Molecular Characterization of Dermal Lymphatic Endothelial Cells from Primary Lymphedema Skin

Abstract

Background:

Lymphatic endothelial cells from primary lymphedema skin have never been cultured nor characterized. A subgroup of patients with primary lymphedema undergo surgery to bring about an improvement in their quality of life. The aim of this study was to culture and characterize LECs from the skin of these patients.

Methods and Results:

Lymphatic endothelial cells were isolated and cultured from the skin of patients with primary lymphedema and from normal skin. The isolated cells were compared in their ability to form microvascular networks in a three-dimensional culture medium, and in their response to treatment with vascular endothelial growth factors A, C, and D. Whole tissue transcriptional profiling was carried out on two pools of isolated lymphatic endothelial cells—one from primary lymphedema skin and the other from normal skin. Lymphatic endothelial cells from primary lymphedema skin form tubule-like structures when cultured in three-dimensional media. They respond in a similar fashion to stimulation with the vascular endothelial growth factors A, C, and D. Comparative analysis between lymphedema tissue and normal tissue (fold change >2) showed differential expression of 2793 genes (5% of all transcripts), 2184 upregulated, and 609 downregulated. Genes involved in cellular apoptosis (vascular endothelial growth inhibitor, zinc finger protein), extracellular matrix turnover (matrix metalloproteinase inhibitor-16), and type IV collagen deposition were upregulated. Various pro-inflammatory genes (interleukin-6, interleukin-8, interleukin-32, E-selectin) were downregulated.

Conclusion:

Cellular adhesion, apoptosis, and increased extracellular matrix turnover play a more prominent role in primary lymphedema than previously thought. In addition, the acute inflammatory response is attenuated as evidenced by the downregulation of various pro-inflammatory genes.This sheds further light on the interplay of the various pathological processes taking place in primary lymphedema.

Introduction

Lymphedema is the result of excessive accumulation of extracellular fluid in the interstitial space as a consequence of defective drainage through the lymph-conducting system (lymphatics and lymph glands).¹ Patients with primary lymphedema have an intrinsic abnormality of the lymphatic system.¹

Nonne, in 1891, was the first to describe congenital lymphedema of the lower limbs.² Milroy, in 1892, described the same condition affecting 26 persons in a single family, spanning six generations.³ This form of lymphedema, characterized by unilateral or bilateral lower limb swelling, with a family history, was classified as primary or congenital lymphedema. Other forms of heritable lymphedema have since been described, including Meige's disease and lymphedema distichiasis. Meige's disease is primary lymphedema of the lower limbs, presenting in puberty, with a genetic predisposition in 30 percent of cases.⁴ The lymphedema is usually mild, rarely extends above the knee and is generally bilateral. Lymphedema distichiasis is congenital lymphedema associated with developmental abnormalities of the eyelids.⁵ Milroy disease⁶ and lymphedema distichiasis⁷ have been linked to specific gene mutations on chromosomes five (5q34–q35) and 16 (16q24.3), respectively. Defective vascular endothelial growth factor receptor three (VEGFR-3) signaling has been identified as the cause of congenital hereditary lymphedema in Milroy disease.⁸ The VEGFR-3 receptor binds the ligands vascular endothelial growth factor (VEGF)-C and VEGF-D. Lymphedema distichiasis, on the other hand, is caused by a mutation in the gene expressing the transcription factor FOXC2.⁹

Lymphatic disease is quite prevalent, and often not well clinically characterized.^10,11 Apart from lymphedema, other human diseases directly or indirectly affect or alter lymphatic structure and function, with a wide array of clinical signs and symptoms.¹² Progress in our understanding of the molecular mechanisms controlling the lymphatic system has lead to new insights in the role of lymphatic endothelium in inflammation, vascular disorders, and cancer dissemination.^13–15

This study focused on a small subgroup of patients with severe lower limb or genital primary lymphedema. These patients with massive limb swelling or genital lymphedema are often offered surgery, after an initial trial of conservative management. Limb and genital reduction surgery has been shown, in patients with severe primary lymphedema, to provide symptomatic relief, and an improvement in quality of life.¹⁶ Lymphatic endothelial cells have been isolated from animals and normal human skin, using antibodies to lymphatic endothelial cell markers.^17–23 The hyaluronan receptor LYVE-1 and the transmembrane glycoprotein podoplanin are expressed by LECs in vitro^17–21 and antibodies to these cell markers have been used to isolate LECs from normal human and bovine skin.

The aims of this study were to isolate and characterize LECs from the skin of patients with primary lymphedema. The ability of these LECs to form tubules, their response to stimulation with growth factors, and the gene profile expressed by these cells was then compared to that of LECs isolated from normal skin.

Patients, Materials and Methods

The Academic Department of Surgery, St. Thomas Hospital, London, is a tertiary referral center for patients being considered for lymphedema surgery. Patients with primary lymphedema that were referred to the surgical outpatient clinic for consideration for surgery fell into two main groups:

those with massive unilateral or bilateral limb swelling.

those with genital swelling.

Control skin was obtained from patients undergoing cosmetic plastic surgical procedures, such as breast reduction surgery or abdominoplasty procedures (excision of redundant abdominal wall skin). Normal breast and abdominal skin is not an ideal ‘control’ for limb and genital lymphedema tissue, however, there are few procedures carried out on the legs that would permit the excision of skin for research purposes. Samples of skin were obtained with ethical approval and informed consent (St. Thomas's Hospital ethics approval number EC 02/064) from a total of 23 patients.

Isolation and cell culture of LECs

Skin samples were obtained and incubated overnight with dispase (50 U/ml; Gibco) for 30 min at 37°C, with subsequent removal of the epidermis as previously described.¹⁷ Endothelial cells were released from the skin by squeezing and scraping. The endothelial cells were then pelleted, and re-suspended in endothelial cell growth medium (EBM-2, Cambrex Bio Science, UK). Cells were then cultured on gelatin-coated flasks to obtain a colony of cobblestone cells.

Purification of LECs

Cells were detached from the culture flasks by incubation for 10 min at 37°C with accutase (PAA laboratories, Somerset, UK). 10 ml of MACS buffer (2 mM EDTA, 0.5% bovine serum albumin in PBS; Miltenyi Biotech, Surrey, UK) was then added to the cell solution to neutralize the action of the accutase. Cells were incubated with 340 μg/ml of the lymphatic specific antibody rabbit anti-LYVE1 polyclonal antibody and rabbit anti-podoplanin antibody (Research Diagnostics, Inc.) at 4°C for 30 min. The cells were centrifuged at 210 g force and then re-suspended in 2 ml of MACS buffer. The antibody-coated LECs were purified from this suspension by incubation with goat anti-rabbit IgG magnetic micro beads (MACS micro beads; Miltenyi Biotech) at 4°C for 30 min. The cell suspension was then passed through a magnetic column (Miltenyi Biotech) and re-suspended in 5 ml of EBM-2. This was then cultured on gelatin-coated flasks for 2–5 days until a confluent colony of cells was obtained.

Immunocytochemical staining of isolated cells

Antibodies to the endothelial cell markers CD31 and CD44 and lymphatic specific markers (anti-LYVE 1 and anti-podoplanin) were used. Although CD44 is a common protein expressed on most blood vascular endothelial cells, blood cells, epithelial cells and stromal cells, its absence from LECs makes it a useful negative marker.^13,17 Cells purified and cultured as described above, were detached from culture flasks by incubation for 10 min at 37°C with accutase. 0.5 ml of cell suspension was seeded onto separate chambers on an 8-chamber slide (Sigma, UK), and cultured for 48 h to reach confluence. Endothelial cells were then washed with phosphate-buffered saline (PBS; Sigma, UK), and incubated for 30 min at room temperature with rabbit anti-LYVE1 antibody (2 μg/ml) and one of either mouse anti-podoplanin antibody (2 μg/ml, Research Diagnostics, Inc), mouse anti-CD31 antibody (8 μg/ml in PBS, Dako, Cambridge, UK) or mouse anti-CD44 antibody (8 μg/ml. in PBS, Dako). Cells were then washed with PBS, and incubated with Alexa fluor 568 conjugated goat anti-rabbit IgG and Alexa fluor 488 conjugated goat anti-mouse (2 μg/ml in PBS, Molecular Probes, UK) for 30 min at room temperature. The slides were then washed with PBS, and mounted with vectashield containing the blue fluorescent nuclear stain DAPI (Vector Laboratories, Peterborough, UK). Images were captured with a Leica (Leitz DMRB) fluorescent microscope at X 200 eyepiece magnification.

Tubule formation assays

Lymphatic endothelial cells from all three sources (normal skin, limb, and genital lymphedema skin) were cultured on a three-dimensional medium (Matrigel). This was to assess whether the abnormality in this group of patients with primary lymphedema was due to an inability of the LECs to adhere and interact with each other in a three-dimensional matrix. A positive finding would be the inability of LECs isolated from limb or genital lymphedema skin to form capillary-like networks when cultured on Matrigel.

To assess in vitro tubule formation, confluent LECs (2.5 × 10⁵) were seeded onto Matrigel-coated 24-well culture plates (BD Biosciences, Oxford, UK).²⁴ Culture plates were inspected under phase contrast light microscopy at 24 h to assess for the formation of tubules.

Proliferation assays with XTT

Lymphatic endothelial cells from different sources were compared in their ability to respond to the growth factors VEGF-A, VEGF-C, and VEGF-D. The controls for each of these experiments were the same set of LECs, untreated with any growth factor. This was to determine if defective VEGF-C, -D/VEGFR-3 signaling was linked to the etiology of primary lymphedema in this subgroup of patients. The authors expected to find significant differences (normal response) in the proliferation curves of LECs from all sources skin treated with VEGF-A (comparing treated and untreated LECs). In addition, the authors expected to find a similar response in the proliferation curve of LECs from normal skin treated with VEGF-C and VEGF-D. However, in the case of LECs from limb and genital lymphedema skin treated with VEGF-C and VEGF-D, the authors expected to find an impaired response (assuming the defect in this subgroup of patients was linked to defective VEGF-C, -D/VEGFR-3 signaling).

LECs (5 × 10⁴) were seeded onto fibronectin-coated 96-well plates. Triplicate dishes were treated without or with 100 ng/ml of VEGF-A, VEGF-C, and VEGF-D (R & D Systems, Abingdon, UK) in endothelial growth medium. Culture plates were then incubated at 37°C, and the rate of cellular proliferation measured by the addition of XTT and estimation of the color change on a colorimeter.²⁵

The cell solution (90 μl), containing 2 × 10⁴ LECs /ml was pipetted into 96-well plates. Endothelial growth medium (10 μl) containing 100 ng of growth factor (one of either VEGF A, VEGF C, or VEGF D) or 10 μl of medium alone (controls) was added into each well. Four identical plates were set up for each LEC isolate, and the number of cells in each well measured on days 0, 2, 4, and 7 using the XTT ({2, 3-bis (2-methoxy-4-nitro-5-sulfophenyl)-5-[.(phenylamino) carbonyl ]-2H-tetrazolium hydroxide}) method 2.²⁵

Proliferation curves and statistical analysis

To estimate the rate of proliferation of each LEC type, the XTT absorbance ratio was plotted against time, using the graph pad software package PRISM. The XTT absorbance ratio was obtained by dividing the values for the absorbance of XTT on days 0, 2, 4, and 7, by the value obtained on day 0. The mean absorbance ratio (and standard error of the mean) for each LEC type (normal LECs and LECs from limb or genital lymphedema skin) was then calculated at the different time points (days 2, 4, and 7).

This produced different proliferation curves for each LEC type treated with the growth factors VEGF-A, VEGF-C, VEGF-D, and for untreated LECs. The treatment of each group of the three LEC types with growth factor VEGF-A, -C, or -D was then compared with untreated LECs by a two-way analysis of variance (ANOVA), with the graph pad software package PRISM.

RNA extraction

Isolated LECs from five lymphedema tissue samples and five normal skin samples were pooled into two separate samples, respectively. This was due to the fact that it was difficult to obtain adequate RNA for microarray analysis from a single sample from one individual. Total RNA was isolated using the manufacturer's protocol (RNeasy Mini kit, Qiagen, West Sussex, UK). The cell pellets were centrifuged for 5 min at 300 g and the supernatant carefully removed by aspiration. The cells were then lysed using buffer RLT (Qiagen). The lysate was homogenized by passing it through a blunt 20-gauge needle connected to an RNAse-free syringe, repeating the process five times. 70% ethanol was then added to the lysate and the sample transferred to an RNeasy spin column which was then centrifuged and the flow-through discarded. Buffer RW1 (Qiagen) was then added to the spin column and then centrifuged at 8000 g for 15 sec. DNAse I incubation mix was then added directly to the spin column and left to stand for 15 min at room temperature. This optional step was added in order to decrease possible contamination with genomic DNA, which could affect the specificity and sensitivity of the downstream microarray experiment. The spin column was then washed by centrifugation with buffer RPE (Qiagen). Total RNA was then eluted by adding 50 μl of RNAse-free water to the spin column and stored at −80°C. The purity of the RNA was then analyzed using the Bioanalyser 2100 (Agilent, Santa Clara, CA) and the Nanodrop (ND-100).

Microarray analysis

The microarray analysis was performed to look for a differential fold change between LECs from lymphedema skin and normal skin, respectively. The aim of the microarray experiment was to look for putative genes involved in the pathogenesis of primary lymphedema. Biotin-labeled cDNA was prepared from the total RNA using reagents and protocols provided by Affymetrix (Affymetrix, Santa Clara, CA). The labeled cDNA samples were then hybridized to the Human Genome U133 Plus 2.0 microarray (Affymetrix). This genechip is an oligonucleotide microarray that allows detection of up to 54,675 transcripts encompassing the whole human genome. The genechip is then stained with streptavidin–phycoerythrin which binds to biotin and fluoresces under the laser in the scanner. The intensity of the fluorescence reflects the level of expression of the gene. The quality of the raw data from the genechips were analyzed with the GCOS system (Affymetrix).

The microarray data from the two samples were imported into the Genespring software (Agilent Technologies). Differential gene expression between the two samples was analyzed. The differentially expressed genes were then classified according to the following molecular functions: apoptosis, proteolysis, inflammation, and angiogenesis. Gene ontology analysis was carried out by uploading the differentially expressed data into GeneCoDis2 (www.genecodis.dacya.ucm.es), a web-based microarray analysis tool.²⁶ The GeneCoDis2 program determines annotations (cellular component, biological process, and molecular function) that are over-represented in the differentially expressed group of genes with respect to a reference set of genes. Statistical analysis is carried out using hypergeometric distribution²⁷ and significant annotations with a p value <0.05 are computed. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was performed for the differentially expressed genes and scored for significance.

Results

Isolation from primary lymphedema skin

Fifteen patients with primary lymphedema consented to the use of skin samples from tissue excised at surgery. LECs were successfully isolated from the skin of 13 out of these 15 patients (Table 1). Eight had limb lymphedema and five had genital lymphedema. In addition, LECs were successfully isolated from 10 normal skin samples. The initial process of isolation of LECs from primary lymphedema skin yielded a confluent layer of cobblestone cells in 2–5 days (Fig. 1). Immunofluorescence staining of this unselected cell population, confirmed co-localization of the lymphatic marker LYVE-1 and the pan-endothelial cell marker CD31 to a subset of cells (Figs. 2A–2C). Purification of this unselected cell population yielded a colony of confluent colony of cells that stained positive for both the lymphatic markers LYVE-1 and podoplanin (Figs. 3A–3C).

FIG. 1.

Primary culture of unselected endothelial cells from primary lymphedema skin. Bar = 100 μm.

FIG. 2.

Immunofluorescent staining of a slide of endothelial cells isolated from primary lymphedema skin. The lymphatic marker LYVE-1 (2A) and the panoendothelial marker CD31 (2B) co-localized to a subset of cells (2C). Bar = 100 μm. This figure is available in color in the online version of this article at www.liebertpub.com.

FIG. 3.

A slide of purified LECs from primary lymphedema skin, stained with the lymphatic markers LYVE-1 (3A; green) and podoplanin (3B; red). The merged image (3C) shows that all the endothelial cells take up at least on lymphatic specific stain (LYVE-1 or podoplanin). Bar = 100 μm. This figure is available in color in the online version of this article at www.liebertpub.com.

Table 1.

Patient Characteristics

	Age	Sex	Tissue Type	Genetic Abnormalities
1	26	Male	Limb skin (lymphedema)	None
2	25	Male	Limb skin (lymphedema)	None
3	38	Female	Limb skin (lymphedema)	None
4	69	Female	Limb skin (lymphedema)	None
5	32	Female	Limb skin (lymphedema)	None
6	45	Female	Limb skin (lymphedema)	None
7	40	Female	Limb skin (lymphedema)	None
8	35	Male	Limb skin (lymphedema)	None
9	46	Male	Scrotal skin (lymphedema)	None
10	60	Male	Penile skin (lymphedema)	None
11	26	Male	Scrotal skin (lymphedema)	None
12	48	Male	Scrotal skin (lymphedema)	None
13	48	Male	Scrotal skin (lymphedema)	None
14	35	Female	Breast Skin (normal)	None
15	36	Female	Breast Skin (normal)	None
16	28	Male	Genital skin (normal)	None
17	33	Female	Abdominal skin (normal)	None
18	29	Female	Breast Skin (normal)	None
19	61	Female	Breast Skin (normal)	None
20	25	Female	Breast Skin (normal)	None
21	28	Female	Breast Skin (normal)	None
22	26	Female	Breast Skin (normal)	None
23	60	Male	Back skin (normal)	None

Primary lymphedema LECs form vascular tubes on Matrigel

There was no difference in the ability of LECs from all sources to form networks on three-dimensional matrix (Matrigel) (Figs. 4A–4C).

FIG. 4.

Lymphatic endothelial cells from normal skin (4A), genital (4B), and limb (4C) lymphedema skin are capable of forming tubular networks when cultured on matrigel.

Proliferation of LECs in response to VEGF treatment

VEGF-A

VEGF-A treatment significantly increased the rate of proliferation of all three LEC types (P < 0.01, Figs. 5A–5C).

FIG. 5.

The effect of VEGF-A on LECs isolated from (A) normal skin (10 patients), (B) limb lymphedema skin (8 patients), and (C) genital lymphedema skin (5 patients). Controls were LECs from all three sources not treated with any VEGF-A (broken lines). LECs from all three sources responded to stimulation with VEGF-A.

VEGF-C and VEGF-D

VEGF-C and VEGF-D treatment on all LECs isolated from lymphedema and normal skin had very little effect on cellular proliferation (Figs 6A–6C and Figs. 7A–7C).

FIG. 6.

The effect of VEGF-C on LECs isolated from (A) normal skin (10 patients), (B) limb lymphedema skin (8 patients), and (C) genital lymphedema skin (5 patients). Controls were LECs from all three sources not treated with any VEGF-C (broken lines). LECs from all three sources showed no response to stimulation with VEGF-C.

FIG. 7.

The effect of VEGF-D on LECs isolated from (A) normal skin (10 patients), (B) limb lymphedema skin (8 patients), and (C) genital lymphedema skin (5 patients). Controls were LECs from all three sources not treated with any VEGF-D (broken lines). LECs from all three sources responded to stimulation with VEGF-D.

RNA extraction

The RNA extracted for both samples were of high quality. Both samples had an RNA integrity number (RIN) of 9.4 (RIN > 5.5, pure RNA). The quality of the RNA as measured by the A260/280 ratio on the nanodrop was 1.9 in each case (1.9–2.1, pure RNA). The nanodrop results also confirm minimal contamination with genomic DNA, protein, and phenol.

Microarray analysis

cDNA were successfully prepared from the RNA samples. Comparative analysis between lymphedema tissue and normal tissue (fold change > 2) showed differential expression of 2793 genes (5% of all transcripts), 2184 genes were upregulated, and 609 downregulated. Differentially expressed genes (3-fold or more) comprised genes involved in the processes of inflammation, cellular adhesion and motility, apoptosis, extracellular matrix turnover, and collagen deposition (Table 2).

Table 2.

All Genes

Probe set ID	Gene Acession No.	Gene name	Fold Change
Upregulated genes
219452_at	NM_022355	dipeptidase 2	56
228492_at	NM_004654	ubiquitin specific peptidase 9, Y-linked (fat facets-like, Drosophila)	25
221085_at	NM_005118	tumor necrosis factor (ligand) superfamily, member 15	18
208166_at	NM_005941	matrix metallopeptidase 16 (membrane-inserted)	15
240135_x_at		TIMP metallopeptidase inhibitor 3	13
222043_at	NM_001831	clusterin	12
209474_s_at	NM_001098175	ectonucleoside triphosphate diphosphohydrolase 1	11
222073_at	NM_000091	collagen, type IV, alpha 3 (Goodpasture antigen)	11
225660_at	NM_020796	sema domain, transmembrane domain ™	10
236191_at	NM_001775.2	CD38 antigen (p45)	10
236491_at	NM_020396	BCL2-like 10 (apoptosis facilitator)	10
204442_x_at	NM_001042544	latent transforming growth factor beta binding protein 4	9
239778_x_at	NM_014296	Calpain 7	9
1552365_at	NM_001112706	scinderin	9
215561_s_at	NM_000877	interleukin 1 receptor, type I	9
235166_at	NM_021964	Zinc finger protein 148 (pHZ-52)	9
227863_at	NM_001909	cathepsin D (lysosomal aspartyl peptidase)	8
235688_s_at	NM_004295	TNF receptor-associated factor 4	8
216320_x_at	NM_02099	macrophage stimulating 1 (hepatocyte growth factor-like)	8
211298_s_at	NM_000477	albumin	7
238252_at	NM_001001976.1	Arginyltransferase 1	7
233645_s_at	NM_016546	complement component 1, r subcomponent-like	7
234179_at	NM_001177506.1	Acyloxyacyl hydrolase (neutrophil)	7
243520_x_at	NM_021794	ADAM metallopeptidase domain 30	7
205392_s_at	NM_004166	chemokine (C-C motif) ligand 14 ; chemokine (C-C motif) ligand 15	7
1554997_a_at	NM_000963	prostaglandin-endoperoxide synthase 2	7
242628_at	NM_002259.4	Killer cell lectin-like receptor subfamily B, member 1	7
32625_at	NM_000906	natriuretic peptide receptor A/guanylate cyclase A	6
229004_at	NM_139055.2	ADAM metallopeptidase with thrombospondin type 1 motif, 15	6
222032_s_at	NM_003470	Ubiquitin specific peptidase 7 (herpes virus-associated)	6
219304_s_at	NM_025208	platelet derived growth factor D	6
205756_s_at	NM_000132	coagulation factor VIII, procoagulant component (hemophilia A)	6
240397_x_at	NM_130386.2	Collectin sub-family member 12	6
201902_s_at	NM_003403	YY1 transcription factor	6
203919_at	NM_003195	transcription elongation factor A (SII), 2	6
231724_at	NM_004831	cofactor required for Sp1 transcriptional activation, subunit 7, 70kDa	6
206685_at	NR_002139	HLA complex group 4	6
240392_at	NM_024110.2	Caspase recruitment domain family, member 14	6
215976_at	NM_014618.2	Deleted in bladder cancer 1	6
209230_s_at	NM_001042483	p8 protein (candidate of metastasis 1)	6
208536_s_at	NM_006538	BCL2-like 11 (apoptosis facilitator)	6
1566709_at	NM_001077654.1	Tumor necrosis factor, alpha-induced protein 8	5
209619_at	NM_001025158	CD74 antigen	5
210944_s_at	NM_000070	calpain 3, (p94)	5
239101_at	NM_031483	itchy homolog E3 ubiquitin protein ligase (mouse)	5
200766_at	NM_001909	cathepsin D (lysosomal aspartyl peptidase)	5
228982_s_at	NM_032172	Ubiquitin specific peptidase 42	5
213176_s_at	NM_001042544	latent transforming growth factor beta binding protein 4	5
221019_s_at	NM_130386	collectin sub-family member 12 ; collectin sub-family member 12	5
222860_s_at	NM_025208	platelet derived growth factor D	5
239101_at	NM_031483	itchy homolog E3 ubiquitin protein ligase (mouse)	5
218831_s_at	NM_001136019	Fc fragment of IgG, receptor, transporter, alpha	5
209619_at	NM_001025158	CD74 antigen	5
216248_s_at	NM_006186	nuclear receptor subfamily 4, group A, member 2	5
221019_s_at	NM_130386	collectin sub-family member 12	5
210374_x_at	NM_000957	prostaglandin E receptor 3 (subtype EP3)	5
241340_at	NM_001225.3	Caspase 4, apoptosis-related cysteine peptidase	4
211478_s_at	NM_001935	dipeptidylpeptidase 4 (CD26, adenosine deaminase complexing protein 2)	4
227463_at	NM_000789	angiotensin I converting enzyme (peptidyl-dipeptidase A) 1	4
213397_x_at	NM_002937	ribonuclease, RNase A family, 4	4
213257_at	NM_015077	sterile alpha and TIR motif containing 1	4
206693_at	NM_000880	interleukin 7	4
206682_at	NM_006344	C-type lectin domain family 10, member A	4
236897_at	NM_018725.3	Interleukin 17 receptor B	4
211478_s_at	NM_001935	dipeptidylpeptidase 4 (CD26, adenosine deaminase complexing protein 2)	4
204936_at	NM_004579	mitogen-activated protein kinase kinase kinase kinase 2	4
204623_at	NM_003226	trefoil factor 3 (intestinal)	4
241340_at	NM_001225.3	Caspase 4, apoptosis-related cysteine peptidase	4
215388_s_at	NM_000186	complement factor H ; complement factor H-related 1	4
202014_at	NM_014330	protein phosphatase 1, regulatory (inhibitor) subunit 15A	4
206360_s_at	NM_003955	suppressor of cytokine signaling 3	4
235745_at	NM_001433	endoplasmic reticulum to nucleus signalling 1	4
203657_s_at	NM_003793	cathepsin F	3
206415_at	NM_012464	tolloid-like 1	3
227256_at	NM_020718	ubiquitin specific peptidase 31	3
203501_at	NM_016134	plasma glutamate carboxypeptidase	3
201212_at	NM_001008530	legumain	3
210628_x_at	NM_001042544	latent transforming growth factor beta binding protein 4	3
214674_at	NM_006677	ubiquitin specific peptidase 19	3
204262_s_at	NM_000447	presenilin 2 (Alzheimer disease 4)	3
220735_s_at	NM_001077203	SUMO1/sentrin specific peptidase 7	3
202087_s_at	NM_001912	cathepsin L	3
217865_at	NM_018434	ring finger protein 130	3
215013_s_at	NM_014709	ubiquitin specific peptidase 34	3
1564736_a_at	NR_000035	caspase 12 pseudogene 1	3
1558117_s_at	NM_020718	ubiquitin specific peptidase 31	3
202701_at	NM_001199	bone morphogenetic protein 1	3
226619_at	NM_014554	SUMO1/sentrin specific peptidase 1	3
215204_at	NM_001100409.1	SUMO1/sentrin specific peptidase 6	3
39582_at	NM_001042355	Cylindromatosis (turban tumor syndrome)	3
225214_at	NR_027406	Proteasome (prosome, macropain) subunit, beta type, 7	3
240295_at	NM_002581	Pregnancy-associated plasma protein A, pappalysin 1	3
205141_at	NM_001097577	angiogenin, ribonuclease, RNase A family, 5	3
211889_x_at	NM_001024912	carcinoembryonic antigen-related cell adhesion molecule 1	3
219257_s_at	NM_001142601	sphingosine kinase 1	3
224027_at	NM_148672	chemokine (C-C motif) ligand 28	3
201998_at	NM_003032	ST6 beta-galactosamide alpha-2,6-sialyltranferase 1	3
211332_x_at	NM_000410	hemochromatosis	3
203528_at	NM_001142287	sema domain, immunoglobulin domain (Ig),	3
200904_at	NM_005516	major histocompatibility complex, class I, E	3
217502_at	NM_001547	interferon-induced protein with tetratricopeptide repeats 2	3
238735_at	NM_003205.3	Transcription factor 12 (HTF4, helix-loop-helix transcription factors 4)	3
209648_x_at	NM_014011	suppressor of cytokine signaling 5	3
207416_s_at	NM_004555	nuclear factor of activated T-cells, cytoplasmic, calcineurin-dependent 3	3
204863_s_at	NM_002184	interleukin 6 signal transducer (gp130, oncostatin M receptor)	3
235255_at	NM_012463	ATPase, H + transporting, lysosomal V0 subunit a isoform 2	3
228813_at	NM_006037	histone deacetylase 4	3
210481_s_at	NM_001144904	C-type lectin domain family 4, member M	3
222354_at	NM_016946	F11 receptor	3
211087_x_at	NM_001315	mitogen-activated protein kinase 14 ; mitogen-activated protein kinase 14	3
238846_at	NM_003839	tumor necrosis factor receptor superfamily, member 11a, NFKB activator	3
205419_at	NM_004951	Epstein-Barr virus induced gene 2	3
226757_at	NM_001547	interferon-induced protein with tetratricopeptide repeats 2	3
1564733_at	NM_001143826.1	Microtubule-associated protein, RP/EB family, member 2	3
223508_at	NM_017617	Notch homolog 1, translocation-associated (Drosophila)	3
211530_x_at	NM_002127	HLA-G histocompatibility antigen, class I, G	3
209647_s_at	NM_014011	suppressor of cytokine signaling 5	3
204057_at	NM_002163	interferon regulatory factor 8 ; interferon regulatory factor 8	3
211610_at	NM_001160124	Kruppel-like factor 6 ; Kruppel-like factor 6	3
210514_x_at	NM_002127	HLA-G histocompatibility antigen, class I, G	3
232466_at	NM_001008895	Cullin 4A	3
204655_at	NM_002985	chemokine (C-C motif) ligand 5	3
241808_at	NM_016010	Interleukin 7	3
202257_s_at	NM_006110	CD2 antigen (cytoplasmic tail) binding protein 2	3
210555_s_at	NM_004555	nuclear factor of activated T-cells, cytoplasmic, calcineurin-dependent 3	3
216557_x_at	XM_001718037	immunoglobulin heavy constant alpha 1	3
200905_x_at	NM_005516	major histocompatibility complex, class I, E	3
202455_at	NM_001015053	histone deacetylase 5	3
208127_s_at	NM_014011	suppressor of cytokine signaling 5	3
212508_at	NM_022151	modulator of apoptosis 1	3
210025_s_at	NM_014550	caspase recruitment domain family, member 10	3
202390_s_at	NM_002111	huntington (Huntington disease)	3
1558143_a_at	NM_006538	BCL2-like 11 (apoptosis facilitator)	3
225078_at	NM_001424	epithelial membrane protein 2	3
210148_at	NM_001048200	homeodomain interacting protein kinase 3	3
222564_at	NM_017542	pogo transposable element with KRAB domain	3
219257_s_at	NM_001142601	sphingosine kinase 1	3
225079_at	NM_001424	epithelial membrane protein 2	3
209878_s_at	NM_001145138	v-rel reticuloendotheliosis viral oncogene homolog A	3
240964_at	NM_000314.4	Phosphatase and tensin homolog	3
207002_s_at	NM_001080951	pleiomorphic adenoma gene-like 1	3
216766_at	NM_005400.2	Protein kinase C, epsilon	3
204262_s_at	NM_000447	presenilin 2 (Alzheimer disease 4)	3
202723_s_at	NM_002015	forkhead box O1A (rhabdomyosarcoma)	3
208588_at	NM_021631	apoptosis inhibitor	3
217865_at	NM_018434	ring finger protein 130	3
219500_at	NM_013246	cardiotrophin-like cytokine factor 1	3
1564736_a_at	NR_000035	caspase 12 pseudogene 1	3
221560_at	NM_031417	MAP/microtubule affinity-regulating kinase 4	3
210655_s_at	NM_001455	forkhead box O3A	3
232753_at	NM_012279	Zinc finger protein 346	3
201964_at	NM_015046	amyotrophic lateral sclerosis 4	3
202512_s_at	NM_004849	ATG5 autophagy related 5 homolog (S. cerevisiae)	3
210026_s_at	NM_014550	caspase recruitment domain family, member 10	3
211658_at	NM_005809	peroxiredoxin 2	3
222533_at	NM_016302	cereblon	3
Down regulated genes
204614_at	NM_001143818	serpin peptidase inhibitor, clade B (ovalbumin), member 2	111
206211_at	NM_000450	selectin E (endothelial adhesion molecule 1)	21
242791_at	NM_012175	F-box protein 3	14
201860_s_at	NM_000930	plasminogen activator, tissue	14
211884_s_at	NM_000246	class II, major histocompatibility complex, transactivator	13
210511_s_at	NM_002192	inhibin, beta A (activin A, activin AB alpha polypeptide)	12
226218_at	NM_002185	Interleukin 7 receptor	11
204475_at	NM_001145938	matrix metallopeptidase 1 (interstitial collagenase)	9
204472_at	NM_005261	GTP binding protein overexpressed in skeletal muscle	9
220390_at	NM_024783	ATP/GTP binding protein-like 2	8
205083_at	NM_001159	aldehyde oxidase 1	7
205680_at	NM_002425	matrix metallopeptidase 10 (stromelysin 2)	7
211668_s_at	NM_001145031	plasminogen activator, urokinase	6
206639_x_at	NM_002159	histatin 1	6
206786_at	NM_000200	histatin 3 ; cingulin-like 1	6
237837_at	NM_000633.2	B-cell CLL/lymphoma 2	6
204534_at	NM_000638	vitronectin	5
209795_at	NM_001781	CD69 antigen (p60, early T-cell activation antigen)	5
227759_at	NM_174936	proprotein convertase subtilisin/kexin type 9	5
207857_at	NM_001130917	leukocyte immunoglobulin-like receptor	5
228598_at	NM_001004360	dipeptidylpeptidase 10	4
234131_at	NM_001110.2	ADAM metallopeptidase domain 10	4
205832_at	NM_001163446	carboxypeptidase A4	4
1562915_at	NM_006438.3	Collectin sub-family member 10 (C-type lectin)	4
222856_at	NM_017413	apelin, AGTRL1 ligand	4
211506_s_at	NM_000584	interleukin 8	4
204972_at	NM_001032731	2′-5′-oligoadenylate synthetase 2, 69/71kDa	4
241640_at	NM_001008405	B-cell receptor-associated protein 29	4
201601_x_at	NM_003641	interferon induced transmembrane protein 1 (9-27)	4
211506_s_at	NM_000584	interleukin 8	4
205927_s_at	NM_001910	cathepsin E	3
204575_s_at	NM_002429	matrix metallopeptidase 19	3
222162_s_at	NM_006988	ADAM metallopeptidase with thrombospondin type 1 motif, 1	3
212190_at	NM_001136528	serpin peptidase inhibitor, clade E	3
239554_at	NM_007282.4	Ring finger protein 13	3
202870_s_at	NM_001255	CDC20 cell division cycle 20 homolog (S. cerevisiae)	3
204575_s_at	NM_002429	matrix metallopeptidase 19	3
242397_at	NM_001172632.1	Oxidised low density lipoprotein (lectin-like) receptor 1	3
201601_x_at	NM_003641	interferon induced transmembrane protein 1 (9-27)	4
202411_at	NM_001130080	interferon, alpha-inducible protein 27	4
211193_at	NM_001126240.1	tumor protein p73-like	4
221371_at	NM_005092	tumor necrosis factor (ligand) superfamily, member 18	4
201601_x_at	NM_003641	interferon induced transmembrane protein 1 (9-27)	4
235236_at	NM_006566	Dedicator of cytokinesis 2	3
207733_x_at	NM_002784	pregnancy specific beta-1-glycoprotein 9	3
207526_s_at	NM_003856	interleukin 1 receptor-like 1	3
1553574_at	NM_176891	interferon epsilon 1	3
205579_at	NM_000861	histamine receptor H1	3
228176_at	NM_005226	endothelial differentiation, sphingolipid G-protein-coupled receptor, 3	3
205659_at	NM_014707	histone deacetylase 9	3
1555812_a_at	NM_001175	Rho GDP dissociation inhibitor (GDI) beta	3
203416_at	NM_000560	CD53 antigen	3
203828_s_at	NM_001012631	interleukin 32 ; interleukin 32	3
208306_x_at	NM_002124.2	Major histocompatibility complex, class II, DR beta 3	3
205207_at	NM_000600	interleukin 6 (interferon, beta 2)	3
207315_at	NM_006566	CD226 antigen	3
201288_at	NM_001175	Rho GDP dissociation inhibitor (GDI) beta	3
212353_at	NM_001128204	sulfatase 1	3
201710_at	NM_002466	v-myb myeloblastosis viral oncogene homolog (avian)-like 2	3
217139_at	NM_003374	voltage-dependent anion channel 1	3
202095_s_at	NM_001012270	baculoviral IAP repeat-containing 5 (survivin)	3
217607_x_at	NM_001042559	eukaryotic translation initiation factor 4 gamma, 2	3
242397_at	NM_001172632.1	Oxidised low density lipoprotein (lectin-like) receptor 1	3

KEGG pathway analysis was also used to group the differentially expressed genes into known biochemical or functional pathways (Fig. 8 and Fig. 9). Genes that were upregulated in LECs from primary lymphedema tissue were significantly involved in the following pathways: cancer, cytokine–cytokine, chemokine, insulin signaling, and calcium signaling. Genes involved in the VEGF signaling pathway, known to play an important role in lymphangiogenesis,²⁸ were also upregulated in LECs from lymphedema skin, as opposed to LECs from normal skin (Table 2).

FIG. 8.

KEGG pathway summary for upregulated genes in LECs isolated from primary lymphedema skin. Genes are grouped into known biochemical or functional pathways. This figure is available in color in the online version of this article at www.liebertpub.com.

FIG. 9.

KEGG pathway summary for downregulated genes in LECs isolated from primary lymphedema skin. Genes are grouped into known biochemical or functional pathways. This figure is available in color in the online version of this article at www.liebertpub.com.

Discussion

Techniques for isolating LECs using the lymphatic specific antibodies anti-LYVE1 with anti-podoplanin have allowed us to separate and culture LECs from primary lymphedema skin and compare it to LECs from normal skin.

LECs isolated from both normal and primary lymphedema skin have the ability to form capillary-like tube structures when cultured for 24 h on growth factor-depleted matrigel. This is independent of the effect of any of the vascular endothelial growth factors. These findings are consistent with those obtained by Podgrabinska et al,²⁰ who studied tube formation in LECs and BECs (blood endothelial cells) from neonatal foreskin, cultured on a collagen gel ‘sandwich’ assay for 48 h. It appears therefore that the lymphedema in these patients may not be the result of an inherent inability of their LECs to form tubules.

Normal, genital, and limb lymphedema LECs, cultured in endothelial growth medium, all respond in a similar fashion to VEGF A. Veikkola et al.²¹ reported similar findings with normal LECs isolated from human dermal microvascular endothelial cells treated with VEGF-A. They showed that the proliferation of these LECs, in response to VEGF-A, was two-fold higher than untreated cells.

VEGF-C and VEGF-D appeared to have little to no effect in isolation on LECs isolated from any of the three sources, when cultured in vitro. Veikkola et al.²¹ reported small differences (1–2 fold increase) in cell proliferation on stimulation of LECs with VEGF-C in vitro. This variation in their results compared to those from this study may be due to differences in the constituent composition of the endothelial culture mediums.

It would therefore appear that LECs from primary lymphedema skin respond in a similar fashion to the growth factors VEGF-A, -C, and -D, as LECs from normal skin. It may be that defective VEGFR-2 and VEGFR-3 signaling may not play a role in the etiology of this subgroup of primary lymphedema. The presentation and management of this subgroup of patients with primary lymphedema is completely different from the much rarer Milroy's disease phenotype.¹⁶ It is therefore reasonable to guess that the etiology of lymphedema in these patients may be unrelated to VEGFR-3 signaling.

In arriving at the results of the proliferation studies, we have made two assumptions. First, that LECs isolated from different patients, with the same clinical subtype of primary lymphedema, will respond in a similar fashion to the growth factors VEGF-A, -C, and -D. This may apparently not be the case as there may be variability in the individual responses to stimulation. In addition, we have also assumed that normal LECs from different sites in the body (breast and abdominal skin) will respond to growth factors in a similar fashion. Unfortunately, there are very few procedures carried out on legs that would permit excision of normal skin for research purposes.

The results from the microarray analysis point to differential expression of genes involved in the extracellular matrix (ECM), inflammation, and the processes of cellular turnover and adhesion, in LECs from lymphedema skin compared to normal skin.

Nelson et al,²⁹ in studies on the differential gene expression of primary lymphatic endothelial cells (compared to blood endothelial cells), from normal skin, reported the upregulation of genes linked to calcium signaling (15 genes) and cell adhesion (5 genes). Our microarray results from primary lymphedema skin point to a more prominent role for all these processes, as evidenced by the larger number of KEGG pathway-associated genes involved in calcium signaling and cell adhesion that were upregulated or downregulated. Hypoxia has been shown to alter the adhesive properties of LECs cultured in vitro,³⁰ by increasing attachment of LECs to the extracellular matrix. Genes involved in the processes of cell adhesion, apoptosis, and chemotaxis have been shown to be differentially upregulated or downregulated by LECs under hypoxic conditions.³⁰ It may be that hypoxia plays a major role in the clinical manifestations seen in primary lymphedema skin. Differential expression of genes involved in cell turnover, basic metabolism, and the cytoskeleton has been shown to be linked to the propagation of LECs cultured in vitro.³¹ This was associated with the downregulation of genes encoding extracellular matrix components.³¹ Our results from LECs from primary lymphedema skin favor reduced turnover in the components of the extracellular matrix. TIMP metallopeptidase inhibitor 3, an inhibitor of the matrix metalloproteinases, was upregulated. The matrix metalloproteinases are involved in the degradation of the extracellular matrix.³² MMP-16 was upregulated 15-fold in LECs from lymphedema skin. Also upregulated was the gene encoding for the major structural protein of the basement membrane, collagen type IV alpha 3³³ and collagen type IV alpha 5. It is tempting to speculate that the combination of reduced breakdown of the extracellular matrix, and increased deposition of certain types of collagen, may in part account for the thick, warty appearance of severe primary lymphedema tissue.

The genes encoding for the interleukins-6, -8, and -32 (IL-6, IL-8, and IL-32) were all downregulated in LECs from primary lymphedema skin. IL-8 is the major interleukin of the inflammatory process³⁴ and has been implicated in angiogenesis, chemotaxis, and cancer metastasis.³⁵ In addition, the gene encoding for the interleukin-7 receptor, known to play an important role in the activation of T-cells, B cells, and macrophages in inflammatory diseases like rheumatoid arthritis and psoriasis,^36,37 was also downregulated. Interestingly, the cell adhesion molecule, E-selectin, known to play a crucial role in leukocyte recruitment during acute inflammation³⁸ was also downregulated. E-selectin had been shown to be expressed in a dose dependent manner by cultured human LECs stimulated with lipopolysaccharides.³⁹ This pattern points towards a general suppression of the genes involved in the acute inflammatory process.

Differential gene expression of primary cultured LECs and blood vascular endothelial cells (BECs) have previously shown the upregulation of pro-inflammatory genes like selectin and chemokine (C-X-C motif) ligand 11.²⁹ In our study, E-selectin was downregulated 21-fold. A possible explanation for this may be the chronic nature of the inflammatory process occurring in this subset of patients with chronic lower limb and genital swelling. E-selectin is one of the cell adhesion molecules known to be expressed in response to the local release of cytokines.^38,39 It may be that the pathological process occurring in this subset of patients with primary lymphedema may be more of a chronic inflammatory process, with the suppression of pro-inflammatory genes. Another possible explanation may be differential patterns of expression of E-selectin in vitro. The expression of E-selectin by lymphatic endothelial cells in vitro has been shown to depend on how the LECs are cultured.⁴⁰ Tumor necrosis factor, member 15 (vascular endothelial growth inhibitor), upregulated 17-fold in lymphedema LEC isolates, has been implicated to have an inhibitory effect on cellular motility and adhesion in bladder cancer cells, and is thought to function as a negative regulator of aggressiveness during the development and progression of bladder cancer.^41,42 It has also been shown to induce apoptosis in endothelial cells and inhibit tumor neovascularization and tumor progression in cellular and animal models.^43,44 Zinc finger protein has been linked to the induction of apoptosis and enhancement of the effect of drugs on hepatocellular carcinoma cells.⁴⁵ Its expression has been linked to the suppression of E-cadherin and loss of cell–cell adhesion in human carcinoma cell lines, thereby favoring tumor invasion.⁴⁶ Taken together, this pattern of gene expression points towards increased cell turnover, consistent with transcriptomal studies of human dermal LECs propagated in vitro.³¹

In summary, this study has isolated and characterized for the first time, LECs from primary lymphedema skin (patients with massive limb and genital swelling). It is also the first time a whole transcriptome study has been performed on human primary lymphedema skin. LECs isolated from primary lymphedema skin respond to the growth factors VEGF-A, VEGF-C, and VEGF-D in a similar fashion to LECs from normal skin. These LECs are also capable of forming microtubules in culture. Transcriptional profiling of these LECs has shown a large number differentially expressed genes. Genes involved in cellular and extracellular matrix turnover and type IV collagen deposition were upregulated. Various pro-inflammatory genes were downregulated. Abnormalities in the basement membrane component, chronic inflammation, and intrinsic defects in extracellular matrix turnover may play a major role in the pathogenesis of primary lymphedema.

Future research work should entail in vivo studies with LECs cultured from primary lymphedema tissue, to see if the results above are replicated under physiological conditions.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Browse

, Stewart

. Lymphoedema: Pathophysiology and classification. J Cardiovasc Surg (Torino), 1985; 26:91–106.

Nonne

. Vier Falle von Elephantiasis congenital hereditaria. Virchows Archiv., 1891; 125:189–196.

Milroy

. An undescribed variety of hereditary oedema. NY Med J, 1892; 56:505–508.

Meige

. Dystrophie oedemateuse hereditaire. Press Medicale, 1898; 6:431–433.

Falls

. A new syndrome combining pterygium colli with developmental anomalies of the eyelids and lymphatics of the lower extremeties. Trans Am Ophthalmol, 1965; 62:248–275.

Evans

, Brice

, Sotirova

et al. Mapping of primary congenital lymphoedema to the 5q35.3 region. Am J Human Genet, 1999; 64:547–555.

Mangion

, Rahman

, Mansour

et al. A gene for lymphoedema-distichiasis maps to 16q24.3. Am J Human Genet, 1999; 65:427–432.

Karkkainen

, Ferrell

, Lawrence

et al. Missense mutations interfere with VEGFR-3 signalling in primary lymphoedema. Nat Genetics, 2000; 25:153–159.

Fang

, Dagenais

, Erickson

et al. Mutations in FOXC2 (MFH-1), a forkhead family transcription factor, are responsible for the hereditary lymphoedema-distichiasis syndrome. Am J Human Genet, 2000; 67:1382–1388.

10.

Rockson

. Lymphoedema. Am J Med, 2001; 110:288–295.

11.

Szuba

, Rockson

. Lymphoedema: Anatomy, physiology and pathogenesis. Vasc Med, 1997; 2:321–326.

12.

Radhakrishnan

, Rockson

. The clinical spectrum of lymphatic disease. Ann NY Acad Sci, 2008; 1131:155–184.

13.

Jurisic

, Detmar

. Lymphatic endothelium in health and disease. Cell Tissue Res, 2009; 335:97–108.

14.

Tammela

, Alitalo

. Lymphangiogenesis: Molecular mechanisms and future promise. Cell, 2010; 140:460–476.

15.

Cueni

, Detmar

. The lymphatic system in health and disease. Lymphat Res Biol, 2008; 6:109–122.

16.

Ogunbiyi

, Modarai

, Smith

et al. Quality of life after surgical reduction for severe primary lymphoedema of the limbs and genitalia. Br J Surg, 2009; 96:1274–1279.

17.

Kriehuber

, Breiteneder–Geleff

, Groeger

et al. Isolation and characterisation of dermal lymphatic and blood endothelial cells reveal stable and functionally specialized cell lineages. J Exp Med, 2001; 194:797–808.

18.

Leak

, Jones

. Lymphatic endothelium isolation, characterisation and longterm culture. Anat Rec, 1993; 236:641–652.

19.

Mizuno

, Yokoyama

, Ono

, Ikomi

, Ohhashi

. Establishment of rat lymphatic endothelial cell line. Microcirculation, 2003; 10:127–131.

20.

Podgrabinski

, Braun

, Velasco

et al. Molecular characterisation of lymphatic endothelial cells. Proc Natl Acad Sci USA, 2002; 99:16069–16074.

21.

Veikkola

, Lohela

, Ikenberg

et al. Intrinsic versus micro environmental regulation of lymphatic endothelial cell phenotype and function. FASEB J, 2003; 17:2006–2013.

22.

Weber

, Lorenzoni

, Lozzi

. Cytochemical differentiation between blood and lymphatic endothelium: Bovine blood and lymphatic large vessels and endothelial cells in culture. J Histochem Cytochem, 1994; 42:1109–1115.

23.

Whitehurst

, Eversgerd

, Flister

et al. Molecular profile and proliferative responses of rat lymphatic endothelial cells in culture. Lymphat Res Biol, 2006; 4:119–142.

24.

Hirakawa

, Hong

Y–K

, Harvey

et al. Identification of vascular lineage-specific genes by transcriptional profiling of isolated blood vascular and lymphatic endothelial cells. Am J Path, 2003; 162:575–586.

25.

Altman

. Tetrazolium salts and formazans. Prog Histochem Cytochem, 1976; 9:1–56.

26.

Nogales–Cadenas

, Carmona–Saez

, Vazquez

et al. GeneCodis: Interpreting gene lists through enrichment analysis and integration of diverse biological information. Nucleic Acids Res, 2009; 37Web Server issueW317–W322.

27.

Carmona–Saez

, Chagoyen

, Rodriguez

. et al. Integrated analysis of gene expression by association rules discovery. BMC Bioinformat, 2006; 7:54.

28.

Lohela

, Bry

, Tammela

, Alitalo

. VEGFs and receptors involved in angiogenesis versus lymphangiogenesis. Curr Opin Cell Biol, 2009; 21:154–165.

29.

Nelson

, Padera

, Garkavtsev

et al. Differential gene expression of primary cultured lymphatic and blood vascular endothelial cells. Neoplasia, 2007; 9:1038–1045.

30.

Irigoyen

, Anso

, Martinez

et al. Hypoxia alters the adhesive properties of lymphatic endothelial cells. A transcriptional and functional study. Biochim Biophys Acta, 2007; 1773:880–890.

31.

Wick

, Saharinen

et al. Transcriptomal comparison of human dermal lymphatic endothelial cells ex-vivo and in vitro. Physiol Genom, 2007; 28:179–192.

32.

Van Lint

, Libert

. Chemokine and cytokine processing by matrix metalloproteinases and its effect on leukocyte migration and inflammation. J Leukoc Biol, 2007; 82,6:1375–1381.

33.

Turner

, Mason

, Brown

et al. Molecular cloning of the Goodpasture antigen demonstrates it to be the alpha 3 chain of type IV collagen. J Clin Invest, 1992; 89:592–601.

34.

Baggiolini

, Dewald

, Moser

. Interleukin-8 and related chemotactic cytokines CXC and CC chemokines. Adv Immunol, 1993; 55:97–179.

35.

Keping

. Interleukin-8 and human cancer biology. Cytokine Growth Factor Rev, 2001; 12:375–389.

36.

Goodwin

, Lupton

, Schmeirer

et al. Human interleukin-7: Molecular cloning and growth factor activity on human and murine B-cell lineage. Proc Natl Acad Sci USA, 1989; 86:302–306.

37.

Van Lint

, Libert

. Chemokine and cytokine processing by matrix metalloproteinases and its effect on leukocyte migration and inflammation. J Leukoc Biol, 2007; 82,6:1375–1381.

38.

Ley

, Kansas

. Selectins in T-cell recruitment to non-lymphoid tissues and sites of inflammation. Nature Rev Immunol, 2004; 4:325–336.

39.

Yoshihiko

, Tsuruga

. The expression of E-selectin and chemokines in the cultured human lymphatic endothelium with lipopolysaccharides. J Anat, 2008; 212:654–663.

40.

Vestweber

, Blanks

. Mechanisms that regulate the function of the selectins and their ligands. Physiol Rev, 1999; 79:181–213.

41.

Zhang

, Sanders

, Ye

, Kynaston

, Jiang

. Expression of vascular endothelial growth inhibitor (VEGI) in human urothelial cancer of the bladder and its effects on the adhesion and migration of bladder cancer cells in vitro. Anticancer Res, 2010; 30:87–95.

42.

Zhai

, Ni

, Jiang

et al. VEGI, a novel cytokine of the tumour necrosis factor family, is an angiogenesis inhibitor that suppresses the growth of colon carcinomas in vivo. FASEB J, 1999; 13:181–189.

43.

Chew

, Pan

, Yu

et al. A novel secreted splice variant of vascular endothelial cell growth inhibitor. FASEB J, 2002; 16:742–744.

44.

Yue

, Ni

, Romanic

et al. TL1, a novel tumour necrosis factor-like cytokine, induces apoptosis in endothelial cells. Involvement of activation of stress protein kinases (stress-activated protein kinase and p38 mitogen-activated protein kinase) and caspase-3-like protease. J Biol Chem, 1999; 274:1479–1486.

45.

Chen

, Chan

, Bai

, Fung

, Tessier

, To

, Merchant

, Lai

. ZBP-89 reduces cell death threshold in hepatocellular carcinoma cell by increasing caspase-6 and S-phase cell cycle arrest. Cancer Lett, 2009; 28,283:52–58.

46.

Comijn

, Berx

, Vermassen

et al. The two-handed E-box binding zinc finger protein SIP1 down regulates E-cadherin and induces invasion. Mol Cell, 2001; 7:1267–1278.