As described in our previous study (13), shRNA primers targeting CLDN3 (sense, 5′-TCCCGCAACATCATCACGTCGCATTCAAGACGTGC GACGTGATGTGTTGCTTTTTTG-3′ and antisense, 5′-AGC TCAAAAAAGCAACATCATCACGTCGCACGTCTTGAA TGCGACGTGATGATGTTGC-3′) were designed and synthesized. We used the HK sequence which has no homology with any of the known mammalian gene sequences as the negative control. The above sequences were then transferred into the pGenesil-2.1 vector (Genesil, Wuhan, China), which contains a kanamycin-resistance gene. The resulting recombinant plasmids were named pshCLDN3 or pshHK, respectively. Both of the two constructs were validated by DNA sequencing. The plasmids were purified using the EndoFree Plasmid Giga kit (Qiagen, Chatsworth, CA, USA) from DH5α Escherichia coli transformants.

The human ovarian serous cystadenocarcinoma cell line SKOV3 (ATCC, Manassas, VA, USA) was cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and antibiotics (100 units/ml penicillin and 100 μg/ml streptomycin) in a humidified atmosphere containing 5% CO₂ at 37°C.

Plasmid transfection was carried out using HPEI nanogels synthesized at the State Key Laboratory of Biotherapy and Cancer Center as previously described (21). Briefly, SKOV3 cells (2×10⁵/well) were seeded in 6-well plates one day prior to transfection to achieve 80% confluence at the time of transfection. Plasmid (pshCLDN3 or pshHK)/HPEI complexes (1 μg plasmid/10 μg HPEI) were prepared in 1 ml DMEM without serum and antibiotics. An equal volume of normal saline (NS) prepared in 1 ml DMEM was used as a control agent. Cells were incubated with NS, HPEI nanogels, pshHK/HPEI complexes or pshCLDN3/HPEI complexes for 6 h, and then the medium was replaced with 2 ml complete medium (DMEM containing 10% FBS) and further incubated for 72 h.

Cells or tumor tissue samples were lysed in RIPA lysis buffer containing proteinase inhibitor (1 mM cocktail plus 1 mM PMSF). The supernatant was collected, and the protein concentration was quantified using Pierce BCA protein assay kit (Thermo Scientific, Rockford, IL, USA). Equal amounts of protein (20 μg) were loaded onto 12% SDS-PAGE gel for electrophoresis and blotted onto a PVDF membrane (Millipore, Bedford, MA, USA). Subsequently, the membrane was blocked in 5% skimmed milk for 2 h, and then incubated with rabbit anti-human polyclonal antibody against CLDN3 (1:200; Invitrogen) at 4°C overnight, followed by horseradish peroxidase-conjugated secondary antibody. The immunoreactive bands were visualized by chemiluminescence detection. GAPDH served as the protein loading control.

The following procedures for the animal experiments were approved by the Institutional Animal Care and Use Committee of Sichuan University. Pathogen-free female athymic BALB/c nude mice, 6–8 weeks of age, were used to establish the intraperitoneal carcinomatosis model as previously described (24).

Briefly, SKOV3 cells (5×10⁶) in 0.1 ml serum-free DMEM were injected s.c. into the right flank of 5 mice. Tumors were collected and minced into small particles (diameter ≤1 mm), when the tumor diameter reached ~1 cm. The small tumor particles were then resuspended in serum-free DMEM to reach a final volume of 15 ml. Thirty nude mice were inoculated i.p. with 0.5 ml of the above tumor particle suspension, respectively.

To explore the therapeutic efficacy of pshCLDN3/HPEI plus cisplatin, we treated the mice 7 days after inoculation. The mice were randomly divided into 6 groups (5/group), and received the following intraperitoneal (i.p.) administration: (i) 100 μl normal saline (NS); (ii) 50 μg HPEI nanogels in 100 μl NS, every two days for 12 times; (iii) 5 μg pshHK/50 μg HPEI complexes in 100 μl NS every two days for 12 times; (iv) 5 μg pshCLDN3/50 μg HPEI complexes in 100 μl NS, every two days for 12 times; (v) 100 μl of cisplatin, weekly for 4 times (3 mg/kg; Qinu Pharmacy Corporation, China); (vi) 5 μg pshCLDN3/50 μg HPEI complexes in 100 μl NS for 12 times and 100 μl of cisplatin (3 mg/kg) for 4 times. The dose of cisplatin used in this study was 3 mg/kg, since it was previously reported that treatment with 3 mg/kg cisplatin was well tolerated, and did not lead to a complete response but significantly inhibited tumor growth in an ovarian cancer model (25,26).

Two days after the last intraperitoneal injection, mice were sacrificed, and the intraperitoneal tumors were collected and weighed. At the same time following sacrifice, the location of the macroscopic tumors of each mouse was carefully observed and recorded. Tumors were divided into two parts. One was fixed in 10% formalin (pH 7.0) and embedded in paraffin. The other was stored at −80°C for protein extraction.

Paraffin-embedded intraperitoneal tumors were cut into 3- to 5-μm sections. Apoptosis was evaluated by terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) assay according to the manufacturer’s instructions (Promega, Madison, WI, USA). Cell nuclei presenting dark green fluorescence were considered to be TUNEL-positive nuclei. The apoptosis index was determined by calculating the average percentage of TUNEL-positive cells in 5 random fields from 3 different sections at a magnification of ×400.

CLDN3 and Ki-67 immunostaining were carried out with rabbit anti-human CLDN3 antibody (1:100; Invitrogen) and rabbit anti-human Ki-67 antibody (1:100; Thermo Scientific), respectively. Briefly, tumor sections were first deparaffinized, rehydrated and incubated with 3% H₂O₂ for 10 min. Then, antigen retrieval was conducted using 10 mM citrate buffer (pH 8.0). After blocking for 15 min with normal rabbit serum, sections were incubated with the primary antibody, followed by biotinylated secondary antibody and the streptavidin-biotin complex, successively. Detection was performed using diaminobenzidine (DAB) peroxide solution, and then the cellular nuclei were counterstained with hematoxylin. The proliferation index was determined by calculating the average percentage of Ki-67-positive cells with brown-staining nuclei in 5 random fields from 3 different sections at a magnification of ×400.

Health correlated indices such as weight loss, anorexia, diarrhea, cachexia, skin ulcerations or toxic death were evaluated every 4 days to evaluate the possible side-effects. Additionally, at the termination of the animal studies, the main organs (heart, liver, spleen, lung and kidney) were fixed in 10% formalin (pH 7.0), embedded in paraffin, and then cut into 3- to 5-μm sections for H&E staining.

Values are expressed as means ± SD. ANOVA and the Student-Newman-Keuls test were used for comparisons. P<0.05 was considered to indicate a statistically significant result.

Evaluation of CLDN3 expression in the SKOV3 cells was performed by western blot analysis 72 h after transfection. As shown in Fig. 1A, a marked reduction in CLDN3 expression was noted in the cells transfected with the pshCLDN3/HPEI complexes, whereas no apparent alteration in CLDN3 expression was observed in cells transfected with the pshHK/HPEI complexes or the HPEI nanogels, when compared with that in the blank control (NS).

We investigated whether pshCLDN3/HPEI complexes reduce the expression of CLDN3 in vivo. The intraperitoneal tumors were harvested for western blot analysis and immunohistochemistry at the termination of the animal experiments. Similar to the in vitro culture, the pshCLDN3/HPEI complexes markedly reduced the expression of CLDN3 (Fig. 1B), whereas the pshHK/HPEI complexes or HPEI nanogels had no discernible effect on CLDN3 expression. Similar results were observed with immunohistochemistry (Fig. 1C).

At the termination of the animal studies, the mice were sacrificed, and the location of the macroscopic tumors was carefully observed. In the NS, HPEI and pshHK/HPEI groups, each mouse developed intraperitoneally macroscopic tumor nodules scattered on various viscera. A number of tumors were observed not only deposited on the surface of the liver, but also invading the parenchyma. In the other 3 groups, the intraperitoneal tumor nodules were localized. Tumor invasion in tissues and organs was not obvious in these groups.

Next, the intraperitoneal tumors were harvested and weighed. As showed in Fig. 2, the mean tumor weight was 1.35±0.21, 1.29±0.18, 1.16±0.14, 0.45±0.13, 0.40±0.11 and 0.17±0.07 g in the NS, HPEI nanogel, pshHK/HPEI complex, pshCLDN3/HPEI complex, low-dose cisplatin and pshCLDN3/HPEI complex plus low-dose cisplatin group, respectively. The data showed that the pshCLDN3/HPEI complexes or low-dose cisplatin alone significantly inhibited tumor growth, compared with the control therapies (P<0.05). Furthermore, the combination of pshCLDN3/HPEI and cisplatin had a superior antitumor effect, compared with pshCLDN3/HPEI or cisplatin alone (P<0.05). No significant difference in tumor weight was found among the control groups (P>0.05).

Ki-67 immunostaining was used to evaluate the tumor cell proliferation in each group. As shown in Fig. 3A and B, both the pshCLDN3/HPEI complexes and low-dose cisplatin monotherapy group exhibited weak staining for Ki-67, when compared with the control groups (P<0.05). No significant difference in Ki-67 staining was visible in the two groups (P>0.05). Moreover, the percentage of Ki-67-positive cells was significantly reduced in the combination therapy group, when compared with the percentage in the pshCLDN3/HPEI or low-dose cisplatin monotherapy group (P<0.05).

We used TUNEL assay to detect apoptotic cells in the tumor tissues of each group. The results showed that both the pshCLDN3/HPEI and cisplatin monotherapy resulted in a significant increase in apoptotic tumor cells when compared with that following treatment with the control agents (P<0.05). Tumors in the combination therapy group showed an increased number of positive nuclei, when compared with the number in the pshCLDN3/HPEI or low-dose cisplatin monotherapy group (P<0.05). However, positive nuclei were rare in the control groups (Fig. 3A and C).

In the present study, no gross abnormalities were observed in any group. As shown in Fig. 4, although the mean body weights in the two cisplatin-treated groups were less than that in the other four groups after the treatment course, the differences did not reach statistical significance (P>0.05). Furthermore, no pathologic change in the heart, liver, spleen, lung or kidney was noted by microscopic examination after the administration of pshCLDN3/HPEI or pshCLDN3/HPEI plus cisplatin.