Mean %

BR

CA

or

RA

D51expre

ssio

n

(%B

RC

A o

r R

ad51/b

eta

-actin)

A.

****

****

**** p<0.0001, *** p:0.0009, ** p:0.0024, * p:0.0467 paired t-test

**

*

*email for more information: asli.x.muvaffak@gsk.com

Elaborating the Mechanism of PARP Synthetic Lethality Following ATM Loss in DLD-1 Cell Line

Asli Muvaffak*, Kevin G. Coleman

GlaxoSmithKline, Synthetic Lethality Research Unit, Waltham, MA, US

● ATM (-/-) KO in DLD-1 cell line caused a marked increase in niraparib sensitivity (80-fold), and induced

downregulation of Rad51, indicative of impairment of HR-capacity in DLD-1 ATM (-/-) KO cell line.

● ATM biallelic mutants are particularly sensitive to niraparib in NSCLC PDX models, and niraparib sensitivity

extends beyond BRCA1/2 genes and mainly driven by ATM, BAP1, & MRE11A biallelic loss in NSCLC PDX

models.

● Loss of ATM is associated with decreased BRCA1/2 expression in both DLD-1 ATM (-/-) and ATM biallelic

mutant NSLC PDX models. Preliminary data indicates ATM (-/-) loss leads to inhibition of phosphorylation of

MRN complex proteins in DLD-1 ATM (-/-) KO cell line, suggestive of downstream impact on HR-impairment

is regulated through BRCA1 and BRCA2 genes in DLD-1 ATM (-/-) KO cell line.

● The presenting author, Asli Muvaffak & co-author Kevin G Coleman are both employees of GSK

● We would like to express special thanks to Olivia Li & Cindy Liu from Applied Stem Cell, Inc. for their contributions on the

generation of the DLD-1 BRCA2 and ATM (-/-), (+/-) CRISPR/Cas9 KO cell lines, Amin Maier, Thomas Metz, Tobias

Deigner, Konstantin Lashuk from Oncotest/CRL for supporting the in vitro & in vivo pharmacology studies, and Joanna

Gawden-Bone and Simon Scrace from Horizon Discovery for performing the Rad51/H2AX high content imaging assay.

Background and Highlights of the Findings

Methods

Conclusions

References

Acknowledgements

Presented at the AACR 2020 Virtual Meeting, June 22-24th, 2020

Abstract No. 1699, Session Title: Basic Mechanisms of Genome Integrity

PARP (Poly ADP-Ribose Polymerase) inhibitors have transformed the treatment for ovarian cancer,

which is known to have high incidence of homologous recombination (HR) deficiency through the

loss of BRCA1 or BRCA2 genes. 30-35% of ovarian cancer patients have defects in BRCA1/2 genes

caused by germline, somatic or epigenetic alterations. PARP inhibitors induce cell death in HR-

deficient cells through the mechanism of synthetic lethality where cancer cells cannot tolerate the loss

of both SSB (Single Strand Break) and DSB (Double Strand Break) repair machinery1-2. Although

genetic or epigenetic alterations in other pathways such as Fanconi anemia, or in non-authentic

homologous recombination repair (HRR) genes such as PTEN, IDH, and CDK12 have been

reported to contribute to HR deficiency, mechanistic understanding of HR loss remains to be

determined. A key DNA damage repair (DDR) regulator protein, ATM, a kinase that phosphorylates

and activates major DNA damage checkpoints, i.e. CHK2 & BRCA1, has being increasingly

recognized recently as a synthetic lethal partner with PARP inhibitors in multiple tumor indications

such as prostate, lung and colorectal cancers3.

Non-Small Cell Lung Cancer (NSCLC) is the leading cause of cancer related mortality worldwide,

and accounts for 80-85% of all lung cancer diagnoses. While targeted therapies against epidermal

growth factor receptor (EGFR) mutations and anaplastic lymphoma kinase (ALK) fusions have

proven to be effective for the treatment of NSCLC, only a fraction of NSCLC patients (i.e. <20%)

benefit from these targeted agents. Additionally, although immunotherapy approaches, specifically

PD-1:PD-L1 blockade, appear to be more broadly efficacious, there are still a substantial proportion of

patients who may not benefit from these treatments3-5. To investigate the potential of targeting the

DNA Damage Response (DDR) pathway in lung cancer, as an alternative therapeutic approach for

these patients, we sought to identify whether ATM loss could be synthetically lethal with niraparib

monotherapy in NSCLC PDX models, and in ATM KO cell lines. Approximately 3% of lung cancers

harbor mutations in ATM (i.e. 3.5% in lung adenocarcinomas & 1.4% in lung squamous cell

carcinoma), which is involved in HRR as well as multiple other DNArepair and checkpoint functions.

ATM functions in both cell cycle arrest and DNA repair in response to DSBs, however how cells

respond to inhibition of PARP in the absence of ATM is not yet fully understood. In this study, we

aimed to identify PARPi synthetic lethality in ATM loss, i.e. via generating an ATM KO by

CRISPR/Cas9 in DLD-1 cell line which was tested in the 3D spheroid and 2D colony formation

assays, and characterize HR-deficiency in ATM KO (-/-) cell lines using a high-content

immunofluorescence imaging assay. We demonstrated that loss of ATM was associated with low

BRCA1 and BRCA2 protein expression suggestive of downstream effect on impairment of HR-

mediated DNA checkpoint signaling. Our preliminary data indicated that loss of ATM leads to

inhibition of phosphorylation of MRN (Mre11-Rad50-Nbs1) complex proteins which is

required for ATM activation and downstream phosphorylation of p53, BRCA1, and CHK2.

This data provided additional evidence suggestive of downstream effect on impairment of HR-

mediated DNArepair in ATM loss is regulated through BRCA1 and BRCA2 genes.

Our findings highlight that PARPi synthetic lethality phenotype observed in tumors with ATM loss

could be regulated through the effect coming from the loss of ATM and a subsequent effect on

BRCA1/2 expression. ATM has a critical role in DNA repair and in synthetic lethality with PARP

inhibitors.

Method I: Niraparib single agent activity was evaluated in 10 lung cancer PDX models with bialleic mutations in a

set of clinically relevant HRR genes. Niraparib was administered at 50mg/kg dose over a period of 28 days or

longer orally, once daily. Tumor growth was monitored twice per week.

Method II: Colony formation assay: 500 µL cell suspension was seeded in 24 well plates (PerkinElmer, cat. no.

1450-605) and treatment with 7 point dose titrations of niraparib, i.e. 50 μM to 50 pM (1:10 dilution) and vehicle(DMSO) were incubated at 37°C, 5% CO2 for 14 days. Media with compounds were replaced every 3-4days/week.

CellMaskTM deep red plasma membrane stain (ThermoFisher, cat. no. C10046) was used for cell labeling & image

acquisition was performed using IN Cell Analyzer 2200 (GE Healthcare).

Method III & IV: HR (homologous repair) capacity for DNA DSB (Double Strand Break) repair was assessed using

a high content immunofluorescence imaging assay to quantify nuclear Rad51 and H2AX-foci in BRCA2 & ATM

deficient DLD-1 cell lines after inducing DNA SSBs following hydrogen peroxide treatment and a 3µM niraparib

treatment for 18hrs; WB analysis of HR function, Rad51 and H2AX protein levels in DLD-1 BRCA2(-/-) KO, DLD-1

ATM (-/-) KO, ATM/BAP1/MRE11A/XRCC2/RAD51D mutant NSCLC PDX models were analyzed from vehicle and

niraparib treated tumors at termination .

Results I. Homozygous Loss of ATM Causes a Marked Increase in the

In Vitro & In Vivo Sensitivity of Tumor Cells to Niraparib

▪ In Vitro: Niraparib showed a 80-fold difference in sensitivity in ATM (-/-) cell line

compared to the DLD-1 parental

Niraparib Activity Summary (CFA, 14d)

Niraparib

IC50 (M)

hBRCA2

(-/-) KO

hATM

(-/-) KO

hATM

(+/-) KO

6.33x10 -9 (B4) 19.25x10-9 2.93x10-6

▪ In Vivo: Niraparib monotherapy inhibited tumor growth in DLD-1 ATM (-/-) KO

tumor xenograft model

▪ Niraparib in vivo activity correlates well with its in vitro activity from 3D-

clonogenic assay in DLD-1 BRCA2 (-/-), ATM (+/-) & (-/-) KO cell lines

Figure 2. Dose response curves for niraparib in BRCA2 (-/-), ATM (-/-), & ATM (+/-) KO cell lines and niraparib sensitivity summary and sensitivity ranking using 2D colony formation assay (A),

niraparib in vivo activity in DLD-1 hBRCA2 (-/-), DLD-1 hATM (-/-) KO & DLD-1 hATM (+/-) KO cell lines (B), Comparison of In vitro vs in vivo activity of

niraparib in DLD-1 BRCA2 and ATM (+/-), (-/-) KO cell lines (C).

A.

DLD-1 BRCA2 (-/-) ATM (-/-) ATM +/-)

TGI: 88% TGI: 60% TGI: 20%Inactive

Niraparib Activity

In Vitro vs In Vivo

B. C.

DLD-1 BRCA2 & ATM

(-/-) HOMO KO

Results IV. Differing Degrees of Niraparib Sensitivity is Observed in ATM

Biallelic Mutant NSCLC PDX Models

Niraparib In Vivo Activiity in HRR Bialleic Mutant

NSCLC PDX ModelsATM Expression is Lost in ATM

Bialleic Mutants

n=6

Non-ATM HRR_Bi ATM_Bi

* *PDX ID: #1 #2 #3 #4 #5 #6

p53 -/- -/- WT del(-/0) WT -/- del(-/0) -/- WT -/-

ND: Not Determined

Del(-/0): deleted, hemizygous

▪ ATM deficiency is associated with markedly increased sensitivity against PARP inhibition

in both p53 deficient and proficient lung adeno- & squamous cell carcinoma PDX models

A.

Figure 3. WB analysis of HR function: H2AX and Rad51 protein levels in DLD-1 BRCA2 (-/-) & DLD-1 ATM (-/-) KO, from niraparib treated in vivo tumors at termination

(A) in vitro Rad51-foci & H2AX-foci assay using DLD-1 BRCA2 (-/-) & DLD-1 ATM (-/-) KO cells, vehicle & niraparib treated samples (B).

In vitro ▪ ATM loss abrogates DSB repair and induces HR-deficiency in DLD-1 ATM (-/-)

Vehicle

Niraparib

(3µM)

Rad51 foci

H2AX foci

Hoechst

hATM (-/-) BRCA2 (-/-)WTDLD-1

Merged Images of Rad51 & H2AX foci

ATM is the primary DDR target that phosphorylates H2AX in response to DSBs, and its loss leads to poor H2AX foci in ATM -/- cell line (Burma, S. et al, J Biol Chem 2001)

Results II. HR Capacity is Impaired in Both DLD-1 hATM (-/-) & hBRCA2 (-/-)

HOMO KO Cell Lines

**

(n=2)(n=3)

*

Mean

%

H2A

X e

xp

ressio

n

(%H

2A

X /

beta

-acti

n)

In vivo

**** p<0.0001, ** p<0.0096, * p:0.036, unpaired t-test

▪ Rad51 foci formation is inhibited in hATM (-/-) & hBRCA2 (-/-) cell lines

(n=2)(n=3)

Mean

% R

AD

51exp

ressio

n

(%R

ad

51/b

eta

-acti

n)

*Rad51 levels normalized to Dld-1 parental cell line levels

**** **

A.

In vivo activity, Tumor Growth Inhibition% (TGI%): {1-(∆T/∆C)}*100; Response Criteria: According to NCI standards, a T/C ≤ 42% is the minimum level of anti-tumor activity & a T/C <10% is considered a

high anti-tumor activity level; In vitro activity (3D clonogenic), TGI%: {1-(TIC50/CIC50)}*100, where TIC50: niraparib IC50 on DLD-1 ATM KO or BRCA2 KO, CIC50: niraparib IC50 on DLD-1 parental)

1. Lord, C. J. & Ashworth, Nat. Rev. Cancer 2016; 16: 110–120.

2. Lord, C. J. & Ashworth, A. Nature 2012; 481: 287–294.

3. Schmitt A, Knittel G., Welcker D., Yang T.P., George J., Nowak M., Leeser U., Büttner R., Perner S., Peifer M., Reinhardt H.C. Cancer

Res. 2017; Jun 1;77(11):3040-3056.

4. Burma S., Chen B.P., Murphy M., Kurimasa A., Chen D.J. J Biol Chem 2001; 276 (45): 42462–42467.

5. Birkelbach, M. et al. J Thorac Oncol 2013; 8 (3): 279-286.

An inverse relationship is observed between BRCA1/2 expression levels and niraparib

sensitivity in ATM biallelic mutant NSCLC PDX models

Figure 5. Niraparib monotherapy in vivo activity summary in HRR mutant NSCLC PDX models (A), & ATM expression profiling in ATM biallelic mutant PDX models

by WB analysis (B), Correlation analysis of tumor growth inhibition (TGI%) and downregulation of BRCA1/2 expression in 5ATM biallelic mutant PDX models; TGI%:

{1-(∆T/∆C)}*100, following niraparib monotherapy at 50mg/kg PO, QD for 28 days or longer, and baseline BRCA1 and 2 protein expression profiling by WB analysis;

%Rad51 expression in PDX samples by WB analysis, compared to levels in Dld-1 parental cell line (%Rad51/beta-actin), (C).

Tu

mo

r G

row

th In

hib

itio

n,

TG

I (%

)

PDX #6 #5 #4 #3 #2

30% ND ND ND 37%

% B

RC

A1

or B

RC

A2

ex

pre

ss

ion

(%B

RC

A/b

eta

-ac

tin)

**** p<0.0001, *** p:0.0009, ** p:0.0024, * p:0.0467 paired t-test

Preliminary data indicates loss of ATM leads to inhibition of phosphorylation of MRN complex proteins, i.e. RAD50 &

NBN, in DLD-1 ATM (-/-) cell line and provides further evidence on impairment of HR-mediated DNA repair checkpoint

signaling in ATM loss and is suggestive of regulation through BRCA1 and BRCA2 genes in this cell line

Figure 4. Change in BRCA1, BRCA2 & RAD51 protein levels (by WB analysis) following vehicle and niraparib treatment (50mg/kg PO, QD 28 days or longer) in tumors from

DLD-1 ATM (-/-) KO cell line tumor model, n=6 mice/arm (A); %pRad50(Ser638)/Rad50 and pNBN (Ser343)/NBN expression in Dld-1 hATM HOMO KO cell line by WB analysis,

compared to levels in Dld-1 parental cell line (%pRad50(Rad50)/beta-actin), (n=1), (B) and (%pNBN(NBN)/beta-actin), (n=1) (C), (Niraparib at 3µM & Etoposide at 40µM).

B.

4.4-fold

1.6-fold

C.

21-fold

2.7-fold

Results III. Loss of ATM is Associated with Low BRCA & Rad51 Expression, and

Inhibition of Phosphorylation of MRN Complex Proteins in DLD-1 ATM (-/-) KO Cell Line

B.

B.

C.

Role of ATM in DNA Damage Response & Double Strand Break Repair

Yoshida, K and Miki, Y Cancer Sci 2004; Brandsma, I et al. Expert Opin Investig Drugs. 2017

Pathway Overview

Figure 1. DDR pathway overview, and the role of key checkpoint regulators, ATM, ATR, and DNA-PK in DSB repair processes.

Yoshida, K and Miki, Y Cancer Sci 2004; Brandsma, I et al. Expert Opin Investig Drugs 2017

ATM mediated phosphorylation of BRCA proteins is required for proper functioning of homology directed repair of DSBs