Functional phenotypic screening of bi-specific antibodies that combine RTK and
stem cell targets in genetically heterogeneous colon carcinoma organoids
1OcellO
B.V. Leiden, The Netherlands;
2Merus
B.V. Utrecht, The Netherlands;
3Hubrecht
Organoid Technology (HUB), The Netherlands; 4Hubrecht
Institute, The Netherlands; 5suppresSTEM consortium, EC FP7 #601876
target stem cells to kill the tumour
combine stem cell targeting and RTK blocking
Organoids are stem cell-derived human epithelial ‘mini-organs’ from tissues of patients with various
diseases including cancer. In culture, organoids stably maintain the genotype/phenotype of the patient
tissue, enabling their use for preclinical drug discovery and validation and a tool for precision medicine.
receptor tyrosine kinase targeting
stem cell targeting
* van de Wetering et al. Cell. 2015 7;161(4):933-45
P8T
1. Broad morphological and mutational heterogeneity of CRC organoids
X
X
X
X
X
X
X
X
X
X
X
X
X
X
P19Ta
P18T
P19Tb
P14T
P11T
P7T
X X
X X
X X
X
X
X
X
X
X
X X X
X X
X
X
X
X
X
X
X
X
X
X
X X
X X
X X X
X
X
X
X X
X
X
X
X
P28T
X X
X X X
X
X
wild type organoids
X
P10N
P11N
P7N
X
X
X
X
X
P20N
X
P14N
P27N
X
X
X
X
X
X
X
X
X
X
X
X X X
The broad heterogeneity of mutation
spectra of the organoids is reflected in
a broad heterogeneity of organoid
phenotypes. Some CRC organoids form
well differentiated spheroids with a
single lumen that resembles the
phenotype of normal wild type
organoids, whereas others exhibit
multiple lumens or are poorly
differentiated without a luminal cavity.
Different growth conditions and different
patient samples can be discriminated in
multi-dimensional feature space
Nuclei mask
Rhodamine channel
Patient #18 + EGF
Patient #18 no GF
Organoid mask
Single nuclei identification Organoid branching
Morphological features contribute to the detection
of specific drug actions and can outperform
measurements of proliferation
Lumen mask
Lumen branching
Single cell mask
10
125
CellTiter-Glo®:
nucleus count:
lumen complexity:
100
IC50 = 222 nM
IC50 = 197 nM
IC50 = 43 nM
RNF43
ZNRF3
mAb-T.01
mAb-E.01
EGFR
mAb-E.02
mAb-E.03
mAb-E.04
mAb-H.01
HER3
mAb-H.02
mAb-H.03
mAb-H.04
Potent EGFR-inhibiting antibodies
mAb-T.01
mAb-E.01
mAb-E.02
mAb-E.03
mAb-E.04
mAb-H.01
mAb-H.02
mAb-H.03
mAb-H.04
control
(no EGF)
control
(with EGF)
AG1478
BEZ235
Dasatinib
Erlotinib
-10
-20
PHA-665752
-30
10 0
10 1
10 2
10 3
10 4
10 5
Concentration (nM)
75
10
50
Patient 26
control
(no EGF)
control
(with EGF)
AG1478
BEZ235
Dasatinib
Erlotinib
0
25
-10
Antibodies that do not affect organoid outgrowth
PHA-665752
0
• A 3D image stack is obtained from fixed and stained
organoids
• A 3D reconstruction of the micro-tissues in the microplate
well is generated
• After segmentation, features are extracted such as number,
shape and size of microtissues, number shape and size of
cells, number shape and size of nuclei.
• Tissue/disease-specific features are also collected, such as
spikes, protrusions and network formation in invasive
tumours, lumen formation and planar polarity in well
differentiated tumours
10 - 6
10
10 - 5
10
10 - 4
10
10 - 3
10
10 - 2
10
-20
10 0
10 1
no EGF
10 2
10 3
10 4
0
Concentration
[erlotinib]
(mM) (mM)
Patient 14
control
(no EGF)
control
(with EGF)
AG1478
BEZ235
Dasatinib
Erlotinib
-5
Patient #14 + EGF
Patient #14 no GF
Patient #14 + HRG
control
20nM Erlotinib
10 5
Concentration (nM)
2mM Erlotinib
-10
-15
PHA-665752
-20
-25
10 0
10 1
10 2
10 3
10 4
10 5
Concentration (nM)
Some features, such as extent of lumen formation, are more sensitive to drug treatment than cell proliferation-associated features (e.g. organoid size or
number of cells per organoid). This improves the sensitivity of the assay to detect active molecules. A set of 10 features is selected to create a drug
response profile. We observed that the presence of activating mutations does not always correlate with sensitivity to corresponding pathway inhibitors,
underscoring the need for empirical testing of drugs to predict patient sensitivity.
To reflect key mutational subtypes, from the panel of 20 CRC organoids, one
APC mutant (P18T), one APC mutant/SMAD4 mutant (P14T) and one APC
wildtype/RNF43 mutant (P19Tb) CRC organoid model were selected to
screen a bispecific antibody panel that targets EGFR or HER3 in combination
with ZNRF3, RNF43, LGR5 or LGR4. Several bispecific antibodies were
identified that potently inhibited growth of the CRC organoids with differing
sensitivities dependent on the mutational background.
Conclusions. These results demonstrate that high content screening of CRC
organoids is an effective strategy to identify novel inhibitors of CRC tumor
growth and enable identification of bispecific antibodies that target
colorectal cancer stem cells with different mutational backgrounds.
Lumen complexity mask
5. Screening and scoring of RTK/WNT bispecific antibodies in three different CRC organoids grown in three different growth factor conditions
APC mutant CRC P18T
APC & SMAD4 mutant CRC P14T
APC wildtype CRC P19Tb
Inhibition
(rel. score)
No GF
Patient 18
0
X
Overlay of segmentation
TT
Phenotypic responses to SoC
treatments in patient CRC organoids
Patient #18 + HRG
X
RGB: focus on lumen
LGR5
Identification of functional RTK arms
control
2. Morphological analysis of CRC organoids
DAPI channel
LGR4
Antibody production matrix
Potent HER3-inhibiting antibodies
X
X
TT
P20T
KR
AS
PIK
3C
FB A
XW
SM 7
AD
P5 4
3
TC
F7
L
BR 2
AF
AR
ID1
AR A
ID2
ER
BB
AX 3
IN2
SM
AD
NR 2
AS
CT
NN
B1
PO
LE
RN
F4
3
P25T
P17T
Stem cell targets
3. Morphological evaluation of drug effects in CRC organoids
PCA 0
P24Ta
P9T
P31T
P26T
P6T
P10T
P16T
P17T
P11T
P20T
P8T
P24Tb
P23T
P25T
P19Ta
P19Tb
P5T
P27T
P14T
P18T
P7T
P28T
AP
C
Gene
mutations:
P23T
P16T
P14T
PCA 0
P24Tb
P8T
P10T
P6T
P26T
P31T
P18T
Activity (%)
P9T
P24Ta
4. Screening of RTK/WNT bispecific antibodies in CRC organoids
Why use bi-specific antibodies?
 A panel of organoids from >40 colorectal cancer patients has been developed from biopsies *
 Organoids show different phenotypes and have known different mutational backgrounds
 Screening results can be related back to the genetic profile of each organoid for stratification
mAb-T.01
mAb-L.01
mAb-L.02
mAb-L.03
mAb-L.04
mAb-L.05
mAb-L.06
mAb-L.07
mAb-L.08
mAb-L.09
mAb-L.10
mAb-L.11
mAb-L.12
mAb-L.13
mAb-L.14
mAb-L.15
mAb-L.16
mAb-L.17
mAb-L.18
mAb-L.19
mAb-L.20
mAb-L.21
mAb-L.22
mAb-L.23
mAb-L.24
mAb-L.25
mAb-L.26
mAb-R.01
mAb-R.02
mAb-R.03
mAb-R.04
mAb-R.05
mAb-R.06
mAb-R.07
mAb-R.08
mAb-R.09
mAb-Z.01
mAb-Z.02
mAb-Z.03
mAb-Z.04
mAb-Z.05
mAb-Z.06
mAb-Z.07
mAb-Z.08
mAb-Z.09
mAb-Z.10
mAb-Z.11
mAb-Z.12
mAb-Z.13
mAb-Z.14
mAb-Z.15
mAb-Z.16
mAb-Z.17
mAb-Z.18
Robert Vries3,4,5, Mark Throsby2,5 and Leo Price1,5
Background. The relationship between cell and tissue morphology and disease state forms the basis of modern histopathology. In an in vitro experimental setting, cell and tissue morphology can be used to
identify potential drugs, discriminate different modes of drug action and discriminate between therapeutic activity and toxicity. 3D cultured tissues show gene expression, differentiation and functional
characteristics which more closely reflect the situation in vivo. Furthermore, complex 3D tissue architecture exemplified by organoid cultures increases the scope for discriminating different drug responses. We
used 3D image analysis of a panel of 20 sequenced colorectal cancer organoids to characterize the genotype-phenotype relationship and the response to a broad spectrum of inhibitors of signaling pathways.
This high throughput approach was then used as a primary screen to evaluate the functional activity of a panel of 545 bispecific antibodies comprised of a HER3 or EGFR targeting arm combined with a LGR4,
LGR5, ZNRF3 or RNF43 targeting arm to target stem cells (antibody generation poster C21).
RTK targets
Bram Herpers1,5, Rob Roovers2,5, Berina Eppink2,5, Marc van de
Wetering3,4,5, Kuan Yan1,5, Lucia Salinaro1,5 Wim de Lau4,5, Hans Clevers4,5,
PCA 0
C24
Neg_ctrl
Cetuximab
mAb-E.01
[email protected]
mAb-E.02
mAb-E.03
www.OcellO.nl
mAb-E.04
mAb-H.01
mAb-H.02
mAb-H.03
mAb-H.04
mAb-T.01
Size : dose
Colours:
Green: Heregulin
Red: EGF
Blue: Wnt3A
Black: negative controls
Orange: Reference Ab’s
TT: non-targeting Ab arm
 OcellO’s screening platform uses high content 3D imaging for
sensitive detection of therapeutic responses
 Here, the platform was applied to a panel of patient derived
CRC organoids
 A panel of 545 novel antibodies (both synthetic and from
immunization in Merus MeMo™ mice) were screened
 Several hits were identified and validated
 This high content screening platform can be used to screen
panels of therapeutic molecules in panels of patient tissues
from diverse tumour types.