Method for in vivo imaging of lymph node lymphangiogenesis by Immuno-Positron Emission Tomography and markers therefore

TECHNICAL FIELD

The invention relates to methods for in vivo imaging of the lymphatic system such as lymph nodes, e.g. for detection and/or monitoring of lymphangiogenesis, using Immuno- Positron Emission Tomography and other imaging methods, to markers for these imaging methods and to a procedure for obtaining such marker systems.

PRIOR ART

Metastatic spread is a characteristic trait of most tumors types and is the cause for the majority of cancer deaths. In many human cancers, metastasis to the regional lymphatic system, i.e. in particular the lymph nodes, is the first step of tumor dissemination and a prognostic indicator for the progression of the disease. The presence of tumor cells in the lymph nodes is a major determinant for the clinical management of cancer patients. Currently, regional lymph nodes, or in cases of breast cancer and melanoma patients only the tumor draining (sentinel) lymph nodes, are invasively analyzed, i.e. are dissected and sections are analyzed for metastases. However, this procedure is elaborate and associated with significant morbidity and costs and raises the demand for a sensitive, non-invasive, and simpler method to detect metastasis to the lymph nodes.

SUMMARY OF THE INVENTION Detecting tumor-induced changes in the stroma can be more sensitive and less laborious in detecting cancer metastasis than to look for the cancer cells themselves. We and others have found that tumors induce the expansion of the lymphatic vasculature (lymphangiogenesis) in tumor draining lymph nodes. Most importantly, primary cancers were shown to induce lymphangiogenesis within the sentinel lymph nodes even before the on-set of metastasis. Once the metastases arrived in the lymph nodes, lymphangiogenesis was further enhanced and promoted metastasis to distant lymph nodes and organs. Expanded lymphatic networks were also found in the metastatic lymph nodes of human melanoma patients and in metastatic lymph nodes of human breast cancer patients, correlating with distant metastasis. These studies show that lymph node lymphangiogenesis is a significant process in humans.

The factors driving lymphangiogenesis identified today are presumed to be vascular endothelial growth factor VEGF-A and VEGF-C, amongst others, that are produced by tumor and tumor-associated stromal ceils. These factors are drained into the lymph nodes or are produced in there once the cancer cells have arrived.

Lymph node lymphangiogenesis can be used as a prognostic indicator to screen for cancer metastasis. Therefore the aim of this invention is to establish methods to image this process non-invasively in vivo using methods such as positron emission tomography (PET), single photon emission computed tomography (SPECT), fluorescence-mediated-tomography (FMT), magnetic resonance imaging (MRI), ultrasound, bioluminescence imaging, and the like. The non-invasive in vivo imaging is carried out using labelled antibodies (or optionally also labeled non-antibody binders) to lymphatic specific epitopes. Labelled means in this context that the labels attached to and/or integrated into the antibodies are able to facilitate and/or make possible imaging using any of the above- mentioned methods.

In case of the use of PET imaging, correspondingly radiolabeled antibodies or other non- antibody binders are used. Radiolabelling in this context means that a system is attached or integrated into the antibody which emits positrons. PET imaging visualizes the distribution of a positron (e ⁺ ) emitting radiotracer administered to an animal or a person in three- dimensions at very high sensitivity. Emitted positrons move through the tissue until they have lost enough energy to react with a tissue electron. The photons resulting from the annihilation process of positrons and electrons are then detected and used for image reconstruction. There are additional methods that could be used to image the accumulation of an antibody to a lymphatic marker in the lymph nodes non-invasively in vivo. These are:

Single photon emission computed tomography (SPECT), if the antibody is coupled to a radionuclide applicable for SPECT.

Fluorescence-mediated-tomography (FMT), if the antibody is coupled to an applicable fluorophor for FMT. Magnetic resonance imaging (MRI), if the antibody is coupled to a suitable MRI-contrast agent.

Ultrasound, if the antibody is linked to a suitable ultrasound active contrast agent such as microbubbles. Bioluminescence imaging, if the antibody is coupled to a suitable molecule able to produce bioluminescence.

The object of the present invention is therefore a method for non-invasive in vivo imaging of the lymphatic system, comprising the steps of delivery of at least one labeled antibody to or labeled non-antibody binder to at least one non-intracellular, surface accessible protein present in or expressed on lymphatic vessels, or in the extracellular matrix of lymphatic vessels, to a patient; allowing the at least one labeled antibody or labeled non-antibody binder to distribute and eventually accumulate in the lymphatic system (it is noted that during this time span further and continued delivery of the labelled antibody or labelled non-antibody binder is possible); acquiring at least one image associated with the label of the labeled antibody or labeled non-antibody binder of the lymphatic system. Preferentially labeled antibody or labeled non-antibody binder should be selected such that it does not inhibit lymphangiogenesis to avoid interference with the object of observation. Preferentially a system is chosen as the protein for the antibody or non-antibody binder which has a redundant function. On the other hand it is possible to intentionally select a labeled antibody or labeled non-antibody binder which has a treatment effect and monitor it efficacy directly. Imaging or data acquisition for image generation can be carried out on a single (two- dimensional or three-dimensional) picture basis but also the time development can be analysed, be it under continuing delivery of the labelled antibody or labelled non-antibody binder or after a single shot delivery of such a system. In this context, the expression antibody shall be meant to include (for example IgG but also IgM) antibodies to image lymph node lymphangiogenesis, other antibody isotypes or formats, e.g. small immuno proteins, single chain variable fragment antibodies (scFv), diabodies or minbodies etc.

According to a first preferred embodiment of the invention, the delivery of the of at least one labeled antibody or labeled non-antibody binder is carried out by, subcutaneous, , intravenous, and/or intraperitoneal injection, preferably by intravenous injection. Typically delivery is carried out by intravenous injection of the labeled antibody or labeled non- antibody binder. The at least one labeled antibody to or labeled non-antibody binder can be labeled with a radiolabel and the image can be acquired using positron emission tomography. In the alternative, the at least one labeled antibody or labeled non-antibody binder can be labeled with a radionuclide applicable for single photon emission computed tomography and the image can be acquired using single photon emission computed tomography. So single photon emission computed tomography (SPECT) can be used, if the antibody is coupled to a radionuclide applicable for SPECT.

In yet another alternative, the at least one labeled antibody to or labeled non-antibody binder can be labeled with a fluorophor applicable for fluorescence-mediated-tomography and the image can be acquired using fluorescence-mediated-tomography. So fluorescence- mediated-tomography (FMT) can be used, if the antibody is coupled to an applicable fluorophor for FMT.

In yet another alternative embodiment, the at least one labeled antibody or labeled non- antibody binder can be labeled with a magnetic resonance imaging contrast agent and the image can be acquired using magnetic resonance imaging. So magnetic resonance imaging (MRI) can be used, if the antibody is coupled to a suitable MRI-contrast agent.

In yet another alternative embodiment, the at least one labeled antibody or labeled non- antibody binder can be labeled with a ultrasound imaging contrast agent and the image can be acquired using ultrasound imaging. So ultrasound imaging can be used, if the antibody is linked to a suitable system such as a microbubble. In yet another alternative embodiment, the at least one labeled antibody or labeled non- antibody binder can be labeled with a system able to produce bioluminescence and the image can be acquired using bioluminescence imaging. So bioluminescence imaging can be used, if the antibody is coupled to a suitable molecule able to produce bioluminescence. According to yet another preferred embodiment of the proposed method, the at least one labeled antibody or labeled non-antibody binder to a protein is one to a protein expressed on the cell surface membrane by lymphatic endothelial cells or which is enriched in the extracellular matrix, and which preferably is essentially not or only weakly expressed on B or T cells or tissues adjacent to the lymph nodes. Preferentially the step of allowing the at least one labeled antibody or labeled non-antibody binder to distribute and eventually accumulate in the lymphatic system is carried out for at least 1 h, preferably for at least 2 or at least 3 h, even more preferably for at least 12 h. Generally and preferentially, the step of acquiring at least one image associated with the label of the labeled antibody or labeled non-antibody binder of the lymphatic system is carried out in a time window of 1-6Oh, preferably in a time window of 15-50 h, after initiation of delivery of at least one labeled antibody to or labeled non-antibody binder to at least one non-intracellular, surface accessible protein expressed on lymphatic vessels, or secreted into the extracellular matrix by lymphatic endothelial cells to a patient. The at least one labeled antibody can be a ¹⁸ F, ⁴⁵ Ti, ⁵² Fe, ⁵⁵ Co, ⁶¹ Cu, ⁶² Zn, ⁶⁴ Cu, ⁶⁸ Ga, ⁷¹ As,

⁷⁴ As, ⁷⁶ Br, ⁷⁷ Br, ⁸⁶ Y, ⁸⁹ Zr, ¹¹⁰ In, ¹²⁴ I, ¹²² Xe labeled antibody or non-antibody binder to a protein expressed by lymphatic endothelial cells and the image is acquired using positron emission tomography. Preferred are systems labelled with the given iodine isotope or systems like ⁸⁹ Zr ⁶⁴ Cu which offer high resolution. As concerns possible candidate proteins for the antibody or non-antibody binder system, the non-intracellular, surface accessible protein expressed on lymphatic vessels, or secreted into the extracellular matrix by lymphatic endothelial cells can be preferentially selected from the following group:

LYVE-I, Activin A receptor, type 1 (ACVRl), Collagen type IV Al (COL4A1), Collagen type IV A2 (COL4A2), EPH receptor Bl (EPHBl), Insulin-like growth factor 1 receptor,

Matrix metalloproteinase 13 (MMP- 13), Neogenin, Plasminogen activator urokinase

(PLAU), Semaphorin 3A (SEMA 3A), Tenascin-C (TNC).

Further molecules expressed by lymphatic endothelial cells in vitro or in vivo, which can be used as targets to image lymph node lymphangiogenesis are as follows: Aquaporin-1 (Gannon, B.J., and CJ. Carati. 2003. Lymphatic research and biology 1 :55-

66; Verkman, A.S. 2006. Kidney international 69:1120-1123),

Ajuba (Clasper, S., D. et al. 2008. Cancer research 68:7293-7303),

B-chemokine receptor D6 (Nibbs, R.J. et al 2001. The American journal of pathology

158:867-877), Carcinoembryonic antigen CEA-CAM (Hirakawa, S., et al, 2003, The American journal of pathology 162:575-586),

Coxsackie-and adenovirus receptor CAR (Vigl, B., et al. 2009, Experimental cell research

315:336-347),

CLEVER-I (Irjala, H., K. et al. 2003. European journal of immunology 33:815-824), CDl 12 and CD200 (Clasper, S., D. et al. 2008),

DC-SIGN CD209 (Martens, J.H., et al. 2006. The Journal of pathology 208:574-589),

Desmoplakin-1 (Hirakawa, S., et al. 2003),

Endoglin (Clasper, S., D. et al. 2008 ;Hirakawa, S., et al, 2003), Endothelial-specific-molecule-l ESM-I (Shin, J.W., et al. 2008. Blood 1 12:2318-2326.),

Ephrin B4 receptor (Makinen, T., et al. 2005. Genes & development 19:397-410),

Emilin-1 (Danussi, C, P., et. al 2008. Molecular and cellular biology 28 -.4026-4039),

Endothelial cell adhesion molecule (ESAM) (Clasper, S., D. et al. 2008. Cancer research 68:7293-7303, Baluk, P., et al. 2007. The Journal of experimental medicine 204:2349-

2362),

Fibroblast growth factor receptor-3 (Shin, J.W., et al. 2006. MoI Biol Cell 17:576-584),

CGRP type 1 receptor (Hirakawa S., et al. 2003. The american journal of pathology

162 :575-586), Galectin 8 (Hirakawa S., et al. 2003. The American journal of pathology 162 :575-586),

Glycine receptor beta subunit (Hirakawa S., et al. 2003. The American journal of pathology 162 :575-586),

Hepatocyte growth factor receptor (c-met) (Kajiya, K., et al. 2005. Embo J 24:2885-2895),

Integrin alpha 1 (Hong, Y.K., et al. 2004, Faseb J 18 :1111-1113), Integrin alpha 2 (Hong, Y.K., et al. 2004, Faseb J 18 :11 11-1113),

Integrin alpha 4 beta 1 (Garmy-Susini, B., et al. 2007. Methods in enzymology 426 :415-

438),

Integrin alpha 5 and alpha v (Dietrich, T., 2007. The American journal of pathology

171 :361-372), Integrin alpha 6B (Hirakawa S., et al. 2003. The American journal of pathology 162 :575-

586),

Integrin alpha 9 (Huang, X.Z., et al. 2000. Molecular and cellular biology 20:5208-5215),

Intracellular adhesion molecule 1 (ICAM-I) (Johnson, L.A., et al. 2006. The Journal of experimental medicine 203:2763-2777), Jagged-1 (Hirakawa S., et al. 2003. The American journal of pathology 162:575-586),

LA 102 (Ezaki, T., et al. 2006. Anatomy and Embryology 211 :379-393),

Leptin receptor (Clasper, S., et al. 2008. Cancer research 68:7293-7303),

Macrophage mannose receptor (Takahashi, K., et al. 1998. Cell and tissue research

292:311-323), Membrane glycoprotein gp 130 (Hirakawa S., et al. 2003. The American journal of pathology 162:575-586),

Microfibrillar associated protein 3 (MF AP3) (Hirakawa S., et al. 2003. The American journal of pathology 162:575-586), Multimerin-1 (Roesli, C, et al. 2008. Faseb J),

Neuropilin-2 (Yuan, L., et al. 2002. Development 129 :4797-4806 / Caunt, M., et al. 2008.

Cancer cell 13:331-342),

Occludin (Baluk, P., et al. 2007. Tne Journal of experimental medicine 204:2349-2362), Orphan G protein-coupled receptor (Hirakawa S., et al. 2003. The American journal of pathology 162:575-586),

Perlecan (Rutkowski, J.M., et al. 2006. American journal of physiology 291:H1402-1410),

Platelet-endothelial cell adhesion molecule-1 (Baluk, P., et al. 2008. Annals of the New

York Academy of Sciences 1131 :1-12), Podoplanin (Breiteneder-Geleff, S., et al. 1999. Verh Dtsch Ges Pathol 83:270-275 /

Schacht, V., et al. 2003. Embo J 22:3546-3556),

Plakoglobin, (Hirakawa S., et al. 2003. The American journal of pathology 162:575-586).

Proteinase-activated receptor-2 (Hirakawa S., et al. 2003. The American journal of pathology 162:575-586), Reelin (Hirakawa S., et al. 2003. The American journal of pathology 162:575-586),

Stabilin-1 (Martens, J.H., et al. 2006. The Journal of pathology 208:574-589),

Stabilin-2 (Martens, J.H., et al. 2006. The Journal of pathology 208:574-589),

Tenascin-C (Clasper, S., et al. 2008. Cancer research 68 :7293-7303),

Toll-like receptors 1, -2, -3, -4, -5, 6, -10 (Pegu, A., et al. 2008. J Immunol 180:3399- 3405),

Tyrosine kinase with immunoglobulin and epidermal growth factor homology domains 1

(Tiel) (Iljin, K., et al. 2002. Faseb J 16:1764-1774),

Tyrosine kinase with immunoglobulin and epidermal growth factor homology domains 2

(Tie 2) (Morisada, T., et al. 2005. Blood 105:4649-4656), Vascular cell adhesion molecule 1 (VCAM-I) (Johnson, L. A., et al. 2006. The Journal of experimental medicine 203 :2763-2777),

Vascular-endothelial cadherin (VE-cadherin) (Baluk, P., et al. ^' 2007. The Journal of experimental medicine 204 -.2349-2362),

Vascular endothelial growth factor receptor-3 (VEGFR-3) (Kaipainen, A., et. al. 1995. Proceedings of the National Academy of Sciences of the United States of America

92:3566-3570).

According to one embodiment, the at least one labeled antibody is a ¹²⁴ I labeled antibody to the lymphatic epitope LYVE-I and the image is acquired using positron emission tomography. In this case, normally and preferably the antibody is an IgG antibody. Furthermore preferably it is intravenously injected, the antibody is allowed to distribute and accumulate during at least 5 h, preferably at least 10 hours after initiation of injection, and the positron emission tomography image is taken subsequently. According to a further preferred embodiment, the labeled antibody or labeled non-antibody binder is delivered to the patient in an amount to saturate the lymphatic vessels. Typically, and in particular in case of using a ¹²⁴ I labeled antibody to the lymphatic epitope LYVE-I, this means intravenous injection in the range of 0.05 - 10 microgram, preferably 1 - 5 microgram, more preferably 0.2 — 3.5 microgram, antibody or labeled non-antibody binder per g bodyweight of the patient.

Furthermore the present invention relates to a method for the making of labeled antibodies or labeled non-antibody binders to at least one non-intracellular, surface accessible protein present in or expressed on lymphatic vessels, or in the extracellular matrix of lymphatic vessels preferably for use in a as described above. Generally speaking, in this method the following steps are carried out: in a first step, candidate proteins expressed by lymphatic endothelial cells are selected either from literature or by transcriptional or proteomic profiling of lymphatic endothelial cells, wherein preferably one of the candidate proteins systems as detailed above is selected, in a second step the staining patterns of antibodies or non-antibody binders to candidate proteins are evaluated on frozen tissue sections of test organs, including lymph nodes with lymphangiogenic lymphatic vessels, wherein, if high affinity monoclonal antibodies or high-affinity non-antibody binders against candidate proteins are not available, high affinity antibodies are produced preferably by antibody-phage display or hybridoma technology, in a third step antibodies or non-antibody binders that show preferential staining of lymphatic vessels or their extracellular matrix are selected and delivered to living organisms bearing lymph node lymphangiogenesis and then detected in tissue sections for evaluation whether they accumulate specifically in lymphatic vessels or their extracellular matrix compared to control antibodies or non-antibody binders, and whether they enrich in the activated lymphatic vessels in the lymph nodes, in a next step, antibodies or non-antibody binders are labeled by a imaging enabling label, e.g. by an I radionuclide, and then selected whether they maintain their binding affinity, preferably by evaluating the immunoreactivity on antigen loaded columns or by injection into a living test organism, in a next optional step, a biodistribution analysis is performed to select for and quantify the in vivo targeting performance of the labeled antibodies or non-antibody binders to candidate proteins compared to labeled control antibodies or non-antibody binders in living test organisms bearing lymph node lymphangiogenesis and antibodies or non-antibody binders are selected, the activity concentration of which to candidate molecules in the lymph node with on-going lymph node lymphangiogenesis is higher, preferably at least twice as high, as in the adjacent tissues and as well preferably twice as high as the enrichment of the injected control antibody or non-antibody binder, optionally supplemented with the analysis of sections of organs adjacent to lymph nodes with ongoing lymphangiogenesis and control lymph nodes by microradiography to evaluate the enrichment pattern in these organs.

More specifically speaking, this method, in particular but not limited to the situation of the case of positron emission tomography can be described for a more specific embodiment as follows:

In a first step, candidate proteins need to be selected. Candidate proteins are expressed by lymphatic endothelial cells (LECs), the cell type lining the lumen of lymphatic vessels. They are selected either from the literature or are identified by transcriptional or proteomic profiling of LECs as described below.

Identification of candidate genes by transcriptional profiling: LECs are isolated from mouse tissue by tissue digestion and fluorescence-activated cell sorting (FACS) as described (Halin, C, et al. 2007. Blood 110:3158-3167.). The transcriptional profile of LECs is evaluated by microarray analysis as described (Shin, J.W., et al. 2008. Blood 112:2318-2326.). Genes up-regulated on activated

(i.e. expanding) compared to resting LECs are identified by comparing the transcriptional profiles of LECs isolated from normal tissue and LECs from tissue containing activated LECs, e.g. inflamed or tumor-associated tissue. Alternatively, resting or activated LECs can also be isolated from normal tissue or tissue with activated lymphatic vessels by laser capture microdissection (LCM) as described (Burgoon, M. Pet al. 2005. Proceedings of the National Academy of Sciences of the United States of America 102:7245-7250.) and analyzed by microarray analysis. Identification of candidate proteins by proteomic profiling: The surface accessible proteome of cultured LECs can be identified by two dimensional peptide mapping as described (Roesli, C, V. et al. 2008.. Faseb J). Proteins up-regulated on activated compared to resting LECs are identified by comparing the surface accessible proteomes of LECs cultured with lymphangiogenic growth factors and LECs cultured in medium depleted of lymphangiogenic growth factors.

Candidate proteins should be surface accessible, i.e. expressed on the cell surface membrane, or be enriched in the extracellular matrix and must not be intracellular. They have to be consistently identified on LECs. Ideally, they are up-regulated on activated (i.e. expanding) compared to control LECs. By literature research, information about tissue and cell specificity is gathered to select candidate proteins whose expression is detected predominantly on lymphatic vessels. It is probable that no protein is exclusively present on activated lymphatic vessels and has a non-redundant function. Thus candidates that fit these criteria as best are selected. Candidate protein must not be strongly expressed on B or T cells or tissues adjacent to the lymph nodes that are aimed to be imaged in vivo.

In a next step, the staining patterns of antibodies to candidate proteins are evaluated on frozen tissue sections of various mouse organs, including lymph nodes with lymphangiogenic lymphatic vessels. Sections of a mouse model of tumor- or inflammation-induced lymph node lymphangiogenesis can be used (Halin et al, 2007). If there are no suitable, high affinity monoclonal antibodies against candidate proteins available, they can be produced, for instance by antibody-phage display or hybridoma technology (Kohler, G., and C. Milstein. 1975. Nature 256:495-497; McCafferty, J., 1990, Nature 348:552-554).

The staining pattern of the antibodies should be at possible restricted to lymphatic vessels or their extracellular matrix. If an antibody stains other structures than lymphatic vessels, equally well, this may indicate either off-target binding of the antibody or expression of the target aside lymphatic vessels. To address the first issue, another antibody against the target protein with a higher specificity could be tested. The candidate protein is dismissed if distinction between the additionally stained structures and the lymphatic vessels in lymph nodes is unlikely to be achieved by in vivo imaging.

Next, antibodies that show preferential staining of lymphatic vessels or their extracellular matrix are intravenously injected into mice bearing lymph node lymphangiogenesis. Again either a mouse model of tumor- or inflammation-induced lymph node lymphangiogenesis can be used (Halin et al, 2007).

Injected antibodies to the candidate protein are detected in tissue sections and it is evaluated whether they accumulated specifically in lymphatic vessels or their extracellular matrix compared to control IgG, and whether they enrich in the activated lymphatic vessels or their extracellular matrix in the lymph nodes. If this is not the case, antibody dosing and time point of observation can be adjusted. If no enrichment in the lymphatic vessels or in their extracellular matrix is achieved, the candidate protein is dismissed. In general, an antibody that stains its target well in frozen tissue sections does not imperatively bind its target as well in vivo. If an antibody stains other structures than lymphatic vessels or their extracellular matrix equally well, this may indicate either off-target binding of the antibody or expression of the target aside lymphatic vessels. Either an antibody with higher specificity for the candidate protein is tested or the candidate protein is dismissed if structures around the lymph nodes are stained so strongly that distinction between them and the lymph nodes is unlikely to be achieved by in vivo imaging, or if non-lymphatic sinusoid/lymphatic vessel structures in the lymph node are stained at an intensity that is likely to cover the signal of antibody accumulation in lymphatic vessels in in vivo imaging.

In a next step, antibodies are labeled, e.g. an I radionuclide such as radionuclide ¹²⁵ I, and then tested whether they maintain their binding affinity. This can be done by evaluating the immunoreactivity on antigen loaded columns. Alternatively, if no antigen loaded columns are available, the antibodies are injected into mice. Then tissue sections of the mice are analyzed by microradiography to analyze whether the iodinated antibodies still accumulate in the lymphatic vessels or their extracellular matrix. In case an iodinated antibody does not bind to its target anymore, the iodination protocol is adjusted or another antibody is applied.

Biodistribution analysis are performed to quantify the in vivo targeting performance of the ¹²⁵ I-lableled antibodies to candidate proteins compared to ^l25 I-labeled control antibodies in mice bearing lymph node lymphangiogenesis. The activity concentration of antibodies to candidate molecules in the lymph node with on-going lymph node lymphangiogenesis needs to be at least twice as high as in the adjacent tissues and as well twice as high as the enrichment of the injected control IgG. Preferentially, accumulation in other organs than lymph nodes should be as low as the background of injected control antibody. Sections of organs adjacent to lymph nodes with on-going lymphangiogenesis and control lymph nodes are additionally analyzed by microradiography to evaluate the enrichment pattern in these organs. If there is an antibody accumulation in these adjacent organs that could impede the distinction of these organs from the lymph node by in vivo imaging, the candidate protein is dismissed. By biodistribution analysis, optimal antibody doses and time points for imaging are evaluated. Optimal doses and time points of imaging are characterized by target saturation in the lymph nodes with on-going lymphangiogenesis while other organs should have low activity concentrations.

If there is specific accumulation of antibodies in the lymph nodes with expanded lymphatic vessels, antibodies are further tested for their ability to detect lymph node lymphangiogenesis in vivo by positron emission tomography (PET) imaging. To this end, they are labeled with a radionuclide suitable for PET imaging (12). In case non-iodine radionuclides are selected, the binding affinity of the labeled antibody has to be tested as described above. Radiolabeled antibodies to candidate molecules and control antibody are injected into mice bearing lymph node lymphangiogenesis and PET imaging is performed. If specific signals at the sites of lymph nodes with expanded lymphatic networks can be visualized by PET imaging, the mice are sacrifized and re-scanned by PET with these lymph nodes placed next to them. If in this scan only the dissected lymph nodes can be visualized but not the original location of the dissected lymph nodes in the mouse, it is confirmed that lymph nodes were imaged in vivo. If in in vivo PET scans the lymph node bearing expanded lymphatic networks gives a stronger signal than control lymph nodes, it is confirmed, that lymph node lymphangiogenesis is imaged.

If lymph nodes cannot be imaged in vivo, the specific activity of the labeled antibodies may have been too low. In this case, the radiolabeling procedure can be adjusted. Furthermore, changing to a radionuclide with better resolution may help to increase contrast and resolution.

In case antibody accumulation in non-lymph node organs is so strong that it out-shines lymph nodes of interest, visualization of the lymph nodes may be achieved by lowering the injected antibody dose or if possible by saturating organs with unlabeled antibody before injection of the radiolabeled antibody. Furthermore the present invention relates to the use of an antibody or labeled non-antibody binder as determined in a method as given above for non-invasive in vivo imaging of the lymphatic system, preferably by positron emission tomography. More specifically, the invention pertains to the use of an IgG based ' I labeled antibody to the lymphatic epitope LYVE-I for non-invasive in vivo imaging with positron emission tomography.

Furthermore the invention in another embodiment relates to the use of a method as described above for imaging , preferably using a antibody or non-antibody binding system as made using a method as given above, for the detection of lymphangiogenesis, cancer and/or inflammation.

Notwithstanding the above, the invention furthermore pertains to the use of a method for imaging as given above, for the monitoring and the control (determination and/or adaptation) of the medical treatment such as cancer radioimmunotherapy of a patient. Last but not least the present invention relates to an IgG based ¹²⁴ I labeled antibody to the lymphatic epitope LYVE-I in particular for non-invasive in vivo imaging with positron emission tomography.

For the more detailed studies as outlined below, the model of inflammation-induced lymph node lymphangiogenesis in Kl 4 VEGF+/- mice was used. By chemical stimulation the ears and ear draining (auricular) lymph nodes of these mice develop a chronic inflammation accompanied by lymphangiogenesis.

In a first step, we established the proof-of-principle that systemically injected antibodies to lymphatic epitopes accumulate in the lymphatic vasculature. In a next step, we were looking for an antibody that accumulates sufficiently in the lymphatic vessels to be imaged by PET. Using such an antibody we could not only visualize lymphatic vessels in the lymph nodes but also lymph node lymphangiogenesis non-invasively in vivo. In this document we show, that lymphatic vessels can unexpectedly be targeted and imaged by antibodies and present for the first time the proof of non-invasive imaging of lymph node lymphangiogenesis in vivo. This novel method may open up a future road of cancer metastasis diagnosis and cancer radioimmunotherapy, and may image lymphangiogenesis as a biomarker for the progression of the numerous medical conditions associated with lymphangiogenesis.

Harrell et al. (Harrell et al, American Journal of Pathology, 2007, 170: 774 - 786) imaged increased lymph flow correlating with lymph node lymphangiogenesis in tumor draining lymph nodes compared to control lymph nodes in a mouse tumor model. To this end, they injected tumor bearing mice interstitially with nanoparticles that were drained into the tumor draining lymph nodes and imaged flow of the nanoparticles non-invasively in vivo by near infrared imaging. Ruddell et al. (Ruddell et al., Neoplasia, 2008, 10: 706-713) imaged increased lymph flow in tumor draining lymph nodes compared to control lymph nodes in the same mouse model as Harrell et al. They injected the mice interstitially with a contrast agent that was drained into the tumor draining lymph nodes. Following, the flow of the contrast agent was imaged non-invasively in vivo by magnetic resonance imaging (MRI).

Harrell et al. and Ruddell et al. detected lymph node lymphangiogenesis only indirectly by measuring increased lymph flow. We were imaging lymph node lymphangiogenesis directly. Furthermore they used other methods than PET for in vivo imaging, namely near infrared imaging and MRI. In the studies of Harrell et al and Ruddell et al., tracer molecules were injected interstitially. In case lymphatic vessels were clogged by metastases it would not be possible to image lymph node lymhangiogenesis because the tracer molecules would not reach the lymph nodes. In contrast, our tracer systems, molecules such as ¹²⁴ I-labeled anti- LYVE-I antibody can be injected intravenously and reach the lymphatic vessels in the lymph nodes as well by extravasation of the blood vessels inside the lymph node and not only through tissue drainage.

In the studies of Harrell et al. and Ruddell et al., imaging had to be performed right after tracer injection to enable quantification of lymph flow. Our method offers more flexibility since we can image lymph node lymphangiogenesis for at least two days after tracer injection.

Our new method of in vivo imaging of lymph node lymphangiogenesis may be applied in the following fields of medicine:

Detection of cancer metastasis. As described above, currently, sentinel or local lymph nodes of cancer patients are dissected and processed for cancer metastasis detection in an elaborate way associated with side effects (e.g. tissue edema) and significant costs. Lymph node lymphangiogenesis has been found to correlate with and even precede cancer metastasis. There-fore we proposed lymph node lymphangiogenesis to be a novel biomarker for cancer metastasis. Imaging lymph node lymphangiogenesis non-invasively in vivo in cancer patients would circumvent sentinel or local lymph node dissection and its side effects and may even be more sensitive in detecting cancer metastasis than the currently applied method. Biomarker observation. Many pathological conditions (e.g. chronic inflammation, rheumatoid arthritis, and cancer) are associated with an expansion of the lymphatic vasculature. Using our approach the expansion of the lymphatic vasculature could be imaged and by this implemented as a biomarker for the progression/regression of these conditions. - Cancer treatment. Many cancer types have been shown to induce lymphangiogenesis in and around the tumor. Radiolabeled antibodies to lymphatic endothelial cell specific epitopes like LYVE-I or molecules that are specifically up-regulated on growing lymphatic vessels may be used for radioimmunotherapy of cancers. Further embodiments of the present invention are outlined in the dependent claims.

BRIEF DESCRIPTION OF THE DRAWINGS

Preferred embodiments of the invention are described in the following with reference to the drawings, which are for the purpose of illustrating the present preferred embodiments of the invention and not for the purpose of limiting the same. In the drawings,

Fig. 1 : shows how a systemically injected radiolabeled anti-LYVE-1 antibody accumulates in the lymphatic vasculature; (A - N) show microradiographies of tissue sections of 1 ²⁵ I-anti-LYVE-l antibody- and control IgG-injected mice; the radio signal of the injected ^l25 I-anti-LYVE-l antibody (A, C, E, G, I, K, M) but not of -control IgG (B, D, F, H, J, L, N) was detected in sections of control (A, B) and inflamed auricular lymph nodes (C, D), control (E, F) and inflamed ears (G, H), lung (I, J), and intestine (M, N); In liver sections (K, L), both ¹²⁵ I-anti-LYVE-l antibody and control IgG could not be detected. (O - R) Serial sections of an inflamed auricular lymph node (O, P) and ear (Q, R) of an ¹²⁵ I-anti-LYVE-l injected mouse. The radioactive signal of the ¹²⁵ I-anti-LYVE-l antibody (O, Q) overlapped with immunofluorescent staining for LYVE-I -positive lymphatic vessels (P, R). Scale bars = 100 μm.

Fig. 2. shows the dose dependent accumulation of ^l25 I-anti-LYVE-l antibody in the lymph nodes; biodistribution experiments of different doses of ^l25 I-anti-LYVE-l antibody or -control IgG in mice; the results are expressed as the percentage of the injected radioactivity dose per gram of tissue (%ID/g) ± SD; (A) seven microgram injected 1 ²⁵ I-anti-L YVE-I antibody resulted primarily in targeting of the lungs; (B) thirty- five microgram injected ¹²⁵ I-anti-LYVE-l antibody resulted in increased targeting of lymph nodes compared to seven microgram injected dose; (C) ninety micrograms injected ¹²⁵ I-anti-LYVE-l antibody did not further increase the targeting of the lymph nodes compared to other organs; enhanced accumulation of antibody in the blood compared to thirty-five microgram injected dose suggested that the lymph nodes were saturated with ^l25 I-anti-LYVE-l antibody.

Fig. 3. shows the accumulation of ¹²⁵ I-anti-LYVE-l in the different organs dropped continuously over time; biodistribution experiments of ¹²⁵ I-anti-LYVE-l antibody at day 1 (approximately 24 h), day 2 (approximately 48 h), and day 3 (approximately 72 h) after antibody injection; the results are expressed as the percentage of the injected radioactivity dose per gram of tissue (%ID/g) ± SD; accumulation of injected I-anti-LYVE-1 antibody in the analyzed organs and the blood dropped uniformly over time; the data from day 1 after antibody injection were from Figure 2;

Fig. 4. shows in vivo imaging of inflammation-induced lymph node lymphangio genesis by PET using ^l24 I-anti-LYVE-l antibody; (A, B) maximal intensity projections (MIP) of in vivo scanned mice injected with ^l24 I-anti-LYVE-l antibody (A) or -control IgG (B); (A) The inflamed auricular lymph node with on-going lymphangiogenesis (black arrow) accumulated more ¹²⁴ I-anti-LYVE-l antibody than the contra-lateral control auricular lymph node (grey arrow); brachial and axillary lymph nodes were as well clearly visible (arrow heads); (B) in vivo PET image of an ¹²⁴ I-control IgG injected mouse; most of the antibody accumulated in the blood pool and visualized the heart (black arrow); (C, D) in order to compare antibody accumulation in lymph nodes of ' I-anti-LYVE-1 antibody (C) and -control IgG (D) injected mice, coronal PET image sections were displayed with a fixed grey scale normalized to the injected dose per body weight; (C) a coronal section through the auricular lymph nodes displayed ' ⁴ I-anti-LYVE-l antibody accumulation in the inflamed auricular lymph node (black arrow), its contra-lateral control auricular lymph node (grey arrow), brachial and axillary lymph nodes (arrow heads); (D) in a corresponding section of an ¹²⁴ I -control IgG injected mouse, no signals of lymph nodes were detected; only the heart and the carotid arteries (black arrows) were visible; (E, F) to prove that the injected ¹²⁴ I-anti-LYVE-l antibody visualized the auricular lymph nodes, ^l24 I-anti-LYVE-l injected animals were sacrificed and imaged ex vivo by PET, with their auricular lymph nodes placed next to their heads; coronal PET image sections of these animals were normalized as above for direct comparison to the in vivo sections; (E) a representative coronal section through the auricular lymph nodes of the ^l24 I-anti-LYVE-l antibody injected mouse from panel C re-scanned ex vivo; the strong radio signal in the throat region visible in panel C was no longer present; (F) instead, the dissected auricular lymph nodes

(black arrows) depicted a radio signal; the radio signals of the brachial or axillary lymph nodes were still present and comparable to those of the in vivo scan (grey arrow); and

Fig. 5. shows the accumulation of ¹²⁴ I-anti-LYVE-l antibody in the lymphatic vessels of the lymph node; sections of an inflamed auricular lymph node of a ¹²⁴ I-anti-LYVE-

1 antibody injected mouse (A-D) or ^I24 I-control IgG injected mouse (E and F), sacrificed two days after antibody injection (approx. 48 h); the radio signal of the injected ¹²⁴ I-anti-LYVE-l antibody (A, black) overlaps with the fluorescence signals of secondary antibodies to the injected anti-LYVE-1 antibody (B, red), and with external immunofluorescence staining for LYVE-I -positive lymphatic vessels

(C, green) in serial sections; (D) merged panels B and C; (E) no radio signal of the 1 ²⁴ I-control IgG was detected by microradiography; the black streak through the picture is a fold of the microradiography film; (F) serial section to panel E, displaying immunofluorescence staining for LYVE-I positive lymphatic vessels; scale bars = 100 μm.

Fig. 6. shows in vivo PET imaging of tumor-induced lymphangiogenesis in popliteal lymph nodes using ¹²⁴ I-anti-LYVE-l antibody, (a, b) MIPs of in vivo scanned mice injected with ^l24 I-anti-LYVE-l antibody (a) or ¹²⁴ I-control IgG (b). (a) The tumor draining popliteal lymph node with on-going lymphangiogenesis (black arrow) is clearly visible, in contrast to the contra-lateral control popliteal lymph node, (b) In vivo PET image of a mouse injected with ^l24 I-control IgG. Most of the antibody accumulated in the blood pool, (c, d) Normalized coronal PET image sections of mice injected with ^l24 I-anti-LYVE-l antibody (c) or ¹²⁴ I-control IgG (d) reveal ¹²⁴ I- anti-LYVE-1 antibody accumulation in the tumor draining popliteal lymph node (c; black arrow), whereas in corresponding sections of a ¹²⁴ I-control IgG injected mouse, no signals in lymph nodes were detected, (e, f) LYVE-I -positive lymphatic sinuses in sections of tumor draining (e) and contra-lateral control (f) popliteal lymph nodes. Scale bars = 100 μm. Fig. 7. shows increased accumulation of ^l24 I-anti-LYVE-l antibody in tumor draining popliteal lymph nodes compared to control lymph nodes, (a) Biodistribution analysis of mice injected with 30 μg ^l24 I-anti-LYVE-l antibody or ¹²⁴ I-control IgG (n=2) subsequent to PET imaging at day 1 after antibody injection, (b) Tumor draining popliteal lymph nodes accumulate more ^l24 I-anti-LYVE-l antibody than contra-lateral control popliteal lymph nodes. LN = lymph node, cpm = counts of gamma quantum stemming from ¹²⁴ I decays per minute.

DESCRIPTION OF PREFERRED EMBODIMENTS

Metastasis to regional lymph nodes represents the first step of tumor dissemination and serves as a prognostic indicator for disease progression. There is a great demand for sensitive and non-invasive methods to detect metastasis to the lymph nodes. While conventional in vivo imaging approaches have focused on the detection of cancer cells, recent evidence indicates that expansion of lymphatic networks within tumor-draining lymph nodes might be the earliest sign of metastasis. The present invention establishes a method to image lymph node lymphangiogenesis non-invasively. In a mouse model of lymph node lymphangiogenesis, it is found that systemically injected antibodies to lymphatic epitopes accumulate in the lymphatic vasculature in tissues and lymph nodes, determined by biodistribution and microradiography analyses. Using a ¹²⁴ I-labeled antibody against the lymphatic vessel endothelial hyaluronan receptor- 1 (LYVE-I), it was possible to specifically image lymphangiogenesis within tumor-and inflammation-draining lymph nodes by in vivo PET imaging. This is the first demonstration of non-invasive imaging of lymph node lymphangiogenesis in vivo. This method opens up new avenues for the early detection of metastases and for radioimmunotherapy and the images obtained can be used as biomarkers for the progression of diseases associated with lymphangiogenesis. Experimental Results:

A systemically injected antibody to a lymphatic epitope accumulates in the lymphatic vasculature: To assess whether antibodies can be used to target and image lymphatic vessels in vivo, it was first tested whether following systemic injection of an antibody against a lymphatic-specific epitope, the antibody accumulates in the lymphatic vasculature. Since antibodies are large-sized molecules (150 kD), leaky blood vessels, which are a hallmark of tumors and inflammation, can promote the antibodies' extravasation from blood vessels and consecutive uptake by and accumulation in the lymphatic vessels. Therefore, an established mouse model of chronic skin inflammation was used. In this model, K14/VEGF transgenic mice are subjected to a cutaneous delayed- type hypersensitivity reaction, induced by topical application of the contact sensitizer oxazolone to the skin. In contrast to wild-type mice, Kl 4/VEGF mice develop a chronic skin inflammation that is associated with vascular hyperpermeability, prominent lymphangiogenesis in the skin and, in particular, in the draining lymph nodes. The Kl 4/VEGF mice were first given systemic injections of a rat antibody to the vascular endothelial growth factor receptor-3 (VEGFR-3) that is expressed by lymphatic endothelium or a control immunoglobulin (Ig)G.

At 48 h after injection, mice were killed and tissue sections were stained with a fluorescently labeled secondary antibody against rat IgG. The anti-VEGFR-3 antibody was specifically detected on lymphatic vessels in the skin, lung, intestine, and tongue; there were high levels of reactivity with the inflamed and control draining auricular lymph nodes. The VEGFR-3 immunoreactivity co-localized with that of the lymphatic marker LYVE-I . However, VEGFR-3 was not detected on blood vessels that were strongly positive for the blood vessel marker Meca32. In contrast, the injected control IgG was not detected on lymphatic vessels and only showed a weak and diffuse staining of LYVE-I- negative areas in the lymph nodes. So, the injected anti-VEGFR-3 antibody specifically accumulated in the lymphatic vessels; surprisingly, normal blood vessels in organs such as the skin and the intestine did not prevent antibody extravasation, thus enabling the detection of lymphatic vessel even in healthy organs.

Iodination of anti-VEGFR-3 antibody does not inhibit its binding to lymphatic vessels in vivo: It was aimed to use antibodies against lymphatic epitopes that were labeled with the radionuclide ¹²⁵ I in the biodistribution experiments. However, iodination of antibodies can decrease their binding affinity for antigen. Thus, it was next investigated whether a systemically injected ^l25 I-anti- VEGFR-3 antibody accumulates in the lymphatic vessels of the Kl 4/VEGF mice with unilateral lymphangiogenesis. Microradiography of tissue sections obtained 48 h after antibody injection into mice revealed strong localization of the ¹²⁵ I-anti- VEGFR-3 antibody at vessel-like structures in the skin and in the draining lymph nodes. In contrast, no signal was detected in samples obtained from mice injected with ¹²⁵ I- control IgG. The localization pattern of the ¹²⁵ I-anti-VEGFR-3 antibody determined by microautoradiography completely overlapped with the localization pattern of the lymphatic marker LYVE-I detected by immunohistochemical stains of serial sections. Together, these findings indicate that iodination of the anti- VEGFR-3 antibody did not interfere with its ability to bind to VEGFR-3 on lymphatic vessels.

Systemically injected anti-LYVE-1 antibody accumulates in the lymphatic vasculature in vivo: To establish a reliable method for the in vivo imaging of lymphatic vessels a different antibody than that against VEGFR-3 was desirable, because the anti- VEGFR-3 antibody (mF4-31Cl) strongly inhibits lymphangiogenesis in vivo. Therefore, the use of an antibody against LYVE-I was investigated, a molecule that is specifically expressed by lymphatic vessels but not by blood vessels. No biological activity of LYVE-I has been identified in vzVo. It was investigated whether systemically injected anti-LYVE-1 antibody reached the lymphatic vasculature of lymph nodes and accumulated at sites of lymphangiogenesis using the K14/VEGF mice and a rat anti-mouse LYVE-I antibody that was injected intravenously. In tissue sections of auricular lymph nodes and ears obtained 24 h after injection, the anti-LYVE-1 antibody was detected by a fluorescently labeled anti-rat IgG antibody. Importantly, the localization of the injected anti-LYVE-1 antibody overlapped with the localization of LYVE-I by external co-staining with a rabbit antibody to LYVE-I, followed by immunofluorescence detection; this confirms that the injected antibody accumulates in lymphatic vessels. Lymphatic-specific expression of the injected anti- LYVE-1 antibody was also detected in other organs, including the intestine, tongue, and salivary glands. However, the anti-LYVE-1 antibody did not co-localize with blood vessels that strongly expressed the vascular marker Meca32. In contrast, the injected rat IgG control antibody was not detected in lymph nodes, skin or other organs. These results indicate that the anti-LYVE-1 antibody accumulates specifically in the lymphatic vasculature of inflamed and normal tissue. Next the rat anti-LYVE-1 antibody and control rat IgG were radiolabeled with ¹²⁵ I and injected systemically into K14/VEGF mice bearing unilateral lymph node lymphangiogenesis. The injected radiolabeled anti-LYVE-1 antibody, but not the injected radiolabeled rat IgG, was clearly detected by microradiography of tissue sections from auricular lymph nodes, ears, lungs, and intestines • (Fig. 1) and was localized at vessel-like structures. LYVE-I staining obtained with a rabbit antibody overlapped with the radio signal of the injected anti-LYVE-1 antibody on serial sections of the ears and auricular lymph nodes (Fig. l),indicating that the affinity of anti- LYVE-1 antibody for its target was not inhibited by radioiodination. Systemically injected, radiolabeled anti-LYVE-1 antibody accumulates in a dose- dependent manner in the lymphatic vessels of the lymph nodes: Biodistribution experiments were performed with the ^l25 I-anti-LYVE-l antibody and ^l25 I-control rat IgG to evaluate their in vivo targeting to sites of lymphangiogenesis within lymph nodes. To this end, different amounts of radiolabeled antibodies (7 μg, 35 μg, or 90 μg) were injected intravenously into K14/VEGF mice with unilateral, inflammation-induced lymph node lymphangiogenesis. At 1 day after injection of 7 μg anti-LYVE-1 antibody, preferential accumulation of the antibody in the lung (89% of the injected dose [ID]/g of tissue) was found (Fig. 2 A and Table Sl), whereas only about 6%-8% ID/g accumulated in the auricular lymph nodes that drained inflamed or non-inflamed skin. Only small amounts of radioactivity were detected in other tissues (Fig. 2 A and Table Sl). Surprisingly, accumulation in the auricular lymph nodes increased to 15%-18% ID/g when 35 μg of anti-LYVE-1 antibody were injected, whereas the accumulation in the lungs decreased to 30% ID/g (Fig. 2 B and Table Sl). The auxiliary lymph nodes accumulated larger amounts of anti-LYVE-1 antibody than the auricular lymph nodes (20%-24% ID/g). In mice injected with control IgG, concentrations of radioactivity in the lymph nodes (between 2%-3 %ID/g) were less than in mice injected with the anti-LYVE-1 antibody, indicating the specificity of the anti-LYVE-1 antibody for lymphatic tissues. Increasing the amount of injected anti-LYVE-1 antibody from 35 μg to 90 μg did not increase the uptake of anti- LYVE-1 antibody in the lymph nodes, compared to other organs, but increased the amount of radioactivity measured in the blood, indicating saturation of the target (Fig. 2 C and Table Sl).

These results demonstrate that the saturation dose of ^l25 I-anti-LYVE-l antibody in the lymphatic vessels in the lymph nodes is in between 7 μg and 90 μg, which is equivalent to 0.2 μg-3.5 μg antibody/g bodyweight. One observes specific targeting to lymphatic vessels by three different concentrations of the ¹²⁵ I-anti-LYVE-l antibody, compared to ¹²⁵ I- control IgG, by microradiography analysis. There were comparable concentrations of radioactivity in the non-inflamed auricular lymph nodes and the inflamed lymph nodes with on-going lymphangiogenesis (Table S2). However, the inflamed lymph nodes weighed approximately 1.8-fold that of the control lymph nodes (Table S2). Table S1. Biodistribution experiments of ^12S l-labeled anti-LYVE-1 antibody and -control IgG

7 μg 35 μg 90 μg

^l25 !-aπti-i_YVE-1 antibody control ear 2.69 ± 1.15 4.79 ± 0.50 4.81 ± 1.07 inflamed ear 4.51 ± 1.44 4.06 ± 2.20 6.39 ± 0.62 control auricular lymph node 7.62 ± 5.79 18.37 ± 1.16 12.59 ± 2.13 inflamed auricular lymph node 6.07 ± 3.04 14.54 ± 1.26 11.69 ± 0.90 left axillary lymph node 2.76 ± 1.27 20.33 ± 6.28 13.91 ± 1.64 right axillary lymph node 3.28 ± 1.70 23.65 ± 5.54 18.91 ± 1.57 left inguinal lymph node 1.72 ± 0.82 9.23 ± 2.20 8.14 ± 1.36 right inguinal lymph node 2.74 ± 1.72 10.11 ± 3.58 8.11 ± 0.18 left submaxillary gland 0.44 ± 0.07 1.68 ± 0.36 2.40 ± 0.50 right submaxillary gland 0.44 ± 0.08 1.35 ± 0.20 2.16 ± 0.34 left sublingual gland 0.97 ± 0.33 5.28 ± 1.25 4.97 ± 0.96 right sublingual gland 1.01 ± 0.36 4.65 ± 1.24 4.18 ± 0.53 left parotid gland 1.10 ± 0.52 4.14 ± 0.25 5.25 ± 0.56 right parotid gland 1.38 ± 0.19 3.91 ± 1.02 4.55 ± 1.32 liver 14.60 ± 2.26 3.33 ± 0.22 3.38 ± 0.28 lung 88.52 ± 7.43 29.64 ± 3.36 24.12 ± 8.64 spleen 1.25 ± 0.40 1.58 ± 0.30 1.97 ± 0.07 heart 4.31 ± 0.52 3.38 ± 0.22 4.32 ± 0.43 kidney 0.63 ± 0.07 ^' 1.17 ± 0.27 3.78 ± 0.68 intestine 2.69 ± 2.33 5.28 ± 0.59 3.58 ± 0.19 skin 0.61 ± 0.26 1.58 ± 0.52 4.30 ± 1.85 blood 0.70 ± 0.06 4.42 ± 0.75 11.54 ± 3.22 ^t25 l-control IgG control ear 3.52 ± 0.60 3.05 ± 0.23 3.26 ± 0.36 inflamed ear 4.93 ± 0.47 4.44 ± 0.59 3.94 ± 0.33 control auricular lymph node 3.02 ± 0.26 2.68 ± 0.19 3.02 ± 0.13 inflamed auricular lymph node 2.55 ± 0.32 2.51 ± 0.33 2.55 ± 0.15 left axillary lymph node 3.03 ± 0.47 2.90 ± 0.44 2.93 ± 0.37 right axillary lymph node 3.02 ± 0.39 3.17 ± 0.43 3.02 ± 0.28 left inguinal lymph node 2.23 ± 0.38 2.23 ± 0.31 2.42 ± 0.45 right inguinal lymph node 2.02 ± 0.44 2.14 ± 0.21 2.40 ± 0.24 left submaxillary gland 2.50 ± 0.39 2.22 ± 0.37 2.81 ± 0.34 right submaxillary gland 2.48 ± 0.60 1.99 ± 0.11 2.65 ± 0.36 left sublingual gland 2.34 ± 0.21 2.25 ± 0.28 3.12 ± 0.88 right sublingual gland 2.40 ± 0.44 2.45 ± 0.39 2.83 ± 0.54 left parotid gland 4.63 ± 1.06 3.93 ± 0.54 4.94 ± 0.66 right parotid gland 497 ± 099 378 ± 081 514 ± 028 liver 309 ± 058 255 i 062 280 ± 056 lung 641 ± 138 434 ± 160 584 ± 018 spleen 196 ± 030 181 i 011 170 ± 019 heart 358 042 334 ± 050 356 x 057 kidney 391 ± 071 362 i 055 438 ± 048 intestine 126 ± 026 134 ± 023 1 17 ± 014 skin 222 ± 077 245 ± 095 273 ± 051 blood 1631 ± 267 1279 ± 186 1236 ± 1 16

Table Sl: Dose dependent accumulation of ¹²⁵ I-anti-LYVE-l antibody in the lymph nodes. Biodistribution experiments of different doses of ¹²⁵ I-anti-LYVE-l antibody or -control IgG in mice. The results are expressed as the percentage of the injected radioactivity dose per gram of tissue (%ID/g) ± SD.

Table S2. Increase in weight and radioactivity in auricular lymph nodes upon inflammation

Amount of injected ^l2S l-labeled antibody 7_μg 35 μg 90 μg

Ratio weight of inflamed to control auricular lymph node

¹²⁵ l-antι-LYVE-1 antibody 18±02 178±03 22±09

^12C| -control IgG 19±08 15±04 14±04

Ratio radioactivity of inflamed to control auricular lymph node

'^ 'll--aannttιι--LLYYVVEE--11 aannttiibbooddyy 17±06 14±01 20±09 " l-control IgG 16±06 12±03 12±03

The results are expressed as the ratio ± SD.

Table S2: Increase in weight and radioactivity in auricular lymph nodes upon inflammation Upon inflammation weight and radioactivity of the auricular lymph nodes were increased. Data are derived of the biodistribution analysis of Figure 2.

Increased accumulation of injected, radiolabeled anti-LYVE-1 antibody in the auricular lymph nodes, compared to the surrounding tissues, over at least 3 days: For reliable PET imaging, it is important that there is a sufficient difference in activity concentration of the region of interest, compared to the neighboring tissue. The auricular lymph nodes are surrounded by salivary glands and the large carotid arteries. The biodistribution experiments revealed that the sublingual gland had the highest % ID/g values of all surrounding tissues (Fig. 2 B). Importantly, at 1 day after antibody injection, the accumulation of radioactivity was 3- to 4-fold higher in the auricular lymph nodes than in the sublingual gland (Table 1). This ratio did not change significantly over time (Fig. 3, Table S3 and Table 1); microradiograpy confirmes that the ^l25 I-anti-LYVE-l antibody is specifically enriched in lymphatic vessels at all three timepoints. The auricular lymph node/blood ratios increased approximately 3-fold by 2 days after injection and 5-fold thereby 3 days, compared to with the first day after injection (Fig. 3, Table S3 and Table

1).

Since the ratio of radioactivity in lymph nodes versus salivary glands did not change significantly over time, and since the radioactivity that accumulated within the lymph nodes diminished from day 1 to day 3, day 1 is chosen as the best timepoint for the consecutive in vivo imaging studies.

Table 1. Auricular lymph node - adjacent organ ratios of the %ID/g of ¹²⁵ l-anti-LYVE-1 antibody

Day 1 Day 2 Day 3 control auricular LN - submaxillary gland left 11.3 ± 2.4 11.1 ± 1.6 7.8 ± 1.5 inflamed auricular LN - submaxillary gland right 11.1 ± 2.6 7.7 ± 2.3 7.7 ± 2.0 control auricular LN - sublingual gland left 3.6 ± 0.7 4.4 ± 0.1 3.4 ± 0.5 inflamed auricular LN - sublingual gland right 3.4 ± 1.5 4.2 ± 1.8 4.4 ± 1.3 control auricular LN - parotid gland left 4.4 ± 0.4 6.5 ± 1.6 3.7 ± 1.4 inflamed auricular LN - parotid gland right 4.0 ± 1.3 3.3 ± 1.2 2.8 ± 0.7 control auricular LN - blood 4.2 ± 0.6 14.8 ± 3.4 19.7 ± 5.2 inflamed auricular LN - blood 3.4 ± 0.6 10.1 ± 1.1 16.2 ± 5.2

The results are expressed as the ratio ± SD.

Table 1 : Accumulation of radiolabeled anti-LYVE-1 antibody in the auricular lymph nodes compared to adjacent tissues and the blood. Table S3. Biodistribution experiments of ¹²⁵ l-anti-LYVE-1antibody at different time points post injection

24 h 48 h 72 h

control ear 4 79 ± 0 50 3 60 ± 0 24 2 22 ± 0 14 inflamed ear 4 06 ± 2 20 2 64 ± 0 31 1 33 ± 0 16 control auricular lymph node 18 37 ± 1 16 13 54 ± 0 49 6 95 ± 0 72 inflamed auricular lymph node 14 54 ± 1 26 9 38 ± 1 45 5 68 ± 0 80 left axillary lymph node 20 33 ± 6 28 16 91 ± 2 07 9 69 ± 1 67 right axillary lymph node 23 65 ± 5 54 18 24 ± 2 30 10 09 ± 1 42 left inguinal lymph node 9 23 ± 2 20 6 78 ± 1 32 3 83 ± 1 91 right inguinal lymph node 10 11 ± 3 58 7 64 ± 1 39 4 78 ± 0 43 left submaxillary gland 1 68 ± 0 36 1 24 ± 0 24 0 90 ± 0 09 right submaxillary gland 1 35 ± 0 20 1 26 ± 0 23 0 76 ± 0 11 left sublingual gland 5 28 ± 1 25 3 11 ± 0 15 2 04 ± 0 26 right sublingual gland 4 65 ± 1 24 2 49 ± 0 89 1 35 ± 0 20 left parotid gland 4 14 ± 0 25 2 19 ± 0 66 2 08 ± 0 66 right parotid gland 3 91 ± 1 02 3 00 ± 0 75 2 09 ± 0 42 liver 3 33 ± 0 22 2 29 ± 0 33 2 02 ± 0 44 lung 29 64 ± 3 36 22 52 ± 3 35 12 98 ± 3 01 spleen 1 58 ± 0 30 1 02 ± 0 23 0 45 ± 0 09 heart 3 38 ± 0 22 2 19 ± 0 37 1 22 ± 0 15 kidney 1 17 ± 0 27 0 55 ± 0 02 0 31 ± 0 04 intestine 5 28 ± 0 59 3 83 ± 0 02 2 54 ± 0 15

Skin 1 58 ± 0 52 1 52 ± 0 48 1 18 ± 0 30 blood 4 42 ± 0 75 0 94 ± 0 18 0 36 ± 0 06

The results are expressed as the percentage of the injected radioactive dose per gram of tissue (%ID/g) ± SD

Table S3: Accumulation of ^I25 I-anti-LYVE-l in the different organs drops continuously over time. Biodistribution experiments of ' ^s I-anti-LYVE-l antibody at day one (approximately 24 h), day two (approximately 48 h), and at day three (approximately 72 h) 5 after antibody injection.

Imaging lymph node lymphangiogenesis in vivo by PET imaging: Based on the encouraging biodistribution results, it was verified that it is possible to visualize lymphatic vessels within lymph nodes by in vivo PET imaging of mice. To this end, the anti-LYVE-10 antibody and control rat IgG were radiolabeled with the positron emitter ¹²⁴ I. ¹²⁴ I and ¹²⁵ I are isotopes that have identical chemical behaviors, so the results obtained from the biodistribution analysis with ¹²⁵ I-anti-LYVE-l antibody could be used to determine the amount of ^l24 I-anti-LYVE-l antibody for injection and the timepoint for in vivo imaging. We injected 38 μg of the radiolabeled anti-LYVE-1 antibody or control IgG intravenously5 into K14/VEGF mice bearing unilateral lymph node lymphangiogenesis. At day 1 after injection, the mice were scanned by PET to obtain in vivo images. In the mice injected with ¹²⁴ I-anti-LYVE-l antibody, strong radio signals were produced at the sites of the auricular, brachial and axillary lymph nodes (Fig. 4 A). The accumulation of ²⁴ I-anti-LYVE-l antibodies in other organs was similar or weaker, compared to the lymph nodes sites. In particular, there was a low level of accumulation in liver, kidney and intestine. However, the most important finding was that the inflamed auricular lymph nodes with on-going lymph node lymphangiogenesis produce much stronger radio signals than the contra-lateral (control) lymph nodes (Fig. 4 A). A stronger accumulation of radiotracer in the auricular lymph node that was undergoing lymph node lymphangiogenesis, compared to the contra- lateral lymph node, was also imaged at day 2 after antibody injection. The radiolabeled control IgG had a different distribution pattern; it was mainly localized in the blood, resulting in visualization of large blood vessels and the heart (Fig. 4 B). It was next investigated whether there was increased accumulation of the ^l24 I-anti-LYVE-l antibody, compared with ¹²⁴ I-control IgG, in the lymph nodes of the mice. Direct comparisons of maximal intensity projections (MIPs) from different animals are not possible, since the gray scales of MIPs are applied depending on the range of antibody accumulation, which can differ between mice. Therefore, PET image sections were displayed with a fixed gray scale that had been normalized to the injected dose per body weight; normalized coronal PET image sections in the planes of the auricular, axillary and brachial lymph nodes were compared. The ¹²⁴ I-anti-LYVE-l antibody accumulation was clearly visible in planes of auricular, axillary, and brachial lymph nodes (Fig. 4 C). Meanwhile, no signal was observed from the corresponding sections of mice injected with ¹²⁴ I -control IgG (Fig. 4 D). For a proof that the ¹²⁴ I-anti-LYVE-l antibody is specifically localized to the auricular lymph nodes, rather than to structures surrounding the lymph nodes, the imaged animals were sacrificed and imaged, ex vivo, by PET, with the dissected auricular lymph nodes placed next to the heads of the mice. Coronal sections of these data sets were normalized to compare them with the in vivo sections. The strong radio signals that were detected in vivo in the regions of the auricular lymph nodes were no longer detectable in the sections of the ex vivo scans (Fig. 4 E). Instead, the auricular lymph nodes that were placed next to the head of the animals emitted a strong radio signal (Fig. 4 F). The radio signals from the regions of the axillary and brachial lymph nodes in the corresponding sections were comparable between the in vivo and ex vivo PET scans (Fig. 4 C and F). In ex vivo scans of mice injected with the control IgG, there were no detectable signals in the dissected lymph nodes. The specific binding of the ^l24 I-anti-L YVE-I antibody to lymphatic vessels, compared to the ^l24 I-control IgG, was confirmed by microradiography analysis and by staining of serial sections for LYVE-I (Fig. 5). This is the first demonstration that lymph node lymphangiogenesis within auricular lymph nodes can be imaged in vivo by PET following systemic injection of a radiolabeled, lymphatic-specific antibody. By the above it is shown that lymphatic vessels can be targeted and imaged by antibodies Without being bound to any theoretical explanation, it is shown for the first time that lymphatic vessels can be targeted and imaged by specific antibodies and the proof-of-principle is provided for the non-invasive in vivo imaging of lymph node lymphangiogenesis by PET.

In vivo imaging of lymphangiogenesis in tumor draining lymph nodes by PET: After the proof-of-concept in a model of inflammation-induced lymph node lymphangiogenesis, we next set out to image lymphatic vessel expansion in tumor draining lymph nodes. Nineteen days after implantation of B 16-Fl melanoma cells that stably overexpressed the lymphangiogenic factor VEGF-C into the footpads of C57BL/6N mice, tumors reached volumes of approximately 110 mm and in vivo PET imaging was performed one day after intravenous injection of 30 μg ' ²⁴ I- anti-L YVE-I antibody or ¹²⁴ I-control IgG. Strikingly, in ^l24 I-anti-L YVE-I antibody injected mice, tumor-draining popliteal lymph nodes were clearly visible by PET imaging, in contrast to contra-lateral control popliteal lymph nodes (Fig. 6 a). In contrast, ^l24 I-control IgG was mainly localized in the blood (Fig. 6 b). Normalized serial sections confirmed that ' ²⁴ I- anti-LY VE-I antibody but not ¹²⁴ I-control IgG accumulated in the tumor draining lymph nodes. In agreement with the observed accumulation of injected radiolabeled anti-LYVE-1 antibody, immunofluorescence analysis of popliteal lymph node sections showed expansion of LYVE-I -positive lymphatic vessels in tumor draining lymph nodes compared to control lymph nodes (Fig. 6 e and f). Tissue distributions of ¹²⁴ I-anti-LYVE-l antibody and ¹²⁴ I-control IgG were quantified directly after PET imaging revealing similar antibody-distribution patterns as observed in the inflammation model (Fig. 7 a), with a 4.1- to 5.5-fold enhanced ' ⁴ I-anti- LYVE-I antibody accumulation in tumor draining lymph nodes compared to control lymph nodes (Fig. 7 b).

Together, these results demonstrate, for the first time, the feasibility of non-invasive in vivo imaging of expanded lymphatic networks within tumor draining lymph nodes. To determine whether a systemically injected antibody against a lymphatic epitope could localize to and be imaged in the lymphatic vessels, a previously described rat anti-mouse VEGFR-3 antibody is chosen, because of VEGFR-3's specific expression at lymphatic vessels. Although VEGFR-3 has been detected on subsets of activated macrophages, monocytes and dendritic cells, and is expressed by blood vascular endothelial cells during embryogenesis, and, likely, on some angiogenic blood vessels associated with tumors and wounds, the combined microautoradiography and immunofluorescence studies reveal that an injected anti- VEGFR-3 antibody specifically accumulates in LYVE-I -positive lymphatic vessels of the skin and the lymph nodes, but not in Meca32-positive blood vessels.

Importantly, the accumulation of injected anti- VEGFR-3 antibody is not caused by unspecific uptake of Igs by lymphatic vessels, because injection of equal amount of control IgG does not lead to any detectable accumulation in lymphatic vessels. Thus, after intravenous administration, circulating anti- VEGFR-3 antibodies seem to be extravasated from leaky blood vessels at sites of inflammation and drained by lymphatic vessels, where they bound to VEGFR-3 expressed on the luminal surface. Similarly, an intraperitoneally injected anti-ICAM-1 antibody was recently detected by immunofluorescence analysis of inflamed lymphatic vessels. It is unlikely that binding of the anti-VEGFR-3 antibody to macrophages or dendritic cells, which might also express VEGFR-3 and/or LYVE-I, followed by migration of these cells from the tissues into lymphatic vessels and lymph nodes contributed to the specific antibody accumulation. Both the LYVE-I immunofluorescence analyses and the microautoradiography clearly depicted signals of the injected antibody associated with endothelial cells in vessel-like structures, but only occasionally with single cells.

One of the most surprising findings of the study is that the systemically injected, radiolabeled or fluorescently labeled anti-VEGFR-3 antibody also accumulates in lymphatic vessels of normal, non-inflamed tissues, including the skin, lymph nodes, tongue and intestine. These findings reveal that the labeled antibodies efficiently extravasated also from normal blood vessels, and not only at sites of vascular hyperpermeability. These effects were confirmed in both FVB and BALB/c nude mice and indicate that it might be possible to perform lymphangiographic analyses after systemic injection of a lymphatic- specific antibody. One potential obstacle regarding the clinical use of an anti-VEGFR-3 antibody for imaging is its ability to block VEGFR-3 signaling, which could inhibit lymphangiogenesis. Therefore, an antibody against LYVE-I, a hyaluronan receptor that is specifically expressed by lymphatic endothelium but not by blood vessels, is used. Apart from its lymphatic expression, LYVE-I is only present on liver sinusoidal endothelial cells and a subset of macrophages and embryonic blood vessels. Most importantly, however, no non- redundant function of LYVE-I has been identified, despite intense research on this most widely used marker for lymphatic vessels. Importantly, it was found that following intravenous injection, the anti-LYVE-1 antibody efficiently accumulates in lymphatic vessels of different organs and tissues (skin, lymph nodes, lung, intestine), indicating bioavailability comparable with that of the anti- VEGFR- 3 antibody.

Unexpectedly, however, the biodistribution analyses of radiolabeled anti-LYVE-1 antibody reveals dose-dependent changes in the accumulation pattern. Whereas injection of a low dose of the antibody (7 μg) results primarily in targeting to the lungs, higher doses accumulate in lymph nodes. LYVE-I is expressed in the lung tissue; molecules in the pulmonary endothelium appear to be the first to bind the systemically injected antibody. When injected at higher doses (35 μg and 90 μg), the amount of anti-LYVE-1 that accumulate in the lymph nodes, compared to the lungs, increases, indicating that lung LYVE-I binding is saturated in the lungs and that the remaining antibody accumulated in the lymph nodes.

The finding that 90 μg of anti-LYVE-1 antibody did not further increase specific accumulation in lymph nodes over that observed after 35 μg, but led to increased blood levels, indicates that the accessible LYVE-I molecules were saturated at an with antibody dose of between 7 and 90 μg. Importantly, the biodistribution data revealed that anti- LYVE-1 did not accumulate in organs besides lymph nodes or lung, suggesting the potential low tissue toxicity of radiolabeled LYVE-I antibodies.

It was chosen to label anti-LYVE with ¹²⁴ I for immuno-PET analysis because of its long half-life (100 hours), which would allow imaging for several days after injection. Moreover, the biodistribution data generated from studies with the I-anti-LYVE-1 antibody could be applied to the ¹²⁴ I-labeled antibody. Although the relative drawbacks of ¹²⁴ I are its intrinsic high energy γ-radiation and the high energy of the emitted positrons, which had the potential to increase background signals and lower resolution, we were able to specifically image lymph node lymphangiogenesis.

Perhaps one of the most important findings of this study is that signal produced by the inflamed auricular lymph node, which had on-going lymph node lymphangiogenesis, is stronger than that of the uninflamed control lymph node. The increased accumulation of the antibody at this site is likely to result from an increased number of LYVE-I molecules on the expanded lymphatic networks within the lymph nodes that drain the skin inflammation at this site. Microradiography analysis shows an increased number of lymph node lymphatics in this tissue and immunoblot analyses of lymph node lysates also revealed an increased amount of LYVE-I protein in inflammation-draining lymph nodes, compared to normal lymph nodes.

Most importantly, ¹²⁴ I-anti-LYVE-l antibody-based PET imaging enabled detection of lymphatic vessel expansion within tumor-draining lymph nodes, using an established model of melanoma tumors growing in the footpad. This novel method of imaging lymph node lymphangiogenesis provides a new strategy for the early detection of metastases in lymph nodes and has several advantages over currently used methods. Lymph node lymphangiogenesis has been identified as an early marker of metastasis to lymph nodes in experimental models. Moreover, expansion of lymphatic networks in sentinel lymph nodes has also been observed in patients with melanoma or breast cancer and found to be a significant predictor of distant metastasis. Thus, the use of PET imaging of radiolabeled antibodies, in particular of anti-LYVEl antibodies, to detect lymph node lymphangiogenesis represents a less invasive, simpler and potentially more sensitive method to identify patients with lymph node metastases than current approaches, including sentinel lymph node dissection. The method also avoids of the need to inject dyes around tumors; this technique does not always lead to the detection of all draining lymph nodes, due to the location of the injection or the clogging of lymphatic vessels by metastatic tumor cells. Moreover, it is possible, in the case of non-resectable tumors, to predict which tumors are most likely to metastasize in the future, based on the expansion of the lymphatic network of the lymph nodes that occurs before the onset of metastasis. To our knowledge, at present no other imaging modality offers the combined advantages of immuno-PET for detecting lymph node lymphangiogenesis. Magnetic resonance imaging (MRI) and computed tomography (CT) might provide even better resolution than PET imaging, but it is a challenge to specifically image lymphatics with these technologies because targeted contrast agents are not small enough to extravasate from blood vessels. Increased lymph flow in tumor sentinel lymph nodes in mice can be visualized by near infrared imaging using nanoparticles and MRI using a gadolinium contrast agent. However, contrast agents and nanoparticles need normally to be injected interstitially and the use of near infrared dyes for imaging deeper lymph nodes is limited. Furthermore, the difference in flow decreases quickly after injection, demanding immediate imaging for accurate quantification. In contrast, immuno-PET using ^l24 I-anti-LYVE-l offers flexibility in time of image collection; it is e.g. possible to image lymph node lymphangiogenesis 2 days after antibody injection. Immuno-PET imaging of lymph node lymphangiogenesis can be applied to medical fields beyond oncology, because many pathological conditions (e.g. chronic inflammatory diseases including rheumatoid arthritis) are associated with lymphangiogenesis. Thus, lymphangiogenesis can be imaged and used as a biomarker for disease progression or response to therapy. Radiolabeled antibodies against lymphatic endothelial cell epitopes such as LYVE-I or other molecules that are specifically up-regulated on growing lymphatic vessels can also be used for radioimmunotherapy of metastases associated with tumor lymphangiogenesis. Materials and Methods

Mouse model of inflammation-induced lymph node lymphangiogenesis: Delayed-type hypersensitivity reactions were induced in the ear skin of female hemizygous transgenic FVB mice that overexpress VEGF-A 164 in the epidermis under control of the human keratin 14 promoter (K14/VEGF mice) or in male C57BL/6J K14/VEGF mice. For all studies, age matched 9- to 21 -week-old hemizygous VEGF transgenic mice that had no pre-existing inflammatory skin lesions were used (22 - 35 g). Mice were sensitized by topical application of a 2% oxazolone (4-ethoxymethylene-2-phenyl-2-oxazoline-5-one; Sigma- Aldrich, St. Louis, MO, USA) solution in acetone/olive oil (4:1 vol/vol) to the shaved abdomen (50 μl) and to each paw (5 μl) as described. Five days after sensitization, the right ears were challenged by topical application of 20 μl of a 1% oxazolone solution. The resulting inflammation in the ear and the ear draining (auricular) lymph node is accompanied by lymphangiogenesis. All animal experiments were approved by the cantonal veterinarian office Zurich.

Mouse model of tumor-induced lymph node lymphangiogenesis: B16-F1 murine melanoma cells (kindly provided by Dr. S. Hemmi, University of Zurich, Switzerland) were cultured in D-MEM (Invitrogen, Carlsbad, CA, USA) containing 10% FBS and were transfected by electroporation with fiill-length human-VEGF-C subcloned into the pcDNA3.1 (Invitrogen) vector. Stable clones (B 16-Fl -VEGF-C cells) were selected and VEGF-C expression was confirmed by RT-PCR and ELISA. 2xlO ⁵ B16-F1-VEGF-C cells in 10 μl PBS were injected into the left footpads of 9 to 11 -week-old female C57BL/6 N mice (Charles River Laboratories, Wilmington, MA, USA) as described.

Ex vivo fluorescence experiments: Four hundred and fifty micrograms of rat anti -mouse VEGFR-3 antibody (mF4-31Cl) or rat control IgG were injected intraperitoneally into 10- week-old C57BL/6J K14/VEGF transgenic mice (one mouse per treatment), one day after challenging one ear with oxazolone. Forty-eight hours after injection, the animals were sacrificed and organs were frozen in optimal cutting temperature (OCT) compound (Sakura Finetec, Zoeterwoude, Netherlands). In an additional experiment, 85 μg of rat anti- mouse LYVE-I antibody (clone 223322, R&D Systems, Minneapolis, MN, USA) or isotype-matched rat control IgG were injected into the tail veins of 10- week-old K14/VEGF mice (one mouse per treatment) 1 day after challenging one ear with oxazolone. Twenty- four hours after injection, the animals were sacrificed and organs were frozen in OCT compound. Seven-micrometer frozen sections were cut and fixed with 4% paraformaldehyde in PBS for 15 min at 4°C. To detect binding of the injected antibodies, the sections were incubated with an Alexa Fluor 594 conjugated donkey anti-rat IgG antibody (Invitrogen, Carlsbad, CA, USA). The sections were co-stained with a rabbit anti- mouse LYVE-I antibody (Angiobio, Del Mar, CA, USA) that was detected by an Alexa Fluor 488 donkey anti-rabbit IgG antibody (Invitrogen), or co-stained with a biotinylated rat Meca32 antibody (specific staining of blood vessels; BD Pharmingen, Franklin Lakes, New Jersey, USA) that was detected by Alexa Fluor 488 streptavidin (Invitrogen). All sections were counterstained with Hoechst 33342 (Invitrogen), mounted with Vectashield (Vector Laboratories, Inc., Burlingame, CA, USA), and analyzed with an AxioScop2 mot plus microscope (Zeiss, Oberkochen, Germany). Images were captured with an AxioCam MRc camera (Zeiss) using the Axio-Vision 4.7 software. Biodistribution studies: The in vivo targeting performance of a rat anti-mouse LYVE-I antibody (clone 2233.22, R&D Systems) or isotype-matched rat control IgG (AbD serotec, Oxford, UK) was evaluated by quantitative biodistribution analysis. Antibodies were radiolabeled with Na ¹²⁵ I (Perkin Elmer, Waltham, MA, USA) by adapting the standard chloramine-T method. Briefly, 70 - 380 μCi Na I and 5 μl of 5 mg/ml chloramine T (Sigma- Aldrich) were added per 100 μg antibody in PBS, concentrated to 1 mg/ml. After 2 minutes, the radiolabeled antibodies were separated from free ¹²⁵ I using PDlO columns (GE-Healthcare, Chalfont St. Giles, UK) pretreated with 1 ml 0.1% BSA and equilibrated in PBS. The radioactivity of the samples was determined using a γ-counter (Cobra Autogamma, Packard Instrument Comp., Meriden, CT, USA).

Normal uptake of radiolabeled iodine by the thyroid glands was blocked by administration of potassium iodide (Fluka, Buchs, Switzerland) in the drinking water starting four days before an experiment. Binding of radiolabeled iodine by iodine symporters in the intestine was blocked by oral administration of sodium perchlorate (Fluka) one hour before antibody injection. For the concentration experiments, 20-week-old K14/VEGF mice were injected intravenously with either 7 μg (16 μCi, n=5), 35 μg (80 μCi, n=5), or 90 μg (55 μCi, n=3) of radiolabeled anti-LYVE-1 antibody, or with equal amounts of radiolabeled rat control IgG at 6 or 8 days after challenging one ear with oxazolone. Mice were sacrificed 24 h after injection. Organs were weighed and radioactivity was counted with a γ-counter. The radioactivity content of representative organs was expressed as the percentage of the injected dose per gram of tissue (%ID/g). For additional timecourse experiments, 20-week- old K14/VEGF mice were given intravenous injections of 37 μg (52 μCi) of radiolabeled anti-LYVE-1 antibody 13 days after oxazolone challenge. Four animals each were sacrificed at 2 days and at 3 days after injection and tissues were processed as described above.

Microradiography: Microradiography was performed with organs from mice injected with ¹²⁵ I- anti-VEGER-3 antibody (mF4-31Cl) or rat control IgG. In additional experiments, mice were injected with ¹²⁵ I- or ¹²⁴ I-radiolabeled anti-LYVE-1 antibodies (clone 223322, R&D Systems) or rat control IgG. Radiolabeling was performed as described above. Radiolabeled anti-VEGFR-3 antibody (18 μg, 47 μCi) or rat control IgG (8 μg, 66 μCi) were injected intravenously into 12-week-old K14/VEGF mice at 10 days after challenging one ear with oxazolone (n=2 per group). Forty-eight hours after injection, mice were sacrificed and organs, including lymph nodes, were frozen in OCT compound. Anti- LYVE-1 antibody (35 μg, 150 μCi) or rat control IgG (35 μg, 120 μCi) were injected intravenously into 16-week-old K14/VEGF mice at 10 days after the challenge (n =2 per group); mice were sacrificed 24 h later. Seven-micrometer frozen sections of the tissues were fixed with 4% paraformaldehyde in PBS at 4°C for 15 min. Air-dried sections were coated with KODAK autoradiography emulsion type NTB (CARESTREAM HEALTH, INC., Rochester, NY, USA), dried and stored at room temperature for 2 weeks ( ¹²⁵ I-Or ¹²⁴ I- anti-LYVE-1 and corresponding control IgG sections) or for 3 weeks (' ⁵ I-anti-VEGFR-3 and corresponding control IgG sections) as described (45). The sections were developed in KODAK Developer Dl 9 (C ARESTREAM HEALTH, INC.) for 4 min and fixed with RA 3000 fixer (Kodak) for 5 min. Finally, the sections taken from mice that had received injections of anti-VEGFR-3 or the corresponding control IgG were counterstained with hematoxylin (Richard Allan Scientific, Kalamazoo, MI, USA).

Serial sections of the microradiography sections were stained with a rabbit anti-mouse LYVE-I antibody (Angiobio) and visualized either by immunohistochemistry using biotinylated anti-rabbit IgG antibodies, the VECTASTAIN ABC Kit and the AEC substrate kit (all from Vector Laboratories, Inc), or by immunofluorescence with an Alexa Fluor 488-labeled secondary antibody. PET imaging: One hundred eighty μg of anti-LYVE-1 antibody (R&D Systems) or rat control IgG were labeled with 2-2.7 mCi Na ^{1 4} I (IBA Molecular, Louvain-La-Neuve, BE) by the chloramide T method described above. ^l24 I-anti- LYVE-I (38 μg, 0.37-0.42 mCi) or ¹²⁴ I-labelled rat control IgG (38 μg, 0.34-0.36 mCi) were injected intravenously into 17- week-old K14/VEGF mice after oxazolone treatment of the ear skin (3 mice per treatment). Injected doses were quantified by measuring syringes before and after injection using a VEENSTRA dose calibrator (Siemens, Erlangen, Germany). Normal uptake of radiolabeled iodine by the thyroid glands was blocked by administration of potassium iodide starting four days prior to antibody injection and the mice were given sodium perchlorate perorally one hour before antibody injection. PET scans were performed approximately 24 h (2 mice per group) or 48 h (1 mouse per group) after intravenous radiotracer injection using the GE Vista/CT camera (GE Healthcare) as described previously. For in vivo PET scanning, mice were anesthesized with isoflurane (Abbott Laboratories, Abbott Park, IL, USA) in an air/oxygen mixture and monitored during PET scanning as described previously. After termination of the acquisition, mice were sacrificed by cervical dislocation. Auricular lymph nodes were dissected and the mice were re-scanned, with the dissected auricular lymph nodes positioned next to their heads. Positrons emitted by ' I move in average 2.3 mm away from their source until they annihilate with an electron. Since lymph nodes are only a few cubic millimeters in size, accumulation of radiotracer in isolated extracted lymph nodes can be missed, because there is little mass for positron slowdown. To circumvent this, the dissected lymph nodes were placed in agar blocks to surround them with mass. For the in vivo and ex vivo approaches, whole-body PET data were acquired in two bed positions (30 min acquisition time per bed position) and were reconstructed in a single timeframe, with pixel sizes of 0.3875 mm and 0.775 mm in the transverse and axial directions, respectively. Series of coronal image slices and maximum intensity projections (MIPs) as well as MIP movies, were generated using the dedicated software PMOD (PMOD Technologies Ltd., Adliswil, Switzerland). MIPs render three-dimensional images into two dimensions by displaying the most intensive value from each voxel stack. For visual inspection and comparison of antibody uptake in different mice, PET image sections were displayed with a fixed grey scale normalized to the injected dose per body weight.

For PET imaging of tumor-induced lymph node lymphangiogenesis, 180 μg of anti-LYVE- 1 antibody or rat control IgG were labeled with 2-2.7 mCi Na ¹²⁴ I as described above. ¹²⁴ I- anti-LYVE-1 (30 μg, 0.26-0.38 mCi) or ¹²⁴ I-labelled rat control IgG (30 μg, 0.32-0.33 mCi) were injected intravenously into tumor bearing mice 19 . days after tumor cell injection. PET imaging and biodistribution analysis of selected organs were performed as described above.