In Vitro Evaluation of PARP1 Inhibitor Olaparib‐Cyanine Dye Conjugate for the Treatment of Glioblastoma

Abstract

Inhibition of DNA damage repair is a common mechanism to induce cancer cell death and, therefore, to treat cancers such as breast, prostate and ovarian cancer. This approach, although very attractive, has not been feasible for treating brain tumours due to the stringent structural features necessary for pharmaceutical agents to cross the blood–brain barrier (BBB). This study presents the proof of concept in vitro cellular activity of PARP1 (Poly(ADP‐ribose) polymerase 1) enzyme inhibitor olaparib conjugated with a known BBB‐crossing heptamethine cyanine dye. We also present evidence for cellular uptake and PARP1 inhibition by the conjugate, which contributes to the observed improvement in potency.

1. Background

Glioblastoma (GBM), despite significant clinical advancements, remains a near‐fatal diagnosis for patients, leading to immense emotional and financial strain on their families. The standard of care for GBM patients is similar in most countries, with survival rates stagnating at ≈15 months for over 30 years since the approval of Temozolomide (TMZ) as the primary chemotherapy treatment. Despite years of research to identify effective therapies for GBM, treatments other than TMZ have not shown substantial benefits for the patient population. Drugs such as platinum‐based agents [1], nitrosoureas [2], and more recently, bevacizumab [3] have been effective in some patient populations, but none of them provide a meaningful extension of overall survival compared to TMZ alone. Tumour‐treating fields (TTF), when combined with TMZ therapy, have shown to extend overall survival by ≈3 months [4]. TTF is not available in all countries, and even if it is made available, it may not be viable for all patients due to the high cost of treatment for the marginal increment in survival that the treatment provides [5]. The modest survival benefit afforded by TTF also makes it difficult for national health agencies to subsidise the treatment.

A shift in how we approach GBM drug development is critical, especially given the disease's unique features and differences from other cancers. The aggressive and infiltrative nature of GBM makes cancer therapies successful in treating other cancers, mostly redundant when applied to GBM [6, 7]. Surgery followed by radiation to remove residual tumour left behind, followed by chemotherapy, cannot eliminate all cancer cells, eventually leading to recurrence and progression of disease, leading to untimely death. Therefore, treatments tailored to take into account the exclusive features of the disease are urgently required. Medicine development also has to take into account the presence of the blood–brain barrier (BBB) and the blood–tumour barrier (BTB), which restricts the delivery of many therapeutic agents. Most chemotherapies are not designed to cross these barriers, thus negating the cancer‐killing abilities of these agents [8]. One could argue that the leaky nature of BTB allows the passage of chemotherapy agents into the tumour tissues, but the presence of efflux pumps considerably impacts the therapeutic efficacy of drugs, as most of the drugs are eliminated from the tumour tissue, reducing their bioavailability [9]. Thus, the introduction of structural features and in vitro screening of compounds in the early stages of drug discovery to ascertain the BBB and BTB permeability is crucial for the successful translation of treatment into humans at a later stage, as overcoming these barriers directly correlates with the efficacy of treatment.

Research teams, including ours, have developed a specialised class of heptamethine cyanine dyes (HMCDs) designed to target brain tumours selectively with the added advantage of crossing the BBB. These dyes are reported to have high specificity for tumour cells and, hence, could be used as a drug delivery agent in the context of various cancers, including brain cancers, as evidenced by their use in animal models of the disease [10, 11, 12, 13, 14, 15]. Animal studies showed preferential uptake of HMCDs in brain tumours but not normal tissues without inducing systemic toxicity [16]. We also elucidated that specific organic anion‐transporting polypeptides (OATPs) subtypes, such as OATP1A2, OATPIC1, OATP2B1, are overexpressed in GBM and play a pivotal role in the uptake of HMCDs by brain tumours [17]. The chemical conjugation of various anticancer agents with these dyes and testing these conjugates in mouse models of the disease has shown effective and enhanced tumour burden reduction compared to the anticancer agent alone [18, 19, 20, 21, 22]. One such drug‐dye conjugate has undergone a phase 1 clinical trial for treating solid tumours and lymphoma and has entered phase 2 trials [23, 24]. These previous findings have provided strong evidence showing the prospect of using HMCDs in developing targeted therapies by conjugating them to various therapeutic agents.

Poly (ADP‐ribose) polymerase 1 (PARP1) is an enzyme involved in DNA damage repair mechanisms, particularly in the repair of single‐strand breaks through the base excision repair pathway and, therefore, targeting PARP1 inhibition resulting in the accumulation of DNA damage in tumour cells, promoting cell death has emerged as a beneficial strategy with multiple FDA approved drugs targeting PARP1 for treating breast, ovarian cancer, and prostate cancer [25]. Overactivation of PARP1, leading to excessive DNA repair, is reported to contribute to the resistance of GBM cells to chemotherapy and radiation [26, 27]. Targeting PARP1 has emerged as a therapeutic strategy in GBM, especially with PARP inhibitors like olaparib, which can block its repair function [28]. PARP1 inhibitors are also trialled in combination with other therapies in vitro and in vivo to overcome GBM's resistance mechanisms, potentially leading to improved treatment outcomes in the future [29]. A recent literature described the conjugation of PARP1 inhibitor rucaparib to HMCD dye MHI‐148, with mouse studies showing preferential uptake of this conjugate in an orthotopic GBM tumour model [30]. Although the potency of the conjugate was significantly less than the parent drug rucaparib due to the interruption of the interaction of the drug (when conjugated to the dye) with amino acids in the catalytic pocket of PARP1, the tumour‐specific accumulation shows the potential of using such conjugates in the context of GBM treatment. Thus, molecular design strategies that preserve the interaction of the drug with the catalytic pocket of PARP1 could result in conjugates with greater potencies than the parent drug.

2. Results

Olaparib is a PARP inhibitor approved for the treatment of various cancers, including breast cancer with a BRCA1 or BRCA2 mutation [31]. Olaparib has undergone clinical trials in recurrent GBM patients and has shown promising safety and tolerability profiles that warrant its further development [32]. Therefore, we hypothesised attaching olaparib to HMCD dye with the intention to improve its tumour specificity and BBB penetrability. Analysis of the published crystal structure data of olaparib in the catalytic pocket of the PARP1 enzyme revealed key interactions that need to be maintained to retain its PARP1 inhibitory activity when attaching to the dye (Figure 1) [34]. The olaparib molecule was modified at the solvent‐exposed region to preserve its hydrophobic interaction with the protein and for conjugation to HMCD dye (Scheme 1(a)). MHI‐148 was chosen as the HMCD carrier for conjugation due to its known BBB crossing abilities and non‐toxicity [12]. The modified olaparib analogue B is reported to exhibit PARP1 inhibition similar to that of olaparib [35]. Briefly, commercially available carboxylic acid 2 was coupled with Boc protected piperazine using standard peptide coupling techniques to afford 3, which was subjected to acid deprotection to obtain olaparib analogue B. This analogue was coupled with indolinium salt 4 and used as an intermediate to afford the drug‐dye conjugate 1 as shown in Scheme 1(b) using a protocol previously reported by our group [36].

FIGURE 1.

Interaction of olaparib with various amino acids in the catalytic pocket of PARP1 [33].

SCHEME 1.

Synthesis of (a) olaparib analogue (b) Conjugate 1.

The cytostatic activity of these compounds was tested with commonly utilised commercial GBM cell lines, U87 MG and KNS42. Conjugate 1 demonstrated superior cytostatic activity (IC₅₀ values) relative to MHI‐148 and olaparib alone in U87 MG cell line, leading to a significantly lower overall cell number (EC₅₀ values) by the 96‐hour time point (Table 1, Figure S1). Notably, MHI‐148 exhibited minimal activity, only reaching maximal effect until concentrations as high as 100 µM, indicating that the dye itself does not significantly contribute to a reduction in cell numbers, which is corroborated by similar studies published previously [36]. Similarly Olaparib displayed minimal activity in the cells, failing to reach maximal effect even at highest 100 µM concentration tested. Conjugate 1 showed an improvement to single digit micromolar potencies, suggesting that the conjugation of olaparib to the dye significantly improved its cytostatic ability. The difference in potency of conjugate 1 between the two cell lines likely stems from biological differences: U87 MG represents adult GBM biology, whereas KNS42 represents paediatric glioma biology, which differ in signalling pathways and drug responsiveness [37].

TABLE 1.

Relative potency of conjugate 1 and controls with and without radiation pretreatment.

	No radiation		3 Gy radiation
	EC 50, μM	IC 50, μM	EC 50, μM	IC 50, μM
MHI‐148	>100	>100	>100	>100
Olaparib analogue (B)	>100	>100	>100	>100
Conjugate 1 (U87 MG)	7.588	1.544	8.412	1.677
Conjugate 1 (KNS42)	63.89	19.62	—	—

Values derived from vehicle‐normalised U87 MG and KNS42 potency curves, n = 3.

Given that olaparib works by inhibiting the PARP enzyme involved in DNA damage repair, we hypothesised that inducing DNA damage through ionising radiation might enhance the cytotoxic effect of conjugate 1. To test this, U87 MG cells were exposed to a 3 Gy dose of radiation, a dose comparable to that administered to GBM patients [38], for 4 h prior to treatments with serial dilutions of MHI‐148, olaparib, and conjugate 1, and fixed 96 h later. However, no significant changes in IC₅₀ values were observed following irradiation, indicating that the DNA damage induced by a single 3 Gy dose was insufficient to enhance the potencies of these compounds under these experimental conditions.

Representative images at 4 µM of these compounds visually demonstrate that conjugate 1 had the greatest impact on cell proliferation (Figure 2). The impact of radiation is modest, with a generally lower cell count and proliferation, but no significant enhancement of the compound's activity was detected (Figure S1c).

FIGURE 2.

Representative images of cytostatic effects of conjugate 1 with and without radiation. 3 Gy radiation pretreatment on U87 MG cells at 4 µM and 96‐h time point. Cell nuclei (blue) and proliferating EdU‐positive cells (red). Scale bar = 100 µM.

To assess the cellular uptake of conjugate 1, we performed single‐cell flow cytometry analysis using U87 MG cells. The intrinsic fluorescence of conjugate 1 enabled direct detection of its internalisation without the need for additional labelling. Following incubation at 500 nM, conjugate 1 was promptly taken up by the cells and remained detectable for up to 96 h, indicating persistent intracellular retention (Figure S2). Importantly, maximal fluorescence intensity was observed as early as 6 h post treatment. This intensity remained stable throughout the 96‐h observation period, suggesting that conjugate 1 is not only efficiently internalised but also retained within the cells over an extended period of time. These findings support the potential of conjugate 1 for prolonged intracellular activity and may contribute to the observed improvement in potency (Figure S2).

To determine the activity of conjugate 1, we employed an in vitro auto‐ADP‐ribosylation (ADPr) assay using recombinant PARP1. In the presence of NAD⁺, PARP1 undergoes auto‐polyADPr modification, which was inhibited by olaparib and conjugate 1 in a dose‐dependent manner (Figure 3A). The dye alone (MHI‐148), on the other hand, remained inactive at the tested concentration (10 µM). It has been shown that histone‐PARylation factor 1 (HPF1) forms a complex with PARP1 and switches its activity towards serine ADP‐ribosylation on histones [33, 37]. Moreover, a recent study showed that the potency of FDA‐approved PARP inhibitors directly correlates with their dissociation from the PARP1/HPF complex [38]. Thus, we evaluated the effect of conjugate 1 on PARP1/HPF1‐mediated ADP‐ribosylation of the histones within mononucleosome substrates. We observed that conjugate 1 inhibited PARP1‐HPF1‐mediated ADP ribosylation on H3 histones in a dose‐dependent manner (Figure 3B).

FIGURE 3.

Conjugate 1 exhibits dose‐dependent activity against recombinant PARP1 and PARP1/HPF1. (A) Immunoblot analysis of PARP1 auto‐ADPr reactions with the indicated compounds). (B) Immunoblot analysis of PARP1/HPF1 ADPr reaction with nucleosomes with the indicated compounds. For both assays n = 2 experimental replicates were performed. MHI‐148 was used at 10 µM.

To determine whether conjugate 1 inhibits PARP1's ADPr activity in cells, we employed the hydrogen peroxide‐induced DNA‐damage assay in the U87 MG cells. We titrated olaparib and conjugate 1 in cells and measured the levels of PARP1 automodification by immunoblot upon hydrogen peroxide challenge. Conjugate 1 exhibited a dose‐dependent inhibition of PARP1 automodification in the cells (Figure 4A) with an IC₅₀ value of 27 nM, which was about 5‐fold less than what we observed for olaparib (IC₅₀: 5 nM) (Figure 4C). Additionally, conjugate 1 was also effective in inhibiting the PARP1‐mediated modification of histones. Importantly, we see no reduction in PARP1 activity in cells by MHI‐148 at the highest tested concentration (10 μM) in line with our biochemical experiments.

FIGURE 4.

Conjugate 1 inhibits PARP1‐mediated ADPr in U87 MG cells following hydrogen peroxide challenge. (A) Immunoblot analysis of ADPr levels on PARP1 following hydrogen peroxide treatment of U87 MG cells; n = 3 independent biological replicates. (B) Immunoblot analysis of ADPr levels on acid‐extracted histones following hydrogen peroxide treatment of U87 MG cells; MHI‐148 was used at 10 µM, n = 2 independent biological replicates. (C) Quantification of the data shown in panel A.; error bars represent mean ± SEM.

3. Discussion

In this study, we designed and synthesised a novel drug‐dye conjugate 1 by covalently linking the PARP inhibitor olaparib to the HMCD MHI‐148. The rationale for this design was guided by structural insights obtained from the crystal structure of olaparib bound within the catalytic pocket of the PARP1 enzyme. This structural information enabled us to intentionally select an appropriate site on olaparib for conjugation to the dye, ensuring that the modification would not interfere with its binding affinity or inhibitory activity. In our previous study, we reported an HMCD–PARPi conjugate in which the PARP inhibitor rucaparib was linked to the HMCD dye IR786 via chloro displacement of the IR786 dye. Although this conjugate exhibited an EC₅₀ of 128 nM, this apparent potency may be partly attributable to the inherent cytotoxicity of the IR786 dye itself (EC₅₀ = 1.7 µM), thereby limiting further development of this conjugate [15]. In contrast, a literature‐reported HMCD conjugate of rucaparib with the dye MHI‐148 was considerably less cytostatic (IC₅₀ = 12 µM), indirectly supporting the notion that IR786 contributes substantially to the observed potency of the rucaparib–IR786 conjugate [30]. Accordingly, we selected MHI‐148 for the present study, as it is far less cytotoxic (EC₅₀ > 100 µM), with the expectation that it would contribute minimally to the observed potency of conjugate 1. Olaparib was chosen based on encouraging clinical trial safety profiles in GBM patients treated in combination with standard chemotherapy [39, 40]. Olaparib is reported to possess poor BBB penetration [41]; therefore, conjugation of olaparib to an HMCD dye to enhance BBB penetration is expected to increase its bioavailability within the brain parenchyma, ultimately improving therapeutic efficacy. As noted above, the drug was conjugated to the carboxylic acid side chain of the dye via a solvent‐exposed region of the drug, to ensure its interaction with the PARP enzyme's catalytic pocket.

Following synthesis, we evaluated the biological activity of conjugate 1 in the extensively utilised U87 MG GBM cell line. Comparative cytostatic assays revealed that conjugate 1 exhibited significantly greater potency in GBM cells compared to olaparib alone. In contrast, the dye MHI‐148 displayed minimal cytostatic activity, underscoring the application of this class of dyes as viable drug carrier molecules.

To assess whether radiation could enhance the efficacy of the compounds, we exposed U87 cells to a single 3 Gy dose of ionising radiation for 4 h prior to treatment with either conjugate 1 or olaparib. However, this pretreatment did not result in a significant increase in cytotoxic and cytostatic activity for either compound. This suggests that a single 3 Gy dose, while clinically relevant, may be insufficient to induce the level of DNA damage required to sensitise cells to PARP inhibition, as well as the 96 h timeframe tested may have missed the impact of radiation if the effects were transient. These experiments, however, are exploratory, as we have only attempted a limited time exposure (4 h) and therefore do not imply that PARP inhibition cannot enhance radiotherapy in GBM. Extending irradiation duration or using fractionated radiation regimens with adjusted time points may yield different outcomes.

To assess PARP1 interaction of our compounds, ADP‐ribosylation assays were employed, which confirmed that both olaparib and conjugate 1 indeed inhibited recombinant PARP1 in a dose‐dependent manner. Further validation using a hydrogen peroxide‐induced DNA damage assay demonstrated that conjugate 1 also inhibited PARP1 activity in U87 MG cells (both automodification and ADPr of histones). However, conjugate 1 displayed lower inhibition levels compared to olaparib, suggesting that the potent cytostatic activity observed with the compound is not solely driven by PARP1 inhibition and alternative mechanisms may also be present. PARP enzyme assays were conducted in cell‐free systems; therefore, a carrier moiety such as an HMCD would not be expected to enhance activity in these assays. Consistent with this expectation, compound 1 exhibits reduced PARP inhibitory activity relative to the parent drug. In contrast, in the U87 cell line, the conjugate displays greater efficacy. This discrepancy appears to arise from a mechanism independent of PARP1 ADP‐ribosyltransferase inhibition. One possibility is that the conjugate more effectively traps PARP1, modulates additional target(s), or alters PARP1 function through an as yet unappreciated mechanism. Notably, we assessed ADP‐ribosylation activity only under H₂O₂‐induced conditions, and it is therefore possible that other PARP1‐dependent processes contribute to the observed effects. As such, our experiments primarily rule out the most prominent and well characterised PARP1 associated mechanisms, while leaving open the possibility of alternative modes of action. Therefore, we explicitly acknowledge the possibility that a component of the observed biological activity may be independent of PARP1 inhibition. We utilised the intrinsic fluorescence of the MHI‐148 dye to monitor the cellular uptake and retention of conjugate 1 using flow cytometry. The results demonstrated efficient uptake of the conjugate by U87 MG cells, with peak intracellular fluorescence observed at 6 h post‐treatment. The fluorescence signal persisted for up to 96 h, indicating prolonged intracellular retention. This prolonged retention time most likely contributes to the enhanced potency of conjugate 1 relative to olaparib, which is known to have a shorter intracellular half‐life [42].

Large, cationic drug‐dye conjugates are known to interact with cellular membranes, mitochondria, and other intracellular compartments in a manner that can produce cytotoxic effects unrelated to the intended molecular target [43, 44]. As such, off‐target or PARP1‐independent mechanisms cannot be excluded at this stage. This represents an important limitation of the current study, as the relative contributions of PARP‐dependent versus PARP‐independent mechanisms to the overall cytotoxic phenotype remain unclear. Future work will therefore be required to decouple target engagement from nonspecific toxicity, including comparative studies in PARP1‐deficient GBM knockout cell lines. Until such studies are completed, the increased efficacy of the conjugate should be interpreted as resulting from enhanced delivery and potentially additional PARP1 independent effects.

The findings presented here support the potential of drug‐dye conjugates such as conjugate 1 as a promising strategy for targeted GBM therapy. While this study primarily focused on the synthetic feasibility and in vitro characterisation of the conjugate, future work will involve a thorough evaluation in multiple patient derived GBM cell lines and in vivo models, including patient‐derived xenografts. As a relatively large, cationic drug‐dye conjugate, compound 1 will potentially face several pharmacokinetic and biodistribution challenges, including rapid plasma clearance and potential accumulation in tissues such as the liver and kidneys. Despite these possibilities, majority of reported drug dye conjugates accumulate in the tumour tissues compared to other organs. These are inferred from previous studies on drug‐dye conjugates [18, 20, 45]. The cationic nature of the HMCD moiety may promote nonspecific interactions with serum proteins and cellular membranes, potentially impacting systemic exposure and distribution. In addition, steric bulk and polarity are likely to limit passive diffusion across the BBB, necessitating reliance on the neo vasculature and the brain interstitium. These factors suggest that conventional PK parameters for drug‐dye conjugates based on HMCDs may differ from those of the parent drug.

To address these challenges and establish in vivo relevance, we will test the conjugate in orthotopic glioma models that closely recapitulate BBB disruption and tumour microenvironment features observed clinically. Initial studies will be conducted in orthotopic U87 xenograft models, enabling direct comparison with in vitro efficacy data obtained in this study. Compound 1 will be administered via systemic routes (intravenous and intraperitoneal) to determine circulation stability and brain delivery. Biodistribution and brain tumour uptake will be quantified using the intrinsic near‐infrared fluorescence of the HMCD dye, enabling whole‐body imaging using the IVIS system as reported previously [12]. Pharmacodynamic engagement will be assessed by measuring PARP pathway modulation in tumour tissue, including biomarkers of PARP inhibition and DNA damage response activation (e.g., PAR levels, γH2AX), and by comparing these findings with the parent drug. These studies will allow us to determine whether efficacy improves when the drug is conjugated to the HMCD dye. Together, these strategies will confirm whether the conjugate achieves therapeutically meaningful concentrations in the tumour tissue. These studies will be critical for determining the translational potential of this approach and for optimising the therapeutic utility of such conjugates in the context of GBM treatment.

4. Conclusions

GBM continues to remain an extremely challenging cancer to treat, and new treatment modalities are desperately needed to improve survival outcomes for patients. PARP inhibitors, with many being approved for treating various cancers, could potentially benefit GBM patients. The in vitro findings reported in this manuscript are encouraging and support the notion that conjugates of PARP inhibitors appended to BBB crossing and tumour tissue‐specific HMCDs could be explored to broaden the application of PARP inhibitors for treating GBM. Our current and future studies are focused on evaluating conjugate 1 in patient derived GBM cell lines and patient derived xenograft mouse models of GBM.

Supporting Information

Additional supporting information can be found online in the Supporting Information section.

Author Contributions

Carina Dumo: investigation, validation, writing – review and editing, methodology. Faheem: methodology, validation. Peter J. Choi: investigation, writing – review and editing, methodology. Nicholas J. Pederson: validation, methodology. Yuge Liu: validation, methodology. Mike Dragunow: writing – review and editing, resources. Thomas I‐H. Park: writing – review and editing, funding acquisition, supervision. Katharine L. Diehl: supervision, writing – review and editing, resources, data curation, methodology. Jiney Jose: conceptualization, funding acquisition, writing – original draft, supervision, resources, project administration.

Funding

This study was supported by Neurological Foundation of New Zealand (2127 PRG); Cure Kids (5500); National Institutes of Health (R35GM143080).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supplementary Material

Data Availability Statement

The data that supports the findings of this study are available in the supplementary material of this article.