| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH |
|
|
|
|||||||
|
|||||||||
Electronic Letters to:
|
|
Electronic letters published:
![[Read eLetter]](/icons/misc/sectionDOWN.gif){kind=icon}
Bevacizumab and Murine VEGF-A
Author Response: Bevacizumab and Murine VEGF-A
|
|
|||
|
Claus Cursiefen
Send letter to journal:
claus.cursiefen{at}uk-erlangen.de Claus Cursiefen
|
With respect to the recently published article by Yu et al., "Interaction between bevacizumab and murine VEGF-A: a reassessment,"1 we first would like to thank the authors for their strong interest in our article "Bevacizumab as a potent inhibitor of inflammatory corneal angiogenesis and lymphangiogenesis".2 Most importantly, we were pleased to notice that Yu et al. were able to support our finding that bevacizumab binds to murine VEGF, albeit with low affinity. We also fully agree with Yu et al. that Avastin, given its low affinity to murine VEGF-A, is not the ideal instrument to neutralize VEGF-A in murine experiments. Nevertheless, in our work, we were able to demonstrate an antihem- and antilymphangiogenic effect of bevacizumab in a murine model of suture-induced inflammatory corneal neovascularization in vivo. In that model, both the ingrowth of blood and lymphatic vessels into the normally avascular cornea is stimulated by placing three interrupted sutures intrastromally. Due to the transparent and avascular nature of the normal cornea, this assay enables a very sensitive detection of blood and lymphatic vessels with CD31 as panendothelial and LYVE-1 as specific lymphendothelial marker. To measure the effects on the outgrowth of blood and lymphatic vessels into the cornea, we used a semi-automatic, quantitative method, which is very sensitive and allowed the detection of the small but significant antihem- and antilymphangiogenic effect of bevacizumab in the murine cornea. To reassess that effect of bevacizumab in vivo, Yu et al. used a laser-induced choroidal neovascularization model and tumor experiments implanting B16F1 melanoma cells in nude mice. The above-mentioned low-binding affinity of bevacizumab to murine VEGF-A may explain why they do not see effects on their tumor and choroidal neovascularization experiments (also using less sophisticated image analysis methods). In order to assess the binding of bevacizumab to murine VEGF-A, we used several different approaches to analyze the protein-protein interaction. The assays were designed to measure the proposed low affinity of bevacizumab to murine VEGF-A in different orientations (ELISA with immobilized VEGF, BIAcore with immobilized bevacizumab) to exclude any unspecific interaction due to surface-mediated artifacts. In both assays, a specific but obviously significantly weaker interaction than to human VEGF-A was detected. Additionally, to avoid any overinterpretation of binding events due to unspecific interaction in all assays, immunoglobulin isotype controls were used as negative control. Yu et al. criticized the high coating density for the BIAcore experiments (12,000 RU in our work2 compared to 1,000 RU in the work of Yu et al.1). This condition was specifically chosen to detect the proposed low affinity interaction with the less sensitive BIAcoreX system. Due to the high coating density a kinetic evaluation of data (calculation of dissociation constant) would be not scientifically valid. This approach was merely selected to assess the interaction qualitatively.3,4 The previous BIAcore study5 with Avastin referred to by Yu et al. was performed under different conditions: in that study, VEGF was immobilized and bevacizumab the soluble component. It is well known that chemical coupling to surfaces can affect the coated proteins. Additionally, we used higher protein concentrations to detect low affinity interaction. We agree with Yu et al. that the ELISA assay shows also a specific but significantly weaker interaction of bevacizumab to murine VEGF-A than to human VEGF-A. Also, this assay was designed to characterize the interaction qualitatively, to compare the human and murine VEGF-A binding of bevacizumab with an isotype-matched negative control. Due to the low binding affinity, an EC50 quantification would be also scientifically questionable. Our recent findings about the antiangiogenic effects of bevacizumab in rodent systems are supported by other recent papers describing measurable effects of this drug in the rodent model.6-10 Heiduschka et al.11 recently investigated the effect of bevacizumab on the retinal function and were able to demonstrate binding of bevacizumab to murine VEGF-A by immunohistochemistry. Gérard et al.12 demonstrated that administration of bevacizumab induced a significant decrease in the relative volume of vessels in the context of expansion of thyroid microvasculature is the earliest event during goiter formation in mice. Our main point was to demonstrate for the first time that bevacizumab is not only an inhibitor of hemangiogenesis but also of lymphangiogenesis. This novel antilymphangiogenic effect of bevacizumab is important for the growing and very successful off-label use of Avastin eye drops in patients with progressive corneal neovascularization in the context of corneal transplantation, but also in cancer patients.13-15 In summary, we agree with Yu et al. that bevacizumab - binds to murine VEGF-A, Most importantly, our experimental data suggest that topical application of Avastin eye drops in patients (with much higher affinity of bevacizumab to human VEGF-A) not only inhibit hem- but also lymphangiogenesis. That has important implications for the use of Avastin in cancer patients (interference with lymphatic metastasis) and for the off-label use of bevacizumab eye drops in the treatment of pre- and postoperative neovascularization in the context of corneal transplantation.13-15 Felix Bock1 1Department of Ophthalmology, University of Erlangen-Nürnberg,
Erlangen, Germany References 1. Yu L, Wu X, Cheng Z, et al. Interaction between
bevacizumab and murine VEGF-A: a reassessment. Invest Ophthalmol Vis Sci. 2008;49:522-527. |
|||
|
|
|||
|
Napoleone Ferrara
Send letter to journal:
nf{at}gene.com Napoleone Ferrara
|
We thank Bock et al. for their interest in our manuscript.1 We appreciate their points, but we are frankly surprised that they persist in the claim that bevacizumab "binds" murine VEGF-A. According to the authors' own ELISA data,2 the binding of bevacizumab to murine VEGF-A amounts to <0.1% of the binding of bevacizumab to human VEGF-A. In fact, Bock et al. now acknowledge in their letter that the binding is so weak that they could not generate any plausible EC50 or dissociation constant values by ELISA or BIAcore. Bevacizumab fails to inhibit the activity of murine VEGF-A even at a vast molar excess.1 Neutralization of VEGF-A activities is the basis of the therapeutic effects of bevacizumab.3 Therefore, we conclude that the binding of bevacizumab to murine VEGF-A is practically negligible. We thank the authors for specifying the amounts of immobilized bevacizumab in their BIAcore experiments and providing further methodological details. However, 12,000 RU reflect a very large amount of protein, confirming and even heightening our earlier concerns.1 The authors imply in their manuscript2 and even in their letter that showing some qualitative binding, no matter how weak, is enough to provide a rationale to undertake or interpret pharmacological experiments in animals. We are confident that this view is far from correct. It is well established that immobilized proteins, especially at high concentrations, may display interactions that are not biologically relevant.4,5 Therefore, it is imperative to validate a putative antigen-antibody interaction not only by quantitative binding analysis but, importantly, also through function blocking studies, verifying that the antibody inhibits its target in a specific and dose-dependent manner. In the absence of such evidences, we do not understand how Bock et al. can conclude that their findings in the corneal suture model2 reflect a specific neutralization of murine VEGF-A by bevacizumab. It is well known that animal models may display considerable variability and that the experimental results can be affected by a variety of factors unrelated to a specific interaction between the drug and the target, potentially leading to inconsistent and controversial findings. Only a rigorous and quantitative approach, employing appropriate controls and verifying that the drug effectively inhibits its target at the dose tested, will ensure that the results are meaningful and can be correctly interpreted.6 It seems to us that testing a single, arbitrarily selected, dosing regimen of a pharmacological agent, whose interaction with its intended target is weak to the point of being unquantifiable, does not meet such criteria. It is difficult to comment on the two studies cited in the letter by Bock et al. in support of their claims. One could raise some issues, which are similar to those that we raised regarding the paper by Bock et al.2 On the other hand, we would like to point the authors' attention to a very recent study7 from the laboratory of Ronald Crystal, describing an adenoviral vector encoding the cDNAs for the light and heavy chains of Mab A.4.6.1, the murine precursor of bevacizumab. Mab A.4.6.1 and bevacizumab have equivalent binding characteristics.3 The authors performed all the appropriate controls, and their data clearly show that the antibody binds to the 121- and 165-amino acid isoforms of human VEGF-A but not to murine VEGF-A.7 We are encouraged by the following statement in the letter by Bock et al.: "We also fully agree with Yu et al. that due to the low affinity of Avastin to murine VEGF-A, in animal experiments, other more potent inhibitors should be used wherever feasible." At least it is a step in the right direction from their initial claim that bevacizumab is a potent inhibitor of inflammatory corneal angiogenesis and lymphangiogenesis in the mouse.2 Lanlan Yu Genentech, Inc, South San Francisco, California References 1. Yu L, Wu X, Cheng Z, et al. Interaction between bevacizumab and murine VEGF-A: a reassessment. Invest Ophthalmol Vis Sci. 2008;49:522-527. |
|||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH |
eLetters: