Inhibition of neuronal nitric oxide synthase activity promotes migration of human-induced pluripotent stem cell-derived neural stem cells toward cancer cells


Address correspondence and reprint requests to Dr Shu Wang, Department of Biological Sciences, National University of Singapore, Singapore 117543, Singapore. E-mails:;


The breakthrough in derivation of human-induced pluripotent stem cells (hiPSCs) provides an approach that may help overcome ethical and allergenic challenges posed in numerous medical applications involving human cells, including neural stem/progenitor cells (NSCs). Considering the great potential of NSCs in targeted cancer gene therapy, we investigated in this study the tumor tropism of hiPSC-derived NSCs and attempted to enhance the tropism by manipulation of biological activities of proteins that are involved in regulating the migration of NSCs toward cancer cells. We first demonstrated that hiPSC-NSCs displayed tropism for both glioblastoma cells and breast cancer cells in vitro and in vivo. We then compared gene expression profiles between migratory and non-migratory hiPSC-NSCs toward these cancer cells and observed that the gene encoding neuronal nitric oxide synthase (nNOS) was down-regulated in migratory hiPSC-NSCs. Using nNOS inhibitors and nNOS siRNAs, we demonstrated that this protein is a relevant regulator in controlling migration of hiPSC-NSCs toward cancer cells, and that inhibition of its activity or down-regulation of its expression can sensitize poorly migratory NSCs and be used to improve their tumor tropism. These findings suggest a novel application of nNOS inhibitors in neural stem cell-mediated cancer therapy.

Abbreviations used

embryonic stem cells


human embryonic stem cells


human foreskin fibroblasts


human-induced pluripotent stem cells


neuronal nitric oxide synthase


nitric oxide


neural stem/progenitor cells



Neural stem/progenitor cells (NSCs) are defined by their ability to self-renew and to differentiate into three fundamental neural lineages, neurons, astrocytes, and oligodendrocytes. These stem cells have found their way into many medical applications, especially in treating brain illnesses where tissues cannot ordinarily regenerate, such as Alzheimer's and Parkinson's diseases. Although endogenous and exogenous NSCs migrate through established migratory routes in a healthy brain, they can also move through atypical migratory routes to reach the tumor microenvironment in the brain of cancer patients (Aboody et al. 2000, 2006, 2008). In in vivo studies on mice, NSCs are observed to surround the tumor border as well as “chasing down” infiltrating tumor cells regardless of NSC injection location and its distance to the tumor. When administrated intravenously, NSCs are also able to migrate to various extra-cranial locations, such as the liver, ovaries, and bone marrow, to target disseminated tumor cells of both neural and non-neural origin (Aboody et al. 2000, 2006; Brown et al. 2003; Yang et al. 2012). These findings have suggested a powerful therapeutic approach of using NSCs to target not only the major tumor mass but also distant tumor outgrowths.

So far different NSCs have been tested in animal tumor models, including NSCs derived from fetal or adult human brain tissues, immortalized human NSC lines, and NSCs derived from human embryonic stem cells (hESCs) (Taupin 2007; Aboody et al. 2008; Kosztowski et al. 2009; Zhao et al. 2012). The great potential of NSCs in cancer therapy emphasizes the necessity for a renewable and steady supply of NSCs for future clinical use. Although NSCs derived from healthy areas of the patient's brain and expanded in vitro for use in autologous transplantation would eliminate a need for immunosuppressive drugs as well as concerns over donor and host compatibility, the risks associated with invasive surgical procedures, such as damaging healthy brain tissue, have limited the clinical application of this approach (Taupin 2007). Furthermore, NSCs derived from these human tissues vary in quality and have limited passaging capacity (Taupin 2007; Aboody et al. 2008; Kosztowski et al. 2009). Although oncogene-mediated immortalization provides a solution to the limited life span of NSCs, this method poses various safety concerns over the transforming and oncogenic potential of the oncogenes used (Taupin 2007; Aboody et al. 2008; Kosztowski et al. 2009).

An alternative cell source for NSC generation is induced pluripotent stem cells (iPSCs) (Takahashi et al. 2007). iPSCs are generated synthetically through the expression of specific transcription factors in differentiated adult somatic cells, creating cells that possess the ability to differentiate into any cell type in the body. Human-induced pluripotent stem cells (hiPSCs) express genes and surface proteins similar to hESCs (Takahashi et al. 2007; Patel and Yang 2010; Wu and Hochedlinger 2011). hESCs have been used to derive NSCs that are similar to naturally occurring NSCs in their genetic make-up, surface markers, and potential to differentiate into various neural cells, yet have the advantage of displaying a larger proliferative capacity (Zhao et al. 2012). hiPSCs could be even more advantageous than embryonic stem cells in generating NSCs as they possess similar characteristics but circumvent the bioethical controversies associated with isolating embryonic stem cells (Nishikawa et al. 2008; Patel and Yang 2010). Moreover, somatic cells isolated from patients could theoretically be reprogrammed into iPSCs that can be used to generate NSCs for autologous transplantation, hence minimizing host versus graft reaction. Likewise, human leukocyte antigen-typed iPSCs could be generated to make iPSC-derived cells that are more patient-specific than other cell sources (Nakatsuji 2010).

Regardless of the NSCs source used for cancer therapy, the migratory capacity of NSCs is crucial for the success in treating disseminated tumors. Cell migration is a highly complex process. The factors that regulate the procedure range from various extracellular environmental cues, cytoskeleton re-structuring after sensing the cues, dynamic regulation of cell attachment and detachment to extracellular matrix to signaling coordination between all these parts to control cell movement (Muller et al. 2006; Aboody et al. 2008; Jurvansuu et al. 2008). A proper understanding of the basic molecular mechanisms of NSC migration toward tumors, especially identification of key cellular regulators of the migration, will have important implications in improving the effectiveness of engineering and employing NSCs as tumor therapy agents. As an initial attempt to address the issue, this study used cDNA microarray technology to analyze gene expression patterns of NSCs collected from in vitro migration assays.

Materials and methods

Generation and characterization of hiPSC-derived NSCs

A Cre-excisable polycistronic lentiviral vector containing Oct4, Klf4, Sox2, and c-Myc genes (Millipore, Bedford, MA) was utilized to generate hiPSCs from human foreskin fibroblasts (HFF, Millipore). Once derived, hiPSCs were maintained in the undifferentiated state by scraping off differentiated cells with a pipette or alternatively by mechanical passage of individual colonies of undifferentiated cells. Prior to differentiation of hiPSCs into NSCs, hiPSC colonies cultured on Matrigel (BD, Franklin Lakes, NJ, USA) were dissociated to single cells. The cells were then seeded onto a 0.1% gelatin-coated six-well cell culture plates (Nalge Nunc International, Rochester, NY, USA) at a density of 200 000 cells per cm2. After 1 month of passaging, adherent monolayer cell populations with characteristic bipolar or tripolar NSC morphology that were morphologically homogeneous were achieved and characterized. The details of generation and characterization of hiPSCs and hiPSC-derived NSCs are provided in Supplementary data.

NSC migration analysis

To assess the in vitro migration capacity of hiPSC-NSCs toward tumor, migration assays were performed in Boyden chambers with the BD Falcon HTS FluoroBlok cell culture inserts with 8-μm pore size for 24-well plates (BD Biosciences, San Jose, CA, USA) in Opti-MEM (Invitrogen, Carlsbad, CA, USA). To assess effects of neuronal nitric oxide synthase (nNOS) inhibitors on cell migration, hiPSC-NSCs, U87 cells, or 4T1 cells were pre-treated with nNOS inhibitors S-methyl-L-thiocitrulline (SMTC, Sigma-Aldrich, St Louis, MO, USA) at 500 μM or (4S)-N-(4-Amino-5[aminoethyl]aminopentyl)-N’-nitroguanidine (nNOS inhibitor I, EMD-Merck, Gibbstown, NJ, USA) at 20 μM for one day. The next day, 50,000 treated cells were seeded into the upper inserted chamber in Opti-MEM containing 500 μM SMTC or 20 μM nNOS inhibitor I. For hiPSC-NSC migration assays, U87 cells or 4T1 cells were seeded into the lower receiver well and used as attractants. For U87 and 4T1 tumor cell migration assays, 600 μl Opti-MEM with 10% fetal bovine serum (FBS) and 20% FBS was added, respectively, into the lower chamber as chemical attractants. The number of migratory cells attached to the bottom of the chamber membrane was counted after migration for 6 h.

For in vivo migration analysis, female specific pathogen-free athymic nude (nu/nu) BALB/c mice (Biological Resource Centre, Singapore, weight 20 g; aged 6–8 weeks) were used. Green fluorescent carbocyanine dye DiO (Invitrogen) labeled U87 cells [5 × 105 cells in 1 μL phosphate-buffered saline (PBS)] were used to generate an orthotopic glioma model. On day 7 post-tumor inoculation, the red fluorescent dye CM-DiI (Invitrogen) labeled hiPSC-NSCs (1 × 106 in 1 μL PBS) were injected into the same hemisphere 2 mm away from the tumor inoculation site. To generate an orthotopic breast cancer model, 4T1 cells were inoculated into the mammary fat pad of anesthetized mice at a dose of 1 × 105 cells in 50 μL PBS per animal. One week later, a lipophilic, near-infrared fluorescent dye 1,1′-dioctadecyl-3,3,3′,3′-tetramethyl indotricarbocyanine Iodide (DiR, Caliper Life Sciences, Alameda, California, USA) was use to label hiPSC-NSCs. The labeled cells, 1 × 106 in 200 μL PBS per animal, were injected into mice through the tail vein. All handling and care of animals was carried out according to the Guidelines on the Care and Use of Animals for Scientific Purposes issued by the National Advisory Committee for Laboratory Animal Research, Singapore.

Microarray analysis followed by verification with real-time PCR

cDNA microarray assays were preformed to compare the gene expression profiles between migratory and non-migratory hiPSC-NSCs. GeneChip® Human Genome U133 Plus 2.0 Array (Affymetrix, Santa Clara, CA, USA) containing more than 54000 probe sets representing all known genes in the human genome was used. Microarray raw data were acquired using the GCOS software (Affymetrix). The data were normalized using the default setting of GeneSpring. To analyze the data, the genes were grouped according to migratory or non-migratory NSCs toward different tumor attractants. The data were filtered according to their flags in such a way that 3 of 3 sample replicates fulfilled the conditions of a cut-off of ≥ twofold difference in expression level and a p-value of less than 0.05. The data were then analyzed using the DAVID functional annotation clustering tool (DAVID, Bioinformatics Resources 6.7, National Institute of Allergy and Infectious Diseases, Bethesda, MD) with a cut-off at an enrichment score of 1.3 (=p < 0.05). Microarray data described in this article have been deposited in NCBI Gene Expression Omnibus (, with accession number GSE40751, in a MIAME-compliant format. The relative mRNA expression of 9 genes selected from microarray assays was analyzed by real-time PCR. Primer sequences used are listed in Supplementary Table 1.

nNOS knockdown experiments using siRNAs

Two siRNAs against human nNOS (SI02624006 and SI03028823) and negative control siRNA were purchased from Qiagen. Each siRNA targets a distinct region of target mRNA. siRNA transfection was performed in 6-well plates using Lipofectamine® 2000 Reagent (Invitrogen). Forty-eight hours after transfection, the cells were harvested for PCR, real-time PCR and western blot analyses or cell migration assay. Goat polyclonal anti-nNOS antibody (1 : 500; Abcam, Cambridge, MA) was used for Western blot analysis.

Statistical analysis

All data are represented as mean ± SD. The statistical significance of differences was determined by unpaired Student's t-test. p-value of <0.05 was considered statistically significant.


Generation and characterization of hiPSCs and NSCs

Generation of hiPSCs from HFFs was achieved by lentivirus-mediated transduction of four transcription factors, Oct4, Sox2, Klf4, and c-Myc (Sommer et al. 2009, 2010). iPSC colonies were formed in 11–15 days (Fig. 1a), with an efficiency of 0.88% (264 putative colonies from 30 000 HFFs). These colonies could be expanded on either mouse embryonic fibroblasts feeder layer in iPS cell medium or feeder-free matrigel in mTeSR medium. hiPSC colonies displayed a typical morphology indistinguishable from hESC colonies and were positive for alkaline phosphatase (AP) staining (Fig. 1a). When the iPS cells were seeded in a bacterial culture dish and cultured in an embryoid body differentiation medium, spherical embryoid body aggregates were formed spontaneously (Fig. 1b).

Figure 1.

Generation and characterization of human-induced pluripotent stem cells (hiPSCs) by re-programming human fibroblasts and derivation of neural stem/progenitor cells (NSCs) from hiPSCs. (a) Formation of hiPSC colonies from human foreskin fibroblasts (HFFs). Generated colonies were visible on day 11, maintained and expanded on mouse embryonic fibroblast feeder layer or under feeder-free conditions on Matrigel, and stained positive for alkaline phosphatase (AP). P: Cell passage. (b) Cell cultures of embryoid bodies and hiPSC-NSCs. (c) RT-PCR analysis of expression of NSC markers in hiPSCs and derived NSCs. GAPDH was used as a loading control. (d) Immunocytochemistry analysis of the multipotency of hiPSC-NSCs. When cultured under defined conditions, hiPSC-NSCs differentiated into βIII tubulin-positive neurons, GFAP-positive astrocytes, and O4-positive oligodendrocytes.

To further examine the differentiation potential of these hiPSCs, we used an adherent monoculture method (Zhao et al. 2012) to generate NSCs in a defined culture condition containing bFGF and epidermal growth factor. hiPSCs propagated on Matrigel were used, since these hiPSCs have a higher survival rate and no contamination from mouse embryonic fibroblasts. On day 0, hiPSC colonies were dissociated, re-suspended in an NSC medium with ROCK inhibitor Y27632, and seeded onto 0.1% gelatin-coated 6-well plates at a density of 200 000 cells per cm2. The high seeding density was essential for cell survival in this step, while lower seeding densities usually resulted in poor cell attachment and low levels of cell survival. On day 1, many spherical cell clusters formed, but only a few of them adhere to the culture plate. More NSC clusters attached after an addition 2–3 weeks of culturing. These adherent clusters were then dissociated and re-seeded at a density of 50 000 cells per cm2. A lower seeding density was seen to trigger differentiation of NSCs. When the cells reached confluency, adherent monolayer with typical bipolar or tripolar NSC morphology could be observed (Fig. 1b).

The generated adherent NSCs were able to form neurospheres when transferred to suspension culture in a low-cell binding plate (Fig. 1b). These adherent cells expressed mRNA transcripts associated with NSCs, including CD133, Nestin, Pax6, Sox1, and Sox2 (Fig. 1c) and were stained positively with antibodies against Sox2, Nestin, Pax6, and Musashi (data not shown). Importantly, these cells displayed the functional hallmark of NSCs: differentiation into β3-tubulin-positive neurons, GFAP-positive astrocytes, and O4-positive oligodendrocytes (Fig. 1d).

hiPSC-NSCs initially grew with a doubling time of approximately 3–4 days. After 10 passages, the doubling rate of hiPSC-NSCs slowed down and the cells were subcultured every 4–5 days instead. After about 15 passages, the cells almost stopped proliferating and approached senescence. Thus, we mainly used NSCs obtained from early passage (less than passage 10) in the following experiments.

Tumor tropism of hiPSC-NSCs

The tumor tropism of primary NSCs, NSC lines, and hESC-derived NSCs has been well documented (Aboody et al. 2008; Zhao et al. 2012). To examine the in vitro migration capacity of hiPSC-NSCs toward tumors, we conducted in vitro transwell cell migration assays using Boyden chambers, which use a transmembrane with a cell culture insert in each well allowing cells move from the top to bottom chamber through the membrane. We employed human U87 glioblastoma cells and mouse 4T1 breast cancer cells, as well as their conditioned medium, as attractants in the bottom chamber and counted the number of migratory cells attached to the bottom side of the membrane. As shown in Fig. 2, hiPSC-NSCs displayed much higher migration capacities toward the tumor cells: the percentage of migratory cells was up to 35%, while the percentage of cells migrating toward plain Opti-MEM cell culture medium was around 5%.

Figure 2.

In vitro migration of human-induced pluripotent stem cells (hiPSC)-neural stem/progenitor cells (NSCs) toward tumor cells. Human U87 glioblastoma cells and mouse 4T1 breast cancer cells, as well as their conditioned media (CM), were used as attractants in Boyden chamber assays. Cell migration toward the Opti-MEM medium was included as a control. (a) Images of migratory hiPSC-NSCs. Scale bar, 1 mm in Calcein-AM images and 200 μm in 4′,6-diamidino-2-phenylindole (DAPI) images. (b) Quantification of hiPSC-NSCs migration. The results are expressed as the percentage of migratory cells to total cells. Values are mean ± SD (n = 3).

To investigate the in vivo tumor tropism of hiPSC-NSCs, an intracranial glioma model was generated by injection of green fluorescent carbocyanine dye DiO-labeled human U87 glioma cells into the right striatum of nude mice. Seven days later, red fluorescent dye CM-DiI labeled hiPSC-NSCs were injected into the same striatum 2 mm to the left of the tumor inoculation site. Using a confocal microscope to examine brain sections, we observed the accumulation of hiPSC-NSCs along the boundary of established tumor mass on day 5 after hiPSC-NSC injection. By day 10, extensive co-localization of hiPSC-NSCs with the tumor mass demonstrated deep penetration of the cells into the tumor (Fig. 3a). We further examined whether hiPSC-NSCs could migrate through the brain parenchyma to reach the tumor site in the opposite hemisphere. U87 glioma cells were pre-labeled with CM-DiI (red) and injected into the right striatum in nude mice. DiO-labeled hiPSC-NSCs (green) were injected into the left striatum 7 days later. Brain sections were collected on day 21 post-tumor inoculation for histological examination. We observed extensive co-localization of hiPSC-NSCs with the tumor mass in the collected brain sections (Fig. 3b). While most of hiPSC-NSCs infiltrated into around 50 μm, some NSCs had penetrated 100 μm into the tumor mass.

Figure 3.

In vivo tumor tropism of human-induced pluripotent stem cells (hiPSC)-neural stem/progenitor cells (NSCs). (a) Glioma tropism. U87 cells were pre-labeled with green fluorescent carbocyanine dye DiO and injected into the right striatum in nude mice. hiPSC-NSCs, pre-labeled with red fluorescent dye CM-DiI, were injected into the same hemisphere 2 mm away from the tumor inoculation site 7 days later. Brain sections were collected on day 5 and day 10 after NSC injection. (b) Glioma tropism after contralateral injection of hiPSC-NSCs. U87 cells were pre-labeled with CM-DiI (red) and injected into the right striatum. DiO-labeled hiPSC-NSCs (green) were injected into the hemisphere contralateral to the tumor inoculation site 7 days later. Brain sections were collected on day 21 post-tumor inoculation. A region with a U87 tumor is shown. (c) Breast cancer tropism. Immunodeficient NSG mice were inoculated with 4T1 breast cancer cells in the mammary fat pad. At day 7 post-tumor inoculation, the near-infrared fluorescent dye DiR-labeled hiPSC-NSCs were injected through the tail vein into the tumor-bearing mice. The mammary fat pads were harvested from the animals at day 14 post-NSC injection (day 21 post-tumor inoculation). Sections of the collected organs were labeled with DAPI and images were captured with a near-infrared confocal microscope.

To investigate the homing capabilities of systemically injected hiPSC-NSCs to tumors, we used the near-infrared fluorescent dye DiR to label the cells and injected them through tail vein into mice pre-inoculated with breast cancer cells in the mammary fat pad. At day 14 post-NSC injection, we examined DiR signals in the tissue sections of the mammary fat pad under a near-infrared confocal microscope and observed significantly increased DiR signals in the tumor inoculation site in tumor-bearing mice as compared with the signals in mice without tumors (Fig. 3c), demonstrating NSC accumulation in the tumor in this orthotopic breast cancer model in mice.

cDNA microarray analysis of gene expression profiles in migratory and non-migratory hiPSC-NSCs

To elucidate the molecular mechanisms underlying NSC tropism for tumor, migratory cells were collected from the bottom side of the filter membrane in Boyden Chambers and non-migratory cells from the top side. RNA was isolated from migratory and non-migratory cells, reverse transcribed into cDNA, and converted into labeled cRNA ‘targets’ to interrogate Affymetrix whole genome microarrays. Twelve microarray chips for four different RNA sources with three experimental replicates per sample were used in this study. The present gene probes on the three chips tested for the cells migrating toward the 4T1 tumor cell-conditioned medium were 49.1%, 50.2%, and 49.8%, respectively, whereas the present gene probes on the three chips tested for non-migratory cells isolated from the same Boyden Chamber migration assays were 48.4%, 49.1%, and 49.0%, respectively. In cells collected from the migration assay toward U87 cell-conditioned medium, 43.8%, 44.1%, and 44.2% of the gene probes on the three chips were present in the migratory NSCs and 45.8%, 45.7% and 46.8% in the non-migratory cells.

After the microarray data was filtered, there were 253 up-regulated and 232 down-regulated probe sets in NSCs migrating toward U87-conditoned medium and 53 up-regulated and 101 down-regulated probe sets in NSCs migrating toward the 4T1-conditioned medium (Supplementary Table 2). Among these were 31 over-expressed and 35 under-expressed probe sets that overlapped between samples with U87- and 4T1-conditioned media as attractants (Supplementary Table 3), indicating that there are common molecules that regulate NSC migration toward both U87 glioma cells and 4T1 breast cancer cells. To investigate whether the affected genes are associated with a specific cellular procedure or organelle that are crucial for NSC migration toward tumor cells, up- and down-regulated probe sets were subjected to DAVID functional annotation clustering analysis. A cluster of genes related to intracellular non-membrane-bounded organelles were detected to be enriched in both NSCs migrating toward 4T1-conditioned medium (19 down-regulated genes with enrichment score of 1.46) and NSCs migrating toward U87-conditioned medium (37 down-regulated genes with enrichment score of 2.13). Furthermore, there were 9 down-regulated genes overlapping between the two gene lists (Table 1).

Table 1. Down-regulated genes related to intracellular non-membrane-bounded organelles were detected in migratory neural stem cells
Gene SymbolGene NameFold Change (Toward 4T1)Fold Change (Toward U87)Basic Gene Function Summary
NOS1Nitric oxide synthase 1 (neuronal)2.923.86Synthesizes nitric oxide from l-arginine
ATRXAlpha thalassemia/mental retardation syndrome X-linked2.753.70A chromatin remodeling protein
HIST1H2BGHistone cluster 1, H2bg2.122.18Responsible for nucleosome structure in the eukaryotic chromosomal fibers
SMC3Structural maintenance of chromosomes 32.052.04Part of the cohesion complex that holds together sister chromatids during mitosis
HOOK1Hook homolog 1 (Drosophila)2.573.43Interacts with several members of the Rab GTPase family, associated with endocytosis. It is believed to link endocytic membrane trafficking to the microtubule cytoskeleton
MLPHMelanophilin2.232.71Encodes a protein that forms a ternary complex with the Ras-related GTPase Rab27A and the motor protein myosin Va
EXOSC6Exosome component 62.653.08One of the subunits of exosome that mediates mRNA degradation
SULT1C4Sulfotransferase family, cytosolic, 1C, member 42.592.91Belongs to the SULT1 subfamily, responsible for transferring a sulfo moiety from PAPS to phenol-containing compounds
APBB1IPAmyloid beta (A4) precursor protein-binding, family B, member 1 interacting protein2.186.30No listed gene function

Quantitative real-time PCR was then performed to verify the relative expression of the above 9 genes. As shown in Fig. 4a and b, four genes from the list, that is, nNOS, Hook homolog 1, Melanophilin, and APBB1IP, were confirmed to exhibit lower levels of expression in migratory hiPSC-NSCs, below 50% of the expression levels in non-migratory cells under both testing conditions. nNOS, neuronal nitric oxide synthase, was the most significantly down-regulated gene in migratory hiPSC-NSCs: 89% down-regulation in the assay using U87-conditioned medium and 84% in the assay using 4T1-conditioned medium. To investigate the differentiation status of migratory and non-migratory hiPSC-NSCs, we examined the expression of NSC markers, CD133, Nestin, and Sox2, and glial precursor cells markers, PDGFR and NG2, using RT-PCR and detected no difference in expression of the differentiation marker genes between migratory and non-migratory hiPSC-NSCs (Fig. 4c). Quantitative real-time PCR analysis further confirmed the observation (data not shown). Our findings indicate that while nNOS signaling might be involved in regulating NSC migration, nNOS expression levels are not an obvious indicator for NSC differentiation.

Figure 4.

Gene expression in migratory and non-migratory human-induced pluripotent stem cells (hiPSC)-neural stem/progenitor cells (NSCs). (a) and (b) Verification of expression of down-regulated genes in migratory hiPSC-NSCs. hiPSC-NSCs migrat-ing toward U87 (a) and 4T1 cells (b) and their paired non-migratory cells were collected. Nine down-regulated genes detected with microarray analysis were analyzed using quantitative PCR. The data were normalized to internal controls GAPDH, β-actin and beta-2 microglobulin (B2M). Values are mean ± SD (n = 3). (c) NSC marker expression in migratory and non-migratory hiPSC-NSCs. Migrating (M) and non-migratory (N) hiPSC-NSCs toward 4T1 or U87 cancer cells were collected for RT-PCR analysis. B2M was included as an internal control.

Effects of nNOS inhibitors on migration of NSCs toward cancer cells

On the basis of our microarray and qPCR results, we moved forward to test whether inhibition of nNOS activity could promote the migration of NSCs toward cancer cells. hiPSC-NSCs were pre-treated with nNOS inhibitor SMTC (Rakshit et al. 2009) or nNOS inhibitor I (Overend and Martin 2007) for 1 day and tested for their migration capacity in Boyden chamber assays. We observed that the pre-treatment with both nNOS inhibitors led to a significant increase in migration of hiPSC-NSCs toward U87 or 4T1 cancer cells and there was no significant difference between the effects of the two tested drugs. As previously reported, nNOS inhibition could promote cell proliferation (Maeda et al. 2000; Ciani et al. 2004). To minimize the interference of this effect on cell migration results, we reduced the cell migration time from 16 h to 6 h and could still observe the increased migration of hiPSC-NSCs, from 15% of migratory cells in the control group without treatment to close to 25% in the groups treated with the nNOS inhibitors (Fig. 5a). Since the doubling time of hiPSC-NSCs is 3–4 days, the increase in the number of migratory cells in 6 h was unlikely to have resulted from the increase in cell proliferation within such a short period. In a cell proliferation assay, we did not observe any significant difference in the change of cell number between hiPSC-NSCs treated with the nNOS inhibitors and the control without treatment within 24 h (data not shown). We also examined whether treatment with nNOS inhibitor SMTC and nNOS inhibitor I would possibly increase migration of 4T1 and U87 tumor cells and observed no obvious differences in cell migration in Boyden chamber assays between treated and non-treated tumor cells (Fig. 5b).

Figure 5.

Effects of neuronal nitric oxide synthase (nNOS) inhibitors on cell migration in Boyden chamber. (a) Effects of nNOS on i human-induced pluripotent stem cells-neural stem/progenitor cells (NSC) migration toward tumor cells. NSCs were cultured in the upper well in a Boyden chamber in the presence or absence of nNOS inhibitors (S-methyl-L-thiocitrulline, 500 μM or nNOS inhibitor I, 20 μM). (b) Effects of nNOS on tumor cell migration. 4T1 or U87 cancer cells were cultured in the upper well in a Boyden chamber and treated with nNOS inhibitors as described above. DAPI images of migratory cells in (a) and (b) were taken after migration for 6 h and the number of migratory cancer cells was counted. Scale bar, 500 μm. Values are mean ± SD (n = 3). Error bars in (a) and (b) indicate SD (n = 3). **p < 0.01, ***p < 0.001 by t-test versus Control.

We further tested whether knock-down of nNOS expression could promote the migration of NSCs toward cancer cells. Two specific siRNAs targeting two different regions of four transcript variants of human nNOS mRNA were transfected into hiPSC-NSCs. A reduction of nNOS mRNA level of about 60% to 80% was observed 48 h later (Fig. 6a and b). The knockdown effects on protein expression by both siRNAs were validated by western blotting (Fig. 6c). The biological effects of nNOS knockdown were confirmed in in vitro Boyden chamber assays, showing statistically significant increase in migration of the siRNA-transfected hiPSC-NSCs toward both 4T1 and U87 tumor cells (Fig. 6d and e), suggesting that nNOS functions as a suppressor of tumor tropism of NSCs. Overall, our findings indicate that inhibition of nNOS activity may sensitize poorly migratory NSCs and enhance the migration of these cells toward tumor cells.

Figure 6.

Effects of neuronal nitric oxide synthase (nNOS) siRNAs on migration of human-induced pluripotent stem cells-neural stem/progenitor cells (NSCs) in Boyden chamber. (a, b, c) Knockdown of nNOS by two siRNAs (40 nM) validated by RT-PCR (a), quantitative real-time PCR (b) and Western blotting (c). B2M was used as specificity control in (a) and β-Actin was used as loading control in (c). Densitometric analysis in (c) was done by Image J software (National Institutes of Health, Bethesda, MD), and values were normalized to the control. (d, e) Knockdown of nNOS by two siRNAs promotes human-induced pluripotent stem cells-NSC migration toward 4T1 (d) and U87 (e) cancer cells. DAPI images of migratory NSCs were taken and the number of migratory NSCs was counted after 6 h. Scale bar, 500 μm. Values are mean ± SD (n = 3). Error bars in (b), (d) and (e) indicate SD (n = 3). *p < 0.05, **p < 0.01 by t-test versus si-Ctrl.


The ability of malignant tumors to infiltrate in the human body and the ineffectiveness of conventional therapies, such as surgery, radio- and chemotherapy, to eliminate disseminated tumors partially attribute to poor prognosis of these tumors. There is an urgent need to develop novel treatments to overcome the limitations of current cancer therapies. NSCs are able to home in on not just brain tumors but also solid tumors of a non-neural origin (Aboody et al. 2000, 2006; Brown et al. 2003; Yang et al. 2012). Especially, the tumor tropism of NSCs has been utilized for targeted delivery of anti-cancer genes to both major tumor mass and distant tumor outgrowths in animal tumor models. A phase 1 clinical trial of gene therapy for recurrent glioblastoma multiforme using a human NSC cell line to deliver a suicide gene is currently ongoing at the City of Hope Medical Center in Duarte, California. We have previously reported that mouse iPSC-NSCs, hESC-NSCs and hiPSC-NSCs can be used for targeted gene therapy for glioma and breast cancer (Lee et al. 2011; Yang et al. 2012; Zhao et al. 2012). We investigated in this study putative molecular mechanisms underlying tumor tropism of iPSC-NSCs and whether the tumor tropism can be enhanced by chemical drug manipulation.

This study is the first attempt to use microarray-based gene expression profiling technology, a technology that allows gene expression analysis with a high-throughput, to map gene expression profiles in hiPSC-NSCs. Given the fact that limited information is available so far on the molecular mechanisms underlying NSC migration toward tumor, we used Affymetrix high density array chips for analysis of up to 45,000 transcripts in migratory or non-migratory NSCs collected from migration assays toward either human glioma tumor cell-conditioned medium or mouse breast cancer cell-conditioned medium.

Comparing the up- and down-regulated genes in hiPSC-NSCs migrating toward 4T1 and U87 conditioned medium, we identified several commonly down-regulated genes associated with intracellular non-membrane-bounded organelles, suggesting that these genes may be significant in regulating the migration of NSCs toward tumors. Intracellular non-membrane-bounded organelles are organized cellular structures with distinctive morphology and functions but not bounded by lipid bilayer membranes (DAVID, Bioinformatics Resources 6.7). These organelles include chromosomes, ribosomes, and the cytoskeleton. Although general functions of chromosomes and ribosomes are not directly related to cell migration, cytoskeleton dynamics plays important roles in cell migration including NSC migration (Kasper et al. 2009). Of the genes associated with intracellular non-membrane-bounded organelles that were confirmed with qPCR analysis, NOS appears highly interesting, as it has been investigated in previous studies for its functions associated with cell migration.

Nitric oxide (NO) is a radical molecule involved in various physiological and pathophysiological processes. It is synthesized from L-arginine by NOS, which exists in three isoforms based on the tissue source, inducible-NOS, endothelial-NOS and finally nNOS (Moreno-Lopez et al. 2000; Su et al. 2005; Zhang et al. 2007). These three isoforms of NOS are co-localized with the cytoskeleton including microtubules, actin microfilaments, and intermediate filaments directly or indirectly. This association assists NOS in performing their function. Re-organization of cytoskeleton, which can be induced by extracellular signals, will indirectly affect NO production. Thus, altered NO production by cytoskeletal reorganization might be important in physiological and pathophysiological conditions. In the nervous system, nNOS takes part in various functions such as vascular homeostasis, neurotransmission, host defense and angiogenesis (Su et al. 2005).

Several studies have demonstrated a connection between NOS expression and migration of neuronal cells. One such study performed on rats subject to cerebral ischemia indicated a decline in nNOS expression post-ischemia in migrating cells originated from the ependymal and subventricular zone, the regions in the mammalian brain that retain the ability for neurogenesis in adulthood because of the presence of neural stem cells (Zhang et al. 2007). The study also noted a decreased level of nNOS expression in cerebral regions were the migrating cells passed through, suggesting that decreased nNOS expression in the migration routes could be related to the cell migration. These findings were consistent with an earlier study by Moreno-López et al. (2000), in which they found that neural precursor cells did not express nNOS either in proliferation or migration regions of adult mouse brains. Furthermore, endothelial-NOS was found to attenuate pulmonary leukocyte migration via inhibition of endothelial adhesion molecule expression (Kaminski et al. 2004).

Also reported previously, there exists a positive relationship between an increased level of NO and NSC differentiation (Moreno-Lopez et al. 2000; Luo et al. 2010) and NOS inhibition enhances NSC proliferation but decreases their differentiation (Maeda et al. 2000; Ciani et al. 2004; Luo et al. 2010). In our study, older passages of NSCs, which are more differentiated and hence might have a higher level of NO, have a decreased migratory capacity (data not shown). The findings from our NSCs migration study correlate well with our microarray results, supporting a hypothesis that up-regulation of nNOS increases NO production and thus differentiation, which in turn might reduce cytoskeletal dynamics, leading to reduced cell migration. Thus, down-regulation of nNOS would on the other hand, decrease intracellular NO level, acting as an effective way to enhance NSC migration toward tumors.

However, other studies performed on rats have indicated an increased proliferation followed by increased migration of neural precursor cells post-ischemia when NO donor was administrated (Zhang et al. 2001). Publications by Tegenge et al. (2011) have indicated that NO stimulates migration of human neural progenitor cells under healthy conditions via a cGMP-mediated signal transduction pathway by activating the soluble enzyme, guanylyl cyclase using NO (Tegenge and Bicker 2009; Tegenge et al. 2011). These studies also showed that the migration of these cells is blocked by the use of nNOS inhibitors and enhanced by the use of an NO donor. The discrepancies between these studies and those showing increased NSC migration associated with decreased nNOS expression could be explained by the difference between physiological and pathological conditions and/or different types of pathological conditions, the species of biological samples used, as well as the neural precursor cell type, age, and NO source. In a previous report, regulation of neurogenesis by nNOS was found to be bi-directional in a way that neuron-derived nNOS and exogenous NO donors decrease neurogenesis and neuronal differentiation, whereas NSC-derived nNOS enhances neurogenesis (Luo et al. 2010). Hence, altered nNOS source may result in altered functions. Since no previous studies have been conducted to investigate the effects of NO on the migration of human NSCs toward tumors, our findings on nNOS down-regulation in migratory NSCs deserve further scrutiny. An interesting topic for future research would be whether nNOS inhibitors can increase NSCs tumor tropism in vivo.

The generation of hiPSCs starts from adult cells, circumventing the bioethical controversies associated with the derivation of hESCs from human embryos (Narsinh et al. 2011). When somatic cells isolated from patients are re-programmed into iPS cells, transplantation of the differentiated progeny of the generated iPSCs will less likely elicit immune rejection responses in the patients. In the case that a hiPSC bank consisting of various human leukocyte antigen types is established (Tamaoki et al. 2010), iPSCs selected from the cell bank can be alternatively used to generate tumor-targeting NSCs with reduced likelihood of immune-mediated rejection. Furthermore, the approach of using iPSCs to derive NSCs is applicable for large-scale mass manufacture of uniform batches of cellular products that are sufficient for repeated patient treatments. Importantly, this approach will help eliminate variability in the quality of cellular products, thus facilitating reliable comparative analysis of clinical outcomes. Large-scale manufacturing of human cells from hiPSCs, instead of collecting and expanding primary cells from individual patients, will increase cost-effectiveness by reducing the laboriousness and simplifying the logistics of cell culture operations. These NSCs can be further prepared as commercial products in a cryopreserved, ready-to-go format. Hence, iPSC-NSCs hold great potential for cancer treatment.

It should be noted that hiPSCs used in this study were generated by transduction with a polycistronic lentiviral vector containing reprogramming factor genes, Oct4, Klf4, Sox2, and c-Myc genes. This expression cassette is excisable by cre-recombinase (Sommer et al. 2010). Although the presence of the reprogramming factor genes in hiPSCs did not impede neural differentiation to generate tumor-tropic NSCs as shown in this study, excision of the expression cassette could possibly improve the developmental potential of hiPSCs and significantly augments their capacity to undergo directed differentiation (Sommer et al. 2010). More importantly, since several re-programming factors, for example, c-Myc, Klf4, and Oct4, have the potential to induce genomic instability and possess oncogenic activity, great caution should be exercised by removing the expression cassette before therapeutic applications of hiPSC-derived cells.


This study was supported by the Singapore Ministry of Health's National Medical Research Council (NMRC/IRG10Nov122), the Singapore Ministry of Education (MOE2011-T2-1-056), Singapore Agency for Science, Technology and Research Joint Council (11/03/FG/07/02), and Institute of Bioengineering and Nanotechnology (Biomedical Research Council, Agency for Science, Technology and Research, Singapore). The authors have declared that no conflict of interest exists.