Impact of laparotomy and liver resection on the peritoneal concentrations of fibroblast growth factor 2, vascular endothelial growth factor and hepatocyte growth factor

Abstract

Purpose: Some data have suggested that major surgery is associated with the post-operative growth of residual tumour masses but the mechanism of this is unknown. This study was designed to determine the relationship between intraperitoneal (IP) cytokine levels, and laparotomy in benign and malignant settings. Methods: Intraperitoneal fluid specimens were obtained at the start and at the end of laparotomy in patients with benign conditions (n=10) and in others undergoing resection of hepatic metastases from colorectal cancer (n=10). Using ELISA the concentration of the angiogenic cytokines, HGF, VEGF-A, VEGF-C, VEGF-D and FGF-2 was determined. Results: The data show that in 16 of 20 patients there was a significant increase (P=0.006) in the IP concentration of hepatocyte growth factor (HGF) but not in the other growth factors by the end of the operation. The mean increase in HGF concentration was 821.5 pg/ml (95% CI: 11.0–6,426.0). Neither the groups (malignant and non-malignant) nor the length of operation correlated with greater or lesser increases in HGF. Conclusion: The observation that the increase in HGF occurred in both the cancer and non-cancer groups suggests that it is the surgery rather than the disease that is associated with the increased cytokine concentration. As HGF is a potent endothelial, epithelial and mesenchymal mitogen the data highlight HGF as a potential target for anti-cancer treatments in the peri-operative period. However, investigators should closely monitor wound healing as this may be compromised by this new class of drugs.

Introduction

Surgery is the initial modality of treatment for most solid tumours. Some data suggest that removal of the primary tumour can be associated with enhanced growth of distant metastases (Slooter et al. 1995) but the mechanism is unclear. On the one hand, surgical excision may remove the source of a variety of growth inhibitors resulting in angiogenesis and subsequent growth of previously dormant metastases (O’Reilly et al. 1994; Folkman 1995). Alternatively, residual tumour cells might bear receptors for cytokines that are released in response to surgery.

Very few data exist on the changes in peritoneal cytokine concentrations following abdominal surgery. Those studies that have investigated the issue have revealed increases in inflammatory cytokines such as tumour necrosis factor-α, interleukin-1β (IL-1β)(Tsukada et al. 1993) and IL-6 (Badia et al. 1996) with the suggestion that the length or stress of the surgery is associated with increased concentrations (Grande et al. 2002; Balague et al. 1999; van Berge Henegouwen et al. 1998; Odegard et al. 2000; Bown et al. 2001; Tashiro et al. 1999). As the change in peritoneal cytokine concentration is greater than that seen in the plasma (Badia et al. 1996) the implication is that the post-operative cytokine response may originate largely from the peritoneal cavity.

We were interested in the impact of surgery on the changes in the intraperitoneal (IP) concentrations of the angiogenic cytokines VEGF-A, -C and -D, hepatocyte growth factor (HGF) and fibroblast growth factor 2 (FGF2) which have been reported at higher concentration in the sera of cancer patients in comparison with healthy individuals (von Schweinitz et al. 2000; Dirix et al. 1996, 1997). These observations suggest that VEGF, FGF2 and HGF may play a role in the growth of residual tumour cells but their role and importance remain to be defined. No studies have investigated changes in peritoneal cytokines following liver resection for hepatic colorectal cancer metastases (HCRM). The aim of the present study was to compare the intra- and post-operative peritoneal concentrations of VEGF, FGF2 and HGF in patients undergoing laparotomy for partial hepatectomy for HCRM with those undergoing laparotomy for a benign condition.

Materials and methods

Patient selection and recruitment

Nine patients with a diagnosis of colorectal cancer who had been selected for partial hepatectomy to remove HCRM entered the study. Another patient who was undergoing a planned cholecystectomy for suspected benign disease was found to have cholangiocarcinoma and her data were analysed with those of the other cancer patients. Ten patients who were undergoing laparotomy for benign conditions participated as controls. All patients underwent surgery at North Manchester General Hospital, Manchester, UK. The study was approved by the ethics committees of North and South Manchester and has therefore been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki. All patients gave informed consent.

Sample collection and processing

Peritoneal fluid was collected intra-operatively at initial laparotomy prior to any surgical procedure being performed and a second sample was taken post-operatively just prior to closure. The pre-operative peritoneal sample was important especially for the HCRM patients (n=9) as it was taken prior to any hepatic mobilisation. Peritoneal fluid samples were collected in sterile containers and immediately centrifuged for 5 min at 1,500 g. The resulting supernatant was removed and stored in additive free bottles at −80°C until assayed. Each sample underwent only a single freeze--thaw cycle.

ELISA for VEGF, FGF2 and HGF

The concentrations of VEGF, VEGF-D, FGF2 and HGF in peritoneal fluid were measured using commercially available sandwich ELISA kits (Quantikine Kit, R&D Systems, Abingdon, Oxon, UK) following the manufacturers’ guidelines. All samples were assayed in duplicate and where necessary, samples were diluted in buffer from the kits. The assay sensitivity was 9 pg/ml for VEGF, 40 pg/ml for HGF and 1.0 pg/ml for FGF-2. VEGF-C was assayed using an in-house ELISA assay that has been previously described (Duff et al. 2003). Statistical analysis was performed using the SPSS package for Windows version 11.5.

Results

Patient characteristics

We analysed ten patients with the diagnosis of cancer (group 1) and ten patients with benign diseases (group 2). The median age was 57 years old (39–77) for group 1 and 59 years old (22–69) for group 2. There were 12 male patients (eight in group 1 and four in group 2). Our patients follow the guidelines for hepatic metastasis resection and therefore had disease limited to one lobe of the liver with no extra-hepatic disease found. A range of different sizes and number of metastases were present in this group. The duration of the surgery was not significantly different between the two groups (median of 150 min for group 1; range 90–320; median of 83 min for group 2; range 60–280; P=0.055).

Intra-peritoneal cytokines

We measured the concentrations of VEGF-A, -C and -D, FGF-2 and HGF in the fluid aspirated upon entering the patients’ abdomens and a second sample was taken upon closure. No further samples were taken as the drains were withdrawn from the patients’ abdomens in accordance with standard practice and we wished to avoid degradation of the growth factors in in-dwelling drains. Table 1 shows that there were no significant differences in the levels of cytokines between the two groups, considering pre-resection and post-operative levels. There was a statistically significant rise in the IP concentration of HGF in the immediate post-operative phase in patients with both non-malignant and malignant conditions. Sixteen of the 20 patients in the trial had increased concentrations of HGF (P=0.006; Wilcoxon Signed Rank Test). In contrast, no differences were seen when VEGF-A, -C, -D and FGF-2 were studied in the early post-operative period (Mann-Whitney U tests using difference data).

Table 1.

Intra-peritoneal concentrations of cytokines before and after resection, considering the two groups of patients

Mean concentration of cytokines (pg/ml)a	Cancer (n=10)	Control (n=10)
VEGFA
Before resection	175.78	5.89
Post-operative	516.67	67.22
VEGFC
Before resection	2.46	1.43
Post-operative	3.27	1.48
VEGFD
Before resection	126.33	59.00
Post-operative	247.22	44.44
HGF
Before resection	1019.78	362.33
Post-operative	2683.44	1676.11
FGF2
Before resection	596.11	658.56
Post-operative	620.11	696.89

^aThere were no significant differences between the groups considering the concentration of each cytokine; there were no significant differences in each group when comparing cytokines levels before and after resection

The median increase in HGF concentration when all the patients were analysed together was 821.5 pg/ml (95% CI: 11.0–6,426.0). Figure 1 shows the individual changes in IP HGF concentrations in the two groups.

Fig. 1.

Changes in intra-peritoneal HGF concentrations according to the two groups, presented as the post-operative value minus the intra-operative value (P=0.796)

Interestingly, there did not seem to be a relationship between changes in cytokine concentration and the duration of the operation. The patients undergoing resection of liver metastases underwent longer operations in general than the other group, but there was no evidence of a greater rise in IP HGF concentration.

Discussion

This investigation has shown that laparotomy is associated with a significant increase in the peritoneal fluid concentration of HGF, a potent epithelial, mesenchymal and angiogenic cytokine (Birchmeier et al. 2003). Receptors for the cytokine have been detected in a number of tumours but particularly in colorectal carcinoma (Fazekas et al. 2000) suggesting that an increase in concentration of this magnitude might stimulate the growth of residual tumour cells (Portera et al. 1998; Liu et al. 1992).

The observation that the increase in HGF occurred in both the cancer and non-cancer groups suggests that it is the surgery rather than the disease that is associated with the increased cytokine concentration. While it is well established that HGF plays a role in liver regeneration (Dtuzniewska et al. 2002; Huh et al. 2004) our finding that the rise in post-operative HGF levels was independent of liver surgery suggests that the data may reflect part of the normal early healing process, which, in the context of patients with cancer, results in a change in the cytokine milieu that could stimulate cancer regrowth. One caveat might be that some patients in the control group might have inflammation that was associated with the increase in HGF. Thus the difference between cancer and non-cancer patients was obscured. Nevertheless, the critical point is that irrespective of the indication for laparotomy 80% of the patients experienced an increase in peritoneal fluid HGF concentration.

Our sample collection was limited by the withdrawal of the abdominal drains in the immediate post-operative period. Therefore it will be important, when extending this study, to take samples at later time points if possible to define more precisely the area under the concentration-time curve for HGF. This might also reveal late changes in VEGFs or FGF that were not observed in this study. Such samples might be available through abdominal drain peritoneal fluid. However, in this study we wished to avoid the possibility that cytokines might degrade in abdominal drains and therefore confined the initial investigation to samples that were immediately available, thereby avoiding any artefactual changes in growth factor concentrations.

Around the world there is an increasing focus on the development of inhibitors of HGF, a potent angiogenic and epithelial mitogen and motogen. Our results suggest that the cytokine might be a suitable target for the abrogation of post-operative tumour growth and for the first time define a new indication for therapeutic anti-cancer intervention. However, investigators should closely monitor wound healing as this may be compromised by this new class of drugs.