Dow

nloa

ded

By:

[Sta

cker

, Ste

ven

A.]

At:

01:0

5 31

Mar

ch 2

008

A system for quantifying the patterning of the lymphatic vasculature

RAMIN SHAYAN1,2,3, TARA KARNEZIS1, EVELYN TSANTIKOS1,2,

STEVEN P. WILLIAMS1, ANDREW S. RUNTING1, MARK W. ASHTON3,

MARC G. ACHEN1, MARGARET L. HIBBS1, & STEVEN A. STACKER1

1Melbourne Tumor Biology Branch, Ludwig Institute for Cancer Research, Royal Melbourne Hospital, Victoria, Australia,2Department of Surgery, Royal Melbourne Hospital, University of Melbourne, Victoria, Australia, and 3Jack Brockhoff

Reconstructive Plastic Surgery Research Unit, RoyalMelbourne Hospital and Department of Anatomy and Cell Biology, Faculty

of Medicine, Dentistry and Health Sciences, University of Melbourne, Victoria, Australia

(Received 20 December 2007; revised 14 January 2008; accepted 22 January 2008)

AbstractThe lymphatic vasculature is critical for immunity and interstitial fluid homeostasis, playing important roles in diseases such aslymphedema and metastatic cancer. Animal models have been generated to explore the role of lymphatics andlymphangiogenic growth factors in such diseases, and to study lymphatic development. However, analysis of lymphatic vesselshas primary been restricted to counting lymphatics in two-dimensional tissue slices, due to a lack of more sophisticatedmethodologies. In order to accurately examine lymphatic dysfunction in these models, and analyse the effects oflymphangiogenic growth factors on the lymphatic vasculature, it is essential to quantify the morphology and patterning of thedistinct lymphatic vessels types in three-dimensional tissues. Here, we describe a method for performing such analyses,integrating user-operated image-analysis software with an approach that considers important morphological, anatomical andpatterning features of the distinct lymphatic vessel subtypes. This efficient, reproducible technique is validated by analysinghealthy and pathological tissues.

Keywords: Lymphatics, lymphangiogenesis, VEGF-D, lymphedema, quantification

Introduction

The lymphatic vasculature consists of distinct types of

lymphatic vessels with absorptive or transport func-

tions (Skobe and Detmar, 2000; Baldwin et al. 2002).

The blind-ended “initial” or “capillary” lymphatics

lack a distinct vessel wall and absorb lymph fluid and

cellular infiltrate from the peripheral tissues, before

draining into more deeply located pre-collector

lymphatic vessels (Skobe and Detmar 2000; Scavelli

et al. 2004). These pre-collector lymphatics have well-

defined vessel walls and valves that direct lymph flow

to the collecting lymphatics, the subcutaneous,

muscular conduits that transport lymph to lymph

nodes and ultimately back to the venous system

(Scavelli et al. 2004; Muthuchamy et al. 2003).

Cancer cells may also metastasise via the lymphatics

(Stacker et al. 2002), and alterations to normal

lymphatic function can result in lymph fluid accumu-

lation (lymphedema) (Baldwin et al. 2002). In

addition, lymphatic vessels give rise to tumours such

as lymphangioma and Kaposi’s sarcoma (Fukunaga

2005) and have been implicated in asthma, psoriasis,

rheumatoid arthritis, transplant rejection, and other

inflammatory conditions (Baluk et al. 2005; Alitalo

et al. 2005). Animal models have been developed to

explore molecular mechanisms underlying such

diseases (Stacker et al. 2002; Karkkainen et al. 2001;

Wirzenius et al. 2007), and lymphatic markers

ISSN 0897-7194 print/ISSN 1029-2292 online q 2007 Informa UK Ltd.

DOI: 10.1080/08977190801932550

Correspondence: S. A. Stacker, Ludwig Institute for Cancer Research, Post Office Box 2008, Royal Melbourne Hospital, Victoria 3050,Australia. Tel: 61 3 93413155. Fax: 61 3 93413107. E-mail: Steven.Stacker@ludwig.edu.au

Growth Factors, December 2007; 25(6): 417–425

Dow

nloa

ded

By:

[Sta

cker

, Ste

ven

A.]

At:

01:0

5 31

Mar

ch 2

008

(Breiteneder-Geleff et al. 1999; Banerji et al. 1999;

Wigle and Oliver 1999) have been identified for the

detection of lymphatics in animal tissues and human

samples (Karpanen et al. 2001; Stacker et al. 2001;

Van der Auwera et al. 2006). In the past, the analysis

of lymphatic vessels in tissue has been largely

restricted to measuring the abundance of immuno-

histochemically stained lymphatics, in two-dimen-

sional tissue slices (Van der Auwera et al. 2006). Yet

this approach does not take into account the distinct

types of lymphatic vessels located at different tissue

depths, and other anatomically and structurally

important features of the lymphatic vasculature,

such as branching, blind ended sacs (BES) or vessel

loops; and does not easily allow patterning parameters

such as vessel dimensions and inter-vessel spacing to

be assessed (Skobe and Detmar 2000; Scavelli et al.

2004; Makinen et al. 2005). As lymphatic morphology

and patterning are intimately related to function, these

parameters could each influence the capacity for fluid

or cellular absorption, and the magnitude or direction

of lymph flow within the lymphatic vasculature (Skobe

and Detmar 2000; Scavelli et al. 2004; Muthuchamy

et al. 2003). It is therefore important to quantify the

different features of the distinct types of lymphatic

vessels in situ, within normal and pathological tissues,

to pinpoint their role in these conditions. Further, the

recent major focus on embryonic lymphatic develop-

ment (Karkkainen et al. 2000; Karkkainen et al. 2004)

requires a methodology enabling screening and

detailed characterisation of functionally important

developmental alterations that result from engineered

genetic mutations. This will facilitate exploration of

the roles of protein growth factors, cell surface

receptors and other signaling molecules in lymphatic

development and biology (Karkkainen et al. 2000;

Karkkainen et al. 2004). Here, we describe such an

approach, combined with an accessible user-friendly

computer interface that is useful for analyzing

lymphatic vessels during development and in disease,

and for monitoring the efficacy of experimental

therapeutics that target the lymphatic vasculature

(Skobe and Detmar 2000; Baldwin et al. 2002; Adams

and Alitalo 2007).

Results

A computer-aided system for quantifying lymphatic vessel

patterning

In order to perform detailed in situ three-dimensional

visualisation of intact lymphatics in relation to other

vessels, we analysed whole-mount-stained (unsec-

tioned) tissue specimens, using lymphatic-specific

markers (see Methods). Due to their different

anatomical locations within the tissue, adjustable

upright microscopes with the capacity to visualise

different specific focal depths within tissues, were used

to image the fluorescently labelled lymphatics

(Figure 1a–c). To accurately quantify lymphatic

vessel patterning, we developed the Lymphatic Vessel

Analysis Protocol (LVAP), a plug-in designed for

quantifying lymphatic vessels using ImageJ (Abramoff

et al. 2004), software commonly used for high-

throughput image analysis, on multiple computer

platforms (Supplemental Figure 1a and b). Due to the

complex architecture of whole-mount-stained lym-

phatic vessels, the design of LVAP incorporates a

strong user-operated component in preference to a

fully automated system.

LVAP takes into account important morphologic

and patterning characteristics (number of BES,

lymphatic branches and loops, vessel diameter,

density, and inter-lymphatic vessel distance (ILVD)),

to study lymphatic vessels. We have selected several

animal models for validation of LVAP, in which the

lymphatic vasculature is altered due to physiological or

pathological stimuli. In order to demonstrate the

utility of LVAP in distinguishing distinct features of

different types of lymphatic vessels, the program was

used to characterise the normal adult and embryonic

skin lymphatics in the mouse. Our analysis demon-

strated increased branching and loop formation in the

embryonic lymphatics, compared with the regular

lymphatic capillaries in the adult skin (Figure 1a–d).

To determine if LVAP may also be used to study

abnormal lymphatic vessels as a result of develop-

mental defects and/or inflammation, we investigated a

naturally occurring mutant, the “motheaten-viable”

(Me v/Me v) mouse (Shultz et al. 1984), in which a

genetic alteration in a signaling molecule involved in

immune function (SHP-1) (Tsui et al. 1993), results

in systemic inflammation and in inflamed, “mothea-

ten” ears (Supplemental Figure 2a and b). Quantifica-

tion analysis of fluorescently labelled lymphatics in the

motheaten ears (Figure 1e) revealed abnormalities in

several morphological and patterning parameters

(Figure 1f). While the average density of lymphatics

and numbers of BES were not statistically different

compared with wild-type mice, the motheaten mouse

ears did have statistically significant increases in

lymphatic vessel width, numbers of lymphatic vessel

branching points and loop structures, compared with

wild-type mice (Figure 1f).

Use of LVAP in pathological models

The remodeling of blood vessels and lymphatics is a

key component of several human pathologies (Adams

and Alitalo, 2007). Two important examples are

tumourigenesis and wound healing. The protein

growth factors vascular endothelial growth factor

(VEGF)-C (Joukov et al. 1996) and VEGF-D (Achen

et al. 1998) drive the proliferation of lymphatic vessels

(lymphangiogenesis), and have been used to create

animal models in which changes to lymphatic vessels

R. Shayan et al.418

Dow

nloa

ded

By:

[Sta

cker

, Ste

ven

A.]

At:

01:0

5 31

Mar

ch 2

008

can be studied (Karpanen et al. 2001; Stacker et al.

2001; Karkkainen et al. 2004; Mandriota et al. 2001;

Baldwin et al. 2005). We previously used a mouse

flank tumour model that secreted VEGF-D (Stacker

et al. 2001), to show that local lymphangiogenesis at

the primary tumour was associated with spread of

tumour cells to lymph nodes. We have shown that

mice harboring these tumours have high circulating

levels of VEGF-D (data not shown), however, it is not

known whether this may have an effect on the systemic

lymphatic network. Here, using the ear lymphatics

(Figure 2a and b) to represent the systemic lymphatic

network in the same flank tumour model, we have

examined whether the LVAP can detect subtle

systemic changes to the lymphatic vasculature that

may occur in the presence of high systemic levels

Figure 1. Analysis of lymphatics in the skin of adult mice and embryos. (a) Immunofluorescent whole-mount staining of adult mouse ear

capillary lymphatics in a wild-type animal using anti-LYVE-1 antibody (green). (b) Fluorescent labelling of pre-collector (open arrow) and

collecting lymphatics (filled arrow) using podoplanin labelling (green) in the skin of an adult wild-type mouse ear. (c) Embryonic (E18.5) skin

lymphatics fluorescently labelled using anti-LYVE-1 antibody (green). (d) Quantification of branching points, lymphatic loops, blind endings

sacs (BES), inter-lymphatic vessel distance (ILVD), lymphatic vessel width and lymphatic vessel density, as determined using LVAP, in

lymphatic capillary, pre-collector and embryonic vessel networks in wild-type mice. (e) Fluorescent labelling of abnormal ear skin lymphatics

in the “motheaten viable” (Me v/Me v) mutant mouse using anti-LYVE-1 antibody (green). (f) Quantification of lymphatic vessel

characteristics by LVAP as measured in (a) wild-type and (e) motheaten viable (Me v/Me v) mice. Bar, 500 mm; asterisks, lymphatic loop;

circle, blind ending sac. Asterisks above bar graphs indicate statistical significance (*p value ,0.05; **p value ,0.01). Error bars represent

SEM.

A system for quantifying the patterning of the lymphatic vasculature 419

Dow

nloa

ded

By:

[Sta

cker

, Ste

ven

A.]

At:

01:0

5 31

Mar

ch 2

008

R. Shayan et al.420

Dow

nloa

ded

By:

[Sta

cker

, Ste

ven

A.]

At:

01:0

5 31

Mar

ch 2

008

of VEGF-D. We also analysed a similar tumour model

which secreted VEGF-C. Overall, we found minimal

effects as a result of the high circulating levels of

VEGF-D (Figure 2a and c), despite significant local

lymphangiogenesis in the primary tumours (Stacker

et al. 2001). In contrast, the enhanced circulating

levels of VEGF-C resulted in alterations to ear

lymphatics, in particular by inducing a small but

significant increase in the lymphatic vessel width

(Figure 2b and c), and consistent with this, the ILVD

was diminished reciprocally in these mice (Figure 2c).

This experiment is consistent with the reported effects

of VEGF-C (Wirzenius et al. 2007), and demonstrates

the ability of the LVAP to detect subtle variations in

the lymphatic vasculature.

In order to evaluate the utility of LVAP to examine

models of ‘local’ effects of lymphangiogenic stimuli,

the lymphatic vessels in two pathological models were

studied. The same VEGF-D-secreting tumour cells as

described above were grown in the mouse ear, and the

resulting lymphatic vessels were compared with those

generated in a simple ear-wound model (Saaristo et al.

2006) (Figure 2d–h). To investigate lymphangiogenic

activity in a three-dimensional tissue specimen using

the LVAP, the plug-in parameters were extended to

address lymphatic vessel sprouting and loops, which

are significant features in the formation of the neo-

lymphatics (see Methods). The average number of

lymphatic sprouts in the tumours (Figure 2d, e, and h)

was significantly higher than at the edges of the wounds

(Figure 2f–h), whereas no lymphatic sprouts were

observed in healthy ear tissue (Figure 1a). In addition,

the number of ‘sprout-tips’ per sprouting neo-

lymphatic vessel (see Figure 2g for example of sprout-

tip) was also significantly higher in tumours than in

wounds (Figure 2d–h). Interestingly, however, the

length of sprouts was greater in the wound-generated

lymphatics (Figure 2 h). While the tumours had

significantly higher numbers of lymphatic loops, the

lymphatics in the wound healing model grew in a more

orderly fashion (Figure 2e, g, and h). Therefore, the

LVAP system highlighted differences between lym-

phatic sprouting in two different pathological models,

which may have biological and functional significance.

These represent differences that could be overlooked

using standard methods of quantification and two-

dimensional tissue analysis.

Evaluation of organ-specific lymphatic vessels

As lymphatics are integral to disease states in many

organs (Skobe and Detmar, 2000; Alitalo et al. 2005),

we compared lymphatic morphology and patterning in

whole-mounted tissue specimens from normal healthy

organs that are relevant to some important diseases in

which lymphatics are implicated. We analysed the

gastrointestinal and respiratory tracts, and the skin,

which are prone to develop metastatic cancers and

inflammatory disorders, involving the lymphatic

system (Stacker et al. 2002; Baluk et al. 2005; Alitalo

et al. 2005). We found significant variations in the

nature of the lymphatics in the different organs.

The characteristic pattern of skin lymphatic capillaries

(Figure 3a) is distinct from the more uniform,

truncated form of lymphatic vessel, the lacteal, in

each villus of the small-bowel mucosa (Figure 3b),

whilst there is a dense, highly branched and looped

pre-collecting lymphatic network deeper within the

bowel mucosa (Figure 3c). Dual staining with

the blood vessel marker PECAM-1 also reveals the

intimate relationship between the blood and lym-

phatic vascular networks (Figure 3b). In contrast, the

trachea boasts few lymphatic loops, branching points

or BES and the lymphatics in this tissue are arranged

circumferentially between the cartilaginous tracheal

rings (Figure 3d–f). Both the width and the ILVD of

the lymphatics in the trachea differ significantly from

those in the skin, however the overall lymphatic

density was not statistically different across the three

tissues assessed (Figure 3f).

Discussion

We have developed a computer-aided approach for

quantitative evaluation of the morphology and

patterning of the lymphatic vasculature. When applied

to fluorescently labelled lymphatics in whole tissue

specimens, this method allows the comparison of a

number of parameters, which describe the form, and

relate to the function, of lymphatic vessels. The

methodology was validated on healthy and pathologi-

cal tissues, including models in which lymphatics were

subjected to a spectrum of stimuli. This methodology

allows a more comprehensive and quantitative analysis

of the lymphatic vasculature than previous approaches

Figure 2. Analysis of the lymphatics with LVAP in tumour and wound healing models. Low power image of the lymphatics in the anterior ear

skin of mice bearing (a) VEGF-D or (b) VEGF-C-secreting flank tumours, labelled with anti-LYVE-1 antibody (green). (c) Quantitative

comparison of ear skin lymphatic vessels from mice bearing VEGF-D or VEGF-C-secreting flank tumours, using LVAP. (d) Low and (e) high-

power representative images of whole-mounted, fluorescently labelled lymphatics (anti-LYVE-1) (green) in an ear tumour; and (f) low and (g)

high-power representative images of whole-mounted, fluorescently labelled skin lymphatics in an ear wound. Open arrow indicates a sprout

with a single tip, filled arrows indicate sprout with multiple tips. (h) Comparative quantification of average number of sprouts, tips per sprout,

and lymphatic loops formed, and average sprout length in the tumour and wounding models, using the LVAP. Bar, 500 mm; asterisks,

lymphatic loops; dotted line, wound edge; T, tumour. Asterisks above bar graphs indicate statistical significance (*p value ,0.05); error bars

represent SEM.

R

A system for quantifying the patterning of the lymphatic vasculature 421

Dow

nloa

ded

By:

[Sta

cker

, Ste

ven

A.]

At:

01:0

5 31

Mar

ch 2

008

that are based on counting lymphatic vessels in

two-dimensional tissue slices (Stacker et al. 2002; Van

der Auwera et al. 2006) or are semi-quantitative in

nature. Another advantage of our approach is that

subtle lymphatic abnormalities, which may have

previously been overlooked, can now be identified

and evaluated. Importantly, the LVAP system also

allows the user to store a record of each individual

vessel parameter count performed, aiding validation

and allowing retrospective comparisons of data as

knowledge in the field of lymphatics, including

definitions of vessel structures, continues to evolve.

The LVAP will also be useful for quantifying the

results of in vitro assays of lymphangiogenesis

Figure 3. Comparison of lymphatics in different organs using LVAP. (a) Low power representative image of whole-mount fluorescent

labelled lymphatics in wild-type adult mouse ear skin, using anti-LYVE-1 antibody (green). Whole-mount gastrointestinal (small bowel)

mucosa depicting fluorescent-labelled lymphatics in (b) high-power, side view of lacteal in villus (filled arrows) and pre-collectors (open

arrows); and (c) low-power view of small bowel mucosa stained using anti-LYVE-1 antibody (green) to show lymphatics; and anti-PECAM-1

antibody (red) to label blood vessels. (d) Low and (e) high-power images of whole-mount fluorescent labelling of lymphatics in the respiratory

tract (trachea) mucosa using anti-LYVE-1 antibody (green) (open arrows show cartilaginous tracheal rings). (f) Comparative quantification of

average number of branching points, loops, BES and average lymphatic vessel width, ILVD and lymphatic vessel density in ear skin,

gastrointestinal tract (small bowel from (c)), and trachea. Bar, 500 mm; Asterisks above bar graphs indicates statistical significance (*p value

,0.05; **p value ,0.01); error bars represent SEM.

R. Shayan et al.422

Dow

nloa

ded

By:

[Sta

cker

, Ste

ven

A.]

At:

01:0

5 31

Mar

ch 2

008

or lymphatic tube formation (for example see Supple-

mental Figure 3a and b).

In designing the LVAP, we chose to examine

parameters that reflect the definitive structural

features of lymphatic vessels and those subject to

change during physiological or pathological chal-

lenges. When the lymphatics respond to altered tissue

demands or stimulatory growth factors, the capacity

for fluid absorption and transport, and the dimensions

of the lymphatics and surrounding interstitial space

can change (Skobe and Detmar 2000; Wirzenius et al.

2007; He et al. 2005). In developmental defects,

signals which specify the correct branching, caliber or

spacing of vessels can become deranged (Karkkainen

et al. 2001; Makinen et al. 2005). In pathological

conditions such as wound healing (Saaristo et al.

2006) and tumour formation (Karpanen et al. 2001;

Stacker et al. 2001; Mandriota et al. 2001; Stacker

et al. 2004; Achen et al. 2006), the expression of

lymphangiogenic growth factors can cause abnormal

proliferation, and possibly differentiation, of lym-

phatic vessels resulting in a mixture of vessel changes,

including vessel sprouting and vessel dilatation

(Wirzenius et al. 2007; He et al. 2005). These features

of lymphatics are assessed in the current version of the

LVAP program, which has the capacity to be updated

in the future, as our understanding of, and capacity to

manipulate, lymphatics evolves.

The LVAP program has a number of applications

for examining lymphatic vessels. As a tool for research

it will facilitate the analysis of lymphatic defects (such

Supplemental Figure 1. (a) ‘Screen grab of ‘Lymphatic Vessel Analysis Protocol’ plug-in and Image J tool bar (Abramoff et al. 2004). (b)

‘Screen grab of ‘Lymphatic Vessel Analysis Protocol’ graph plug-in layover on picture, to enable quantification of whole-mounted,

fluorescently labelled lymphatic vessels.

Supplemental Figure 3. (a) Representative image of tube formation

by lymphatic endothelial cells in culture. (b) Quantitative

comparison of lymphatic endothelial cell culture in vitro, pictured

in Supplemental Figure 3a, compared to in vivo lymphatic vessel

formation (as seen in Supplementary Figure 1b), using LVAP.

(2mu*p value ,0.05); error bars represent SEM. Bar, 200 mm.

Supplemental Figure 2. (a) Macroscopic pictures of adult wild-type

mouse ear. (b) Macroscopic picture of swollen motheaten ear of age-

matched “motheaten viable (Me v/Me v)” mutant mouse.

A system for quantifying the patterning of the lymphatic vasculature 423

Dow

nloa

ded

By:

[Sta

cker

, Ste

ven

A.]

At:

01:0

5 31

Mar

ch 2

008

as lymphedema) in mice with experimentally gener-

ated genetic mutations (Karkkainen et al. 2001;

Karkkainen et al. 2000). This could include large

populations of mutant mice generated in functional

genetic screens. Finally, the capacity to evaluate local

and systemic lymphatic responses and complimentary

in vitro assays, makes the LVAP valuable as an efficient

and reproducible platform, with which to comprehen-

sively analyze the therapeutic agents that target

lymphatics, in animal models of disease.

Methods

Mouse mutant, tumour and wounding experiments

293EBNA-1 cells stably transfected to express VEGF-

C or VEGF-D, were xenografted subcutaneously to

the flank or ear of SCID/NOD mice (Stacker et al.

2001). Animals were sacrificed on reaching the

tumour size limit stipulated by the guidelines of the

Ludwig Institute Animal Ethics Committee, and

serum samples were taken, before harvesting both

ears. Embryos used were embryonic day 18.5.

Homozygous Motheaten viable (Me v/Me v

(C57BL/6)) mutant mice (Shultz et al. 1984) and

wild-type, age-matched littermates were obtained

from The Walter and Eliza Hall Institute of Medical

Research. Whole-mounted tissues were imaged after

fluorescent labelling with lymphatic and/or blood-

vessel markers. Ear wounds were generated using an

ear skin punch, as described (Saaristo et al. 2006), and

the wound-edge imaged circumferentially.

Fluorescent lymphatic labelling

Skin from embryos and whole ears from adult mice

were dissected and fixed (6 h) in 4% paraformalde-

hyde, before incubation in blocking solution (1% BSA,

5% goat serum in 0.3% Triton in PBS), followed by

primary antibody (16 h, 48C) (Makinen et al. 2005).

Samples were washed (6 h) then incubated with

fluorescently conjugated secondary antibodies (16 h,

48C), before mounting in Vectashield (Vector Lab-

oratories). Samples were imaged and analysed using a

Nikon Eclipse 90i upright fluorescent microscope and

Nikon DXM1200c digital camera.

Antibodies

Antibodies used: hamster anti-mouse podoplanin

(RDI), 1:1000; rabbit polyclonal anti-mouse LYVE-

1 (Fitzgerald Industries), 1:1000; rat anti-mouse

CD31 (BD Pharmingen), 1:200. Anti-rat IgG Alexa

594, anti-rabbit IgG Alexa 488, and anti-hamster IgG

Alexa 488 fluorescent secondary antibodies (Molecu-

lar Probes) were used at 1:200 dilution.

Quantification of lymphatic vessels

Apical photographs of mouse ears were taken ( £ 4

magnification), opened in the Image J program

(Abramoff et al. 2004), and the image ‘initialised’ in

the ‘Lymphatic Vessel Analysis Protocol’ (LVAP) plug-

in. Images were overlaid with a 200 pixel grid using a

specific plug-in. (Supplemental Figure 1b). The

operator selects a parameter (variable of lymphatic

vessel morphology or patterning e.g. vessel width)

then, aided by the grid, proceeds systematically

through the image. Locating the cursor on the

parameter (or “event”) the mouse button is clicked,

both counting the event and automatically recording a

marker at that point, to avoid re-counting. In the case

of the lymphatic vessel width and ILVD, the operator

follows the horizontal grid lines from left to right across

the screen, and places a marker at the intersection

between this line and a lymphatic. The second click on

the other side of the vessel demarcates the end of the

vessel and this distance is measured. In the case of the

ILVD, the markers are placed at the beginning and end

of the space between vessels. The lymphatic vessel

density was the total number of intersections between a

horizontal grid line and a lymphatic vessel throughout

the image, divided by the number of lines. All these

measurements are made, and the values tallied, before

the data are integrated into an Excel spreadsheet for

statistical analysis. The number of lymphatic vessel

branching points, loops, and BES were calculated

using the appropriate buttons in the LVAP tool, in a

blinded fashion, and averaged over each image.

Averaged field counts for each parameter were collated

for all sections and used to obtain averages for each ear

assessed, before data for all mice were de-identified and

grouped to generate averages for each mouse type, and

their respective statistical significance determined via

the Student’s T-test (Microsoft Excel). Parameters

chosen in wounding experiments were: number of (1)

lymphatic sprouts, (2) tips per sprout, (3) loops and (4)

sprout length (Figure 2f–h).

Software design

The LVAP was designed as a plug-in for ImageJ, a

Java-based image analysis application for high-

throughput processing and image manipulation on

multiple computer platforms (Abramoff et al. 2004).

ImageJ is free to download and use, with numerous

plug-ins that extend the functionality of the program,

which have been contributed by the scientific

community. As there were no plug-ins that were

sufficient for the needs of this project, two existing

plug-ins were modified as required. An on-line tutorial

on the use of LVAP, the modified source code and

plug-ins are available from the Ludwig Institute for

Cancer Research (Melbourne Branch) website: http://

www.ludwig.edu.au/archive.

R. Shayan et al.424

Dow

nloa

ded

By:

[Sta

cker

, Ste

ven

A.]

At:

01:0

5 31

Mar

ch 2

008

Acknowledgements

The authors would like to acknowledge Tony Burgess

for critical reading of this manuscript, and thank Jee-

Wong Chew for his IT expertise, Kari Alitalo for the

generous provision of a VEGF-C cDNA, Janna Taylor

and Pierre Smith for assistance with the generation of

figures, and Stephen Cody for assistance with

microscopy. This work was supported by Program

and Project Grants from the National Health and

Medical Research Council of Australia (NHMRC).

R.S. is supported by the Surgical Scientist

Scholarship Program of the Royal Australasian

College of Surgeons; E.T. by an Australian Post-

graduate Award, M.G.A. and M.L.H. by Senior

NHMRC Research Fellowships and S.A.S. by a

Foundation Fellowships and a Senior NHMRC

Research Fellowship from Pfizer.

References

Abramoff MD, Magelhaes PJ, Ram SJ. 2004. Image processing with

image. J Biophotonics Intl 11:36–42.

Achen MG, et al. 1998. Vascular endothelial growth factor D

(VEGF-D) is a ligand for the tyrosine kinases VEGF receptor 2

(Flk1) and VEGF receptor 3 (Flt4). Proc Natl Acad Sci USA

95:548–553.

Achen MG, Mann GB, Stacker SA. 2006. Targetting lymphangio-

genesis to prevent tumour metastasis. Br J Cancer 94:1355–60.

Adams RH, Alitalo K. 2007. Molecular regulation of angiogenesis

and lymphangiogenesis. Nat Rev Mol Cell Biol 8:464–478.

Alitalo K, Tammela T, Petrova TV. 2005. Lymphangiogenesis in

development and human disease. Nature 438:946–953.

Baldwin ME, Stacker SA, Achen MG. 2002. Molecular control of

lymphangiogenesis. Bioessays 24:1030–1040.

Baldwin ME, et al. 2005. Vascular endothelial growth factor D is

dispensable for development of the lymphatic system. Mol Cell

Biol 25:2441–2449.

Baluk P, et al. 2005. Pathogenesis of persistent lymphatic vessel

hyperplasia in chronic airway inflammation. J Clin Invest 115:

247–257.

Banerji S, et al. 1999. LYVE-1, a new homologue of the CD44

glycoprotein, is a lymph-specific receptor for hyaluronan. J Cell

Biol 144:789–801.

Breiteneder-Geleff S, et al. 1999. Angiosarcomas express mixed

endothelial phenotypes of blood and lymphatic capillaries:

Podoplanin as a specific marker for lymphatic endothelium. Am J

Pathol 154:385–394.

Fukunaga M. 2005. Expression of D2-40 in lymphatic endothelium

of normal tissues and in vascular tumours. Histopathology 46:

396–402.

He Y, et al. 2005. Vascular endothelial cell growth factor receptor 3-

mediated activation of lymphatic endothelium is crucial for

tumor cell entry and spread via lymphatic vessels. Cancer Res

65:4739–4746.

Joukov V, et al. 1996. A novel vascular endothelial growth factor,

VEGF-C, is a ligand for the Flt-4 (VEGFR-3) and KDR

(VEGFR-2) receptor tyrosine kinases. EMBO J 15:290–298.

Karkkainen MJ, et al. 2000. Missense mutations interfere with

VEGFR-3 signalling in primary lymphoedema. Nat Genet 25:

153–159.

Karkkainen MJ, et al. 2001. A model for gene therapy of human

hereditary lymphedema. Proc Natl Acad Sci USA 98:

12677–12682.

Karkkainen MJ, et al. 2004. Vascular endothelial growth factor C is

required for sprouting of the first lymphatic vessels from

embryonic veins. Nat Immunol 5:74–80.

Karpanen T, et al. 2001. Vascular endothelial growth factor C

promotes tumor lymphangiogenesis and intralymphatic tumor

growth. Cancer Res 61:1786–1790.

Makinen T, et al. 2005. PDZ interaction site in ephrinB2 is required

for the remodeling of lymphatic vasculature. Genes Dev

19:397–410.

Mandriota SJ, et al. 2001. Vascular endothelial growth factor-C-

mediated lymphangiogenesis promotes tumour metastasis.

EMBO J 20:672–682.

Muthuchamy M, Gashev A, Boswell N, Dawson N, Zawieja D.

2003. Molecular and functional analyses of the contractile

apparatus in lymphatic muscle. FASEB J 17:920–922.

Saaristo A, et al. 2006. Vascular endothelial growth factor-C

accelerates diabetic wound healing. Am J Pathol 169:

1080–1087.

Scavelli C, et al. 2004. Lymphatics at the crossroads of angiogenesis

and lymphangiogenesis. J Anat 204:433–449.

Shultz LD, Coman DR, Bailey CL, Beamer WG, Sidman CL. 1984.

“Viable motheaten”, a new allele at the motheaten locus. Am J

Pathol 116:179–192.

Skobe M, Detmar M. 2000. Structure, function, and molecular

control of the skin lymphatic system. J Invest Dermatol

Symp Proc 5:14–19.

Stacker SA, et al. 2001. VEGF-D promotes the metastatic spread of

tumor cells via the lymphatics. Nat Med 7:186–191.

Stacker SA, Achen MG, Jussila L, Baldwin ME, Alitalo K. 2002.

Lymphangiogenesis and cancer metastasis. Nat Rev Cancer

2:573–583.

Stacker SA, Williams RA, Achen MG. 2004. Lymphangiogenic

growth factors as markers of tumour metastasis. APMIS

112:539–49.

Tsui HW, Siminovitch KA, de Souza L, Tsui FW. 1993. Motheaten

and viable motheaten mice have mutations in the hematopoietic

cell phosphatase gene. Nat Genet 4:124–129.

Van der Auwera I, et al. 2006. First international consensus on the

methodology of lymphangiogenesis quantification in solid

human tumours. Br J Cancer 95:1611–1625.

Wigle JT, Oliver G. 1999. Prox1 function is required for the

development of the murine lymphatic system. Cell 98:769–778.

Wirzenius M, et al. 2007. Distinct vascular endothelial growth

factor signals for lymphatic vessel enlargement and sprouting.

J Exp Med 204:1431–1440.

Author Contributions R. Shayan and T. Karnezis concept

and method development, experimental work and manu-

script preparation; E. Tsantikos and S.P. Williams exper-

imental work and manuscript preparation; A.S. Runting

software design and testing; M.W. Ashton, M.G. Achen,

M.L. Hibbs, and S.A. Stacker concept and manuscript

preparation.

Competing Interest Statement S.A. Stacker and M.G.

Achen are consultants for Vegenics Ltd.

A system for quantifying the patterning of the lymphatic vasculature 425