Abstract
Background: Randomized studies have shown that bevacizumab combined with taxane-based regimens increases response rates and prolongs progression-free survival (PFS) of patients with metastatic breast cancer (MBC). However predictive or prognostic biological markers that identify the appropriate target population, thus improving the cost-effectiveness ratio of this treatment, are still needed. Patients and Methods: Retrospectively, 124 patients with MBC treated either with paclitaxel 90 mg/m2 weekly x12 plus bevacizumab 10 μg/kg every 2 weeks or 15 μg/kg every 3 weeks (85 patients) or paclitaxel 175 mg/m2 plus bevacizumab 15 μg/kg every 3 weeks for 6 cycles (36 patients) were identified. Additionally, the prognostic significance of a panel of key biological markers was evaluated centrally by immunohistochemistry (IHC) in 88 evaluable patients. Results: More than two thirds of the patients completed chemotherapy, as planned. The response rate was almost identical (55.3% vs. 55.6%) in the patients treated with weekly or 3-weekly paclitaxel, respectively. After a median follow-up time of 23 months, the median PFS of the study population was 13 months, while median survival had not yet been reached. Common severe adverse events were neutropenia (33%), neuropathy (18.6%) and metabolic disturbances (17.6%). The incidence of hypertension of all grades was 28.1%. High expression of vascular endothelial growth factor (VEGF) receptor 3 (VEGFR3) was associated with clinical response, while high expression of VEGFR1 was associated with poor survival. Conclusion: The safety and activity of the combination of bevacizumab with paclitaxel given either weekly or 3-weekly in patients with MBC is confirmed.
Metastatic breast cancer (MBC) is an incurable disease. Systemic treatments i.e. hormonal therapy or chemotherapy have mainly a palliative role, even though recently published meta-analyses suggest that newer chemotherapeutic agents, such as taxanes or trastuzumab may contribute to a significant prolongation of survival (1, 2). Furthermore, two large randomized studies clearly showed that weekly paclitaxel should probably be one of the preferred regimens, since it is cost-effective in comparison with other commonly used chemotherapeutic regimens (3-5).
Angiogenesis is essential for the development and metastasis of breast cancer (6). Preclinical studies have strongly indicated that vascular endothelial growth factor A (VEGF-A) is a major mediator in breast cancer angiogenesis (7). Through binding to its receptors, VEGF receptor 1 (VEGFR1), and mainly through VEGFR2, among other actions, VEGF-A stimulates endothelial proliferation, migration and motility (8). Furthermore, many tumors including breast carcinoma produce VEGF-A and express VEGFRs suggesting a possible autonomous autocrine mechanism contributing to tumor growth and survival (9).
Anti-angiogenic therapy is a promising new therapeutic strategy in human epidermal growth factor receptor 2 (HER2)-negative breast cancer (10). Bevacizumab, a humanized anti-VEGF-A monoclonal antibody (11), in combination with either paclitaxel or docetaxel resulted in prolongation of progression-free survival (PFS) and increased overall response rate (ORR) in patients with MBC (12, 13). Furthermore, in vitro studies have demonstrated a synergistic effect from the administration of bevacizumab with paclitaxel (14). Following the publication of these results and the approval of bevacizumab in combination with a taxane in this group of patients, investigators within the Hellenic Cooperative Oncology Group (HeCOG) adopted this treatment in patients with HER2-negative MBC.
However, it is evident from the two randomized, phase III trials (12, 13) and daily clinical experience that the majority of patients do not respond to the combined treatment and, eventually, tumor progression is expected to occur in all patients. Probably any targeted treatment is cost-effective only when the anti-target agent is administered preferentially to those patients who actually benefit from the treatment and not indiscriminately to all patients, who are, thus, unnecessarily subjected to its potentially hazardous effects.
Therefore, there is an imperative need to identify biological markers that will select the patients most likely to respond to the targeted treatment.
In the present retrospective clinical-translational research study, firstly, the clinical activity and safety of the combination of paclitaxel and bevacizumab in patients with MBC was evaluated and secondly, the potential prognostic value of a number of biological markers, mostly related to tumor angiogenesis, the target of bevacizumab, or cell division, the target of taxanes was explored.
Patients and Methods
The database of the HeCOG was searched retrospectively for patients with MBC, treated with paclitaxel and bevacizumab, as first line combination chemotherapy, between August 2007 and August 2009. Eligibility criteria for the present study included histologically confirmed MBC, performance status (PS) of 0-2 of the Eastern Cooperative Oncology Group scale and age ≥18 years. Previous hormonal therapy or radiation therapy was allowed. Previous adjuvant chemotherapy was allowed providing that the interval between completion of adjuvant chemotherapy and first relapse was >6 months. The acquisition of any type of biological material was not a prerequisite for inclusion in the study but this possibility was intensively explored. According to our established policy, each patient included in the present study had signed an informed consent form before her first treatment if she wished to provide biological material for research purposes. The research protocol was approved by the Bioethics Committee of Aristotle University of Thessaloniki School of Medicine.
In total, 124 patients fulfilling the eligibility criteria were identified in 13 certified HeCOG centers. Among the eligible patients, two received docetaxel (instead of weekly paclitaxel) with bevacizumab and one switched from weekly to the 3-weekly schedule of paclitaxel. These three patients were excluded from the analysis of response and toxicity, but were included in the translational analysis along with the group of patients treated with the weekly schedule.
Following the publication of the results from two large randomized trials showing the superiority of weekly paclitaxel in terms of survival of patients with MBC compared to 3-weekly paclitaxel or to other taxane-based combinations (3, 4), this schedule was widely adopted by members in our Group. Consequently, most of the patients (88 patients, 71%) received weekly paclitaxel 90 mg/m2 and bevacizumab 10 μg/kg every 2 weeks or 15 μg/kg every 3 weeks for at least 12 weeks. Nevertheless, a considerable number of patients (36 patients, 29%) had been treated with 3-weekly paclitaxel 175 mg/m2 and bevacizumab 15 μg/kg for 6 cycles. According to our Group's standard policy, the recommendation was that bevacizumab be continued following the completion of chemotherapy, as maintenance treatment, until disease progression, unacceptable toxicity or voluntary withdrawal, whatever occurred first. Patients with estrogen receptor (ER)-positive tumors received appropriate hormonal therapy post-chemotherapy. Patients with osseous metastases usually received treatment with bisphosphonates during the chemotherapy period.
Response evaluation and follow-up. Response was routinely assessed after the third and sixth cycle in those patients treated with the 3-weekly schedule and after the twelfth cycle in those treated with the weekly schedule. After the completion of chemotherapy, all the patients were followed, according to our Group's guidelines, every 3-4 months with physical examination, full blood count, complete biochemistry and imaging examinations, as indicated. Notably, central evaluation of imaging material pertinent to tumor response was not performed in this study. Toxicity was assessed according to the National Cancer Institute Common Terminology Criteria Version 3.0 (http://ctep.cancer.gov) by expert HeCOG personnel.
Biological material. Formalin-fixed paraffin-embedded (FFPE) tumor tissue blocks were retrieved from the HeCOG Tumor Tissue Repository for 97 patients. In total, 9 blocks were excluded from the translational analysis because the tumor tissue was inadequate for translational analysis.
FFPE tumor tissue from 88 patients was therefore tested for protein expression by IHC. In 67 cases, the tumor tissue in the paraffin block was adequate for the construction of tissue microarrays (TMAs) containing 2 cores of 1.5 mm in diameter per case. Each TMA block also included tissue cores of the same diameter in the first and the last column, which served as positive and negative controls for the tested antibodies. The remaining 21 cases not included in the TMAs were evaluated as whole tissue sections. In 76 cases the examined tissue was derived from the primary tumor, while in the remaining 12 cases it was from a metastatic site.
Immunohistochemistry. Serial 2.5 μm thick sections from the original blocks or the TMA blocks were cut at the Laboratory of Molecular Oncology of the Hellenic Foundation of Cancer Research, School of Medicine, Aristotle University of Thessaloniki, mounted on adhesive microscope slides and subjected to IHC labelling using Bond Max™ (Leica Microsystems, Germany/Menarini Diagnostics Athens, Hellas) and i6000 (Biogenex, San Ramon, CA) autostainers. The IHC protocols for ER, progesterone receptor (PgR), HER2, Ki67, microtubule-associated protein-Tau (MAPT), VEGF-A, VEGF-C, VEGFR1, VEGFR2, VEGFR3, class III β-tubulin (TUBB3) and γ-tubulin were performed as described in Table I.
Fluorescent in situ hybridization (FISH) analysis. The cases with equivocal expression of HER2 (2+ by IHC) were further analyzed by FISH using a ZytoLight® SPEC HER2/CEN17 Dual color probe kit (ZytoVision GmbH, Bremerhaven, Germany). Cases were considered amplified for HER2 when the respective gene probe/centromere probe ratio was ≥2.2 (15).
Interpretation of IHC. The evaluation of all IHC sections was performed by an experienced breast cancer pathologist (HPK), blinded as to the patients' clinical characteristics and survival data, according to previously proposed or established criteria. Briefly, for ER, PgR, VEGF-A, VEGF-C, VEGFR1, VEGFR2, VEGFR3, MAPT and TUBB3 the histological score (H-score) was calculated by the following method: H-score=(1 x percentage of weakly positive cells) + (2 x percentage of moderately strong positive cells) + (3 x percentage of strongly positive cells). The cases were considered positive for ER and PgR if the H-score was ≥1 (16). The median H-score values for VEGF-A (180), VEGF-C (50), VEGFR1 (100), VEGFR2 (35), VEGFR3 (100), MAPT (170) and TUBB3 (130) were used as cut-offs to assign low/high expression. The proliferative index (Ki67 index) was defined as low (0-13%) and high (≥14%) based on the percentage of stained/unstained nuclei from the entire tumor area (17). The expression of HER2 protein was scored according to recently proposed criteria (15). Due to the homogenous expression of γ-tubulin, only the intensity of cytoplasmic staining was evaluated as weak=1, moderate=2 or strong=3. Weak and moderate staining (1 and 2) were considered low expression, while strong staining (3) was considered high expression.
Primary antibodies, source and staining conditions used in the present study.
Given the limited literature regarding the pathways mediating the action of VEGFs in neoplastic cells of breast tumors, coexpression patterns between VEGF-A and VEGFR1, VEGF-A and VEGFR2 and finally VEGF-C and VEGFR3 were sought. In order to investigate possible coactivation of VEGF-A and VEGF-C downstream pathways, coexpression of VEGFR1 and/or VEGFR2 with VEGFR3 was also investigated.
Statistical analysis. The data on the patients, tumor characteristics and acute toxicity were obtained from the medical records. Summary statistics were provided for all the parameters. Categorical variables were summarized as frequencies and corresponding percentages and continuous variables as median and range. Associations of the assessed biomarkers with response to treatment, as well as the associations between them were evaluated using the Fisher's exact test. Survival was measured from the date of the initiation of treatment with bevacizumab and paclitaxel to the date of the patient's death or last contact. PFS was measured from the date of treatment initiation to documented disease progression, death without prior documented progression or last contact. Survival status was updated in June of 2010.
Univariate Cox regression analysis, Kaplan-Meier analysis and the log-rank test were used to demonstrate differences in survival and PFS between the two treatment schedules of paclitaxel (weekly vs. 3-weekly), as well as for examining the prognostic significance of the VEGF-family and microtubule-associated biomarkers. The coexpression patterns of the VEGFRs were also examined for prognostic value. For the multivariate Cox regression model, a backward selection procedure with a removal criterion of p>0.10 was used to select the predictors among the following: group (weekly vs. 3 weekly), age, menopausal status (pre vs. post), PS (0 vs. 1-2), visceral metastasis (no vs. yes), number of metastatic sites (1-2 vs. ≥3), adjuvant chemotherapy (no vs. yes), adjuvant hormonal therapy (no vs. yes), ER status (negative vs. positive), PgR status (negative vs. positive), Ki67 (low vs. high), all microtubule-associated markers (low vs. high) and all VEGF-family markers (low vs. high), as well as the predefined coexpression of VEGF-A and VEGFR1, VEGF-A and VEGFR2, VEGF-C and VEGFR3 (both positive vs. everything else) and finally the coexpression of VEGFR1 and/or VEGFR2 with VEGFR3 (both pathways positive vs. everything else). The reported p-values are two-sided, and p<0.05 was considered statistically significant. The results are presented in accordance with the REMARK criteria (18).
Patient and tumor characteristics.
Treatment characteristics (N=121).
Results
The patient and tumor characteristics of the 124 patients are presented in Table II. The median age was 62 years (range 32-84). The majority of the patients were postmenopausal, with hormone receptor (HR)-positive invasive ductal carcinomas, PS of 0-1 and visceral metastases. A history of adjuvant chemotherapy was recorded in 61 women (49%), 52 of whom had received an anthracycline. There was no significant difference in the major patient and tumor characteristics between the patients treated with weekly or 3-weekly paclitaxel or between the whole patient population and those with available tumor tissue.
Compliance to treatment was generally satisfactory in both schedules of paclitaxel administration, as depicted in Table III. Treatment characteristics, except for the number of cycles delivered, were calculated for 12 cycles, in the case of weekly and for 6 cycles, in the case of 3-weekly treatment. The majority of the patients received all cycles at full dose. The relative dose intensity of paclitaxel in the weekly and 3-weekly schedules was 0.87 and 0.98, respectively.
Four deaths occurred, all of them in the weekly schedule. Analytically, one patient died of febrile neutropenia and sepsis following the second infusion of paclitaxel, while a second patient died of acute respiratory infection after the seventh infusion. Two additional patients died of rapidly progressing tumor, following the first and eleventh paclitaxel infusion respectively. In general, 28 patients (22.6%) discontinued their treatment with paclitaxel. The reasons for treatment discontinuation were non-fatal toxicity in 13 patients (10.1%), death in 4 (3.2%), progression of the disease in 7 (5.6%), physician's decision in 2 and voluntary withdrawal in 2 patients (1.6%).
In total, 64 (52%) of the 124 patients continued bevacizumab after the completion of the treatment with paclitaxel, as maintenance treatment. The median duration of the bevacizumab administration was 22 weeks (range 1-71).
Response was evaluated in 109 (90%) of the 121 patients. Three of the 124 patients were excluded from the analysis of response and toxicity for reasons stated in the methods section. Out of the remaining 12 non-evaluable patients, one did not have measurable lesion at the time of the initiation of treatment, two died probably from treatment-related reasons and two from rapid tumor progression, as stated above; six demonstrated tumor progression before evaluation, while for one patient response data were missing. The ORR was almost identical in the patients treated with weekly or 3-weekly paclitaxel (55.3% vs. 55.6%). Six patients (7.1%) in the weekly and 2 (5.6%) in the 3-weekly schedule achieved complete response. Furthermore, stabilization of the disease was noticed in 22 (25.9%) and 12 (33.3%) of the patients treated with the weekly and 3-weekly schedules, respectively.
Acute side-effects, as recorded in the medical records, according to the schedule of paclitaxel administration are shown in Table IV. Severe toxic effects were infrequently seen. However, two grade 5 events occurred in the weekly treatment group, as previously stated. The rates of severe neutropenia (p=0.047), and metabolic disturbances (p=0.005) were higher with weekly paclitaxel.
In total, 44 patients (35%) received second line chemotherapy, 11 of them with a single agent and the rest with two- or three-drug combinations. The most frequently used drugs were vinorelbine, capecitabine, gemcitabine, carboplatin and anthracyclines.
Incidence of toxicities according to paclitaxel schedule (N=121).
After a median follow-up of 23 months (range, 4-32), 73 patients (59%) demonstrated disease progression and 37 (30%) had died. The median PFS was 13 months (95% CI 9-16), with a 2-year PFS rate of 30% (95% CI 20%-39%). The median survival has not been reached yet, while the 2-year survival rate was 72% (95% CI 63%-81%) (Figure 1). Interestingly, an unplanned analysis revealed that median PFS was significantly higher with the 3-weekly (20 months, 95% CI 13-28) compared with the weekly paclitaxel (10 months, 95% CI 7.5-13, p=0.040) (Figure 1). However, significantly more patients treated with the 3-weekly paclitaxel regimen continued bevacizumab, as maintenance treatment, than with weekly paclitaxel (40 [45%] in the weekly and 24 [67%] in the 3-weekly, p=0.047), which might explain the difference in PFS. No differences were observed in the survival of patients between treatment groups (p=0.14).
The expression of VEGFs, VEGFRs and microtubule-associated proteins was predominantly cytoplasmic. High expression was observed in 44 out of 82 tumors (53.7%) for VEGF-A and VEGF-C, 60 out of 82 (73.2%) for VEGFR1, 41 out of 82 (50%) for VEGFR2, 45 out of 82 (54.9%) for VEGFR3, 42 out of 82 (51.2%) for MAPT and 48 out of 85 (56.5%) for TUBB3. High expression was also observed in 34 out of 82 tumors (41.5%) for γ-tubulin, 69 out of 83 (83.1%) for ER, 53 out of 84 (63.1%) for PgR and 68 out of 83 (81.9%) for Ki67. One out of 85 tumors (1.2%) was HER2 positive. Nuclear expression in a small number of neoplastic cells was also noted for VEGFR1 and VEGFR2 in 38 out of 82 (46.3%) and 56 out of 82 (68.3%) of the cases, respectively. For all the VEGFs and VEGFRs, positivity was also observed in endothelial cells, stromal cells and lymphocytes/plasma cells in the tumor microenvironment.
Examining the association between VEGF ligands and receptors, high expression of VEGF-A was associated with high expression of VEGFR1 (p=0.001) and VEGFR2 (p=0.014) in the neoplastic cells, while VEGF-C was not associated with VEGFR3 (p=0.12). None of the VEGF-family markers were associated with ER or PgR protein status, while only VEGF-C was marginally associated with Ki67 (p=0.042). More specifically, high expression of VEGF-C was more frequent in the high (59%) than in low (27%) Ki67 expressing tumors. High expression of MAPT was more frequently seen in the ER-positive tumors (59% in ER-positive vs. 21% in ER-negative, p=0.017) or PgR-positive tumors (67% in PgR-positive vs. 27% in PgR-negative, p=0.001), while no association was observed with Ki67 (p=0.57). No statistical significance was observed between the other microtubule-associated biomarkers and hormone receptors or Ki67.
Multivariate Cox regression analysis.
Among the examined markers only VEGFR3 was significantly associated with response to treatment, with the high expressing tumors being more frequent in responders (67% in responders vs. 42% in non-responders, p=0.038). Concerning the prognostic significance of markers, only VEGFR1 was significantly associated with survival, but not with PFS. More specifically, 23 deaths (38%) occurred in the 60 high expressing tumors versus 3 (14%) in the 22 low expressing VEGFR1 tumors. Univariate Cox regression revealed an increased risk of death in the high VEGFR1 expressing tumors (hazard ratio (HR): 3.98, 95% CI: 1.19-13.32, Wald's p=0.025) in comparison to tumors with low VEGFR1 expression. None of the remaining VEGF-family members or microtubule-associated biological markers was significantly associated with either survival or PFS.
Examining the coexpression of VEGF ligands and their corresponding receptors, tumors highly expressing both VEGF-A and VEGFR1 (VEGF-A+/VEGFR1+) or VEGF-A and VEGFR2 (VEGF-A+/VEGFR2+) were associated with shorter survival (HR=2.56, 95% CI: 1.14-5.74, Wald's p=0.022 for VEGF-A+/VEGFR1+ and HR=2.36, 95% CI: 1.07-5.21, Wald's p=0.033 for VEGF-A+/VEGFR2+ coexpression), while no association was found in terms of PFS (HR=1.17, 95% CI: 0.68-2.01, Wald's p=0.57 for VEGF-A+/VEGFR1+ and HR=1.73, 95% CI: 0.99-3.03, Wald's p=0.054 for VEGF-A+/VEGFR2+). These coexpressions were also not associated with response to treatment (p=0.50 and p=0.81 respectively). Finally, high expression of either VEGFR1 or VEGFR2 in combination with high expression of VEGFR3 was associated with response (58% in responders vs. 32% in non-responders, p=0.037). Among 38 cases highly expressing both “receptor-pathways”, 17 patients (45%) died in comparison to 9 deaths (20%) among 44 cases highly expressing none or only one pathway. Univariate Cox regression revealed an increased risk of death for tumors with high expression of VEGFR1 or VEGFR2 in addition to VEGFR3 (HR=2.97, 95% CI: 1.32-6.70, Wald's p=0.009). No such association was found in terms of PFS (HR=1.50, 95% CI: 0.88-2.58, Wald's p=0.14).
Multivariate Cox regression analysis (Table V), revealed that the number of metastatic sites, PS and adjuvant hormonal therapy remained independent predictors for both PFS and survival, while among all the examined biomarkers only VEGFR1 independently predicted survival. Of note, treatment schedule remained a significant prognostic factor for PFS, as well. Regarding PFS, 3-weekly paclitaxel was associated with reduced risk for relapse (HR=0.56, 95% CI: 0.32-0.96, Wald's p=0.037), while risk for relapse was increased in patients with higher PS (HR=1.82, 95% CI: 1.10-3.02, Wald's p=0.020), history of adjuvant hormonal treatment (HR=1.28, 95% CI: 1.28-3.21, Wald's p=0.003) and presence of three or more metastatic sites (HR=2.07, 95% CI: 1.28-3.56, Wald's p=0.003). Regarding survival, high expression of VEGFR1 was associated with increased risk of death (HR=7.59, 95% CI: 1.97-29.25, Wald's p=0.003). Increased risk for death was also seen in patients with higher PS (HR=3.55, 95% CI: 1.27-9.95, Wald's p=0.016), history of adjuvant hormonal treatment (HR=2.64, 95% CI: 1.04-6.70, Wald's p=0.040) and presence of three or more metastatic sites (HR=5.25, 95% CI: 2.10-13.13, Wald's p<0.001).
Discussion
The combination of weekly or 3-weekly paclitaxel and bevacizumab gave an ORR of 55% in this study, which was remarkably similar to the 49% of patients with measurable lesions in the pivotal E2100 trial (12). Likewise, the median PFS in the present patients was 13 months (11.8 months in the E2100 trial), while 1-year survival was 85% (81% in the E2100 trial).
Toxicity in the present study was generally manageable. Severe neutropenia was significantly more frequent with the weekly than the 3-weekly paclitaxel regimen, at a rate similar to that of the ATHENA trial (19), in which the majority of the patients were treated with bevacizumab and taxane-based regimens and severe neutropenia and hypertension were noticed in 5.4% and 4% of the 2251 patients, respectively. Among the present patients hypertension was frequently observed, although it was usually controlled with appropriate treatment. Finally, the rates of neuropathy, infection and fatigue were very close to those reported in the E2100 trial.
Angiogenesis is of seminal importance for both tumor growth and the development of metastases. One of the most important angiogenic factors is VEGF-A, which functions mainly in a paracrine manner, through binding to VEGFR2 expressed on endothelial cells. This binding leads to signaling in several downstream pathways, regulating endothelial cell proliferation and apoptosis, and promoting angiogenesis. VEGF-A also binds with high affinity to VEGFR1, which has however weak signal transducing properties, and its function remains unclear (9). VEGF-C mediates lymphangiogenesis through binding to VEGFR3 and angiogenesis through binding to VEGFR2. Breast carcinoma cells have been shown by IHC to express VEGF-A (20-32), VEGF-C (22, 31, 33, 34), VEGFR1 (21, 23, 32, 35-37), VEGFR2 (21-23, 28, 29, 35, 36, 38) and VEGFR3 (22, 33) but their roles in neoplastic cells are still elusive. These factors were relatively homogeneously expressed in the present study, not only in the tumor cells, but also in cells of the tumor microenvironment. The clone 55B11 for VEGFR2, used in the present study, has been found to be specific, but the clone KLT9 for VEGFR3, used herein, was previously characterized as non-specific (39). Petrova and coworkers also addressed the non-specificity of VEGFR3 antibodies, however they did not include clone KLT9 among those examined in their study (40). Nonetheless, the issue of specificity of anti-VEGFR antibodies merits further investigation, particularly in light of the potential therapeutic implications.
In this study, VEGF-A protein overexpression did not reach statistical significance with regard to prognostic value. Existing data regarding the prognostic significance of VEGF-A, when studied by IHC, are controversial. No prognostic significance was found in some studies (24, 26, 27, 30, 32, 41), although others observed independent prognostic value for relapse-free survival (42), or association with decreased overall survival in univariate analysis (31).
VEGFR1 was associated with worse prognosis in some studies (32, 35, 36). Of note, VEGFR1 is also among the genes of poor prognosis in the 70-gene prognosis signature (43). It appears that, unlike its role as a decoy receptor in endothelial cells, VEGFR1 may have a more active role in neoplastic cells. Activation of VEGFR1 results in tumor growth and survival through activation of the mitogen-activated protein kinase (MAPK) and phosphatidylinositol-3 kinase (PI3K)/AKT pathways (44) and VEGFR1, but not VEGFR2, participates in autocrine survival signals of neoplastic cells through activation of the PI3K/Akt pathway (37). In the present study, high protein expression of VEGFR1, as assessed by IHC, was associated with shorter survival, in both univariate and multivariate analyses, in agreement with the existing literature.
In the analysis of factor coexpression, an association of VEGF-A with VEGFR1 was observed, unlike a previous report (32), and an association of VEGF-A and VEGFR2, as has previously been described (21, 28). The association of VEGF-A with VEGFR1 and VEGFR2 in neoplastic cells is suggestive of an autocrine regulatory mechanism, supporting cancer cell autonomy, as has previously been suggested (9). The expression of VEGFR1 and VEGFR2 in the cytoplasm, rather than on the cell membrane, favors an “intracrine” mechanism, which has been previously described as a tumor cell survival system (37). Indeed, VEGF-A, VEGFR1 and VEGFR2 are not related to the proliferation marker Ki67 however, either alone, or in various high-expression combinations, they seem to be associated with survival. These findings suggest that the effects of VEGF-A/VEGFR1/ VEGFR2 may rather be mediated through apoptotic control mechanisms, a hypothesis that needs to be explored further.
VEGF-C has previously been associated with survival and PFS (31). In the present study, VEGF-C was not associated with survival, in agreement with other reports (45, 46). However, VEGF-C was associated with the proliferation index Ki67, thus a possible mechanism of action may be through control of the proliferation of neoplastic cells.
In the present study, VEGFR3 was not found to be associated with VEGF-C, unlike in a previous study (22). Although VEGFR3 alone does not seem to have prognostic value, high expression of VEGFR3 was found to be associated with response to treatment indicating possible predictive value. Furthermore, high expression of VEGFR3 in combination with high expression of either VEGFR1 or VEGFR2 was found to be associated with poor survival, suggesting that activation of two of the VEGFR pathways may be needed for tumor cell proliferation and survival and that blockade of more than one receptor might be necessary to more efficiently compromise the neoplastic cell. In addition, these findings might provide further support to the hypothesis of complementary functions of the angiogenesis pathways in the neoplastic cell and underline the importance of the receptors, rather than the growth factors, an observation with potential important therapeutic implications.
The microtubule-associated proteins MAPT, TUBB3 and γ-tubulin were not associated with clinicopathological parameters or found to have prognostic value in this study, unlike previous findings in breast carcinoma (47, 48).
In summary, the safety and similar activity of the combination of bevacizumab with weekly or 3-weekly paclitaxel is confirmed. The angiogenesis factors VEGFR1 and VEGFR3 appear to play an important role in tumor growth through “intracrine/autocrine” regulatory mechanisms. A prognostic value of VEGFR1 protein expression was also demonstrated in both univariate and multivariate analyses, while VEGFR3 was found to be associated with response to treatment. For the first time, the prognostic significance of the coexpression of VEGFRs, particularly VEGFR1 and VEGFR3, in the neoplastic cells was demonstrated. If supported by larger studies, this may have potential implications in the evaluation of MBC patients for different antiangiogenic therapies and appears to suggest that targeting combinations of VEGFRs, rather than a single VEGFR, might have more optimal therapeutic results. The identification and verification of predictive factors for bevacizumab treatment in MBC patients is of paramount importance, especially since the Food and Drug Administration, but not the European Medicines Agency announcement in December of 2010, that it will “take the first steps to remove the indication for bevacizumab in metastatic breast cancer”.
A: PFS (green line) and survival (brown line) of the whole study population. B: PFS of patients treated with bevacizumab and weekly paclitaxel (blue line) or 3-weekly paclitaxel (red line). C: Survival of patients treated with bevacizumab and weekly paclitaxel (blue line) or 3-weekly paclitaxel (red line).
Protein expression detected by IHC in tissue microarrays from breast carcinoma cases. VEGF-A (a); VEGF-C (b); VEGFR1 (c); VEGFR2 (d); VEGFR3 (e); TUBB3 (f); γ-tubulin (g); MAPT (h). Original magnification ×40.
Acknowledgements
The Authors wish to thank Ms. Evita Fragou and Ms. Dimitra Katsala for monitoring the study, Ms. Maria Moschoni for coordinating the data management, Ms. Thalia Spinari for tissue sample collection, Ms. Genovefa Polychronidou MD for molecular data management and Ms. Sophia Chrisafi Ph.D. for her excellent technical assistance with the IHC and FISH assays. On behalf of the Hellenic Foundation for Cancer Research, Athens, Greece, the principal investigator has pending patent applications with Siemens Healthcare Diagnostics, Tarrytown, NY, USA.
The translational research was supported by a HeCOG research grant: HE TRANS_07. GF received Commercial Research Funding by Roche Hellas SA and Genesis Pharma SA, Athens, Greece.
- Received April 12, 2011.
- Revision received June 29, 2011.
- Accepted June 30, 2011.
- Copyright© 2011 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- The Prognostic Value of the Immunohistochemical Expression of Phosphorylated RB and p16 Proteins in Association with Cyclin D1 and the p53 Pathway in a Large Cohort of Patients with Breast Cancer Treated with Taxane-based Adjuvant Chemotherapy
- Immunohistochemical Study of the Epithelial-Mesenchymal Transition Phenotype in Cancer of Unknown Primary: Incidence, Correlations and Prognostic Utility