Thalidomide was kindly provided by Changzhou Corp. It was dissolved in 0.5% sodium carboxylmethyl cellulose (CMC) at a final concentration of 20 mg/ml. Animals Inbred, clean, black, 6–8 week old C57BL/6 mice were obtained from Institute of Hematology and Blood Disease Hospital affiliated Chinese Academy of Medical Science(License number: SCXK(Jin)2004-0001). They arrived in Animal Centre of Tianjin Cancer Hospital one week before the experiment and were bred under SPF. This study was approved by the Medical University of Tianjin's Animal Welfare Committee.

C57BL/6 mice (n = 30) were injected subcutaneously with 2 × 10⁶ murine melanoma B16F10 cells into the lower left groin. Thirty mice were randomly divided into treatment group and the control group, with 15 mice in the treatment group and 15 mice in the control group. Thalidomide was dissolved in 0.5% sodium carboxyl methylcellulose. Five days after B16F10 melanoma inoculation, 0.2 ml thalidomide (200 mg/kg/d) was administered via interperitoneal injection once a day in the treatment group and an equivalent volume of 0.5% CMC alone was administered in the control group. After administration, there was not obviously symptom of disorder of digestive tract in the mice of the treatment group. The length and width of the subcutaneous tumors were measured using a caliper every day. The tumor size was calculated according to the following formula: Tumor volume (cm³) = (length × width²)/2. The tumor growth curve was drown based on tumor size. On the 19^th day after inoculation, all mice were sacrificed, and the tumor tissues were harvested. Part of the tumor without necrosis were collected and stored at -80°C and the remainder of the tumors were fixed with formalin and embedded in paraffin.

Total RNA was extracted with Trizol reagent according to the manufacturer's instructions. 1% agarose electrophoresis and detection of OD260/OD280 ratio were performed to identify the integration and purity of isolated RNA. Complementary DNA (cDNA) was synthesized and amplified from total RNA using the Access real time PCR system (TaKaRa). The primer sequences used for matrix metalloproteinase (MMP)-2(GeneID: 17390) detection were 5'- GATAACCTGGATGCCGTCGTG -3' (sense) and 5'- CTTCACGCTCTTGAGACTTTGGTTC -3' (anti-sense). The primer sequences used for MMP-9 (GeneID: 17395) detection were 5'-GCCCTGGAACTCACACGACA-3' (sense) and 5'-TTGGAAACTCACACGCCAGAAG-3' (anti-sense). The primers used to amplify β-actin were 5'-CATCCGTAAAGACCTCTATGCCAAC-3' (sense) and 5'-ATGGAGCCACCGATCCACA-3' (antisense). The resultant cDNA products of MMP-2, MMP-9 and β-actin were 109, 86 and 174 base pairs, respectively. Real time PCR products were analyzed with the Gene AMP PCR System 5700 Sequence Detector and the C_T values were evaluated. The CT value (the cycle number at which the fluorescence crosses the threshold) was determined and 2^ΔCT where ΔCT = ΔCT_MMPs-ΔCT_β-actin was defined as the relative quantity of the amplified fragment. Every sample was tested in triplicate and the mean value was used. The products of real-time PCR were validated with 1.5% agarose electrophoresis.

For histopathology studies, tumor tissues were cut in the center to obtain the largest section and indicate the information of the whole tumor. They were then fixed in 10% buffered formalin, dehydrated, and embedded in paraffin using routine methods. For immunohistochemical staining, formalin-fixed, paraffin-embedded tissue was sectioned and dried overnight at 65°C and deparaffinized in xylene. The sections were rehydrated through graded alcohols into water. Endogenous peroxidase was blocked with 3% hydrogen peroxide in 50% methanol for 10 min at room temperature. After rehydrating, the sections were washed with PBS and then pretreated with citrate buffer (0.01 M citric acid, pH 6.0) for 20 min at 100°C in a microwave oven. After rinsing with PBS, slides were incubated overnight at 4°C with primary polyclonal antibodies (rabbit anti-human, mouse and rat), including the antibodies against vascular endothelial growth factor(VEGF), proliferating cell nuclear antigen(PCNA) (Boster Biological technology Ltd, Wuhan, China, dilution 1:150), nuclear factor-κB (NF-κB, Upstate, New York, USA, dilution 1:100), MMP-2 (BA0596, Boster Biological technology Ltd, Wuhan, China, dilution 1:100), and MMP-9 (BAO573, Boster Biological technology Ltd, Wuhan, China, dilution 1:100). The sections were then washed with PBS and incubated with the second antibody for 30 min at 37°C. The sections were incubated with secondary antibody (Non-Biotin HRP detection system, Zhongshan goden bridge biotechnology CO., Ltd, Beijing, China) for 30 min at 37°C after the PBS washes. Visualization was performed using a DAB Kit(DC 10, Boster Biological technology Ltd, Wuhan, China) under microscope. The nuclei were counterstained with hematoxylin, followed by dehydration and coverslip mounting. Appropriate positive and negative controls without primary antibody were included.

Antibodies used in this study were mouse monoclonal anti-CD34 (Sigma Chemical Co., St. Louis, Mo, USA) and mouse monoclonal anti-HMB45(BM0093, Boster Biological technology Ltd, Wuhan, China, dilution), which were used at dilutions: 1:400. After immunohistochemical staining, Sections were exposure to 1% sodium periodate for 10 min. The sections were then rinsed with distilled water for 5 min and incubated with periodic acid-Schiff (PAS) for 15 min. Finally, all of the sections were counterstained with hematoxylin, dehydrated and mounted. Normal human stomach mucous membrane was the positive control.

When stained for VEGF, NF-κB, MMP-2 and MMP-9, tumor cells with brown cytoplasm were considered positive, and when stained for PCNA, tumor cells with brown nuclei were considered positive. We observed 10 fields per section at 400× magnification, and positive cell numbers were counted in 100 random melanoma cells in every field. The mean percentage of positive cells was used to determine the expression of the proteins in a section. EVs, MVs and VM channels were also counted. All these counts were blindly performed in at least 3 randomly chosen sections from each mouse. The mean value of each type of microvessel in 10 fields was the final outcome.

Statistical software SPSS 10.0 (Chicago, Illinois) was used in the analysis. A P value less than 0.05 was considered statistically significant. Two-way ANOVA was performed to evaluate melanoma growth of two groups. Counts of distinguished blood vessels in the treatment group and the control group were evaluated with one-way ANOVA. Differences of protein and mRNA expressions between two groups were compared using an unpaired t test.

Ninth days after melanoma cell injection, engrafted tumors were palpated on the mouse groin areas and detected tumors were removed. Soft globular tumors were observed and there were numerous network vessels on the tumor surfaces. Some melanoma cells invaded into the skeletal muscle and showed a spindle configuration. The engrafted melanomas in the thalidomide treatment group grew slower than those in the control group (Figure 1). Fourteen days after the thalidomide administration, the average size of the tumor in the treatment group was 6.912 cm³, while in the control group it was 4.164 cm³. The tumor volume in the treatment group was decreased compared with the control group and there was statistical significance between these two groups (P = 0.019) after the mice were sacrificed. In the early stages of the experiment, there were no significant differences in tumor volume, but from day 12 to day 19 the tumor volume in the treatment group showed a obvious reduction compared with the control group (Figure 1 and Figure 2A). There were large areas of necrosis in tumor tissues of the treatment group, while necrosis was not obvious in the control group. Numerous melanoma cells of the treatment group were characterized as cell degeneration including vacuoles within the cytoplasm and nuclei. However, in the control group, there are only some tumor cell degeneration in the centre of tumor mass.

We performed PAS staining and staining for the endothelial cell markers CD34. CD34 is a marker of endothelial cells and the base membrane is positive for PAS. CD34 immunohistochemical and PAS histochemical double-staining was used to distinguish VM, MVs and EVs. HMB45 immunohistochemical and PAS histochemical double-Staining was used to identify the origin of tumor cells lining VM channels and the structure of VM as well. Cells lining VM channels were negative for CD34 and positive for HMB45 confirmed that cells around the channels were not composed of endothelium but melnaoma cells (Figure 2B and 2C). Some channels lined with both CD34-positive/PAS positive and HMB45-positive/PAS positive cells were MVs. Red cells in the centre of VM channels indicated that they may be connected with EVs. Under microscope with 400× magnification, EVs, MVs and VM channels were counted and the three microcirculation patterns could co-exist in melanoma(Figure 2D). The number of VM(P = 0.03) and MVs(P = 0.004) in the treatment group were decreased compared with the control group. There were fewer EVs in the treatment group compared with the control group, but the difference was not statistically significant (P = 0.068) (Table 1).

Both the treatment group and the control group contained positive tumor cells. The positive positions of tumor cells for VEGF, NF-κB, MMP-2 and MMP-9 staining were located in the cytoplasm and for PCNA in the nucleus. VEGF(P = 0.004), NF-κB(P = 0.009), PCNA(P = 0.002), MMP-2(P = 0.000), and MMP-9(P = 0.002) expression in the treatment group was significantly decreased compared with the control group, including overall number and staining intensity of the positive cells (Figure 3 and Table 2). The number of tumor cells positive for PCNA staining in the treatment group was less than observed in the control group. Results of PCNA immunohistochemical staining indicated that the tumor cells in the treatment group had a lower proliferative ability than that in the control group.

Real time PCR demonstrated that the expression of MMP-2 mRNA in the treatment group was decreased compared with the control group, which was similar to MMP-2, MMP-9 protein expression. The C_T value of MMP-2 and MMP-9 in the control group was lower than that in the treatment group. The results of real time PCR show that thalidomide could down-regulate MMP-2 and MMP-9 mRNA expression, suggesting that thalidomide can inhibit tumor growth via down-regulation of the MMP-2 and MMP-9 mRNA level(Table 3). There were statistical significance for MMP-2(P = 0.000) and MMP-9(P = 0.000) mRNA levels between the treatment group and the control group.