Downregulating CD26/DPPIV by apigenin modulates the interplay between Akt and Snail/Slug signaling to restrain metastasis of lung cancer with multiple EGFR statuses

API (A3145) and dimethyl sulfoxide (DMSO) were purchased from Sigma-Aldrich (St. Louis, MO). Fetal bovine serum (FBS), antibiotics, molecular weight standards, trypsin-EDTA, trypan blue stain, and all medium additives were obtained from Life Technologies (Gaithersburg, MD). An enhanced chemiluminescence kit was purchased from Amersham (Arlington Heights, IL). An anti-matrix metalloproteinase (MMP)-3 antibody was purchased from Epitomic (Burlington, CA). Antibodies specific for fibronectin and MMP-9 were obtained from Abcam (Cambridge, MA). Antibodies specific for CD26, MMP-2, Snail, Slug, Twist, phosphatase and tensin homolog (PTEN), and unphosphorylated and phosphorylated (p-) forms of the corresponding Akt and epidermal growth factor receptor (EGFR) were obtained from Cell Signaling Technology (Danvers, MA). Antibodies specific for vimentin and N-cadherin were purchased from BD Biosciences (San Jose, CA). Antibodies specific for presenilin-1 and β-actin were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Polyvinylidene fluoride (PVDF) membranes for Western blotting were purchased from Bio-Rad (Hercules, CA). Unless otherwise specified, other chemicals used in this study were purchased from Sigma Chemical (St. Louis, MO).

The A549, H1975, and HCC827 NSCLC cell lines were purchased from American Type Culture Collection (ATCC, Manassas, VA), and a series of NSCLC cell lines, CL1–0, CL1–3, and CL1–5, in ascending order of invasiveness, were established in the National Health Research Institute laboratory [31]. PC9 cells were developed by Lee and colleagues at National Cancer Center Hospital (Tokyo, Japan) [32]. All cells were maintained in RPMI 1640 supplemented with 10% FBS and 1% penicillin-streptomycin-glutamine. All cells were incubated at 37 °C in a humidified 5% CO₂ atmosphere.

A549, CL1–5, HCC827, and H1975 NSCLC cells (5 × 10³) were seeded in 96-well plates, treated with various concentrations of API (5~ 80 μM) for 24 and 48 h, and then subjected to a cell-viability assay (MTS assay; Promega, Madison WI) according to the manufacturer’s instructions. Data were collected from three replicates.

NSCLC cells were diluted and seeded at 1000 cells/well in six-well plates and incubated for 24 h. Subsequently, various concentrations (5~ 40 μM) of API were added for 24 h, and then continuously incubated in new fresh medium at 37 °C. After incubation for 7~ 10 days, cells were stained with crystal violet, and a colony was defined as consisting of more than 50 cells.

Migration and invasion assays were performed according to our previous study [33]. Briefly, 3 × 10⁴ cells were plated in a uncoated top chamber (24-well insert; pore size, 8 μm; Corning Costar, Corning, NY) for the transwell migration assay. The invasion assay used 4 × 10⁴ cells (A549 and HCC827) or 3 × 10⁴ cells (CL1–5 and H1975) plated in a Matrigel (BD Biosciences, Bedford, MA)-coated top chamber. In both assays, cells which had been pretreated for 24 h with API (5~ 40 μM) were plated in medium without serum or growth factors, and medium supplemented with 10% serum was used as a chemoattractant in the lower chamber. Cells that were allowed to migrate or invade for 24 h were fixed with methanol and stained with crystal violet. The number of cells migrating through or invading through the membrane was counted under a light microscope (× 100 or × 200, three random fields per well).

IF techniques were used to observe actin rearrangement after treatment of NSCLC cells with API or the vehicle. Cells were grown on coverslips and fixed in 4% paraformaldehyde, permeabilized with 0.1% Triton X-100, and stained with Alexa Fluor 594 Phalloidin (Thermo Fisher Scientific, Rockford, IL) in the dark. Slides were examined and photographed using a Zeiss Axiophot fluorescence microscope (Carl Zeiss Microimaging, Gottingen, Germany). Nuclei were counterstained with 4′,6-diamino-2-phenylindole (DAPI).

Protein lysates were prepared as described previously [34]. A Western blot analysis was performed with indicated primary antibodies and horseradish peroxidase-conjugated secondary antibodies. After washing, blots were incubated with the Western blotting reagent ECL (TOOLS, New Taipei City, Taiwan), and chemiluminescence was detected by the chemiluminescence imaging system, MultiGel-21 (TOP BIO, New Taipei City, Taiwan). Furthermore, the same blots were stripped by stripping buffer (TOOLS, New Taipei City, Taiwan) and reprobed with the β-actin antibody as an internal control.

Short hairpin (sh) RNAs were purchased from the National RNAi Core Facility at Academic Sinica (Taipei, Taiwan). The target sequence of CD26 shRNA was 5’-ACACTCTAACTGATTACTTAA-3. The shRNA lentivirus was produced as previously described [35].

Messenger (m) RNA was isolated and amplified as described previously [35]. Primer sequences of CD26 were F: 5’-GAATGCCAGGAGGAAGGAATCT-3′ and R: 5’-TATTCCACACTTGAACACGCCA-3′.

All animal experiments were performed under a protocol approved by the Institutional Animal Care and Use Committee of Taipei Medical University. For the CD26 overexpression and knockdown experiments in an orthotopic xenograft model, 5-week-old nonobese diabetic (NOD)-SCID male mice were anesthetized with isoflurane and placed in the right lateral decubitus position; then the A549-mock-luciferase, A549-CD26-luciferase, or A549-sh-CD26-luciferase (sh-775 and sh-777) stable cell lines (10⁶ cells) were resuspended in a 1:1 mixture of phosphate-buffered saline (PBS) and GFR-Matrigel and injected into the left lung parenchyma of a NOD-SCID mouse using a 30-gauge needle. After 7 days, the mice were randomized into six groups according to bioluminescence images taken using the Xenogen IVIS-Spectrum system (Caliper; Xenogen, CA), and treatment was initiated according to similar mean tumor sizes in each group. Subsequently, mice were intraperitoneally (IP) administered 3 mg/kg API or the vehicle (10% DMSO in PBS) 5 days/week. The day after API treatment, the mice were injected with D-luciferin and imaged for 1~5 min using this live imaging device to monitor the tumor size and location in real time. After 28 days, A549-injected mice were sacrificed, and luciferase activities in the excised lungs were further determined using the In Vivo Imaging System (IVIS)-Spectrum system. Mouse lungs were also fixed, sectioned, and stained with hematoxylin and eosin (H&E).

Values are presented as the mean ± standard deviation (SD). The statistical analysis was performed using Statistical Package for Social Science software, vers. 16 (SPSS, Chicago, IL). Data were analyzed using Student’s t-test when two groups were compared. A one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test was used to analyze three or more groups. Statistical analyses of the correlation between CD26 and the invasiveness of NSCLC cells were performed using Spearman rank correlations. Differences were considered significant at a 95% confidence interval (p < 0.05).

Recent reports indicated that API can inhibit the in vitro proliferation and motility of human A549 NSCLC cells which harbor the WT EGFR [30]. To further investigate the broad therapeutic potential of API against NSCLC, we examined the effects of API on cell migration and invasion of NSCLC cells harboring different EGFR statuses including the exon 19 deletion (del E746-A750) and L858R/T790 M. First of all, four NSCLC cell lines including A549 and CL1–5 (WT EGFR), HCC827 (del E746-A750), and H1975 (L858R/T790 M) were treated with various concentrations (0~ 80 μM) of API for 24 or 48 h to determine the cytotoxic effects of API. An MTS analysis indicated that only at the highest concentration of 80 μM, API partially suppressed or did not alter the viability of these NSCC cell lines after 24 h of treatment, compared to that of the controls (Fig. 1a). In contrast to short-term treatment (24 h), the colony-formation assay showed that long-term treatment (7 days) with API significantly inhibited the colony-forming abilities of A549, CL1–5, HCC 827, and H1975 cells even at lower concentrations of 10~20 μM (Fig. 1b, Additional file 1: Figure S1). These results indicated that 24-h treatment with API (0~ 40 μM) exhibited very low cytotoxicity in all tested NSCLC cells, but long-term treatment with API efficiently inhibited the in vitro tumorigenicity of these cells. To further determine the effects of short-term API treatment on the cell motility of NSCLC cells, transwell-migration and Matrigel-invasion assays were performed. As shown in Fig. 1c and d, after 24 h of treatment of A549, CL1–5, HCC827, and H1975 cells with API (0~ 40 μM), the migratory and invasive abilities of these cells were all concentration-dependently suppressed, suggesting that even non-toxic concentrations of API can abolish the cell motility of these NSCLC cells.

The EMT and extracellular matrix (ECM) degradation are two key steps in cancer metastasis and were also shown to play critical roles in regulating the invasion and migration of NSCLC [36, 37]. Previous studies indicated that actin remodeling is an important upstream regulator of the EMT in metastatic cancer cells [38].

To determine whether API suppresses cell motility accompanied by changes in the actin cytoskeleton, Alexa Fluor 594 phalloidin and DAPI were used to respectively stain actin filaments and nuclei. As shown in Fig. 2a, A549 and HCC827 control cells displayed well-organized F-actin-containing microfilament bundles within the cytoplasm, whereas API-treated cells contained few microfilament bundles, indicating that F-actin-containing microfilament bundle rearrangements may be involved in API-modulated cell motility. Furthermore, we determined the effects of API on the EMT progression of NSCLC cells. In A549 and HCC827 cells, expressions of mesenchymal marker, N-cadherin, and the EMT-promoting transcription factors, Snail and Slug, were all concentration-dependently downregulated after treating cells with API (0~ 40 μM) for 24 h (Fig. 2b). These data suggested that downregulating Snail family members (Snail and Slug) might be critical for the API-mediated suppression of the EMT in EGFR WT and EGFR mutant NSCLC cells. Next, to determine whether Snail family members play critical roles in API-modulated cell motility, we overexpressed Snail or Slug in A549 cells to reverse the inhibitory effect of API on the Snail family (Fig. 2c). Meanwhile, the migratory ability was promoted in Snail- or Slug-overexpressing A549 cells compared to control A549 cells (A549/Neo), and the API-mediated suppression of the migratory ability in A549 cells was respectively reversed when Snail or Slug was overexpressed in cells (Fig. 2d). To elucidate the clinical relevance of Snail family members, a cohort of 499 lung cancer cases from The Cancer Genome Atlas (TCGA) was analyzed. Results from a Kaplan-Meier plot showed that patients with tumors exhibiting high expression levels of both Snail and Slug had significantly shorter survival times compared to patients with tumors exhibiting low expression levels of Snail and Slug (Fig. 2e). These data suggest that downregulating Snail and Slug is crucial for the API-mediated inhibition of EMT progression and cell invasion, and those high levels of Snail and Slug predicted a poor prognosis in patients with NSCLC.

Previous studies indicated that the Akt signaling pathway plays a crucial role in most API-mediated anticancer processes and increases the stability and transcription of Snail family members during the EMT of cancer cells [15‐17, 29]. Herein, we found that API inhibited activation of Akt at indicated time points (2, 4, and 24 h) in A549, CL1–5, and H1975 cells (Fig. 3a). Moreover, we further overexpressed the constitutively activated Akt, myr-Akt, in A549 and H1975 cells and found that overexpressing myr-Akt significantly increased Snail and Slug expression levels in both cells (Fig. 3b). Functionally, we observed that overexpressing activated Akt significantly reversed API-induced inhibition of invasion in A549 cells (Fig. 3c). These data suggest that API suppression of Snail family-mediated cell motility is attributable to its capacity to inhibit Akt activation. In contrast to Akt-mediated Snail and Slug expressions, Snail was also reported to induce the EMT through activating the Akt pathway in NSCLC [39]. Our results showed that overexpression of Snail and Slug could reverse the API-mediated suppression of Akt activation (Fig. 3d). Moreover, overexpression of Snail and Slug also inhibited expression of the PI3K/Akt pathway suppressor, PTEN (Fig. 3e), suggesting that API may inhibit Snail family members to maintain PTEN expression and further suppress the PI3K/Akt pathway. Taken together, these data suggest that a regulatory network between Akt activation and Snail family expressions exists in API-mediated inhibition of cell motility.

Recent findings indicated that several proteases, including MMP-2, MMP-3, MMP-9, presenilin 1, and CD26, are promoters and mediators of EMT processes in different cancer types [24, 40, 41]. We next screened the effect of API on these proteases and found that MMP-3 and CD26 were considerably downregulated in CL1–5 cells after API treatment (Fig. 4a). In a further check of the inhibitory effect of API on proteases in other NSCLC cell lines (A549 and H1975), Western blot data showed that only CD26 was consistently suppressed by API in all tested NSCLC cells (Fig. 4b, Additional file 1: Figure S2). Moreover, results showed that API concentration-dependently suppressed CD26 expression at concentrations of 10~ 80 μM in CL1–5 cells (Fig. 4b). In addition to the protein level, CD26 mRNA expression was also inhibited by API in A549, CL1–5, and H1975 cells (Fig. 4c). These data suggest that inhibition of CD26 by API may be a general phenomenon in NSCLC cells. To further determine the functional role of CD26 in NSCLC, we investigated the correlation between the CD26 expression and the cell invasive ability. We first detected expression levels of CD26 in a set of NSCLC cell lines (A549, CL1–0, CL1–3, CL1–5, HCC827, PC9, and H1975) using Western blotting and an RT-PCR (Fig. 4d). Next, the invasive abilities of these cell lines were further evaluated, and we found a variety of invasive abilities, with A549 and H1975 cells exhibiting relatively high invasive abilities (Fig. 4e). Moreover, we observed that H1975 and A549 cells also expressed relatively high CD 26 protein and mRNA levels among these cell lines (Fig. 4d). We further combined results from Fig. 4d and e and found a significant correlation (correlation coefficient r = 0.773, p = 0.042) between CD26 protein levels and the invasive abilities of these cell lines (Fig. 4f). Next, to determine whether CD26 modulates the cell invasive ability, we knocked-down CD26 in A549, H1975, and CL1–5 cells. The knockdown efficiency of CD26-specific siRNA was detected by a Western blot analysis (Fig. 4g, left panel). After knocking-down CD26 in these cells, we found their invasive abilities were all significantly attenuated compared to control cells (Fig. 4g, right panel). In comparison, overexpression of CD26 significantly increased the cell-invasive abilities of poorly invasive CL1–0 cells and also reversed the API-mediated suppression of the invasive ability of CL1–0 cells (Fig. 4h). In the clinic, we analyzed CD26 gene expression data obtained from TCGA and found that CD26 expression was inversely correlated with recurrence-free survival of lung cancer patients (Fig. 4i). These results suggested that CD26 may play a critical role in the API-regulated invasive ability of NSCLC cells and may be associated with the poor prognosis of patients with lung cancer.

Next, to dissect the relationship between CD26 and the Akt-Snail/Slug pathway modulated by API, we first overexpressed Snail or Slug in A549 cells. Figure 5a shows that overexpression of Snail or Slug increased Akt activation, but had no significant effect on CD26 expression. Figure 5b further shows that the ectopic expression of myr-Akt significantly reversed the API-mediated inhibition of Akt activation, but not inhibition of CD26 expression. These results suggest that the Akt-Snail/Slug pathway does not modulate CD26 expression in API-treated NSCLC cells. In contrast, to determine whether CD26 modulates Akt activity and the Snail/Slug-mediated EMT, we knocked-down CD26 with CD26-specific shRNA in H1975 and A549 cells. Figure 5c shows that CD26 knockdown significantly decreased Akt activation and Snail, Slug, and fibronectin expressions, and the knockdown efficiency of CD26-specific shRNA was also shown here. Similar inhibitory effects of CD26 shRNA on Akt activation and Snail/Slug expression were also observed in CL1–5 cells (Additional file 1: Figure S3). From TCGA database, we observed that patients with CD26^high/Akt^high lung tumors had the shortest recurrence-free survival times compared to the CD26^high/Akt^low, CD26^low/Akt^high, and CD26^low/Akt^low groups (Fig. 5d). Taken together, these results revealed that CD26 may be vital in mediating the Akt-Snail/Slug-induced EMT in NSCLC cells, and API can target this pathway. Clinical data indicated that upregulation of CD26 and Akt may be critical events in promoting lung cancer progression.

To investigate the role of CD26 in tumor progression and in API-mediated anticancer activity in vivo, we established an orthotopic lung tumor-bearing model by transplanting luciferase-tagged cells, A549-mock-luciferase, A549-CD26-luciferase, or A549-shCD26-luciferase (sh-775 and sh-777), into NOD-SCID mice and allowed them to become established for 7 days before initiating treatment. A549–1-Luc orthotopic graft mice were IP-administered API or the vehicle 5 days/week. The effects of API administration or CD26 knockdown and overexpression on tumor growth and metastasis were monitored by bioluminescence imaging. A schematic timeline of this experimental design and setup is shown in upper panel of Fig. 6a. In vivo photon emission detection revealed that both API treatment and CD26 knockdown in A549 cells attenuated tumor growth compared to the control group (Fig. 6a, b). The tumorigenic ability was promoted in CD26-overexpressing A549 cells compared to control A549 cells and significantly reversed the antitumor growth effect of API (Fig. 6a, b). Similar to the in vivo data, after the mice were sacrificed, ex vivo imaging of their lungs respectively showed lower and higher photon intensities in A549/shCD26- and A549/CD26-injected mice compared to control mice, in both orthotopic tumors (left lung) and metastatic tumors (right lung) (Fig. 6c). The growth- and metastasis-inhibitory effects of API were also reversed in cancer cells overexpressing CD26 (Fig. 6c). Moreover, histological data also confirmed that mice bearing A549/shCD26 tumors or receiving API treatment all developed fewer metastatic nodules (as indicated by a black arrow) in the right lung compared to control mice (Fig. 6d). Taken together, these results revealed that API treatment suppressed the in vivo metastasis of NSCLC through targeting CD26.