Novel N-acetyl-Glycol-split heparin biotin-conjugates endowed with anti-heparanase activity

Article abstract of DOI:10.1016/j.ejmech.2019.111831

Heparanase is regarded as a promising target for anticancer drugs and Ronepastat is one of the most promising heparanase inhibitors insert in clinical study for Multiple Myeloma Therapy. To improve its pharmacokinetic/pharmacodynamic profile, as well to have an antidote able to neutralize its activity in case of over dosages or intolerance, a new class of its derivatives was obtained inserting non-carbohydrate moieties of different length between the polysaccharide chain and biotin or its derivatives. In vitro these novel derivatives maintain the anti-heparanase activity without induced toxicity. The newly synthesized compounds retained the ability to attenuate the growth of CAG myeloma tumors in mice with potency similar, or in one case even higher than that of the reference compound Roneparstat as well as inhibited metastatic dissemination (lung colonization) of murine B16-F10 melanoma cells in vivo.

Full text of DOI:10.1016/j.ejmech.2019.111831

Journal Pre-proof
Novel N-acetyl-Glycol-split heparin biotin-conjugates endowed with anti-heparanase
activity
Emiliano Esposito, Israel Vlodavsky, Uri Barash, Giuseppe Roscilli, Ferdinando M.
Milazzo, Giuseppe Giannini, Annamaria Naggi
PII:
S0223-5234(19)30983-3
DOI:
https://doi.org/10.1016/j.ejmech.2019.111831
EJMECH 111831
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 17 June 2019
Revised Date: 10 October 2019
Accepted Date: 28 October 2019
Please cite this article as: E. Esposito, I. Vlodavsky, U. Barash, G. Roscilli, F.M. Milazzo, G. Giannini,
A. Naggi, Novel N-acetyl-Glycol-split heparin biotin-conjugates endowed with anti-heparanase activity,
European Journal of Medicinal Chemistry (2019), doi: https://doi.org/10.1016/j.ejmech.2019.111831.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
©
2019 Published by Elsevier Masson SAS.

NRE
RE
U2X
ANY6X
CO2-
CH OX
2
O
O
Major disaccharide
repart unit of Heparin
OH
R
O
-
OH
X=SO or H
3
O
-
-
Y= SO or CH CO
3
3
R= H or Polysaccharide chain
OX
NHY
n
[
A]
[C]
[
B]
COO-
CH OSO -
2
3
O
CH OSO -
2 3
O
N
N
N
O
O
OH
OH
O
O
CO2-
O
O
O
N
OH
O
OH
HN NH
H
O
HN NH
X'H
NHAc
H
HN
OSO3-
NHAc
spacers
N
H
N
S
spacers
S
O
CH OSO -
2
3
C1-3
O
O
A1-7
CO2-
OH
O
OH
O
O
O
HN NH
OSO3-
HN
O
H
spacers
N
S
B1-3
O

Novel N-acetyl-Glycol-split heparin biotin-
conjugates endowed with anti-heparanase activity
a
b
b
c
d
Emiliano Esposito, Israel Vlodavsky, Uri Barash, Giuseppe Roscilli, Ferdinando M. Milazzo,
d
a
Giuseppe Giannini, * Annamaria Naggi. *
a
Istituto di Ricerche Chimiche e Biochimiche G. Ronzoni, via G. Colombo, 81, 20133 Milan,
Italy
b
Cancer and Vascular Biology Research Center, Rappaport Faculty of Medicine, Technion,
Haifa, Israel
c
Takis s.r.l., Via Castel Romano 100, I-00128 Roma, Italy
d
R&D Alfasigma S.p.A., Via Pontina Km 30,400, Pomezia, I-00071 Roma, Italy
KEYWORDS Heparanase inhibitors, Anticancer agents, Antimetastatic agents, Roneparstat,
Biotinylated agents.
ABSTRACT Heparanase is regarded as a promising target for anticancer drugs and Ronepastat is
one of the most promising heparanase inhibitors insert in clinical study for Multiple Myeloma
Therapy. To improve its pharmacokinetic / pharmacodynamic profile, as well to have an
1

antidote able to neutralize its activity in case of over dosages or intolerance, a new class of its
derivatives was obtained inserting non-carbohydrate moieties of different length between the
polysaccharide chain and biotin or its derivatives. In vitro these novel derivatives maintain the
anti-heparanase activity without induced toxicity. The newly synthesized compounds retained
the ability to attenuate the growth of CAG myeloma tumors in mice with potency similar, or in
one case even higher than that of the reference compound Roneparstat as well as inhibited
metastatic dissemination (lung colonization) of murine B16-F10 melanoma cells in vivo.
INTRODUCTION
1
Roneparstat, N-acetylated glycol split (gs) heparin (also named SST0001) is one of the most
2
3,4
promising heparanase-inhibiting drugs, possessing antitumor activity. Heparanase appears to
be involved in several human pathologies other than tumors such as chronic inflammation,
5
diabetic nephropathy, bone osteolysis, thrombosis, atherosclerosis, as well as rare diseases.
6
,7
Roneparstat has been promisingly tested in a phase I clinical study nevertheless it is
considered useful to improve its pharmacokinetic / pharmacodynamic profile to reduce the
number of daily subcutaneous administrations as well as to have an antidote able to neutralize its
8
,9
activity in case of over dosages or intolerance . Therefore, new Roneparstat-like derivatives
were prepared inserting non-carbohydrate moieties of different length between the
polysaccharide chain and biotin or its derivatives. Exploiting different functional groups of the
polysaccharide, our aim was to preserve the heparanase inhibitory activity, improve
bioavailability and get ready biotin derivatives able to be neutralized by avidin endowed with a
1
0
strong affinity for biotin . In addition, studies have disclosed that folate receptors and more
2

1
1
specifically biotin receptor (BR) are overexpressed in a large number of cancer cells lines and
1
2
solid tumors, and hence may contribute to tumor-directed drug delivery . Besides, biotin
provides functional groups easily detectable by chemical-physical methods or interaction tests
with avidin, thus allowing an easier monitoring of plasma concentrations of the drug.
Roneparstat has been prepared from heparin by periodate oxidation followed by borohydride
1
3
reduction of fully N-desulfated and N-acetylated heparin as represented in Figure. 1.
Figure 1
In this study biotinylation, through a linker of different functional groups of heparin derivatives,
has been investigated by three distinct approaches (A, B, C) exploiting appropriate sequences of
reactions, as schematized in Figure 2.
Figure 2
RESULTS AND DISCUSSION
In the first approach (route [A]) the intermediate (I-3), common to Roneparstat preparation
obtained by periodate oxidation, was used to introduce a biotin-spacer through a reductive
amination of the aldehyde groups, derived from the oxidized diol of uronic acid, with the amine
2
group present on a biotin conjugated linker (abbreviated as biotin-spacer-NH ). In the second
approach (route [B]) starting from N-desulfated heparin, the first intermediate of Roneparstat
synthesis (I-1), unsubstituted glucosamine residues, were N-acylated with biotinylated esters
activated with N-hydroxysuccinimide (abbreviated as Biotin-spacer- NHS). The third method
3

(route [C]) consists of exploiting the aldehyde group of the reducing end (RE) residues of
heparin chain to conjugate a Biotin-linker (abbreviated as Biotin-spacer-N ) through “click
3
1
4
“
chemistry . In particular, incorporation of biotin derivatives at the reducing terminus was
achieved by a two-step procedure: (i) first, N-desulfated, N-acetylated heparin (I-2) was
reductively aminated with an alkyne-amine such as propargylamine, and then subjected to
glycol-splitting reaction to give end functionalized gs N-acetyl heparin (I-6). (ii) Second,
reductively aminated heparin was linked with biotinylated spacers containing an azide group,
using copper (I)-catalyzed azide-alkyne 1,3 dipolar-cycloaddition (CuAAC), so called “clik
1
5
reaction”. To build-up a library of compounds, which differ not only in the position of the
substituents, but also in their length and hydrophilicity, seven compounds have been used
(
Figure 3) classified into three types according to the functional group employed for the coupling
with the polysaccharide. One of the items (1) is commercially available the others (2-7) were
synthesized as described in the experimental section.
Figure 3
Biotinylated spacers 1, 4 and 6 contain only a simple 1,5 diamino-pentyl chain, as spacer arm
between biotin moiety and heparin chain thus, moving away the biotin residue to reduce the
steric hindrance that could interfere with biotin-binding assays. Differently, spacer 2 and 7
contain, in addition to the diamino pentyl moiety, a hydrophilic short chain of polyethylene
glycol (PEG), incorporated in the final compound [Biotin-PEG-X]. The polypeptide spacer 3,
showing a length double than the biotin-Peg chain of 2 is constituted by the amino acidic
sequence Gly-Gly-Gly-Ser-Gly-Gly-Gly-Ser-Gly-Gly-Gly. Such sequence is well known and
4

widely used in protein fusions, conferring a high flexibility and, due to the absence of proteases
1
6
active on this sequence, is also quite stable .
Biotinylated-linker 2 [Biotin-PEG
dicyclohexylcarbodiimide (DCC), from the commercial products 5-(biotinamide)pentinylamine
1), and 15-(tert-Butyloxycarbonyl)amino-4,7,10,13-tetra-oxa-pentanoic acid (Boc-NH-PEG4-
4 2
-NH ] was synthesized, in presence of N,N'-
(
CO H), after activation with N-hydroxysuccinimide (NHS) followed by deprotection of t-
2
butyloxycarbonyl group with trifluoroacetic acid. (Fig. S1, Supporting Information).
Biotinylated-linker 3 [Biotin-AA-NH ] was prepared starting the commercial product 1, which
2
was coupled with protected linear precursor peptide 9 (Fig. S2, Supporting Information) using
the conventional uronum-based reagent O-(6-chlorobenzotriazol-1-yl)-N,N,N’,N’-
1
7
tretramethyluronium (HCTU) ]; the latter obtained, after cleavage from the resin, in solid phase
synthesis MW-SPPS heating assisted.
Biotinylated-linker 4 [Biotin-N ] was synthetized for direct diazo-transfer of the terminal
3
1
8
primary amine present on biotinylated linker 1 to give the correspondent azide . In the same
way we prepared the glycine-azide (17) obtained from the corresponding commercial glycine
(16), subsequently linked with 1 through standard coupling conditions to obtain Biotinylated-
linker 5 [Biotine-glycine-N ]. (Fig. S3, Supporting Information)
3
4
Biotinylated-linker 7 [Biotin-PEG -NHS] was synthesized starting from commercial 6-
biotinylamino-hexanoic acid 12 (Biotin-Ahx-OH) and 15-Amino-4,7,10,13-tetraoxa-
pentadecanoic acid t-butyl ester 13 (tBuO-Peg -NH ) through standard coupling conditions
4
2
reactions (HCTU) and subsequently, after the deprotection of t-butyloxycarbonyl group with
trifluoroacetic acid, activated with N-hydroxysuccinimide. (Fig. S4, Supporting Information).
5

Synthesis of Roneparstat-like derivatives
Preparation of novel biotinylated N-acetyl-glycol split heparins assembled with distinct
synthetic strategy, involving the reacting groups of three different residues of the heparin chain is
schematized in Fig. 2, and their chemical-physical properties and heparanase inhibition activity
summarized in Table 1.
Conjugation on gs-uronic acid unit - Derivatives substituted at the uronic acid unit (D-
glucuronic, GlcA, or L-iduronic, IdoA) were synthesized through derivatization of glycol split
moiety. As an example, in Fig. 4 the scheme of reaction for compound A3 is reported. N-
acetylated heparin (I2) was oxidized with periodate obtaining the cleavage of C3-C2 bond of the
pyranose ring of non-sulfated uronic acids. The resulting oxy-N acetylated heparin (I-3), instead
of being stabilized immediately by borohydride reduction, leading to Ronepastat, was subjected
to reductive amination reactions with different stoichiometric ratios of glycol-split residues and
biotin-spacers (1, 2 or 3) to obtain derivatives with different substitution degrees (SD), followed
4
by NaBH reduction of the unreacted aldehyde groups.
Table 1
Figure 4
Given the stoichiometric ratio between the reagents, the degree of substitution (DS), express as
percentage of biotinylated disaccharide versus their total number, can be modulated (Tab. 1).
6

Considering that the glycol splitting can occur only at the level of non-sulfated GlcA and IdoA
residues, which represent about 25 % of total uronic acid in a heparin chain, the biotinylation
maximum DS obtainable is about 50% when all the gs generated aldehydes were substituted. As
shown in Table 1 N-acetyl gs-heparin biotinylated derivatives show a wide range of DS values
ranging from 3.0 - 41 %.
Conjugation on N-desulfated glucosamine units - Conjugation of biotinylated linkers (4,5) to N-
desulfated glucosamine unit was achieved through their N-hydroxysuccinimide (NHS) esters.
The synthetic process shown in Fig. 5 comprises five steps. N-desulfated heparin (I-1) was
oxidized with periodate under controlled conditions to avoid concomitant glycol-splitting at level
of glucosamine C2-C3 linkage bearing vicinal amino and hydroxyl group and significant
1
9,20
depolymerization.
For this purpose pH was maintained at about 3.5 and the reaction time was
reduced to 6 h. By this way, almost all the amino functions present on the heparin chains were
preserved and available, after reduction of aldehyde groups, to be conjugate with the biotinylated
linkers-succinimide derivative (6 or 7). Finally, the unreacted amine groups have been
exhaustively N-acetylated in the presence of acetic anhydride to obtain the N-acetylated-N-
biotinylated-glycol split heparins (B1-3). As in the case of uronic acids derivatization, the DS
can be modulated as a function of the stoichiometric ratio between the reagents.
Figure 5
7

A first attempt to incorporate biotin-spacer 1 at the glucosamine units was carried out by
acylation of the N-desulfated heparin intermediate (I-1), before the periodate oxidation and
reduction with borohydride. Even though conjugates were obtained with good yields, this path
was abandoned due to a partially periodate oxidation of the tetrahydrothiophene moiety (sulfinyl
group SO) of biotin (data not shown)
Conjugation on chain reducing end. To conjugate the biotinylated spacer (4, 5) on the heparin
terminal reducing-end, the N-acetylated heparin (I-2) was reductively aminated with
propargylamine in order to introduce an alkyne function (I-5) (Fig. 6). The modification at the
1
reducing end of I-4 was confirmed by its H-NMR spectrum that clearly revealed the presence of
the characteristic acetylenic proton at 2.95 ppm. The alkyne intermediate was oxidized with
periodate and reduced with borohydride in order to obtain intermediate compound I-6.
Figure 6
Click reaction between the intermediate N-acetylated-glycol split heparin alkyne aminated at
RE (I-6) and biotinylated spacers functionalized with a terminal azide group (biotin-spacer-N ,4
3
or 5) was carried out, at rt in a mixture of H O/tBuOH (2:1, v/v), in the presence of CuSO and
2
4
sodium ascorbate. The excess of spacer used to ensure the complete reaction with the alkyne
amine function was easily removed by extensive dialysis (membrane with cut-off of 3500 Da).
In conclusion approaches [A] and [B] led to a random biotinylation along the heparin chains
while the “clik reaction” [C] allowed derivatization at the RE units.
8

Characterization by NMR spectrometry
Purity and substitution degrees of all products, including intermediates, were mainly evaluated
by one- and two-dimension NMR spectroscopy. The N-acetylation degree was determined by
1
3
3
integration of C-NMR signal at 60 and 56 ppm corresponding to the C-2 of GlcNSO and
GlcNAc, respectively. (Fig. S5, Supporting Information). The percent of glycol-split residues (gs)
present in all biotinylated compounds was calculated by using the HSQC-NMR spectroscopy
2
1
approach previously developed for the structural characterization of unmodified heparins
namely from the H1/C1 signals in the anomeric region, typical for the glycol-split residues, at
.94/106.9 ppm (gsIdoA, I) and 4.71/106.5 ppm (gsGlcA, G) versus the area of the anomeric
signal of the 2-O sulfated IdoA at 5.2/102 ppm (I B) as follows: gsIdoA= I/(I+G+ I ) x 100 and
4
2
s
2s
respectively for the glucuronic unit gsGlcA= G/(I+G+ I2s) x 100.
1
Degree of biotinylation per disaccharide unit (DS) was determined by H-NMR spectroscopy
comparing the integrated peak area of the characteristic geminal diastereotopic proton 2.75 ppm
(
H-3´; α) of the tetrahydrothiophene ring present in all biotin-linkers versus that of methyl
protons of the N-acetyl group of the polysaccharide, at 2.01 ppm (β), as follows: DS= α/(β/3) x
00. (Fig. S6 Supporting Information). For compounds bearing all glucosamine residues N-
1
acetylated (pathway [A] and [C]) the N-acetyl signal represent all the disaccharide units. In the
case of different N-derivatives of glucosamine (Compound B1-3), the formula has been changed
as follows: DS= α/ (α + β/3) x 100 where the sum of the integrals of the protons (α + β/3)
represents the sum of the disaccharide units as reported in Table 1.
9

1
The click reaction was confirmed by H-NMR for the appearance of the aromatic proton
singlet at δ 8.2-8.05 ppm typical of triazole ring.
For three derivatives (A7, B3 and C3) the molecular ratio of biotin incorporation was also
determined applying the classical HABA (4'-hydroxyazobenzene-2-carboxylic acid)-Avidin
assay and the results confirmed the degree of biotinylation measured by spectroscopy analysis
(data not shown).
Heparanase inhibitory activity. To assess the inhibition of the heparanase catalytic activity,
2
2
compounds were tested by an in vitro assay originally developed by Hammond and coworkers
based on heparanase-mediated cleavage of the synthetic heparin fragment, the pentasaccharide
Fondaparinux (AGA*IA). The results for newly synthesized compounds, expressed as IC50
values generated from the dose-response curves, are reported in Table 1 with that of the
reference compound Roneparstat.
According to their IC values, all the biotinylated derivatives were shown to inhibit the
5
0
heparanase enzymatic activity with a potency (low nanomolar concentrations) similar to that of
Roneparstat. Moreover, although a slight decrease of activity was observed for some derivatives
i.e. A3, B2, B3, C1 and C3) the results indicated that, independently of the chemical approach,
(
these structural modifications did not significantly affect the inhibitory potency respect to the
reference compound.
Evaluation of Biological Activities
10

Three new heparanase inhibitors, the biotinylated derivative A7, B3 and C3, representative of
each type of derivatization and showing a similar degree of substitution, were further
characterized in vitro by biological assays, with respect to the reference compound
(Roneparstat), in order to evaluate their activity against selected human cancer cell lines.
Proliferation Assay. The selected compounds were tested in vitro by a cell proliferation assay
to assess their effect on the growth of three human tumor cell lines expressing different levels of
heparanase: HT1080 (human fibrosarcoma), U87MG (human glioblastoma astrocytoma) and
2
3
U2OS (human osteosarcoma)
2
4,25
.
Cells were treated for 72h with increasing concentrations (up to 100 µM) of each
heparanase inhibitor, including Roneparstat. Results, expressed as percentage of proliferation
with respect to vehicle-treated (control) cells, are summarized in Table 2.
Table 2
Overall, Ronepastat proved to be moderately active in inhibiting proliferation of two cell lines
(U87MG and U2OS) out of the three and only at the highest concentration (100 µM) tested, and
to be marginally or no active at lower concentrations. It is likely that this effect could be
unrelated to heparanase inhibition due to the very high amount of compound needed to affect cell
proliferation. Regarding the derivatives, only a very marginal antiproliferative activity was
observed mainly at the highest concentration, indicating that the introduced substituents did not
increased toxicity of compounds.
11

Invasion Assay. According to the pivotal role of heparanase in promoting cancer cell invasion
and metastasis, we sought to assess the ability of derivatives to affect the invasive potential of
HT1080, U87MG and U2OS human cell lines. To this aim, we used the Matrigel cell invasion
assay. All the compounds were assessed at not toxic concentrations, below the relative IC of
1
0
the anti-proliferative values measured for each cell line. Data are expressed in Table 3 as the
percent invasion with respect to cell invasion in absence of compounds. Although in this assay
Roneparstat, at least at the dose (10 µM) tested, failed to affect the invasive potential of the three
cell lines, surprisingly, derivative B3 significantly blocked invasion of HT1080 cells (> 50%
inhibition) and, at lesser extent (< 50% inhibition), also of U87MG cells. In contrast, the other
two derivatives behaved as the reference compound and were ineffective in inhibiting cell
invasion.
Table 3
2
6
Cell Adhesion Assay. Cell surface heparan sulfate proteoglycans (HSPGs) appear to be
involved in cell-extracellular matrix (ECM) adhesive interactions. Thus, to evaluate whether the
new derivatives could interfere with adhesion properties of cancer cells, compounds were tested
at not toxic concentrations on U87MG glioblastoma cells, by means of a classical cell adhesion
assay, and the results (expressed as percent cell adhesion with respect to vehicle-treated cells) are
presented in Table 4. All the compounds markedly inhibited adhesion of U87MG cells with
potency similar to that of Roneparstat. Only compound C3 showed a minor potency when tested
at the lowest dose.
12

Table 4
Expression of proangiogenic factors. Translocation of fraction of active heparanase into the cell
nucleus and its involvement in controlling transcription of genes encoding for pro-angiogenic
2
7,28,29
and pro-invasion genes, upon degradation of the nuclear heparan sulfate, has been reported.
Therefore, we wondered whether the new heparanase inhibitors could also affect transcription of
genes encoding for proangiogenic factors in tumor cell lines. To this aim, HT1080 human
fibrosarcoma cells were treated for 24 h with compound B3 compared to the reference compound
Roneparstat, both at 10 µM. Then, the mRNA levels related to genes encoding for proangiogenic
factors, such as FGF1/2, VEGF, MMP-9, and for HPSE-1 were measured by a quantitative Real-
Time PCR. As shown in Figure 7, the biotinylated derivative B3 inhibited the expression of these
genes to an extent comparable to Roneparstat.
Figure 7
3
0
Experimental metastasis model with B16-F10 mouse melanoma cells. Roneparstat and the three
biotinylated derivatives (A7, B3 and C3) were tested for their antimetastatic activity against
murine B16-F10 melanoma cells. Mice were pre-treated with each compound (50 µg/mouse),
administered 1h prior to the vein tail injection of the melanoma cells. The number of melanoma
13

colonies per lung (lung metastases) was determined after 18 days. Results are shown in Figure 8,
and highlight a profound anti-metastatic activity of all derivatives, with potency comparable to
that of the reference compound. Compound A7 appears somewhat better than Roneparstat
(p=0.06, compound A7 vs Roneparstat), but at the concentration used (50 µg/mouse) the
inhibition was >80% with all compounds making it difficult to assess minor differences.
Figure 8
Effect on growth of CAG myeloma. The heparanase inhibitor Roneparstat was previously shown
to strongly inhibit the growth of myeloma tumors in vivo (subcutaneous tumors formed by CAG
human myeloma cells) in part by targeting the tumor microenvironment, including a significant
3
1
inhibition of tumor angiogenesis. According to this data, we evaluated the growth inhibition of
CAG myeloma cells, injected subcutaneously into NOD/SCID mice, upon treatment with three
biotinylated compounds (C3; B3 and A7), using Roneparstat as reference. Results shown in
Figure 9 indicate that all the biotinylated-derivatives inhibited CAG myeloma-tumor growth with
potency similar to that of Roneparstat. The inhibitory effect of compound A-7 reached statistical
significance (p=0.05).
Figure 9
14

CONCLUSIONS
Thirteen heparin derivatives, analogues of the well-studied heparanase inhibitor Roneparstat
were prepared. All were conjugate with a non-carbohydrate molecule containing a biotin moiety
and having different length and hydrophilicity. Independent of the nature and number of the
heparin substituent (from 0.4 to 11 for chain) the anti-heparanase activity was conserved (IC in
5
0
anomolar range (1.4 to 6.9 nM), similar to that of Roneparstat (Table 1).
The synthetic approaches preserved the glycol split (gs) residues able to interact with the active
site of heparanase, as well as the sulfation of position 2 of iduronic acid and of position 6 of
3
2
glucosamine residues, which are involved in enzyme recognition . Moreover, it appears that, in
this range of substitution, the substituents do not mask the negative charges able to interact with
the heparan sulfate recognition sites of the protein.
In particular, the synthesis route A does not influence the degree of polymerization of the
compound, leaving the inhibition capacity almost unchanged with respect to the Roneparstat,
regardless of the substituent structure and the degree of substitution. On the contrary, synthesis
route B shows a slight decrease in heparanase activity (up to 6.9 nM) which highlights the
importance of the type of substitute of the amino function of glucosamine, which can be assumed
to slightly interfere with binding to the protein. But it should be noted that the samples of the B
series have a lower molecular weight than Roneparstat and of the samples of the other two series,
this means that if the inhibition activity is expressed in weight and not in moles this observed
difference disappears. On the other hand, the derivative B3 turned out to be able to inhibit the
transcription of genes encoding for proangiogenic factors (MMP-9, VEGF, FGFs, HPSE) in
tumor cells, likely due to inhibition of nuclear heparanase HS degrading activity.
15

Three derivatives (A7, B3, C3; SST0761NA1, SST0635NA1, SST0764NA1, Lab code
respectively) with a nearly one substituent per chain and different chemical nature and length of
the spacer were selected for further biological characterization. Briefly, A7 (SST0761NA1) is
substituted on oxidized glucosamine residues, the linker between N-acetyl gs-heparin and biotin
is a polypeptide chain made of a total of 33 atoms. B3 (SST0635NA1) is substituted on the
amino group of glucosamine residue with a polyethylene glycol spacer made of a total of 17
atoms. C3 (SST0764NA1) is substituted only on the terminal reducing end residue with a glycine
as spacer. These selected substituents did not induce significant effect on cell-proliferation as
demonstrate with three human tumor cell lines. Notably, Roneparstat and its derivatives were
endowed with significant anti-adhesion properties as clearly shown applying an in vitro adhesion
assay and human U87MG glioblastoma cells. This effect may be attributed to an adjuvant effect
of the PEG chain.
Importantly, the newly synthesized compounds retained the ability to attenuate the growth of
CAG myeloma tumors in mice with potency similar, or even higher (in the case of derivative B3)
than that of the reference compound Roneparstat. Furthermore, compound B3 inhibited cell
invasion through a reconstituted basement membrane (Matrigel) in vitro and all three derivatives,
as well as Roneparstat, profoundly inhibited metastatic dissemination (lung colonization) of
murine B16-F10 melanoma cells in vivo.
It should be noted that the in vivo studies were performed only to gain initial information on
the anti-tumor activity of the newly synthesized biotinylated derivatives. In-depth analysis of
PK/PD properties will be the matter of subsequent studies, limited to the compound(s) that will
be selected for further analyses.
16

In conclusion we have demonstrated that is possible to introduce even complex molecules
without impair the anti-heparanase activity of the N-acetylated glycol-split heparin (Roneparstat)
using sequence of reactions that maintain the structural feature needed for heparanase
recognition. Only in-depth in vivo studies, which are not the subject of this study, will highlight
differences due to the greater or lesser bioavailability of these samples. The present study
supports the possibility of introducing biotin moiety in the substituent to allow the monitoring of
pharmacokinetic and pharmacodynamic properties of Roneparstat, and to target specific folate
receptors that may improve drug delivery into the tumor but, using the same approach the biotin
moiety could be replaced by other active molecules to yield a synergistic inhibition of tumor
growth.
EXPERIMENTAL SECTION
Materials:
Chemicals, reagent-grade obtained from Sigma Aldrich SRL (Milan, Italy) and Iris Biotech
(
GmbH, Germany), were used as received. Porcine heparin with an average Mw of 17 kDa was a
preparation of porcine mucosal heparin sodium salt from Laboratori Derivati Organici,
L.D.O.),Trino Vercellese, Italy. N-desulfated, N-acetylated glycol split (gs) heparin, also named
(
as Roneparstat taken as reference compound, was from Leadiant Bioscience S.A.; former Sigma-
Tau Research, Switzerland.
NMR spectra were measured at 500 and/or 600 MHz on Bruker Avance spectrometers
2 2
equipped with a 5-mm TCI cryoprobe at 298 K from D O solutions (15 -25 mg/0.5 mL of D O,
17

1
9
9.99% D). H NMR spectra were recorded with 32 scans and were Fourier transformed after
exponential multiplication (line broadening of 0.3 Hz) and zero filling (from 32 k to 128 k
points), phased, baseline corrected and calibrated on the 3-(Trimethylsilyl)propionic-2,2,3,3-d4
acid sodium salt (TSP) signal. The solvent residual peaks of MeOD d or DMSO d were used as
4
6
1
3
internal standards, at 3.31 (298K) ppm and 2.48 ppm (298K) respectively. C NMR spectra
were recorded with a 10-mm multi-nuclear broad band probe, using a Bruker pulse program
zgig. The spectra were acquired with relaxation delay 1 s, acquisition time 0.327 s, spectral
width 299 ppm. at 303 K and the applied transient number was 20400. The transmitter frequency
off-set was 90 ppm. The spectrum size used was128k with zero filling. Before Fourier
transformation line broadening of 4 Hz was applied to decrease noise. To recorde HSQC-NMR
spectra, heparin samples (20 mg) and its derivatives were dissolved in 0.6 mL of D O.
2
Acquisition of the spectra was based on 24 scans, 32 dummy scans, 2s pulse delay, at 303 K (150
Hz1JC-H). Integration of peak volumes in the HSQC spectra was performed using standard
Bruker TOPSPIN 3.0 software. Heparin signal assignment and the relative content of
monosaccharide residues were calculated from the corresponding anomeric cross-peaks
3
3
identified by HSQC according to published methods.
Gel permeation chromatography (GPC) measurements, by high pressure size exclusion
3
4
chromatography coupled with triple array detector (HP-SEC-TDA), were employed to obtain
the molecular weight parameters Mw, Mn, and polydispersion degree. A Viscotek (Houston,
Texas) instrument equipped with a triple detector array302 with refraction index (RI),
viscometer, and light-scattering (90°and 7°) detector was used. SEC was performed on G2500
and G3000 (7.8 mm x 30 cm TSK GMPWxl) Tosoh columns with 0.1 M NaNO as eluent at a
3
18

flow-rate 0.6 mL/min, by injecting 100 µL of 5 mg/mL solution of each samples. Peak
integration and data processing were performed using OMNISEC 4.1 software.
Desalting of biotinylated derivatives and intermediates was carried out by dialysis against H
2
O
through 3,5 kDa cut/off tubes (Spectrum) or by purification on a 2.5 x 100 cm Sephadex G10
column, with 10 % EtOH in H O as eluent and UV detection at 210 nm. Microwave-assisted
2
reactions were performed on CEM Corporation Discover instrument, and mass spectra on Bruker
Autoflex Maldi-Tof or Micro Q-Tof instruments. Flash chromatography was carried out on silica
gel (Merck 230-400 mesh) and monitored by TLC silica gel 60 F254 plates.
Synthesis of heparin intermediates
N-desulfated heparin: (I-1). Heparin (8.01g) was dissolved in 60 mL of H O and converted in
2
+
the acidic form with Amberlite IR 120 (H form). An excess of pyridine was added to the
aqueous solution then concentrated under reduced pressure. The resulting pyridinium salt was
dissolved in 80 mL of a mixture of DMSO /H O (95:5) and stirred at rt overnight, to obtain an
2
almost complete N-desulfation. Then, the solution was diluted with an equal volume of H O and
2
dialyzed for 72 h. Subsequently, product was isolated by concentration under reduced pressure
and freeze drying. The structure of the product, used as such in the next step, was confirmed by
1
3
C-NMR and SEC-TDA (15400 Da) analysis. (Fig.S5, Supporting Information)
N-acetylated heparin: (I-2). N-desulfated heparin I-1 (2.5 g), dissolved in 57 mL of NaHCO
3
2
saturated solution was cooled to 4°C, then under vigorous stirring Ac O, (2.1 mL) was added and
stirring was continued at 4°C for 30 min maintaining pH at about 8 by adding NaHCO . Finally,
3
other 2.1 mL of Ac O were added and the mixture was maintained at 4°C for another 1.5 h, then,
2
after dilution with an equal volume of H O was dialyzed for 72 h. The product was isolated by
2
19

concentration under reduced pressure and freeze dried. The structure of the final product, (2.1 g;
1
3
yield=84%) used as such in the next step, was confirmed by C-NMR and SEC-TDA (Mw=
1
5.400 Da) analysis (Fig.S5, Supporting Information)
N-acetylated-gs-heparin-Aldehyde (I-3). N-acetylated heparin I-2 (1.15 g), dissolved in 33.5
mL of H O and cooled to 4°C, was added at an equal volume of a 0.2 M NaIO solution and kept
2
4
under magnetic stirring in the dark for 20 h. The excess of periodate was quenched by adding
ethylene glycol (3 mL) and after 1 hr the reaction mixture was desalted by dialysis at 4°C in the
dark for 16 hrs. The retentate was recovered, gently concentrated to half the initial volume and
lyophilized to give the final intermediate I-3 (1.04 g, yield=90%) to be used as such in the next
step.
N-desulfated-gs heparin (I-4). N-desulfated heparin I-1 (1.2 g), in 35 mL of H O and cooled to
2
4
°C, was added of an equal volume of aqueous 0.4 M NaIO (33.5 mL). The pH value was
4
adjusted to 3.5 with HCl (2% v/v, about 0.5 mL) and kept under stirring at rt in the dark for 6 h.
Then, the oxidized heparin was precipitated by addition of 160 mL of EtOH (96% v/v)
3
containing 3% (w/v) of CH COONa. The precipitate was collected by centrifugation (12.000
rpm) and washed several times with EtOH to remove excess of reagents and sodium acetate. To
crude residue dissolved in 33 mL of H O, NaBH (731 mg) was added in several portions, and
2
4
the mixture was stirred for 3 hr at rt. Next, under stirring at 4 °C the pH value was adjusted to 4
with 5% (v/v) AcOH solution to quench the NaBH in excess. After 10 min stirring, the mixture
4
was neutralized with 0.1 N NaOH. After dialysis in membrane (cut-off 3500 Da at rt),
concentration under reduced pressure and freeze drying the final product was obtained (1.05 g
20

1
yield=87%). Content of gs-residues was confirmed by H, HSQC NMR spectra and molecular
weight evaluated by SEC-TDA (Mw = 7500 Da).
Reductive amination with alkyne amine of N-acetyl heparin reducing-end-units (I-5). N
acetylated heparin I-2 (1.21 g, 80 µmol), solubilized in 15 mL of phosphate buffer (15 mM
pH=5.5) under stirring at rt was added of propargylamine x HCl (266 mg, 2.9 mmol, 36 equiv.)
and reaction mixture was heated up to 40°C and kept under stirring for 3 h. Then NaBH CN (130
3
mg, 2.09 mmol) was added and the mixture stirred at 40°C for additional 24 h. A further 20
equiv. 130 mg NaBH CN was added and the mixture heated for another day. The reaction
3
mixture was diluted with H O and dialyzed against distilled H O at rt for 2 days. The retentate
2
2
concentrated and freeze dried, from aqueous solution it was purified by alcoholic precipitation
followed by SEC on G-10 Sephadex using H O/ethanol (9:1). After freeze-drying, 1.15 g of
2
alkyne derivative was obtained with 95% (w/w) yield. The percentage of amino group attached
1
to the reducing end was confirmed by H-NMR and molecular weight evaluated by SEC-TDA
(Mw = 15.300 Da) analysis.
N-acetyl-gs-heparin reducing end-alkylamino-functionalized (I-6). Alkylamino-N acetylated
heparin- I-5 (1.21 g), dissolved in 32 mL of H O and cooled to 4°C, was added of an equal
2
volume of aqueous 0.2 M NaIO , and kept under stirring in the dark for 20 h. The excess of
4
periodate was quenched by adding ethylene glycol (3.4 mL) and after 1 hr the reaction mixture
was desalted by dialysis at 4°C for 16 hrs. The retentate was treated with NaBH (740 mg),
4
stirred for 3 hr at rt then, cooled at 4 °C adjusting the pH value to 4 with 5% (v/v) AcOH for
4
quenching the NaBH in excess. After 10 min stirring, Then the solution was neutralized with 0.1
21

N NaOH and dialysed at rt, concentrated and freeze dried. The crude gs-derivative was desalted
on a Sephadex G-10 column (900 x 25mm) using H O /EtOH 9:1 (v/v) as a mobile phase at 2.1
2
mL/min. and monitoring the absorbance at 210 and 190 nm. After freeze-drying, 1.01 g of I-6
1
was obtained with a 85% (w/w) yield. Content of gs-residues was confirmed by H, HSQC NMR
spectra and molecular weight evaluated by SEC-TDA (Mw = 14.700 Da) analysis.
Biotinylation of N-acetyl heparin glycol split units
General preparation (herein described for compound A7). N-acetyl-gs-heparin intermediate I-
3
(0.95g, ca. 0.8 mmoli of aldehyde function) was dissolved in 15 mL phosphate buffer (15 mM,
pH=5.5) under stirring at rt. Biotinylated linker 3, shown in Fig 3, (210 mg, 0.2 mmol, about
.25 equiv. for each aldehyde function present on the polysaccharide chain) was solubilized in
0
phosphate buffer (1 mL) and added to I-3 solution that was heated up to 40°C and kept under
stirring for 1 h, then treated with NaBH CN (502 mg, 10 equiv.) the mixture was maintained at
3
4
0°C for overnight. After cooling to rt 10 equiv. of NaBH (296 mg) were added and the reaction
4
mixture was maintained at rt for other 3 h. After cooling at 4°C the pH value was adjusted to 4
with 5% (v/v) AcOH for quenching the NaBH excess, and after 10 min stirring, neutralized with
.1 N NaOH. After dilution with 10 mL of NaCl 2M, dialysis against distilled H O at rt for 1
4
0
2
day, the product was isolated by concentration and freeze drying to obtain the biotinylated-gs-
heparin A7, which was purified by gel permeation chromatography (Shephadex G-10) to give the
1
final product (about 810 mg, yield=84%). It was characterized by NMR ( H, HSQC) and SEC-
1
TDA analysis. Molecular weight and degree of biotinylation was evaluated by H-NMR (Fig.
S13 Supporting Information) as described in the characterization section and shown in the table
22

1
. Other derivatives (A1-6) with different biotinylation contents (DS) were prepared by same
procedure. The type of linkers, and molar ratio used to obtain each biotinylate compound are
given in table 1. (Fig. S14 and S15 Supporting Information.)
Biotinylation of the glucosamine free amino group
General procedure for N-desulfated-N-acyl-biotinylated gs-heparin (herein described for
compound B-3). N-desulfated-gs-heparin I-4 (1.1g, ca 1.8 mmol) dissolved in 20 mL of 0.1N
3 4
NaHCO and cooled to 4°C was treated with Biotin-Peg -NHS 7 (320 mg 0.9 mmoli, about 0.5
equivalents for each glucosamine free amine function) in 1 mL of DMF and the mixture was
stirred at rt for 2 h. Then further 0.2 equiv. of biotinylating agent 7 (ca 128 mg in 0.5 ml of
DMF) was added and stirred for additional 12 h. The reaction medium was diluted with 10 mL of
NaCl 2 M and dialyzed at rt for 1 day. The retentate was concentrated up to ca 15 mL and
purified by alcoholic precipitation by addition of 30 mL of EtOH (96% v/v). The precipitate was
collected by centrifugation (12.000 rpm) and washed several times with EtOH then redissolved
in 15 mL of H
2
O, and the alcoholic precipitation process repeated twice. The final biotin-gs-
1
conjugated compound I-7 was characterized by NMR ( H, HSQC) and SEC-TDA analysis and
used as such in the next step. Another biotinylated intermediate compound bearing the biotin-
linker 6 was prepared by the same procedure.
N-Acetylation of N-Biotinylated-gs-N-desulfated heparin (B-3). N-desulfated-gs-N-
biotinylated heparin I-7 (0.85 g, 1.4 mmoli), dissolved in 23 mL of saturated solution NaHCO
3
23

was cooled to 4°C then, under vigorously stirring, added Ac
°C for 30 min. During the reaction, pH was controlled and maintained at ca 8 by adding
NaHCO . After a further addition of Ac O (1.1 mL) the mixture was maintained at 4°C for
2
O, (1.1mL, 11 mmol) and stirred at
4
3
2
another 1.5 h, then, diluted with an equal volume of H O and dialyzed for 2 days. Concentration
2
and freeze drying gave the crude N-biotinylated-gs-heparin derivative, which was purified by gel
permeation chromatography (Sephadex G-10). The final product B-3 (720 mg, yield=84%) was
1
characterized by NMR ( H, HSQC, Fig. S18 Supporting Information) and SEC-TDA
(MW=9400 Da) analysis. The degree of biotinylation (DS= 6.7 for disaccharide unit) was
1
evaluated by H-NMR. Biotinylated compound (B-1), bearing the biotin spacer 6 was prepared
by the same procedure described for B-3, using 1 equivalent of linker 7 instead of 0.5 with an
overall yield of 87% showing a DS of 15.3 for disaccharide unit, while biotinylated compound
(B-2), DS of 6.6.4 for disaccharide unit, was prepared using 0.5 equiv. of linker 7 instead of
linker 6 with an overall yield of 89°%.
Reducing end unit conjugation by “click” chemistry approach (compounds C1, C2 and
C3)
N-Acetylated-gs- heparin- reducing end triazole biotinylated (C3). N-acetyl-gs-heparin-alkyne
amine derivative I-6 (1.01g, 0.067 mmol) in a H O/t-BuOH (12 mL, 70:30 v/v %) under
2
magnetic stirring for 10 min was treated dropwise with biotinylated linker azide derivative 5 (54
mg, 0.13 mmol, about 2 equivalents for each alkyne function present on the polysaccharide
chain) solubilized in a H
equiv.) and CuSO 5 H O (300 µl, aqueous 0.1M, 0.5 equiv.) were added, and the reaction
mixture was stirred at rt for 2 h. Finally, other 2 equiv. of azide, 1 equiv. of sodium ascorbate and
.5 equiv. of CuSO 5 H O were added, and the mixture maintained at rt for further 4 h, then
2
O/t-BuOH (1 mL, v/v 1:1). Sodium ascorbate (300 µl, aqueous 0.2M, 1
4
2
0
4
2
24

dialyzed for 48 h. The retentate concentrated to 15-20 mL and freed from Cu using 3 mL of
Chelex 100 resin, then neutralized with 0.1 N NaHCO . After dialysis, concentration under
3
reduced pressure and freeze drying compound C3 was obtained (ca 920 mg, yield=87%). Its
1
structure was confirmed by NMR ( H, HSQC, Fig. S19 Supporting Information) DS=1.9 % for
disaccharidic units and SEC-TDA analysis (MW = 14200 Da).
Biotinylated compound C2 was prepared by the same procedure with an overall yield of 86%
1
and characterized by H-HSQC showing a DS of 2.1% for disaccharide unit and by SEC-TDA
analysis (MW= 9600 Da).
Biotinylated compound C1, bearing the biotinylated linker azide derivative 4 was prepared by
the similar procedures described for C3, using 3 equivalents of biotinylating agent 5 instead of 2
with an overall yield of 84% and a DS of 3.4% for disaccharide unit.
Anti-heparanase activity. In order to test the activity of the heparanase inhibitors, a
homogenous assay based on the cleavage of the synthetic heparin pentasaccharide Fondaparinux
(Arixtra; Aspen) has been used. The assay, which yields a disaccharide that is colorimetrically
2
1
determinated (560 nm), was essentially performed as described by Hammond . Serial dilutions
of the compounds and of the reference Roneparstat were tested in triplicate with standard
deviations (SDs) less than 10%. IC values for each compound, versus heparanase, were
5
0
ultimately calculated by GraphPad software.
Cell lines and maintenance. HT1080 (fibrosarcoma), U87MG (glioblastoma) and U2OS
(osteosarcoma) human cell lines and B16-F10 mouse melanoma cells were purchased from the
American Type Culture Collection (ATCC; Manassas, VA) and cultured according to the ATCC
recommendations. Briefly, HT1080, U87MG and mouse B16-F10 cells were grown in
25

Dulbecco's Modified Eagle's Medium (DMEM; Thermo Fisher Scientific). Luciferase-labelled
CAG human myeloma cells (kindly provided by Dr. R. Sanderson, University of Alabama at
Birmingham) were cultured in RPMI 1640 medium, and U2OS cells were grown in McCoy's 5a
medium (Thermo Fisher Scientific). All media were supplemented with 10% FCS (Thermo
Fisher Scientific), 100 U/mL penicillin and 100 µg/mL streptomycin (Thermo Fisher Scientific),
and 2 mM L-Glutamine (Thermo Fisher Scientific). All cell lines were maintained at 37°C in a
2
humidified 5% CO atmosphere.
Proliferation Assay. HT1080, U87MG and U2OS exponentially growing cells were seeded
into 96-well plates and then, 24h later, treated with test compounds at different concentrations or
the solvent alone. Inhibition of cell proliferation was measured upon 3 days of treatment by
means of a classical sulforhodamine B (SRB) assay performed in triplicate. Results were
expressed as percentage of tumor cell proliferation as compared to vehicle-treated (control) cells.
Matrigel invasion assay. Cells were pretreated with test compounds at the indicated
concentrations, in complete medium for 24h. Then, cells were detached, resuspended in serum-
4
free medium and added (2.5 x10 cells/filter) to the upper chamber of 24-well Transwell plates
(
Costar, Corning Inc., Corning, NY) previously coated with Growth Factor Reduced Matrigel
BD Biosciences, San Jose, CA). The same compound concentration used for cell pretreatment
(
was added to both the upper and lower chambers. Cells were allowed to invade for 24 h at 37°C
in a CO incubator and then not-migrated cells were removed from the upper chamber with a
2
cotton swab. The invading cells were fixed, stained with SRB, and then counted under a
microscope, at 40X magnification. Activity of test compounds was compared to the reference
compound.
26

Cell Adhesion Assay. U87MG (human glioblastoma astrocytoma) cells were treated for 18-20h
with test compounds (B3, A7 and C3) in complete medium supplemented with 10% FCS. Then,
cells were collected, washed in complete medium, and added to matrigel coated wells in
triplicate, at previously identified optimal seeding density, and incubated at 37 °C in complete
2
5
medium for times ranging from 15 minutes to 4h . After incubation, the wells were washed
three times with serum-free medium and the remaining firmly attached cells were stained with
crystal violet 0.1 reagent. Optical density was measure at 500 nm by Victor 3 (Perkin-Elmer)
plate analyzer. Each experiment was repeated three times with standard deviations (SDs) less
than 10%.
6
CAG myeloma Assay. Luciferase-labelled CAG myeloma cells (5x10 ) were injected
subcutaneously into NOD/SCID mice (n=5). 8-weeks old female mice were treated with
biotinilated compounds C3, B3 and A7 (1000 µg/mouse, injected intraperitoneally twice a day,
3
5
starting on day 9 of cell inoculation) . Vehicle (PBS) alone was similarly injected as control
(CTL).
Tumor growth was examined by IVIS imaging prior to the first injection of compounds and
once a week thereafter. The experiment was terminated on day twenty-nine.
All the animals were handled as per protocols and procedures approved by the Technion
Institutional Animal Care and Use Committee.
Bioluminescent imaging of the luciferase-expressing tumors was performed with a highly
sensitive, cooled charge coupled device (CCD) camera mounted in a light-tight specimen box
(
IVIS; Xenogen Corp., Waltham, MA). Briefly, mice were injected intraperitoneally with D-
luciferin substrate at 150 mg/kg, anesthetized and placed onto a stage inside the light-tight
27

camera box, with continuous exposure to isoflurane (EZAnesthesia, Palmer, PA). Light emitted
from the bioluminescent cells was detected by the IVIS camera system with images quantified
4
7
for tumor burden using a log-scale color range set at 5×10 to 1×10 and measurement of total
photon counts per second (PPS) using Living Image software (Xenogen).
2
9, 36 37
Experimental metastasis. B16-F10 mouse melanoma cells
:
5
B16-F10 mouse melanoma (2x10 ) cells were injected into the tail vein of C57BL/6 6-weeks
old female mice without or together with the test compound (50 µg/mouse), essentially as
2
9,35,36
described
. In some experiments, compounds were also injected (50 µg/mouse;
intraperitoneally) to the mice 30 min prior to the tail vein injection of cells. Lungs were
harvested 18 days thereafter, fixed in Bouin solution and metastases occurrence was determined
by gross examination and counting of colonies using a stereomicroscope.
Real-Time Quantitative PCR (qPCR) Assay. Total RNA was extracted from HT1080 cells,
after 24 h treatment with test compounds (10 µM), and then retrotranscribed using the iScript™
Advanced cDNA Synthesis Kit for RT-qPCR (Bio-Rad Laboratories Inc., Hercules, CA)
according to the manufacturer’s instructions. Real Time quantitative PCR analysis was
performed using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad Laboratories Inc.,
Hercules, CA) and the following PrimePCR SYBR Green Assays: qHsaCID0015228 (HPSE-1),
qHsaCED0002206 (FGF-1), qHsaCID0015510 (FGF-2), qHsaCID0011597 (MMP-9),
qHsaCED0043454 (VEGF-A) (Bio-Rad Laboratories Inc., Hercules, CA). The 7900HT
28