 2h, Supplementary Figure 3). We also analyzed the karyotype (Fig. 2i, Supplementary Figure 4) and confirmed the presence of chromosome 21 trisomy in cultures derived from DS donors, as well as the absence of additional chromosomal aberrations that could occur during the reprogramming process.
Differentiation of iPS cells into neurons
We differentiated our iPS cell lines into neurons in two stages (Fig. 1). First, we induced neural differentiation manifested by the formation of the so-called neural rosettes (Fig. 3a), consisting of NSCs positive for nestin (Fig. 3b). After selective isolation and expansion of neural progenitors, cells were reseeded at low density and subjected to the second stage of differentiation to produce the mature neuronal cultures (Fig. 3c). As a result, we derived three cultures of neurons from DS donors (N-AFS-DS1, N-AFS-DS2, N-AFS-DS3) and three cultures of neurons from healthy donors (N-AFS1, N-AFS2, N-AFS3). The cultures of neurons were positive for mature neuronal markers such as beta-III-tubulin (Fig. 3d, Supplementary Figure 4), microtubule associated protein tau (MAPT) (Fig. 3f), Synapsin 1 (Fig. 3g), Synaptophysin (Fig. 3h), and neural cell adhesion molecule (hNCAM) (Fig. 3i). Also, a small fraction of cells (0.1–5%) was positive for the glial cells marker glial fibrillary acidic protein (GFAP) (Fig. 3e, Supplementary Figure 4). We have not noticed a significant difference in the effectiveness of neural differentiation between cultures of DS and healthy control cells. We also confirmed neural cell differentiation using the RT-PCR analysis of genes expression typical of pluripotent stem cells (Oct4, Sox2, Nanog, developmental pluripotency-associated protein 4 (DPPA4)) and neural cells (Nestin, b-III-tubulin, neuron specific enolase (NSE), GFAP) (Fig. 4b). We observed significant peak of pluripotency markers expression in the course of reprogramming donor cells toward iPS cells followed by substantial decrease during differentiation into NSCs and neurons. The only exception was the Sox2, which was previously shown to be highly expressed in neural cells [19]. Neural markers expression levels were significantly increased in neural stem cells and neurons.
We examined the ability of differentiated cells to behave like neurons, i.e., generate action potentials. Figure 4a shows an example recording of electrophysiological responses of differentiated cells to hyperpolarization and depolarization steps; depolarization induces a train of action potentials typical of mature neurons.
Neural cells from DS donors demonstrate signs of AD-type pathology in vitro
After confirming the neural state of derived cells, we performed a series of experiments designed to detect signs of abnormal Aβ metabolism in neurons of DS donors and their relation to AD-type pathology. Accumulation of pathological Aβ isoform consisting of 42 amino acid residues (Aβ42) is generally accepted as a marker of such processes. In contrast to the normal isoform with 40 residues (Aβ40), Aβ42 has a high tendency to aggregation and precipitation into the deposits which further develop into Aβ plaques [20, 21]. Immunohistochemical analysis of DS neuronal cell cultures using antibodies against Aβ42 showed accumulation of intensively stained granules in tightly arranged neurons (Fig. 5). Such clusters were not found in cells from healthy donors although there were weak staining background and single granules (Fig. 5). We have not noticed a significant difference between the DS neurons and healthy controls in immunostaining on AβPP and Aβ40 (Fig. 5).
We have analyzed the concentration of intracellular Aβ40 and Aβ42 isoforms by quantitative western blotting. Our results show that DS neurons have increased content of both isoforms of Aβ, on average, as compared to healthy control although it should be noted that results demonstrate high heterogeneity, that the cells N-AFS-DS1 showed values above the average healthy control by a factor of 3–3.5, while N-AFS-DS2 and N-AFS-DS3— approximately by a factor of 1.5–2 (Fig. 6a). It should be noted that we used non-denaturing lysis conditions, so the main band was located at 70 kDa. This corresponds to protein conformation of oligomeric Aβ that is generally consistent with the data reported previously [22, 23]. We have observed multiple bands corresponding to different oligomeric forms of Aβ but these bands were stained significantly less brightly than the main band at 70 kDa (data not shown). Using an ELISA assay, we measured the concentration of pathological Aβ42 secreted isoform into the culture medium. The results showed overall increased Aβ42 secretion in cultures of neuronal cells with DS as compared to the healthy control (Fig. 6b). Noteworthy, there was also a significant difference in values both between groups and within the group of DS cultures.
Analysis of AD associated gene expression
Observing signs of abnormal Aβ metabolism in neural cell cultures with DS, we decided to evaluate gene expression levels traditionally associated with AD by RT-PCR analysis. The list of genes included members of Aβ metabolism cycle: the precursor of amyloid protein gene (APP), beta-secretase gene type 1 (BACE1), beta-secretase gene type 2 (BACE2), and genes encoding components of gamma-secretase— PSEN1, PSEN2, PSENEN. We also looked for gamma-secretase regulators: CD147 (also known as EMMPRIN) and TMED10 (also known as P23); genes which are located in a specific region of chromosome 21— Down Syndrome critical region— and associated with AD: DYRK (Dual-specificity tyrosine-regulated kinases), RCAN1 (regulator of calcineurin 1), ETS2 (v-ets avian erythroblastosis virus E26 oncogene homolog 2); genes encoding proteins which are usually associated with neuronal differentiation and AD: CREB (cAMP responsive element binding protein 1) and MAPT (microtubule-associated protein tau). Results of RT-PCR measurements showed a significant spread of the expression levels. As in the previous section, we found a significant difference between both groups (DS and healthy controls), and between cell lines inside the groups (Fig. 7). In our opinion, this effect is related to the individual characteristics of each cell line and the status of neurons in culture at the moment of RNA extraction taking into account that we took average results of at least three different independent experiments for each point. Therefore, we considered the gene to be differently expressed if only values in compared groups significantly differed from each other with a p-value level less than 0.05. Applying this criterion, we found that of the 13 genes examined, the expression levels of only 5 genes (APP, BACE2, TMED10, RCAN1, and ETS2) differed significantly in DS neurons in comparison with healthy controls (Fig. 7). Noteworthy, four of these genes are located on chromosome 21 and the average expression of these genes was more than two times higher in cells with DS.
DISCUSSION
A lot of studies have confirmed that iPS cells generated from donors with inherited diseases are useful tools for biomedicine because they can reproduce the pathological processes in vitro and provide a basis for test systems for drug screening. For example, iPS cells from donors with hereditary forms of suchneurodegenerative diseases as AD, Parkinson’s disease, and Huntington’s disease show the ability to reproduce certain pathological processes in vitro [24–26]. iPS cells derived from donors with DS may be a good example of such test systems due to the high risk of AD-type dementia in patients with DS after 40 years. In our study, we obtained three cultures from different donors with DS and analyzed their abilities to reproduce the pathological cellular characteristics of AD-type pathology in DS in vitro in comparison with cultures obtained from three healthy donors with normal karyotype. We consider the research using not one but several lines of iPS cells to be a very important point as it has been shown that many iPS cell lines can acquire unique characteristics in the process of reprogramming of epigenetic circuit that is manifested by, for example, tendency to differentiate into certain lineage and other differences [12–14]. There is a possibility that such processes of chromatin changes may affect other important characteristics, in this case, responsible for the metabolism of Aβ. Thus, a study performed using one iPSC line may yield inadequate results, especially as having only one control line with normal karyotype. In addition, it seemed to us very important to analyze three iPSC lines from different donors, rather than three clonal lines from a single donor, as in this case, we also cover the different genetic background to compare.
Our results show that all three cultures from donors with DS have an increased secretion of pathological Aβ isoforms in comparison with healthy control. Immunocytochemical reaction for pathological Aβ42 isoform revealed accumulation of Aβ granules within the clusters of neurons from donors with DS. In healthy control neurons, we usually observed a few brightly stained granules. This is in agreement with the situation in vivo when sparsely arranged amyloid plaques can be usually detected in the brain tissue specimens of healthy individuals [27]. In our hands, iPS cell lines differed greatly by their characteristics including the lines within the DS group. For example, level of Aβ42 secretion in the group with DS was significantly different by 45% (between N-AFS-DS1 and N-AFS-DS3), that can reflect both the real variability in the concentration of Aβ secretion of different donors [28, 29] or show the individual characteristics of iPS cell lines acquired during reprogramming.
In all cell lines derived in the study, we evaluated expression levels of genes associated with AD. Currently, most researchers tend to explain hyper-amyloidosis in patients with DS by the fact that APP is located on triplicate chromosome 21. In favor of this hypothesis, many studies have been published on the familial cases of early AD, with the duplicated gene APP [30, 31]. On the contrary, sometimes the expression of genes located on chromosome 21 is reduced [32]. It is also worth of mentioning that the mechanism of AβPP to Aβ conversion includes several stages which may limit the process. Therefore, increased AβPP content does not necessarily increase the content of Aβ and vice versa; high content of Aβ is not always related to high APP expression. For example, Yagi and co-authors [24] showed that the concentration of APP mRNA in cells of AD patients was indistinguishable from that in healthy controls. Additionally, as it was previously demonstrated in multiple studies, it is the ratio of different Aβ isoforms (39, 40, 41, and 42 amino acid residues) that is very important for the development of AD. For example, in hereditary forms of AD (mutations in APP and PSEN1), shift of secretion balance in favor of Aβ42 isoform of Aβ instead of Aβ40 is observed without increase in the overall level of APP expression and Aβ secretion. This shift leads to an extremely high probability of AD at an early age [33, 34]. On the other hand, it is supposed that AP-type pathology in DS represents a special case of hyper-amyloidosis and may be caused by rather different mechanisms as compared to other cases of AD.
Taking into account all mentioned above, it seemed to us very important to determine the expression of other genes potentially affecting the metabolism of Aβ and its secretion level. In our experiments, five of 13 genes examined had altered expression in DS cultures being overexpressed. Quite expectedly, four of them (APP, BACE2, RCAN1, and ETS2) are located on chromosome 21. Two of these latter genes, APP and BACE2, are directly involved in the production of Aβ; their increased expression may upregulate Aβ secretion. Interestingly, it was previously reported that BACE2 normally contributes less into the metabolism of Aβ than BACE1 does [35]. It may be proposed that in our case it is BACE2overexpression that causes a shift of AβPP cleavage in favor of the beta-cleavage.
How RCAN1, ETS2, and TMED10 (also known as TMP21) can influence the hypersecretion of amyloid in DS patients is not yet known but the parallel between dysregulation of these genes and the development of AD may be found in many reports [36–40].
CONCLUSIONS
The cultures of neural cells from DS donors have abnormal processing of Aβ, compared to healthy controls. This is manifested by accumulation of Aβ42 granules in clusters of neurons in vitro, and, in this way, can reproduce the early stages of AD. Cultures of neural cells from DS donors can be used as a basis for creating test systems for screening drugs for AD-type pathologies treatment as iPS cells have unlimited proliferative potential which allows scaling the research. We will further evaluate the detailed mechanism of Aβ hypersecretion in neurons from donors with DS. The contribution of APP, BACE2, RCAN1, ETS2, and TMED10 to this process will also be further studied.
