The following reagents were purchased: cisplatin (Jiangsu Hansoh Pharmaceutical Co., Ltd., Jiangsu, China); RNAiso Plus, RT reagent kit with gDNA Eraser, and SYBR Premix Ex Taq II (Takara Bio Inc., Dalian, China); anti-VEGF-D (Bioworld Technology, Inc); immunohistochemistry (IHC) kit (ZSGB-Bio., Beijing, China); and blood urea nitrogen (BUN) and serum creatinine (SCr) kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China).

All protocols followed the guidelines of ethics regarding the treatment of animals used in experiments. This project was approved by the Dalian Medical University Ethics Committee in China. Female C57BL/6J mice (6 weeks old; n=5–6 mice/group) were obtained from the Dalian Medical University Laboratory Animal Center and were randomly divided into 6 groups. Tumor-bearing mice were treated with saline, 2 mg/kg of cisplatin, 12.5 mg/kg cisplatin with or without 500 μg/kg G31P or 500 μg/kg G31P alone. A sixth group of healthy mice served as the normal control animals. We had previously optimized the dose of human G31P required for use in rodents (Gordon JR, unpublished data), as we had in multiple other species.

Tumor-bearing mice received 5×10⁶ H22 mouse hepatoma cells (0.2-ml volumes, s.c.). G31P treatment mice were administered G31P s.c. on alternative days, commencing 7 days after tumor implantation, while the remaining groups received normal saline in the same manner. Cisplatin treatment mice were administered either 12.5 mg/kg (high dose) or 2 mg/kg (low dose) cisplatin i.p., while the remaining mice received an equal volume of saline i.p. The clinical status and the body weight of each mouse were evaluated daily, and on day 21 all mice were sacrificed for assessment of tumor progression. The primary tumors were resected from each animal and weighed.

Blood samples were collected from the angular vein of the mice under pentobarbital sodium (50 μg/kg) at 4, 8, 12, 24, 48 and 72 h and at day 21 post-tumor cell implantation. Serum samples were stored at −20°C for blood urea nitrogen (BUN) and serum creatinine (SCr) detection. BUN and SCr values were determined according to the supplier’s instructions, as standard measures of renal function (2). In addition, the renal coefficient (RC) was determined for each animal as a surrogate measure of general kidney health (20). To do so, both kidneys were excised and weighed and the RC was calculated using the following formula: Kidney weight/body weight. The kidneys were then snap-frozen in liquid nitrogen, and stored at −80°C prior to real-time PCR analysis.

RNA of renal tissues was isolated using RNAiso Plus kits according to the supplier’s protocol. Total RNA was reverse transcribed into cDNA using Takara RT reagent kit with gDNA Eraser. Briefly, first the genomic DNA was eliminated from each sample by incubation at 42°C for 5 min and then the reverse-transcription reactions were performed. PCR parameters were set as 37°C for 15 min and 85°C for 5 sec. The transcriptional levels of IL-1β, KC and MIP-2 were determined by quantitative real-time PCR (qRT-PCR) analysis with Takara SYBR Premix Ex Taq II. Melting curve analyses were used to verify the accuracy of the PCR products following the amplification reactions. The PCR-primer sequences are indicated in Table I. The PCR cycling parameters (40 cycles) were set as: pre-denaturation (95°C for 30 sec); denaturation (95°C for 5 sec); annealing (59°C for 30 sec); and extension (72°C for 1 min). The relative expression values were normalized to that of β-actin. The 2^−ΔΔCT method was used for analysis of the data.

Mice were sacrificed after 21 days of treatment and each tumor was resected and weighed. Solid tumor samples were fixed in 4% paraformaldehyde and embedded in paraffin using standard procedures in preparation for immunohistochemical (IHC) detection of VEGF-D. Briefly, endogenous peroxidase was blocked with 3% H₂O₂ for 10 min at room temperature. Afterward, the samples were rinsed three times with PBS (5 min each), incubated for 20 min at room temperature with normal goat serum as a blocking agent, washed as above with PBS and then incubated overnight at 4°C with the primary antibody (1:100 dilution). The tissues were again washed with PBS (3 times for 2 min each), incubated for 1 h at 37°C with the secondary antibody, washed again with PBS as above, incubated in diaminobenzidine (DAB) solution for 10 min and counterstained with hematoxylin. The staining results were assessed by image analysis using the Microsoft Image-Pro 6.0.

Statistical analysis of the data and inter-group comparisons were carried out using the Student’s t-test (two-tailed). Probability values of <0.05 were considered to indicate statistically significant results. The results are expressed as means ± SEM.

Upon examination, we found that, overall, the tumor-bearing mice gained less weight than the healthy normal control mice (Fig. 1A). All cisplatin-treated animals displayed reduced activity, hypotrichosis, lackluster attitude and anorexia during the first 72 h following drug treatment. The 12.5 mg/kg dose cisplatin treatment led to substantial loss of body mass during the first 6 days post-treatment, whether or not the animals were co-treated with G31P, while the mice treated with G31P only or a 2 mg/kg dose of cisplatin lost no weight during this period. Weight gain in the high-dose cisplatin-treated tumor-bearing animals remained suppressed for the first 2 weeks, while G31P co-treatment ameliorated this effect somewhat. On day 21, all tumor-bearing mice weighed significantly less than the healthy control mice (P≤0.05; Fig. 1A).

On day 21 we also assessed the weight of the primary tumor in each animal and found that G31P treatment alone reduced the sizes of the primary tumors by 19%, while treatment with 2 mg/kg cisplatin reduced the tumor mass by 29%. High-dose cisplatin treatment alone reduced the size of the tumors in our mice by 52% (Fig. 1B; P≤0.05), and the addition of G31P co-therapy to this cisplatin treatment further enhanced the growth-suppressive effect observed, such that these tumors were 71% smaller than those in the saline-treated tumor-bearing mice (P≤0.05; Fig. 1B). The effect of high-dose cisplatin and G31P co-treatment on the progression of the primary tumors was additive.

As VEGF-D expression is strongly linked to tumor dissemination via metastasis (21,22), we assessed expression of VEGF-D in our H22 hepatoma cell tumors using immunohistochemical methods and image analysis (Fig. 2). We observed strong expression of VEGF-D in the tumors of the saline-treated mice, and neither treatment with the ELR-CXC chemokine antagonist G31P alone nor low-dose cisplatin therapy had discernible effects on this expression. In contrast, high-dose cisplatin therapy, either alone or in combination with G31P co-therapy, reduced expression of VEGF-D in the tumors by ~50% (Fig 2B). Given the associations between VEGF-D production and metastatic disease, and the prognostic implications of extrahepatic metastasis in hepatocellular carcinoma (23), we also assessed the impact of ELR-CXC chemokine antagonism and cisplatin therapy on lymph node metastasis in our model. We resected the inguinal lymph nodes of our mice at the time of sacrifice and examined them directly for evidence of metastases using H&E-stained tissue sections (Fig. 3). We found abundant metastases in the lymph nodes of the saline-treated H22 hepatoma cell tumor-bearing mice, and evidence of metastases in the G31P only and low-dose cisplatin-treated animals. Few tumors were noted in the mice treated with high-dose cisplatin and particularly with the combination of G31P/high-dose cisplatin (Fig. 3).

High-dose cisplatin therapy has been reported to induce renal inflammation and toxicity (2). Therefore, we assessed the expression of inflammatory mediators in our cisplatin and G31P-treated mice (Fig. 4), as well as the renal function, as determined by serum levels of blood urea nitrogen (BUN) and serum creatinine (SCr) and the renal coefficients (Fig. 5). We found that the saline- and G31P-treated mice displayed slight renal IL-1, KC and MIP2 expression thoughout the 3 days of observation (Fig. 4). Notably, the high- and low-dose cisplatin-treated mice expressed substantial levels of IL-1 during the 48 h following drug infusion, and the G31P co-treatment only modestly reduced this response. By day 3 the IL-1 response in the mice treated with low-dose cisplatin and G31P/high-dose cisplatin was again at background, while there remained a small IL-1 response at this time in the high-dose cisplatin-treated animals (Fig. 4A). The low-dose cisplatin treatment had a slight effect on inducing an ELR-CXC chemokine (i.e., KC or MIP2) response, while the high-dose cisplatin therapy markedly elevated the expression of both chemokines, albeit with distinct kinetics (Fig. 4B and C). The G31P co-treatment appeared to fully off-set the 8-h KC and MIP2 response to cisplatin, but by 48 h these animals also expressed substantial levels of both of these potent neutrophil agonists.

We also ascertained whether the mice were suffering from cisplatin-induced nephrotoxicity and whether CXCR1/CXCR2 antagonism had an impact. Neither the saline- nor the G31P only-treated hepatoma tumor-bearing mice displayed discernible evidence of kidney damage, while administration of cisplatin led to dose-dependent nephrotoxicity, as determined by increases in BUN, SCr and RC values (Fig. 5), with the high-dose therapy being overtly toxic. While G31P co-treatment significantly augmented the therapeutic outcome of high-dose cisplatin in terms of reducing tumor progression in our model, this combined therapy had a substantial beneficial impact on nephrotoxicity, as measured by serum levels of BUN and SCr and the RC values (P≤0.05, for G31P/high-dose cisplatin vs. high-dose cisplatin therapy alone). Indeed, the BUN and SCr levels in the mice treated with G31P/12.5 mg/kg cisplatin were approximately equivalent to those in the low-dose cisplatin-treated mice. Moreover, the RC values in the mice treated with G31P/high-dose cisplatin were lower overall than these values in the mice treated with low-dose cisplatin, indicating that co-delivery of G31P allowed the use of higher, more tumor-growth suppressive doses of cisplatin while at the same time protecting the animals against these cisplatin-associated toxic side-effects.