Summary:

Effective options are limited for patients with small-cell lung cancer who develop progressive disease during or after etoposide plus platinum-based therapy. In this issue of Cancer Discovery, Farago and colleagues highlight the data for temozolomide plus olaparib in this patient population and demonstrate the potential to accelerate biomarker discovery through co-clinical trials utilizing patient-derived xenografts.

See related article by Farago et al., p. 1372.

Small-cell lung cancer (SCLC) is an aggressive neuroendocrine tumor characterized by loss of function of p53 and RB (1). Most patients are highly sensitive to initial etoposide plus platinum-based therapy (EP). Unfortunately relapses are nearly universal, and effective options after EP-based treatments are few. Reliable predictors of treatment efficacy have not been developed, despite increasing evidence that biologically distinct subsets of SCLC exist (2).

Topotecan is the only FDA-approved second-line therapy for SCLC. The objective response rate (ORR) to this agent is only 10% to 24%, median progression-free survival (PFS) is 3 to 3.5 months, and median overall survival (OS) is 6 to 8 months (3). PD-1 axis inhibitors (nivolumab with or without ipilimumab and pembrolizumab) have been incorporated into guidelines and FDA-approved as third-line therapies for SCLC. However, these PD-1 axis inhibitors are unlikely to be effective following progression on EP plus PD-1 axis inhibition, now a standard of care in the first-line setting (3). Thus, new therapies are urgently needed.

PARP is highly expressed in SCLC (4). PARP inhibition prevents DNA-damage repair and can increase levels of replication stress beyond what SCLC cells can withstand, thus representing a therapeutic vulnerability in this disease. Unfortunately, PARP inhibitors have shown only minimal activity as monotherapy in previously treated patients and, to date, have not demonstrated improved OS in unselected patient populations when combined with EP or as monotherapy maintenance (1, 3). Recently, we demonstrated in preclinical models that PARP inhibitors activate the cGAS–STING pathway and synergize with PD-1 axis inhibitors (5); however, early clinical data in relapsed SCLC suggest that the benefit of these combinations may be limited to a subset of patients (1, 6).

Temozolomide is an oral alkylating agent with good central nervous system penetration. It has similar efficacy as other single-agent chemotherapies in SCLC (3). Preclinical data suggest that PARP inhibition may potentiate the effect of temozolomide. Consistent with this, we found in a randomized phase II trial a higher ORR with temozolomide plus veliparib (39%) versus temozolomide plus placebo (14%; ref. 7). Although survival was not improved for all comers treated with the combination, IHC for SLFN11 (a protein that causes cells to arrest and undergo apoptosis when DNA is damaged) showed promise for identifying patients who may derive a survival benefit from the addition of a PARP inhibitor.

In this issue of Cancer Discovery, Farago and colleagues present a phase I/II trial of temozolomide and the PARP inhibitor olaparib in previously treated, extensive-stage SCLC. Olaparib plus temozolomide (OT) was well tolerated, with the main adverse events consisting of cytopenias. The recommended phase II dose (RP2D) was determined to be 200 mg olaparib twice daily and temozolomide 75 mg/m2 daily on days 1 to 7 of a 21-day cycle. The ORR at the RP2D was notable at 44% (16/36 patients; ref. 1), especially given that nearly half of the patients had received ≥2 prior lines of therapy and 40% had brain metastases (including 6/15 patients with untreated brain metastases).

For patients treated at the RP2D of OT, the median PFS was 4.2 months and the median OS was 6.7 months. In patients treated at any dose (n = 50), the median PFS was 4.5 months and the median OS was 8.5 months (1). The numerical differences in OS between all patients and those treated at the RP2D may be a product of sample size and the relatively short median follow-up of 7 months. Seventy-three percent of patients treated with OT were platinum-sensitive, a characteristic associated with improved prognosis and with greater PARP inhibitor sensitivity. As expected, patients with platinum-sensitive SCLC had a higher ORR to OT (47% in platinum-sensitive vs. 29% in platinum-resistant). However, survival among patients was not statistically different when comparing platinum-sensitive versus platinum-resistant patients (1).

Other therapies evaluated in a similar setting as OT have historically demonstrated a lower ORR of approximately 5% to 25%, with the exceptions being lurbinectedin and temozolomide plus veliparib (3, 7). Lurbinectedin, an inhibitor of RNA polymerase II–mediated transcription, recently demonstrated promising activity in relapsed SCLC (ORR of 35%, median PFS of 3.9 months, and median OS of 9.3 months; ref. 8). However, unlike the Farago and colleagues trial (which allowed untreated, asymptomatic brain metastases < 1 cm), the lurbinectedin study excluded patients with brain metastases. Unfortunately, information was not available for intracranial ORR to the OT combination or the comparative efficacy of OT in patients with and without baseline brain metastases. However, it is noteworthy that despite including a large number of patients with baseline brain metastases (a poor prognostic indicator), the outcomes with OT were similar in cross-trial comparisons to second-line lurbinectedin (1, 8).

Much of the molecular data on SCLC has been developed in cell lines, genetically engineered mouse models, or a small number of available human tissues (2, 4). More recently, patient-derived xenografts (PDX) have been used to learn about this malignancy. PDXs are beneficial for studying human disease, as they are established directly from human tumors and/or circulating tumor cells (CTC) and represent the heterogeneity of the disease. Furthermore, those PDXs established after disease relapse provide a rare opportunity to study treatment-resistant SCLC.

Farago and colleagues previously demonstrated the ability to rapidly and efficiently generate SCLC PDX models (9). In this study, they leverage their cohort of PDXs to conduct a coclinical trial exploring response to OT and candidate predictive biomarkers. Specifically, 32 PDX models constructed from 22 patients were studied. This included 6 PDXs generated from 4 patients treated with OT. Among the 4 OT-treated patients, only 2 patients had paired models generated before OT and after progression (1). Thus, there were insufficient post-treatment models to propose potential acquired resistance mechanisms to the combination.

Consistent with preclinical studies of SCLC and clinical findings from other tumor types, Farago and colleagues demonstrate in their PDX models a correlation between sensitivity to EP and to OT, as well as overlapping markers of EP and PARP inhibitor sensitivity. These findings suggest that these predictive biomarkers may be more broadly associated with response to certain types of DNA damage, rather than with a specific drug or combination. However, biomarkers may also reflect the dosing regimen and PARP trapping caused by olaparib (because PARP trapping contributes to a cytotoxic, chemotherapy-like effect).

Using a gene set enrichment analysis of RNA profiles from their PDXs, Farago and colleagues found that inflammatory gene signatures, including increased expression of CEACAM1, TNFSF10 (TRAIL), TGIF1, and OAS1, were associated with greater in vivo sensitivity to OT and EP. This finding was further validated in 11 additional PDX models (1). Performance characteristics of these markers suggested that these gene-expression predictors may be more robust than SLFN11, a top candidate biomarker for PARP inhibitor and chemotherapy response (1). Given the limitations of modeling immune responses in mice lacking an adaptive immune system, it is uncertain whether these inflammatory markers will ultimately predict benefit in patients treated with EP or OT. Unfortunately, in this study, there is no information about whether these inflammatory markers predicted clinical efficacy in the OT-treated patients. However, in another recent phase II trial, patients who responded to a combination of olaparib with the anti–PD-L1 antibody durvalumab had pretreatment biopsies showing an inflamed tumor phenotype (6).

On the basis of these results, the authors propose that the inflammatory signature genes (especially those related to IFN signaling or TGFβ) may outperform existing hypothesis-driven biomarkers. This is true for MGMT methylation and PARP1 levels; however, the differences in performance between the inflammatory/IFN signature genes, SLFN11, and epithelial-to-mesenchymal transition (EMT)–related markers may be explained by technical considerations. For example, in the published clinical study investigating SLFN11, it was measured by IHC and categorized as SLFN11-positive (H-score ≥ 1) versus SLFN11-negative. Differences in RNA versus protein levels, cutoff points or methodology for biomarker scoring, or imbalance between patient or PDX cohorts could all contribute to observed differences between the studies. Furthermore, high SLFN11 expression is significantly correlated with increased expression of IFN response pathway genes, including OAS1 in SCLC (10). This association between SLFN11 and IFN response genes suggests that these distinct biomarkers may be different ways of measuring a common and biologically important mechanism of PARP inhibitor or chemotherapy sensitivity. Similarly, the association between EMT regulators [TGFβ and SNAI2 (SLUG)] and OT resistance warrants further study, especially given the established role for EMT in drug resistance and immune escape. Practically speaking, the optimal biomarker needs to be both robust and clinically feasible. Although gene expression signatures may be highly predictive, biomarkers that can be assessed by IHC (which is widely available and used for biomarkers such as PD-L1) may be more feasible to incorporate into practice at this time.

The investigators in this study demonstrate the potential for PDX models to predict patient responses in the setting of a co-clinical trial (1). Although there were only 4 PDXs from patients treated on the OT trial, these models faithfully recapitulated their clinical responses to OT. Given recent efforts to better categorize SCLC into specific subtypes based on key transcription factors, the investigators also explored differences in OT response between subtypes. Although they did not observe a signal of greater sensitivity to OT or EP in a particular SCLC subtype (i.e., ASCL1-high, NEUROD1-high, YAP1, or POU2F3), the analysis was limited by small numbers of PDXs for some subsets (e.g., only one PDX model in the POU2F3 subtype). Additional PDX models representing the full spectrum of SCLC subtypes will benefit future biomarker studies, as will investigation of subset-specific biomarkers in instances where drugs have activity in more than one subtype.

Temozolomide plus PARP inhibition is active in SCLC. Candidate predictors of benefit include SLFN11 overexpression by IHC and an inflammatory gene signature (Fig. 1; refs. 1, 7, 10). However, we still need to improve on these biomarkers to optimize patient selection for PARP inhibitor combinations. Generating PDX models from patients with SCLC and running a co-clinical trial treating these models with OT was innovative and suggests a potentially powerful translational tool to advance drug development. In the future, longitudinal construction of PDX models from samples taken before treatment, at different timepoints during treatment, and at development of progressive disease will further our ability to personalize treatment of SCLC and provide us with information on how resistance develops so that we may better combat it.

Figure 1.

Candidate predictors of benefit to temozolomide and PARP inhibition in SCLC. ITV, initial tumor volume; TMZ, temozolomide; Res, resistant; Sen, sensitive; STING, stimulator of interferon genes (refs. 1, 7, 10).

Figure 1.

Candidate predictors of benefit to temozolomide and PARP inhibition in SCLC. ITV, initial tumor volume; TMZ, temozolomide; Res, resistant; Sen, sensitive; STING, stimulator of interferon genes (refs. 1, 7, 10).

Close modal

J.M. Pacheco is a paid consultant at AstraZeneca, Gerson Lehrman Group, Novartis, Pfizer, and Takeda, and has received speakers bureau honoraria from Genentech and Takeda. L.A. Byers is a paid consultant at AstraZeneca, AbbVie, PharmaMar, SA, Sierra Oncology, and Bristol-Myers Squibb, and has received commercial research grants from AbbVie, AstraZeneca, Tolero Pharmaceuticals, and Sierra Oncology. No other potential conflicts of interest were disclosed.

L.A. Byers acknowledges support from NIH/NCI R01-CA207295, NIH/NCI U01-CA213273, NIH/NCI P5-CA070907, and an Andrew Sabin Family Fellowship.

1.
Farago
AF
,
Yeap
BY
,
Stanzione
M
,
Hung
YP
,
Heist
RS
,
Marcoux
JP
, et al
Combination olaparib and temozolomide in relapsed small-cell lung cancer
.
Cancer Discov
2019
;
9
:
1372
87
.
2.
Rudin
CM
,
Poirier
JT
,
Byers
LA
,
Dive
C
,
Dowlati
A
,
George
J
, et al
Molecular subtypes of small cell lung cancer: a synthesis of human and mouse model data
.
Nat Rev Cancer
2019
;
19
:
289
97
.
3.
Pacheco
JM
,
Bunn
PA
. 
Advancements in small-cell lung cancer: the changing landscape following IMpower-133
.
Clin Lung Cancer
2019
;
20
:
148
60
.
4.
Byers
LA
,
Wang
J
,
Nilsson
MB
,
Fujimoto
J
,
Saintigny
P
,
Yordy
J
, et al
Proteomic profiling identifies dysregulated pathways in small cell lung cancer and novel therapeutic targets including PARP1
.
Cancer Discov
2012
;
2
:
798
811
.
5.
Sen
T
,
Rodriguez
BL
,
Chen
L
,
Corte
CMD
,
Morikawa
N
,
Fujimoto
J
, et al
Targeting DNA damage response promotes antitumor immunity through STING-mediated T-cell activation in small cell lung cancer
.
Cancer Discov
2019
;
9
:
646
61
.
6.
Thomas
A
,
Vilimas
R
,
Trindade
C
,
Erwin-Cohen
R
,
Roper
N
,
Xi
L
, et al
Durvalumab in combination with olaparib in patients with relapsed SCLC: results from a phase II study
.
J Thorac Oncol
2019
;
14
:
1447
57
.
7.
Pietanza
MC
,
Waqar
SN
,
Krug
LM
,
Dowlati
A
,
Hann
CL
,
Chiappori
A
, et al
Randomized, double-blind, phase II study of temozolomide in combination with either veliparib or placebo in patients with relapsed-sensitive or refractory small-cell lung cancer
.
J Clin Oncol
2018
;
36
:
2386
94
.
8.
Paz-Ares
LG
,
Trigo Perez
JM
,
Besse
B
,
Moreno
V
,
Lopez
R
,
Sala
MA
, et al
Efficacy and safety profile of lurbinectedin in second-line SCLC patients: results from a phase II single-agent trial
.
J Clin Oncol
37
:
15s
, 
2019
(
suppl; abstr 8506
).
9.
Drapkin
BJ
,
George
J
,
Christensen
CL
,
Mino-Kenudson
M
,
Dries
R
,
Sundaresan
T
, et al
Genomic and functional fidelity of small cell lung cancer patient-derived xenografts
.
Cancer Discov
2018
;
8
:
600
15
.
10.
Allison Stewart
C
,
Tong
P
,
Cardnell
RJ
,
Sen
T
,
Li
L
,
Gay
CM
, et al
Dynamic variations in epithelial-to-mesenchymal transition (EMT), ATM and SLFN11 govern response to PARP inhibitors and cisplatin in small cell lung cancer
.
Oncotarget
2017
;
8
:
28575
87
.