**2. TFPI-2 expression in normal hepatic and hepatocarcinoma tissues**

#### **2.1 Tissue specimens**

Human hepatocarcinoma tissues and tumor-adjacent normal hepatic tissues were obtained from HCC patients admitted to Shenzhen People's Hospital. They were stored frozen at - 75C until use.

#### **2.2 In situ hybridization**

Tumor specimens were fixed in formalin overnight and embedded in paraffin using standard procedures. Series sections (4um) were deparaffinized with xylene, rehydrated in a graded series of ethanol, and washed in PBS. Human TFPI-2 mRNA was detected using the

 \* Xiaolin Qin2, Jinjing Zhou2, Zhiguang Tu2, Xiao Bi3, Wuxian Li2, Xiaoqing Fan2 and Yi Zhang3

*<sup>1</sup>Pingshan People's Hospital, Shenzhen, Guangdong, P.R. China 2Key Laboratory of Laboratory Medical Diagnostics, Ministry of Education, Chongqing Medical University, Chongqing, P.R. China* 

*<sup>3</sup>The Second Clinical Medical College, Jinan University, Shenzhen People's Hospital, Guangdong, P.R. China*

Tissue Factor Pathway Inhibitor-2 Inhibits the Growth and Invasion of

**3. TFPI-2 plasmid expression vector construct** 

used as a positive control.

**3.1 TFPI-2 gene sequencing** 

Hepatocellular Carcinoma Cells and is Inactivated in Human Hepatocellular Carcinoma 163

Detection Kit (Zhongshan Goldbridge, Beijing, China) according to the manufacturer's instructions. Staining was developed with DAB, slides were counterstained with hematoxylin, dehydrated, and mounted. The primary antibody replaced with PBS was used as a negative control.Mature placenta tissue, known to express large amounts of TFPI-2,was

*In situ* hybridization with TFPI-2 probe demonstrated that little or no TFPI-2 mRNA was detected in hepatocarcinoma tissue sections, while a high level of TFPI-2 mRNA was detected in tumor-adjacent normal hepatic tissue sections (Fig. 1A, B). The positive and negative controls confirmed the specifility of hybridization liquid replaced with PBS used as a negative control confirmed the absence of a specific hybridization signal (data not shown). Further immunohistochemical analysis confirmed that TFPI-2 protein was stained strongly positive in normal hepatic tissues but was weakly stained in hepatocarcinoma tissues (Fig. 1C, D). The TFPI-2 immunostaining scores for normal hepatic tissues and hepatocarcinoma tissues were 46.60±1.80 and 22.54±1.22, respectively (P<0.05). Taken together, these data indicate that the expression of TFPI-2 was markedly reduced in hepatocarcinoma tissues.

The RNA from hepatic tissue of human fetor (Shenzhen People's Hospital) was isolated and full length TFPI-2 cDNA was amplified with RT-PCR kit (TaKaRa). The cloned gene was inserted into plasmid pcDNA2.1 (Chinagen, Shenzhen, China), and sequenced from forward and reverse direction at Shanghai Biotechnology (China), then it was inserted into eukaryotic expression vetor pcDNA3.1, a gift from Dr Tiyuan Li (Central Laboratory,

Shenzhen People's Hospital) verified by enzyme digestion and sequencing.

In Situ Hybridization Detection Kit (Boster, Wuhan, China) according to the manufacturer's instructions. Briefly, the sections were hybridized in prehybridization buffer supplemented with 0.1 ug/ml digoxigenin-labeled, 1.2-kb antisenese TFPI-2 probe overnight at 37C and incubated with biotinylated mouse antidigoxigenin antibody (1:1000 dilution), then incubated with biotinylated peroxidase. Staining was developed with DAB. Slides were counterstained with hematoxylin, dehydrated, and mounted. The number of cells stained brown (indicating the presense of TFPI-2 mRNA) were assessed by light microscopy. The hybridization probe replaced with PBS was used as a negative control. Mature placenta tissue, known to express large amounts of TFPI-2, was used as a positive control.

### **2.3 Immunohistochemistry**

Tissue sections were prepared in the same manner as above. Then the expression of TFPI-2 was determined by incubation with a mouse polyclonal antibody against human TFPI-2 (Santa Cruz, CA), horseradish peroxidase (HRP)-conjugated sheep anti-mouse lgG secondary antibodies (Chinagen, Shenzhen, China), and final detection using the non Biotin-labeled

Fig. 1. TFPI-2 expression in normal hepatic and hepatocarcinoma tissues. Expression of (A) TFPI-2 mRNA in hepatocarcinoma tissue and (B) tumor-adjacent normal hepatic tissue was examined by *in situ* hybridization with a digoxigenin-labeled TFPI-2 probe. Expression of (C) TFPI-2 protein in hepatocarcinoma tissue and (D) tumor-adjacent normal hepatic tissue was examined by immunohistochemical analyses with TFPI-2 antibody. Magnification, x400.

Detection Kit (Zhongshan Goldbridge, Beijing, China) according to the manufacturer's instructions. Staining was developed with DAB, slides were counterstained with hematoxylin, dehydrated, and mounted. The primary antibody replaced with PBS was used as a negative control.Mature placenta tissue, known to express large amounts of TFPI-2,was used as a positive control.

*In situ* hybridization with TFPI-2 probe demonstrated that little or no TFPI-2 mRNA was detected in hepatocarcinoma tissue sections, while a high level of TFPI-2 mRNA was detected in tumor-adjacent normal hepatic tissue sections (Fig. 1A, B). The positive and negative controls confirmed the specifility of hybridization liquid replaced with PBS used as a negative control confirmed the absence of a specific hybridization signal (data not shown).

Further immunohistochemical analysis confirmed that TFPI-2 protein was stained strongly positive in normal hepatic tissues but was weakly stained in hepatocarcinoma tissues (Fig. 1C, D). The TFPI-2 immunostaining scores for normal hepatic tissues and hepatocarcinoma tissues were 46.60±1.80 and 22.54±1.22, respectively (P<0.05). Taken together, these data indicate that the expression of TFPI-2 was markedly reduced in hepatocarcinoma tissues.
