Novel targeted strategies to overcome resistance in small-cell lung cancer: focus on PARP inhibitors and rovalpituzumab tesirine

ABSTRACT

Introduction: Small-cell lung cancer (SCLC) is a highly aggressive neuroendocrine tumour, and its outcome is strongly conditioned by the rapid onset of resistance to conventional chemotherapeutics. First-line treatment with a combination of platinum agents and topoisomerase inhibitors has been the standard of care for over 30 years, with disappointing clinical outcome caused by early-acquired chemoresistance. In this disheartening scenario, novel treatment strategies are being implemented in order to either revert or bypass resistance mechanisms.

Areas covered: The general mechanism of action of the standard frontline treatment regimens for SCLC, as well as the known resistance mechanisms to these drugs, is reviewed. Moreover, we focus on the current preclinical and clinical evidence on the potential role of PARP inhibitors and rovalpituzumab tesirine (Rova-T) to tackle chemoresistance in SCLC.

Expert opinion: Preliminary evidence supports PARP inhibitors and Rova-T as two promising approaches to either revert or bypass chemoresistance in SCLC, respectively. The identification of potential predictive biomarkers of response to these innovative treatments (SLFN11 and DLL3) has shortened the gap between SCLC and personalized targeted therapy. Further large-scale clinical studies are urgently needed for a better designation of PARP inhibitors and Rova-T in the therapeutic algorithm of SCLC patients.

KEYWORDS:

• Small-cell lung cancer (SCLC)
• chemoresistance
• PARP inhibitors
• Notch pathway
• DLL3
• rovalpituzumab tesirine
• targeted therapy

1. Introduction

Small Cell Lung Cancer (SCLC) is a poorly differentiated neuroendocrine tumour that accounts for 10% to 15% of all lung cancer cases [1]. Tobacco smoking is the leading cause of SCLC, and smoking signature contributes to the high mutational burden observed in this subtype of tumour [2,3]. While targetable genomic aberrations are very rare in SCLC, this tumour almost universally harbours inactivation of TP53 and Rb1 genes [3]. Key features of SCLC include rapid tumour growth, a propensity towards early metastatic dissemination and high genomic instability. Despite being highly responsive to initial chemotherapy and radiotherapy, SCLC is often diagnosed at a late stage and develops resistance to conventional treatments within a few months. Taken together, these characteristics make SCLC the most aggressive lung malignancy, with a 5-year survival rate lower than 7% [4]. Historically, SCLC can be classified according to the extension of the disease rather than the traditional TNM staging used for most of the malignant neoplasms. This classification recognizes limited disease (LD), defined as SCLC confined to one hemithorax, and extensive disease (ED), in the case of spreading beyond ipsilateral hemithorax, including malignant pleural effusion and hematogenous metastases [5]. In the past 30 years, no new effective strategies have been developed for SCLC treatment. Since its first proven efficacy in 1985 [6], platinum-based combination chemotherapy has remained the standard of care for the first-line treatment of SCLC, with the option of concomitant radiation therapy for LD-SCLC. Most patients respond to combined chemotherapy, which consists of cisplatin or carboplatin in combination with etoposide, with an objective response rate (ORR) exceeding 60–70% in patients with ED-SCLC [7]. Platinum compounds are also being combined with irinotecan in the frontline setting [8], even though improvement in ORR and overall survival (OS) over standard platinum-etoposide has not been demonstrated [9,10]. Despite initially-high chemosensitivity, resistance against frontline chemotherapy constitutes a dismal limitation for all the patients [10,11]. Remarkably, a treatment-free interval shorter than 2 months identifies patients who will likely be refractory to salvage second-line chemotherapy and have a poor prognosis [12]. On the other hand, patients who relapse > 2 months after the first-line treatment are classified as sensitive relapse and are thought to be responsive to subsequent chemotherapy, if the physical conditions are suitable. To date, recommended second-line options in SCLC are limited to topotecan monotherapy, whose efficacy is strongly dependent on the duration of response to frontline platinum salts [13]. Irinotecan monotherapy is still another viable option in this setting [8], but data about irinotecan outcomes are limited, because there are country-related differences [14–16]. Nevertheless, in this disheartening scenario, novel therapeutic approaches for SCLC are forthcoming. The PD-L1 inhibitor atezolizumab has achieved the milestone of being the first drug to grant a 2-months OS advantage when combined to standard upfront carboplatin and etoposide in patients with ED-SCLC, with a 30% reduction in the risk of death [17]. Moreover, anti-PD-1 nivolumab ± anti-CTLA-4 ipilimumab and anti-PD-1 pembrolizumab have shown promising activity in relapsed SCLC [18–20]. Besides the results achieved with immunotherapy, novel targeted agents are currently undergoing clinical investigation in SCLC. Through a better understanding of SCLC biology and specific pathways’ aberrations, poli [ADP-ribose] polymerase (PARP) and delta-like canonical Notch ligand 3 (DLL3) have recently been identified as promising drug-gable targets in SCLC [21]. Against this background, innovative targeted approaches with PARP inhibitors and the anti-DLL3 antibody-drug conjugate rovalpituzumab tesirine (Rova-T) have been developed in order to either bypass or reverse resistance to conventional first-line chemotherapy. The aim of this review is to provide an overview of the mechanism of action and general resistance mechanisms to the standard frontline treatment regimens for SCLC, being cisplatin, carboplatin and etoposide, including irinotecan for comprehensiveness, and discuss the current evidence on the ability of PARP inhibitors and Rova-T to tackle chemoresistance in SCLC.

2. Cisplatin and carboplatin

2.1. Cellular mechanism of action of cisplatin and carboplatin

Cisplatin and carboplatin are chemotherapeutic agents that act by causing crosslinking of DNA which interferes with cellular DNA replication and/or repair, lastly resulting in the kill of rapid proliferating cells (Figure 1). In order to generate a cytotoxic response, platinum salts first need to be activated by the replacement of their cis-chloro ligands with water molecules [22]. Besides the similar mechanism of action, that relies on the same adduct structure, cisplatin and carboplatin differ for their leaving group [23]. Once activated, they can interact with N7 position of guanine, leading to DNA adduct formation, after which interstrand and intrastrand crosslinks are established [24]. Intrastrand adducts are responsible for most of the cytotoxic effect of cisplatin [25], accounting for 85–90% of total lesions [26]. Following the DNA damage induced by platinum salts, several downstream effects are documented.

Figure 1. Mechanism of action of platinum drugs, topoisomerase inhibitors, PARP inhibitors and rovalpituzumab tesirine. Platinum derivatives (cisplatin, carboplatin) form DNA adducts that in turn cause crosslinking of DNA, thus interfering with DNA replication and/or repair. Etoposide is a TOPIIA inhibitor, while irinotecan is a TOPIB inhibitor. Topotecan shows a similar effect but is less potent as SN-38. PARP inhibitors interfere with DNA damage repair by blocking PARP enzyme. Rovalpituzumab tesirine is an antibody-drug conjugate that conveys a toxin inside DLL3-expressig SCLC cells and activates Notch signalling.

The cisplatin-induced DNA damage mainly activates cell cycle checkpoints. This event induces a transient cellular S-phase arrest, and later a durable G2/M-phase arrest, caused by the inhibition of Cdc2-cyclin A or B kinases [27]. Normally, cell cycle arrest is associated with an inhibitory effect on the cytotoxicity of cisplatin as it is needed to enable nucleotide excision repair (NER) complex to eliminate the formed adducts and promote cell survival. NER can indeed recognize the DNA adduct via XPA protein, which leads to the recruitment of proteins to form a complex. ERCC2 and ERCC3 helicases dissociate the two DNA strands, after which ERCC4 and ERCC5 endonucleases can cleave the damaged strand at 3ʹ and 5ʹ sides. This enables POL-δ and POL-ε to resynthesize the excised strand. Ligase III can next re-ligate the newly synthesized DNA to the main DNA chain [23]. When cisplatin induces extensive damage to the DNA and the repair via NER is incomplete, apoptosis will occur [28].

Cisplatin can also directly activate extracellular-regulated kinases (ERK), c-Jun N-terminal kinases (JNKs) and p38 kinases, which constitute key proteins within the mitogen-activated protein kinase (MAPK) pathway [29]. Noteworthy, cisplatin induces the strongest activation of p38 kinase, as compared to other chemotherapeutics, such as doxorubicin and paclitaxel [30]. The cisplatin-induced activation of MAPK cascade leads to integration of extracellular signals to regulate cell proliferation, differentiation, survival and apoptosis [31,32]. Moreover, cisplatin has been found to indirectly activate the tumour suppressor p53, by stimulation of ATM and ATR kinases, which are known to phosphorylate p53 at serine-15 [33]. Activated p53 can transactivate several genes associated with cell cycle arrest, DNA repair and apoptosis, such as GADD45A [34]. In turn, Gadd45a protein enhances NER activity via association with proliferating cell nuclear antigen (PCNA) [35].

When DNA damage induced by cisplatin is extensive, apoptosis is favoured. Apoptosis is mediated via translocation of Bax from the cytosol to the mitochondria, leading to release of apoptogenic factors, which activate the caspase 9-caspase 3 pathway [36,37]. This process is regulated by the Bax/Bcl-2 ratio. If this ratio is increased, apoptosis will be induced, since Bcl-2 is the antiapoptotic counterpart of Bax [38]. Cisplatin-induced apoptosis can also be mediated via the interaction of Fas with FasL, which activates the caspase 8-caspase 3 pathway [37,39]. In addition, cisplatin can induce accumulation of p73, which is a nuclear p53-related protein that acts as a pro-apoptotic factor. Of note, it has been shown that cellular accumulation of p73 is dependent on a proficient mismatch repair (MMR). When the cells are deficient in MMR proteins, cisplatin-induced accumulation of p73 does not occur [40]. In addition, a proficient MMR is critical for triggering the apoptotic cascade after the recognition of cisplatin-induced DNA damage [41].

2.2. Cellular mechanism of resistance to cisplatin and carboplatin

Overall, resistance to platinum drugs can emerge due to prevention of cisplatin interaction with DNA, interference with DNA damage signals to prevent apoptosis, or both. Resistance is generally multifactorial, involving several mechanisms at the same time in one tumour cell [42].

One of the main mechanisms of resistance to cisplatin relies on the reduction of intracellular drug accumulation, due to inhibition of drug uptake, increase of drug efflux, or both [28]. Cisplatin efflux can be mediated by multidrug resistance-associated protein 2 (MRP2), which is an ABC-family membrane protein found to be overexpressed in cisplatin-resistant cells [43], but this has not been confirmed in later studies. Similarly, ATP7A and ATP7B, copper-transporting ATPase proteins, are overexpressed in cisplatin-resistant cells and mediate cisplatin efflux [44]. Additionally, the copper transporter CTR1 plays a role in the uptake of platinum drugs by cells. CTR1 normally regulates the cellular absorption of copper, but CTR1 can also mediate uptake of platinum compounds. Deletion of the CTR1 gene results in a reduced intracellular drug accumulation and therefore an increased drug resistance [45,46].

Cisplatin resistance can also occur due to increased drug inactivation by thiol-containing molecules, such as glutathione (GSH), that reduce the amount of cisplatin available to interact with DNA [26]. Elevated expression of the γ-GCS (γ-glutamylcysteine synthetase) gene results in the increase of GSH, and is pointed out as a resistance mechanism [47]. GSTπ (GSH-S-transferase π) [48], which catalyses the conjugation of active GSH with cisplatin, is also increased in platinum-resistant tumour cells [49]. Moreover, GSH can increase DNA repair or inhibit apoptosis to further mediate resistance to cisplatin [26]. This resistance mechanism has been investigated in SCLC cell lines and xenografts. Kurokawa et al. [50]investigated a γ-GCS transfected SCLC cell line, SBC-3/GCS, in which platinum levels were reduced and GSH levels were increased compared to the control SBC-3 cell line. Similarly, Meijer et al. [51]found increased levels of GSH in the cisplatin resistant SCLC cell line GLC₄-CDDP. Other studies also found higher levels of GST in cisplatin resistant cell lines or xenografts [52]. These results suggest that GST and GSH play a role in the resistance to cisplatin in SCLC. However, not all studies showed positive results. Campling et al. [53]found that GSH levels are not associated with cisplatin resistance in SCLC cell lines. GSH and GST probably play a role in resistance to cisplatin in patients with SCLC, but the extent is not yet clear.

Another redundant mechanism of resistance to platinum drugs consists of the increased repair of cisplatin-induced DNA damage. This is mainly exerted by NER, whose activity is increased in resistant cells, as documented by the enhanced ERCC1 and XPA expression [28]. An increased ERCC1 expression was associated with a shorter survival in patients with NSCLC [54], but its role as a predictive biomarker has never been fully validated. A recent study implicated p53 as a potential confounding variable via promoting a compensation for the loss of ERCC1 through a modulation of error-prone mechanisms affecting interstrand crosslinks [55]. Among the proteins responsible for DNA damage repair, MMR deficiency is associated with cisplatin resistance, by impairing the ability of the cells to detect both the damage and activate apoptotic cascade [56]. On the other hand, HMG1 can protect the DNA adducts from being recognized by DNA repair machinery, thereby increasing cisplatin sensitivity [57].

Lastly, aberrations within the apoptotic cascade can lead to platinum resistance. TP53 mutations in exons 4 to 9 account for most of the genetic aberrations causing the loss of the cellular ability to induce apoptosis in response to platinum drugs. Gene mutation results in the inability of p53 to bind to DNA or transactivate p53-dependent genes, such as BAX [58]. Wild-type TP53 can also result in resistance when cisplatin-dependent pathways that activate p53 are impaired [59]. Besides p53, XIAP and survivin are anti-apoptotic molecules that are commonly overexpressed in cisplatin-resistant tumour cells [60]. These inhibitory molecules affect the activation of caspases, which are critical in cisplatin sensitivity [28]. However, the anti-apoptotic protein bcl-2 can also prevent apoptosis by inducing an increase in GSH [61]. Bcl-2 is overexpressed in 80% of SCLC specimens [62]. In vivo studies showed a remarkable preclinical activity of ABT-263 (Navitoclax; ABT-263), which is a potent inhibitor of both Bcl-2 and Bcl-x(L), in multiple SCLC xenografts [63]. However, the same drug did not show clinical activity in a phase IIa trial in patients with recurrent and progressive SCLC after at least one prior therapy. These controversial results might be explained by the use of preclinical models which did not recapitulate SCLC phenotype, since primary xenografts, derived from direct transfer of human SCLC into recipient mice, showed low-level activity of the closely related parental drug, ABT-737 [64]. These results suggested the need of further studies on biomarkers which should be used to guide treatment with BCL-2 inhibitors in SCLC as well as in other tumor types [65,66].

3. Topoisomerase inhibitors: etoposide and irinotecan

Topoisomerases are enzymes that provide relaxation of supercoiled DNA and unwind DNA helices and strands, which in turn help DNA replication [67]. Six different types of DNA topoisomerases exist in human cells, and they can be divided in three subgroups: IB enzymes (TOP1 and TOP1mt), IIA enzymes (TOP2α and TOP2β) and IA enzymes (TOP3α and TOP3β). Subgroups IA and IB belong to topoisomerase type 1 enzymes, which are involved in single-strand cleavage of DNA. Subgroup IIA constitutes the topoisomerase type 2 enzyme, which cleaves two strands of DNA. Topoisomerase I and II inhibitors work synergistically with platinum compounds, due to their ability to enhance the inhibition of DNA elongation [68], and are routinely used in combination with platinum salts for the treatment of SCLC.

3.1. Cellular mechanism of action of etoposide and irinotecan

Etoposide is a TOPIIA inhibitor, which can cleave two strands of DNA and form a ternary TOPIIA-DNA-drug complex. This means that it can prevent re-ligation of the DNA strands, causing double-strand breaks in the DNA. A build-up of double-strand breaks will ultimately lead to cell death (Figure 1) [69].

Topotecan and Irinotecan are TOPIB inhibitors that primarily prevent re-ligation after topoisomerase IB cleavage. Irinotecan is a prodrug that is activated into 7-ethyl-10-hydroxycamptothecin (SN-38) by carboxylesterases 1 and 2 (CES 1/2) in the liver. Irinotecan can also be inactivated in the liver by cytochrome P450 isoforms 3A4 and 3A5 (CYP3A4 and CYP3A5) by degradation to 7-ethyl-10-[4-N-(5-aminopentanoic acid)-1-piperidino] carbonyloxycamptothecin (APC) and 7-ethyl-10-(4-amino-1-piperidino) carbonyloxycamptothecin (NPC) metabolites. Once active, irinotecan forms a ternary TOP1B-DNA-drug complex, that causes a misalignment in the DNA, leading to an accumulation of single-strand breaks in the DNA. These single-strand breaks alone are not toxic to the cell, because they can re-ligate after removal of the drug. The collision of the DNA replication fork together with the ternary complex induced by irinotecan cause irreversible double-strand breaks that eventually lead to cell death [70].

3.2. Cellular mechanism of resistance to etoposide and irinotecan

Resistance to etoposide can occur due to different mechanisms. The most frequent cellular mechanism of resistance relies on the decreased expression of TOPIIA, which is the direct target of etoposide. TOPIIA expression was indeed reduced in etoposide-resistant cell lines (MCF-7/1E and MCF-7/4E), as shown by PCR analysis [71]. In the same study, the activity of TOPIIA was not significantly decreased, but the resistance was induced by the alteration in etoposide binding site [71]. Increased expression of MRP1 (ABCC1), which is a gene that encodes for one of the ATP Binding Cassette (ABC) transporters, can also mediate efflux of etoposide from the cells, thus leading to decreased drug accumulation and drug resistance [71]. Overexpression of ABC-transporters as a resistance mechanism to etoposide has been widely studied in SCLC cell lines. Overexpression of the MRP and MDR1 gene in etoposide resistant SCLC cell lines with reduced levels of drug accumulation, suggest that ABC-transporters play a role in the resistance to etoposide [72–75]. Moreover, downregulation of MMR genes has been pointed out as an additional mechanism of resistance to etoposide. Among MMR genes, MSH2 levels were found to be decreased in etoposide-resistant cells. Other studies suggested that the effects of the decreased expression of MLH1 were more important. In the presence of a deficient MMR machinery, etoposide-induced apoptosis is impaired due to a reduced recognition of the damaged DNA [76]. Moreover, cells that harbour a decrease or deficiency in the expression of MMR genes will ultimately gain point mutations, including mutations in TOPIIA [77].

Resistance to irinotecan can be caused by either reduction of intracellular drug accumulation by increased activity of the above-mentioned P-glycoprotein (MDR1, ABCB1), alteration in the target or impaired drug metabolism. In particular, prevention of formation of irinotecan ternary complexes can be due to TOPOI mutations, reduced expression of TOPOI, post-translational alterations of TOPOI, and interaction of TOPOI with other proteins [78]. Regarding irinotecan metabolism, an increase in CYP3A4/CYP3A5 activity and a decrease in CES 1/2 activity are associated with insensitivity to irinotecan, resulting in increased inactivation and decreased drug activation, respectively. Moreover, a crucial step into irinotecan metabolism is the glucuronidation of the active metabolite SN-38 by uridine diphosphate glucuronosyltransferase 1As (UGT1As), such as UGT1A1, which inactivates SN-38 by forming SN-38 glucuronide (SN-38G) [79]. In this regard, polymorphisms in UGT1A1 have an influence on the efficacy and toxicity of irinotecan treatment. UGT1A1*1 is the wild-type allele and comprises six thymine-adenine (TA) repeats in the promoter region of UGT1A1, the TATA box. The number of TA repeats determines the polymorphism of the UGT1A1 gene. Indeed, an increased number of TA repeats is associated with a reduced activity of the UGT1A1 enzyme, which is the case for UGT1A1*37 with eight TA repeats, whereas UGT1A1*36 is proficient with five TA repeats [80]. Proficient UGT1A1 polymorphisms, such as UGT1A1*1 are typically associated with insensitivity to irinotecan, due to an increased inactivation of SN-38. On the contrary, UGT1A1*28 and UGT1A1*6 are associated to reduced catalytic activity of the enzyme, resulting in increased accumulation of active irinotecan metabolites that can ultimately cause severe toxicities, such as neutropenia [81,82]. Compared to patients with the UGT1A1*1 wild type allele, homozygous UGT1A1*28 patients had significantly higher systemic exposure to SN38 and a lower plasma SN38-G/SN-38 ratio [80,83].

4. Novel strategies to tackle chemoresistance in SCLC

Taking into account the multitude of cellular resistance mechanisms to conventional chemotherapeutics that cause early onset of clinical relapse/progression of SCLC, novel pharmacological approaches need to be implemented in this highly lethal malignancy. To date, innovative strategies are aimed at either reversing or bypassing the chemoresistance, by selectively targeting proteins involved in resistance or alternative signalling pathways, respectively. In this scenario, PARP inhibitors showed potential to reverse resistance in SCLC, while Rova-T demonstrated preliminary clinical activity in relapsed SCLC because of its targeted effects and potential to activate an alternative tumour suppressor pathway (Notch pathway). In the subsequent sections the mechanisms of action of these two classes of drugs are summarized as well as current evidence of their potential in the clinical management of SCLC.

4.1. PARP inhibitors: mechanism of action and preclinical evidence

PARP is a critical enzyme involved in the process of DNA repair, and it catalyses the synthesis of poly(ADP-ribose) (PAR) chain after recognition of single-strand DNA breaks [84]. PARP binds strongly to DNA-strand breaks and blocks the access of DNA-repair enzymes to the DNA-strand breaks. After auto-modification of PAR protein, PARP unbinds the DNA and allows other DNA-repair enzymes to detect the PAR chain and trigger the repair process [84]. Since the increased repair of drug-induced DNA damage constitutes a relevant mechanism of resistance to standard chemotherapy, it was reasoned that PARP inhibitors could be an effective strategy to re-sensitize cells or avoid the onset of resistance. Remarkably, from solid malignancies, SCLC cell lines showed the highest expression of PARP1 mRNA. Moreover, compared to NSCLC tumours or normal lung, SCLC tumours had significantly higher rates of PARP1 mRNA (p = 0.005) [85]. In SCLC cell lines, PARP1 showed the highest relative levels among all DNA repair proteins. When SCLC cell lines (H69, H82, H524) and the NSCLC cell line (A549) were treated with the PARP inhibitor AZD2281 (olaparib) for 24 hours, the SCLC cell lines showed the highest drug sensitivity and reduction of PAR levels. In the same research study, AZD2281 was further evaluated in combination with irinotecan in H82 cells, showing a significant decrease in tumour cell viability compared to irinotecan alone, and indicating a synergistic effect of the addition of a PARP inhibitor to this TOP1B inhibitor. Interestingly, immunohistochemical (IHC) staining showed that PARP1 expression was present in SCLC specimens. These results suggest that PARP1 could be a potential drug target in SCLC [85].

Regarding other preclinical evidence of PARP inhibitors in SCLC cell lines (Table 1), the efficacy of talazoparib in combination with etoposide/carboplatin has been tested on 63 SCLC cell lines, but only 10–15% of the SCLC cell lines had an increased response to this combination compared to etoposide/carboplatin [86]. More promising results were obtained by Owonikoko and collaborators [87] on veliparib, in which the combination of veliparib, cisplatin and etoposide was more effective than etoposide and cisplatin in vivo, providing a solid rationale for further development of this combination in humans. Interestingly, the concentration of veliparib in the tumour was two-fold higher than the plasma concentration [87]. Lallo and colleagues [88] demonstrated that the combination of olaparib and the WEE1 kinase inhibitor AZD1775 is a therapeutic option for SCLC, mice with chemosensitive SCLC tumours receiving this combination therapy showed complete and durable tumour regressions, with five out of seven mice having a 1-year tumour free survival [88].

Table 1. Overview of PARP inhibitors implemented in the clinic.

Drug	Specificity	IC50	Company	Dose	FDA indication
Olaparib (AZD-2281)	PARP1 and PARP2	5 nM (PARP1) 1 nM (PARP2)	AstraZeneca	300 mg/day	First-line maintenance treatment of BRCA-mutated ovarian cancer Pre-treated HER2-negative germline BRCA-mutated metastatic breast cancer
Talazoparib (BMN-673)	PARP1 and PARP2	0.57 nM (PARP1)-	Pfizer	1 mg/day	Pre-treated HER2-negative germline BRCA-mutated metastatic breast cancer
Rucaparib (DB12332)	PARP1, PARP2 and PARP3	Ki = 1.4 nM	Clovis Oncology	600 mg/day	Maintenance treatment of recurrent BRCA-mutated ovarian cancer
Niraparib (MK-4827)	PARP1 and PARP2	3.8 nM (PARP1) 2.1 nM (PARP2)	Tesaro	300 mg/day	Maintenance treatment of recurrent ovarian cancer
Veliparib (ABT-888)	PARP1 and PARP2	5.2 nM (PARP1) 2.9 nM (PARP2)	AbbVie	-	-

4.2. PARP inhibitors: clinical implication for the treatment of SCLC

An overview of the available clinical evidence of PARP inhibitors in SCLC is summarized in Table 2. First data came from two phase I clinical trials, in which veliparib was tested in combination with conventional platinum-based chemotherapy [89,90]. In these studies, the addition of veliparib to the platinum-etoposide combination was safe, with the most common adverse events being haematological, followed by nausea and fatigue [89,90]. Moreover, both studies showed promising results in efficacy, with good documented responses [89,90]. De Bono and collaborators explored talazoparib monotherapy in 110 patients with advanced germline BRCA1/2 mutations and selected sporadic cancers, among which 23 had ED-SCLC [91]. In this early-phase trial, talazoparib demonstrated single-agent antitumour activity and was well tolerated [91]. Veliparib was further tested in two multicentric placebo-controlled randomized phase II trials. Owonikoko and colleagues [92] evaluated the addition of veliparib to cisplatin-etoposide as first-line treatment in patients with ED-SCLC. This trial met its primary endpoint and showed a significant improvement in PFS with the addition of veliparib (Hazard Ratio [HR], 0.63, p = 0.01). Moreover, a trend towards a better OS was documented in the veliparib arm, but this was not statistically significant (10.3 versus 8.9 months; HR, 0.83, p = 0.17) [92]. Interestingly, a subgroup analysis demonstrated that males with elevated LDH benefit more from the addition of veliparib compared to the other subgroups in terms of PFS [92].

Table 2. Clinical studies of PARP inhibitors in SCLC.

Reference	Study population	Phase	Treatment arm(s)	Primary outcome(s)
Owonikoko et al. (87)	9 patients, 8 of which had ED-SCLC	I	Cisplatin + etoposide + veliparib	RP2D: 100mg/day; Tolerable safety profile Median OS: 9.0 months
Atrafi et al. (89)	39 patients, 25 of which had ED-SCLC	I	Carboplatin + etoposide + veliparib	RP2D: 240mg/day Tolerable safety profile ED SCLC patients: CR/PR in 64% (16/25) Median PFS in RP2D group (n = 13): 5.6 months
De Bono et al. (91)	110 patients, 23 of which had ED-SCLC	I	Talazoparib	RP2D: 1.0 mg/day Tolerable safety profile Median PFS: 11.1 weeks
Pietanza et al. (93)	104 ED-SCLC patients, among which 4 were treatment-naïve	II	Arm A: Temozolomide + veliparib Arm B: Temozolomide + placebo	Arm A: 4-months PFS 36%, median PFS 3.8 months, median OS 8.2 months, ORR 39% Arm B: 4-months PFS 27%, median PFS 2.0 months, median OS 7.0 months, ORR 14%
Owonikoko et al. (92)	128 ED-SCLC patients	II	Arm A: Cisplatin + etoposide + veliparib Arm B: Cisplatin + etoposide + placebo	Arm A: PFS 6.1 months, OS 10.3 months, ORR 72% Arm B: PFS 5.5 months, OS 8.9 months, ORR 66%

Abbreviations: CR, complete response; ED-SCLC, extensive disease-small-cell lung cancer; ORR; overall response rate; OS, overall survival; PFS, progression-free survival; PR, partial response; RP2D, recommended phase 2 dose.

Veliparib efficacy has also been assessed in combination with temozolomide, a DNA-methylating agent, in recurrent SCLC patients enrolled in a placebo-controlled randomized phase II study by Pietanza and collaborators [93]. In this study, the ORR was significantly higher when veliparib was added to temozolomide compared to temozolomide monotherapy (39% vs 14%, p = 0.016), but no differences were observed both in terms of 4-months PFS (36% vs 27%, p = 0.19) and OS (8.2 vs 7.0 months, p = 0.50) [93]. In the same trial, the investigators explored the IHC expression of SLFN11, which is a protein that blocks replication as a reaction to replication stress, by binding to chromatin [94]. Interestingly, the authors found that patients treated with veliparib and temozolomide with SLFN11-positive tumours had significantly prolonged PFS (5.7 months vs. 3.6 months, p = 0.009) and OS (12.2 months vs. 7.5 months, p = 0.014) compared to patients negative for SLFN11- [93]. This result demonstrated that high levels of SLFN11 contribute to a higher sensitivity to PARP inhibitors [95]. High levels of SLFN11 mRNA were also associated significantly with talazoparib response in SCLC patient-derived xenografts (PDXs) [96], and IHC analysis of protein expression showed similar results [95,96]. Moreover, low expression of SLFN11 in SCLC cell lines was associated with resistance to PARP inhibitors [95,96]. These results indicate that SLFN11 could potentially be used as a biomarker for selecting SCLC patients who will likely benefit from treatment with PARP inhibitors. Several trials with PARP inhibitors, both in monotherapy and in combination, are currently ongoing and they are summarized in Table 3.

Table 3. Ongoing studies of PARP inhibitors in SCLC.

Study title	Clinical phase	Condition	Experimental drug
Study in patients with SCLC of veliparib in combination with topotecan (NCT03227016)	Phase I	SCLC	Veliparib in combination with topotecan
Study of talazoparib, a PARP inhibitor, in patients with advanced or recurrent solid tumours (NCT01286987)	Phase I	Breast Neoplasms Ovarian Cancer Ewing Sarcoma SCLC Prostate Cancer Pancreas Cancer	Talazoparib
A Phase I/II study of MEDI4736 in combination with olaparib in patients with advanced solid tumours. (MEDIOLA) (NCT02734004)	Phase I/II	Ovarian Cancer Breast Cancer SCLC Gastric Cancers	MEDI4736 in combination with olaparib and MEDI4736 in combination with olaparib and bevacizumab
Trial of CRLX101, a nanoparticle camptothecin with olaparib in people with relapsed/refractory small cell lung cancer (NCT02769962)	Phase I/II	Solid Tumours SCLC NSCLC Lung Neoplasms	CRLX101 in combination with olaparib
Talazoparib and low-dose temozolomide in treating patients with relapsed or refractory extensive-stage small cell lung cancer (NCT03672773)	Phase II	Recurrent Extensive Stage SCLC Refractory Extensive Stage SCLC	Talazoparib in combination with temozolomide
Temozolomide with or without veliparib in treating patients with relapsed or refractory small cell lung cancer (NCT01638546)	Phase II	Recurrent SCLC	Temozolomide in combination with veliparib or placebo
A phase 2 study of cediranib in combination with olaparib in advanced solid tumours (NCT02498613)	Phase II	Pancreatic Cancer Breast Cancer NSCLC SCLC	Cediranib maleate in combination with olaparib
A study of niraparib as maintenance therapy following first-line platinum-based chemotherapy in patients with extensive stage small cell lung cancer (NCT03516084)	Phase III	Extensive Stage SCLC	Niraparib or placebo

4.3. Rova-T: mechanism of action and preclinical evidence

Rova-T is injected intravenously and is an antibody-drug conjugate made of a humanized monoclonal antibody (SC16) directed against DLL3 that acts by attacking a DNA-damaging pyrrolobenzodiazepine dimer toxin SC-DR002 (D6.5) in DLL3-positive tumour cells. Since SCLC is frequently characterized by high expression of DLL3, Rova-T has selectivity against these tumours, by delivering and thus internalizing the toxin only in DLL3-expressing cells [97–99]. In humans, DLL3 is important for promoting neurogenesis and inhibiting gliogenesis in the human brain [100], and it is a crucial inhibitor of Notch pathway [100,101]. Importantly, Notch pathway activation shows opposite effects in SCLC and NSCLC, acting as a tumour suppressor as well as a tumour promoter, respectively [102]. The seminal study by George and colleagues [3] investigated mutations in 110 SCLC tumours via genome sequencing. Inactivating mutations were documented in NOTCH1, and in the majority of tumours (53/69, 77%) high levels of ASCL1 were found. ASCL1 is a lineage oncogene of neuroendocrine cells and can be inhibited by an active Notch signalling pathway [3]. This suggests that SCLC tumours commonly have a suppressed Notch signalling pathway, part of which can be accounted for high DLL3 expression.

DLL3 is often highly expressed in SCLC tumours [97–99]. In a study performed on 63 tissue specimens from SCLC patients, 52 (83%) were DLL3-positive, according to IHC staining. Moreover, SCLC cell lines showed a higher DLL3 expression than NSCLC cell lines, suggesting that DLL3 is preferentially expressed in SCLC tumours [99]. Furthermore, DLL3 is expressed on the surface of SCLC cells, making it a potential target for anti-DLL3 antibody-drug conjugate therapy [98,99]. Efficacy of Rova-T has been demonstrated in vivo in multiple SCLC PDX models, and the drug was able to eliminate cancer stem cells in this preclinical model [99]. It has also been supposed that DLL3 downregulation to evade Rova-T could result in Notch pathway reactivation, which in turn ultimately leads to slowed tumour growth [98].

4.4. Rova-T: clinical implication for the treatment of SCLC

The first phase I clinical trial that evaluated the tolerability and preliminary efficacy of Rova-T in SCLC patients was conducted by Rudin and collaborators (Table 4) [103]. This dose-escalating study included 82 patients, of which 74 were pretreated SCLC patients, all of whom received at least one dose of intravenous Rova-T ranging from 0.05 mg/kg to 0.8 mg/kg every three-six weeks, followed by 0.3–0.4 mg/kg every six weeks or 0.2 mg/kg every three weeks. Noteworthy, in 47% of these cases, Rova-T was administered as a third-line treatment. The maximum tolerated dose was 0.4 mg/kg every three weeks and dose-limiting toxicities occurred at the dose of 0.8 mg/kg every 3 weeks, including grade 4 thrombocytopenia and liver function test abnormalities. The recommended schedule for further investigation was set at 0.3 mg/kg for two cycles every six weeks. Among the 65 SCLC patients evaluable for the objective response, 41 (68%) achieved disease control (11 patients had partial response and 30 patients obtained a stability of disease). Median duration of response in 65 assessable patients was 5.6 months (95% CI, 2.5–8.3) and median PFS was 2.8 months (95% CI, 2.5–4.0). Concerning toxicities, grade 3 or worse treatment-related adverse events occurred in 38% of patients, and the most frequent groups included thrombocytopenia (12%), pleural effusion (11%) and skin reactions (8%). A retrospective IHC analysis for DLL3 was performed on the available tissues, and an exploratory threshold of 50% DLL3-positive tumour cells was used to divide samples into DLL3-low (< 50%) and DLL3-high (≥ 50%). Remarkably, disease control rate was higher among DLL3-high compared to DLL3-low patients (88% vs 50%, respectively), and a trend towards a better OS was observed in DLL3-high population (5.8 vs 2.7 months). These results strongly suggested that DLL3 expression could be used as a biomarker for Rova-T treatment [103].

Table 4. Clinical studies of Rova-T in SCLC.

Reference	Study population	Phase	Treatment arm(s)	Primary outcome(s)
Rudin et al. (103)	74 ED-SCLC patients	I	Rova-T	RP2D: 0.3 mg/kg for two cycles every 6 weeks DCR: 68% (50% in DLL3-low, 88% in DLL3-high) ORR: 18% (0% in DLL3-low, 38% in DLL3-high) mPFS: 2.8 months (2.2 months in DLL3-low, 4.3 months in DLL3-high) mOS: 2.2 months in DLL3-low, 5.8 months in DLL3-high
Carbone et al. (104)	339 ED-SCLC patients	II	Rova-T	ORR: 18% (19.7% in DLL3-high) mPFS: 3.8 months in DLL3-high mOS: 5.6 months (5.7 months in DLL3-high)

Abbreviations: DCR, disease control rate; ED-SCLC, extensive disease-small-cell lung cancer; ORR; overall response rate; OS, overall survival; PFS, progression-free survival; RP2D, recommended phase 2 dose.

In the subsequent open-label, single arm, phase II trial (TRINITY), DLL3-positive expression was used as an inclusion criteria, and DLL3-high expression was defined as ≥ 25% tumour cells expressing DLL3 by IHC [104]. In this study, the activity of Rova-T in DLL3-expressing SCLC patients after failure to ≥ 2 previous line of treatment, including at least one treatment with platinum-based chemotherapy, was explored [104]. However, the results of this trial were disappointing, as documented by an ORR of 18% in the overall population, and Abbvie decided not to pursue accelerated approval of the experimental drug in the third-line setting [105]. When evaluating DLL3-high patients, median PFS and OS were 3.8 and 5.6 months, respectively. This subgroup of patients had a higher benefit from Rova-T treatment compared to non-DLL3-high, both in terms of ORR (16% vs 6%), best overall response (24% vs 14%) and clinical benefit rate (72% vs 57%) when the data were analysed by an independent review committee. Regarding DLL3-high group, patients who were treated in third line were the most likely to respond to Rova-T. The toxicity profile of Rova-T was in line with the results of the phase I study, with ≥ grade 3 treatment-related adverse events documented in 40% of patients. Noteworthy, ten treatment-related deaths were documented [104].

Following the results of Rova-T in heavily pretreated SCLC patients, a phase III trial comparing Rova-T with topotecan for the second-line treatment of patients with DLL3-high (≥ 75%) ED-SCLC was performed (TAHOE). Disappointingly, the independent data monitoring committee has stopped further enrolment in this trial, on the basis of a detrimental effect of Rova-T on OS compared to topotecan. Among the ongoing Rova-T trials, a randomized placebo-controlled phase III study of Rova-T as a maintenance therapy after first-line chemotherapy is currently recruiting patients (NCT03033511).

5. Expert opinion

Resistance to conventional chemotherapy is almost inevitable in SCLC and strongly reduces patients’ prognosis. Although the high proliferation rate of this rapidly progressing malignancy makes it responsive to standard chemotherapeutics, this efficacy is further limited by the onset of acquired resistance. Standard regimens used for the first-line treatment of SCLC consist of cisplatin/carboplatin, etoposide and irinotecan. Resistance to these compounds is heterogeneous and mainly mediated by reduced drug accumulation, increased drug inactivation, increased DNA damage repair, aberrations in drug metabolism and inhibition of apoptosis. Therefore, drug discovery has focused on development of new treatment strategies that would be able to overcome SCLC mechanisms of resistance. One of the strategies to counteract chemotherapy resistance relies on the inhibition of the DNA repair system. Therefore, PARP inhibitors appear to be potential candidates to either enhance the efficacy of DNA-damaging agents or resensitize resistant cells. Indeed, early clinical studies showed that the combination of PARP inhibitors with platinum compounds, as well as temozolomide, was effective and safe in ED-SCLC [90,92,93]. However, due to the early phase of the trials and the only modest benefit documented, these results should be interpreted with caution, and further large-scale trials are needed to strengthen these findings. More interestingly, the identification of SLFN11, a key protein that regulates response to DNA damage has been found to predict response to PARP inhibitors in SCLC in vitro and in vivo. This might open new horizons in the proper selection of patients. Furthermore, we can speculate that future studies will need to evaluate PARP inhibitors efficacy in SCLC patients based on this specific biomarker.

Besides the promising activity of PARP inhibitors in tackling the DNA repair machinery, another developmental pathway dysregulated in SCLC is the Notch pathway. DLL3 surface expression is high in SCLC and is related to a strong Notch pathway inactivation. Therefore, DLL3 expression could be used to selectively apply an antibody-drug conjugate to SCLC cells in order to exert its cytotoxic function, thus bypassing resistance. Rova-T indeed successfully achieved this goal in vitro and in vivo. However, the debatable efficacy results in pre-treated SCLC patients and the recent halt of the phase III TAHOE trial raised doubts about the further implementation of Rova-T in the management of SCLC. Moreover, the drug-related toxicities should be taken into account, especially considering the rapid deterioration of clinical performance status in SCLC patients along the natural course of disease. In the same perspective, the role of DLL3 IHC expression as a predictive factor to Rova-T treatment needs more elucidation and large-scale evaluation.

In conclusion, preliminary studies of both PARP inhibitors and Rova-T supported their potential in (partially) overcoming chemoresistance in SCLC. Unfortunately, the promising preclinical profile could not yet be translated to the clinic. Survival advantages for these drugs (and combinations) are limited to a few months at most, similar to immunotherapy. In order to show efficacy, immunotherapy needs to be combined with conventional chemotherapy. However, it seems, despite considerable advances in recent years, that the understanding of the biology of SCLC, and the mechanisms of resistance in patients are still poorly understood. The role of the driving forces in SCLC needs further characterization in order to develop treatment algorithms of SCLC that will ultimately lead to improved treatment with prolonged survival. and/or higher response rates with acceptable toxicity. In addition, the identification of two novel biomarkers, SLFN11 and DLL3, has potential for personalized targeted therapy for SCLC, and will guide the clinicians to properly design further investigational studies.

Article highlights

• Small-cell lung cancer (SCLC) is an aggressive disease characterized by a rapid onset of resistance to chemotherapeutics and a 5-year overall survival rate lower than 7%;

• Resistance mechanisms to conventional cytotoxic drugs (platinum analogs and topoisomerase inhibitors) mainly rely on reduced drug accumulation, increased drug inactivation, increased DNA damage repair, aberrations in drug metabolism and inhibition of apoptosis;

• PARP inhibitors and rovalpituzumab tesirine have shown promising preclinical and clinical activity in either reverting or bypassing chemoresistance, respectively;

• Large-scale randomized trials and a better characterization of biomarkers that can predict response to these novel treatments are largely awaited.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This work was supported by Italian Association for Cancer Research (AIRC grant IG2017-20074 to M Tiseo, and Start-Up grant to E Giovannetti).