Animal experiments were performed in compliance with the rules of the Animal Use Committee of Fujian Medical University. The present study was approved by the Institutional Ethics Board of Fujian Medical University Union Hospital.

Northern blotting was performed to detect miRNAs as previously described (16). Briefly, we used RNAiso Plus to extract 10–20 µg total RNA from cells. The RNA was then separated on 10% denaturing polyacrylamide gels and electrotransferred to nylon membranes. DNA oligonucleotides that were antisense to mature miRNAs and end-labeled with ³²P were used as probes. U6 snRNA was used as an internal control. A miR-19a probe (Pr009177336, 5′-UCAGUUUUGCAUAGAUUUGCACA-3′) was used to detect miR-19a.

The human CRC cell line HCT116 (a gift from Professor Pan Jinshui) was cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS; Gibco Life Technologies, Gaithersburg, MD, USA) and was maintained at 37°C in a humidified 5% CO₂ atmosphere. The miR-19a mimic (5′-UGUGCAAAUCUAUGCAAAACUGA-3′), inhibitor and negative control were purchased from RiBoBio (Guangzhou, China). For transfection, cells (3×10⁵/well) were cultured in a six-well plate until they reached 40–50% confluency, and the mimics were transfected using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions.

Hematoxylin and eosin (H&E) staining was performed using standard protocols. Briefly, deparaffinized rehydrated sections were stained with hematoxylin for 5 min, incubated in 0.1% acid alcohol, and then counterstained in 0.5% eosin Y solution. Sections were finally dehydrated and examined by a skilled pathologist. For immunohistochemistry, sections were deparaffinized and quenched with 3% H₂O₂. After the antigen was retrieved in 10 mM sodium citrate buffer and blocked with goat serum (ZLI-9022; ZSGB-BIO, Xicheng, Beijing, China), the sections were incubated with antibodies recognizing CD31 (ab28364; Abcam, Cambridge, UK) or KRAS (12063–1-AP; Proteintech, Chicago, IL, USA) and then with secondary antibodies conjugated with horseradish peroxidase (PV-6101; ZSGB-BIO). Sections were processed according to the manufacturer's instructions and counterstained with hematoxylin.

We performed sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) to separate total cell lysates, and then transferred the proteins to polyvinylidene difluoride membranes (EMD Millipore, Billerica, MA, USA). Western blotting was performed with the appropriate antibodies, which were visualized by enhanced chemiluminescence in accordance with the manufacturer's instructions (ECL; Millipore).

Reverse transcription of total RNA was performed using 4 µg total RNA from HCT116 cells. M-MLV reverse transcriptase (BGI, Shenzhen, China) was used to generate cDNAs. The mRNA level of KRAS was assessed by qRT-PCR using SYBR-Green I on a CFX96 real-time RT-PCR detection system (Bio-Rad, Hercules, CA, USA). The sequences of the primers used in the present study are listed in Table I. PCR was performed for 45 cycles using the following conditions: denaturation at 95°C for 20 sec, annealing at 58°C for 20 sec, and elongation at 72°C for 20 sec. All values were normalized to the expression of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA, and the relative expression was calculated according to the ΔΔCT method (17), where ΔΔCT = ΔCT_sample - ΔCT_GAPDH.

The tube formation efficiency of human umbilical vein endothelial cells (HUVECs) was assessed by an angiogenesis assay on Matrigel (BD Biosciences, Franklin Lakes, NJ, USA). First, we prepared the GFP-labeled HUVECs, as described by Guo et al (18), and then 36 h after transfection with miR-19a, KRAS/sh-KRAS-2# or VEGFA expression constructs, the HUVECs (1×10⁴ cells) were added to a 24-well plate coated with 100 µl Matrigel basement membrane matrix (BD Biosciences), which was derived from an Engelbreth-Holm-Swarm tumor. After culturing for 12 h, we recorded tube formation with an IX71 inverted fluorescence microscope (Olympus, Tokyo, Japan) at a low magnification of (×10). Four wells for each treatment were photographed. The images were saved as JPEG files and analyzed using TCS Cellworks AngioSys 1.0 software (Botolph Claydon, Buckingham, UK) to quantify the effects on angiogenesis. Data were analyzed using unpaired Student's t-test of GraphPad Prism 5.0 (GraphPad Software, Inc., La Jolla, CA, USA).

For lentivirus packaging, 1.5 µg lentivirus vector was mixed with 1.5 µg packaging plasmids (0.75 µg pMDL, 0.45 µg VSV-G and 0.3 µg REV) in six-well plates. The mixed plasmids were transfected into 293T cells at a density of ~80% confluence using calcium phosphate precipitation. Lentivirus particles were harvested 48 h after transfection. The plasmid pBOB-EGFP was used as a positive control.

HCT116 cells were infected using the appropriate amount of lentivirus, and then subjected to amplification. Infected cells were trypsinized and suspended in FBS-free DMEM-F12 medium at a density of 6×10⁶ cells/0.1 ml. The suspended cells were then implanted into the backs of 27 five-week-old BALB/c female nude mice divided into 3 groups, with each group containing nine mice for initiating tumor xenografts. Tumor diameters were measured every 3 days.

The plv vector carrying a copy of miR-19a or VEGFA was a kind gift from Dr Lixin Hong (School of Life Science, Xiamen University, Fujian, China). The pGL3 control vector carrying a 1,923-bp fragment of the 3′UTR of human VEGFA mRNA was directly purchased from RiBoBio. The KRAS 3′UTR harboring predicted target sites for miR-19a (Fig. 1A) was PCR-amplified using primers containing XbaI sites and cloned into the XbaI-digested pGL3 control vector. The recombinant plasmids were constructed using ligation-independent combination (LIC). F1 and R1 primers were used to generate the KRAS wild-type 3′UTR. To construct 3′UTRs with deletion mutations, two sets of PCR were performed using the primer pairs F1/R2 and F2/R1 to generate two fragments. Then, the two fragments were cloned into the XbaI-digested pGL3 control vector using LIC.

The DNA oligos encoding shRNA sequences were designed and cloned into the expression vector pLV-H1-EF1α-puro using single oligonucleotide RNAi technology developed by Biosettia (San Diego, CA, USA). All lentiviral-shRNA vectors were constructed following the manufacturer's protocol. The primers used in the present study are listed in Table I.

HCT116 cells were plated in six-well culture plates 24 h before transfection. The cells were then transfected with the recombinant pGL3 vector (1.6 µg), plv-miR-19a (0.2 µg) and the pRLTK vector (0.2 µg) for normalization of transfection efficiency. Cell lysates were collected and assayed 48 h after transfection. Firefly and Renilla luciferase activities were determined using a Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA). Firefly luciferase activities were calculated as the mean ± standard deviation (SD) after normalization to Renilla luciferase activity. Three independent experiments were performed.

Mann-Whitney tests were performed to compare the differences between two groups when variances were unequal. Analysis of variance with post hoc tests were performed when comparing >2 groups. For qualitative data, non-parametric tests such as Kruskal-Wallis tests were performed.

We identified the conserved miR-19a-binding site in the 3′UTR region of human KRAS using the microRNA.org website (Fig. 1A). The KRAS 3′UTR-Mut was generated by deleting the predicted binding site. We observed that a reporter gene containing the KRAS 3′UTR was suppressed by miR-19a, while removal of the predicted miR-19a-binding site resulted in insensitivity of the reporter gene to miR-19a in HCT116 cells (Fig. 1B).

To ascertain the effect of miR-19a on KRAS expression in HCT116 cells, the mature miR-19a mimic/inhibitor was transfected in HCT116 cells. The results revealed that the endogenous KRAS mRNA and protein levels were significantly downregulated in the miR-19a mimic cells and upregulated in the miR-19a inhibitor cells compared with those in the untransfected cells (Fig. 2A and B). Accordingly, the VEGFA protein level was also downregulated in the miR-19a mimic cells and upregulated in the miR-19a inhibitor cells (Fig. 2B).

Next, in order to construct stable miR-19a-overexpressing HCT116 cell lines for subsequent studies, we transfected HCT116 cells with a lentivirus expressing miR-19a. Then, via northern blotting, we confirmed miR-19a overexpression in HCT116 cells (Fig. 2C and D). Next, we compared the levels of KRAS expressed in HCT116 cells with and without lentiviral transfection for overexpression of miR-19a. The resulting in vitro data revealed that overexpression of miR-19a decreased KRAS mRNA and protein levels in HCT116 cells compared with that in the control (miR-Con; Fig. 2E and F).

TargetScan analysis revealed that VEGFA was not a miR-19a target. Since the sites that match the miRNA seed sequence are located in 3′UTRs in most mammalian mRNAs, we performed the dual-luciferase reporter assay to detect whether miR-19a bound to the 3′UTR of VEGFA. The results revealed that the reporter gene containing the VEGFA 3′UTR was not suppressed by miR-19a, which confirmed that VEGFA was not directly regulated by miR-19a in HCT116 cells (Fig. 2G).

Since high levels of miR-19a downregulates KRAS, we hypothesized that miR-19a may affect angiogenesis. Hence, we performed tube formation assays using HUVECs, which differentiate and form capillary-like structures on Matrigel, to mimic the process via which endothelial cells form capillaries in vivo (19). The results revealed that overexpression of miR-19a significantly (P<0.01) inhibited angiogenesis in the Matrigel assay, and that rescuing KRAS expression could restore angiogenesis inhibited by miR-19a (Fig. 3A and B). This confirmed our hypothesis that miR-19a suppressed angiogenesis by targeting KRAS.

A previous study revealed that the vascular endothelial growth factor receptor (VEGFR) signaling pathway plays a major role in angiogenesis (20) and that KRAS plays a key role in VEGF signaling (21). To ascertain these results, we knocked down KRAS using shRNA and determined the expression level of VEGFA in HCT116 cells. The results revealed that two KRAS shRNAs reduced KRAS mRNA and protein levels significantly, and consequently, VEGFA was also downregulated (Fig. 3C and D). Furthermore, we knocked down KRAS using shRNA in HUVECs and performed the tube formation assay. As shown in Fig. 3E, KRAS knockdown disrupted tube formation, whereas re-expression of VEGFA rescued the defect. Thus, the results demonstrated that KRAS was a key regulator of VEGFA expression and angiogenesis.

The above data indicated that miR-19a overexpression could target KRAS in HCT116 cells and inhibit tube formation. Therefore, we next examined whether overexpression of miR-19a could suppress CRC progression by blocking angiogenesis via targeting of KRAS using a xenograft model.

The results revealed that the sizes of xenograft tumors that developed from HCT116 cells expressing miR-19a were much smaller than those from HCT116 cells transfected with the control lentivirus (Fig. 4A and B). In addition, tumors that developed from miR-19a-expressing lentivirus-transfected cells tended to be paler in color than xenograft tumors developed from control lentivirus-transfected cells, indicating decreased blood flow in miR-19a-expressing tumors (Fig. 4A).

We observed that tumors that developed from miR-19a-expressing cells had fewer blood vessels than the control tumors. Immunohistochemical staining for CD31, a marker of vascular endotheliocytes, also revealed that the blood vessels in tumors that developed from miR-19a-expressing cells were smaller than those in the control group (Fig. 4D). Moreover, immunohistochemical staining for KRAS and western blotting analysis confirmed that the tumors that developed from miR-19a-expressing cells exhibited reduced KRAS expression compared with those in the control group (Fig. 4C and D). Moreover, western blotting analysis revealed that the expression level of VEGFA in tumors that developed from miR-19a-expressing cells was also reduced (Fig. 4C). These results indicated that overexpression of miR-19a inhibited KRAS expression in CRC, which may reduce the production of VEGFA and suppress tumor angiogenesis, thereby delaying tumor growth.