Title:
Loss of the endothelial glycocalyx is associated with increased E-selectin mediated adhesion of
lung tumour cells to the brain microvascular endothelium
Authors:
Srijana Rai1, Zaynab Nejadhamzeeigilani1, Nicholas J. Gutowski1, 2, and Jacqueline L. Whatmore1.
1
Institute of Biomedical and Clinical Science, University of Exeter Medical School, St. Luke’s campus, Exeter,
United Kingdom, EX1 2LU. s.rai@exeter.ac.uk, zaynabarya@gmail.com
2
Royal Devon and Exeter NHS Foundation Trust, Barrack road, Exeter, United Kingdom, EX2 5DW.
N.J.Gutowski@exeter.ac.uk
Corresponding author:
Dr Jacqueline L. Whatmore
University of Exeter Medical School, St Luke's Campus, Exeter, EX1 2LU
Tel: (+44) 01392 262944, Fax: (+44) 01392 722926. Email: j.l.whatmore@exeter.ac.uk
1
Abstract
Background
Arrest of metastasising lung cancer cells to the brain microvasculature maybe mediated by interactions between
ligands on circulating tumour cells and endothelial E-selectin adhesion molecules; a process likely to be
regulated by the endothelial glycocalyx. Using human cerebral microvascular endothelial cells and non-small
cell lung cancer (NSCLC) cell lines, we describe how factors secreted by NSCLC cells i.e. cystatin C, cathepsin
L, insulin-like growth factor-binding protein 7 (IGFBP7), vascular endothelial growth factor (VEGF) and
tumour necrosis factor-alpha (TNF-α), damage the glycocalyx and enhance initial contacts between lung tumour
and cerebral endothelial cells.
Methods
Endothelial cells were treated with tumour secreted-proteins or lung tumour conditioned medium (CM). Surface
levels of E-selectin were quantified by ELISA. Adhesion of A549 and SK-MES-1 cells was examined under
flow conditions (1 dyne/cm2). Alterations in the endothelial glycocalyx were quantified by binding of
fluorescein isothiocyanate-linked wheat germ agglutinin (WGA-FITC).
Results
A549 and SK-MES-1 CM and secreted-proteins significantly enhanced endothelial surface E-selectin levels
after 30 minutes and 4 hours and tumour cell adhesion after 30 minutes, 4 and 24 hours. Both coincided with
significant glycocalyx degradation; A549 and SK-MES-1 CM removing 55 ± 12% and 58 ± 18.7% of WGAFITC binding, respectively. Inhibition of E-selectin binding by monoclonal anti-E-selectin antibody completely
attenuated tumour cell adhesion.
Conclusion
These data suggest that metastasising lung cancer cells facilitate their own adhesion to the brain endothelium by
secreting factors that damage the endothelial glycocalyx, resulting in exposure of the previously shielded
adhesion molecules and engagement of the E-selectin-mediated adhesion axis.
Keywords:
adhesion, endothelial cells, E-selectin, glycocalyx, lung cancer
2
Introduction
Brain metastases arising from primary lung cancer contribute significantly to the morbidity and mortality
of the disease. These tumours present a challenging clinical scenario since they are generally resistant to
conventional chemotherapy due to the inability of these drugs to cross the blood-brain barrier (BBB) [1].
Insights into the molecular mechanisms of the metastatic process will aid development of more effective
therapeutic strategies.
In order to form metastases, tumour cells must successfully complete a sequence of key, inter-related steps
initiated by extensive proliferation of the primary tumour cells and their invasion of the surrounding
extracellular matrix (ECM). These malignant cells then dissociate from the primary site, intravasate the
circulation and form tumour microemboli, which eventually arrest along the microvasculature of a target organ.
Tumour cells attach to the endothelium, extravasate and ultimately colonize to form secondary metastatic
lesions. Although, the exact mechanisms of tumour cell adhesion in the vasculature of specific organs remain
unknown, parallels have been drawn with the leukocyte-endothelial cell (EC) interactions that occur during
inflammation. Here, initial leukocyte ‘tethering’ and ‘rolling’ is primarily mediated by endothelial surface
proteins from the selectin family (E- and P- selectin) whereas stronger bonds resulting in leukocyte arrest, firm
adhesion and finally transmigration are determined by members of the endothelial immunoglobulin superfamily
[intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1)]. These same
adhesion molecules may be involved in metastatic tumour cell adhesion [2, 3]. Several studies [4, 5] have shown
that altered E-selectin expression on activated endothelial cells greatly influences metastasis formation and
subsequent colonization, and increased circulating E-selectin levels have been associated with metastasis of both
breast and colon carcinoma cells to the liver [6, 7]. Data on the role of endothelial adhesion molecules in lung
tumour metastasis to the brain are very limited; however Soto et al. did recently report a role for VCAM-1 and
activated leukocyte cell adhesion molecule (ALCAM) in the early stages of brain metastasis in a mouse model
[8].
It is now becoming recognised that a major physiological regulator of adhesion of circulating cells to the
endothelium is the endothelial surface layer (glycocalyx); a dense intraluminal layer of transmembrane and
membrane-bound molecules [9]. It is physically linked to the endothelium through key ‘backbone’
proteoglycans and glycoproteins, which together form a mesh into which both endothelium- and plasma-derived
soluble molecules are incorporated [10]. The glycocalyx is a dynamic structure maintained by a tightly
controlled equilibrium between degradation and synthesis of the principal components.
3
If lung tumour cells utilise the cerebral endothelial adhesion molecules to initially adhere to the brain
microvasculature during metastasis, then the glycocalyx may play a critical role in this process by sterically
influencing the ligand-receptor binding dynamics. In health, the EC adhesion molecules are embedded within
the glycocalyx layer since glycocalyx thickness has been reported up to 1µm; considerably larger than the 10nm
span of EC adhesion molecules and receptors [11]. Thus, endothelial adhesion molecules are exposed by loss of
the glycocalyx. This raises the intriguing possibility that circulating lung tumour microemboli arrested in the
cerebral microcirculation have a metastatic potential that enables them to secrete factors that locally degrade the
glycocalyx; enhancing endothelial adhesion molecule exposure and facilitating their own adhesion and
ultimately transmigration. Thus the objective of this study was to identify key pro-metastatic factors secreted by
lung tumour cells and to investigate the effect of these factors on cerebral EC adhesion molecule exposure,
glycocalyx integrity and binding of lung tumour cells.
Here, we report for the first time that human non-small cell lung cancer (NSCLC) cells (A549;
adenocarcinoma and SK-MES-1; squamous cell carcinoma) secrete factors that enhance human cerebral
endothelial adhesion molecule exposure and increase adhesion of tumour cells to the cerebral endothelium.
Importantly, we identify E-selectin as the key adhesion molecule involved in the initial tethering of the lung
tumour cells and for the first time demonstrate that loss of integrity of the cerebral endothelial glycocalyx
enhances exposure of E-selectin and tumour cell adhesion. Our results highlight a potentially important
mechanism by which malignant lung tumour cells can enhance their own metastatic potential; raising the
possibility that interventions designed to maintain glycocalyx integrity may have important clinical benefit in
reducing brain metastases in primary lung tumour patients.
Materials and Methods
Cell culture and preparation of tumour conditioned-medium
A transformed human cerebral microvascular endothelial cell line, hCMEC/D3, was kindly provided by
Dr. Pierre-Olivier Couraud, Institut Cochin, INSERM U1016, France. This cell line has been fully characterized
by Dr Couraud and shown to express a range of endothelial and BBB-specific markers [12]. Cells were cultured
and grown to confluence in rat tail collagen type I coated tissue culture flasks at 37oC under 5% CO2 atmosphere
using Endothelial Cell Basal MV2 Medium (C-22221, PromoCell, Germany) supplemented with MV2
supplement pack (C-39221, PromoCell, Germany). For experiments, hCMEC/D3 cells were between passage 25
and 35.
4
Human NSCLC cell lines, A549 and SK-MES-1, were purchased from the European Collection of Cell
Cultures (ECACC; Salisbury, UK) and cultured in Dulbecco’s Modified Eagle Medium (DMEM, Lonza,
Slough, UK) supplemented with L-glutamine (2mM), gentamicin (50µg/ml) and 10% (v/v) FBS (Sigma
Aldrich, Gillingham, UK).
For collection of lung tumour conditioned-medium (CM), A549 and SK-MES-1 cells at 80% to 90%
confluence were washed with PBS (x3) and incubated with chemically defined DMEM-BS medium [13] (1ml of
medium per 7.5cm2 of growth area) for 24 hours at 37oC. Media were then centrifuged (600g, 10 minutes, 4oC)
and stored at -80oC for further experiments.
Mass Spectrometry Analysis and ELISA
ELISA and LC-MALDI TOF/TOF mass spectrometry (MS) were carried out to screen the secretome (as
CM) of A549 and SK-MES-1 cells (see Additional file 1 for experimental details). Analyses were also carried
out on DMEM-BS as a control. DMEM-BS is a serum-free chemically defined medium specially formulated by
Bottenstein & Sato from DMEM [13]. Since the formulation is known, any additional factors secreted into CM
could be easily identified. Significant results were determined by selecting proteins that matched with two or
more peptides with total ion CIs of >95%, followed by a literature search to determine proteins with a possible
role in metastasis. The following tumour secreted metastasis-related proteins were identified from the available
literature: tumour necrosis factor-alpha (TNF-α) [14], cathepsin L (CL) [15], cystatin C (CC) [16], vascular
endothelial growth factor (VEGF) [17] and insulin-like growth factor binding protein-7 (IGFBP-7) [18].
Secreted levels of TNF-α, CL, CC, VEGF-A and IGFBP-7 in the A549 and SK-MES-1 CM were measured
by ELISA (TNF-α and CC: Bender MedSystems, Vienna, Austria; CL: R&D systems, Abingdon, UK; VEGFA: PeproTech®, London, UK) and IGFBP-7: in house ELISA previously described in [18].
Treatment conditions
During the various investigations, with the exception of E-selectin ELISA, hCMEC/D3 cells were
incubated with either complete lung tumour CM (A549 CM and SK-MES-1 CM) or individual tumour secretedproteins (at concentrations detected during the previous ELISAs) – 80ng/ml CC (Abcam, Cambridge, UK),
10ng/ml CL (Sigma Aldrich, Gillingham, UK), 200ng/ml IGFBP-7 (Abcam, Cambridge, UK), 0.2ng/ml VEGFA (R&D Systems, Abingdon, UK) and 160pg/ml TNF-α (Enzo Life sciences, Exeter, UK). All of the
aforementioned solutions were prepared in DMEM-BS, and control cells received DMEM-BS only.
Cell surface based ELISA (E- and P-selectin)
5
Surface levels of E- and P-selectin on human cerebral endothelial cells (ECs) were assessed using a semiquantitative cell based ELISA as previously described in [19]. Cells were cultured in collagen I coated 96-well
plates; once a monolayer had formed they were serum starved overnight and treated with increasing
concentrations of identified tumour secreted-proteins (in DMEM-BS) for 4 hours; CC – 8, 80 and 800ng/ml, CL
– 1, 10 and 100ng/ml, IGFBP-7 – 100, 300 and 900ng/ml, VEGF-A – 0.2, 10 and 20ng/ml and TNF-α – 100,
500 and 2500pg/ml and/or lung tumour CM. Exposed selectin molecule was recognised by an anti-E- or Pselectin antibody and all subsequent ELISA incubation and washing steps are described in Additional file 1.
Laminar flow adhesion assay
The adhesion of lung cancer cells to the brain endothelium was assessed using an in vitro flow model,
where treated hCMEC/D3 monolayers in gas-permeable flow chambers were perfused with a suspension of lung
tumour cells. ECs were cultured and treated as required in collagen I coated flow chambers (ibidi µ-slides I and
VI0.4, Thistle Scientific, Glasgow UK). After 4 and 24 hours, A549 or SK-MES-1 cells (1 x 106 cells/mL)
stained with 0.5µM calcein acetoxymethyl ester (Calcein AM; Life Technologies Ltd, Paisley UK) were flowed
over the monolayers at a flow rate calculated to provide a shear stress of 1dyne/cm 2, i.e. low shear stress
comparable to that found in post-capillary venules where the majority of adhesion occurs, for 10 minutes at
37oC [20]. To remove any non-adherent tumour cells, the slides were rinsed with PBS for 2.5 minutes at 37oC
before being fixed with 4% paraformaldehyde. The flow slides were then examined under a Nikon Eclipse
TS100 fluorescence microscope and the number of adherent A549 or SK-MES-1 cells in 10 random fields of
view were counted, concentrating on central areas rather than borders.
For the E-selectin inhibition experiments, endothelial monolayers were treated for 24h with CM and
secreted-proteins and then exposed to monoclonal anti-E-selectin antibody (5µg/ml, clone 1.2B6, Sigma
Aldrich, Gillingham, UK) for 30 minutes prior to perfusion with calcein AM stained A549 and SK-MES-1 cell
suspensions (as described above). The optimum conditions for anti-E-selectin treatment were determined
previously through time- and concentration-dependent inhibitory experiments.
Flow cytometry analysis of adhesion molecule counter ligands/receptors
Expression of adhesion molecule counter ligands and receptors on A549 and SK-MES-1 lung tumour cells
were investigated by flow cytometry (FC); specifically the E-selectin ligand sialyl Lewis X (sLeX), ICAM-1
receptors, lymphocyte function-associated antigen 1 (LFA-1α/CD11a) and macrophage-1 (Mac-1α/CD11b) and
VCAM-1 receptor, very late antigen-4 (VLA-4α/CD49d). See Additional file 1 for experimental details.
6
Cell based fluorescence assay
A cell-based fluorescence (CBF) assay, as previously described by Singh and colleagues [21], was
performed to quantify changes within the glycocalyx using fluorescein isothiocyanate-labelled wheat germ
agglutinin (WGA-FITC), which binds to N-acetyl neuraminic acid and N-acetyl glucosamine residues of
proteoglycans and glycoproteins present in the glycocalyx.
Lectin-binding (as WGA-FITC binding) was also observed (as previously described [22]), before and after
1 hour treatment with lung tumour CM, with a confocal laser scanning microscope (Leica DMi8 TCS SP8
Confocal, Leica Microsystems, Germany). The CLSM images (1024 x 1024 pixels) were acquired with a 10x
tube lens and processed using the Leica LAS X Core software.
Glycocalyx constituents ELISA – Hyaluronic acid and Syndecan-1
Hyaluronic acid (HA) and syndecan-1 (CD138) levels in CM-treated endothelial supernates were measured
using the R&D Systems (Abingdon, UK) Quantikine Hyaluronan Immunoassay kit (DHYAL0) and the
Syndecan-1 Human ELISA kit (ab46506) by Abcam (Cambridge, UK). See Additional file 1 for experimental
details.
Statistical analysis
Data are presented as mean ± standard deviation (SD) and statistical differences between control and
treatment groups were assessed using Mann-Whitney non-parametric test. P-value ≤ 0.05 is considered
statistically significant.
Results
Lung cancer cells secrete a range of factors with a potential role in mediating brain metastasis
Initially, MS was used to screen the secretome (as CM) of A549 and SK-MES-1 cells for the presence of
metastasis-related proteins. As expected both cell types secreted a large range of proteins not present in the
DMEM-BS control (see tables A1, A2 and A3 in Additional file 2). However, only the following three proteins
were identified to have a possible role in metastasis and investigated further: CC, CL, and IGFBP-7.
Surprisingly, no growth factors or cytokines were detected using this approach, presumably due to limitations of
MS in identifying proteins in complex mixtures, such as CM, and instrument limitation. Thus, ELISAs were
carried out to quantify the secreted levels of the proteins identified above and to investigate whether the tumour
cells secreted other known pro-angiogenic factors including VEGF-A and TNF-α. Figure 1 shows the
7
concentrations of VEGF-A, TNF-α, CC, CL, and IGFBP-7 in CM from both cancer cell lines. The
concentrations of measured proteins did not vary greatly between A549 and SK-MES-1 CM.
Lung tumour secreted-proteins induce increased exposure of selectins on the surface of human cerebral
ECs
Metastasis critically requires adhesion of the circulating tumour cells (CTCs) to the endothelium of
secondary organs, possibly utilising endothelial selectins. Thus, the influence of the lung tumour secreted
proteins (at a concentration range based on protein levels measured in CM) on cerebral EC E- and P-selectin
surface levels was investigated. Enhanced E-selectin surface expression typically requires de novo synthesis
usually triggered by an inflammatory stimulus and peaks approximately 4-6 hours after activation. After 4
hours, all tumour secreted-proteins induced a significant increase in surface E-selectin at either higher
concentrations tested (CC, IGFBP-7) or over the whole concentration range (CL, VEGF-A, TNF-α) (Fig. 2).
Based on the obtained data and the protein concentrations measured in CM, subsequent studies used only one
concentration of each protein: CC 80ng/ml; CL 10ng/ml; IGFBP-7 200ng/ml; VEGF-A 0.2ng/ml and TNF-α
160pg/ml. At these concentrations the proteins significantly increased detected endothelial E-selectin levels to
226±73.1%, 285±129.2%, 210±70.9%, 229±45.1% and 189±59.1% compared to control (100%) (p≤0.01).
Cerebral ECs exposed to both A549 and SK-MES-1 CM also demonstrated enhanced levels of surface Eselectin at 4 hours (272±96.1% and 177±70.4% respectively). Interestingly, a significant increase was detected
even after 30 minutes treatment (Fig. 2a) (132±4.6% and 125±7.2%) (p≤0.05). This was also evident with the
tumour-secreted proteins at their chosen concentrations (CC: 146±42.3%, CL: 153±33.4%, IGFBP-7:
143±28.5%, VEGF-A: 146±41.0% and TNF-α: 122±11.2%) (Fig. 2b).
Furthermore, at 30 minutes A549 CM also appeared to significantly enhance the exposure of P-selectin on
the surface of hCMEC/D3 cells (128±30.6% v control 100%) (p≤0.037,n=3). There was no increase in Pselectin levels detected with SK-MES-1 CM.
Lung tumour secreted-proteins enhance adhesion of lung tumour cells to human cerebral microvascular
endothelial cells
The observation that A549 and SK-MES-1 lung tumour cells secrete factors that enhance surface levels of
E-selectin led us to investigate whether this was reflected in altered adhesion of lung tumour cells (over 10
minutes) to similarly treated endothelial monolayers. Using a flow adhesion assay to mimic microvascular flow
conditions, it was observed that adhesion of both lung tumour cell types to cerebral EC monolayers was
8
significantly increased by previous exposure to relevant whole CM. With 24 hours treatment, adhesion was
enhanced to 358±20.5% (A549) and 476±45.1% (SK-MES-1 CM) compared to control levels (100%) (p≤0.05,
Fig. 3a & c) Interestingly, a reduced, but still significant, increase in tumour cell adhesion was also observed
after only 30 minutes of A549 and SK-MES-1 CM treatment; 137±17.3% and 170±85.4%, respectively.
Treatment of cerebral ECs with all of the identified tumour secreted-proteins significantly increased adhesion of
both lung tumour cell types to the monolayer. 24 hour treatment with CC, CL, IGFBP-7, VEGF-A and TNF-α
enhanced A549 cell attachment to 568±40.5%, 535±22%, 424±36.1%, 386±18.6% and 356±18.8%,
respectively. Again, a reduced but still significant effect was observed at the shorter treatment time studied.
Overall, SK-MES-1 cell adhesion was more pronounced in response to the secreted-proteins; at 24 hours this
rise was approximately 8-fold with IGFBP-7 and VEGF-A (788±302.3% and 817±167.6%). CC, CL and TNF-α
also increased adhesion to 462±4.6%, 501±10.7% and 530±15.6%, respectively.
Inhibiting E-selectin binding fully attenuated lung cancer cell adhesion to cerebral ECs
These data suggest that lung cancer secreted-factors influence tumour cell adhesion to cerebral endothelial
monolayers by inducing enhanced surface levels of E-selectin. To verify the role of E-selectin, blocking studies
using a monoclonal antibody against E-selectin were performed. Treatment of the endothelial monolayer with
anti-E-selectin antibody fully attenuated the increased adhesion of A549 and SK-MES-1 lung cancer cells (Fig.
4) induced by the tumour secreted-proteins. Isotype-matched control (mouse monoclonal IgG1) studies
confirmed the specificity of the inhibition to the anti-e-selectin antibody (see Additional Figure S1).
Since these data strongly indicate a role for E-selectin in the adhesion of lung tumour cells to the cerebral
endothelium, the expression of counter adhesion ligands and receptors on the tumour cells was investigated i.e.
the E-selectin ligands sialyl Lewis X (sLeX) and P-selectin glycoprotein ligand-1 (PSGL-1) , ICAM-1 receptors,
lymphocyte function-associated antigen 1 (LFA-1α/CD11a) and macrophage-1 (Mac-1α/CD11b) and VCAM-1
receptor, very late antigen-4 (VLA-4α/CD49d). Both A549 and SK-MES-1 cells displayed relatively high levels
of sialyl Lewis X compared with the other antigens investigated (Fig. 4c). Indeed, levels of VLA-4α, LFA-1α,
PSGL-1 and Mac-1α were all very low (<5%) on both cell types. Isotype-matched (IgM, κ) and positive controls
corroborated these findings. The expression of VLA-4α, LFA-1α and Mac-1α was clearly evident (>30%) on
U937 cells and freshly isolated human monocytes, included as a positive control (see Additional Figure S2).
The cerebral endothelial glycocalyx is significantly altered by lung tumour CM and secreted-proteins
9
E-selectin is transcriptionally regulated and increased surface expression is reported to take a minimum of
2 hours. Thus, the observation that lung tumour secreted-factors significantly enhance both exposure of cerebral
endothelium surface E-selectin and tumour /endothelial cell interactions within 30 minutes suggests that an
additional, alternative mechanism contributes to regulation of tumour/endothelial interactions. The endothelial
glycocalyx may influence cell adhesion interactions through steric effects on adhesion molecule/ligand binding
dynamics. We therefore hypothesised that the lung tumour secretome could alter the integrity of the cerebral
endothelial glycocalyx, leading to the increased exposure of previously ‘hidden’ E-selectin molecules and
facilitating enhanced lung tumour cell adhesion.
A cell-based fluorescence assay was used to quantify changes within the glycocalyx. Significant reduction
in glycocalyx staining of cerebral ECs occurred rapidly i.e. within 30 minutes of lung tumour CM treatment
(A549 CM: 84±1.7% and SK-MES-1 CM: 76±7.1%, Figure 5a, d), suggesting loss of glycocalyx integrity. An
even greater reduction in glycocalyx staining was observed after 24 hours to 45±10.9% and 42±17.1% after
exposure to A549 and SK-MES-1 CM, respectively vs control (100%). Interestingly, the effect of CM was more
pronounced than that of the F. heparinum heparinase III enzyme (positive control, 71±15.9%), which
exclusively cleaves heparan sulphate glycosaminoglycans within the glycocalyx (Fig. 5a). Treatment of the
endothelial monolayer with CC, CL, IGFBP-7, VEGF-A and TNF-α also significantly altered the cerebral
endothelial glycocalyx at both time points, with the greatest reduction in staining observed with 24h VEGF-A
and TNF-α treatment, respectively (52±26.5% and 57±28.3% vs control 100%, Fig. 5b).
Disruption or shedding of the glycocalyx can also be assessed by measuring the release/loss of key
components such as syndecan, heparan sulphate and hyaluronan. Syndecan-1 is a proteoglycan core protein
while hyaluronan or hyaluronic acid (HA) is one of the glycosaminoglycans (GAG) and both contribute to the
stability of the glycocalyx structure [23]. HA or syndecan-1 release into the supernatant during lung tumour CM
exposure was used as an additional measure of glycocalyx degradation; determined respectively by hyaluronan
and syndecan-1 ELISA. After 30 minutes, significantly increased levels of HA were detected in the SK-MES-1
CM treated endothelial supernate (56.94 ± 23.6ng/ml vs control, 0.74 ± 0.04ng/ml) and raised syndecan-1 levels
(1.15 ± 1.09ng/ml vs control [not detected]) were observed in the A549 CM exposed endothelial supernates
(Fig. 5c).
Discussion
Although, the exact mechanisms involved in the arrest and extravasation of CTCs during metastasis are
still not fully understood, adhesion to, and transmigration through, the vascular endothelium is critically
10
required. In the brain, the endothelium is an integral component of the blood-brain barrier (BBB), which due to
its highly impermeable nature, presents a formidable challenge to the passage of metastasising CTCs.
An increasing number of reports highlight a role for endothelial adhesion molecules in the tethering of
CTCs to the endothelium. This process may be additionally regulated by the steric effects of the endothelial
glycocalyx which, in health, shields the adhesion molecules from circulating cells. Glycocalyx loss in response
to metastasising tumour cells could favour attachment of tumour cells to the vessel wall, facilitating metastatic
progression [24]. We hypothesised that both adenocarcinoma and squamous human lung tumour cells
metastasising to the brain actually secrete factors that facilitate their own adhesion to the cerebral endothelium,
thus enhancing their metastatic potential. Our data indicate, for the first time, that both adenocarcinoma and
squamous human lung tumour cells secrete proteins that not only disturb the integrity of the human cerebral
endothelial glycocalyx, but concomitantly increase endothelial surface exposure of E-selectin and the binding of
flowing tumour cells to the cerebral endothelium. Additionally, human lung adenocarcinoma tumour cells
secrete proteins that enhanced P-selectin exposure. These data suggest a critical role for the glycocalyx and
cerebral endothelial adhesion molecules in the initial stages of extravasation during lung tumour metastasis to
the brain.
Initial proteomic analysis of the secretome of two NSCLC subtypes, A549 (adenocarcinoma) and SKMES-1 (squamous cell carcinoma), detected several secreted proteins with a potential role in tumour
development or metastasis i.e. CC, CL, IGFBP-7, VEGF-A and TNF-α. It is interesting to note that the
measured secreted levels of VEGF were relatively low compared to some of the other proteins detected.
However, comparable, picogram levels of VEGF-A have been reported in the plasma of cancer patients [25]
and to have significant effects on endothelial function [26]. The pathogenic role of VEGF in tumour growth is
well documented (reviewed in [17]) and not only has enhanced VEGF expression been observed in all forms of
NSCLC, but also high serum levels have been correlated with poor prognosis in patients [27, 28]. TNF-α,
however, has been reported to have both pro- and anti-tumour effects [29]. For instance, in NSCLC a positive
correlation has been identified between high TNF-α expression and favourable prognosis [30]. In contrast, not
only have pro-tumourigenic effects of TNF-α been reported in various cancers e.g. breast [14], but interestingly,
anti-TNF-α treatment of inflammatory diseases has led to a significant rise in the rates of malignancies in
patients [31]. CC is a potent inhibitor of cysteine proteases. It also has contradictory roles in cancer
pathogenesis; while expression of CC has been shown to be down-regulated in various aggressive metastatic
tumours including prostate [32], increased plasma levels have been reported in patients suffering from a range of
11
malignancies including lung cancer [16]. CL is an endopeptidase that degrades a range of proteins including
collagen I and IV and elastin. It has been reported to promote cancer invasiveness by catalysing breakdown of
the interstitial matrix and basement membrane [15]. IGFBP-7 is a member of the IGFBP superfamily [33] and
binds to insulin-like growth factors (IGFs) with high affinity; regulating their activity by preventing
ligand/receptor interactions. We and others have demonstrated a pro-angiogenic role for IGFBP-7. Exogenous
IGFBP-7 induced proliferation, migration and tubule formation in omental microvascular endothelial cells [18]
and induced tube-like structure formation in brain ECs [34]. Additionally, IGFBP-7 is strongly up-regulated in
tumour microvasculature of a range of tumours including lung squamous cell carcinoma [35] and head and neck
squamous cell carcinoma [36]. All five of these identified secreted-proteins and complete tumour CM
significantly enhanced exposure of E-selectin on human cerebral microvascular endothelial monolayers and
importantly, these changes were reflected in increased binding of the relevant lung tumour cells to the
endothelial monolayer during 10 minute perfusion.
The role of E-selectin in the lung tumour/endothelial adhesion was confirmed by blocking experiments
using an anti-E-selectin antibody which completely attenuated the enhanced tumour/endothelial cell binding
induced by the lung tumour secreted-factors. Additionally, flow cytometry analysis confirmed the expression of
the E-selectin ligand sLeX on the surface of both lung tumour cell types. The observation that sLeX expression
was higher in the A549 cell population than in SK-MES-1 cells is consistent with other studies that have
demonstrated greater expression of sLeX antigens in lung adenocarcinomas than in squamous cell carcinomas
[37]. The enzymes responsible for synthesizing sLeX such as α (1, 3) - fucosyltransferases are up-regulated in
lung carcinomas, which could contribute to the adhesive and metastatic potential of lung cancer cells [38].
The observed role for E-selectin in lung tumour metastasis to the brain is in agreement with studies
implicating E-selectin in metastasis in other cancer models. For instance, recent studies in a mouse model have
demonstrated endothelial and tumour cell expression of E-selectin in brain metastases induced by metastatic
murine mammary carcinoma cells [8]. The rapid (i.e. within 30 minutes) enhancement of lung tumour adhesion
and surface exposure of E-selectin in both lung tumour cell types suggested that the tumour cells triggered
additional adhesive mechanisms other than simply stimulating a transcriptionally regulated increase in surface
expression of E-selectin. This raised the possibility that the interaction between the lung cancer cells and the
endothelium is additionally regulated by the endothelial glycocalyx. Since, in health the length of glycocalyx
components often exceeds that of the endothelial adhesion molecules, we investigated whether factors secreted
from lung tumour cells could induce the rapid loss of cerebral endothelial glycocalyx; exposing the previously
12
shielded adhesion molecules and facilitating adhesion. We show for the first time that factors secreted from both
lung cancer subtypes rapidly and significantly induced degradation of the cerebral glycocalyx layer as assessed
by WGA-FITC binding and shedding of glycocalyx components. Our previous data has confirmed that this
lectin almost exclusively stains the luminal glycocalyx in cultured ECs [22]. The mechanism of glycocalyx
modification or degradation in response to all of the tumour secreted-proteins can only be speculated at this
point. These data are the first to show that CL, CC and IGFBP-7 induce glycocalyx changes and for the latter
two proteins there is no available literature describing a role in glycocalyx degradation. CL may mediate
glycocalyx damage indirectly by activating other glycocalyx degrading enzymes, e.g. heparanase, since
formation of the active form of this enzyme from its precursor is exclusively catalysed by CL. However, since
CL induced such rapid glycocalyx changes in cerebral endothelial cells an as yet unidentified, direct role on
glycocalyx components cannot be ruled out.
TNF-α and VEGF have both been reported to alter the glycocalyx in other endothelial cell types. Henry
and Duling suggested that TNF-α could directly stimulate the endothelium to either shed glycocalyx
components, release free radicals, or activate surface-bound proteases or phospholipases that degrade it, such as
endothelial-derived metalloproteases [39]. VEGF is able to directly regulate synthesis and shedding of
glycocalyx components e.g. in glomerular endothelial cells [40] and has also been reported to induce changes in
vessel permeability by causing structural glycocalyx alterations [41].
It is interesting to note that while significant and functionally relevant glycocalyx changes were observed
within 30 minutes, the effects of longer treatment of cerebral endothelial cells with lung tumour secreted-factors
induced even greater levels of E-selectin exposure and tumour cell adhesion. It is thus possible that the effects of
lung tumour secreted factors is two-fold – not only inducing rapid glycocalyx perturbations which facilitate
functional responses i.e. increased adhesion, over the shorter term, but inducing enhanced longer term responses
resulting from even greater glycocalyx loss augmented by induction of increased cellular E-selectin protein
expression and translocation to the cell surface.
Conclusions
In summary, our data indicate a critical role for the cerebral EC glycocalyx and E-selectin/sLeX axis in
regulating the adhesion of circulating lung tumour cells to the brain microvasculature; a key process in
metastatic colonisation. We propose a cascade initiated by the secretion of pro-metastatic proteins from the lung
tumour cells, which alter the cerebral endothelial glycocalyx structure revealing the E-selectin molecules on the
endothelial cell surface. The E-selectin is then available to bind to sLeX ligand on the tumour cells facilitating
13
the stronger bonds required for tethering and capture, ultimately enabling extravasation and secondary growth in
the brain. Our data suggest that combining inhibitors capable of interrupting the contacts between E-selectin and
its ligand along with preservation of glycocalyx integrity may represent a novel treatment modality for reducing
brain metastasis in primary lung tumour patients.
Additional files
Additional file 1: Additional file 1.docx, Supplementary methods (detailed description of MS Analysis, Eselectin ELISA, flow cytometry and legends for Additional figures S1 and S2).
Additional file 2: Additional file 2.docx, Supplementary tables A1, A2 and A3.
Additional figure S1: Additional figure S1.png, Supplementary cell adhesion figure depicting specificity of antiE-selectin antibody.
Additional figure S2: Additional figure S2.png, Supplementary flow cytometry scatter plots for isotype-matched
(IgM, κ) and positive controls.
Abbreviations
EC, endothelial cell(s); CM, Conditioned-medium; IGFBP-7, Insulin-like growth factor-binding protein; ICAM1, Intercellular adhesion molecule-1; NSCLC, Non-small cell lung cancer; sLeX, Sialyl lewis X; TNF-α,
Tumour necrosis factor-alpha; VCAM-1, Vascular cell adhesion molecule-1; VEGF, Vascular endothelial
growth factor; WGA-FITC, Wheat germ agglutinin conjugated to fluorescein isothiocyanate.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
SR performed experiments, collected results, did the data analysis and drafted the manuscript, ZN analysed
collected MS and ELISA data, N.J.G and J.L.W supervised all the work and revised the manuscript. All the
authors have read and approved the manuscript.
Acknowledgements
We thank Drs Boleslaw Winiarski and Lesley Maskell for their help with ELISAs, Dr Paul Eggleton for his
advice on mass spectrometry and Dr Pia Leete for assistance with confocal microscopy. This work was funded
by FORCE Cancer Charity, Exeter, Grant number (RR103273-101). The funding source had no involvement in
the study design, in the writing of the report or in the decision to submit the article for publication.
14
References
1. Nathoo N, Chahlavi A, Barnett GH, Toms SA. Pathobiology of brain metastases. J Clin Pathol.
2005;58(3):237-42.
2. Witz I. The selectin–selectin ligand axis in tumor progression. Cancer and Metastasis Reviews.
2008;27(1):19-30. doi:10.1007/s10555-007-9101-z.
3. Desgrosellier JS, Cheresh DA. Integrins in cancer: biological implications and therapeutic opportunities. Nat
Rev Cancer. 2010;10(1):9-22.
4. Kohler S, Ullrich S, Richter U, Schumacher U. E-/P-selectins and colon carcinoma metastasis: first in vivo
evidence for their crucial role in a clinically relevant model of spontaneous metastasis formation in the lung. Br
J Cancer. 2010;102(3):602-9.
5. Kobayashi K, Matsumoto S, Morishima T, Kawabe T, Okamoto T. Cimetidine inhibits cancer cell adhesion
to endothelial cells and prevents metastasis by blocking E-selectin expression. Cancer Res. 2000;60(14):397884.
6. Wittig BM, Kaulen H, Thees R, Schmitt C, Knolle P, Stock J et al. Elevated serum E-selectin in patients with
liver metastases of colorectal cancer. Eur J Cancer. 1996;7(8):1215-8.
7. Eichbaum C, Meyer AS, Wang N, Bischofs E, Steinborn A, Bruckner T et al. Breast cancer cell-derived
cytokines, macrophages and cell adhesion: implications for metastasis. Anticancer Res. 2011;31(10):3219-27.
8. Soto MS, Serres S, Anthony DC, Sibson NR. Functional role of endothelial adhesion molecules in the early
stages of brain metastasis. Neuro Oncol. 2014;16(4):540-51.
9. Constantinescu AA, Vink H, Spaan JA. Endothelial cell glycocalyx modulates immobilization of leukocytes
at the endothelial surface. Arterioscler Thromb Vasc Biol. 2003;23(9):1541-7.
10. Reitsma S, Slaaf DW, Vink H, van Zandvoort MA, oude Egbrink MG. The endothelial glycocalyx:
composition, functions, and visualization. Pflugers Arch. 2007;454(3):345-59.
11. Becker BF, Chappell D, Bruegger D, Annecke T, Jacob M. Therapeutic strategies targeting the endothelial
glycocalyx: acute deficits, but great potential. Cardiovasc Res. 2010;87(2):300-10.
12. Weksler BB, Subileau EA, Perriere N, Charneau P, Holloway K, Leveque M et al. Blood-brain barrierspecific properties of a human adult brain endothelial cell line. FASEB J. 2005;19(13):1872-4.
13. Bottenstein JE, Sato GH. Growth of a rat neuroblastoma cell line in serum-free supplemented medium.
Proceedings of the National Academy of Sciences of the United States of America. 1979;76(1):514-7.
15
14. Miles DW, Happerfield LC, Naylor MS, Bobrow LG, Rubens RD, Balkwill FR. Expression of tumour
necrosis factor (TNF alpha) and its receptors in benign and malignant breast tissue. International journal of
cancer Journal international du cancer. 1994;56(6):777-82.
15. Tan GJ, Peng ZK, Lu JP, Tang FQ. Cathepsins mediate tumor metastasis. World journal of biological
chemistry. 2013;4(4):91-101. doi:10.4331/wjbc.v4.i4.91.
16. Chen Q, Fei J, Wu L, Jiang Z, Wu Y, Zheng Y et al. Detection of cathepsin B, cathepsin L, cystatin C,
urokinase plasminogen activator and urokinase plasminogen activator receptor in the sera of lung cancer
patients. Oncol Lett. 2011;2(4):693-9. doi:10.3892/ol.2011.302.
17. Rapisarda A, Melillo G. Role of the VEGF/VEGFR axis in cancer biology and therapy. Adv Cancer Res.
2012;114:237-67. doi:10.1016/b978-0-12-386503-8.00006-5.
18. Winiarski BK, Wolanska KI, Rai S, Ahmed T, Acheson N, Gutowski NJ et al. Epithelial ovarian cancerinduced angiogenic phenotype of human omental microvascular endothelial cells may occur independently of
VEGF signaling. Translational oncology. 2013;6(6):703-14.
19. Kendall AC, Whatmore JL, Winyard PG, Smerdon GR, Eggleton P. Hyperbaric oxygen treatment reduces
neutrophil-endothelial adhesion in chronic wound conditions through S-nitrosation. Wound repair and
regeneration : official publication of the Wound Healing Society [and] the European Tissue Repair Society.
2013;21(6):860-8. doi:10.1111/wrr.12108.
20. Cucullo L, Hossain M, Tierney W, Janigro D. A new dynamic in vitro modular capillaries-venules modular
system: cerebrovascular physiology in a box. BMC neuroscience. 2013;14:18. doi:10.1186/1471-2202-14-18.
21. Singh A, Satchell SC, Neal CR, McKenzie EA, Tooke JE, Mathieson PW. Glomerular endothelial
glycocalyx constitutes a barrier to protein permeability. J Am Soc Nephrol. 2007;18(11):2885-93.
22. Barker AL, Konopatskaya O, Neal CR, Macpherson JV, Whatmore JL, Winlove CP et al. Observation and
characterisation of the glycocalyx of viable human endothelial cells using confocal laser scanning microscopy.
Physical Chemistry Chemical Physics. 2004;6(5):1006-11. doi:10.1039/B312189E.
23. Zeng Y, Ebong EE, Fu BM, Tarbell JM. The structural stability of the endothelial glycocalyx after
enzymatic removal of glycosaminoglycans. PloS one. 2012;7(8):e43168. doi:10.1371/journal.pone.0043168.
24. Cai B, Fan J, Zeng M, Zhang L, Fu BM. Adhesion of malignant mammary tumor cells MDA-MB-231 to
microvessel wall increases microvascular permeability via degradation of endothelial surface glycocalyx. J Appl
Physiol. 2012;113(7):1141-53.
16
25. Santos LV, Cruz MR, Lopes Gde L, Lima JP. VEGF-A levels in bevacizumab-treated breast cancer patients:
a systematic review and meta-analysis. Breast cancer research and treatment. 2015;151(3):481-9.
doi:10.1007/s10549-015-3410-7.
26. Reddy CL, Yosef N, Ubogu EE. VEGF-A165 potently induces human blood-nerve barrier endothelial cell
proliferation, angiogenesis, and wound healing in vitro. Cellular and molecular neurobiology. 2013;33(6):789801. doi:10.1007/s10571-013-9946-3.
27. Herbst RS, Prager D, Hermann R, Fehrenbacher L, Johnson BE, Sandler A et al. TRIBUTE: a phase III trial
of erlotinib hydrochloride (OSI-774) combined with carboplatin and paclitaxel chemotherapy in advanced nonsmall-cell lung cancer. J Clin Oncol. 2005;23(25):5892-9. doi:10.1200/jco.2005.02.840.
28. Das M, Wakelee H. Targeting VEGF in lung cancer. Expert opinion on therapeutic targets. 2012;16(4):395406. doi:10.1517/14728222.2012.669752.
29. Mocellin S, Nitti D. TNF and cancer: the two sides of the coin. Front Biosci. 2008;13:2774-83.
30. Boldrini L, Gisfredi S, Ursino S, Lucchi M, Melfi F, Mussi A et al. Tumour necrosis factor-alpha:
prognostic role and relationship with interleukin-8 and endothelin-1 in non-small cell lung cancer. International
journal of molecular medicine. 2006;17(5):887-92.
31. Bongartz T, Sutton AJ, Sweeting MJ, Buchan I, Matteson EL, Montori V. Anti-TNF antibody therapy in
rheumatoid arthritis and the risk of serious infections and malignancies: systematic review and meta-analysis of
rare harmful effects in randomized controlled trials. JAMA : the journal of the American Medical Association.
2006;295(19):2275-85. doi:10.1001/jama.295.19.2275.
32. Wegiel B, Jiborn T, Abrahamson M, Helczynski L, Otterbein L, Persson JL et al. Cystatin C is
downregulated in prostate cancer and modulates invasion of prostate cancer cells via MAPK/Erk and androgen
receptor pathways. PLoS One. 2009;4(11):0007953.
33. Burger AM, Leyland-Jones B, Banerjee K, Spyropoulos DD, Seth AK. Essential roles of IGFBP-3 and
IGFBP-rP1 in breast cancer. Eur J Cancer. 2005;41(11):1515-27. doi:10.1016/j.ejca.2005.04.023.
34. Pen A, Moreno MJ, Durocher Y, Deb-Rinker P, Stanimirovic DB. Glioblastoma-secreted factors induce
IGFBP7 and angiogenesis by modulating Smad-2-dependent TGF-beta signaling. Oncogene. 2008;27(54):683444. doi:10.1038/onc.2008.287.
35. Akaogi K, Okabe Y, Sato J, Nagashima Y, Yasumitsu H, Sugahara K et al. Specific accumulation of tumorderived adhesion factor in tumor blood vessels and in capillary tube-like structures of cultured vascular
endothelial cells. Proc Natl Acad Sci U S A. 1996;93(16):8384-9.
17
36. Chen LH, Liu DW, Chang JL, Chen PR, Hsu LP, Lin HY et al. Methylation status of insulin-like growth
factor-binding protein 7 concurs with the malignance of oral tongue cancer. Journal of experimental & clinical
cancer research : CR. 2015;34:20. doi:10.1186/s13046-015-0138-5.
37. Ogawa J, Sano A, Inoue H, Koide S. Expression of Lewis-related antigen and prognosis in stage I non-small
cell lung cancer. The Annals of thoracic surgery. 1995;59(2):412-5.
38. Martin-Satue M, Marrugat R, Cancelas JA, Blanco J. Enhanced expression of alpha(1,3)-fucosyltransferase
genes correlates with E-selectin-mediated adhesion and metastatic potential of human lung adenocarcinoma
cells. Cancer Res. 1998;58(7):1544-50.
39. Henry CB, Duling BR. TNF-alpha increases entry of macromolecules into luminal endothelial cell
glycocalyx. Am J Physiol Heart Circ Physiol. 2000;279(6):H2815-23.
40. Foster RR, Armstrong L, Baker S, Wong DW, Wylie EC, Ramnath R et al. Glycosaminoglycan regulation
by VEGFA and VEGFC of the glomerular microvascular endothelial cell glycocalyx in vitro. Am J Pathol.
2013;183(2):604-16. doi:10.1016/j.ajpath.2013.04.019.
41. Fu BM, Shen S. Structural mechanisms of acute VEGF effect on microvessel permeability. Am J Physiol
Heart Circ Physiol. 2003;284(6):30.
18
Figure 1
Fig. 1 Potential prometastatic proteins were identified and quantified in human NSCLC cell secretome. ELISA
analysis of CC, CL, IGFBP-7, TNF-α and VEGFA concentration in A549 and SK-MES-1 CM. All experiments
were carried out in duplicate on three separate samples.
19
Figure 2
Fig. 2 Increased levels of E-selectin detected on the surface of human cerebral microvascular endothelial cells in
response to lung tumour CM and secreted-proteins. Confluent hCMEC/D3 cells were stimulated for 4hrs with
A549 and SK-MES-1 CM (a) along with CC (c), CL (d), IGFBP-7 (e), VEGF (f) and TNF-α (g). E-selectin
levels from 3 separate experiments, carried out in quadruplets, were statistically analysed (n≥12, mean ± SD,
*p≤0.05 and **p≤0.01, ***p≤0.005 and ****p≤0.0001). Enhanced E-selectin was also detected within 30
minutes of CM (a) and secreted-proteins (b) exposure (*p≤0.05 and **p≤0.01, n=6).
20
Figure 3
Fig. 3 Lung tumour CM and secreted-proteins enhance adhesion of A549 and SK-MES-1 cells to human brain
endothelial monolayers. Confluent hCMEC/D3 cells in µ-slides were exposed to A549 CM or SK-MES-1 CM
for 30min and 24hr prior to perfusion with calcein AM stained A549 (a) or SK-MES-1 cells (c). A549 and SKMES-1 adhesion to hCMEC/D3 cells treated with 80ng/ml CC, 10ng/ml CL, 200ng/ml IGFBP-7, 0.2ng/ml
VEGF and 160pg/ml TNF-α at 4 and 24hrs (b) and (d). Mean values from three separate experiments were
statistically analysed. * p≤0.05 versus control levels (100%). (e) Representative images of adherent calcein-AM
stained A549 [(i) and (ii)] and SK-MES-1 [(iii) and (iv)] cells to DMEM-BS [(i) and (iii)] and CL-treated [(ii)
and (iv)] brain endothelial monolayers. Scale bar = 75µm.
21
Figure 4
Fig. 4 E-selectin inhibition attenuated adhesion of A549 and SKMES-1 lung cancer cells to human cerebral
microvascular endothelial cells. Confluent hCMEC/D3 cells were treated with A549 CM (a), SK-MES-1 CM
(b) and secreted-proteins for 24hr prior to exposing the relevant channels with anti-E-selectin antibody for
30mins. The bars represent cancer cell adhesion to brain endothelial monolayer calculated as a percentage of
control obtained on 3 separate occasions. (* p≤0.05 vs control, #p≤0.05, ##p≤0.01 vs anti-E-selectin). (c) FC
analysis of adhesion molecule ligands and receptors, CD49d, CD11a, CD11b, PSGL-1 and sLeX, on lung
tumour cells. The presented data was collected from 3 different experiments.
22
Figure 5
Fig. 5 Exposure of lung tumour CM and secreted-proteins significantly altered the brain endothelial glycocalyx.
(a) and (b) Integrity of the hCMEC/D3 glycocalyx was assessed by a CBF assay following exposure of ECs for
30min or 24h to A549 and SKMES-1 CM (a), or 80ng/ml CC, 10ng/ml CL, 200ng/ml IGFBP-7, 0.2ng/ml
VEGF and 160pg/ml TNF-α (b). Data from 6 independent experiments, carried out in sextuplicate, is presented
as WGA-FITC fluorescence/μg protein and these values were calculated as a percentage of control. *p≤0.05,
**p≤0.01 vs control level. (c) Hyaluronan (HA) and syndecan-1 levels were measured by ELISA in hCMEC/D3
growth medium following 30 minute treatment with fresh DMEM-BS (control), A549 or SK-MES-1 CM. SKMES-1 CM vs control = p=0.007, n=6. nd= not detected (d) and (e) Representative confocal images of FITCWGA stained hCMEC/D3 cells before (a) and after (b) 30 minute exposure to SKMES-1 CM. Scale bar =
40µM.
23
Additional File 1 - Supplementary methods (detailed description of MS Analysis, E-selectin
ELISA, flow cytometry and legends for Additional file 3: Figure S1 and Additional file 4:
Figure S2).
Supplementary Methods
Mass Spectrometry Analysis
LC-MALDI TOF/TOF mass spectrometry was carried out by Dr Heidi Fuller (The Robert Jones and
Agnes Hunt Orthopaedic and District Hospital, Oswestry, UK). One millilitre of either A549 or SK-MES-1 CM
or DMEM-BS was mixed with 10l of StrataCleanTM Resin (Stratagene). The mixture was vortexed and
centrifuged for 1 minute at 2000rpm. The supernatant was removed and the pelleted samples were digested
overnight at 30C with 20l of trypsin (20g/ml, sequencing grade, Promega, Southampton, UK). Samples were
then separated by liquid chromatography on an UltiMate 3000 Dionex system where they were loaded onto a
C18 trapping column (10l) and then eluted onto a C18 PepMap TM analytical column. Peptides were eluted
using a 40 minute acetonitrile (MeCN) gradient (2%-50% MeCN/ H2O) followed by a further elution at 90%
MeCN for 10 minutes. The column was washed and equilibrated for 10 minutes. Samples were spotted at 10
second intervals using a ProbotTM Microfraction Collector with alpha-cyano-4-hydroxycinnamic acid (α-CHCA,
3mg/ml in 70% MeCN/ 0.1% TFA) at a continuous flow rate of 1.2l/min. Sample plates were then placed in a
4800 MALDI TOF/TOFTM Analyser (Applied Biosystems, Warrington, UK) for mass spectrophotometry and
close external standards were used to calibrate the instrument. Proteins were identified using Peak Explorer TM
software. Identified proteins with a total ion score (Confidence Interval, C.I) of >95%, with two or more
identified peptides, were deemed as true interactions. However, any proteins with lower scores could indeed be
true if their presence could be confirmed by further biochemical methods. Total ion scores were calculated from
weighted ion scores for individual peptides that were matched to a given protein.
Cell surface based E- and P-selectin ELISA
Following treatment, hCMEC/D3 cells were washed with DMEM-BS and incubated with mouse antihuman antibody against E- or P-selectin (5µg/ml in PBS-1% BSA, 50µl per well) for 1 hour at 4 oC. Following
three washes with PBS-BSA, a biotinylated goat anti-mouse secondary antibody (0.64µg/ml in PBS-BSA, 50µl
per well) (Sigma Aldrich, Gillingham, UK) was applied to the wells for another hour at 4 oC. After washing as
above, the cells were incubated with a solution of streptavidin alkaline phosphatase (2µg/ml in Tris buffered
24
saline, 100µl per well) for 30 minutes at 4 oC, washed and incubated with para-nitrophenyl phosphate (p-NPP)
substrate solution for 30 minutes in the dark. The enzyme reaction was stopped with 3M NaOH prior to reading
the absorbance on a plate reader (BMG Labtech FLUOstar OPTIMA, BMG LABTECH Ltd. Aylesbury, UK) at
405nm.
Flow cytometry analysis of adhesion molecule counter ligands/receptors
Lung tumour cells at approximately 80% confluence were detached using non-enzymatic cell dissociation
solution (Sigma Aldrich, Gillingham, UK) and centrifuged (600g, 5 minutes) before incubation with 10% (v/v)
blocking serum (Sigma Aldrich, Gillingham, UK) for 30 minutes at room temperature (RT). After centrifugation
(600g, 10 minutes), A549 and SK-MES-1 cells were incubated with 1µg mouse anti-human primary antibodies
against CD49d, CD11a, CD11b, P-selectin glycoprotein ligand-1 (PSGL-1) (all ImmunoTools, Germany) and
sialyl Lewis X (sLeX, BD Biosciences, Oxford, UK) for 30 minutes on ice. The cells were then washed in PBS
prior to incubation with anti-mouse secondary antibody conjugated to FITC (1:300) (Sigma Aldrich,
Gillingham, UK) for 30 minutes in the dark at RT. Following washes, samples (in 500µl PBS) were analysed on
a Beckman Coulter Cell Lab Quanta SC flow cytometer and data for 10,000 events were collected. The
percentage of positively stained cells was identified and the results were recorded. Cells stained with relevant
isotype control primary antibody (IgM, κ) were subtracted from positive cell staining with test antibody to
control for test antibody specificity and avoidance of false-positive results.
Hyaluronan Quantikine® Immunoassay and Syndecan-1 ELISA
Fully confluent hCMEC/D3 cells, grown in 12 well plates, were treated for 30 minutes with either
DMEM-BS or freshly prepared A549 or SK-MES-1 CM at 37oC. The supernatants (500µl), after centrifugation
at 600g for 10 minutes and 4oC, were collected and analysed immediately using the Hyaluronan Quantikine
ELISA kit (R&D Systems, Abingdon, UK) or the Syndecan-1 (CD138) ELISA kit (Abcam, Cambridge, UK)
according to the manufacturer’s instructions. Preliminary experiments were performed to determine optimum
dilution requirements. In order to determine the exact amount of hyaluronan being released from the brain
endothelial glycocalyx, HA and syndecan-1levels were also measured in both A549 and SK-MES-1 CM alone.
The latter was then deducted from the hCMEC/D3 supernate values to give the final concentrations.
25
Figure S1 Effects of IgG1 isotype control antibody on the adhesion of A549 and SK-MES-1 cells to TNF-α
stimulated hCMEC/D3 monolayers. Here, confluent brain endothelial monolayers were treated with 160pg/ml
TNF-α for 24hrs prior to incubation with 5µg/ml mouse monoclonal IgG1 antibody for 30minutes at 37oC.
Thereafter, hCMEC/D3 monolayers were perfused with Calcein AM-stained A549 or SK-MES-1 cells for
10minutes. Exposure of the isotype-matched antibody had no significant effects on the increased adhesion of
A549 and SK-MES-1 cells induced by TNF-α.
Figure S2 Isotype-matched (IgM, κ) and positive controls for flow cytometry analysis of adhesion molecule
ligands/receptors. The EV-SS plot along with their corresponding histogram and tables obtained for the isotype
control antibody indicate that the positive staining observed during sialyl Lewis X expression studies was not
due to any non-specific binding of the primary antibody in both A549 (a) and SK-MES-1 (b) lung tumour cells.
Expression profiles of CD11a, CD49d and CD11b in U937, a monocytic cell line, and freshly isolated
monocytes (positive controls). U937 cells were incubated with 1μg CD11a (c) and CD49d (d) primary
antibodies and FITC-conjugated secondary antibody prior to expression analysis by flow cytometry. However,
U937 displayed very low levels of CD11b and as such freshly isolated monocytes were used as a positive
control for this antigen (e).
26
Additional file 2
Rank
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Protein Name
Transferrin [Homo sapiens]
fibronectin precursor [Homo sapiens]
actin, beta [Homo sapiens]
glucose-6-phosphate dehydrogenase [Homo
sapiens]
nucleobindin 1 [Homo sapiens]
cathepsin L1 preproprotein
[Homo sapiens]
beta amyloid peptide precursor
Chain A, Human Cystatin C; dimeric form
With 3D domain Swapping
aldo-keto reductase family 1, member B10
[Homo sapiens]
FLJ00343 protein [Homo sapiens]
enolase 1 [Homo sapiens]
Heterogeneous nuclear ribonucleo- protein
A2/B1 isoform B1
[Homo sapiens]
NUANCE [Homo sapiens]
Heterogeneous nuclear ribonucleo protein
A1 (Helix-destabilizing protein) (Singlestrand RNA-binding)
insulin-like growth factor binding protein 7
[Homo sapiens]
cingulin [Homo sapiens]
aldo-keto reductase family 1, member D1
[Homo sapiens]
PREDICTED: cathepsin H isoform 6 [Pan
troglodytes]
BIGH3 [Homo sapiens]
dihydrodiol dehydrogenase isoform DD 1
[Homo sapiens]
laminin alpha5 chain precursor
[Homo sapiens]
Titin [Homo sapiens]
Accession
Number
gi|37747855
Peptide
Count
9
Total Ion Score
C.I. %
100
gi|31397
gi|14250401
gi|452269
9
8
6
100
100
99.99951
gi|20070228
gi|4503155
6
4
99.99811
99.99753
gi|226343
gi|14278690
5
4
99.99565
99.92613
gi|20127592
4
99.90828
gi|21748542
gi|4503571
gi|14043072
6
2
3
99.88059
99.65686
99.18628
gi|17016967
gi|133254
6
1
99.05587
99.01943
gi|4504619
2
98.71626
gi|16262452
gi|5174695
3
2
98.55297
98.3038
gi|114658412
2
98.26029
gi|2996636
gi|556516
2
2
97.89391
97.26803
gi|20147503
4
96.042
gi|17066105
11
95.33925
Table A1 Proteins identified in the A549 CM sample. The protein rank, peptide count and total ion score
C.I% are shown.
27
Rank
Protein Name
1
2
Transferrin [Homo sapiens]
Chain A, Crystal Structure Of Human
Serum Albumin
Insulin-like growth factor binding protein
7
[Homo sapiens]
Beta actin [Homo sapiens]
Heat shock 90kDa protein 1, beta [Homo
sapiens]
Prosaposin isoform b prepro protein
[Homo sapiens]
FLJ00343 protein
[Homo sapiens]
Unnamed protein product [Homo sapiens]
Glucose-6-phosphate
dehydrogenase
[Homo sapiens]
Alpha glucosidase II alpha subunit
isoform 2
[Homo sapiens]
Transforming growth factor, betainduced, 68kDa
[Homo sapiens]
Histone cluster 1, H2ae
[Homo sapiens]
Unnamed protein product [Homo sapiens]
heat shock protein 90kDa alpha
(cytosolic), class A member 1 isoform 2
[Homo sapiens]
Fibronectin precursor
[Homo sapiens]
Tubulin, alpha 1B [Mus musculus]
Sparc/osteonectin, cwcv and kazal-like
domains proteo glycan (testican) 1
[Homo sapiens]
Heat shock 70kDa protein 8 isoform 1
[Homo sapiens]
Collagen, type VI, alpha 1 precursor
[Homo sapiens]
Chain A, Human Cystatin C; Dimeric
Form With 3d Domain Swapping
Heterogeneous nuclear ribo nucleoprotein
A2/B1 isoform B1 [Homo sapiens]
N2B-Titin Isoform
[Homo sapiens]
Unnamed protein product [Homo sapiens]
Titin [Homo sapiens]
Heat shock protein 70 testis variant
[Homo sapiens]
Alpha2-HS glycoprotein [Homo sapiens]
Cytosolic
thyroid
hormone-binding
protein (EC 2.7.1.40)
Nebulin
Golgi antigen gcp372
[Homo sapiens]
Histone H4
Putative [Homo sapiens]
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
28
Accession
Number
gi|37747855
gi|3212456
Peptide
Count
6
2
Total Ion Score
C.I. %
100
100
gi|4504619
4
100
gi|4501885
gi|20149594
5
5
100
100
gi|110224476
2
99.99984051
gi|21748542
5
99.99680322
gi|16552261
gi|452269
3
4
99.99665257
99.99198854
gi|38202257
4
99.98806804
gi|4507467
4
99.96470515
gi|10645195
2
99.95536132
gi|31092
gi|154146191
3
3
99.95127953
99.81132624
gi|31397
3
99.80197910
gi|34740335
gi|21265163
1
2
99.43543841
99.42758439
gi|5729877
2
99.40198550
gi|87196339
3
99.33822373
gi|14278690
2
99.24714502
gi|14043072
3
99.16109757
gi|17066104
8
98.85694535
gi|7020225
gi|17066105
gi|3461866
3
11
2
98.71556746
98.61296195
98.60455938
gi|2521981
gi|338827
1
4
98.24728410
97.80359977
gi|19856971
gi|808869
2
4
97.60521144
97.50201880
gi|223582
gi|553734
1
1
97.20931261
96.08518438
32
33
34
35
KIAA0841 protein
[Homo sapiens]
Acyl-Coenzyme A dehydroge nase, very
long chain isoform 1 precursor [Homo
sapiens]
Protein kinase C substrate 80K-H isoform
1
[Homo sapiens]
Aldolase A [Homo sapiens]
gi|33869699
3
93.72359101
gi|4557235
3
92.47145020
gi|48255889
3
92.11664266
gi|4557305
4
91.29615
Table A2 Proteins identified in the SK-MES-1 CM sample. The protein rank, peptide count and total ion
score C.I% are shown.
Rank
Protein Name
Accession
Number
gi|37747855
Peptide
Count
8
Total Ion Score
C.I. %
100
1
Transferrin [Homo sapiens]
2
gi|15624075
1
98.62989076
gi|393036
3
98.25516197
4
TGF-beta resistance-associated protein
TRAG [Homo sapiens]
human type 3 inositol 1,4,5-trisphosphate
receptor
hypothetical protein [Homo sapiens]
gi|5262501
2
97.51826734
5
centromere protein F [Homo sapiens]
gi|55770834
4
95.93681220
6
restin isoform a [Homo sapiens]
gi|4506751
2
95.43169441
3
Table A3 Proteins identified in DMEM-BS sample. The protein rank, peptide count and total ion score C.I%
are shown.
29
Components
Sodium Chloride
Potassium Chloride
Calcium Chloride 2H20
Magnesium Sulphate 7H20
Sodium Dihydrous Phosphate 2H20
Dextrous Anhydrous
Ferric Nitrate 9H20
L-Glutamine
Sodium Pyruvate
Phenol Red
Sodium Bicarbonate
L-Arginine-HCl
L-Histidine-HCl-H20
L-Isoleucine
L-Leucine
L-Lysine-HCl
L-Methionine
L-Phenylalanine
L-Threonine
L-Tryptophan
L-Valine
Glycine
L-Serine
L-Cystine
L-Tyrosine
Choline-Chloride
Nicotinamide
D-Calcium Pantothenate
Pyridoxine-HCl
Thiamine-HCl
Riboflavin
Folic Acid
I-Inositol
Sodium Selenite
Putrescine
Tri-Iodo L Thyronine
Progesterone
L-Thyroxine
Insulin (Human Recombinant)
Human Transferrin
Contents g/L
6.4
0.4
0.264
0.2
0.141
1
0.0001
0.0116
0.0022
0.0003
0.074
0.084
0.042
0.105
0.105
0.146
0.03
0.066
0.095
0.016
0.094
0.03
0.042
0.048
3.6
0.004
0.004
0.004
0.004
0.004
0.0004
0.004
0.0072
0.0000346
0.0145
0.000303
0.0000566
0.00036
0.001
0.1
Table B Composition of Dulbecco’s Modified Eagles Medium-Bottenstein Sato (DMEM-BS)
30
Additional file 3: Figure S1. Supplementary cell adhesion figure depicting specificity of anti-E-selectin
antibody.
31
Additional file 4: Figure S2. Supplementary flow cytometry scatter plots for isotypematched (IgM, κ) and positive controls
32