Background
Ovarian cancer is the leading cause of mortality among the gynecological diseases occurring in developed countries with 225,500 new cases and 140,200 estimated deaths worldwide [
1,
2]. This high rate of mortality is due in part because most patients presenting epithelial ovarian cancer have an advanced disease at the time of diagnosis [
1]. After surgical resection, residual disease then represents the major pejorative predictive factor for survival [
2]. For optimum results, surgery must be as complete as possible which generally means performing en-bloc resections of bowel, reproductive organs and genital tract, and peritonectomies.
The corollary of these “radical” surgeries is the induction of intra-abdominal adhesions that lead to the sticking together of two tissues normally moving freely past each other. As well as causing bowel obstruction and chronic pain, these adhesions can also induce heterogeneity of drug diffusion and compromise the delivery of adjuvant intraperitoneal chemotherapy [
3]. Moreover, they can complicate any new surgical procedures, which in ovarian cancer disease are highly frequent due to the high rate of recurrence (70%). The place of secondary cytoreductive surgery is actually evaluated in several clinical trials with a complete resection as a gold standard. It is clear therefore that the reduction of adhesive disease has many immediate and delayed functional and therapeutic consequences in ovarian cancer bearing patients. An exhaustive review published by
Ward B.C describes all adhesion prevention strategies including non-invasive surgical strategies, cellular strategies, pharmaceutical approaches, fluid and gel barriers, and lastly solid barriers including Hyaluronic Acid-Carboxymethyl Cellulose (HA-CMC) barrier [
4].
HA-CMC barrier is an implantable and bioabsorbable synthetic membrane, composed of hyaluronan (also called hyaluronic acid or HA) and carboxymethylcellulose (CMC) which turns into a gel within 24 hours, before being absorbed from the surgical site within one week for elimination from the organism in less than 30 days. It became available in France in 2003 for indications of benign abdominal surgeries and since 2007, following Oikonomakis [
5] and Kusonoki’s publications [
6] for respectively colon and rectal cancers, is also used for carcinologic surgeries. Since then, many preclinical, animal and clinical works studying HA-CMC barrier in digestive cancers have been published. A meta-analysis performed by
Diamond MP describes this material as « biocompatible, non-mutagenic, non-pyrogenic, non-irritating and non-toxic » [
7]. However, some animal studies have given contradictory results revealing no effect of HA-CMC in colon tumor metastasis but a local increase in tumor growth [
8], and in some cases, an increase in rate of abdominal metastasis [
9].
The major component of HA-CMC barrier, HA, is a physiological component of extracellular matrix in association with collagen, proteoglycans (such as fibronectin, laminin and vitronectin) and elastin. Its main receptor is CD44 which is present at the surface of various normal or tumor cells. Another HA receptor, CD168, also called receptor hyaluronic acid mediated motility (RHAMM), has been described in CD44 knock-out mice [
10]. The HA expression level is modified in the context of ovarian cancer and high levels are correlated with poor prognosis [
11]. In histological retrospective studies, tumor stromal HA concentration has been linked with tumor aggressiveness, and stromal HA has been shown to be an independent prognostic factor for disease free and overall survival [
12,
13]. Moreover, high level of HA synthase 1 (HAS1), but not HAS2 and HAS3, is correlated with reduced overall survival in ovarian cancer [
14]. Furthermore, CD44 expression level of ovarian cancer cells is inversely proportional to patient survival [
15].
Many studies, mainly
in vitro, have allowed us to define more precisely the implication of HA in ovarian cancer. As an example, HA from mesothelial cells, the major cell population in the peritoneum, is an adhesion factor for CD44-positive ovarian cancer cells [
16]. These data have been confirmed by adhesion tests using hyaluronidase, anti-CD44 antibodies and CD44 targeted siRNA [
17‐
19]. Mesothelial extracellular HA is also involved in ovarian carcinoma cell motility and dose-dependent chemotactism. Indeed, Boyden chamber tests highlighted than ovarian cancer cells preferentially migrate towards mesothelial cells as compared to control conditions, and that this effect is abolished using hyaluronidase [
20]. Moreover, HA-CD44 interaction in SKOV-3 ovarian cancer cells can induce F-actin oriented polymerization using two signaling pathways involving Src or N-WASP [
20‐
22].
Among non-gynaecological cancers, a retrospective study showed that HA-CMC barrier use has been associated with a 13% increased risk of abscess after colectomy [
23]. On the contrary, HA-CMC barrier has not been showed to be associated with increased complications in thyroid surgery despite an uncertain anti-adhesive effect [
24].
While some studies regarding HA-CMC barrier utilization in gynecological surgery have been published [
25‐
29] none describe the potential effects of this biomaterial in ovarian metastatic process in animals. Of the mainly clinical studies, most are retrospective. Only one is prospective including 14 patients with HA-CMC barrier, and concern in half of the cases cervical and uterine cancers. Krill
et al. evaluated the risk of postoperative complications related to HA-CMC utilization in ovarian cancers after cytoreductive surgery and showed that this biomaterial is not responsible for major complications but may be associated with increased risk of pelvic abscess [
30]. However, they did not study the effects of HA-CMC on ovarian cancer progression and recurrence.
Our concern was to evaluate the potential effect of HA-CMC in a context where residual cancer cells were present. Indeed, after ovarian surgery, and because 70% of patients will relapse, it is clear that those cells could be affected by HA-CMC presence.
Accordingly, we decided to perform a preclinical study in order to evaluate HA-CMC barrier impact on ovarian tumor progression. In vitro experimentations were performed to analyze the effects of HA-CMC barrier on the proliferation of several ovarian cancer cell lines and on the activation of proliferative, survival and adhesion signaling pathways. In vivo mice xenograft models were used to evaluate the oncological risk of HA-CMC barrier utilization in comparison with expected benefits from adhesion prevention. Our ultimate aim was to provide a clear assessment of the appropriateness of using HA-CMC barrier in the surgical management of ovarian cancer. We demonstrated that in vitro, HA-CMC barrier did not increase tumor cell proliferation and that in vivo, except for anterior and lateral peritoneum implantation, HA-CMC barrier did not induce tumor growth.
Methods
Biomaterials
HA-CMC barrier membranes and absorbable material were granted from the surgical department of Claudius Regaud Institute (Toulouse France) without industrial founds.
Cell culture
NIH ovarian adenocarcinoma cells (OVCAR-3 and SKOV-3) and HeLA cells were obtained from the American Type Culture Collection (ATCC® numbers HTB-161, HTB-77 and CCL-2). IGROV-1 ovarian adenocarcinoma cells were a gift from the Gustave Roussy Institute (Villejuif, France) and the REH lymphoblastic line was a gift from the hematology research unit of Toulouse University Hospital (France). All cells were cultured in RPMI medium supplemented with 10% fetal calf serum (FCS), penicillin/streptomycin (100 IU/mL/100 μg/mL) and 2 mM L-glutamine (Cambrex biosciences, Milan, Italy). Cell lines were routinely checked for mycoplasma.
Cell proliferation
OVCAR-3, IGROV-1 and SKOV-3 cells were seeded in 6-well plates at a concentration of 2 × 105 cells per well. Three different conditions were used: culture with RPMI alone or with 1 cm2 pieces of HA-CMC barrier or a control absorbable material. Cell proliferation kinetics was assessed by counting cells on a hemocytometer after 1, 2, 4 and 6 days of culture. Cell population doubling time during the exponential growth phase was calculated according to the following formula: T(hours) = t × [ln(2)/[ln(C1) ‒ ln(C0)]] in which t corresponds to the exponential growth phase in hours, C0 to the initial cell concentration at the beginning of this phase and C1 to the final cell concentration at the end of the phase.
CD44 expression evaluation by flow cytometry
Each cell line (106 OVCAR-3, IGROV-1, SKOV-3, HeLA or REH lymphoblast cells) was incubated during 45 minutes with a primary antibody directed against CD44 (1/100, [F10-44-2] (ab6124), Abcam, Paris, France), or against a control isotype (κ isotype Ctrl PE Mouse IgG1, BioLegend). Cells were then washed with PBS, centrifuged 5 minutes at 300 g, then incubated 45 minutes with an anti-mouse secondary antibody (1/50, BD Biosciences, Le pont de Claix, France). Cells were washed with PBS and the fluorescence was measured with a FACS Calibur cytometer and analyzed with Cell Quest Pro software.
Animals
Four to five week-old female Swiss Nude athymic mice (Charles River laboratories, l’Abresle, France) were used after approval from Claudius Regaud Institute animal ethics committee (ICR-2012-017). They were housed according to the European Laboratory Animal Science Association standards. Mice experimentations began after one week of quarantine.
HA-CMC barrier mouse absorption validation
A 2.25 cm2 (15 × 15 mm) piece of HA-CMC barrier was inserted into the abdomens of mice by median laparotomy after intraperitoneal general anesthesia (ketamine 50 mg/mL, xylazine 20 mg/mL and NaCl: 10 μg per gram of mouse). Mice were then sacrificed via cervical dislocation 1, 3 or 7 days after surgery in order to proceed to histological analysis.
Subcutaneous (s.c.) xenografts
107 SKOV-3 cells were injected subcutaneously into each of the right and left flanks of 16 mice. The day following this injection (D1), after cutaneous debridement with chlorhexidine, an incision was made under general anesthesia at the two injection sites and a 1 cm2 (10 × 10 mm) piece of HA-CMC barrier was alternatively inserted into one of the two sites. The other side was opened and then closed to reproduce the inflammation caused by the surgical procedure. The well-being of the mice was checked every 2 to 3 days. Tumor lengths (L) and widths (w) were also measured every two to three days and the tumor volume extrapolated using the formula: (π × length × width2)/6. Mice were sacrificed 21 days via cervical dislocation after the injection. Tumors were sampled then conditioned for histological analysis (formol fixation) or western blot (cryopreservation in liquid nitrogen).
Intraperitoneal (i.p.) xenografts
2.5 × 107 SKOV-3 cells were injected intraperitoneally into each of 22 mice. The day following the injection (D1), after cutaneous debridement with chlorhexidine, a laparotomy was performed under general anesthesia according to one of two procedures: (A) 11 mice underwent a “white” laparotomy (incision, abdominal opening and closure) and (B) 11 mice received an i.p. 2.25 cm2 (15 × 15 mm) piece of HA-CMC barrier. For both groups, parietal closure consisted in three separate absorbable material 3/0 sutures. The well-being of the mice was checked during 21 days and their weight was measured three times a week. The mice were sacrificed by cervical dislocation 21 days after the injection of cells. Autopsies were then performed by two double-blinded specialist surgeons in ovarian cancer surgery to evaluate the abdominal carcinosis index and tumors were removed for histological and western blot analysis.
Histological analysis, Ki67 staining and mitotic index determination
Proliferative indices of tumor from s.c. or i.p. injections were assessed by immunohistochemical staining of paraffin embedded tumor sections for the proliferation marker Ki67 using mouse monoclonal MIB1 antibody. Ki67 positive nuclei were counted in random fields. The mitotic index was assessed by evaluating the number of cells in mitosis per high-power field (10 high-power fields per tumor).
Western blot
Western blot was performed to evaluate in vitro protein expression in CD168 cell lines and to analyze in vivo the activation of signaling pathways within the xenografts. In vitro, proteins in OVCAR-3, IGROV-1, SKOV-3, HeLa and REH lymphoblastic cells were studied. In vivo, cryopreserved xenograft tumors were pooled according to protocol (subcutaneous versus intraperitoneal, with or without HA-CMC barrier). Cells and tumor samples were crushed then prepared with RIPA lysis buffer (Tris 1 M at pH 7.4, NaCl 150 mM, Triton, SDS 20%, MgCl2, NaF and protease inhibitors). The concentration of proteins was determined using the « Bradford Assay » (Bio-Rad) before their separation by SDS-PAGE in a 10% polyacrylamide gel and transfer onto a PVDF membrane (Amersham Hybond) previously activated with methanol. These membranes were saturated 45 minutes in TBS (50 mM Tris, 150 mM NaCl) / Tween 0.2% (TBST) supplemented with 5% milk and then incubated 90 minutes with a rabbit primary antibody against CD168 (1/1000, ab108339, Abcam) for cells lines, or Akt, p-Akt, p-ERK or FAK (1/1000, Cell Signaling Technology) for tumor samples. Membranes were washed three times with TBST and incubated 90 minutes with an anti-rabbit secondary antibody coupled with horseradish peroxidase (HRP) (1/2000, Cell Signaling Technology, St Quentin, France). Membranes were washed three times with TBST and twice with TBS before immunocomplexes were revealed with Enhanced Chemoluminescence (GE Healthcare) and visualized with a photon camera (Bio-Rad).
Statistics
Quantitative variables were presented by median and range and qualitative data by frequency and percent. Group comparisons were made using the Wilcoxon signed rank test (matched) or Mann–Whitney test (independent) for non-parametric data and by the Student’s t test for parametric ones. For this entire study, statistical significance was reached for p < 0.05. In vivo, with a standard deviation of 0.1, to reach a statistical difference mean of 15% in terms of tumor proliferation with a statistical power of 80% and α risk of 5%, 22 mice per experimentation were necessary. All analyses were performed using STATA software (Version 12.0; Stata Corporation, College Station, TX, USA).
Discussion
Complete cytoreductive surgery for advanced ovarian cancer can induce extended abdominal adhesions. These adhesions may prevent optimal intraperitoneal treatment delivery and can cause bowel obstruction and pain. The need to prevent such adhesions during abdominal surgery is a unanimously agreed objective. Among the solid barriers placed to decrease adhesive disease, HA-CMC barrier is one of the most widely used, and yet its essential component HA is a prognostic factor of survival in ovarian cancer bearing patients. Despite our in-depth study of the literature, we found no conclusive data regarding the oncological risk of using HA-CMC barrier in these patients.
Our results show that HA-CMC barrier does not activate the proliferation of the ovarian tumor cells OVCAR-3, IGROV-1 or SKOV-3
in vitro. Moreover, it cannot be excluded that the biomaterials may have actually acted as a physical constraint of cell proliferation. Indeed, whether the biomaterial used was HA-CMC barrier or an absorbable material, cell proliferation was less active than without the biomaterial. Particularly with IGROV-1 cells, HA-CMC seemed to present an anti-proliferative effect. Observing the same number of IGROV-1 cells in each condition at day 1 (Figure
2B and D), we did not conclude on an anti-adhesive effect of HA-CMC barrier on IGROV-1 cells. Considering that
in vitro experiments exclude microenvironment participation, we studied
in vivo the potential implication of HA-CMC barrier in ovarian tumor progression in mice.
Nearly all the results obtained
in vivo, whether they concerned subcutaneous tumor cell injections (tumor volume, mitotic index or Ki67 percentage) or intraperitoneal xenografts (mean weight of mice, median carcinosis score of the diaphragm, liver and intestine, mitotic index or Ki67 percentage), we observed no proliferation increase associated with the HA-CMC barrier. Our findings corroborate other studies concerning non gynecological cancer in which HA-CMC barrier was described to protect trocar sites from gall bladder metastasis and abdominal wall implantation [
32] while not decreasing overall survival for colon and rectal cancers [
33].
However, our study does highlight that an i.p. HA-CMC barrier implantation can induce a statistically significant increase in the peritoneum carcinosis score (excluding hemi diaphragm). While this was the only detected pejorative impact of HA-CMC barrier, Hubbard et al. did publish similar findings for KM12-L4 colon cancer xenografts in mice, observing a local increase in tumor volume in the sidewall of the HA-CMC barrier and absorbable material groups. This increase was thought to be due to local inflammation induced by the presence of an exogenous biomaterial [
8].
This result does warrant some concern, but such tumor growth was isolated in our study. Human peritoneal carcinosis extension and resecability evaluation are respectively performed with Sugarbaker and Fagotti scores but there is no equivalent for murine carcinosis experimentation. We decided to elaborate our own scoring system, which is applicable macroscopically without any endoscopic or informatics systems contrary to Matsuzaki published score [
31]. There were some differences between the two double-blinded evaluation but the mean difference concerning the peritoneal carcinosis score, which constitutes the only localization apparently impacted by HA-CMC barrier, was only 6.83%.
Since we were unable to highlight a clear implication of HA-CMC barrier in ovarian tumor progression using macroscopic parameters, we focused on a potential effect of this biomaterial on molecular activation. To this end, we studied potential alteration in the expression of proteins involved in cell survival (total Akt and phospho-Akt), cell proliferation (phospho-ERK) and adhesion (FAK). These signaling pathways have been described in the literature as being activated following AH-CD44 interactions (PI3K-Akt [
34], ERK [
35] or FAK [
36]). However we observed no activation of these different pathways in the HA-CMC barrier group of tumors by comparison with control tumors thus confirming the lack of effect of this biomaterial present on ovarian tumor behavior.
HA-CMC barrier-derived HA must therefore be differentiated from extracellular endogenous HA, the expression of which favors the evolution of ovarian cancer carcinosis. One of the possible explanations for the safety of HA-CMC barrier-derived HA could be the difference in its molecular weight compared to endogenous HA; high weight HA (4000 MDa) is anti-angiogenic and non-immune response-inducing and low weight HA (10 MDa) is pro-inflammatory, pro-angiogenic and immune response-inducing [
37]. Another explanation could be that HA-CMC barrier HA shows less affinity for CD44-expressing tumor cells. To test such a hypothesis, it would be interesting to determine the phosphorylation status of CD44 receptors in response to HA-CMC barrier placement
in vitro. Endogenous HA effects are likely the consequence of several interactions, not only with tumor cells but also with the tumor microenvironment. Moreover, they may also depend on their combined activation of surface receptors with other extracellular matrix components such as vitronectin or fibronectin.
We performed our
in vivo experiments on nude mice models. These animals presented a lack of mature T cells that could modify the response of ovarian tumors to HA-CMC barrier. However, we focused on human tumor cells behavior in nude mice where macrophages, neutrophils and NK cells were present and responsible for a local inflammation. Indeed, we showed an acute inflammatory reaction 3 and 7 days after HA-CMC barrier injection in nude mice. We concluded that the absence of T cells did not alter the inflammation process in our model. Moreover, the describing functions of all T cells subtypes (CD8+ T cells, T helpers cells, regulatory T cells and γδ T lymphocytes) in ovarian cancer progression showed opposite roles [
38]. So, we decided to work with immunodeficient mice allowing the utilization of human tumor cells compared to immunocompetent mice requiring murine cells.
Our results are worrisome because of MCI increase in anterior and lateral peritoneum of HA-CMC barrier treated mice. However, the clinical benefit of HA-CMC barrier utilization is to our mind clearly established. First, ovarian cancer bearing patients need to undergo multiple surgeries that lead to the development of adherences which constitute a complication risk factor. Secondly, intraperitoneal chemotherapy delivery cannot be performed because of adherences which induce heterogeneity of chemotherapy diffusion [
3]. Because of benefit/risk balance, we decided to use this biomaterial for our patients keeping in mind a clinical prospective study remains necessary in order to clearly establish the harmlessness of HA-CMC barrier.
Competing interests
The authors declare that they have no competing interest.
Authors’ contributions
LM and BT carried out the research and prepared the original manuscript. EM carried out the immunohistochemistry experiments. MO was in charge of the statistical analysis. AM and GF carried out the mouse observation for murine carcinosis score obtention. BT, JPD, BC and GF supervised the research and directed the project. All authors approved the final manuscript.