1
23814, an Inhibitory Antibody of Ligand-Mediated Notch1 Activation, Modulates
Angiogenesis and Inhibits Tumor Growth without Gastrointestinal Toxicity
Theresa Proia1, Feng Jiang2, Alisa Bell, Richard Nicoletti, Lingxin Kong, Kelly Kreuter, Laura
Poling3, William M. Winston4, Meghan Flaherty5, Solly Weiler, Samantha Perino, Ronan
O’Hagan6, Jie Lin, Jeno Gyuris, Heidi Okamura
Authors’ Affiliation: AVEO Pharmaceuticals, Inc., 650 E. Kendall Street, Cambridge, MA 02142
Current addresses: 1AstraZeneca, Waltham, MA 02451; 2EMD Serono, Billerica, MA 01821;
3 Mersana Therapeutics, Cambridge, MA 02139; 4Potenza Therapeutics, Cambridge, MA
02138; 5Novartis Institutes for BioMedical Research, Cambridge, MA 01239; 6Merck Research
Laboratories, Boston, MA 02115
Running title: 23814 Notch1 antibody inhibits tumor growth without toxicity
Keywords: Notch1, angiogenesis inhibitor, antibody, tivozanib, toxicity
Corresponding Author: Heidi Okamura, AVEO Pharmaceuticals, Inc., 650 E. Kendall Street,
Cambridge, MA 02142; Phone: 617-299-5897; Fax: 617-995-4995;
E-mail: [email protected]
Conflict of interest: All authors were employees of AVEO Pharmaceuticals when the research
was conducted
Abstract 198
Word count 5371
Tables and figures 6
References 50
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
2
Abstract
Dysregulation of Notch signaling has been implicated in the development of many different
types of cancer. Notch inhibitors are being tested in the clinic, but in most cases gastrointestinal
and other toxicities have limited the dosage and, therefore, the effectiveness of these therapies.
Herein, we describe the generation of a monoclonal antibody against the ligand binding domain
of the Notch1 receptor that specifically blocks ligand-induced activation. This antibody, 23814,
recognizes both human and murine Notch1 with similar affinity, enabling examination of the
effects on both tumor and host tissue in preclinical models. 23814 blocked Notch1 function in
vivo, inhibited functional angiogenesis, and inhibited tumor growth without causing
gastrointestinal toxicity. The lack of toxicity allowed for combination of 23814 and the VEGFR
inhibitor tivozanib, resulting in significant growth inhibition of several VEGFR inhibitor-resistant
tumor models. Analysis of the gene expression profiles of an extensive collection of murine
breast tumors enabled the successful prediction of which tumors were most likely to respond to
the combination of 23814 and tivozanib. Therefore, the use of a specific Notch1 antibody that
does not induce significant toxicity may allow combination treatment with angiogenesis inhibitors
or other targeted agents to achieve enhanced therapeutic benefit.
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
3
Introduction
The Notch signaling pathway regulates many fundamental processes essential for
normal development such as the control of cell differentiation, survival, proliferation, and
angiogenesis [1-3]. Signaling pathways necessary for embryonic development are often found
reactivated during tumorigenesis, and inappropriate activation of Notch signaling has been
implicated in the development of many different types of adult and pediatric cancers [4, 5].
Interestingly, in some tissues, inactivation of Notch signaling may facilitate tumor development,
particularly in the case of squamous cell carcinomas [6-8] and bladder cancer [9]. This
highlights the complexities of Notch signaling, and the importance of understanding the correct
context in which Notch pathway inhibition will be beneficial.
In mammals, there are four Notch receptors (Notch1-4), all with similar functional
domains and structures [10]. The extracellular domains consist of a series of EGF-like repeats
that contain the ligand binding domain (LBD), followed by a negative regulatory region (NRR)
that locks the receptor in the “off” conformation in the absence of ligand [11]. There are five
canonical Notch ligands, including three Delta-like (DLL1, DLL3, DLL4) and two Jagged (Jag1,
Jag2) proteins, which are also membrane-bound. Receptor-ligand activation requires interaction
between neighboring cells, leading to conformational changes that result in sequential cleavage
of the receptor by ADAM and gamma secretase proteases, and release of the Notch
intracellular domain (ICD). The Notch ICD translocates to the nucleus where it complexes with
other transcription factors and coactivators to turn on downstream target genes [10].
Notch1 activation impacts tumorigenesis through multiple mechanisms. Activating
mutations of Notch1 occur in several different types of hematopoietic malignancies, including
>50% of T-cell acute lymphoblastic leukemia (T-ALL) [12] and ~10% of chronic lymphocytic
leukemia (CLL) [13, 14], where Notch1 signaling appears to promote resistance to apoptosis
[15]. Immunohistochemistry shows elevated levels of cleaved Notch1 ICD in several different
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
4
primary human tumor types, suggesting that Notch1 activation is also a common occurrence in
solid tumors [16, 17]. Cancer stem cell maintenance is another area where Notch signaling is
thought to play a key role [18, 19]. Inhibition of Notch1 in mouse models can reduce cancer
stem cell numbers in both hematopoietic and solid tumors, potentially decreasing resistance to
chemotherapy and delaying tumor recurrence [20-22]. Finally, Notch1 activity in the
microenvironment keeps VEGF signaling in check to optimize tumor angiogenesis. Inhibition of
Notch1-DLL4 signaling leads to uncontrolled vascular sprouting, and a significantly greater
number of tumor vessels [23-27]. These vessels, however, are not efficiently perfused [23, 28].
Notch1 inhibition has also been shown to decrease production of endothelial nitric oxide, a
vasodilating agent important for regulating blood flow [29]. Thus, despite increasing vascular
density, Notch1-DLL4 blockade results in impaired vascular function and tumor growth
inhibition, likely due to increased hypoxia.
The multifaceted manner in which Notch1 can facilitate tumor growth underscores the
tremendous potential of targeting this pathway. Gamma-secretase inhibitors (GSI), which inhibit
all four Notch receptors as well as numerous other substrates, have been investigated in the
clinic, where hints of anti-tumor activity have been observed. However, severe gastrointestinal
toxicity due to goblet cell metaplasia, or reduced GSI exposure due to CYP3A4-mediated
deactivation, has forced suboptimal dosing and schedule alterations, limiting their therapeutic
utility [30-33]. Inhibitory antibodies targeting Notch1 have been developed in an effort to limit
toxicity, but gastrointestinal and other toxicities have still been reported in preclinical and clinical
settings [27, 34, 35]. Nevertheless, signs of clinical response have been seen in patients
treated with Notch1 antibody [35], indicating that Notch1 inhibition can be efficacious. Recent
preclinical studies demonstrate that a soluble decoy encompassing a subset of the Notch1
EGF-like repeats can block Notch1 signaling without inducing gut toxicity, suggesting that
inhibition and toxicity can be uncoupled [36].
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
5
Here, we describe the in vitro and in vivo activity of 23814, a human antibody made
against the Notch1 LBD that specifically inhibits ligand-induced activation of the Notch1
receptor. The 23814 antibody possesses cross-reactivity towards human and murine Notch1
with similar affinity, allowing us to examine the effect on both tumor and host tissue. At effective
doses that allow inhibition of Notch1 signaling in both physiological and pathological contexts,
no evidence of gastrointestinal or other toxicity is observed. The lack of toxicity allowed
combination of 23814 with the pan-VEGFR inhibitor tivozanib to achieve significant tumor
growth inhibition in several models that do not respond robustly to tivozanib monotherapy.
These results suggest that treatment with Notch1-specific antibodies may be a viable
therapeutic option in cancer settings resistant to anti-angiogenic treatment.
Materials and Methods
Cell lines and reagents
Karpas 45 was acquired from the German Collection of Microorganisms and Cell
Cultures (DSMZ) cell bank. The HT-1080 cell line was acquired directly from the ATCC and
stored frozen at early passage until use. No additional authentication was performed. Stable
Notch ligand-expressing lines were established by transfecting CHO Flp-In cells (Life
Technologies) with full length Jagged1, Jagged2, DLL1, or DLL4 using Lipofectamine 2000 (Life
Technologies). Notch receptor antibodies for FACS analysis were purchased from BioLegend.
Tumor Xenograft Studies
All mice were treated in accordance with the OLAW Public Health Service Policy on
Human Care and Use of Laboratory Animals and the ILAR Guide for the Care and Use of
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
6
Laboratory Animals. All in vivo studies were conducted following the protocols approved by the
AVEO Pharmaceuticals Institutional Animal Care and Use Committee.
Efficacy studies were performed with 8-10 week old female NCR nude mice. Cells were
resuspended in HBSS containing 50% matrigel and inoculated subcutaneously in the flank of
each mouse. Primary human tumors from surgical resection were obtained through the
Cooperative Human Tissue Network (CHTN). For the Ki-168 patient-derived xenograft (PDX)
study, mice were inoculated with 5x106 cells obtained from a human primary renal clear cell
tumor propagated in vivo and stored frozen prior to use. When tumors approached ~200 mm3,
mice were randomized and received 23814 or control hIgG (Xolair, Novartis AG) at 20 mg/kg
three times weekly, ErbB3 antibody (AV-203) at 20 mg/kg twice weekly, or tivozanib (dissolved
in 0.5% carboxymethylcellulose (CMC)) at 5 mg/kg daily. Mice were inoculated with 3.5x106
HT-1080, or 1x105 BH cells and randomized when tumor size reached 150-200 mm3. Animals
received 23814 or tivozanib as described above, combination treatment with 23814 (20 mg/kg
three times weekly) plus tivozanib (5mg/kg daily), or vehicle control (20 mg/kg hIgG three times
weekly plus daily dosing with 0.5% CMC). Statistical analysis was performed using 1-way
ANOVA and two-tailed t-tests using Graphpad Prism software version 6.
CD31 Staining
Mice were cohorted when tumors approached 250 mm3, and treated with 20 mg/kg
23814 or hIgG antibody on days 1 and 3, and/or dosed daily with 5 mg/kg tivozanib for 3 days.
Tumors were collected on day 4, fixed, processed, and embedded in paraffin. Tumor sections
were immunostained with antibodies to CD31/PECAM1 to detect the vasculature. The average
number of vessels per tumor area (number/μm2) was determined using an Aperio ScanScope
XT (Aperio, Vista, CA) and Microvessel Analysis v1 Parameters Tool image analysis software
(Aperio).
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
7
Thymocyte Depletion
Four week old female mixed background mice (Taconic) were separated into groups of
five, and treated with 20 mg/kg hIgG or 23814 antibody by intraperitoneal injection. Mice were
dosed 2-3 times weekly for 18 days. Thymuses were harvested in 5%FBS/PBS, pushed
through a 100 μm filter, and spun. The cell pellet was resuspended in cold 1X RBC lysis buffer
(eBioscience) for 2 minutes before addition of 5%FBS/PBS. After spinning for 5 minutes at
1000 rpm, the cell pellet was resuspended in 0.5% BSA/PBS. Thymocytes were counted using
the Countess® Automated Cell Counter (Invitrogen).
Alcian Blue Staining
Mice were treated with 23814 or hIgG at 20 mg/kg 3x weekly, or with the GSI
dibenzazepine (DBZ) (Syncom) at 10 µmol/kg daily. All mice were weighed twice weekly. DBZ-
treated animals were taken down after four weeks due to deteriorating health. 23814 and hIgG
arms were treated for eight weeks. For goblet cell detection, sections of small intestine from all
treatment groups were harvested at end of study, fixed in 10% formalin, processed, and stained
using the Alcian Blue pH 2.5 Stain Kit (American Master Tech Scientific) according to the
manufacturer’s instructions.
Immunoassay for 23814 Detection
Serum samples and standards were diluted in fetal bovine serum, and 23814 captured
using recombinant Notch1Fc (R&D Systems) bound to 96-well High Bind plates (Meso Scale
Discovery). Captured 23814 was detected with a SULFO-TAG labeled anti-human F(ab’)2
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
8
(Jackson ImmunoResearch) using electrochemiluminescence with the Meso Scale Discovery
platform.
Luciferase reporter assays
Karpas 45 were infected with lentivirus containing the firefly luciferase gene under
control of the RBP-Jk transcriptional response element (SABiosciences), and selected with
puromycin. Assays were performed by preincubating reporter cells with antibody for 30
minutes, then co-culturing with ligand-expressing CHO Flp-In cells that had been seeded in 96
well plates. 24 hours later, cells were processed using the Bright Glo (Promega) reporter assay
protocol per the manufacturer’s instructions. Lysates were transferred to white–walled 96 well
plates (Greiner Bio-One) and read on a GloMax Luminometer (Promega).
ICD Cleavage and Western Blotting
Recombinant Jag1 mFc and Jag2 mFc were created by fusing cDNA encoding the
extracellular domain of human Jag1 or human Jag2, respectively, to the murine IgG1 Fc
fragment. Constructs were stably transfected into CHOK1SV cells and fusion protein was
purified from the supernatant. 96-well plates were coated with anti-His tag (R&D Sytems) or
anti-mFc (Jackson ImmunoResearch) to capture recombinant DLL4 (R&D Systems), Jag1 mFc,
Jag2 mFc, or mFc (Jackson ImmunoResearch). Karpas 45 were preincubated with 23814 or
hIgG for 30’ before plating onto ligand-coated wells. Cells were harvested after 4 hours, and
lysates analyzed by SDS PAGE and Western blot. Blots were probed with antibody specific for
cleaved Notch1 ICD (Cell Signaling) or β-tubulin (Cell Signaling). Bands were detected using a
ChemiDoc XRS+ imager and ImageLab software (BioRad).
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
9
Expression Analysis
RNA was prepared using Qiashredders and RNeasy miniprep columns (Qiagen), and
cDNA was prepared using High Capacity cDNA reverse transcription kit (Life Technologies).
qPCR was performed on Applied Biosystems HT7900 using TaqMan Gene Expression assays
(DTX1-Hs00269995_m1, HES4-Hs00368353_g1, and NRARP-Hs01104102_s1) and
manufacturer-suggested PCR protocols (Life Technologies). Results were analyzed by the
comparative Ct method (Delta Delta Ct) using β-actin as the loading control, as detailed in
Supplementary Materials and Methods.
Hierarchical Clustering
Unsupervised hierarchical clustering was conducted using Euclidian distance and the
complete linkage method on the BH expression data [37], selecting the top 1,000 most variant
probes as defined by median absolute deviation. The analysis was performed using R version
3.0.2.
Results
Identification and characterization of Notch1-specific antagonistic antibodies
To assess the therapeutic potential of targeting the Notch pathway in cancer, we used
phage display to generate monoclonal antibodies that specifically target EGF-like repeats 11-12,
the LBD of the Notch1 receptor. Candidate clones were screened by ELISA and FACS for
binding to human and murine Notch1, in an effort to identify antibodies that cross-react with the
LBD of both species. To identify antibodies able to inhibit ligand-induced activation, we utilized
a luciferase reporter under the control of an RBP-Jκ transcriptional response element, which
mediates Notch signaling. A stable reporter line was established using Karpas 45, a human T-
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
10
ALL cell line expressing high levels of endogenous Notch1 receptor, and used to screen for
blocking antibodies. This led to the identification of 23814. This antibody was able to block
activation induced by all ligands tested in a dose-dependent manner. Luciferase activity was
inhibited by 23814 when the reporter line was co-cultured with cells expressing Jag1, Jag2,
DLL1, or DLL4 (Fig. 1A). Western blots confirmed that addition of 23814 to Karpas 45 cells
robustly prevented ligand-induced Notch1 ICD cleavage, maintaining Notch1 ICD at background
levels even in the presence of high amounts of ligand (Fig. 1B). Consistent with effective
inhibition of Notch1 signaling, expression of downstream target genes was significantly
decreased in the presence of 23814 (Fig. 1C).
FACS analysis shows that 23814 binds to human and murine Notch1 expressed on the
cell surface, but not to Notch2 or Notch3 (Fig. 1D). Surface plasmon resonance also confirmed
high affinity binding of 23814 to the extracellular domains of human and murine Notch1 (KD =
2.3 nM and 3.7 nM, respectively), but not to those of other Notch family members. These
results demonstrate that 23814 is a neutralizing Notch1-specific antibody that recognizes both
human and murine Notch1 with high affinity.
Treatment of mice with 23814 results in sustained inhibition of Notch1 signaling in vivo
To assess the ability of 23814 to inhibit Notch1 function in vivo, we examined its effect
on thymocyte development, a well-characterized process dependent on Notch1 signaling.
Active Notch1 signaling is required for thymic progenitor cells to commit to the T cell lineage,
and for normal progression through T cell development. Inducible knockout of the Notch1 gene
in neonatal mice has been shown to result in a marked reduction in thymus size, as well as a
five-fold reduction in thymocyte number [38]. Treatment of mice with 23814 resulted in a
dramatic reduction in the number of thymocytes relative to isotype control-treated mice (Fig.
2A). The decrease in thymocyte number was dose dependent and consistent with effective
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
11
sustained inhibition of murine Notch1 activity by 23814. No weight loss or diarrhea was
observed at any of the doses tested (Fig. 2B).
23814 does not induce gut toxicity at efficacious doses
Several Notch inhibitors have entered the clinic, but have limited effectiveness due to the
development of goblet cell metaplasia, an overwhelming conversion of progenitor cells to
secretory goblet cells at the expense of absorptive enterocytes [30, 39]. The ability of 23814 to
recognize murine Notch1 allowed us to investigate its effect on gut toxicity at doses relevant to
the inhibition of Notch1 signaling. To confirm the lack of goblet cell conversion after 23814
treatment, Alcian blue staining was performed on the small intestines of mice treated for 8
weeks with 20 mg/kg of either hIgG or 23814, or for 4 weeks with dibenzazepine (DBZ), a
gamma-secretase inhibitor. In agreement with previously reported results, mice treated with
DBZ developed severe goblet cell metaplasia [30, 40] as indicated by increased Alcian blue
staining of mucin in intestinal crypts (Fig. 2C). In contrast, the small intestines of 23814-treated
mice were histologically similar to those of control mice and did not show increased goblet cell
numbers (Fig. 2C). Moreover, no signs of vascular neoplasm were observed, as has been
reported for some DLL4 and Notch1 inhibitors [34]. 23814 was confirmed to be present at high
concentration in the serum of treated animals at end of study (Fig. 2D). To further investigate
the potential toxicity of Notch1 inhibition, mice were treated twice weekly with 5, 20, or 40 mg/kg
23814 antibody for 4 weeks. Upon termination of the study, all major organs including liver,
heart, spleen, kidney, and thymus were examined for histology. Again, no signs of vascular
neoplasm or vascular proliferation were observed in any of the animals (Supplementary Fig.
S1A-C). Histopathology showed rare instances of single cell necrosis or mitotic figures in the
livers of some 23814-treated animals, but these were not dose-responsive and considered
recoverable. Serum collected at termination showed that clinical pathology analysis was well
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
12
within normal reference range for liver enzymes (Supplementary Table S1), and that 23814
exposure was maintained in a dose-dependent manner (Supplementary Fig. S1D). This
suggests that 23814 treatment for 4 weeks, at doses that achieve functional inhibition of Notch1
signaling in vivo, does not produce severe toxicities.
23814 inhibits tumor growth in kidney PDX model
Expression of DLL4 in endothelial cells is directly induced by VEGF signaling [41].
Therefore, Notch1 activation might be expected in settings where VEGF signaling is high.
Although VEGF is overexpressed in a number of different types of cancer, clear cell renal cell
carcinoma (RCC) is uniquely dependent on VEGF signaling due to the frequency of the von
Hippel-Lindau (VHL) tumor suppressor mutation [42]. To assess the functional activity of 23814
in a tumor model we utilized Ki-168, a PDX model established from a clear cell RCC tumor.
PDX models are thought to retain key characteristics of the original donor tumor and, therefore,
to possess superior potential to predict clinical outcome compared to traditional cell line based
models [43].
Ki-168 PDX tumors were treated with hIgG, 23814, AV-203 (a negative control antibody
against ErbB3), or tivozanib, a potent and selective inhibitor of VEGFR-1, -2 and -3 that has
demonstrated significant activity in the clinic [44]. Blockade of VEGFR signaling due to
tivozanib treatment effectively inhibited Ki-168 tumor growth relative to control. Similarly,
treatment with the 23814 Notch1 antibody resulted in potent tumor growth inhibition of Ki-168. In
contrast, inhibition of ErbB3 signaling with AV-203 had no effect in this model (Fig. 3). These
results show that monotherapy treatment with 23814 is able to inhibit growth of a patient-derived
RCC tumor, a type of tumor especially sensitive to inhibitors of angiogenesis.
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
13
Treatment with 23814 alters tumor vasculature and inhibits tumor growth in models that
do not respond well to VEGFR inhibitors
It has previously been demonstrated that inhibition of the Notch pathway through
blockade of DLL4 can negatively impact tumor growth in preclinical models that are resistant to
anti-VEGF therapy [23, 24, 26, 28, 45]. To see if specific inhibition of the Notch1 receptor can
achieve similar results, we utilized the highly vascular HT-1080 xenograft model, which has
been widely used to study tumor angiogenesis. HT-1080 tumors were treated with 23814, either
alone or in combination with tivozanib. Treatment with either 23814 or tivozanib monotherapy
resulted in partial tumor inhibition relative to control. Combination treatment using 23814 and
tivozanib together resulted in significantly greater efficacy than either drug alone (p<0.001)
(Fig.4A).
To determine whether 23814 was affecting tumor angiogenesis, immunohistochemistry
was performed on HT-1080 tumors. Established tumors that had been treated for three days
with either 23814, tivozanib, or a combination of the two were collected and stained with anti-
CD31 to examine vessel density. Quantification of CD31 staining demonstrates that treatment
with 23814 markedly increased tumor vascular density (Fig. 4B), consistent with specific
inhibition of Notch1-DLL4 signaling [23-25, 27]. As expected, inhibition of VEGF signaling with
tivozanib resulted in greatly decreased tumor vasculature. Combination treatment with 23814
and tivozanib resulted in reduced tumor vessel density relative to vehicle control, similar to that
observed with tivozanib alone (Fig. 4B).
To determine whether 23814 could inhibit tumor growth in other VEGF inhibitor-resistant
models, we examined a collection of breast cancer tumors derived from chimeric mice
engineered to inducibly express mutated HER2 (V659E) [37, 46]. More than 100 individual
primary tumors from these breast Her2 (BH) mice have been propagated, characterized, and
shown to have diverse characteristics across the collection by histopathology, microarray
analysis, and variation in response to drug treatment due to their chimeric nature [37].
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
14
Accordingly, tumors from BH models differ widely in their response to tivozanib monotherapy,
ranging from complete regression to varying degrees of resistance (Fig. 5A, Supplementary Fig.
S2, and data not shown). Several different BH tumor models that did not robustly respond to
tivozanib monotherapy were chosen to examine whether increased efficacy could be achieved
through blockade of Notch1 signaling using 23814. While most of the models tested had a
moderate response to either 23814 or tivozanib alone, all but one model exhibited significantly
increased tumor inhibition when treated with a combination of tivozanib and 23814 (Fig. 5A,
Supplementary Fig. S2, and data not shown). To confirm that 23814 and tivozanib were
effectively inhibiting the Notch and VEGFR signaling pathways, respectively, expression of
downstream genes was examined using TaqMan assays. Analysis of the treated tumors
demonstrated that 23814 inhibited Notch1 signaling, as canonical Notch targets such as Hey1
and Hes1 were downregulated (Fig. 5B). Tivozanib treatment decreased Flt4 (VEGFR3)
expression as expected, but 23814 significantly increased Flt4 expression (Fig. 5B), in
accordance with findings suggesting that increased VEGFR3 signaling is responsible for the
deregulated angiogenesis observed upon Notch1 inhibition [47]. Hypoxia-regulated gene
expression also increased upon treatment with 23814, tivozanib, and combination therapy,
consistent with inhibition of angiogenesis (Fig. 5C). CD31 staining demonstrated a marked
increase in tumor vessels relative to vehicle control in 23814-treated tumors consistent with
Notch1 blockade, while tivozanib and combination treatment greatly decreased vessel number
(Fig. 5D). Interestingly, significant increases in vessel sprouting were also observed in some
BH tumors treated with 23814 monotherapy whose growth was not robustly inhibited
(Supplementary Fig. S2), suggesting that aberrant vessel density is a marker of Notch1
inhibition, but is not predictive of potent tumor growth inhibition. Although addition of 23814 to
tivozanib treatment did not decrease the number of tumor vessels relative to tivozanib alone
(Fig. 5D), the combination of the two inhibitors did significantly increase the amount of tumor
necrosis observed in BH224 (Supplementary Fig. S3).
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
15
Hierarchical clustering predicts response to 23814 and tivozanib combination treatment
As not all BH tumors are equally inhibited by 23814 and tivozanib, it would be desirable
to be able to predict which tumors would best respond to combination treatment. Since BH224
exhibited robust tumor inhibition in response to combination therapy (Fig. 5A),
unsupervised, hierarchical clustering was used to determine which BH tumor models were most
biologically similar to BH224. BH224 clustered in a distinct subgroup that did not include any of
the other BH tumors tested in initial studies with 23814 (Fig. 6A). To see if tumors in this
subgroup share a dependence on Notch1/VEGF signaling, two of the closest neighbors to
BH224, BH270 and BH228, were assessed for response to combination treatment with 23814
and tivozanib. Both BH270 and BH228 tumor models responded strongly to combination
therapy, and exhibited greater sustained inhibition with 23814/tivozanib treatment than
previously tested BH models that clustered in separate subgroups (Fig. 6B). These results
suggest that there is an underlying biology in the tumors clustering within the BH224 subgroup
that predisposes them to dependence on Notch1/VEGF signaling, making them especially
susceptible to inhibitors of these pathways.
Discussion
Here we report the identification of a Notch1-specific antibody that exhibits anti-tumor
efficacy and modulates tumor angiogenesis, without causing gastrointestinal or other toxicities
previously reported for Notch pathway inhibitors. Several pan-Notch inhibitors have entered the
clinic, but a major obstacle in their progress has been the induction of dose-limiting toxicity due
to intestinal goblet cell metaplasia. Studies in conditional knockout mice have suggested that
goblet cell metaplasia requires inhibition of both Notch1 and Notch2 receptors [48]. Therefore,
much effort has gone into the development of antibodies that specifically target various epitopes
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
16
of Notch1 [20, 21, 27, 49, 50], with the expectation that this would eliminate gut toxicity.
Surprisingly, however, it was reported that treatment of mice with antibodies that target Notch1
through binding to the negative regulatory region (NRR) can induce some level of goblet cell
metaplasia [27]. One possible explanation for the difference in toxicity observed between 23814
and the previous report is the differing mechanisms of action. While 23814 binds to the Notch1
LBD and prevents association of the receptor with its activating ligands, the NRR antibody
allows ligand binding but keeps the receptor locked in an “off” state. It is possible that Notch1
receptors bound to an NRR antibody are able to bind and sequester ligands in this inactive
conformation, thereby affecting signaling of other Notch family receptors. Other Notch1-specific
antibodies detailed in the literature either do not cross-react with murine Notch1 or are not
characterized in in vivo experiments, so it is unknown whether induction of goblet cell
metaplasia is a feature of anti-Notch1 NRR antibodies in general, or a specific property of
certain antibodies. Another possible explanation for the lack of toxicity for 23814 versus other
Notch1 antibodies is simply that there is a difference in potency or antibody kinetics that leads to
a larger therapeutic window. Perhaps for inhibitors of targets such as Notch, which regulate
stem cell and normal tissue homeostasis, the key is not to treat at the maximum tolerated dose,
but instead to treat at the minimum efficacious dose.
Our data suggests that Notch1 inhibitors can be effective in settings where anti-
angiogenesis therapies are indicated. We have demonstrated potent tumor inhibition in a
patient-derived primary RCC model using 23814 monotherapy, as well as in combination with
tivozanib in several models that exhibit partial resistance to VEGFR inhibitor alone. Although
monotherapy treatment of the BH tumors with 23814 resulted in only modest tumor inhibition,
23814 clearly potentiated the anti-tumor activity of tivozanib when used in combination.
Addition of 23814 to tivozanib treatment resulted in significantly greater tumor necrosis, as well
as increased induction of hypoxia genes, compared to tivozanib monotherapy (Fig. 5C and
Supplementary Fig. S3). This is consistent with an anti-angiogenic mechanism of inhibition, and
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
17
is similar to what has been observed when DLL4 and VEGF inhibitors are combined [28], further
suggesting that inhibition of the Notch1 and VEGF pathways is not redundant [23, 24, 45]. Use
of a Notch1 antibody may have the added advantage of simultaneously blocking oncogenic
activation of Notch1 signaling not only through DLL4 in the stroma, but also through other
ligands such as Jag1 or Jag2 in the tumor itself, possibly impacting pro-survival signals in
addition to angiogenesis.
Undoubtedly, there are many mechanisms leading to VEGF inhibitor resistance, and
discovering which tumors are likely to respond to Notch1 inhibition is essential for success in the
clinic. Our BH tumors, derived from chimeric primary mouse models, demonstrate a wide range
of responses to combined tivozanib 23814 therapy, likely reflecting what would be observed in
the patient population. It is well-established that the expression level of Notch1 receptor itself is
not predictive of response to Notch inhibitors in the clinic, and response to combination therapy
is unlikely to depend on Notch signaling alone. Therefore, instead of limiting our scope to
differences in canonical Notch pathway genes, we used unsupervised hierarchical clustering in
an attempt to predict a priori which tumor models would respond to tivozanib/23814 treatment.
The two models selected for testing from the same subgroup as the best responder exhibited
robust tumor inhibition in response to tivozanib/23814 combination treatment that significantly
exceeded response to either therapy alone. This underscores the ability to enrich for
responders based on the unique combination of biological pathways that are activated in
subgroups of tumors.
Our findings demonstrate that the 23814 Notch1 antibody exhibits anti-tumor efficacy
that is not redundant with VEGF pathway inhibition, and that angiogenesis inhibitors with
different mechanisms of action can be combined to achieve improved outcome. The lack of
toxicity may allow combination of 23814 with VEGF inhibitors or other therapies at efficacious
doses, and suggests that evaluation of specific Notch1 inhibitors as a treatment option for VEGF
inhibitor-resistant tumors is warranted.
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
18
Acknowledgements
The authors gratefully acknowledge Hamid Tissire, Sara Haserlat, and Sandra Abbott for
their contributions to antibody screening and engineering, Lynette Cook for 23814 purification,
and Tom Magee for toxicology assistance.
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
19
References
1. Artavanis-Tsakonas S and Muskavitch MA. Notch: the past, the present, and the future.
Curr Top Dev Biol 2010;92:1-29.
2. Radtke F and Raj K. The role of Notch in tumorigenesis: oncogene or tumour
suppressor? Nature Reviews Cancer 2003;3:756-767.
3. Bray SJ. Notch signalling: a simple pathway becomes complex. Nat Rev Mol Cell Biol
2006;7:678-89.
4. Ranganathan P, Weaver KL, and Capobianco AJ. Notch signalling in solid tumours: a
little bit of everything but not all the time. Nat Rev Cancer 2011;11:338-51.
5. Ntziachristos P, Lim JS, Sage J, and Aifantis I. From fly wings to targeted cancer
therapies: a centennial for notch signaling. Cancer Cell 2014;25:318-34.
6. Agrawal N, Jiao Y, Bettegowda C, Hutfless SM, Wang Y, David S, et al. Comparative
genomic analysis of esophageal adenocarcinoma and squamous cell carcinoma. Cancer
Discov 2012;2:899-905.
7. Wang NJ, Sanborn Z, Arnett KL, Bayston LJ, Liao W, Proby CM, et al. Loss-of-function
mutations in Notch receptors in cutaneous and lung squamous cell carcinoma. Proc Natl
Acad Sci U S A 2011;108:17761-6.
8. Stransky N, Egloff AM, Tward AD, Kostic AD, Cibulskis K, Sivachenko A, et al. The
mutational landscape of head and neck squamous cell carcinoma. Science
2011;333:1157-60.
9. Rampias T, Vgenopoulou P, Avgeris M, Polyzos A, Stravodimos K, Valavanis C, et al. A
new tumor suppressor role for the Notch pathway in bladder cancer. Nat Med
2014;20:1199-205.
10. Kopan R and Ilagan MX. The canonical Notch signaling pathway: unfolding the
activation mechanism. Cell 2009;137:216-33.
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
20
11. Gordon WR, Vardar-Ulu D, Histen G, Sanchez-Irizarry C, Aster JC, and Blacklow SC.
Structural basis for autoinhibition of Notch. Nat Struct Mol Biol 2007;14:295-300.
12. Weng AP, Ferrando AA, Lee W, Morris JPt, Silverman LB, Sanchez-Irizarry C, et al.
Activating mutations of NOTCH1 in human T cell acute lymphoblastic leukemia. Science
2004;306:269-71.
13. Fabbri G, Rasi S, Rossi D, Trifonov V, Khiabanian H, Ma J, et al. Analysis of the chronic
lymphocytic leukemia coding genome: role of NOTCH1 mutational activation. J Exp Med
2011;208:1389-401.
14. Puente XS, Pinyol M, Quesada V, Conde L, Ordonez GR, Villamor N, et al. Whole-
genome sequencing identifies recurrent mutations in chronic lymphocytic leukaemia.
Nature 2011;475:101-5.
15. Arruga F, Gizdic B, Serra S, Vaisitti T, Ciardullo C, Coscia M, et al. Functional impact of
NOTCH1 mutations in chronic lymphocytic leukemia. Leukemia 2014;28:1060-70.
16. Cancilla B, Cain, J, Wang, M, Beviglia, L, Shah, J, Gurney, A, et al. Anti-Notch1 antibody
(OMP-52M51) impedes tumor growth and cancer stem cell frequency (CSC) in a chemo-
refractory breast cancer xenograft model with an activating Notch1 mutation and
screening for activated Notch1 across multiple solid tumor types. In 104th Annual
Meeting of the Amercian Association for Cancer Research. 2013. Washington, DC.
Philadelphia (PA): AACR; 2013. Abstract 3728.
17. Kluk MJ, Ashworth T, Wang H, Knoechel B, Mason EF, Morgan EA, et al. Gauging
NOTCH1 Activation in Cancer Using Immunohistochemistry. PLoS One 2013;8:e67306.
18. Hoey T, Yen WC, Axelrod F, Basi J, Donigian L, Dylla S, et al. DLL4 blockade inhibits
tumor growth and reduces tumor-initiating cell frequency. Cell Stem Cell 2009;5:168-77.
19. Pannuti A, Foreman K, Rizzo P, Osipo C, Golde T, Osborne B, et al. Targeting Notch to
target cancer stem cells. Clin Cancer Res 2010;16:3141-52.
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
21
20. Qiu M, Peng Q, Jiang I, Carroll C, Han G, Rymer I, et al. Specific inhibition of Notch1
signaling enhances the antitumor efficacy of chemotherapy in triple negative breast
cancer through reduction of cancer stem cells. Cancer Lett 2013;328:261-70.
21. Agnusdei V, Minuzzo S, Frasson C, Grassi A, Axelrod F, Satyal S, et al. Therapeutic
antibody targeting of Notch1 in T-acute lymphoblastic leukemia xenografts. Leukemia
2014;28:278-88.
22. Ma W, Gutierrez A, Goff DJ, Geron I, Sadarangani A, Jamieson CA, et al. NOTCH1
signaling promotes human T-cell acute lymphoblastic leukemia initiating cell
regeneration in supportive niches. PLoS One 2012;7:e39725.
23. Noguera-Troise I, Daly C, Papadopoulos NJ, Coetzee S, Boland P, Gale NW, et al.
Blockade of Dll4 inhibits tumour growth by promoting non-productive angiogenesis.
Nature 2006;444:1032-7.
24. Ridgway J, Zhang G, Wu Y, Stawicki S, Liang WC, Chanthery Y, et al. Inhibition of Dll4
signalling inhibits tumour growth by deregulating angiogenesis. Nature 2006;444:1083-7.
25. Scehnet JS, Jiang W, Kumar SR, Krasnoperov V, Trindade A, Benedito R, et al.
Inhibition of Dll4-mediated signaling induces proliferation of immature vessels and
results in poor tissue perfusion. Blood 2007;109:4753-60.
26. Li JL, Sainson RC, Shi W, Leek R, Harrington LS, Preusser M, et al. Delta-like 4 Notch
ligand regulates tumor angiogenesis, improves tumor vascular function, and promotes
tumor growth in vivo. Cancer Res 2007;67:11244-53.
27. Wu Y, Cain-Hom C, Choy L, Hagenbeek TJ, de Leon GP, Chen Y, et al. Therapeutic
antibody targeting of individual Notch receptors. Nature 2010;464:1052-7.
28. Miles KM, Seshadri M, Ciamporcero E, Adelaiye R, Gillard B, Sotomayor P, et al. Dll4
blockade potentiates the anti-tumor effects of VEGF inhibition in renal cell carcinoma
patient-derived xenografts. PLoS One 2014;9:e112371.
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
22
29. Patenaude A, Fuller M, Chang L, Wong F, Paliouras G, Shaw R, et al. Endothelial-
specific Notch blockade inhibits vascular function and tumor growth through an eNOS-
dependent mechanism. Cancer Res 2014;74:2402-11.
30. Milano J, McKay J, Dagenais C, Foster-Brown L, Pognan F, Gadient R, et al. Modulation
of notch processing by gamma-secretase inhibitors causes intestinal goblet cell
metaplasia and induction of genes known to specify gut secretory lineage differentiation.
Toxicol Sci 2004;82:341-58.
31. Takebe N, Nguyen D, and Yang SX. Targeting notch signaling pathway in cancer:
clinical development advances and challenges. Pharmacol Ther 2014;141:140-9.
32. Krop I, Demuth T, Guthrie T, Wen PY, Mason WP, Chinnaiyan P, et al. Phase I
pharmacologic and pharmacodynamic study of the gamma secretase (Notch) inhibitor
MK-0752 in adult patients with advanced solid tumors. J Clin Oncol 2012;30:2307-13.
33. Tolcher AW, Messersmith WA, Mikulski SM, Papadopoulos KP, Kwak EL, Gibbon DG, et
al. Phase I study of RO4929097, a gamma secretase inhibitor of Notch signaling, in
patients with refractory metastatic or locally advanced solid tumors. J Clin Oncol
2012;30:2348-53.
34. Yan M, Callahan CA, Beyer JC, Allamneni KP, Zhang G, Ridgway JB, et al. Chronic
DLL4 blockade induces vascular neoplasms. Nature 2010;463:E6-7.
35. Patnaik A, LoRusso, P, Munster, P, Tolcher, AW, Davis, SL, Heymach, R, et al. Safety
and early evidence of activity of a first-in-human phase I study of the novel cancer stem
cell (CSC) targeting antibody OMP-52M51 (anti-Notch1) administered intravenously to
patients with certain advanced solid tumors. In 26th EORTC-NCI-AACR Symposium on
Molecular Targets and Cancer Therapeutics. 2014. Barcelona, Spain: European Journal
of Cancer.
36. Kangsamaksin T, Murtomaki A, Kofler NM, Cuervo H, Chaudhri RA, Tattersall IW, et al.
NOTCH Decoys That Selectively Block DLL/NOTCH or JAG/NOTCH Disrupt
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
23
Angiogenesis by Unique Mechanisms to Inhibit Tumor Growth. Cancer Discov
2015;5:182-97.
37. Watters JW, Cheng C, Majumder PK, Wang R, Yalavarthi S, Meeske C, et al. De novo
discovery of a gamma-secretase inhibitor response signature using a novel in vivo
breast tumor model. Cancer Res 2009;69:8949-57.
38. Radtke F, Wilson A, Stark G, Bauer M, van Meerwijk J, MacDonald HR, et al. Deficient T
cell fate specification in mice with an induced inactivation of Notch1. Immunity
1999;10:547-58.
39. van Es JH, van Gijn ME, Riccio O, van den Born M, Vooijs M, Begthel H, et al.
Notch/gamma-secretase inhibition turns proliferative cells in intestinal crypts and
adenomas into goblet cells. Nature 2005;435:959-63.
40. Real PJ, Tosello V, Palomero T, Castillo M, Hernando E, de Stanchina E, et al. Gamma-
secretase inhibitors reverse glucocorticoid resistance in T cell acute lymphoblastic
leukemia. Nat Med 2009;15:50-8.
41. Lobov IB, Renard RA, Papadopoulos N, Gale NW, Thurston G, Yancopoulos GD, et al.
Delta-like ligand 4 (Dll4) is induced by VEGF as a negative regulator of angiogenic
sprouting. Proc Natl Acad Sci U S A 2007;104:3219-24.
42. Kim WY and Kaelin WG. Role of VHL gene mutation in human cancer. J Clin Oncol
2004;22:4991-5004.
43. Hidalgo M, Amant F, Biankin AV, Budinska E, Byrne AT, Caldas C, et al. Patient-derived
xenograft models: an emerging platform for translational cancer research. Cancer
Discov 2014;4:998-1013.
44. Motzer RJ, Nosov D, Eisen T, Bondarenko I, Lesovoy V, Lipatov O, et al. Tivozanib
versus sorafenib as initial targeted therapy for patients with metastatic renal cell
carcinoma: results from a phase III trial. J Clin Oncol 2013;31:3791-9.
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
24
45. Kuhnert F, Kirshner JR, and Thurston G. Dll4-Notch signaling as a therapeutic target in
tumor angiogenesis. Vasc Cell 2011;3:20.
46. Zhou Y, Rideout WM, 3rd, Zi T, Bressel A, Reddypalli S, Rancourt R, et al. Chimeric
mouse tumor models reveal differences in pathway activation between ERBB family-
and KRAS-dependent lung adenocarcinomas. Nat Biotechnol 2010;28:71-8.
47. Benedito R, Rocha SF, Woeste M, Zamykal M, Radtke F, Casanovas O, et al. Notch-
dependent VEGFR3 upregulation allows angiogenesis without VEGF-VEGFR2
signalling. Nature 2012;484:110-4.
48. Riccio O, van Gijn ME, Bezdek AC, Pellegrinet L, van Es JH, Zimber-Strobl U, et al.
Loss of intestinal crypt progenitor cells owing to inactivation of both Notch1 and Notch2
is accompanied by derepression of CDK inhibitors p27Kip1 and p57Kip2. EMBO Rep
2008;9:377-83.
49. Aste-Amezaga M, Zhang N, Lineberger JE, Arnold BA, Toner TJ, Gu M, et al.
Characterization of Notch1 antibodies that inhibit signaling of both normal and mutated
Notch1 receptors. PLoS One 2010;5:e9094.
50. Sharma A, Paranjape AN, Rangarajan A, and Dighe RR. A monoclonal antibody against
human Notch1 ligand-binding domain depletes subpopulation of putative breast cancer
stem-like cells. Mol Cancer Ther 2012;11:77-86.
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
25
Figure Legends
Figure 1. In vitro characterization of 23814 activity. A, Notch1 luciferase reporter line co-
cultured with stable lines expressing Jag1, Jag2, DLL1 and DLL4 ligands in the presence of
23814 or control hIgG antibodies. Results represented as percent inhibition of luciferase activity.
B, Karpas 45 plated on wells coated with recombinant Jag1, Jag2, DLL4 or Fc (no ligand) in the
presence of 23814 or hIgG control (10 ug/ml). Lane 1, no ligand + hIgG; lane 2, ligand + hIgG;
lane 3, ligand + 23814 antibody. Western blots were probed with antibody specific for cleaved
Notch1 ICD (Cell Signaling). Equal amounts of lysate transferred to a separate blot were
probed with β-tubulin antibody as a loading control. C, Karpas 45 co-cultured with ligand-
expressing CHO lines in the presence of 23814 or hIgG. Total RNA was isolated after 24 hours
and gene expression assessed by TaqMan assay (* p<0.05 by t-test). D, FACS analysis of
23814 binding to CHO cells expressing human or murine Notch1, human Notch2, or human
Notch3.
Figure 2. 23814 inhibits Notch1 signaling in vivo and does not cause gut toxicity. A, 23814
treatment of mice (n = 5 animals/group) results in thymocyte depletion (***p<0.001; NS, not
significant). B, at doses where 23814 effectively inhibits Notch1 signaling in vivo, no weight loss
occurs. C, mice were treated with hIgG, 23814 (8 weeks), or the gamma secretase inhibitor
DBZ (4 weeks). Alcian blue staining of small intestines shows significant goblet cell conversion
in DBZ-treated mice, while 23814-treated mice looked similar to control animals. D, 23814 in
serum was captured with recombinant Notch1Fc and detected with SULFO-TAG labeled anti-
human F(ab’)2. A1-A5, mice treated with hIgG; B1-B5, mice treated with 23814.
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
26
Figure 3. 23814 inhibits growth of a renal clear cell PDX tumor. NCR nude mice (n = 10)
received either hIgG control, anti-ErbB3 (AV-203), 23814, or tivozanib. 23814 and tivozanib
monotherapy significantly inhibited tumor growth relative to control (*** p<0.001).
Figure 4. 23814 treatment increases tumor vasculature. A, HT-1080 tumor model (n = 10) was
treated with control hIgG, 23814, tivozanib, or a combination of 23814 plus tivozanib.
Combination treatment results in significantly greater tumor growth inhibition than either
tivozanib or 23814 alone (*** p<0.001). B, CD31 staining and microvessel density quantification
of HT-1080 tumors from mice (n = 3) treated for 3 days with hIgG, tivozanib, and/or 23814 (***
p<0.001, ** p<0.01).
Figure 5. Combined inhibition of Notch1 and VEGFR pathways enhances anti-tumor efficacy in
models partially resistant to VEGFR inhibitor monotherapy. A, BH tumor models (n = 10) exhibit
a range of responses to treatment with 23814, tivozanib, or a combination of the two (***
p<0.001, ** p<0.01, NS= not significant). For B, C, and D, analysis was performed on BH224
tumors harvested after 3 days of treatment. B, TaqMan analysis of Notch1 target gene
expression. Results normalized to the mean expression level of vehicle-treated tumors (n = 3).
*p<0.05. C, TaqMan analysis shows increased expression of the hypoxia-regulated genes ADM
(adrenomedullin), NDRG1 (N-myc downstream-regulated 1) and VEGFA, although statistical
significance was not reached. D, CD31 staining of treated tumors.
Figure 6. Unsupervised hierarchical clustering identifies tumors likely to respond to
23814/tivozanib combination therapy. A, cluster analysis demonstrates that BH tumors
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
27
separate into distinct subgroups. BH413 and BH226 (labeled in blue), representing a non-
responder and partial responder, respectively, cluster in different subgroups from BH224, while
BH228 and BH270 (labeled in green) are biologically similar. B, BH228 and BH270 tumor
growth inhibition is significantly increased by combination treatment with 23814/tivozanib
relative to tivozanib monotherapy (**p<0.01, *p<0.05).
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
Figure 1
D C
Jag2
DLL1
Jag1
DLL4
B
hNotch1 CHO Control CHO
hNotch1
mNotch1 CHO Control CHO
mNotch1 hNotch2
hNotch2 CHO Control CHO
hNotch3
hNotch3 CHO Control CHO
A
0
0.2
0.4
0.6
0.8
1
1.2
DTX1 HES4 NRARP
Rela
tive
Exp
ressio
n
hIgG 23814
* * *
30
40
50
60
80
100
220
30
40
50
60
80
100
220
30
40
50
60
80
100
220
b-Tubulin 50 50 50
Notch1 ICD
1 2 3
DLL4 1 2 3
Jag2 1 2 3
Jag1
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
Figure 2
B A
hIgG 23814 DBZ
C
0
100
200
300
20 mpk 20 mpk 10 mpk 2 mpk
hIgG 23814 Notch1 mAb
Tota
l T
hym
ocyte
Num
ber
(x10
6)
NS
*** ***
Days
Bo
dy
we
igh
t (g
)
0 5 10 15 200
5
10
15
20
25
hIgG (20mpk)
23814 (20mpk)
23814 (10mpk)
23814 (2mpk)
0.00
200.00
400.00
600.00
800.00
A1 A2 A3 A4 A5 B1 B2 B3 B4 B5
hIgG 23814
An
tib
od
y C
on
c [u
g/m
l]
a-Notch1 Ab at end of study
D
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
Figure 3
***
Days
***
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
Figure 4
B
A
tivozanib Vehicle 23814 23814+tivozanib
Days
Tu
mo
r vo
lum
e (
mm
3)
0 5 10 15 200
500
1000
1500
2000
Control
Tivozanib
23814
23814 + Tivozanib
0.00E+00
1.00E-04
2.00E-04
3.00E-04
4.00E-04
5.00E-04
6.00E-04
7.00E-04
Vehicle Tivo 23814 Combo
Nu
mb
er V
esse
ls/m
m2
Microvessel Density
***
** ***
***
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
Figure 5
B
A
0 5 10 15 20 250
500
1000
1500
2000
2500VehicleTivozanib23814Tivozanib+23814
days
tum
or v
olu
me
(m
m3
)
BH224
**
0 5 10 15 200
500
1000
1500
2000
2500
VehicleTivozanib23814Tivozanib+23814
daystu
mo
r v
olu
me
(m
m3
)
BH226
***
0 5 10 15 200
500
1000
1500
2000
2500VehicleTivozanib23814Tivozanib+23814
days
tum
or v
olu
me
(m
m3
)
BH413
NS
C
0
0.5
1
1.5
2
HEY1 HES1 FLT4
Re
lati
ve
Ex
pre
ss
ion
Vehicle
Tivozanib
23814
23814+Tivo
0
1
2
3
4
ADM NDRG1 VEGFA
Re
lati
ve E
xpre
ssio
n
Vehicle
Tivozanib
23814
23814+Tivo
D Vehicle 23814
Tivozanib Tivo+23814
*
* *
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
Figure 6
B
A
0 5 10 15 200
500
1000
1500
2000
2500
VehicleTivozanib23814Tivozanib+23814
days
tum
or
vo
lum
e (
mm
3)
0 5 10 15 200
500
1000
1500
2000
2500
VehicleTivozanib23814Tivozanib+23814
days
tum
or
vo
lum
e (
mm
3)
BH270 BH228
* **
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104
Published OnlineFirst May 20, 2015.Mol Cancer Ther Theresa Proia, Feng Jiang, Alisa Bell, et al. without Gastrointestinal ToxicityActivation, Modulates Angiogenesis and Inhibits Tumor Growth 23814, an Inhibitory Antibody of Ligand-Mediated Notch1
Updated version
10.1158/1535-7163.MCT-14-1104doi:
Access the most recent version of this article at:
Material
Supplementary
http://mct.aacrjournals.org/content/suppl/2015/05/21/1535-7163.MCT-14-1104.DC1
Access the most recent supplemental material at:
Manuscript
Authoredited. Author manuscripts have been peer reviewed and accepted for publication but have not yet been
E-mail alerts related to this article or journal.Sign up to receive free email-alerts
Subscriptions
Reprints and
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Permissions
Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)
.http://mct.aacrjournals.org/content/early/2015/05/20/1535-7163.MCT-14-1104To request permission to re-use all or part of this article, use this link
on October 16, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from
Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited. Author Manuscript Published OnlineFirst on May 20, 2015; DOI: 10.1158/1535-7163.MCT-14-1104