An exploratory biomarker study in metastatic tumors from colorectal cancer patients treated with bevacizumab

Abstract

Aims

Inhibition of angiogenesis is an effective treatment option for metastatic colorectal cancer. Predictive biomarkers to select patients who are most likely to benefit from this therapeutic strategy are lacking. We conducted a pilot, retrospective biomarker study in a cohort of metastatic colorectal cancer patients treated with bevacizumab. The objectives of this study were to evaluate the prognostic value of biomarker expression in metastases and to compare their expression in paired tumor specimens.

Materials and methods

Eligible patients were treated with a bevacizumab-containing therapy; from these patients, tumor tissue from metastases was available. PTEN, PI3K p110a, c-MET, and CAIX were analyzed by immunohistochemistry.

Results

Forty-two patients received bevacizumab, 13 (31%) with first-line and 29 (69%) with second-line chemotherapy. Expression of CAIX, PI3K p110a, and c-MET in metastases did not predict objective response. PTEN loss was associated with response to treatment (p=0.02) and this association remained significant after adjusting for prognostic variables (p=0.006). However, no association with survival outcomes was found. In 32 patients (76%) with available paired specimens, we observed an equal expression between primary tumors and corresponding metastases in 75% of cases for CAIX in epithelial tumor cells, 56% for CAIX in stromal cells, 63% for PTEN, and 87% for c-MET.

Conclusion

PTEN loss in metastases appears to be associated with response to bevacizumab-based therapy. However, larger studies are necessary to confirm the potential role of the PI3K/AKT/mTOR pathway in modulating the therapeutic effect of bevacizumab. Tumor heterogeneity should be taken into consideration when analyzing tumor tissues for biomarker studies.

Keywords

Bevacizumab Biomarkers Colorectal cancer Metastases PTEN

Introduction

Angiogenesis has been demonstrated to play a crucial role in cancer growth and progression (1–2–3), and targeting angiogenesis has become an established therapeutic option in several malignancies. Bevacizumab (Avastin; Genentech Inc, South San Francisco, CA, US) is a recombinant, humanized monoclonal antibody that exerts anti-angiogenic activity by binding to and inhibiting the vascular endothelial growth factor (VEGF), a key mediator in tumor angiogenesis (4, 5). In metastatic colorectal cancer (mCRC), adding bevacizumab to standard first and second-line chemotherapy resulted in a significant improvement in response rates (RR), progression-free survival (PFS), and overall survival (OS) (6–7–8–9–10–11).

Studies investigating predictive factors of responsiveness or resistance to bevacizumab have provided inconsistent results and have largely failed (12–13–14). This failure is likely due to the complexity of angiogenesis and to the intricate network of intracellular signaling pathways underlying tumor response to hypoxia. Indeed, although VEGF plays a key role in promoting tumor cell survival, growth, and invasiveness, several cellular types, growth factors, and intracellular mediators are involved in tumor angiogenesis (5, 15). Moreover, several concerns on biomarker studies still remain, including the lack of validation methodology for biomarker analysis and the use of old samples from primary tumors that are no longer representative of the actual tumor biology (16).

The VHL/HIF-1 axis is not the only mediator of the hypoxia-induced angiogenic response (17, 18). The PI3K/AKT/mTOR pathway via RHO/RHO kinase and the transcriptional factor c-Myc was shown to activate HIF-1-independent, angiogenic pathways (19). Moreover, preclinical data showed that inhibition of PI3K resulted in downregulation of VEGF (19).

c-MET is a receptor tyrosine kinase with high binding affinity for the hepatocyte growth factor (HGF). In human cancers, c-MET and HGF are often deregulated and associated with poor prognosis (20, 21). Even more interestingly, c-MET was shown to promote tumor invasiveness and metastasis (22–23–24) especially following drug-induced hypoxic conditions (25).

Carbonic anhydrase IX (CAIX) is an enzyme involved in the regulation of intracellular and extracellular pH (26). CAIX expression has been shown to correlate with poor outcome in several malignancies. CAIX is upregulated in CRC (27–28–29–30–31) and this upregulation may represent a marker of tumor metabolic adaption during hypoxia and may favor tumor proliferation in acidic conditions.

Based on these data, we performed a pilot, retrospective biomarker study investigating both the primary and metastatic tumor expression and prognostic value of PTEN, PI3K subunit p110a, c-MET, and CAIX in a small cohort of mCRC patients treated with a bevacizumab-containing therapy.

Materials and Methods

This was a multicenter, retrospective study conducted in 2 Italian centers: Humanitas Cancer Center/Humanitas Research Hospital (Rozzano, Milan), and Humanitas Gavazzeni (Bergamo). The study was approved by the institutional review boards.

Patients were eligible for this study if they had: (i) histologically confirmed CRC; (ii) received a bevacizumab-containing therapy for the treatment of metastatic disease; (iii) regular, on-treatment radiological assessment for evaluation of tumor response; (iv) archived tumor tissue from metastatic sites available for biomarker analysis.

Patient medical records were used to collect baseline clinicopathological characteristics including gender, age, site of primary tumor, disease-free survival (DFS), number and sites of metastases, levels of carcinoembrionic antigen (CEA), carbohydrate antigen (CA) 19-9, alkaline phosphatase (ALP), and lactate dehydrogenase (LDH), as well as chemotherapy regimen used in combination with bevacizumab.

Tumor assessment was performed with CT scan every 2-3 months as per clinical practice in the institutions participating in the study. PET or bone scans were not routinely performed and were requested only when clinically indicated. For the purpose of this study, baseline and on-treatment scans were systematically reviewed by 2 radiologists who were blinded to the results of the biomarker analysis. Tumor response was reported according to the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1.

Biomarker analysis

The expression of PTEN, PI3K subunit p110a, c-MET, and CAIX was assessed by immunohistochemistry (IHC) on 4 consecutive 2-µm thick tissue sections of formalin-fixed, paraffin-embedded (FFPE) specimens. Sections were immunostained with biomarker-specific antibodies as follows: PTEN (Cell Signalling 9559; dilution 1:100), PI3K subunit p110a (Cell Signalling 4249; dilution 1:200), CAIX (Abcam AB15086; dilution 1:2,000), and c-MET (Assay Designs ADI-905-076; dilution 1:100). Slides were then incubated with the secondary antibody using the MACH1 Universal HRP-Polymer Detection kit (Biocare Medical). Staining was performed with 3,3′ diaminobenzidine (DAB) as a chromogen and sections were then counterstained with hematoxylin. For each batch, positive and negative control slides without the primary antibody were also included.

Biomarker analysis was performed on archived tissue from metastatic sites and, when available, on tissue from the correspondent primary tumors. For patients with archived tissues from 2 or more metastatic sites, all tumor specimens were analyzed. Nevertheless, only the results from the biomarker analysis conducted on tumor tissue obtained before the start of the bevacizumab-based treatment were used for correlation with the clinical data. All the analyses were performed by pathologists who were blinded to baseline patient characteristics and clinical outcomes.

Immunohistochemistry

For each antibody, immunoreactivity was expressed as membrane (for CAIX), membrane/cytoplasmic (for c-MET and p110a), and cytoplasmic (for PTEN) staining, and its intensity was scored according to a four-tier system: 0, no staining; 1, weak; 2, moderate; 3, strong. The percentage of immunoreactive tumor cells was recorded for each tissue specimen. PTEN, PI3K p110a, and c-MET expression was analyzed only in epithelial tumor cells, while the expression of CAIX was assessed in both epithelial tumor cells and tumor stromal cells (myofibroblasts).

CAIX. Epithelial and stromal cells were scored separately. Specimens were defined as positive if staining occurred in >5% of the tumor epithelial or stromal cells (32). PTEN. Specimens were defined as positive if PTEN expression in the cytoplasm was confirmed irrespective of the intensity of staining and percentage of immunoreactive cells (i.e., any positive tumor cells) (33). PI3K p110a. The evaluation of p110a expression was assed as a continuous trait by multiplying the percentage of stained tumor cells (0-100%) by the staining intensity (0-3) and adding this to a score between 0 and 300. c-MET. Specimens were defined as positive if moderate (+2) or strong (+3) c-MET expression was found in ≥50% of tumor cells (34).

Statistical analysis

The objectives of this retrospective, exploratory study were: to assess the value of PTEN, PI3K p110a, c-MET, and CAIX expressed in tumors from metastatic sites, in predicting the outcome of mCRC patients treated with a bevacizumab-containing therapy, as well as to evaluate the concordance of biomarker expression between primary tumors and metastases.

PFS was calculated as the time from treatment start to progressive disease or death from any cause. OS was calculated as the time from treatment start to death from any cause. Patients who were still alive or alive with no evidence of progressive disease when this analysis was conducted were censored respectively at the time of the last contact or last radiological assessment.

Data are summarized as frequencies and proportions or as medians and ranges, as appropriate. Logistic regression was carried out to determine the association between biomarker expression and therapy responsiveness, calculating the adjusted OR with 95% confidence intervals (95% CI). The Cox proportional hazards models was applied to evaluate the hazard ratio (HR) with corresponding 95% CI, controlling for the effect of possible confounders. The McNemar's test was used to compare the expression between primitive and metastatic tumor sites. A p value less than 0.05 was considered to be the limit of statistical significance. All statistical analyses were performed using the R package.

Results

From August 2004 to March 2011, 42 patients who met the study eligibility criteria were identified in the 2 institutions. Baseline patient characteristics are reported in Table I. Patients included 28 males (66.7%) and 14 females (32.3%), and their median age was 64.4 years (range 42.4-80.0 years). In 32 patients (76.2%) the primary tumor was localized in the colon and in 10 patients (23.8%) in the rectum. The majority of patients (64.3%) presented with metastatic disease at diagnosis. In 13 patients (31%) bevacizumab was given with first-line chemotherapy, while 29 patients (69%) received bevacizumab in association with second-line chemotherapy. In the majority of cases (81%), irinotecan was the companion chemotherapy agent. The median number of cycles of bevacizumab administered with first-line and second-line chemotherapy was 12 (range 3-29) and 10 (range 3-20) respectively.

TABLE I

Baseline Patient Ch Aracteristics

	N (42)	%
Gender
Male	28	66.7
Female	14	33.3
Age
Median years (range)		62.1 (41-77)
Primary tumor
Colon	32	76.2
Rectum	10	23.8
Treatment line
First	13	31
Second	29	69
Type of chemotherapy
Combination chemotherapy	35	83.3
Single agent therapy	7	16.7
Chemotherapy backbone
Irinotecan	34	81
Oxaliplatin	6	14.3
5-FU alone	2	4.7
Number of metastases
1	13	31.0
≥2	29	69.0
Site of metastases
Liver only	7	16.7
Other	35	83.3
CEA at baseline
Normal	13	31.0
Elevated	27	64.3
Unknown	2	4.7
CA 19-9 at baseline
Normal	21	50
Elevated	19	45.3
Unknown	2	4.7

A favorable response to treatment (complete response or partial response) was observed in 76.9% of patients treated in the first-line setting and in 31% of patients treated in the second-line setting. Median PFS was 13.3 months in patients treated in the first-line setting and 8.8 months in patients treated in the second-line setting. The median follow-up was 32 months (range 2.3-49 months). At the time of this analysis, 14 patients (33.3%) were still alive while 28 had died (66.7%). The median OS was 26.4 months for patients receiving bevacizumab with first-line chemotherapy and 17.3 months for patients treated with second-line therapy. Among the baseline clinical factors, number of metastases (1 vs ≥2) and line of treatment (first vs second) were predictive of tumor response, while CA 19-9 level (normal vs high), line of treatment (first vs second), chemotherapy regimen (single agent vs combination therapy), and chemotherapy companion drug (irinotecan vs oxaliplatin) were predictive of PFS and OS.

Biomarker expression levels in tumor samples from metastatic sites are reported in Table II: CAIX expression in epithelial tumor or stromal cells, as well as PI3K p110a and c-MET expression in epithelial tumor cells were not found to predict tumor response. By contrast, PTEN loss was associated with increased response to treatment. In patients with tumors exhibiting PTEN loss the response rate was 71% compared to 32% in patients with PTEN-expressing tumors (p=0.02) (Tab. III). In a multivariate analysis, the association between PTEN loss and tumor response remained significant after adjusting for number of tumor metastases and line of treatment (p=0.006) (Tab. IV). When the survival outcomes were analyzed, none of the biomarkers considered was found to predict PFS or OS in a univariate analysis.

TABLE II

Biomarker Expression And Concordance Between Primary And Metastatic Tumors

	Primary tumors		Metastases		Concordance (%)	p value
	N (30)	%	N (42)	%
PTEN	18/32	56.3	28/42	66.7	62.5	0.248
PI3K p110α	18/32	56.3	24/42	75	67.7	0.527
CAIX (epithelial cells)	20/32	62.5	30/42	71.4	75	0.48
CAIX (stromal cells)	14/32	43.8	21/42	50	56.3	0.109
c-MET	3/32	9.4	1-42	2.4	87.1	0.317

TABLE III

Patients’ Characteristics And Outcome According To Pten Status

	PTEN -		PTEN +		p value
	N = 14	100%	N = 28	100%
Gender					0.358
Male	8	57.1	20	71.4
Female	6	42.9	8	28.6
Age					0.870
Median Years (range)	64.4 (47.7 - 75.6)		64.8 (42.4 - 80.1)
Primary tumor					0.610
Colon	10	71.4	22	78.6
Rectum	4	28.6	6	21.4
Treatment line					0.350
First	3	21.4	10	35.7
Second	11	78.6	18	64.3
Chemotherapy					0.561
Combination	3	21.4	4	14.3
Single agent	11	78.6	24	85.7
Chemo backbone					0.798
Irinotecan	9	64.3	25	89.3
Oxaliplatin	3	21.4	3	10.7
5-FU alone	2	14.3	0	0
N of metastases					0.638
1	5	35.7	8	28.6
≥2	9	64.3	20	71.4
Site of metastases					0.561
Liver only	3	21.4	4	14.3
Other	11	78.6	24	85.7
Baseline CEA					0.309
Normal	6	42.9	7	25
Elevated	8	57.1	19	67.9
Unknown	0	0	2	7.1
Baseline CA 19-9					0.429
Normal	8	57.1	13	46.4
Elevated	5	35.7	14	50
Unknown	1	7.2	1	3.6
Response rate	10	71.4	9	32.1	0.020
Median PFS (mos)	9.2		8.7		0.968
Median OS (mos)	21.1		17.3		0.628

TABLE IV

Association Between Biomarker Expression And Tumor Response

Biomarker	OR (95% CI)	Adjusted¹ OR (95% CI)	p value	Adjusted¹ p value
PTEN	0.19 (0.05-0.77)	0.07 (0.01-0.45)	0.02	0.006
PI3K p110a	–	–	0.654	0.519
CAIX (epithelial cells)	–	–	0.769	0.392
CAIX (stromal cells)	–	–	0.354	0.407
c-MET²	–	–	–	–

Adjusted for treatment line and number of metastatic sites.

Odds ratios and p values for the association between c-MET expression and tumor response are not presented due to the small number of c-MET-positive tumors in the study population.

Thirty-two out of 42 patients (76.2%) had also tissue specimens from the primary tumor available for analysis. The expression of biomarkers in paired specimens is reported in Table III. Concordance between primary tumor and corresponding metastasis was observed in 75% of the cases for CAIX in epithelial tumor cells, 56% for CAIX in stromal cells, 62% for PTEN, and 87% for c-MET. When the expression of biomarkers in the primary tumor was correlated with patients’ response (PFS and OS), the PI3K p110a expression was continuously associated with improved PFS. However, this association did not remain significant after adjusting for other prognostic variables including CA 19-9 level, line of treatment, chemotherapy regimen, and chemotherapy companion drug.

For 3 patients, tissue from metastases taken before and after the administration of bevacizumab was available for analysis. In 2 cases a change in the expression of PTEN was observed (after treatment 1 tumor became PTEN-positive and 1 became PTEN-negative). The expression of CAIX in both epithelial tumor and stromal cells, as well as that of PI3K p110a in epithelial tumor cells changed in 1 case. No changes in the expression of c-MET were observed before and after bevacizumab treatment.

Discussion

In this study we report the results of an exploratory, retrospective biomarker analysis conducted in a small cohort of mCRC patients treated with bevacizumab. The biomarkers here analyzed were selected based on preclinical evidence and preliminary clinical data suggesting their potential involvement in the mechanisms of angiogenesis and hypoxia-induced response.

By restricting our analysis to patients with available tumor tissue from metastatic lesions, we assessed the role of these biomarkers in tissue samples that were likely to be representative of the tumor biology at the time of the bevacizumab-based therapy. Tumor heterogeneity is an established phenomenon that results in significant differences between primary tumors and metastases as well as between different metastases originated from the same tumor (35, 36). Tumor heterogeneity can influence the overall tumor response to treatment and can also negatively affect the ability of a specific biomarker to truly predict the effect of the treatment. Interestingly, when we compared metastases and corresponding primary tumors, the discordance rate for biomarker expression was relatively high, ranging from 13% to 44%. Unfortunately, only for 3 patients metastatic tumor samples before and after the use of bevacizumab were available; this small number of samples did not allow us to investigate potential changes in biomarker expression induced by the use of an anti-angiogenic therapy.

Although the importance of repeated tumor sampling is being recognized in the setting of both clinical trials and routine practice, most of the available information on individual tumor characteristics are still largely based on the assessment of archived tumor tissue taken at the time of the initial diagnosis. We acknowledge that, due to the limited tumor sampling, our study does not take into account the potential contribution of the intralesional heterogeneity to the substantial discordance in biomarker expression between primary tumors and metastases. However, our findings are in line with the recent reports on tumor heterogeneity and clonal evolution, and further support the view that the source of tumor tissue should be an important factor to consider when interpreting the results of tumor biomarker analyses.

We found that lack of PTEN expression in the metastases was an independent predictor of response to treatment with chemotherapy plus bevacizumab. PTEN has already been reported to act as a potential biomarker. In particular, PTEN loss has been suggested to be a negative predictive factor for benefit from other targeted agents such as cetuximab in mCRC and trastuzumab in metastatic breast cancer (36–37–38). To our knowledge, this is the first study to show a potential role for PTEN in predicting response to bevacizumab in mCRC. Given that PTEN loss results in activation of the PI3K/AKT/mTOR pathway, it could be hypothesized that tumors with increased activation of PI3K-dependent angiogenic pathways may be more sensitive to the therapeutic effect of bevacizumab (39). However, it is worth noting that, possibly due to the small sample size, no significant differences in survival outcomes were observed in our population according to the expression of PTEN. Very limited information is currently available on the role of PTEN in predicting the outcome of an anti-angiogenic therapy in CRC. The largest set of data is from a retrospective analysis of the AGITG MAX trial, in which the additional benefit of bevacizumab on PFS and OS was not found to be significantly greater among the patients with tumors exhibiting PTEN loss compared to patients with no PTEN loss in tumors. However, 94% of the tumor tissue analyzed was from the primary site, and all samples were analyzed for PTEN copy number (40).

In contrast with previous studies, CAIX expression, either measured in epithelial tumor cells or stromal cells, was not found to predict tumor response, PFS, or OS (41, 42). Similarly, although preclinical data suggested a role for c-MET in promoting invasiveness and metastasis in hypoxic conditions (25), no association was found between the expression of this biomarker and treatment outcome. However, the association between c-MET and bevacizumab effect could not be properly assessed due to the small number of c-MET-positive tumors in our study.

We also analyzed the expression of PI3K p110a in order to further investigate the potential of the PI3K/AKT/mTOR axis in mediating responsiveness or resistance to an anti-VEGF strategy. PI3K p110a expression in metastatic tumors did not predict treatment outcome; however, its expression in primary tumors was associated with an improvement in PFS in univariate analysis. Although the association was not significant in multivariate analysis, this finding would possibly be in line with the hypothesis of an increased efficacy of bevacizumab in tumors with upregulation of the PI3K/AKT/mTOR pathway.

Studies suggest that high LDH levels are associated with poor prognosis in CRC patients treated with bevacizumab and may represent a potential stratification factor for treatment selection (43, 44). In our study, baseline LDH levels were available for a limited number of patients (n=25). Although a trend towards a better outcome for patients with normal LDH at baseline was observed, this was not statistically significant (data not shown).

We acknowledge that our analysis is negatively affected by several factors. These include the retrospective design of the study, the small sample size and heterogeneity of the study population, the inherent limitations associated with the use of IHC, and the absence of a validated scoring system for any of the biomarkers analyzed. Moreover, our study did not include a control group treated without bevacizumab and, even in the presence of a significant association between PTEN and tumor response, a true predictive (rather than prognostic) value of this biomarker could not be demonstrated.

Conclusions

In this exploratory, retrospective study we showed that significant heterogeneity exists between primary colorectal tumors and metastases with regard to the expression of several biomarkers including CAIX, PTEN, PI3K, and c-MET. Moreover, we found that PTEN loss in the metastases was associated with increased tumor response to a bevacizumab-based therapy. Larger studies are necessary to further investigate whether the PI3K/AKT/mTOR pathway may truly modulate the therapeutic effect of bevacizumab and whether PTEN may represent a valid biomarker for the selection of patients more likely to benefit from an anti-angiogenic therapy.

Footnotes

Informed Consent: No study-specific consent was obtained from patients included in this study. However, the study protocol was approved by the Ethics Committee/local Institutional Review Boards.

Ethical Considerations: The study was approved by the local Institutional Review Boards. All the procedures followed were in accord with the ethical standards of the committee on human experimentation of our institution.

Financial Support: Roche provided a research grant for the biomarker analysis. Roche was not involved in study design, data analysis, or manuscript preparation and had no access to study data.

Conflict of Interest Statements: The authors have declared that no conflict of interest exists.

References

Folkman

Tumor angiogenesis: therapeutic implications.

N Engl J Med 1971 285(21): 1182–1186

Folkman

What is the evidence that tumors are angiogenesis dependent?

J Natl Cancer Inst 1990 82(1): 4–6

Fidler

Ellis

The implications of angiogenesis for the biology and therapy of cancer metastasis.

Cell 1994 79(2): 185–188

Presta

Chen

O'Connor

Humanization of an anti-vascular endothelial growth factor monoclonal antibody for the therapy of solid tumors and other disorders.

Cancer Res 1997 57(20): 4593–4599

Hicklin

Ellis

Role of the vascular endothelial growth factor pathway in tumor growth and angiogenesis.

J Clin Oncol 2005 23(5): 1011–1027

Kabbinavar

Hambleton

Mass

Hurwitz

Bergsland

Sarkar

Combined analysis of efficacy: the addition of bevacizumab to fluorouracil/leucovorin improves survival for patients with metastatic colorectal cancer.

J Clin Oncol 2005 23(16): 3706–3712

Hurwitz

Fehrenbacher

Novotny

Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer.

N Engl J Med 2004 350(23): 2335–2342

Giantonio

Catalano

Meropol

Eastern Cooperative Oncology Group Study E3200. Bevacizumab in combination with oxaliplatin, fluorouracil, and leucovorin (FOLFOX4) for previously treated metastatic colorectal cancer: results from the Eastern Cooperative Oncology Group Study E3200.

J Clin Oncol 2007 25(12): 1539–1544

Kabbinavar

Hurwitz

Fehrenbacher

Phase II, randomized trial comparing bevacizumab plus fluorouracil (FU)/leucovorin (LV) with FU/LV alone in patients with metastatic colorectal cancer.

J Clin Oncol 2003 21(1): 60–65

10.

Saltz

Clarke

Díaz-Rubio

Bevacizumab in combination with oxaliplatin-based chemotherapy as first-line therapy in metastatic colorectal cancer: a randomized phase III study.

J Clin Oncol 2008 26(12): 2013–2019

11.

Loupakis

Bria

Vaccaro

Magnitude of benefit of the addition of bevacizumab to first-line chemotherapy for metastatic colorectal cancer: meta-analysis of randomized clinical trials.

J Exp Clin Cancer Res 2010 29(1): 58

12.

Ince

Jubb

Holden

Association of k-ras, b-raf, and p53 status with the treatment effect of bevacizumab.

J Natl Cancer Inst 2005 97(13): 981–989

13.

Hurwitz

Ince

Novotny

Rosen

The clinical benefit of bevacizumab in metastatic colorectal cancer is independent of K-ras mutation status: analysis of a phase III study of bevacizumab with chemotherapy in previously untreated metastatic colorectal cancer.

Oncologist 2009 14(1): 22–28

14.

Jubb

Hurwitz

Bai

Impact of vascular endothelial growth factor-A expression, thrombospondin-2 expression, and microvessel density on the treatment effect of bevacizumab in metastatic colorectal cancer.

J Clin Oncol 2006 24(2): 217–227

15.

Murukesh

Dive

Jayson

Biomarkers of angiogenesis and their role in the development of VEGF inhibitors.

Br J Cancer 2010 102(1): 8–18

16.

Maru

Venook

Ellis

Predictive biomarkers for bevacizumab: are we there yet?

Clin Cancer Res 2013 19(11): 2824–2827

17.

Mizukami

Kohgo

Chung

Hypoxia inducible factor-1 independent pathways in tumor angiogenesis.

Clin Cancer Res 2007 13(19): 5670–5674

18.

Mizukami

Zhang

Zimmer

Iliopoulos

Chung

Hypoxia-inducible factor-1-independent regulation of vascular endothelial growth factor by hypoxia in colon cancer.

Cancer Res 2004 64(5): 1765–1772

19.

Mizukami

Fujiki

Duerr

Hypoxic regulation of vascular endothelial growth factor through the induction of phosphatidylinositol 3-kinase/Rho/ROCK and c-Myc.

J Biol Chem 2006 281(20): 13957–13963

20.

Qiao

Hung

Tremblay

Constitutive activation of met kinase in non-small-cell lung carcinomas correlates with anchorage-independent cell survival.

J Cell Biochem 2002 86(4): 665–677

21.

Beviglia

Matsumoto

Lin

Ziober

Kramer

Expression of the c-Met/HGF receptor in human breast carcinoma: correlation with tumor progression.

Int J Cancer 1997 74(3): 301–309

22.

Takeuchi

Bilchik

Saha

c-MET expression level in primary colon cancer: a predictor of tumor invasion and lymph node metastases.

Clin Cancer Res 2003 9(4): 1480–1488

23.

Lee

Chow

Lin

Lee

The prognostic significance of RON and MET receptor coexpression in patients with colorectal cancer.

Dis Colon Rectum 2008 51(8): 1268–1274

24.

Di Renzo

Olivero

Giacomini

Overexpression and amplification of the met/HGF receptor gene during the progression of colorectal cancer.

Clin Cancer Res 1995 1(2): 147–154

25.

Pennacchietti

Michieli

Galluzzo

Mazzone

Giordano

Comoglio

Hypoxia promotes invasive growth by transcriptional activation of the met protooncogene.

Cancer Cell 2003 3(4): 347–361

26.

Potter

CPS

Harris

Diagnostic, prognostic and therapeutic implications of carbonic anhydrases in cancer.

Br J Cancer 2003 89(1): 2–7

27.

Saarnio

Parkkila

Immunohistochemical study of colorectal tumors for expression of a novel transmembrane carbonic anhydrase, MN/CA IX, with potential value as a marker of cell proliferation.

Am J Pathol 1998 153(1): 279–285

28.

Kivela

Saarnio

Karttunen

Differential expression of cytoplasmic carbonic anhydrases, CA I and II, and membrane-associated isozymes, CA IX and XII, in normal mucosa of large intestine and in colorectal tumors.

Dig Dis Sci 2001 46(10): 2179–2186

29.

Saarnio

Parkkila

Immunohistochemistry of carbonic anhydrase isozyme IX (MN/CA IX) in human gut reveals polarized expression in the epithelial cells with the highest proliferative capacity.

J Histochem Cytochem 1998 46(4): 497–504

30.

Talvinen

Tuikkala

Grönroos

Biochemical and clinical approaches in evaluating the prognosis of colon cancer.

Anticancer Res 2006 266C 4745–4751

31.

Niemelä

Hynninen

Mecklin

J-P

Carbonic anhydrase IX is highly expressed in hereditary nonpolyposis colorectal cancer.

Cancer Epidemiol Biomarkers Prev 2007 16(9): 1760–1766

32.

Cleven

van Engeland

Wouters

de Bruïne

Stromal expression of hypoxia regulated proteins is an adverse prognostic factor in colorectal carcinomas.

Cell Oncol 2007 29(3): 229–240

33.

Laurent-Puig

Cayre

Manceau

Analysis of PTEN, BRAF, and EGFR status in determining benefit from cetuximab therapy in wild-type KRAS metastatic colon cancer.

J Clin Oncol 2009 27(35): 5924–5930

34.

Santoro

Rimassa

Borbath

Tivantinib for second-line treatment of advanced hepatocellular carcinoma: a randomised, placebo-controlled phase 2 study.

Lancet Oncol 2013 14(1): 55–63

35.

Gerlinger

Rowan

Horswell

Intratumor heterogeneity and branched evolution revealed by multiregion sequencing.

N Engl J Med 2012 366(10): 883–892

36.

Loupakis

Pollina

Stasi

PTEN expression and KRAS mutations on primary tumors and metastases in the prediction of benefit from cetuximab plus irinotecan for patients with metastatic colorectal cancer.

J Clin Oncol 2009 27(16): 2622–2629

37.

Nagata

Lan

K-H

Zhou

PTEN activation contributes to tumor inhibition by trastuzumab, and loss of PTEN predicts trastuzumab resistance in patients.

Cancer Cell 2004 6(2): 117–127

38.

Gallardo

Lerma

Escuin

Increased signalling of EGFR and IGF1R, and deregulation of PTEN/PI3K/Akt pathway are related with trastuzumab resistance in HER2 breast carcinomas.

Br J Cancer 2012 106(8): 1367–1373

39.

Castellino

Muh

Durden

PI-3 kinase-PTEN signaling node: an intercept point for the control of angiogenesis.

Curr Pharm Des 2009 15(4): 380–388

40.

Price

Hardingham

Lee

Prognostic impact and the relevance of PTEN copy number alterations in patients with advanced colorectal cancer (CRC) receiving bevacizumab.

Cancer Med 2013 2(3): 277–285

41.

Korkeila

Talvinen

Jaakkola

Expression of carbonic anhydrase IX suggests poor outcome in rectal cancer.

Br J Cancer 2009 100(6): 874–880

42.

Hong

Cho

Kim

Carbonic anhydrase 9 is a predictive marker of survival benefit from lower dose of bevacizumab in patients with previously treated metastatic colorectal cancer.

BMC Cancer 2009 9(1): 246

43.

Scartozzi

Giampieri

Maccaroni

Pre-treatment lactate dehydrogenase levels as predictor of efficacy of first-line bevacizumab-based therapy in metastatic colorectal cancer patients.

Br J Cancer 2012 106(5): 799–804

44.

Bar

Spencer

Morgan

Correlation of lactate dehydrogenase isoenzyme profile with outcome in patients with advanced colorectal cancer treated with chemotherapy and bevacizumab or cediranib: Retrospective analysis of the HORIZON I study.

Clin Colorectal Cancer 2014 13(1): 46–53