Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2011 Jan 1.
Published in final edited form as: Br J Haematol. 2009 Oct 11;148(2):235–244. doi: 10.1111/j.1365-2141.2009.07942.x

Lymphoma cell VEGFR2 expression detected by immunohistochemistry predicts poor overall survival in diffuse large B cell lymphoma treated with immunochemotherapy (R-CHOP)

Dita Gratzinger 1, Ranjana Advani 2, Shuchun Zhao 1, Neha Talreja 2, Robert J Tibshirani 3, Ragini Shyam 3, Sandra Horning 2, Laurie H Sehn 4, Pedro Farinha 4, Javier Briones 5, Izidore S Lossos 6, Randy D Gascoyne 4, Yasodha Natkunam 1
PMCID: PMC2809124  NIHMSID: NIHMS152438  PMID: 19821819

Abstract

Diffuse large B cell lymphoma (DLBCL) is clinically and biologically heterogeneous. In most cases of DLBCL, lymphoma cells coexpress vascular endothelial growth factor (VEGF) and its receptors VEGFR1 and VEGFR2, suggesting autocrine in addition to angiogenic effects. We enumerated microvessel density and scored lymphoma cell expression of VEGF, VEGFR1, VEGFR2 and phosphorylated VEGFR2 in 162 de novo DLBCL patients treated with R-CHOP (rituximab, cyclophosphamide, vincristine, doxorubicin and prednisone)-like regimens. VEGFR2 expression correlated with shorter overall survival (OS) independent of International Prognostic Index (IPI) (p=0.0028). Phosphorylated VEGFR2 (detected in 13% of cases) correlated with shorter progression-free survival (PFS, p=0.044) and trended toward shorter OS on univariate analysis. VEGFR1 was not predictive of survival on univariate analysis, but it did correlate with better OS on multivariate analysis with VEGF, VEGFR2, and IPI (p=0.036); in patients with weak VEGFR2, lack of VEGFR1 coexpression was significantly correlated with poor OS independent of IPI (p=0.01). These results are concordant with our prior finding of an association of VEGFR1 with longer OS in DLBCL treated with chemotherapy alone. We postulate that VEGFR1 may oppose autocrine VEGFR2 signaling in DLBCL by competing for VEGF binding. In contrast to our prior results with chemotherapy alone, microvessel density was not prognostic of PFS or OS with R-CHOP-like therapy.

Keywords: Non-Hodgkin lymphoma, VEGF, angiogenesis, tumour biology, prognostic factors

Introduction

Immunochemotherapy combining anti-CD20 recombinant antibody with anthracycline-based chemotherapy has significantly improved survival in diffuse large B cell lymphoma (DLBCL) (Mounier et al. 2003),(Sehn et al. 2005),(Pfreundschuh et al. 2006),(Habermann et al. 2006). This improvement in outcomes has also necessitated renewed examination of factors that were previously shown to be prognostic in the setting of chemotherapy alone. Vascular endothelial growth factor (VEGF)-A is the founding member of the family of vascular endothelial growth factors (Cross et al. 2003) and is commonly simply designated VEGF. VEGF has been a successful target of antiangiogenic therapy in the setting of solid tumors,(Kabbinavar et al. 2003) and clinical trials of similar therapies are under way in the setting of non-Hodgkin lymphoma.(Ganjoo et al. 2006),(Stopeck et al. 2009) We have previously found that microvessel density, a static surrogate measure of tumor angiogenesis, is a poor prognostic indicator in DLBCL treated with anthracycline-containing chemotherapy alone (Gratzinger et al. 2008, Natkunam et al. 2008) suggesting that angiogenesis may be clinically important in DLBCL in that therapeutic setting. VEGF signaling however is not limited to the vasculature (Podar and Anderson 2005). In addition to its angiogenic and vascular permeabilizing properties, VEGF also has a direct role in hematolymphoid cell development (Gerber et al. 2002). Cell lines derived from a variety of hematolymphoid malignancies have been shown to express both VEGF and its receptors (Bellamy et al. 1999). VEGF, VEGFR1 and VEGFR2 are coordinately expressed in primary human DLBCL specimens (Gratzinger et al. 2007), suggesting a potential role for VEGF-mediated autocrine signaling in lymphoma. Indeed, DLBCL cell lines in culture proliferate in response to VEGF stimulation (Wang et al. 2004),(Rahimi et al 2009). We therefore hypothesized that VEGF-mediated signaling at the level of the lymphoma cell may be functionally and prognostically relevant in DLBCL. In our previous study of DLBCL treated with anthracycline-containing chemotherapy alone we found that immunohistochemically detectable expression of VEGFR1 by lymphoma cells correlated with improved overall survival (OS); VEGFR2 and VEGF were not independently associated with survival in that setting (Gratzinger et al. 2008).

Both VEGFR1 and VEGFR2 are activated following VEGF ligand binding resulting in intracellular signaling events; tyrosine phosphorylation of VEGFR2 following its activation is strong and immunohistochemically detectable (Olsson et al. 2006). Tyrosine-phosphorylated VEGFR2 (phosphoVEGFR2) has previously been shown to be expressed in non-Hodgkin lymphoma.(Stewart et al. 2003),(Giatromanolaki et al. 2008) We have now assessed the prognostic relevance of microvessel density, VEGF, VEGFR1, VEGFR2 and phosphoVEGFR2 in DLBCL in the setting of the current standard of care, immunochemotherapy (rituximab, cyclophosphamide, vincristine, doxorubicin and prednisone; R-CHOP).

Methods

Patient selection

Pre-treatment biopsies of 162 patients with de novo DLBCL treated with anti-CD20 antibody (rituximab) plus an anthracycline-containing chemotherapy (cyclophosphamide, vincristine, doxorubicin and prednisone [CHOP] or CHOP-like regimens) with clinical follow-up data were used. The biopsy specimens and clinical follow-up data originated from four institutions: University of British Columbia/British Columbia Cancer Agency (77 cases); Stanford University (53 cases); Hospital Santa Creu i Sant Pau. (Universitat Autònoma de Barcelona) (19 cases); and University of Miami (12 cases). Specimens were chosen for this study based on the following criteria: (1) diagnosis of de novo DLBCL; (2) availability of tissue obtained at diagnosis before the initiation of therapy; (3) treatment with R-CHOP; and (4) availability of follow-up and outcome data at the treating institution.

All patients were treated with curative intent. Patients who underwent therapy intensification with stem cell transplantation despite clinical response to initial regimens were excluded from this study. Institutional Review Board approval was obtained from all participating institutions. In all patients chosen for this study, information was available about staging of the disease by physical examination, bone marrow biopsy, and computed tomography of the chest, abdomen, and pelvis. Patients were staged according to the Ann Arbor system. The following clinical and laboratory data at the time of diagnosis was available: age, performance status, stage, number of extranodal sites involved, and levels of serum lactate dehydrogenase (LDH). Based on this information, international prognostic index (IPI) scores could be determined on all patients. Patients were categorized into either a low-risk group (IPI score 0–2) or a high-risk group (IPI score of 3–5). None of the patients had a known history of HIV infection or other forms of immunosuppression. Follow-up information was obtained from the patients’ medical records and included response to initial therapy based on the Cheson criteria,(Cheson et al. 1999) progression-free survival (PFS), and OS. Histological sections were reviewed to confirm the diagnoses. All cases showed a diffuse large cell infiltrate without any evidence of follicles or other low-grade component and were compatible with the histological features of DLBCL according to the World Health Organization classification of haematopoietic tumours (Jaffe et al. 2001).

Tissue Microarray(TMA)

Standardized methods for tissue fixation (10% buffered formalin) and processing for paraffin-embedded sections were used at all participating centres. Tissue microarrays (TMAs) of formalin-fixed, paraffin-embedded tissue samples of DLBCL were obtained from University of British Columbia. TMAs of cases from Stanford University, University of Miami, and Universitat Autònoma de Barcelona were constructed using a tissue arrayer (Beecher Instruments, Silver Spring, MD), as previously described (Natkunam et al. 2005). Tissue cores were selected for TMA by characteristic morphology based on examination of haematoxylin and eosin-stained sections, without prior knowledge of immunohistological stains of individual cases. Two to four representative cores of each case were included to maximize informative cores. Sections of 4–5 micron thickness were cut from the tissue microarrays and placed on glass slides, then baked for 1 h at 60°C.

Immunohistochemistry and microvessel density counting

Antibodies used for immunohistochemistry and staining conditions are listed in Table I. Immunohistochemistry was performed on 4-µm sections, which were placed on glass slides, baked for 1 h at 60°C, deparaffinized in xylene and hydrated in a graded series of alcohol. Endogenous peroxidase was blocked, and the chromogen used was diaminobenzidine.

Table I.

Immunohistochemistry materials and methods.

Antibody Clone Manufacturer Dilution Antigen
retrieval		 Chromogen
development
VEGF A-20 Santa Cruz Biotech,
Santa Cruz, CA		 1:75 Tris EnVision Kit (Dako,
Carpinteria, CA)
VEGFR1 C-17 Santa Cruz Biotech 1:40 Tris EnVision Kit (Dako)
VEGFR2 A-3 Santa Cruz Biotech 1:50 Tris EnVision Kit (Dako)
Phospho
VEGFR2		 Tyr1175 Cell Signaling
Technology, Danvers, MA		 1:30 Citrate EnVision Kit (Dako)
CD34 QBEnd Dako, Carpinteria, CA 1:40 Benchmark stainer, Ventana
Medical Systems, Tucson, AZ
Open in a new tab

Duplicate cores were evaluated in each case; for microvessel density the mean score was reported, and for all other scores the maximum score was reported. Stains for VEGF, VEGFR1, VEGFR2, and phosphoVEGFR2 were scored as follows: >30% of lymphoma cells staining, score=2 (strong); 5–30% of lymphoma cells staining, or weak diffuse staining, score=1 (weak); and <5% of lymphoma cells staining, score=0 (none). Microvessel density was quantitated as the number of CD34+ microvessels (defined as any distinct CD34+ cell or cell cluster)(Weidner et al. 1991) in the entire 0.7mm core at 300x using a model BX45 microscope (Olympus, Center Valley, PA). Microvessel density counts above the median score of 26 were denoted as “high”, and those denoted “low” were below the median score. Noninformative cores (nondiagnostic tissue, necrotic tumor, or obscuring reactivity in non-tumor cells) were excluded from analysis; for purposes of microvessel density counting, which required evaluation of the entire core, cores that were partially detached or only partially diagnostic tissue were also excluded. The percentages of informative cases for each stain were as follows: VEGF 92%, VEGFR1 94%, VEGFR2 95%, phosphoVEGFR2 86%, CD34 (microvessel density counting) 76%. Scoring was performed in a blinded manner by DG and checked by YN. Discrepancies (5–10%) were resolved over a multi-headed microscope. Images of immunohistological staining were acquired using a model BX51 microscope with a 100x/1.25 NA Plan oil objective lens 40x/0.75 NA Plan Fluor objective lens and Microfire™ digital camera with PictureFrame™ software (Olympus). Digitized images were processed using Adobe Photoshop 7 image processing and manipulation software (Adobe Systems, San Jose, CA).

Statistical Analysis

Survival curves were estimated using the product-limit method of Kaplan-Meier and were compared using the log-rank test. Univariate and multivariate analysis was performed according to the Cox proportional hazards regression model. The endpoints used were OS and PFS. The Chi-squared test for independence was used to assess for correlations between categorical variables, with the Yates continuity correction applied where indicated. For numerical variables, the independent samples t-test or the Mann-Whitney test (where normality was accepted via the D’Agostino-Pearson test) were used. Statistical analyses were performed using R: A Language and Environment for Statistical Computing (http://www.R-project.org, Vienna, Austria) and MedCalc for Windows, version 10.3.2.0 (MedCalc Software, Mariakerke, Belgium). P values < 0.05 were considered significant.

Results

Patient characteristics

162 patients with an initial diagnosis of de novo DLBCL were drawn from multiple institutions in the United States, Canada and Europe. This cohort represents all patients treated with R-CHOP with curative intent at those institutions during the study period for whom paraffin blocks were available. The median age of the patients was 59.5 years (range, 20–85 years; 49% >60 years old); patients were initially diagnosed between 2000 and 2006, with the majority diagnosed in 2001 and 2002. Patients were followed for a median of 45 months (range, 1–94 months) from the initial diagnosis of de novo DLBCL; 25% died and 37% suffered disease progression during the follow-up interval. About 2/3 of patients had low IPI scores (0–2) and 1/3 had high IPI scores (3–5). The distribution of IPI risk factors in this cohort is similar to that of previously reported retrospective and prospective DLBCL studies, including the original publication of the IPI model (Shipp et al. 1993). At the end of the follow-up interval 16% of patients with low IPI scores had died and 27% had suffered disease progression; among patients with high IPI scores, 41% had died and 54% had suffered disease progression. Table II presents the baseline clinical characteristics in the patient population overall and with respect to each of the immunohistochemical parameters studied. There was no statistically significant (p<0.05) association among the immunohistochemical parameters studied (VEGF, VEGFR1, VEGFR2, phosphoVEGFR2, and microvessel density) and patient characteristics (age, stage, performance status, LDH, number of extranodal sites, and IPI) except that patients with high IPI scores and high LDH were overrepresented among the VEGFR2 negative cases (p=0.022 for LDH, p=0.016 for IPI). Given the large number of comparisons performed, this was probably a chance finding.

Table II.

Clinical characteristics: All comparisons were statistically non-significant except that high lactate dehydrogenase (LDH; p=0.022) and high International Prognostic Index (IPI; p=0.016) are overrepresented in the no VEGFR2 category. Percentages are in parentheses.

VEGF VEGFR1 VEGFR2 phospho
VEGFR2		 microvessel
density
all none weak strong none weak strong none weak strong none positive <median >median
n 162 7(5) 49(33) 92(62) 17(11) 71(47) 64(42) 12(8) 60(39) 81(53) 120(87) 18(13) 59(50) 60(50)
mean age (years) 58 62 58 57 61 59 55 58 57 58 57 62 56 59
stage
I–II 79 4(6) 25(35) 42(59) 9(12) 34(46) 31(42) 3(4) 30(41) 41(55) 60(88) 8(12) 24(44) 31(56)
III–IV 83 3(4) 24(31) 50(65) 8(10) 37(47) 33(42) 9(11) 30(38) 40(51) 60(86) 10(14) 35(55) 29(45)
Performance Status
0–1 113 6(6) 35(33) 64(61) 12(11) 53(50) 41(39) 6(6) 42(40) 58(55) 86(88) 12(12) 41(51) 40(49)
≥2 49 1(2) 14(33) 28(65) 5(11) 18(39) 23(50) 6(13) 18(38) 23(49) 34(85) 6(15) 18(47) 20(53)
LDH
Normal 85 5(6) 27(34) 47(59) 7(9) 39(49) 33(42) 2(3) 36(46) 41(52) 65(88) 9(12) 31(54) 26(46)
High 77 2(3) 22(32) 45(65) 10(14) 32(44) 31(42) 10(14) 24(32) 40(54) 55(86) 9(14) 28(45) 34(55)
Extranodal Sites
0–1 128 6(5) 36(31) 75(64) 14(12) 56(47) 50(42) 10(8) 49(40) 64(52) 97(88) 13(12) 44(47) 50(53)
≥2 34 1(3) 13(42) 17(55) 3(9) 15(47) 14(44) 2(7) 11(37) 17(57) 23(82) 5(18) 15(60) 10(40)
International
Prognostic Index
0–2 103 5(5) 31(33) 59(62) 10(10) 45(46) 42(43) 3(3) 40(41) 54(56) 80(89) 10(11) 38(50) 38(50)
3–5 59 2(4) 18(34) 33(62) 7(13) 26(47) 22(40) 9(16) 20(36) 27(48) 40(83) 8(17) 21(49) 22(51)
Open in a new tab

Lymphoma characteristics

The majority of DLBCL specimens had immunohistochemically detectable VEGF (95%), VEGFR1 (89%), and VEGFR2 (92%); by contrast, phosphoVEGFR2 was detected in only 13% of the specimens studied. (Fig 1, % immunoreactivity for all markers; Fig 2, representative photomicrographs). PhosphoVEGFR2 in endothelial cells served as an internal positive control and was present in 92% of the specimens studied (Figure 2d). As we have previously demonstrated, VEGF and its receptors VEGFR1 and VEGFR2 are coordinately overexpressed (Gratzinger et al. 2007, Gratzinger et al. 2008) (VEGF vs VEGFR1, p=0.008, X=19; VEGF vs VEGFR2, p=2×10−7, X=37; VEGFR1 vs VEGFR2, p=1×10−11, X=56.5). By contrast, there was no statistically significant relationship between lymphoma cell expression of VEGF, VEGFR1, or VEGFR2 and detection of phosphoVEGFR2. Microvessel density was highly variable (range 5–177, median 26, standard deviation 21) and there was no association between VEGF expression by lymphoma cells and microvessel density.

Figure 1.

Figure 1

Open in a new tab

DLBCL expression of VEGF, VEGFR1, VEGFR2, and phosphoVEGFR2. Whereas the majority of DLBCL cases express VEGF and its receptors VEGFR1 and VEGFR2, a minority of DLBCL cases express immunohistochemically detectable amounts of phosphoVEGFR2.

Figure 2.

Figure 2

Open in a new tab

Representative immunohistochemistry demonstrates DLBCL expression of VEGF (a), VEGFR1 (b), VEGFR2 (c), phosphoVEGFR2 (d, negative; e, nuclear; f, cytoplasmic. Endothelial cell expression of phosphoVEGFR2 (d) acts as an internal positive control in cases where lymphoma cell phosphoVEGFR2 is not detected. Original magnification ×400.

Outcome data

The results of survival analyses for the endpoints of PFS and OS, including hazard ratios (HR) and 95% confidence interval, are summarized in Table III; in the text the 95% confidence interval follows the HR in parentheses. Kaplan-Meier survival curves are presented for significant comparisons in Figure 3 – Figure 5.

Table III.

Logrank analysis and Cox proportional hazards regression for overall survival and progression-free survival.

Logrank Analysis (Univariate)
Overall Survival Progression-free survival
HR 95% CI p-value HR 95% CI p-value
IPI 1.7 1.3–2.1 0.00036 1.4 1.2–1.7 0.00012
VEGF 1.6 0.86–3.1 0.13 1.1 0.67–1.7 0.22
VEGFR1 0.90 0.56–1.4 0.67 0.81 0.56–1.2 0.29
VEGFR2 2.2 1.2–4.0 0.014 1.2 0.78–1.8 0.40
Phospho
VEGFR2 		1.7 0.90–3.3 0.10 1.7 1.0–2.9 0.044
Microvessel
density 		1.0 0.99–1.0 0.66 1.0 0.99–1.0 0.91
Cox proportional hazards model
Overall Survival Progression-free survival
HR 95% CI p-value HR 95% CI p-value
IPI 1.8 1.4–2.3 0.000016 1.4 1.2–1.7 0.00063
VEGF 1.5 0.69–3.2 0.31 1.1 0.63–1.9 0.78
VEGFR1 0.51 0.27–0.96 0.036 0.66 0.40–1.1 0.11
VEGFR2 2.9 1.3–6.4 0.0092 1.5 0.86–2.7 0.15
Open in a new tab

HR: Hazard ratio; CI: Confidence interval

Figure 3.

Figure 3

Open in a new tab

Overall survival and VEGFR2. a) OS is significantly shorter with increasing VEGFR2 [p=0.0028, HR=2.5(1.5–4.6)]. on multivariate analysis with IPI). b) Among patients with low IPI, there is a non-significant trend towards shorter OS (p=0.29, z=1.07). c) Among patients with high IPI, OS is significantly shorter with increasing VEGFR2 (p=0.0054, z=2.78).

Figure 5.

Figure 5

Open in a new tab

Overall survival and VEGFR1, stratified by VEGFR2 status. a) Among patients with no VEGFR2, survival is 100% regardless of VEGFR1 status. b) Among patients with weak VEGFR2, lack of VEGFR1 correlates with shorter OS [*p=0.0077, HR=8.9(1.8–44) for no versus any VEGFR1]. c) Among patients with strong VEGFR2, there is no statistical difference in survival by VEGFR1 status.

As expected, IPI scores were highly predictive of both OS [p=0.000036, HR 1.7(1.3–2.1)] and PFS [p=0.00012, HR 1.4(1.2–1.7)]. Univariate logrank analysis also revealed a statistically significant association between lymphoma cell expression of VEGFR2 and shorter OS [p=0.014, HR 2.2(1.2–4.0)]; on multivariate analysis (Cox proportional hazards regression) this was independent of IPI [p=0.0028, HR=2.5(1.5–4.6)]. Two-year survivals (+/−standard error) were 100%, 84(+/−5)%, and 76(+/−5)% respectively (Kaplan-Meier survival curve, Fig 3A). The relationship between VEGFR2 and OS was strongest among patients with high IPI [p=0.0054,HR=2.9(1.4–6.2)]; in this group two-year survivals were 100%, 67(+/−11)%, and 63(+/−10)% (Fig 3C). Among patients with low IPI, there was a non-significant trend toward poorer OS, with two-year OS of 100%, 92(+/−4)%, and 82(+/−5)% [p=0.29; HR=1.7(0.63–4.7) Fig 3B]. A test for interaction between the IPI risk categories and VEGFR2 was non-significant; the lack of statistical significance for VEGFR2 in the low IPI risk category could therefore be that lymphoma cell VEGFR2 expression did not significantly correlate with PFS, although a non-significant trend toward worse PFS was present.

Like VEGFR2, phosphoVEGFR2 was also associated with poorer survival on univariate logrank analysis (Kaplan-Meier survival curves, Figs 4A and 4B). The association was statistically significant for PFS [p=0.044, HR=1.7(1.0–2.9)] and there was a non-significant trend toward poorer OS in phosphoVEGFR2 positive cases [p=0.10, HR=1.7(0.90–3.30]. Two-year PFS was 74(+/−4)% for patients with phosphoVEGFR2-negative lymphoma versus 54(+/−12)% for those with detectable phosphoVEGFR2. The association of phosphoVEGFR2 with shorter PFS did not reach statistical significance on multivariate analysis with IPI [p=0.14, HR=1.4(0.9–2.6)].

Figure 4.

Figure 4

Open in a new tab

Survival and phosphoVEGFR2. a) Progression-free survival is significantly decreased in patients with detectable phosphoVEGFR2 [p=0.044, HR=1.7(1.0–2.9)]. b) There is a non-significant trend toward decreased overall survival with detectable phosphoVEGFR2 [p=0.1, HR=1.7(0.90–3.30)].

VEGFR1 was not associated with survival on univariate analysis of the group overall. However on multivariate analysis (Cox proportional hazards regression) including VEGF, VEGFR1, VEGFR2, and IPI, an independent association with longer OS emerged [p=0.036; HR=0.51(0.27–0.96)]. Since expression of VEGFR1 and VEGFR2 is tightly correlated, we hypothesized that the association of VEGFR2 with poorer outcomes might mask a weaker association of VEGFR1 with better outcomes. We therefore performed subgroup analysis to look for an association of VEGFR1 with OS when VEGFR2 expression was held constant. The subgroup of patients with no VEGFR2 showed 100% two-year OS and thus no additional effect of VEGFR1 could be expected (Kaplan-Meier survival curve, Fig 5A). Patients with weak VEGFR2, by contrast, had 84(+/−5)% two-year OS as a group; when this group was stratified by VEGFR1 expression, statistically significant differences in survival emerged. –Two-year OS was 68(+/−15)%, 85(+/−6)% and 92(+/−8)%, respectively, with no, weak, or strong VEGFR1 (Kaplan-Meier survival curve, Fig 5B) [p=0.058, HR=0.40 (0.16–1.0)]. The risk was highest among patients lacking VEGFR1, compared to those whose tumours coexpressed VEGFR1 and VEGFR2, [p=0.0077, HR=8.9(1.8–44)] and this finding remained statistically significant on multivariate analysis with IPI (p=0.01). The subgroup of patients with high VEGFR2 did not contain any patients with no VEGFR1, reflecting the strong positive correlation between VEGFR1 and VEGFR2 expression. In this group the difference between two-year OS for weak (74+/−8%) and strong (78+/−6%) VEGFR2 was not statistically significant (p=0.47) (Kaplan-Meier survival curve, Fig 5C).

VEGF expression by lymphoma cells was not predictive of OS [p=0.13, HR=1.6 (0.86–3.1)] or PFS [p=0.22, HR=1.1 (0.67–1.7)] on univariate analysis or on multivariate analysis with VEGFR1, VEGFR2 and IPI. Microvessel density was also not predictive of OS [p=0.66, HR=1.0(0.00–1.0)] or PFS [p=0.91, HR=1.0(0.99–1.0)].

Discussion

With the success of anti-VEGF therapy in colorectal carcinoma and other solid tumors, attention has turned to the potential therapeutic role of VEGF inhibitors in lymphoma. Little is known about the functional role of VEGF signaling in lymphoma. In hematopoietic stem cells, an autocrine loop involving VEGF and its receptor VEGFR1 modulates in vivo hematopoietic stem cell survival and proliferation (Gerber et al. 2002). Cell lines derived from a variety of haematolymphoid malignancies have been shown to express both VEGF and VEGFR1 (Bellamy et al. 1999), suggesting a role for a similar autocrine loop in neoplasia. Indeed, both VEGFR1 and VEGFR2 are capable of mediating DLBCL proliferation in culture, and VEGFR1 signaling promotes growth of human aggressive B cell non-Hodgkin lymphoma xenografts in mice (Wang et al. 2004). In contrast, in the developing vasculature, VEGFR2-mediated signaling is the predominant growth signal, and VEGFR1 appears to largely act as a brake on VEGFR2 activation -- perhaps by sequestering excess VEGF ligand (Roberts et al. 2004).

We previously showed that a majority of DLBCLs coordinately express VEGF, VEGFR1, and VEGFR2 (Gratzinger et al. 2007). Upon binding VEGF, VEGFR2 becomes phosphorylated on a number of cytoplasmic tyrosine residues, prominently including Y1175 and Y1214. The functional significance of the tyrosines differs; for example Y1175 is crucial for vasculogenesis and hematopoiesis while Y1214 appears to be dispensible (Sakurai et al. 2005). Immunohistochemistry was performed with a commercially available phosphospecific antibody directed against Y1175 and it was found that a subset (13%) of DLBCLs express detectable phosphoVEGFR2. Previous studies of phosphoVEGFR2 expression in DLBCL have been published by Stewart et al. (2003) using a monoclonal antibody they had raised against Y1214. A recently published study from this group (Giatromanolaki et al. 2008) reported significantly higher rates of phosphoVEGFR2 positivity: 78% with cytoplasmic staining and 16% with nuclear staining. There are several pertinent differences between the two studies: we employed a commercially available monoclonal antibody against Y1175 at a dilution of 1:30, while the other group used an antibody raised in their laboratory against Y1214 at a dilution of 1:2. It is possible that, due to the antibody used or the staining conditions, we lacked the sensitivity to detect lower levels of cytoplasmic reactivity; alternately, there could be a true difference in the phosphorylation at these two residues in DLBCL.

We previously studied VEGF and VEGF receptor expression as well as microvessel density in patients treated with anthracycline-based chemotherapy (CHOP) in the pre-rituximab era. In this cohort, high expression of VEGFR1 by lymphoma cells correlated with improved OS and microvessel density correlated with poorer OS (Gratzinger et al. 2008). Rituximab in combination with anthracycline-based chemotherapy (R-CHOP) has markedly improved outcomes in DLBCL and has necessitated re-evaluation of the prognostic significance of biomarkers. Indeed, in contrast to CHOP-treated patients, R-CHOP-treated patients with increased microvessel density had no worse survival as a group than those with low microvessel density. The mechanism by which addition of rituximab apparently abrogates the negative prognostic import of higher microvessel density is unclear. By contrast, VEGFR2 correlates with poorer OS in R-CHOP treated patients, and this correlation is independent of IPI on multivariate analysis [p=0.0028, HR=2.5(1.5–4.6)]. While the correlation of VEGFR2 with poorer OS is most evident in patients with high IPI on subgroup analysis (see Fig 3), a test for interaction of IPI category (low, 0–2 versus high, 3–5) with VEGFR2 was non-significant on Cox regression analysis. This does not support true subgroup heterogeneity, and it is likely that the correlation of VEGFR2 with outcome is simply harder to demonstrate statistically in the low IPI subgroup due to the better overall outcomes in this group of patients.

VEGFR1 and VEGFR2 expression was tightly positively correlated (p=1×10−11), in line with our previous studies. This highly reproducible coordinate upregulation of VEGFR1 and VEGFR2 suggests common upstream regulation and possibly functional interaction, yet makes it more difficult to evaluate the association of each individual VEGFR with outcome variables. In CHOP-treated patients, VEGFR1 had been associated with longer OS; univariate analysis of the R-CHOP dataset, however, did not reveal a statistically significant association of VEGFR1 with OS. Multivariate analysis including VEGFR1, VEGFR2, VEGF and IPI unmasked opposing associations of the two VEGFRs with OS. VEGFR2 remained associated with shorter OS on multivariate analysis [p=0.009, HR=2.9 (1.3–6.4)], while VEGFR1 was associated with longer OS [p=0.036, HR=0.51(0.27–0.96)]. When patients were stratified by VEGFR2 expression status, failure to express VEGFR1 was associated with significantly worse OS in the weak VEGFR2 subgroup only. By contrast VEGFR1 was not associated with OS in the subgroup with no VEGFR2 expression; these patients already had 100% OS within the follow-up interval, so no differences could be detected. VEGFR1 was also not associated with OS in the subgroup with high VEGFR2 expression, as all lymphomas coexpressed VEGFR1 and VEGFR2 in this group. The fact that VEGFR1 was only correlated with better OS in the weak VEGFR2 subgroup (see Fig 5) is thus not necessarily attributable to heterogeneity of effect within different VEGFR2 categories, but rather due to the inability to detect such an effect in the no VEGFR2 category (due to 100% survival) and the high VEGFR2 category (due to coexpression of VEGFR1 in all cases). This is underscored by a lack of interaction between VEGFR1 and VEGFR2 on Cox regression analysis.

The opposite prognostic significance of VEGFR1 and VEGFR2 with respect to OS suggests that they may functionally oppose each other in lymphoma cells. While it is possible that such functional opposition occurs at the level of downstream signaling cascades, it is also plausible that VEGFR1 simply functions to sequester VEGF, decreasing its availability to trigger VEGFR2-mediated signaling. This model is attractive both because it already has a well-known precedent in the context of the developing vasculature (Roberts et al. 2004) and because it could accommodate two other findings of this study. First, one would expect that if functional VEGFR-mediated signaling were important to lymphoma biology in a clinically relevant manner, that lymphoma cell expression of VEGF would show some association with survival. This has not been the case in either our previous study of CHOP-treated patients, or in the current R-CHOP cohort. If a large fraction of VEGF were non-productively sequestered, however, then overall lymphoma cell expression of VEGF as assessed by immunohistochemistry would not correlate with downstream events such as VEGFR activation and, ultimately, clinical outcomes. Furthermore, sequestration of VEGF would also help explain why, despite the coexpression of VEGF and VEGFR2 in the majority of cases, only 13% of lymphomas had immunohistochemically detectable phosphoVEGFR2. The low percentage of cases with phosphoVEGFR2 limits the power of this study to detect a difference in outcomes; nevertheless a trend toward shorter OS [p=0.1, HR=1.7(0.9–3.3)], and significantly poorer PFS [p=0.044, HR=1.7(1.0–2.9)] was present. Thus, not just lymphoma cell VEGFR2 expression, but functional activation of VEGFR2, is associated with shorter OS in DLBCL.

Finally, a note of caution regarding the p-values reported in this study. Multiple comparisons were made involving five variables (VEGF, VEGFR1, VEGFR2, phosphoVEGFR2, and microvessel density) and two endpoints (OS and PFS) for a total of 10 comparisons. While unadjusted p-values are generally reported, a conservative approach (Bonferroni correction) would be to adjust the p-values by a factor of 10 to adjust for multiple comparisons. Of the p-values reported, the association of VEGFR2 with OS remains statistically significant even when multiplied by 10 (adjusted p=0.028 for VEGFR2); the p-values associated with VEGFR1 and phosphoVEGFR2 would lose their statistical significance.

Acknowledgments

Y.N. is supported by National Institutes of Health (NIH, Bethesda, MD) CA109335. I.S.L. is supported by NIH CA109335, NIH CA122105, Fidelity Foundation and the Dwoskin Family Foundation (Miami, FL).

References

  1. Bellamy WT, Richter L, Frutiger Y, Grogan TM. Expression of vascular endothelial growth factor and its receptors in hematopoietic malignancies. Cancer Res. 1999;59:728–733. [PubMed] [Google Scholar]
  2. Cheson BD, Horning SJ, Coiffier B, Shipp MA, Fisher RI, Connors JM, Lister TA, Vose J, Grillo-Lopez A, Hagenbeek A, Cabanillas F, Klippensten D, Hiddemann W, Castellino R, Harris NL, Armitage JO, Carter W, Hoppe R, Canellos GP NCI Sponsored International Working Group. Report of an international workshop to standardize response criteria for non-Hodgkin's lymphomas. J Clin Oncol. 1999;17:1244–1244. doi: 10.1200/JCO.1999.17.4.1244. [DOI] [PubMed] [Google Scholar]
  3. Cross MJ, Dixelius J, Matsumoto T, Claesson-Welsh L. VEGF-receptor signal transduction. Trends Biochem Sci. 2003;28:488–494. doi: 10.1016/S0968-0004(03)00193-2. [DOI] [PubMed] [Google Scholar]
  4. Ganjoo KN, An CS, Robertson MJ, Gordon LI, Sen JA, Weisenbach J, Li S, Weller EA, Orazi A, Horning SJ. Rituximab, bevacizumab and CHOP (RA-CHOP) in untreated diffuse large B-cell lymphoma: safety, biomarker and pharmacokinetic analysis. Leuk Lymphoma. 2006;47:998–1005. doi: 10.1080/10428190600563821. [DOI] [PubMed] [Google Scholar]
  5. Gerber HP, Malik AK, Solar GP, Sherman D, Liang XH, Meng G, Hong K, Marsters JC, Ferrara N. VEGF regulates haematopoietic stem cell survival by an internal autocrine loop mechanism. Nature. 2002;417:954–958. doi: 10.1038/nature00821. [DOI] [PubMed] [Google Scholar]
  6. Giatromanolaki A, Koukourakis MI, Pezzella F, Sivridis E, Turley H, Harris AL, Gatter KC. Phosphorylated VEGFR2/KDR receptors are widely expressed in B-cell non-Hodgkin's lymphomas and correlate with hypoxia inducible factor activation. Hematol Oncol. 2008;26:219–224. doi: 10.1002/hon.861. [DOI] [PubMed] [Google Scholar]
  7. Gratzinger D, Zhao S, Marinelli RJ, Kapp AV, Tibshirani RJ, Hammer AS, Hamilton-Dutoit S, Natkunam Y. Microvessel density and expression of vascular endothelial growth factor and its receptors in diffuse large B-cell lymphoma subtypes. Am J Pathol. 2007;170:1362–1369. doi: 10.2353/ajpath.2007.060901. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Gratzinger D, Zhao S, Tibshirani RJ, Hsi ED, Hans CP, Pohlman B, Bast M, Avigdor A, Schiby G, Nagler A, Byrne GE, Jr, Lossos IS, Natkunam Y. Prognostic significance of VEGF, VEGF receptors, and microvessel density in diffuse large B cell lymphoma treated with anthracycline-based chemotherapy. Lab Invest. 2008;88:38–47. doi: 10.1038/labinvest.3700697. [DOI] [PubMed] [Google Scholar]
  9. Habermann TM, Weller EA, Morrison VA, Gascoyne RD, Cassileth PA, Cohn JB, Dakhil SR, Woda B, Fisher RI, Peterson BA, Horning SJ. Rituximab-CHOP versus CHOP alone or with maintenance rituximab in older patients with diffuse large B-cell lymphoma. J Clin Oncol. 2006;24:3121–3127. doi: 10.1200/JCO.2005.05.1003. [DOI] [PubMed] [Google Scholar]
  10. Jaffe ES, Harris NL, Stein H, Vardiman JW. Pathology and genetics of tumors of the hematopoietic and lymphoid tissues. Lyon: IARC Press; 2001. [Google Scholar]
  11. Kabbinavar F, Hurwitz HI, Fehrenbacher L, Meropol NJ, Novotny WF, Lieberman G, Griffing S, Bergsland E. 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;21:60–65. doi: 10.1200/JCO.2003.10.066. [DOI] [PubMed] [Google Scholar]
  12. Mounier N, Briere J, Gisselbrecht C, Emile JF, Lederlin P, Sebban C, Berger F, Bosly A, Morel P, Tilly H, Bouabdallah R, Reyes F, Gaulard P, Coiffier B. Rituximab plus CHOP (R-CHOP) overcomes bcl-2--associated resistance to chemotherapy in elderly patients with diffuse large B-cell lymphoma (DLBCL) Blood. 2003;101:4279–4284. doi: 10.1182/blood-2002-11-3442. [DOI] [PubMed] [Google Scholar]
  13. Natkunam Y, Lossos IS, Taidi B, Zhao S, Lu X, Ding F, Hammer AS, Marafioti T, Byrne GE, Jr, Levy S, Warnke RA, Levy R. Expression of the human germinal center-associated lymphoma (HGAL) protein, a new marker of germinal center B-cell derivation. Blood. 2005;105:3979–3986. doi: 10.1182/blood-2004-08-3112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Natkunam Y, Farinha P, Hsi ED, Hans CP, Tibshirani R, Sehn LH, Connors JM, Gratzinger D, Rosado M, Zhao S, Pohlman B, Wongchaowart N, Bast M, Avigdor A, Schiby G, Nagler A, Byrne GE, Levy R, Gascoyne RD, Lossos IS. LMO2 protein expression predicts survival in patients with diffuse large B-cell lymphoma treated with anthracycline-based chemotherapy with and without rituximab. J Clin Oncol. 2008;26:447–454. doi: 10.1200/JCO.2007.13.0690. [DOI] [PubMed] [Google Scholar]
  15. Olsson AK, Dimberg A, Kreuger J, Claesson-Welsh L. VEGF receptor signalling - in control of vascular function. Nat Rev Mol Cell Biol. 2006;7:359–371. doi: 10.1038/nrm1911. [DOI] [PubMed] [Google Scholar]
  16. Pfreundschuh M, Trumper L, Osterborg A, Pettengell R, Trneny M, Imrie K, Ma D, Gill D, Walewski J, Zinzani PL, Stahel R, Kvaloy S, Shpilberg O, Jaeger U, Hansen M, Lehtinen T, Lopez-Guillermo A, Corrado C, Scheliga A, Milpied N, Mendila M, Rashford M, Kuhnt E, Loeffler M. CHOP-like chemotherapy plus rituximab versus CHOP-like chemotherapy alone in young patients with good-prognosis diffuse large-B-cell lymphoma: a randomized controlled trial by the MabThera International Trial (MInT) Group. Lancet Oncol. 2006;7:379–391. doi: 10.1016/S1470-2045(06)70664-7. [DOI] [PubMed] [Google Scholar]
  17. Podar K, Anderson KC. The pathophysiologic role of VEGF in hematologic malignancies: therapeutic implications. Blood. 2005;105:1383–1395. doi: 10.1182/blood-2004-07-2909. [DOI] [PubMed] [Google Scholar]
  18. Rahimi N, Golde TE, Meyer RD. Identification of ligand-induced proteolytic cleavage and ectodomain shedding of VEGFR-1/FLT1 in leukemic cancer cells. Cancer Res. 2009;69:2607–2614. doi: 10.1158/0008-5472.CAN-08-2905. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Roberts DM, Kearney JB, Johnson JH, Rosenberg MP, Kumar R, Bautch VL. The vascular endothelial growth factor (VEGF) receptor Flt-1 (VEGFR-1) modulates Flk-1 (VEGFR-2) signaling during blood vessel formation. Am J Pathol. 2004;164:1531–1535. doi: 10.1016/S0002-9440(10)63711-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Sakurai Y, Ohgimoto K, Kataoka Y, Yoshida N, Shibuya M. Essential role of Flk-1 (VEGF receptor 2) tyrosine residue 1173 in vasculogenesis in mice. Proc Natl Acad Sci U S A. 2005;102:1076–1081. doi: 10.1073/pnas.0404984102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Sehn LH, Donaldson J, Chhanabhai M, Fitzgerald C, Gill K, Klasa R, MacPherson N, O'Reilly S, Spinelli JJ, Sutherland J, Wilson KS, Gascoyne RD, Connors JM. Introduction of combined CHOP plus rituximab therapy dramatically improved outcome of diffuse large B-cell lymphoma in British Columbia. J Clin Oncol. 2005;23:5027–5033. doi: 10.1200/JCO.2005.09.137. [DOI] [PubMed] [Google Scholar]
  22. Shipp MA, Harrington DP, Anderson JR, Armitage JO, Bonadonna G, Brittinger G, Cabanillas F, Canellos GP, Coiffier B, Connors JM, Cowan RA, Crowther D, Dahlberg S, Engelhard M, Fisher RI, Gisselbrecht C, Horning SJ, Lepage E, Lister TA, Meerwaldt JH, Montserrat E, Nissen NI, Oken MM, Peterson BA, Tondini C, Velasquez WA, Yeap BY. A PREDICTIVE MODEL FOR AGGRESSIVE NON-HODGKINS-LYMPHOMA. New England Journal of Medicine. 1993;329:987–994. [Google Scholar]
  23. Stewart M, Turley H, Cook N, Pezzella F, Pillai G, Ogilvie D, Cartlidge S, Paterson D, Copley C, Kendrew J, Barnes C, Harris AL, Gatter KC. The angiogenic receptor KDR is widely distributed in human tissues and tumours and relocates intracellularly on phosphorylation. An immunohistochemical study. Histopathology. 2003;43:33–39. doi: 10.1046/j.1365-2559.2003.01644.x. [DOI] [PubMed] [Google Scholar]
  24. Stopeck AT, Unger JM, Rimsza LM, Bellamy WT, Iannone M, Persky DO, Leblanc M, Fisher RI, Miller TP. A phase II trial of single agent bevacizumab in patients with relapsed, aggressive non-Hodgkin lymphoma:Southwest oncology group study S0108. Leuk Lymphoma. 2009;50:728–735. doi: 10.1080/10428190902856808. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Wang ES, Teruya-Feldstein J, Wu Y, Zhu Z, Hicklin DJ, Moore MA. Targeting autocrine and paracrine VEGF receptor pathways inhibits human lymphoma xenografts in vivo. Blood. 2004;104:2893–2902. doi: 10.1182/blood-2004-01-0226. [DOI] [PubMed] [Google Scholar]
  26. Weidner N, Semple JP, Welch WR, Folkman J. Tumor angiogenesis and metastasis--correlation in invasive breast carcinoma. N Engl J Med. 1991;324:1–8. doi: 10.1056/NEJM199101033240101. [DOI] [PubMed] [Google Scholar]

RESOURCES