Tumor Survivin Is Downregulated by the Antisense Oligonucleotide LY2181308: A Proof-of-Concept, First-in-Human Dose Study

Survivin, a 16.5-kDa protein encoded by the essential gene BIRC5, was originally identified as an inhibitor of apoptosis (IAP) that exerts its effects through binding to SMAC (second mitochondrial activator of caspases) preventing the sequestration of IAPs by SMAC and inhibition of caspase-dependent apoptosis (1). As a component of the kinetochore-associated complex, survivin also plays an important role in the regulation of late mitosis and cytokinesis (2). Survivin is expressed in a wide range of human cancers and when overexpressed, is associated with a poorer prognosis (3). With the exception of placenta, thymus, activated T cells, gastrointestinal crypt cells, and regenerating liver, survivin is not expressed in normal adult tissue (4). Thus, survivin represents an attractive molecular target for therapeutic intervention. Targeted approaches against survivin include small molecule inhibitors against the survivin protein, gene silencing, and survivin mRNA blockade (5, 6). The 2′-O-methoxyethyl–modified ASO (second generation) LY2181308 is an 18-mer ASO that binds to the translation initiation codon of the survivin transcript. Following ASO hybridization to survivin mRNA, RNase H–dependent cleavage of the duplex ensues with subsequent degradation of the survivin mRNA (7). Thus, survivin protein expression is specifically inhibited without affecting expression of other genes including other IAP proteins (8). Compared with phosphorothioate or first-generation ASOs, second-generation ASOs are more stable, have an improved pharmacokinetic (PK) profile, increased potency, and reduced toxicity (9). Using aggregate data from preclinical pharmacology and toxicology, an integrated PK/PD (pharmacodynamic) model was developed to predict a biologically effective dose range for clinical evaluation (10). On the basis of this, a first-in-human dose (FHD) study was designed to evaluate the biodistribution profile of LY2181308 by measuring the following tumor-specific PD changes: tumor tissue penetration of LY2181308, downregulation of tumor survivin at the mRNA, and protein levels and restoration of tumor apoptosis at a dose and schedule that was safe in humans. To this end, pre- and posttreatment tumor sampling was performed for rigorous assessment of target modulation (i.e., survivin protein and gene expression), the effect of restoring normal apoptosis/cell cycle–related protein activity and biodistribution of LY2181308 including subcellular localization.

This FHD monotherapy study was divided into 3 parts: part A (safety, PK), 1-patient cohorts with initial 100 mg and escalated, by dose doubling, to 400 mg; part B (safety, PK, PD), 3-patient cohorts with 50% dose escalation until dose-limiting toxicity (DLT); part C (dose confirmation cohort, PK, PD). The transition from part A to part B was planned on the basis of either development of toxicity or evidence from PK that the predicted biological effective dose (BED) had been reached. The recommended dose for phase 2 studies was determined from both DLT and the reduction in survivin expression in tumor tissue at the anticipated BED as defined by the preclinical PK/PD model of Callies et al. (10). At the recommended phase 2 dose, 2 companion studies were conducted at the Universities of Oxford (study 1, endobronchial tumor sampling) and Manchester (study 2, [¹¹C]LY2181308 study). The entire study was approved by the Medicines and Healthcare products Regulatory Agency (MHRA) and a Multicentre Research Ethics Committee (MREC).

Inclusion criteria were as follows: at least 18 years of age, confirmed malignancy, exhausted approved standard therapies, tumor accessible for biopsy, written informed consent, Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1, discontinued previous anticancer therapies, absolute neutrophil count ≥ 1.5 × 10⁹/L, platelets ≥ 100 × 10⁹/L, hemoglobin ≥ 9 g/dL, normal serum bilirubin <2.5× upper limit of normal, alanine transaminase (ALT) and aspartate transaminase (AST), calculated creatinine clearance ≥ 50 mL/min, normal activated prothrombin time (aPTT), and prothrombin time (PT), and contraceptive precautions taken. Exclusion criteria were as follows: bleeding diathesis, major surgery within last 4 weeks, pregnant or lactating, symptomatic central nervous system (CNS) neoplasm, taking concomitant anticoagulant therapy, received prior ASO, treatment with an unapproved drug, positive test results for viral infection.

LY2181308 was diluted in 500 mL of normal saline for intravenous injection and infused as a 3-hour infusion on 3 consecutive days as a loading dose, followed by weekly maintenance doses. The dose and dose schedule were based on preclinical safety, PK and pharmacology studies of LY2181308. These were integrated within a PK/PD model that predicted the BED (10).

DLT was defined (CTCAE Version 3.0; ref. 11) as 1 of 3 patients with the following: grade 3/4 hematologic toxicities more than 5 days, aPTT prolongation more than 48 hours after infusion, any grade 3/4 nonhematologic toxicity or any other DLT including those associated with sustained complement activation. Laboratory examinations: C-reactive protein (CRP), complement split products (National Jewish Medical and Research Center), hematology, serum chemistry. Pretreatment tumor biopsy: after enrollment and before first loading. Posttreatment tumor biopsy: window of 48 to 96 hours after the third loading dose. Descriptive statistics were used for all patients receiving at least 1 dose of LY2181308 to evaluate safety, PK, and efficacy.

PK parameters were analyzed (WinNonlin Enterprise, Version 5.2) for: time of maximum concentration (T_max), maximum plasma concentration (C_max), area under the plasma concentration versus time curve (AUC), and clearance (CL) following the third and fifth administration of LY2181308 on days 3 and 15. Plasma PK data were pooled for nonlinear mixed affect modeling analysis (NONMEM, Version 6.1) to determine the compartmental PK parameters: mean value, variance, and subject variability. The interindividual variability was coded as an exponential model and the residual variability as a proportional model.

A polyclonal rabbit antibody against the ASO was generated using a keyhole limpet hemocyanin (KLH)–modified ASO (Lampire Biological Laboratories). It was validated for use in formalin-fixed, paraffin-embedded specimen utilizing Ventana's DiscoveryXTstaining platforms (Ventana).

Formalin-fixed tissue was used for immunohistochemical assessment at Ventana (Tucson) to determine: ASO, survivin (NB500-201; Novus Biologicals), cleaved caspase 3 (CC3; CST9661; Cell Signaling Technologies), and Ki67 (mouse monoclonal antibody; Ventana). All antibodies were detected by appropriate secondary antibodies (Vector Laboratories). Staining intensity was determined using the HSCORE and Ventana's automated imaging software (12). Gene expression analysis (Panomics; ref. 13) was used to quantify Survivin mRNA expression (branched DNA, bDNA). Percentage change in protein and mRNA expression was summarized using medians and approximately 95% confidence limits (14).

Fiberoptic endobronchial tumor sampling is a safe, short, and standard procedure (15). Endobronchial tumors were visualized and sampled by brushing with a 1-mm brush. The samples were processed immediately for flow cytometry assessment.

Briefly, disaggregated tumor cells were fixed in formaldehyde and subsequently stained with phycoerytherin-(PE)conjugated monoclonal antibodies directed against survivin (Clone 91630; R&D Systems). Flow cytometric data (obtained on a CyAN machine, Dako) were analyzed using Summit software (Dako).

After establishing safety in nonhuman primates (16), patients received less than 1 mg and less than 600 MBq of [¹¹C]LY2181308 prior to LY2181308, and during the maintenance infusion on day 15 (17). PET data were collected for 90 minutes following bolus radiotracer injection.

FDG-PET imaging was performed prior to treatment with LY2181308 and on day 22. The imaging was performed as previously recommended (18) and consistent with institutional radiation guidelines.

Between October 2004 and December 2008, 40 patients were enrolled in this monotherapy FHD study, of whom 17 were enrolled in the initial dose-escalation stage with doses ranging from 100 mg to 1,000 mg. Twenty-six patients were treated at the 750 mg dose level, including 6 patients in 2 site-specific studies: study 1 assessed apoptosis and cell-cycle progression changes by flow cytometry of endobronchial tumor cells in 3 patients with NSCLC (non–small cell lung carcinoma); study 2 examined the biodistribution of [¹¹C]LY2181308 measured by PET in 3 patients. The demographics of the study population was typical of phase 1 oncology trials (19) with the majority of patients having received prior chemotherapy (Table 1). Fourteen patients with hepatic metastasis were included, of whom 11 had biopsies taken from liver lesions. LY2181308 was well tolerated with the majority of the patients showing grade 1 or 2 toxicities (29/40, 72.5%). Of the 40 patients, 11 patients (27.5%) had grade 3 or 4 toxicities (Table 2). Two of the 11 patients received the 1,000 mg dose. One patient had a grade 3 hypophosphatemia, whereas the other patient experienced a grade 4 lymphopenia and a prolonged Grade 3 headache, which did not respond to pain medication. This event was defined as the DLT for this study. Concurrently, a sharp rise in CRP was observed in this patient raising the concern of a possible complement-induced cerebrospinal leak syndrome. Three patients were treated at the 900 mg dose level and no DLTs were observed, but CRP, AST/ALT, and moderate flu-like symptoms occured during the loading doses, suggesting continued complement-activation. The most common symptoms were flu-like syndrome (fever, rigor, musculoskeletal pain, nausea), fatigue, and vomiting (Table 2). The most frequent laboratory toxicity was prolongation of aPTT (generally grade 1), which was observed at the time of the infusion in 75% of the patients. Other laboratory abnormalities, not considered as severe adverse events by investigators, included lymphopenia (70%), thrombocytopenia (38%), hypokalemia (38%), and anemia (35%). Transient increases in complement Bb were noted at the end of the third loading dose, but was not associated with adverse clinical events. Given the comparable PK profile of the higher doses of 900 and 1,000 mg to the 750-mg dose, the recommended dose of LY2181308 for further clinical studies using the same dosing schedule was defined as 750 mg, a dose likely to be safe in combination with chemotherapeutics.

The PK properties of LY2181308 and reduction of survivin expression in tumor tissue were assessed by collecting plasma samples and tumor biopsies before and after completing the loading dose. In parts B and C of the protocol, 31 patients were enrolled with 26 (84%) agreeing to have both pre- and posttreatment biopsies. Ten patients had either insufficient tumor material in the posttreatment biopsy or endogenous pigments (bilirubin, hemosiderin, and melanin) that interfered with IHC staining. Hence, in 16 (52%) patients, we obtained sufficient pre- and postdosing tumor tissue to assess the PD effects of LY2181308 by protocol-defined laboratory procedures (Fig. 1A–F).

Survivin protein expression as measured by IHC was reduced by 21%, concomitant with a statistically significant 20% reduction in mRNA expression (P < 0.05) as shown by gene expression analysis (Fig. 1G). Because of the high intensity of nuclear survivin IHC staining prior to the treatment, reduction of nuclear protein expression was more clearly detected than changes seen in the cytoplasm (Fig. 1E–G). Similarly, in the site-specific study in NSCLC, a reduction of survivin protein expression was demonstrated by flow cytometry in endobronchial tumor cells compared with pretreatment levels in all 3 patients studied (Fig. 1I).

The PK of LY2181308 confirmed a multiphasic disposition in plasma with rapid tissue distribution and half-lives of 0.5, 2.5, and 12 hours, and an elimination half-life of 31 days (Fig. 2A). The terminal half-life was the result of a clearance of 20 to 30 L/h and volume of distribution of more than 1,000 L consistent with the tissue distribution of LY2181308 of approximately 90%. As described for other ASOs (9), the primary sites of normal tissue uptake of LY2181308 were renal and hepatic, with less uptake seen in other normal tissue (17). The ASO was detected by IHC in tumor tissue in 10 of 11 patients, either within tumor cells or tumor-associated macrophages (Fig. 2B and C) and in 5 of 11 patients in stroma (Fig. 2D and E). Compatible with the IHC staining, the [¹¹C]LY2181308 PET imaging confirmed that following administration of 1 mg of [¹¹C]LY2181308, tumor penetration of LY2181308 ranged from 21 to 84 ng h/mL tumor with a maximum concentration that ranged from 10 to 60 ng/mL. These levels are comparable with the concentration measured by ELISA in tumor biopsy and with that required for target inhibition (Table 3). The PK profile and tissue distribution of LY2181308 were consistent with the PD effects of the agent described above. However, there was no evident relationship between the localization of the ASO within the tumor tissue compartments, clinical response, or change in survivin expression, perhaps due to the limited numbers of patients included within each cohort.

Of 22 patients assessed for efficacy using the RECIST (response evaluation criteria in solid tumors) criteria (20), a total of 4 patients achieved stable disease, including 1 patient with metastatic melanoma who remained free of disease progression for 18 months. Reduction of survivin expression was associated with an increase in the apoptosis marker CC3 and a reduction in tumor expression of the proliferation marker Ki67 (Fig. 3A and B). In the single patient assessed, a partial metabolic response in several mesothelioma lesions was observed with a 40% reduction in standard uptake values (SUV) on FDG-PET imaging (21; Fig. 3C and D).

Few FHD studies of ASOs have sought to demonstrate PD changes in cancer patients at safe doses such as reduction of the targeted mRNA and protein levels in tumor tissue (22–25). This FHD study of the second-generation ASO LY2181308 was designed to include analysis of serial tumor samples to prove the concept that the ASO was able to inhibit the function of survivin at a safe dose and schedule. We show that the ASO accumulated in tumor tissue, significantly downregulated tumor survivin mRNA and protein expression and enhanced the expression of markers indicative of restored tumor cell apoptosis.

Consistent with previous studies, serial tumor biopsy procedures were safe and accepted by patients (26). Analysis of pre- and posttreatment biopsies demonstrated a reduction of tumor mRNA and protein expression by approximately 20% in a wide range of tumor types. This reduction is lower than that seen in xenograft models, where up to 50% reduction in survivin mRNA and protein were reported (10). In 5 cases, the presence of melanin, bilirubin, or hemosiderin in tumor tissue interfered with the IHC staining and consequently changes in survivin levels after treatment with LY2181308 could not be determined. Tumor samples obtained from patients with NSCLC using fiberoptic bronchoscopy were perhaps even more effective in demonstrating survivin reduction. However, a larger number of patients would need to be included in future studies to confirm our observation. Although survivin levels were reduced, objective clinical responses were not observed and thus direct cytotoxic effects on tumors were difficult to determine. This was an expected finding in the context of a FHD study and because preclinical data suggested that LY2181308 has a cytostatic, rather than a cytotoxic, antitumor effect. Supporting this mode of action were the observations that 1 patient with metastatic melanoma had stable disease while receiving maintenance LY2181308 weekly for 18 months, and 1 patient who had a partial metabolic tumor response based on FDG-PET imaging. Although antitumor efficacy was not a primary endpoint of the study, it was disappointing that no objective responses were seen. Whether this would have been achieved had survivin mRNA and/or protein levels been reduced by more than 20% cannot be addressed by the current study. Combination studies of LY2181308 will assess whether inhibition of survivin enhances antitumor effects of proapoptotic agents such as docetaxel.

An important goal of this study was to demonstrate that PD responses were consistent with the predicted PK profile of LY2181308 in plasma and tissue (Table 3) and its similarity with other second-generation ASOs (9). This prediction was confirmed at the 750 mg dose (Table 3). This is consistent with the observation that ASOs can be scaled successfully from animal to human. The PK profile of LY2181308 and its terminal half-life of about 30 days require a weekly maintenance dose to keep tissue concentrations at levels that were associated with target inhibition in animals. We also demonstrate that the ASO penetrates tumor tissue as detected by IHC and quantitative assessment of ASO levels by ELISA (Table 3). The simulated concentrations in the tumor tissue were confirmed by concentration measurements using ELISA and the [¹¹C]LY2181308 study (Table 3). The range of LY2181308 concentration level observed in the tumor is slightly lower though overlapping relative to the range of LY2181308 through plasma concentration. This could be explained by the fact that other tissues, such as liver and kidney, in addition to the tumor, contribute to the equilibrium between LY2181308 tissues and plasma concentrations The advantage of employing IHC is that it was possible to localize both the disposition of the ASO within tumor tissue and its intracellular distribution. There appeared to be a difference in localization of the ASO within tumor cells and the tumor microenvironment. Whether this observation was a result of the wide range of tumor types studied, the site of the biopsy or the timing after the last dose of LY2181308 is unclear (15 patients had their biopsy taken on day 4; 8 on day 5; 1 on day 6; and 2 on day 7). The variance in timing of posttreatment biopsies resulted in a degree of heterogeneity of results and may have introduced bias. Future studies should define consistent timing of biopsies for all patients and study sites. The use of [¹¹C]LY2181308 PET further supported the IHC-based observation that the ASO penetrated tumor tissue at a pharmacologically relevant concentration. Hence, LY2181308 clearly accumulated in tumor tissue over time (17). This is consistent with studies in animals, where the distribution and stability of the second-generation ASOs was evaluated (9, 27). Other FHD studies evaluating inhibitors of the IAP family have not reported similar PD changes in solid tumor tissue after dosing with ASOs. For example, the small molecular weight survivin inhibitor YM155 was only evaluated for safety and PKs (28), whereas in the clinical trial of the X-linked IAP (XIAP) ASO inhibitor AEG35156 PD changes in PBMCs (peripheral blood mononuclear cells) were reported in addition to safety and PKs (25). Furthermore, the first-generation ASO oblimersen was evaluated for PD activity in melanoma patients only, but the proposed dose and dose schedule was not chosen for future phase 2 studies (22). The second-generation ASO OGX-011 against clusterin did evaluate PD changes in prostate cancer tumor tissue and is the only trial that established a dose and dose regimen based on PD activity in cancer patients that was later used in phase 2 studies (23).

Finally, the dose range at which we observed consistent survivin reduction in tumor tissue had a favorable toxicity profile for future clinical development. Grade 3 and 4 toxicities were present in 11 patients (27.5%) and acute renal failure was not observed during the first cycles of treatment as reported for YM-155, a small molecule survivin inhibitor (28). However, we did observe increased creatinine values (grade 2), which returned to baseline levels after stopping the agent in 1 patient with metastatic melanoma who received LY2157299 at the 750 mg dose for 18 months (29). Although several confounding factors were present, we cannot exclude that LY2157299 treatment was associated with this reversible renal injury. Although not considered medically adverse, lymphopenia was seen in 70% of the patients. Whether this was a PD effect of LY2181308-induced apoptosis of survivin-expressing lymphocytes (30) or reflected an off-target effect of the ASO will require additional investigation. A similar effect on lymphocyte counts was seen with the ASO AEG35156 against XIAP (25). Typical off-target effects of ASO administration were observed for LY2181308 including anemia, thrombocytopenia, and transient prolongation in aPTT (31). In contrast to the phosphothioate ASOs or high doses of second-generation ASOs, the off-target toxicity of LY2181308 was milder and generally limited to grade 1 and 2. This favorable toxicity profile was also recently confirmed in Japanese patients, in particular the loading dose-associated elevation of complement Bb (32).

In conclusion, the integration of PD and PK analyses in this FHD study has provided the proof of concept of effective and specific downregulation of the key molecular target, survivin, in tumor tissue by the second-generation ASO, LY2181308. These findings validate the application of second-generation ASOs for cancer patients. The 750-mg dose and schedule of LY2181308 is currently being evaluated in clinical studies in combination with agents that induce apoptosis such as chemotherapy or radiation (33, 34).

The authors M. Ranson, S. Callies, V. Andr!e, S. Kadam, M. Burgess, and C. Slapak are employees of Eli Lilly and Company, are fully compensated, and hold stock in the company. D. C. Talbot has received other commercial research support, honoraria from Speaker's Bureau, and is on the Advisory Board for Eli Lilly and Company. The other authors declare no potential conflicts of interest.

We thank Drs. C. Prenant, G. Brown, T. Jones, and A. McMahon for [¹¹C]LY2183108 radiolabeling, Dr. M. Slade (Oxford Radcliffe Hospitals) for endobronchial tumor sampling, B. Monia (Isis) for review of the manuscript, and J. Grimes, J. Birkett, C. Leppert, C. Stoner, Helen Desmier and Stacey Maxwell for data validation and study report writing.

The study received funding from Cancer Research UK and the Experimental Cancer Medicine Centres of Oxford, Manchester and London. The study was sponsored by Eli Lilly and Company.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.