Interstitial Flow as a Guide for Lymphangiogenesis
Kendrick C. Boardman, Melody A. Swartz
Abstract—The lymphatic system is important in tissue fluid balance regulation, immune cell trafficking, edema, and cancer
metastasis, yet very little is known about the sequence of events that initiate and coordinate lymphangiogenesis. Here,
we characterize the process of lymphatic regeneration by uniquely correlating interstitial fluid flow and lymphatic
endothelial cell migration with lymphatic function. A new model of skin regeneration using a collagen implant in a
mouse tail has been developed, and it shows that (1) interstitial fluid channels form before lymphatic endothelial cell
organization and (2) lymphatic cell migration, vascular endothelial growth factor-C expression, and lymphatic capillary
network organization are initiated primarily in the direction of lymph flow. These data suggest that interstitial fluid
channeling precedes and may even direct lymphangiogenesis (in contrast to blood angiogenesis, in which fluid flow
proceeds only after the vessel develops); thus, a novel and robust model is introduced for correlating molecular events
with functionality in lymphangiogenesis. (Circ Res. 2003;92:801-808.)
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
Key Words: lymphatic capillary 䡲 interstitial fluid flow 䡲 vascular endothelial growth factor-C
䡲 lymphatic vessel hyaluronan receptor 1 䡲 matrix metalloproteinases
T
he lymphatic system is an important component of the
circulation; however, our understanding of lymphatic
biology lags far behind that of the blood vasculature, even
though the lymphatics are the primary route of metastasis of
many common cancers, including breast, colon, and skin
cancer. Several important molecular regulators have been
implicated in lymphangiogenesis, including vascular endothelial growth factor (VEGF)-C and VEGF-D1– 4 and their
receptor Flt-4,5 as well as Prox16 and basic fibroblast growth
factor7; furthermore, a CD44-like hyaluronan receptor specific for lymphatics (lymphatic vessel hyaluronan receptor 1;
LYVE-1) has been identified.8 –10 However, the morphological, spatial, and temporal aspects of lymphangiogenesis
remain poorly understood. This is due in great part to the lack
of an adequate model in which to study the spatial organization of lymphatic endothelial cells (LECs) relative to fluid
flow as well as the development of functionality and patterning in an emerging lymphatic capillary network.
The extent of similarity between the processes of lymphangiogenesis and blood angiogenesis is poorly understood,
although they share many molecular regulators and serve
complementary functions. The latter, which is much more
extensively studied and characterized, occurs in embryonic
development, wound healing, and tumor growth through a
sprouting process managed by a number of polypeptide
growth factors such as VEGF-A and angiopoietin-2.11 The
primary physiological driving force for blood angiogenesis is
oxygen concentration, which is directly correlated with the
primary function of the blood vasculature. Indeed, a number
of growth factors including VEGF-A and erythropoietin are
expressed under the influence of hypoxia inducing factor-I,
an oxygen-sensitive transcription factor.12 The primary function of the lymphatic system, in contrast to the blood
circulation, is to maintain interstitial fluid balance and provide lymphatic clearance of interstitial fluid and macromolecules, thereby sustaining osmotic and hydrostatic gradients
from blood capillaries through the interstitium and stimulating convection for interstitial protein transport.13 Thus, although it is clear that VEGF-C and VEGF-D are important
molecular regulating factors, it is plausible to suggest that an
important physiological regulating factor of lymphangiogenesis may be the maintenance of interstitial fluid balance and
protein convection (analogous to oxygen transport as a
physiological regulating factor in blood angiogenesis) and,
therefore, that interstitial fluid flow may play a role in
lymphangiogenesis.
Previous research into the physiological regulation of
lymphangiogenesis has been hindered by the lack of appropriate experimental models. In the present study, we introduce a new model of lymphatic development in a regenerating region of skin. The model is unique and robust in that it
allows for the observation of fluid channels and (through the
concurrent use of molecular markers) differential identification of functional lymphatic vessels. We developed this
model in the skin of the mouse tail, which has been used
previously to measure lymph flow velocity,14 to evaluate
tissue fluid balance parameters,15 and to observe lymphatic
architecture in tissue grafts16 and in transgenic mice.3
Original received January 28, 2003; revision received February 25, 2003; accepted February 26, 2003.
From the Biomedical Engineering Department (K.C.B., M.A.S.) and the Chemical Engineering Department (M.A.S.), McCormick School of
Engineering and Applied Sciences, Northwestern University, Evanston, Ill.
Correspondence to Melody A. Swartz, Chemical Engineering Department, McCormick School of Engineering and Applied Sciences, Northwestern
University, 2145 Sheridan, Evanston, IL 60208-3107. E-mail [email protected]
© 2003 American Heart Association, Inc.
Circulation Research is available at http://www.circresaha.org
DOI: 10.1161/01.RES.0000065621.69843.49
801
802
Circulation Research
April 18, 2003
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
Figure 1. Lymphangiogenesis: model construction and features.
a, Lymphatic capillary network located within the dermis of the
normal mouse tail, as seen with FM. Bar⫽1 mm. b, CDE skin
replacement model in mouse tail to investigate morphological
patterns of lymphangiogenic growth. Bar⫽1 cm. c, Close-up
photograph of CDE. d, Schematic diagram of CDE. An annulus
of skin is surgically removed from the tail without disturbing the
tail core (tc, including large blood vessels, tendons, muscle, and
bone). Bar⫽1 mm. The silicone sleeve (s) is placed over the
wound and fixed in place with surgical glue. Collagen is injected
under the sleeve and allowed to gel (g).
To create this model, a collagen “window” is formed in the
tail skin, through which lymph and interstitial fluid necessarily move in a proximal direction. A 2-mm-wide circumferential band of dermal tissue is removed midway up the tail,
leaving the underlying bone, muscle, major blood vessels,
and tendons intact. Blood flow to the tail remains uninterrupted because of the large subcutaneous blood vessels, but
the lymphatic network and the flow of interstitial fluid out of
the tail tip are completely interrupted by the procedure. Left
alone, this interruption of the lymphatic network without a
fluid-bridging mechanism would result in marked edema of
the tail.15,16 Instead, to bridge the distal and proximal portions
of the tail skin and restore the proximally routed unidirectional flow of lymph, a collagen dermal equivalent (CDE) is
inserted into the prepared gap (Figures 1b through 1d). The
CDE provides an effective path for fluid transport, and edema
in the distal portion of the tail is prevented. Furthermore,
because the CDE is initially acellular, any cells or structures
(particularly lymphatic vessels) later observed within the
CDE are presumably the result of cell migration, proliferation, and organization.
This model has enabled us to observe the development of
functional fluid channels and to correlate these functional
features with migrating LECs and blood endothelial cells
(BECs). In the present study, we report the temporal and
spatial relationship between lymphatic development within
CDE and interstitial fluid channeling, showing that (1)
interstitial fluid channeling precedes LEC organization, and
(2) fluid channeling, LEC migration, and functional vessel
formation are initiated only in the direction of lymph flow.
(Note: here and later, the term migration refers to the net
result of cell translocation, which is a function of cell survival
and possibly proliferation as well as migration.) We also
show that matrix protease activity before fluid channeling is
more pronounced at the upstream (distal) rather than downstream (proximal) side of the CDE, further hinting that
interstitial flow may play a role in the mechanism of the
observed directional fluid channel formation and cell migration. Furthermore, we find that expression of VEGF-C is
similarly more pronounced at the distal side of the CDE
before LEC infiltration. These observations are in contrast to
the events that take place during blood angiogenesis, which
occurs equally from all sides of the implanted tissue equivalent and is not preceded by interstitial fluid channeling,
consistent with decades of observations of blood angiogenesis in a variety of models.11,17 Therefore, the process of
lymphangiogenesis seems distinct from that of blood angiogenesis and may involve a new mechanism related to interstitial fluid flow. On the basis of our observations, we suggest
that interstitial flow can drive the formation of fluid channels
along which LECs migrate, proliferate, and finally reorganize
into a functional lymphatic capillary network.
Materials and Methods
Twenty-one Balb/C mice were used in the present study. Animals
were anesthetized with a subcutaneous mixture of ketamine (100
mg/kg) and xylazine (10 mg/kg). Postsurgical analgesic (buprenorphine 2 mg/mL) was administered subcutaneously twice a day for up
to 1 week. All work was conducted with the approval of the
Northwestern University Animal Care and Use Committee.
Surgical Preparation
Midway up the mouse tail, a 2-mm-wide circumferential annulus of
skin was removed from the tail, leaving the underlying bone, muscle,
major blood vessels, and tendons intact (Figure 1). Great care was
taken to ensure that the deeper draining lymphatics running alongside the major blood vessels were also severed. A diameter-matched
silicone sleeve was placed over the wound and fixed in place with
surgical glue to the intact skin on the proximal and distal edges. A
0.3% solution of type I rat tail collagen18 was injected into the wound
under the sleeve and allowed to gel; this formed the CDE. The tail
was then wrapped with self-sticking tape to protect the surgical site
from interference by the animal. The CDE is not meant to mimic the
granulation tissue of a healing wound (in fact, it never scars because
of the silicone sleeve); rather, it is simply meant as a tissue bridge for
interstitial fluid flow and as a scaffold for cell migration and growth.
Microlymphangiography
After removing the silicone sleeve for optimal visualization, we
evaluated the functional lymphatics of the mouse tail by use of
fluorescence microlymphangiography (FM).19 By this technique, a
Boardman and Swartz
Interstitial Flow and Lymphangiogenesis
803
fluorescently labeled macromolecule (in this case, lysine-fixable
FITC-dextran of 2000 kDa; Molecular Probes) was injected intradermally at a constant pressure of 35 mm Hg into the tip of the tail
(which was roughly 25 to 30 mm away from the site of the CDE).
Because of its large size, the tracer was taken up by the lymphatics
but was excluded from the blood vasculature. Because the tracer fills
the network downstream from the injection site, the fluorescent
tracer is clearly visible within lymphatic capillaries and any connecting fluid channels (Figure 1a). The entire procedure was monitored
with a Nikon Eclipse TE 200 inverted microscope; digital images
were captured with a Spot RT slider camera (Diagnostic
Instruments).
Specimen Preparation
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
Perfusion fixation was used to cross-link the lysine-fixable FITCdextran tracer in place, marking the lymphatic vessels and fluid
channels in thin cryosections. Immediately after FM at various time
points (10, 25, and 60 days), the animals were fixed with Zamboni’s
fixative20 (2% formaldehyde, 15% picric acid, and 0.1 mol/L
phosphate buffer, pH 6.9) by perfusion through the abdominal aorta.
After fixation, the animals were then perfused with cyroprotectant
(10% sucrose in PBS, pH 7.2), the tails were removed, and a 6-mm
region centered around the CDE was embedded in freezing media
and flash-frozen in melting isopentane. Longitudinal cryosections
(10 ␮m) were obtained from each specimen.
Immunofluorescence and Immunohistochemistry
We identified various endothelial markers by use of standard
immunofluorescence techniques. Cryosections were incubated with
rat anti-mouse CD31 (PharMingen), rabbit anti-mouse LYVE-1
(kind gift of Dr David Jackson, John Radcliffe Hospital, Headington,
Oxford, UK), goat anti-mouse Flt-4 (R&D Systems), and rabbit
anti-mouse SLC (R&D Systems, kind gift of Dr Gwendalyn
Randolph, Mount Sinai Medical School, New York, NY) antibodies
and labeled with the appropriate rhodamine red-X– conjugated IgG
(Jackson ImmunoResearch Laboratories). The mounting medium
contained DAPI (Vectashield, Vector) for visualizing cell nuclei. We
also identified VEGF-C expression within the CDE using standard
immunohistochemical techniques. Cryosections were incubated with
goat anti-mouse VEGF-C (A-18, Santa Cruz Biotechnology). Sections were further incubated with biotinylated rabbit anti-goat IgG
(Vector) and then visualized with the ABC-AP kit by the use of
Vector Red substrate (Vector). In this case, cell nuclei were
counterstained with hematoxylin.
In Situ Zymography
We mapped the location of matrix protease activity using an in situ
zymographic technique.21 Briefly, DQ Collagen (Molecular Probes)
is heavily loaded with fluorescein such that fluorescence is initially
quenched. Digestion by collagenolytic proteases yields a highly
fluorescent product whose intensity is proportional to proteolytic
activity. We incubated freshly frozen tissue sections overnight at
room temperature with DQ type I collagen in reaction buffer
(50 mmol/L Tris-HCl, 0.15 mol/L NaCl, 5 mmol/L CaCl2, and
0.2 mmol/L sodium azide, pH 7.6). We examined the sections
without washes or fixation.
Results
Overall, 3 different types of markers were used in evaluating
the lymphangiogenic process: fluid channel markers (via
FM), cellular markers (via immunofluorescence), and soluble
molecular markers (via immunohistochemistry and in situ
zymography); furthermore, each of these could be correlated
in space and in time. At each time point, FM was performed
on the live animal before perfusion fixation, cryosectioning,
and immunostaining. Because FM allowed the lysine-fixable
cutaneously injected fluorescent tracer to be tracked from its
interstitial deposit into the local lymphatic capillaries and
Figure 2. FM (green) allows direct visualization of fluid channel
formation and lymphatic network organization within the CDE.
Dashed lines indicate CDE location (n⫽3 per time point). a, At
day 10 after procedure, lymph fluid is diffuse, and channeling is
not observed. b, At day 25, discrete fluid channels are
observed. c, At day 60, a hexagonal fluid channel network, continuous with upstream and downstream lymphatics, is nearly
complete. Lymph and interstitial fluid flows through the CDE
proximally from left (tail tip) to right (tail base). Over time, the
CDE may contract, resulting in a width of ⬍2 mm. Bar⫽1 mm.
then proximally within the lymphatic network and because
the fluid tracer was then fixed in place before sectioning, we
could directly correlate functional fluid channels with LECs
or other molecular markers via immunostaining.
Fluid Channeling Precedes Lymphatic
Network Organization
We used FM to track lymph flow pathways (either as fluid
channels or lymphatic vessels) within the CDE over a period
of 60 days after the procedure. We saw that after an initial
phase of at least 10 days (during which lymph pooled and
moved interstitially through the CDE), fluid channels eventually formed to bridge the distal and proximal portions of the
lymphatic system; finally, those channels gave way to an
organized network of lymphatic capillaries. In normal tail
skin, FM showed a highly regular hexagonal network of
lymphatic vessels (Figure 1a). Ten days after the skin was
removed and replaced with the CDE, the tracer appeared to
move through the construct diffusely and without discrete
channeling, and it pooled uniformly throughout the CDE
(Figure 2a). This was also seen at earlier times, eg, at 3 and
7 days (data not shown). In contrast, by 25 days, the tracer
was seen concentrated within crude but distinct fluid channels
(Figure 2b). Finally, within 60 days, the tracer in the CDE
was present within a continuous network of fluid channels
between the upstream and downstream intact lymphatics
(Figure 2c), restoring lymphatic continuity and lymph flow.
Remarkably, the remodeled fluid channels within the CDE
had recovered the unique hexagonal patterning of a normal
lymphatic network.
Observation of LECs Versus BECs
The ability to identify functional lymphatic vessels and
differentiate LECs from BECs through combined functional
and molecular techniques is a key advantage of our model.
Immediately after FM, the animals were perfusion-fixed, and
the tails were flash-frozen. Later, the tails were cryosectioned
804
Circulation Research
April 18, 2003
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
Figure 3. Immunofluorescence staining of LECs and BECs in longitudinal tail skin cryosections. Lymph flows proximally from left to
right, and dashed lines indicate CDE location (n⫽3 for each time point). Specimens are stained with antibodies against LEC markers
LYVE-1 (panels a through c) and Flt-4 (panel d) or BEC marker CD31 (panels e through h) (red), alone (top) and in combination with the
green fluid tracer (bottom). Cell nuclei are counterstained with DAPI (blue). a and e, LYVE-1 staining in normal or preprocedure specimens is correlated with fluid channels in normal skin (arrows). CD31⫹ cells that do not strongly colocalize with fluid channels are identified as blood vessels (arrowheads), although lymphatic vessels are characterized by a weak CD31 staining (arrows). b and f, LYVE-1 is
not seen at day 10, and the fluid tracer within in the CDE is diffuse, although CD31⫹ cells appear throughout the CDE. c and g, At day
25, fluid channels have formed within the CDE, and LECs (arrows) begin to infiltrate along fluid channels from the upstream side only,
whereas CD31⫹ structures appear more organized and trimmed. d and h, By day 60, the fluid channels have attained a density and
organization similar to that of normal tail skin, and most are colocalized with LECs (arrows). At the proximal edge, an unlined fluid
channel indicates that lymphangiogenesis is not yet complete (arrow with asterisk). The organization of CD31⫹ structures also resembles that of normal skin, but colocalization with fluid channels is minimal. Bar⫽0.5 mm.
and immunostained for various markers to distinguish interstitial fluid channels from endothelium-lined vessels as well
as to differentiate blood from lymphatic vessels. Fixation was
initiated when lymph, saturated with lysine-fixable FITCdextran, was present within the CDE as well as the entire
lymphatic network; thus, the fixation cross-linked the fluid
tracer in place within both those fluid channels and the
lymphatic vessels. This allowed us to distinguish between
interstitial fluid channels (positive for FITC-dextran but
negative for the putative lymphatic markers Flt-4 and LYVE1), functional lymphatic vessels (positive for FITC-dextran
and Flt-4 and/or LYVE-1), nonfunctional lymphatic structures (negative for FITC-dextran but positive for Flt-4 and/or
LYVE-1), and nonlymphatic endothelial structures (negative
for both FITC-dextran and Flt-4 or LYVE-1 but positive for
other endothelial markers such as CD31). Because these
markers are not always specific for LECs and BECs under
certain conditions,22–25 we carefully examined the specificity
of these markers in our model before confirming these
classifications.
In normal skin, 83⫾13% (mean⫾SD) of structures that
were LYVE-1⫹ (Figure 3a) and Flt-4⫹ (not shown) colocalized with tracer-filled fluid channels, whereas only 18⫾9% of
the CD31⫹ structures similarly colocalized with the FITCdextran. Furthermore, the CD31⫹ structures that did colocalize with the fluid channels had a lower staining intensity
relative to the CD31⫹ vessels that did not colocalize with
fluid channels (Figure 3e); this is consistent with the stronger
expression of CD31 by BECs relative to LECs.26 In experimental animals in which normal skin was replaced with the
CDE, CD31⫹ structures were seen distributed equally
throughout the CDE as early as 10 days after the procedure
(Figure 3f), whereas Flt-4⫹ and LYVE-1⫹ structures were
absent from the CDE at day 10 (Figure 3b). At 60 days, both
CD31⫹ and Flt-4⫹/LYVE-1⫹ structures were present within
the CDE (Figures 3d and 3h), but only the Flt-4⫹/LYVE-1⫹
structures strongly colocalized with fluid channels (80⫾15%
of Flt-4⫹/LYVE-1⫹ structures colocalized with fluid tracer,
but only 10⫾2% of CD31⫹ structures colocalized with fluid
tracer). Thus, these additional observations allowed confirmation of the classifications suggested above.
Boardman and Swartz
Interstitial Flow and Lymphangiogenesis
805
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
Figure 4. Endothelial cells and fluid
tracer in upstream and downstream
regions of CDE. CD31, LYVE-1, Flt-4,
and SLC receptors (red) are shown alone
(top) and in merged images with green
fluid tracer (bottom). Cell nuclei were
counterstained with DAPI (blue). Micrographs are of normal mouse tail dermis
and of upstream and downstream
regions of the CDE at 10, 25, and 60
days after procedure (n⫽3 per time
point), and lymph flows proximally from
left to right. Bar⫽50 ␮m.
We observed in the CDE that the development of the
lymphatic system lags behind that of the blood vasculature;
this finding is concordant with findings in earlier studies
involving scar and granulation tissue.27–29 In observing the
CD31 staining patterns, it can be seen that the epithelialization and vascularization of the CDE occurred more quickly
than lymphatic development (Figures 3b and 3f). The entire
CDE was completely vascularized (CD31⫹/FITC-dextran⫺)
by day 10 after the procedure, although organization and
trimming continued through day 60 (Figures 3f through 3h).
By contrast, lymphatic development (LYVE-1⫹/Flt-4⫹ and
FITC-dextran⫹) within the CDE occurred between 25 and 60
days (Figures 3b through 3d). Furthermore, although LEC
migration occurred primarily in the direction of interstitial
fluid flow (at 25 days, 79⫾12% of Flt-4⫹/LYVE-1⫹ cells
were in the upstream half of the CDE), there were no
discernible differences in BEC migration or organization
between the upstream or downstream directions, even at 3
(data not shown) and 10 days, emphasizing the morphological
and physiological differences between blood and lymph
angiogenesis.
LEC Migration Follows Direction of Lymph Flow
Along Fluid Channels
By comparing the relative distribution of fluid channels and
LECs within the CDE in cryosections at various time points,
we observed lymphangiogenesis as a process whereby fluid
channel formation and LEC migration along those channels
preceded the eventual organization of the fluid channels into
endothelium-lined vessels (Figures 3a through 3d) and developed from the upstream edge of the CDE (we assume that the
only differences between the upstream and downstream edges
of the CDE were in the direction of interstitial fluid flow,
inasmuch as reepithelialization and blood vascularization
occurred equally from both ends). At 10 days after the
procedure, the fluid tracer was diffusely dispersed throughout
the CDE interstitium without any LECs present (Figure 3b).
By day 25, we observed discrete fluid channels within the
CDE that bridged the distal and proximal lymphatic networks
of the intact skin, consistent with the microlymphangiography
data in Figure 2. However, we also detected LECs (Flt-4⫹/
LYVE-1⫹ cells) primarily within those fluid channels, but
only as single cells and not as a continuous vessel developing
from the preexisting vessels of the intact skin. Furthermore,
those cells were located mostly on the upstream (distal) edge
(Figure 3c), consistent with the direction of lymph flow
(again, 79% of LYVE-1⫹ structures were located in the
upstream half of the CDE). Finally, by day 60, the fluid tracer
appeared to be present in smaller, more regularly spaced, and
more numerous clumps, and the fluid channels were fully
colocalized with LECs (Figure 3d) similar to those seen in
normal skin (Figure 3a). These observations, together with
806
Circulation Research
April 18, 2003
Figure 5. a and b, Collagen-degrading
MMPs within the CDE, as indicated by
green fluorescence using in situ zymography in specimens at 3 days (panel a)
and 7 days (panel b) after CDE implantation. Dashed lines demarcate edges of
CDE. Flow is from left to right (distal to
proximal). c through f, Enlarged distal
and proximal views of 3- and 7-day tails.
Cell nuclei were counterstained with propidium iodide (red). Bar⫽100 ␮m.
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
those shown in Figure 2, indicate that the fluid channels had
been transformed into a dense network of functional lymphatic vessels with a normal hexagonal architecture.
The observation that LECs translocate along fluid channels
primarily in the direction of lymph flow (from distal to
proximal) is corroborated by comparing the spatial localization of fluid channels with expression patterns of CD31,
LYVE-1, Flt-4, and SLC (another putative lymphatic marker
also known as 6Ckine, exodus-2, and CCL21)30 between the
upstream versus downstream edges of the CDE (Figure 4).
Because lymph flows through the mouse tail skin and the
CDE in a single direction (proximally), fluid channels are
necessarily initiated at the upstream edge of the CDE. At 10
days, CD31⫹ cells were seen equally in both the upstream and
downstream regions (with no apparent relationship to fluid
channels), but we did not observe LYVE-1, Flt-4, or SLC
expression within the CDE. At 25 days, CD31⫹ structures
remained distributed uniformly throughout the CDE, but
LYVE-1, Flt-4, and SLC were seen only in cells migrating
along fluid channels in the upstream region. Finally, at 60
days, upstream and downstream regions showed staining
patterns of CD31, LYVE-1, Flt-4, and SLC similar to those of
normal tissue, ie, colocalizing with several small fluid channels. Thus, the migrating cells within the CDE express these
3 putative LEC markers, and our observations suggest that
they were responsible for creating the organized hexagonal
lymphatic network seen in Figures 2c and 3d.
Patterns of Protease and VEGF-C Expression in
the CDE
Matrix degradation was examined to investigate the relationship between fluid flow and fluid channeling (as seen in
Figures 2b, 3c, and 3g). Because the rate of collagen
hydrolysis is much faster for matrix metalloproteinases
(MMPs) compared with other collagenolytic proteases and
enzymes such as cathepsin K, trypsin, and thermolysin,31 it
was assumed that collagen degradation was due primarily to
MMP activity. Therefore, we mapped the location of
collagen-degrading proteases using in situ zymography21 at
each of our time points and saw the majority of activity only
at early times. After 3 days after implantation, we observed
several points of MMP activity present at both ends of the
CDE (Figures 5a, 5c, and 5d) indicative of individual migrating cells. However, by 7 days, 2 patterns of MMP activity
emerged (Figure 5b): (1) MMPs from adjacent intact dermis
streamed into the CDE in the direction of fluid flow (only at
the distal edge, Figure 5e); this stood in contrast to activity at
the proximal edge of the CDE, where MMP activity did not
infiltrate the CDE against the direction of fluid flow (Figure
5f). (2) A path of MMP activity, independent of that produced
by the intact dermis, was present within and through most of
Figure 6. VEGF-C protein expression
within the CDE using immunohistochemistry. a and d, VEGF-C staining (red)
appeared to be strongest within the CDE
at 10 days, primarily in the distal edge. b
and e, Decreased, but more distributed,
VEGF-C staining was also observed at 25
days after procedure. c and f, VEGF-C
cannot be detected anywhere in the CDE
at 60 days. Flow is from left to right (distal to proximal). Cell nuclei were counterstained with hematoxylin (blue). Bar⫽200
␮m.
Boardman and Swartz
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
the length of the CDE. These observations lead to speculation
that there is at least 1 potential mechanism of unidirectional
interstitial fluid channel formation: proteases, transported by
the interstitial fluid flow, preferentially degrade matrix in a
downstream direction, leading to directed cell migration and
fluid channel formation in the direction of flow.
Finally, we correlated the spatial and temporal expression
of VEGF-C with the lymphangiogenic events occurring
within the CDE by immunohistochemistry. At 10 days after
the procedure, a large amount of VEGF-C was seen at the
distal but not the proximal edge of the CDE (Figures 6a and
6d). The pattern of VEGF-C distribution suggested, in a
similar manner to the location of MMPs, that VEGF-C
transport and expression were likely directed by interstitial
fluid flow. Because VEGF-C is known to be a mitogen for
LECs, this could provide a significant chemotactic factor for
LEC proliferation and migration, preferentially from the
distal, rather than proximal, edge. By 25 days, VEGF-C was
only weakly observed and appeared to be distributed throughout the CDE (Figures 6b and 6e), and by 60 days, VEGF-C
could not be detected within the CDE (Figures 6c and 6f).
Discussion
The data in the present study reveal the spatial and temporal
patterns of interstitial fluid flow and cell migration involved
in lymphangiogenesis in an implanted skin equivalent. The
significance of these data are 2-fold. First, our results introduce a potential new physiological regulator in lymphangiogenesis by implicating interstitial fluid flow in the process
and signify a potential new mechanism of vessel formation.
In this mechanism, lymphatic vessel formation is initiated
along preestablished routes of fluid flow. This is in contrast to
the known mechanisms of blood vessel growth and remodeling (eg, angiogenesis, vasculogenesis, arteriogenesis, and
intussusception), through which fluid flow occurs only after
an intact vascular network has developed. In our proposed
mechanism, fluid channels form within the extracellular
milieu, and then the LECs migrate along these channels to
organize into a functional lymphatic network, contiguous
with the upstream and downstream lymphatic networks.
Indeed, this is consistent with the function of the lymphatic
capillaries of maintaining interstitial fluid balance and
convection.
Second, our results present a new and robust model for
studying lymphangiogenesis. Because the implanted CDE is
circumferential, there is no possibility for fluid shunting
around the implant, and lymph (fluid contained in the
upstream lymphatics) must necessarily pass through the
implant to continue moving in downstream lymphatics.
Therefore, the fluid pathways through the CDE can be
positively tracked with FM originating from the upstream
lymphatics and can later be correlated with LECs by immunohistological examination. This is the first model to our
knowledge in which function can be definitively examined,
because there is no alternative drainage pathway as in other
models; indeed, the question of whether identified LEC-lined
structures are functional lymphatics has been an ongoing
controversy in recent literature.
Interstitial Flow and Lymphangiogenesis
807
The question of the mechanism of fluid channel formation
naturally arises from our observations. Although this issue is
clearly interesting and important and warrants further study,
one potential mechanism can be suggested from the pattern of
MMP activity within the CDE. Skewed by interstitial convection, the MMPs preferentially degrade matrix in a downstream direction, leading to directional fluid flow and cell
migration. Furthermore, the temporal and spatial patterns of
VEGF-C distribution suggest that growth factor distribution
and thus LEC mitogenesis may be influenced by interstitial
fluid flow. This would potentially lead to the preferential
proliferation and migration of LECs from the distal, rather
than the proximal, edge. This may help to at least partially
explain the translocation of LECs along fluid channels (as
evidenced in both FM and immunofluorescent staining)
before the eventual organization of a complete and functional
lymphatic capillary network.
The phenomenological evidence obtained in the present
study for the role of interstitial fluid channeling in functional
lymphatic vessel formation is consistent with and complementary to the rapidly growing literature on molecular
regulation of lymphangiogenesis, which is particularly focused on VEGF-C and its receptor Flt-4. For example, it has
been observed that upregulating the lymphatic growth factor
VEGF-C in the dermis of transgenic mice induces lymphatic
hyperplasia but does not change lymphatic density or architecture3; this is consistent with our findings because the blood
vasculature, extracellular matrix, and thus, presumably, interstitial fluid balance were normal in these transgenic mice.
When Flt-4 was made inactive through mutation32 or was
blocked,33 a decrease in lymphangiogenesis was observed,
along with marked edema (a state of increased fluid accumulation and decreased interstitial flow). Thus, VEGF-C appears
to be required for LEC proliferation, a necessary step in
lymphangiogenesis, but additional factors are necessary to
alter lymphatic architecture or to induce new vessel growth.
In summary, we introduce a new model of lymphangiogenesis in which physiological function can be incorporated
with molecular regulation, and we propose physiological
mechanisms for lymphangiogenesis that complement molecular regulation. We show that fluid channeling and LEC
migration precedes lymphatic capillary formation and that
unlike blood angiogenesis, these events occur primarily in the
direction of flow into the implanted dermal equivalent. In
exploring one possible mechanism for these events, we
observed that in the upstream region of interstitial flow, MMP
activity is increased, which could lead to preferential cell
migration in the direction of flow. We also showed that
increased expression of the LEC mitogen VEGF-C occurs
primarily in upstream regions of the CDE, which (with
subsequent unidirectional transport) could then enhance cell
migration and proliferation even further in the direction of
flow. Thus, interstitial flow may represent an important
transport mechanism to help guide growth and organization
of a developing lymphatic capillary network.
Acknowledgments
This study was supported by grants from the American Cancer
Society (Illinois Division), the US Army Medical Research and
808
Circulation Research
April 18, 2003
Materiel Command, and the National Science Foundation. We are
deeply grateful to Dr David Jackson for kindly giving us the LYVE-1
antibody and Dr Gwendalyn Randolph for the SLC antibody. We
also thank Seth O’Day, Amanda Knudson, Jeffrey Blatnik, and Eva
Mika for technical assistance and Drs Mihaela Skobe and Matthew
Glucksberg for helpful comments.
References
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
1. Joukov V, Pajusola K, Kaipainen A, Chilov D, Lahtinen I, Kukk E,
Saksela O, Kalkkinen N, Alitalo K. A novel vascular endothelial growth
factor, VEGF-C, is a ligand for the Flt4 (VEGFR-3) and KDR
(VEGFR-2) receptor tyrosine kinases. EMBO J. 1996;15:1751.
2. Lee J, Gray A, Yuan J, Luoh SM, Avraham H, Wood WI. Vascular
endothelial growth factor-related protein: a ligand and specific activator
of the tyrosine kinase receptor Flt-4. Proc Natl Acad Sci U S A. 1996;93:
1988 –1992.
3. Jeltsch M, Kaipainen A, Joukov V, Meng X, Lakso M, Rauvala H, Swartz
M, Fukumara D, Jain R, Alitalo K. Hyperplasia of lymphatic vessels in
VEGF-C transgenic mice. Science. 1997;276:1423–1425.
4. Achen MG, Jeltsch M, Kukk E, Makinen T, Vitali A, Wilks AF, Alitalo
K, Stacker SA. 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 U S A. 1998;95:548 –553.
5. Kaipainen A, Korhonen J, Mustonen T, Fang GH, Dumont D, Breitman
M, Alitalo K. Expression of the fms-like tyrosine kinase 4 gene becomes
restricted to lymphatic endothelium during development. Proc Natl Acad
Sci U S A. 1995;92:3566 –3570.
6. Wigle JT, Oliver G. Prox1 function is required for the development of the
murine lymphatic system. Cell. 1999;98:769 –778.
7. Tan Y. Basic fibroblast growth factor-mediated lymphangiogenesis of
lymphatic endothelial cells isolated from dog thoracic ducts: effects of
heparin. Jpn J Physiol. 1998;48:133–141.
8. Banerji S, Ni J, Wang S, Clasper S, Su J, Tammi R, Jones M, Jackson
DG. LYVE-1, a new homologue of the CD44 glycoprotein, is a lymphspecific receptor for hyaluronan. J Cell Biol. 1999;144:789 – 801.
9. Prevo R, Banerji S, Ferguson DJP, Clasper S, Jackson DG. Mouse
LYVE-1 is an endocytic receptor for hyaluronan in lymphatic endothelium. J Biol Chem. 2001;276:19420 –19430.
10. Jackson DG. The lymphatics revisited: new perspectives from the hyaluronan receptor LYVE-1. Trends Cardiovasc Med. 2003;13:1–7.
11. Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature.
2000;407:249 –257.
12. Bunn HF, Poyton RO. Oxygen sensing and molecular adaptation to
hypoxia. Physiol Rev. 1996;76:839 – 885.
13. Swartz MA. The physiology of the lymphatic system. Adv Drug Deliv
Rev. 2001;50:3–20.
14. Swartz MA, Berk DA, Jain RK. Transport in lymphatic capillaries, I:
macroscopic measurement using residence time distribution analysis.
Am J Physiol. 1996;270:H324 –H329.
15. Swartz MA, Kaipainen A, Netti PA, Brekken C, Boucher Y, Grodzinsky
AJ, Jain RK. Mechanics of interstitial-lymphatic fluid transport: theoretical foundation and experimental validation. J Biomech. 1999;32:
1297–1301.
16. Slavin SA, Van den Abbeele AD, Losken A, Swartz MA, Jain RK. Return
of lymphatic function after flap transfer for acute lymphedema. Ann Surg.
1999;229:421– 427.
17. Vailhe B, Vittet D, Feige JJ. In vitro models of vasculogenesis and
angiogenesis. Lab Invest. 2001;81:439 – 452.
18. Bell E, Ivarsson B, Merrill C. Production of a tissue-like structure by
contraction of collagen lattices by human fibroblasts of different proliferative potential in vitro. Proc Natl Acad Sci U S A. 1979;76:1274 –1278.
19. Bollinger A, Jäger K, Sgier F, Seglias J. Fluorescence microlymphangiography. Circulation. 1981;64:1195–1200.
20. Stefanini M, DeMartino C, Zamboni L. Fixation of ejaculated spermatozoa for electron microscopy. Nature. 1967;216:173–174.
21. Oh LYS, Larsen PH, Krekoski CA, Edwards DR, Donovan F, Werb Z,
Yong VW. Matrix metalloproteinase-9/gelatinase B is required for
process outgrowth by oligodendrocytes. J Neurosci. 1999;19:8464 – 8475.
22. Valtola R, Salven P, Heikkila P, Taipale J, Joensuu H, Rehn M, Pihlajaniemi T, Weich H, deWaal R, Alitalo K. VEGFR-3 and its ligand
VEGF-C are associated with angiogenesis in breast cancer. Am J Pathol.
1999;154:1381–1390.
23. Paavonen K, Puolakkainen P, Jussila L, Jahkola T, Alitalo K. Vascular
endothelial growth factor receptor-3 in lymphangiogenesis in wound
healing. Am J Pathol. 2000;156:1499 –1504.
24. Partanen TA, Arola J, Saaristo A, Jussila L, Ora A, Miettinen M, Stacker
SA, Achen MG, Alitalo K. VEGF-C and VEGF-D expression in neuroendocrine cells and their receptor, VEGFR-3, in fenestrated blood
vessels in human tissues. FASEB J. 2000;14:2087–2096.
25. Mouta Carreira C, Nasser SM, di Tomaso E, Padera TP, Boucher Y,
Tomarev SI, Jain RK. LYVE-1 is not restricted to the lymph vessels:
expression in normal liver blood sinusoids and down-regulation in human
liver cancer and cirrhosis. Cancer Res. 2001;61:8079 – 8084.
26. Erhard H, Rietveld FJR, Brocker EB, de Waal RMW, Ruiter DJ. Phenotype of normal cutaneous microvasculature. J Invest Dermatol. 1996;106:
135–140.
27. Reichert FL. The regeneration of the lymphatics. Arch Surg. 1926;13:
871– 881.
28. Clark ER, Clark EL. Observation on the new growth of lymphatic vessels
as seen in transparent chambers introduced into the rabbit’s ear. Am J
Anat. 1932;51:43– 87.
29. Bellman S, Oden B. Regeneration of surgically divided lymph vessels: an
experimental study on the rabbit’s ear. Acta Chir Scand. 1958;116:
99 –117.
30. Gunn MD, Tangemann K, Tam C, Cyster JG, Rosen SD, Williams LT. A
chemokine expressed in lymphoid high endothelial venules promotes the
adhesion and chemotaxis of naive T lymphocytes. Proc Natl Acad Sci
U S A. 1998;95:258 –263.
31. Lauer-Fields JL, Juska D, Fields GB. Matrix metalloproteinases and
collagen catabolism. Biopolymers. 2002;66:19 –32.
32. Karkkainen MJ, Saaristo A, Jussila L, Karila KA, Lawrence EC, Pajusola
K, Bueler H, Eichmann A, Kauppinen R, Kettunen MI, et al. A model for
gene therapy of human hereditary lymphedema. Proc Natl Acad Sci
U S A. 2001;98:12677–12682.
33. Makinen T, Jussila L, Veikkola T, Karpanen T, Kettunen MI, Pulkkanen KJ,
Kauppinen R, Jackson DG, Kubo H, Nishikawa S, Yla-Herttuala S, Alitalo
K. Inhibition of lymphangiogenesis with resulting lymphedema in transgenic
mice expressing soluble VEGF receptor-3. Nat Med. 2001;7:199–205.
Interstitial Flow as a Guide for Lymphangiogenesis
Kendrick C. Boardman and Melody A. Swartz
Downloaded from http://circres.ahajournals.org/ by guest on June 16, 2017
Circ Res. 2003;92:801-808; originally published online March 6, 2003;
doi: 10.1161/01.RES.0000065621.69843.49
Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2003 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7330. Online ISSN: 1524-4571
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circres.ahajournals.org/content/92/7/801
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further information
about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation Research is online at:
http://circres.ahajournals.org//subscriptions/