ONLINE SUPPLEMENT FOR

Chemokine mediated migration of melanoma cells towards lymphatics

– a mechanism contributing to metastasis.

Jacqueline D. shields, Maxine S Emmett, Darryl BA Dunn, Kavita D. Joory, Leslie Sage, †Howard Rigby, #Peter S. Mortimer, •Antonio Orlando, *J. Rodney Levick, David O. Bates.

Methods.

Human amelanotic A375 melanoma cells (CRL-1619, ATCC), A375P melanoma cells – a non metastatic clone of A375 (A kind gift of IJ Fidler, MD Anderson Centre, Texas), A2058 metastatic melanoma cells (CRL-11147, ATCC), Chinese Hamster Ovary (CHO cells and neonatal, human, dermal endothelial cells (HDEC, CC-2516, Cambrex) were maintained as detailed in the supplementary material. HDECs are a mixture of blood endothelial cells (BEC) and lymphatic endothelial cells (LEC). LEC and BEC were separated from the HDEC by double immunomagnetic purification with the lymphatic specific marker podoplanin as previously described(Kriehuber et al., 2001). The separated LEC phenotype was confirmed using lymphatic-specific LYVE-1 staining, which was positive for LEC and negative for BEC. Conversely von Willebrand factor staining was negative for LEC, and positive for BEC, while CD31 staining was positive for both cell types. To produce CM (CM) cells were cultured until 80-90% confluent, washed in PBS and incubated in serum free medium for 24 hours. CM was collected and stored at –20ºC. CCL21 and VEGF-C ELISAs were carried out using commercial Quantikine kits (R&D cat DC600 and DVEC00 respectively)CHO, A2058 and A375 cells were maintained in DMEM supplemented with 10% FBS. HDEC, LEC and BEC were maintained in EGM-2 MV media (CC-3202, Cambrex). A375P, cells were maintained in MEM (Invitrogen) supplemented with 10% fetal bovine serum, sodium pyruvate, nonessential amino acids, L-glutamine, and vitamin solution.

Immunomagnetic purification.

HDECs were separated into lymphatic endothelial cells (LEC) and blood vascular endothelial cells (BEC) by double immunomagnetic purification (Miltenyi Biotech) with the lymphatic specific marker podoplanin protocols as previously described(Kriehuber et al., 2001). Briefly, cells were trypsinised, re-suspended in wash buffer (PBS supplemented with 2 mM EDTA, 0.5% BSA) and incubated with an antibody to the lymphatic specific marker podoplanin (4 µg.ml-1, Research Diagnostics) as previously described (Kriehuber et al., 2001). The suspension was washed and then incubated in the presence of goat anti-rabbit coated beads (Miltenyi Biotech). Labelled cells were passed through a column within a magnetic field. The effluent was collected. This contained unlabelled blood vascular endothelial cells (BEC). The magnetically labelled lymphatic endothelial cells (LEC) were collected by removing the column from the magnetic field and flushing the cells out. Cells were maintained in EGM-2 MV media. Cell phenotype was tested by LYVE-1 staining (positive for LEC, negative for BEC), and von Willebrand Factor (positive for BEC) and CD31 (positive for both cell types).

Condition media was collected after overnight incubation of the appropriate media from T75 flasks containing approximately 80% confluent CHO cells, HDEC, purified LEC, BEC or A375 cells. MAZ51 (5 µM, Calbiochem), and/or CCL21 neutralising antibody (3 µg/ml AF366, R& D systems), were added to the lower chamber during the migration assay.

Migration assays were performed in a modified Boyden chamber consisting of a cell culture insert with an 8 µm pore polycarbonate membrane (Millipore) seated in each well of a 24 well plate. Membranes were coated with 200 µg.ml-1 of collagen-A and air-dried. Cells were seeded at a density of 1 x 105 suspended in 500µl of serum free medium supplemented with 0.2% BSA into the top chamber. The lower chamber contained 500 µl of serum free media and migratory stimulus. The plate was incubated overnight in a humidified atmosphere at 37ºC to allow migration. The inserts were then fixed in methanol and non-migrated cells were removed using a cotton swab. Migrated cells were stained with haematoxylin and the cell culture membranes were then mounted onto slides. Migrated cells could then be counted. The fold increase in migration compared to control was calculated as the migration in the test wells divided by the mean of that in the equivalent media with minimal (0.001%) FCS supplementation. This usually resulted in 1-3% of the cells migrating. To ensure that migration was possible, migration towards 10% FCS was used a positive control. HDEC, LEC, BEC, A375 and A375P cells showed a 42.9 ± 2.9, 37.0 ± 9.8, 4.3±0.8, 54.3 ± 4.3, and 54.4 ± 3.0 fold increase in migration in response to 10% FCS respectively.

Boyden Chamber Migration assays.

Migration was studied using the lymphatic chemokine (15 ng.ml-1, CCL21, R&D systems, UK) and VEGF-C (100ng/ml a kind gift of K Alitalo) as previously described (Bates et al., 2002). Briefly cells were seeded suspended in serum free medium supplemented with 0.2% BSA into the top of a Boyden chamber, in which the lower chamber contained low serum media and/or migratory stimulus and incubated for 24 hours. Inserts were then fixed and non-migrated cells removed. Migrated cells were stained with haematoxylin and then counted. Migration in response to the test condition was expressed as a multiple of the control migration, i. e. migration in the test wells was divided by the mean of that in the control media with minimal (0.001%) FCS supplementation.

Proliferation assays and immunocytochemistry for CCR7

A375 cell were seeded onto 6 well plates, and grown to 50% confluence. Cells were then serum starved overnight, and then treated with either fresh media or LEC CM containing 0.001%fetal calf serum, or fresh media containing 10% FCS. Cells were then allowed to grow overnight, and then cells either trypsinised and counted using a haemocytometer, or fixed as below for immunocytochemistry.

To investigate CCR7 expression, cells grown on collagen coated coverslips were fixed in 4% formaldehyde and permeabilised with 0.2% Triton X-100. Non specific binding was prevented by incubating in 3% BSA (in PBS) for 1hr. Samples were then incubated in primary antibody overnight at 4ºC in goat anti-human CCR7 (4 µg.ml-1). Matched concentrations of IgG were used as negative controls. The primary antibody was detected using biotinylated secondary antibodies (6.7 µg.ml-1). Samples were washed, mounted and visualised on a Nikon E400 microscope using the same exposure time for positively stained cells and negative (IgG) controls. Expression of CCR7 was confirmed by Western blotting.

Human skin samples and archived human melanomas

Control skin was collected from the forearm of twenty patients by 4 mm punch biopsy, with Local Ethics Committee approval (North Bristol NHS Trust). Twenty-four archived samples of primary malignant melanoma (3-12 years old) were studied (follow-up times of 5.7±1.4 years for the metastatic group and, 8.3±1.5 years for the non metastatic group, p>0.1 t test). Of the 24 melanoma patients, 15 had subsequently developed metastasis diagnosed according to normal clinical practice, and 9 were still free of any form of clinically detectable spread by June 2004.

Immunohistochemistry for lymphatic endothelium, proliferating cells. Tumours sections were de-waxed and re-hydrated prior to microwave antigen retrieval (800 watts) in Tris/EDTA (Trizma, 100mM; EDTA, 2mM), pH 9. Endogenous peroxidase activity was blocked by incubation for 5min in 3% hydrogen peroxide. Non-specific binding was prevented by incubation in normal serum for 20 min (for LYVE-1 1.5% goat serum v/v in PBS; for Ki67, 1.5% v/v horse serum in PBS). The test slides were incubated in rabbit anti-human LYVE-1 (4.2 µg.ml-1, Dr D Jackson), mouse anti human CCR7 (4µg. gml-1, Santa Cruz) or mouse anti-human Ki67 to detect proliferating cells (4µg.ml-1, Oncogene) and negative slides were incubated in normal rabbit, or mouse IgG at the same concentration overnight at 4ºC. The primary antibody was detected with a biotinylated goat anti-rabbit (LYVE-1), or horse anti-mouse (ki67, CCR7) secondary antibody (Vector Laboratories) for 30min. Staining was visualised using an avidin-biotin enzyme complex (Vector) for 30 min at room temperature followed by diaminobenzidine (Vector).

CCL21 ELISA (R&D Systems) was carried out according to the manufacturers instructions. To examine CCR7 expression in melanoma cells, lysis buffer (1µl/mg tissue; 1 mmol/L phenyl methyl sulfonyl fluoride (PMSF), 1 mmol/L Va3N04, 0.1µg/mL Leupeptin, 1µg/mL Aprotonin, 1µg/mL Pepstatin, in RIPA buffer) was added to A375, A375P or A2058 cells and placed at 4°C on an agitating rocker plate for 20 minutes, removed and placed in ice for a further 60 minutes and agitated every 10-15 minutes. Samples were centrifuged at 4C for 15 minutes at 13,000rpm, and the supernatant collected and stored at -20C. Protein concentrations were determined by photospectrometry. All protein samples were re-suspended in SDS sample buffer, heated at 100°C for 5 min, and subsequently resolved by SDS-PAGE and transferred to PVDF membrane. The membranes were blocked with PBS, 10% skimmed milk powder (Marvel, Somerfield UK), 0.05% Tween 20, for 1 h at room temperature, followed by overnight incubation at 4oC with rabbit anti-human CCR7 (AB12953, Abcam, 1:2000) by overnight incubation at 4oC. After washing in PBST, the blots were then incubated for 1h with HRP conjugated goat anti-rabbit, (Pierce, 34095,1:5000) antibody, and visualised. using a SuperSignal West Femto Maximum Sensitivity Substrate Kit (Pierce).

Measurement of lymphatic microvessel density (LD) around tumours

Lymphatic density, LD, was calculated from counts on LYVE-1 stained sections using the method described in Shields et al (Shields et al., 2004). LD was defined as the number of lymphatic microvessel profiles per unit area of non-tumour tissue within 350 µm of the tumour edge (1 field of view from the tumour edge).

Measurement of LD in focal areas of superficial normal dermis.

5 m transverse sections from 20 routinely fixed and processed normal human forearm skin samples were stained for LYVE-1 as previously described. Eighty basal keratinocytes were selected at random and digital images were captured with the selected keratinocytes at their centre using EclipseNet software (Nikon Instruments, UK). The greatest lymphatic density that can be measured is in the upper dermis, and therefore the greatest lymphatic density that the melanomas could experience is that in the upper dermis. The sampling area was thus defined by drawing a circle of radius of 350 m centred on the keratinocyte. This circle was overlayed on each digital image. All dermal lymphatic cross-sections (i.e. LYVE-1 positive profiles) within the sampling area were counted. The area of the dermis within the sampling area was defined on the digital image by freehand trace and the value recorded. The area sampled is more superficial than that in deeper layers of the dermis, and hence has a higher mean lymphatic density than previously described(Shields et al., 2004). This was to ensure that we did not under-estimate the lymphatic density by measuring that in sparsely populated deeper dermis.

Animal model.

Male nude mice (25 g) were injected with 1x106 A375 or A2058 (metastatic) or A375P (non metastatic) cells suspended in 100 µl PBS subcutaneously in the dorsal intrascapular region with a 25 gauge needle that had been scored and dipped in Monastral Blue dye. This resulted in a tattoo of the injection site through the dermis. The mice were then injected with 1x105 human lymphatic endothelial cells (LEC) or blood vascular endothelial cells (BEC) purified as described above suspended in 100µl of PBS through a 25 gauge needle scored and dyed as above. The endothelial cell injection site was approximately 10 mm caudal to the melanoma injection site. Mice were left until tumors reached 8-10 mm when measured through the skin. The animals were then killed by cervical dislocation, tumors excised, the skin laid out and photographed. Directional growth of the melanoma was measured by drawing a line on the image from the two tattoo points, then taking a perpendicular axis at the melanoma injection point. Measurements were done blind. The plan area of the tumor below the perpendicular axis (i.e. closer to the endothelial cell injection site) was measured using Image J, and expressed as a percentage of the total area of the tumor. A value of 100% represents complete growth towards the LEC site, a value of 50% no directional growth and a value of 0% complete growth away from the injection site.

Kriehuber E, Breiteneder-Geleff S, Groeger M, Soleiman A, Schoppmann SF, Stingl G, Kerjaschki D and Maurer D. (2001). J Exp Med, 194, 797-808.

Shields JD, Borsetti M, Rigby H, Harper SJ, Mortimer PS, Levick JR, Orlando A and Bates DO. (2004). Br J Cancer, 90, 693-700.