Transcript Abraxane in Metastatic Pancreatic Cancer

SPARC in Pancreatic
Cancer
Nab-paclitaxel properties
• utilises the properties of albumin to reversibly bind
paclitaxel, and transport it across the endothelial cell and
concentrate it in areas of tumor
• avoids the use of Cremophor EL, which contributes to
serious toxicity (e.g. hypersensitivity, axonal degeneration)
and requires special infusion tubing, premedication and
prolonged infusion
• shows improved bioavailability and linear
pharmacokinetics, whereas CrEL forms micelles
entrapping the paclitaxel, leading to decreased unbound
drug fraction, decreased drug clearance and lack of dosedependent antitumour activity
Solvent-based taxanes provoke
formation of micelles in circulation
Large
micelle
• Micelle formation in the
circulation entraps
paclitaxel in plasma
• Resulting non-linear
pharmacokinetics
contribute to a lack
of dose-dependent
antitumour activity
Control plasma
Plasma + solvent-based
paclitaxel
Aapro et al. EJC Suppl. 2008;6:3–11
Hamad et al. Expert Opin Drug Deliv. 2008;5: 205–219
nab -Paclitaxel Results in Higher Tumoral Uptake
Compared With paclitaxel
Tumor Uptake in Nude Mice Xenografts Following 20 mg/kg Dose of Paclitaxel
Paclitaxel (nCi/g)
140
120
AUC
Ka
(nCi•hr/g) (hr1)
nab-paclitaxel 3,632
0.43
Taxol
2,739
0.13
100
80
Tumor AUC nab-paclitaxel =
1.33 x Taxol
p < .0001 ANOVA
60
40
0.01
0.1
1
10
100
Hours
•
The accumulation of paclitaxel in tumors was 33% higher for nab-paclitaxel compared with paclitaxel (P < .0001)1
AUC, area under the curve; KA, absorption rate.
Desai et al. Clin Cancer Res. 2006; 12:1317-1324.
KEY STEPS in albumin-paclitaxel
delivery to tumor
• Albumin initiates the endothelial transcytosis of paclitaxel by binding to
a cell surface receptor: 60-kDa glycoprotein (gp60).
• In turn, gp60 associates with caveolin-1 resulting in the invagination of
the endothelial cell membrane trapping the complex in vesicular
structures called caveolae.
• Clustering of the gp60-albumin complex during vescicle formation
reduces receptor affinity for albumin, which permits the release of
albumin and any bound ligands to the abluminal side of the cell.
• Albumin accumulates in tumors, possibly due, in part, to the secretion
of the albumin-binding protein SPARC (secreted protein, acidic and rich
in cysteine, osteonectin or BM-40), which, in turn, may result in
preferential intatumoural accumulation of albumin-bound molecules.
The Unique Properties of Albumin Improve the
Risk to Benefit Profile of nab-Paclitaxel
• 130-nm sized albumin-paclitaxel complexes1,2
– nab-Paclitaxel is the first nanotechnology-derived
agent approved for the treatment of breast cancer
– Albumin gives nab-paclitaxel linear
pharmacokinetics3 = predictable drug exposure
with dose modification
Albumin
Paclitaxel
nab-Paclitaxel
particle
nab® is a registered trademark
of Celgene Corporation.
2D
Conceptualization
1. Desai et al. SABCS. 2004 [abstract 1071].
2. Kratz et al. J Control Release. 2008;132(3):171-183.
3. Ibrahim et al. Clin Cancer Res. 2002;8(5):1038-1044.
Mechanism of Action of nab-Paclitaxel
SPARC, secreted protein acidic and rich in cysteine.
Investigation of the functional importance of SPARC
with respect to nab-paclitaxel is ongoing.
Albumin-Mediated Transcytosis of Paclitaxel
Endothelial
cells
Subendothelial space
Tumor cells
SPARC, Secreted Protein Acidic and Rich in Cysteine.
Investigation of the functional importance of SPARC with
respect to nab-paclitaxel is ongoing.
SPARC level in heterogeneous tumors affects
relative response to nab-paclitaxel
Control
Abraxane 15 mg/kg q4x3
Saline
Abraxane 15 mg/kg q4dx3
Control
Abraxane, 10 mg/kg, qdx5, 2 cycles
HT29
MDA-MB435
3000
Tumor Volume (mm3)
2500
2000
1500
1000
2500
Tumor volume (mm3 )
Tumor Volume (mm3)
MDA-MB-231
3000
1500
2000
1500
1000
500
1000
500
500
0
0
0
0
10
20
30
Days
Low
SPARC
(TGI = 36%)
40
0
20
40
Days
Medium
SPARC
(TGI = 60%)
60
0
20
40
Days
High
SPARC
(TGI = 81%)
60
Peritumoral Fibroblast SPARC Expression and Patient
Outcome With Resectable Pancreatic Adenocarcinoma
• Stromal SPARC
was associated
with worse
outcome and
poor survival
• Tumoral SPARC
did not correlate
with survival
Infante 2007
Overexpression of SPARC gene in human gastric carcinoma
and its clinic-pathologic significance
Diffuse Type
Intestinal Type
Non-cancerous Mucosa
Gastric Cancer
Wang, C-S et al - British Journal of Cancer (2004) 91, 1924 – 1930
Summary of SPARC as a
marker of poor prognosis
Classification
SPARC Expression/Function
Hepatocellular
Carcinoma
Overexpression by stromal myofibroblasts
correlates well with angiogenesis & tumor
progression
Overexpression in juxtratumoral perivascular
cells but not non-malignant brain vessels
Significant decrease in plasma levels of SPARC
has a prognostic value & shows + correlation
with Hb levels & platelet counts
A diagnostic marker for invasive meningiomas
regardless of grade
High levels of SPARC mRNA & protein as a
marker of CaP metastatic foci
Glioblastoma
Multiplle Myeloma
Meningioma
Prostate Carcinoma
Reference
Lau et al. 2006
Pen et al. 2007
Turk et al. 2005
Remple et al.
1999
Thomas et al.
2000
Head & Neck Cancer
High/Marker of poor prognosis
Chin et al. 2005
Tongue Carcinoma
High/Marker of poor prognosis
Kato et al. 2005
Cervical Carcinoma
High/Marker of poor prognosis
Sova et al. 2006
Non-small cell
lung cancer
High/Marker of poor prognosis
Bladder Cancer
High/Marker of poor prognosis
Melanoma
High levels correlate with metastasis
Esophageal Cancer
High/Marker of poor prognosis
Breast Cancer
High/Marker of poor prognosis; shows +
correlation with stage & grade
Koukourakis
et al. 2003
Yamanaka et al.
2001
Massi et al.
1999
Yamashita et al.
2003
Watkins et al.
2005
SPARC Expression in Microarrays Performed on
Tumors Taken Directly From Patients
Tumor Type
# with SPARC /# studied
(%)
Breast
6/9
67%
Ovary
5/15
33%
Pancreas
13/16
81%
Melanoma
15/17
88%
Adrenal
2/5
40%
Colon
4/15
27%
Total†
76/113
67%
*Increased expression at level of 0.001 versus normal tissue
Slide courtesy Dan Von Hoff, AACR 2006
Preclinical Platform
Rubio et al, CCR 2006
Tumors regressed 50% of its initial size (%)
Average Response Rate in Xenografts (n = 11)
60
50
40
30
20
10
0
GEM
ABI
GEM+ABI
On a Microscopic level
1. Pancreatic cancers are poorly perfused
• have a poor blood supply
2. One factor – an intense fibro inflammatory
reaction – stroma – squeezes out blood supply and
stops infiltration of immunocytes
3. Need to attack the stroma to improve tumor cells
killing
17
Effects of nab-paclitaxel on Tumor Stroma
Collagen Type I Staining
Effects of nab-paclitaxel on Blood Vessels
B
Concentration ng/g tumor (Mean ± SEM)
Effects of nab-paclitaxel on Gemcitabine Delivery
8000
GEM alone
7000
GEM+ABI
6000
5000
4000
3000
2000
1000
0
GEM
Questions in Development
• Role and regulation of SPARC in Pancreatic Cancer.
• SPARC as
• prognostic (which patients need rx)
• predictive biomarker (which patients are likely to benefit
from a specific rx).
• Stromal effects.
• Diagnostic test.
21
Hypothesis:
Albumin-Binding Proteins May Aid in the Uptake of
nab-paclitaxel Into Tumors
• Accumulation of albumin in tumors may be
mediated by the protein SPARC1
– SPARC binds albumin2
– Many tumor types overexpress SPARC compared
with normal tissues3-5
• High SPARC expression has been shown to be a negative
prognostic indicator in many cancer types4,5
SPARC, Secreted Protein Acidic and Rich in
Cysteine.
1. Kratz F, et al. J Controlled Release. 2008; 132:171-183.
2. Schnitzer JE, et al. J Biol Chem. 1994;269(8):6072-6082.
3. Watkins G, et al. Prostaglandins, Leukotrienes and Essential Fatty Acids. 2005;72:267-272.
4. Desai N, et al. Transl Oncol. 2009;2:59-64.
5. Podhajcer OL et al. Cancer Metastasis Rev. 2008;27:691-705.
22
SPARC Expression and Clinical Response
D. von Hoff et al. ASCO 2009 Poster # 4525



SPARC status by IHC was available for 32 RECIST evaluable patients (investigator dataset:
2CR, 14PR, 14SD, 2PD)
Staining of tumor cells (and not stromal fibroblasts) by antibody P showed improved
response for SPARC+ patients (P = 0.027)
Other epitopes of SPARC showed similar response between SPARC+ve and SPARC-ve
groups (P = NS)
Fraction (%) of Patients Responding
SPARC status
SPARC Positive
Antibody P
(Tumor Cell)
(N=32)
8/10 (80%)
2 CR / 6 PR
SPARC Negative 8/22 (36.4%) 0 CR / 8 PR
P-value
Abbreviations:
M = antibody M;
NS = not significant;
P = antibody P
0.027
Antibody M
(Tumor Cell)
(N=32)
Antibody P
(Stromal Fibroblasts)
(N=27)
Antibody M
(Stromal Fibroblasts)
(N=27)
2/5 (40%)
8/16 (50%)
3/5 (60%)
14/27 (52%)
7/11 (64%)
12/22 (55%)
NS
NS
NS
P positive pt # 014
P negative pt # 012
SPARC Expression and Clinical Response
D. von Hoff et al. JCO 2011
 SPARC status was evaluated in 36 patients:
a significant increase in OS was observed for patients in the high-SPARC
group versus the low-SPARC group (median OS: 17.8 vs 8.1 months,
respectively, P=0.0431)
Though high SPARC
expression is typically
a poor prognostic
factor, it actually
predicts an improved
response to nabpaclitaxel in this trial in
terms of overall
survival
1. Celgene data on file.
SPARC Assessment: actual tools
• Several antibodies are available and have been
used in publications assessing SPARC protein
• Antibodies have different staining patterns, thus
need to optimize an assay:
– Compare available antibodies
– Standardize antigen retrieval, antibody dilution,
secondary antibody and automated staining systems.
– Scoring criteria needs to be simplified and
standardized
– Tumor compartments staining with SPARC will vary
based on histology
25
Biomarker Development:
Requirements
Clinician/Lab education
Commercialisation
Auditing for result
consistency
Distribution to laboratories
Validation
Regulatory test approval
Assay refinement & development IVD test
Clinical validation of candidate m. using assay
Discovery
Develop prototype assay
Identification of candidate markers
Generation of a biomarker hypothesis
Biomarker evaluation in CA046
• Archival tumor tissue for SPARC IHC
– Slides and/or blocks for received from 160
patients to date
– Further explore role of SPARC in response to nabpaclitaxel in context of a randomized trial
27
Next steps in developing SPARC
prototype assay
• The prototype assay is based on IHC
methodologies
– a collaboration with EU and USA academic
centers has been established with the goal to
define the IHC platform
– M. Hidalgo, Madrid, will lead this collaboration,
and the kick-off meeting will take place October
17 in Seville
28
SPARC: where we are today
Clinical use
Potentially valuable
Early stage assessment
Exploratory: current standing
of SPARC assessment
In Summary
• Stromal components are strategic targets
in Pancreatic Cancer.
• SPARC is a stromal component.
• Gem-Nab-Paclitaxel promising activity.
• Stromal depletion vs stroma modification.
30
Back-up
31
SPARC IHC method
• Developed at MPI (AZ); transferred to
St.John’s pathology lab (CA)
• Two antibodies used; each slide scored
separately
– Monoclonal R&D #MAB941(1:250; 30 min)
– Monoclonal Haematologica Technologies, Inc,
#A0N5031 (1:250; 30 min)
– Polyclonal also evaluated, but not used
• R&D AF941 (1:150; 30 min)
• DAKO Automated stainer, NO Antigen retrieval
32
Scoring criteria
• Tumor compartments assessed separately
– Tumor, fibroblasts, blood vessels, inflammatory
cells, background stroma tissue, any normal tissue
• Highest intensity (0 to +4) for a compartment
(>10% of cells in that compartment) recorded
• % staining at highest intensity in compartment
recorded
33
34