Introduction

Established cancer cell lines are routinely used to study a range of biological characteristics of cancer, such as proliferation, migration, chemo-resistance, and tumorigenicity (Bal et al. 2017; Lapek et al. 2017; Lu et al. 2017; Nakayama et al. 2017; Vu et al. 2017). The human glioblastoma U87MG cell line is one of the most commonly used glioblastoma cell line. It was established in 1966 by researchers at the University of Uppsala, Sweden using tissues from a 44-year-old glioblastoma patient. The tissues were kept in a cell repository and later transferred to ATCC (Allen et al. 2016). The widespread use of this cell line has facilitated the research of glioma. However, Allen et al. found that the U87MG cell line procured from ATCC was different from that of the original one (Allen et al. 2016), indicating the probable misidentification during serial passage.

Besides of probable contamination and misidentification, serial in vitro passage may also affect the reproducibility of in vitro experiments (Jin et al. 2017; Lin et al. 2003; Moutsatsou et al. 2000). However, the passage number of the cells employed in experiments and the experiment period are seldom considered in scientific papers. Considering this, different passage of cell lines may lead to controversial results among different research teams. But few researchers focused on the changes of cancer cell lines during passage, even though these cell lines may have been passaged for several decades.

To our knowledge, our study provides the first evidence for the different tumorigenicity between early and late passage U87MG cells, and the probable underlying molecular transition between these two cell strains. This emphasizes the necessity to shorten experiment period and accomplish experiments with the same or similar passage cancer cell strains.

Materials and Methods

Cell Culture

The human glioblastoma cell line U87MG was purchased from the Chinese Academy of Sciences (Shanghai, China). The cell line was grown in Dulbecco’s modified Eagle’s medium (DMEM) (Hyclone, Logan, UT) and supplemented with 10% fetal bovine serum (FBS) (Gemini, Shanghai, China). The cells were then cultured at 37 °C in a humidified atmosphere of 5% CO2 environment. During the passage, cells were plated at a density of 1.7–2.0 × 104 cells/cm2, and the media was replaced every 3 days. Cells were then passaged when they reached around 80–90% confluence.

Orthotopic Xenograft

An orthotopic xenograft model was established according to previously described methods (Song et al. 2017). Female athymic mice (4–6 weeks old) were purchased from Sun Yat-sen University (Guangzhou, China), and divided into two groups of five mice per group. After acclimating for 1–2 weeks, the mice were anesthetized with 1% pentobarbital sodium and placed in a stereotactic fixation device (RWD Life Science, Shenzhen, China) prior to cell inoculation. A total of 1 × 106 U87MG cells in 5 ul of sterile PBS were stereotactically implanted into the point located 1 mm posterior to the bregma, 2 mm left of the midline, and 2.5 mm deep into the brain. Cells were slowly injected over 5 min. The mice were sacrificed individually upon they became moribund, and the life span was recorded. The mice brains were then fixed in 4% paraformaldehyde for 24 h and embedded in paraffin. For the use of xenograft models, approval from the Ethics Committees of Nanfang Hospital of Guangdong Province was obtained. All animal studies were conducted in accordance with the principles and procedures outlined in the National Institutes of Health Guide for the Care and Use of Animals.

Immunochemistry (IHC)

For IHC, tissue sections were cut into 4-µm sections from paraffin blocks of mice brains, deparaffinized in xylene, and then rehydrated in degraded ethanol (100, 95, 80, 70%). Tissue sections were deparaffinized as described above. After washing thoroughly with PBS, the slides were treated with 10 mM citrate buffer for 2 min at 100 °C to achieve antigen retrieval, and then treated with 3% hydrogen peroxide to block endogenous peroxidase activity. After washing thoroughly with PBS, slides were blocked with 5% BSA for 1 h at room temperature, and incubated with primary antibodies, including Nestin (ZSGB-bio, China), E-cadherin, and N-cadherin (Cell signaling technology, USA) at 4 °C overnight. After washing thoroughly with PBS, the slides were incubated with biotin-labeled goat anti-mouse or anti-rabbit antibodies (ZSGB-bio, China) for 1 h at room temperature. The sections were finally visualized with DAB, counterstained with hematoxylin, mounted in neutral gum, and imaged using a bright field microscope equipped with a digital camera (Leica, Germany).

Cell Viability Assay

Cell viability was analyzed by MTT assay as described previously (Song et al. 2014a). U87MG cells were seeded in 96-well plates at a density of 1000 cells/ well with five replicates. After most cells were attached, 20 µl of MTT (5 mg/ml in PBS) (Sigma, St Louis, MO) was added into the wells and the cells were incubated for 4 h at 37 °C. After incubation, the supernatants were removed and 150 µl dimethyl sulfoxide (DMSO) (Sigma, St. Louis, MO) was added. The optical density (OD) of each well was detected at 490 nm. For the next 7 days, MTT was performed as described above at the same time of day and the OD was collected.

EdU Assay

Cell viability was analyzed by Edu assay as described previously (Song et al. 2014a). U87MG cells were seeded into 96-well plates at 2000 cells per well for 48 h, and then exposed to 50 µM EdU reagent (Ribobio, China) for 2 h before fixation with 4% formaldehyde and treatment with 0.5% TritonX-100. After washing gently with PBS for three times, the cells were exposed to 100µl 1 × Apollo reaction reagent for 30 min. Subsequently, the nuclei were stained with DAPI (Beyotime, China) for 10 min. Finally, images were captured with fluorescent microscope.

Colony-Forming Assay

The U87MG(I) and U87MG(II) cells were plated in 12-well plates (200 cells per well), respectively, and cultured for 2 weeks. The colonies were then fixed with methanol for 30 min and stained with 1% crystal violet for 1 min. All assays were independently performed in triplicate.

Wound-Healing Assay

The wound-healing assay was conducted as previously described (Song et al. 2017). To assess cell migration, U87MG cells were seeded into 6-well plates. When the cells grew to around 90% confluence, three scratch wounds were made across each well using a P-200 pipette tip. Fresh medium supplemented with 10% FBS was added, and the wound-healing procedure was observed after 12 and 24 h, respectively.

Cell Migration and Invasion Assays

In vitro cell migration and invasion assays were performed according to the previous study (Qi et al. 2012). For the cell migration assay, 1 × 105 cells were seeded into the upper chamber of the Transwell apparatus (Costar, MA) with 100 µl DMEM medium without FBS, and 600 µl DMEM with 10% FBS was added to the lower chamber as a chemoattractant. The cells were then incubated for 2 h at 37 °C in a humidified atmosphere of 5% CO2. The apparatus was then washed with PBS, and the cells adhering to the lower surface of the apparatus were fixed with methanol, stained with Giemsa solution (NJJC Bio, China), and counted under a microscope in four predetermined fields (200×). A cell invasion assay was performed similarly to the cell migration assay, except for that the Transwell membranes were pre-coated with 250 µg/ml Matrigel (R&D Systems, USA). The cells were then incubated for 4 h at 37 °C in a humidified atmosphere of 5% CO2. The cells were treated and counted in the same way as in the Transwell assay. All assays were conducted in triplicate.

Western Blot Analysis

Western blot was carried out as previously described (Song et al. 2014b), using the following primary antibodies: E-cadherin, N-cadherin, β-catenin, Slug, Snail, Vimentin (Cell signaling technology, USA); PI3K, PI3K(phosopho-Tyr467/199), Akt, Akt(phosphor-Ser473) (SAB, USA); Actin (CWbio, China). The antibodies were then immunoblotted with an HRP-conjugated anti-mouse or anti-rabbit immunoglobulin-G antibody (CWBio; 1:2000). The bound antibodies were then detected using enhanced chemiluminescence reagents (Fdbio, China) with a ChemiDoc™ XRS + system (Bio-Rad, USA). Signal intensities were obtained using Image-Pro Plus 6.0. All the Western blot experiments were performed in triplicate.

Statistical Analysis

All quantified data represented an average of at least three replicates. SPSS 13.0 and Graph Pad Prism 5.0 were used for statistical analysis. Data are presented as mean ± SD. A one-way ANOVA or two-tailed Student’s t test was used for comparisons between groups. Survival analysis was performed using the Kaplan–Meier method. Multivariate Cox proportional hazards method was used for analyzing the relationship between the variables and survival time. Differences were considered statistically significant for p < 0.05.

Results

Late Passage U87MG Cells Exhibited Different Morphology and Lower Tumorigenicity

The U87MG cell strain purchased from ATCC was kept under liquid nitrogen cryopreservation. Late passage which had been revived about 1 year before the experiment underwent regular, serial in vitro passage continually, and was designated as U87MG(II). Early passage, which was revived from the same cryopreserved cell strain as U87MG(II), was revived before the experiment started and designated as U87MG(I). The relative difference in passage number was more than 100. To observe cell morphology differences, approximately 1.3–1.5 × 105 of these two cell strains were planted into 100-mm culture dishes, respectively. Morphology and density were then observed and captured under a microscope 12, 24, 48, and 72 h later. No significant difference in density was observed. However, after 48 h, U87MG(I) was neatly arranged and formed obvious network structures with large mesh space, while U87MG (II) netted dispersedly without obvious mesh space (Fig. 1a). To compare the tumorigenicity of these two cell strains, an orthotopic glioma xenograft model was established in female athymic mice within 4–6 weeks, which were divided into two groups of five mice per group (Fig. 1b). The mice were sacrificed individually when they became moribund, and a histological examination of their brains was performed. While both strains exhibited strong tumorigenicity, U87MG(I) showed a higher tumorigenicity than U87MG(II) and led to decreased overall survival (Fig. 1c, d). By recourse to IHC, we further confirmed that N-ca was more highly expressed in the U87MG(I) formed tumors(Fig. 1c, p = 0.0414), indicating that U87MG(I) exhibited more mesenchymal characteristics than U87MG(II). Therefore, these results demonstrated that the morphology and tumorigenicity of the U87MG cell line varied during serial passage, with late passage exhibiting decreased tumorgenicity and less mesenchymal characteristics.

Fig. 1
figure 1

Late passage U87MG cells exhibited different morphology and lower tumorigenicity. a The morphology and density of two cell strains were observed 12, 24, 48, 72 h after seeding. (100×, scale bar = 200 µm). b Schematic diagram of orthotopic glioma xenograft model. c The tumor tissues formed by U87MG(I) and U87MG(II) were examined by IHC staining (200×, scale bar = 100 µm), respectively. d Mice bearing U87MG(I) formed tumor had decreased overall survival (p = 0.0357). Data are presented as mean ± s.d. from three independent experiments (NS no significant difference, *p < 0.05, **p < 0.01)

Full size image

No Significant Differences of in vitro Cell Growth were Observed between the Two Cell Strains

The cell viability differences between the lower and higher passage of U87MG cells were then compared in vitro. The results of the MTT assay revealed no significant difference (p = 0.1038, Fig. 2a) in proliferation rate between U87MG(I) and U87MG(II). In the EdU assay, the mean percentage of positive proliferative cells of U87MG(I) was 28.46% compared to 25% in U87MG(II) (p = 0.412, Fig. 2b). In addition, colony-forming ability was also assessed, and no significant difference was apparent (p = 0.6507, Fig. 2c). It suggested that no significant difference was found in cell proliferation capacity between early and late passage. Therefore, this suggested that other factors leading to the different tumorigenicity of these two strains should be considered.

Fig. 2
figure 2

No significant difference of in vitro cell growth was observed between two cell strains. a MTT assays were performed to detect the proliferation rates of U87MG(I) and U87MG(II), respectively. b EdU assays were performed to detect the positive proliferative cells in U87MG(I) and U87MG(II), respectively (200×, scale bar = 100 µm). c Colony-forming assay was performed to detect the proliferative ability in U87MG(I) and U87MG(II), respectively. Data are presented as mean ± s.d. from three independent experiments (NS no significant difference, *p < 0.05, **p < 0.01)

Full size image

Early Passage Possessed Higher Invasion than Late Passage

Considering that migration and invasion may also involve in tumorigenicity, Transwell, wound-healing, and Boyden assays were performed. In Transwell assays, the number of migrated U87MG(I) cells was approximately the same as the number of migrated U87MG(II) cells (p = 0.9779, Fig. 3b). This result was consistent with that found in the wound-healing assay (p = 0.768, Fig. 3a). However, examining the invasion of these two strains in the Boyden assay revealed higher invasiveness in U87MG(I) (p = 0.0054, Fig. 3c). Therefore, these results demonstrated that decreased invasion might be responsible for the decreased tumorigenicity found in late passage.

Fig. 3
figure 3

Early passage possessed higher invasiveness than late passage. a Wound-healing assays were performed to compare the migration ability between U87MG(I) and U87MG(II) (200×, scale bar = 100 µm). b Transwell assays were performed to compare the migration ability between U87MG(I) and U87MG(II) (200×, scale bar = 100 µm). c Boyden assays were performed to compare the invasiveness between U87MG(I) and U87MG(II) (200×, scale bar = 100 µm). Data are presented as mean ± s.d. from three independent experiments (NS no significant difference, *p < 0.05, **p < 0.01)

Full size image

EMT Pathway, PI3K/Akt Pathway, and TGF-β Pathway Differences Were Detected between U87MG(I) and U87MG(II)

Since less mesenchymal characteristics and decreased invasion were identified in U87MG(II), we further examined epithelial-to-mesenchymal transition (EMT) pathway and two other crucial pathways that involved in invasion, including phosphatidyl inositol-3-kinase(PI3K)/Akt pathway and transforming growth factor beta (TGF-β) pathway in these two cell strains. In the EMT pathway, while E-cadherin was upregulated in U87MG(II), mesenchymal biomarkers including N-cadherin, β-catenin, and snail were significantly downregulated, suggesting that U87MG(II) exhibited less mesenchymal characteristics than U87MG(I). It concurs with that found in in vivo experiments. Conversely, two mesenchymal biomarkers, Vimentin and Slug, were upregulated in U87(II) (Fig. 4a). Similar results were found in the PI3K/Akt pathway. Although no significant differences were found in PI3K nor in p-PI3K, Akt and p-Akt were decreased significantly in U87MG(II) compared to that in U87MG(I) (Fig. 4b). As the level of TGF-β RII was decreased in U87MG(II), both Smad3 and p-Smad3 showed lower expression in U87MG(II) than in U87MG(I). Although Smad2 was increased in U87MG(II), no significant difference of p-Smad2 was detected between early passage and late passage. MMP9, another important protein involved in invasion and regulated by TGF-β pathway, was also decreased in U87MG(II) (Fig. 4c). In conclusion, to some extent, late passage exhibited more epithelial characteristics and decreased expression of PI3K/Akt pathway and TGF-β pathway, demonstrating the probable molecular transitions during serial passage.

Fig. 4
figure 4

EMT, PI3K/Akt, and TGF/β pathways differences were detected between U87MG(I) and U87MG(II). a Expressions of EMT pathway related proteins were analyzed by Western blot in both U87MG(I) and U87MG(II). b Expressions of PI3K/Akt pathway related proteins were analyzed by Western blot in both U87MG(I) and U87MG(II). c Expressions of TGF/β pathway related proteins were analyzed by Western blot in both U87MG(I) and U87MG(II). Data are presented as mean ± s.d. from three independent experiments (NS no significant difference, *p < 0.05, **p < 0.01)

Full size image

Discussion

Glioblastoma cell line U87MG is a well-established cell line that widely used in glioma research, but few focused on the change of its biological characteristics during serial passage. In the present study, the morphologic and tumorigenicity of the cell strains were observed to change after serial passage. Although no significant differences in cell viability and migration were observed between early and late passage, but late passage showed decreased invasion. In particular, we found that serial passage of U87MG may result in the transition to epithelial phenotypes with decreased level of PI3K/Akt pathway and TGF-β pathway. This might explain, to some extent, the differences observed between these two cell strains with different passage, though the precise mechanism is yet to be elucidated.

Although established cancer cell lines are routinely used, two issues should raise researcher’s attention. First, during the passage of cell lines, different cell lines may be cross-polluted or mislabeled, and some manuscripts have been withdrawn due to confusing cell lines during experiments (Zhang et al. 2013). Several scientific journals now require cell line authentication before submitting manuscripts (Allen et al. 2016; Huang et al. 2016). Second, even with cell line authentication, the influence of passaging on cell lines cannot be confirmed and avoided. The passage can affect cells in various ways, and therefore influence the reproducibility of in vitro experiments (Jin et al. 2017). In rat glioma C6 cells, continuous passaging led to different numbers of beta-adrenergic receptors and phosphodiesterase activity, which has been observed to alter the beta-adrenergic responsiveness of C6 high passage (Mallorga et al. 1981). The risk that accumulating genetic and epigenetic alterations during in vitro passage may lead to malignant cell transformation also challenges the cell therapy use of human mesenchymal stem cells (Binato et al. 2013; Maitra et al. 2005). However, the specific changes of U87MG cells during serial passage remain elusive.

Transition between epithelial and mesenchymal phenotypes regulates tumor cell motility and invasion, and subsequently affect its tumorigenic potential (Jie et al. 2017). In particular, EMT plays a vital role in invasion and involves several molecular changes, including upregulation of epithelial markers such as E-cadherin and claudins and the downregulation of mesenchymal markers such as N-cadherin and Vimentin (Jie et al. 2017; Yang and Weinberg 2008). This transition is promoted by several EMT-inducing transcription factors(EMT-TFs) such as Snail and Slug (Nieto et al. 2016). These molecular changes during EMT lead to loosened intercellular junctions and reorganization of adhesive molecules as well as the cytoskeleton, which may convert the tumor cells into spindle-like shape morphology that is more appropriate for invasiveness (Bonnomet et al. 2010; Ikenouchi et al. 2003; Lamballe et al. 2016). Nevertheless, EMT can facilitate the tumor cell tumorigenic potential in hostile environmental conditions by rescuing them from apoptosis, anoikis, and senescence (Singh and Settleman 2010; Tiwari et al. 2012). In our study, we found that late passage adopted morphological change with more cell contacts compared to early passage. Consistent with this phenotypic modification, late passage exhibited more epithelial phenotype and less tumorigenicity in xenograft models. Of note, EMT is rarely fully activated in tumor cells, and mesenchymal markers such as vimentin may not be always overexpressed (Brabletz et al. 2018; Nieto et al. 2016). Moreover, increased Slug in late passage indicates the profound effects of EMT-TFs in glioblastoma (Kahlert et al. 2017). As increased Vimentin was also found in late passage, we proposed that serial passage may lead to partial reversion of EMT in U87MG, during which intercellular junctions were decreased, whereas little cytoskeleton reconstruction was observed.

Several important signaling pathways also involve in EMT and tumor cell invasion, including PI3K/Akt pathway and TGF-β pathway. The PI3K/Akt pathway is a novel regulatory pathway under cellular stress (Datta et al. 1999). When PI3K binds to the phosphor-tyrosine consensus residues of growth factor receptors, second messengers such as Akt/protein kinase(PKB) accumulate and promote the cell cycle progression, cell growth (Porta et al. 2014). Crosstalk between PI3K/Akt and TGF-β pathway can further promote tumor cell invasion (Rumman et al. 2016; Zhou et al. 2017). TGF-β activates the membrane receptor serine/threonine complex formed by TGFβR-II and TGFβR-I, and then phosphorylates Smad2 and Smad3, forming a Smad complex accumulated in the nucleus to regulate the downstream transcription (Eichhorn et al. 2012). Although TGF-β plays as a tumor suppressor in normal epithelial cells and early-stage tumors (Kitamura et al. 2007), activation of TGF-β pathway in advanced tumors is oncogenic and stimulates the expression of matrix metalloproteins (MMPs) family, such as MMP9, and then further enhances tumor invasion and migration (Lu et al. 2011; Park et al. 2016; Shi et al. 2017). Although we observed downregulated PI3K/Akt pathway and TGF-β pathway at protein level in late passage U87MG cells, we only found lower invasion in late passage but no significant differences in proliferation between these two cell strains, implying that a more complicated feedback regulation in tumor cell proliferation should be considered during serial passage.

In conclusion, the biological characteristics of traditionally utilized cancer cell lines may change during serial passage, and partially account for the controversial results between distinct research teams. But few researchers recognized the influence of passage-induced changes on cancer research and our results filled the gap. To minimize these changes and improve the reproducibility of cancer cell line experiments, shortening the experiment-span and accomplishing them with the same or similar passage cell strains should be necessary.