Use of Multifunctional Sigma-2 Receptor Ligand Conjugates to Trigger Cancer-Selective Cell Death Signaling

There have been notable improvements since 1975 in the relative 5-year survival rates for many cancers, excluding pancreatic cancer, for which survival has not improved substantially over the past 30 years (1). From 1999 to 2005, the 5-year survival rate was approximately 4% to 6% (1). This poor outcome reflects primarily the difficulty in achieving early diagnosis and the failure of surgery, radiation, and chemotherapy. At diagnosis, only about 15% of patients are eligible for surgical resection (2). After pancreatectomy, and even when combined with chemotherapy, radiotherapy, and immunotherapy, the majority of patients develop recurrence and median survival is approximately 2 years (3). Novel therapeutic strategies are desperately needed.

There is considerable interest in stimulating apoptosis and inhibiting the survival machinery as components of cancer therapy (4–6). Many oncogenic transformations result from inactivation or loss of proapoptotic genes or the translocation of an antiapoptotic chromosomal segment to the vicinity of highly active promoters (5, 7, 8). The σ-2 receptor is a unique target that can facilitate tumor cell apoptosis for pancreatic cancer when engaged with an appropriate ligand(s). The sigma receptor was initially proposed to represent a subtype of opioid receptors (9). Early receptor binding studies using benzomorphan opioids indicated the existence of at least 2 types of sigma receptors: the σ-1 and σ-2 subtypes (5). These subtypes display different tissue distributions and distinct physiologic and pharmacologic profiles in both the central and peripheral nervous systems. Although the natural ligand for the σ-2 receptor is still unknown, it has been shown that this receptor is overexpressed in a variety of human and rodent tumors (5, 6, 10, 11), including pancreatic cancer with only weak expression in the normal pancreas (12). As such, synthetic ligands to this receptor could play an important role in cancer diagnosis, imaging, and therapy (13). In fact, we have recently reported on progesterone receptor component 1 (PGRMC1) as being part of the σ-2 receptor complex (14). Equally important, especially with regard to our current study, is the notion that related σ-2 ligands are rapidly internalized into cancer cells (11).

Here, we assessed the feasibility by chemical conjugation to deliver death-inducing cargo via the σ-2 receptor-specific ligand SV119 into the cancer cells. We focused on various molecules along the apoptosis pathway such as Bim (Bcl-2 interacting mediator of dell death), a BH3-only proapoptotic protein in the Bcl-2 family (15–18), carboxyl-terminal modulator protein (CTMP), an Akt inhibitor (19–21), and rapamycin, acting on the phosphoinositide 3-kinase (PI3K)/Akt signaling axis (22, 23).

The findings described in our current article represent a major step in the path to personalized cancer medicine and introduce a platform technology with excellent potential for further expansion and improvement. These novel σ-2 conjugates are capable of delivering death-inducing payloads selectively into their target cells. In addition, the cancer-selective cell killing properties are further enhanced in combination with standard therapies.

σ-2–specific ligands SV119, and fluorescently labeled σ-2 ligand, SW120, were synthesized and prepared as previously described (11, 12). The σ-1 receptor ligand, (+)-pentazocine (Sigma Chemical), was used as a control. The peptide-based σ-2 conjugates were synthesized using solid phase peptide synthesis (SPPS) with the standard Fmoc/HOBT chemistry using the intermediate SV119-Asp (Supplementary Fig. S1). The chemical synthesis of S2-rapamycin can be found in the Supplementary data (Supplementary Fig. S2). The conjugates were purified by reverse-phase high-performance liquid chromatography (HPLC) to more than 95% purity and, where applicable, the amino acid composition was verified with matrix-assisted laser desorption/ionization–time-of-flight (MALDI-TOF) mass spectrometry.

The murine pancreatic adenocarcinoma cell line Panc02, was obtained from Bryan Clary (Duke University, Durham, NC) and maintained in RPMI-1640 (24). Human pancreatic adenocarcinoma cell lines (Panc-1, AsPC1, and CFPAC) were obtained from American Type Culture Collection and maintained in Dulbecco's Modified Eagle's Medium (DMEM; ref. 24).

Tumor cells were incubated with 10 nmol/L fluorescently labeled σ-2 receptor ligand SW120 for 30 minutes. To show the specificity of SW120 for σ-2 receptor binding, 10 μmol/L (+)-pentazocine (σ-1 receptor ligand) was added to the cells 30 minutes prior to SW120 treatment. To study whether σ-2-Bim binds to σ-2 receptors in tumor cells, a series of blocking experiments were done with σ-1 and σ-2–selective ligands. The tumor cells were preincubated with S2-Bim, SV119, or (+)-pentazocine at the indicated concentrations. The cells were then treated with SW120 for 30 minutes, and the fluorescence intensity of the labeled cells was analyzed by flow cytometry. Quantitative analysis of SW120 displacement was carried out with a nonlinear regression model (25).

Tumor cells were seeded at a density of approximately 2 × 10⁵ cells per 12-well plates. The various drugs were added to the cultures with a final dimethyl sulfoxide (DMSO) concentration of less than 1%. The extent of apoptosis was subsequently assessed by terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) and caspase-3 staining using flow cytometry (FACScan, BD Biosciences) and data analysis with CellQuest software (BD Biosciences; ref. 11).

Female C57BL/6 mice (8–12 weeks old) were purchased from the National Cancer Institute (NCI). All mice were injected in the right flank with 200 μL single-cell suspension containing 1 × 10⁶ Panc02 cells. Two weeks after tumor implantation, at which point the mean tumor diameter was approximately 5 mm, mice were treated with a single intraperitoneal injection of S2-Bim. Twenty-four hours later, tumors were harvested and minced to 1 mm and digested in a RPMI buffer containing 1 mg/mL collagenase (Sigma-Aldrich) and 0.1 mg/mL DNase (Sigma-Aldrich) for 45 minutes to obtain a single-cell suspension. After filtering, erythrocyte contaminants were lysed in Ammonium Chloride (ACK) buffer, pelleted, and resuspended in PBS (pH 7.4). Single-cell suspensions were fixed with 1% paraformaldehyde. Apoptosis was assessed by staining for activated caspase-3 as described earlier using flow cytometry.

Female C57BL/6 (NCI) and nude mice (nu/nu, Harlan; 8–12 weeks old) were injected as described earlier with 1 × 10⁶ Panc02 (C57BL/6) or 2 × 10⁶ CFPAC cells (nude mice), respectively. Treatment of tumors started approximately 2 weeks postimplantation, at which point the mean tumor diameter was approximately 6 to 7 mm. To evaluate the effect of treatment both systemically and on tumors in vivo, several treated mice were sacrificed and blood cytologic (complete blood count) and biochemical analysis [liver enzymes, bilirubin, amylase, lipase, blood urea nitrogen (BUN), creatinine, glucose] were conducted by the Division of Comparative Medicine at our institution. For survival studies, tumor-bearing mice (n = 12 per group) were treated with S2-Bim and S2-BimX. All mice were euthanized when their tumor ulcerated or reached a mean diameter of 15 mm. All studies were conducted in accordance with an animal protocol approved by the Washington University Institutional Animal Care Facility.

Error bars represent means ± SEM of an experiment with at least 3 biologic replicates. For statistical analysis between 2 groups, the Student t test was applied. For statistical analysis of differences between groups, ANOVA was conducted. For in vivo experiments, Kaplan–Meier survival curves were plotted and differences were compared with a log-rank test. A P value of less than 0.05 was considered significant for all analyses.

We have previously reported that apoptosis-inducing peptides such as the BH3 domain of the Bcl-2 antagonist Bim (18) and a peptide derived from the Akt inhibitor CTMP (CTMP-4; ref. 21) can be delivered to cancer cells via the HIV-1–derived transduction domain TAT. Because this delivery mode is nonselective, we assessed the possibility to deliver proapoptotic effector molecules via SV119 (Fig. 1A) preferentially to pancreatic cancer cells that are characterized by an abundance of receptors for this ligand.

The following compounds were generated as described using standard chemistry (26, 27): S2-Bim and S2-CTMP-4 (Fig. 1B, only the precursor structure without the corresponding peptide is shown) as well as S2-rapamycin (Fig. 1C) along with their nonfunctional, inactive variants (Supplementary Figs. S1 and S2). The binding capacity of the conjugates to the σ-2 receptor was compared with nonconjugated SV119. As a negative control, the σ-1–specific ligand (+)-pentazocine (Fig. 1D) was used (12).

Competitive binding assays were conducted using the fluorescently labeled σ-2 ligand SW120 (11, 25, 28; Fig. 1E). Compared with the nonconjugated SV119 compound, S2-Bim and its inactive variant S2-BimX displaced SW120 in a dose-dependent fashion (Fig. 2). As expected, the σ-1–specific ligand (+)-pentazocine was unable to compete off SW120 and served as a negative control. Similar results were obtained when the same assays were repeated with SV119 conjugated to CTMP-4 and rapamycin (data not shown). These results suggest that covalent attachment of additional small molecule cargoes to the parental molecule SV119 does not alter the specificity of the compounds to the σ-2 receptor.

The drug conjugates were next tested for their ability to induce cancer cell death in vitro. As a surrogate of target cell apoptosis, activation of caspase-3 was used as previously described (12). The mouse pancreatic cancer cell line Panc02 was challenged in vitro with equimolar doses (10 μmol/L) of S2 conjugates, and caspase-3 induction was assessed 24 hours later. Relative to the nonmodified parental σ-2 delivery agent, all compounds linked to an “active” cargo exerted an increase in target cell apoptosis (Fig. 3). Apoptosis induction increased between 1.3-fold (S2-CTMP-4, Fig. 3A) and 3.5-fold (S2-Rap, Fig. 3B) with S2-Bim resulting in an intermediate increase of 1.6-fold compared with SV119 alone (Fig. 3C). The most dramatic effect in overall potency was achieved with S2-Bim with an 11% increase in caspase-3 activation compared with SV119 alone (Fig. 3C). Of note, this activity was not matched by TAT-Bim, a recently described very potent, nonselective inducer of target cell death (18), with S2-Bim being 13% more powerful.

To gain insight into the respective killing mechanisms, we studied the signaling pathways that are targeted by the various effector molecules delivered to the cells via the SV119 moiety. Caspase-3 activation is closely linked to the antagonizing interaction of Bim with Bcl-2 and has been already described earlier (Fig. 3C). CTMP interferes with activation of Akt whereas rapamycin blocks activation of the p70 S6 kinase as part of the mTOR complex.

CFPAC cells were treated with S2-CTMP-4, and the activation status of Akt was monitored 24 hours posttreatment. Assessed on the basis of its phosphorylation status, only S2-CTMP-4 and TAT-CTMP-4 (included as a control), were capable of interfering with Akt activation (Fig. 4). Likewise, when S2-rapamycin was used, a significant reduction in phosphorylation of its downstream target was observed (data not shown). In both cases, SV119 alone was incapable of altering the activation status of these 2 central kinases.

In an attempt to show broader applicability, we assessed the killing capacity of S2-Bim using additional pancreatic cancer cell lines. All the cell lines under investigation responded to this novel drug candidate in a dose-dependent fashion being consistently more potent than SV119 alone and the nonselective death inducer TAT-Bim (Supplementary Fig. S3). Surprisingly, the covalently linked inactive form of Bim (S2-BimX) had virtually no activity in all cell lines tested. The reason for this loss of activity in vitro is currently unknown and might be partially rescued under in vivo settings (see later).

According to the activity profiles of our various σ-2 conjugates, S2-Bim seemed to have the greatest potential as a cancer therapeutic. Consequently, this drug was tested next in vivo, that is, using a syngeneic and a xenogenic subcutaneous mouse model of pancreatic cancer. In an initial study, mouse pancreatic tumor cells (Panc02) were engrafted into the flanks of wild-type C57BL/6 mice. After the tumors were established, the effect of S2-Bim was assessed following a single intraperitoneal injection by immunohistochemistry (TUNEL staining) and by staining for activated caspase-3. Twenty-four hours posttreatment, large areas of TUNEL-positive cells were identified (Fig. 5A). Furthermore, the activity of S2-Bim was dose dependent as doubling the concentration to 400 μg per animal increased the number of caspase-3–positive tumor cells by nearly 100% (Fig. 5B).

In vivo efficacy was studied next in which C57BL/6 mice with established Panc02 tumors were treated intraperitoneally with S2-Bim for 2 weeks every other day. The controls, vehicle alone and the conjugate incorporating the inactive variant of the Bim peptide S2-BimX, had no effect on the tumor growth (Fig. 5C), in agreement with the lack of in vitro cell killing activity (Supplementary Fig. S3). In contrast, S2-Bim prevented tumor growth completely within the first week of treatment (Fig. 5C). However, at his dosing regimen the initial response to the drug ceased and the tumor growth resumed at a similar rate as the controls. Of note, whereas all the control mice (vehicle and S2-BimX) died at 43 days post tumor inoculation, almost half (>40%) were still alive at the end of the experiment (Supplementary Fig. S4).

Using a xenograft tumor model, CFPAC-bearing nude mice were treated with S2-Bim for 1 week on a daily schedule. Tumor growth of mice receiving vehicle only progressed rapidly whereas mice treated with S2-BimX displayed a somewhat reduced tumor growth rate, similar to SV119 alone (Fig. 5D). In stark contrast, the tumor growth curve of mice receiving S2-Bim displayed a negative slope (tumor regression) during the treatment period (Fig. 5D). However, after the treatment was discontinued, the tumor growth progressed, comparable with that of the control group.

In an attempt to assess the potential systemic toxicity of the S2-Bim conjugate, tumor-bearing C57BL/6 mice were injected intraperitoneally with a single dose of the drug, and several key organs were analyzed for apoptosis induction via staining for active caspase-3. The following day, a definitive increase in tumor cell apoptosis was detectable for S2-Bim with only background signals for the vehicle and inactive σ-2 compound S2-BimX (Fig. 6A). Also, significant caspase-3 activity was noted in the pancreas, which occurred only transiently (see later) and further suggests that our delivery vehicle has specificity for this particular organ. All the other organs analyzed were essentially unaffected by this treatment regimen (Fig. 6A).

Because we noticed a slight toxicity of S2-Bim to the pancreas, we assessed next serum amylase and lipase levels. These would increase in case of organ damage to the pancreas following drug treatment. And indeed, at 1 week following a single injection of S2-Bim into normal mice, we detected an increase in the levels of these biomarkers, ranging from 7- to 10-fold relative to their baseline levels (Fig. 6B). However, the mice showed no clinically measurable signs of distress and recovered from this insult quickly as by 1 week after drug cessation both serum concentrations reached baseline levels (Fig. 6C). The same is true for our long-term treatment studies where the mice remained healthy by weight and appetite throughout the course of our experiments (Fig. 5C and D), and we did not encounter unanticipated deaths. In support of these findings, necropsy revealed no gross or microscopic organ damages.

Monotherapies are rarely successful when it comes to the treatment of malignant diseases. For this reason, we asked whether and to what extent a dual therapy would potentially increase the effectiveness of S2-Bim. Two of the most widely applied treatment regiments are radiation and conventional chemotherapy. When CFPAC cells were treated in vitro with a combination of gemcitabine and S2-Bim, an increase of approximately 10% to 15% cell death was induced compared with the same treatment but replacing the active with the inactive form of the σ-2 compound (Fig. 7A). Cotreatment of S2-Bim and radiation in vitro resulted in an almost 100% increase in TUNEL-positive target cells from 30% to nearly 60% (Fig. 7B).

A fundamental limitation of most conventional cancer therapeutics is their nondirected delivery to the target cells. Nonselective delivery often represents the primary reason for toxic side effects and limited treatment efficiencies. If the delivery component would be structurally separated from the therapy component of a drug conjugate, one could build on these modular reagents to tailor therapies that would target particular defects associated with a specific type of cancer subtype. Such a cancer-selective delivery agent could open up the door for retesting many of the drugs that already proved efficacious in vitro but proved to be too toxic for in vivo applications.

In our current article, we investigated this concept by chemically linking a targeting vehicle with high specificity to pancreatic cancer cells with a variety of small molecules that would, upon delivery to the target cell, exert their respective intracellular effector functions (apoptosis induction). It should be noted here that our chosen delivery agent has a limited, intrinsic cytotoxicity toward pancreatic cancer cells. This was, however, considered during the design phase of our dual domain therapeutics, as only the combined activities of the cancer-selective delivery vehicle SV119 and a secondary, complementary death stimulus were expected to result in the most effective drug conjugate. Indeed, we showed here that the overall potency of our various drug conjugates enhanced the intrinsic killing capability of the delivery agent SV119 via specific engagement of the respective therapy component intended to induce a particular apoptotic signaling pathway.

A key requirement for a universally applicable drug delivery platform requires that the target specificity is maintained following covalent linkage of the parental structure to a particular effector moiety. The additional effector molecules did not abrogate binding of the conjugates to pancreatic cancer cells. However, when SV119 was combined with the inactive peptide mutant(s) of, for example, Bim (S2-BimX), a complete loss of killing activity was noticed in vitro. This was unexpected, as we assumed that this variant should have at least SV119 baseline activity. The same reagent elicited SV119 baseline activity in a pancreatic xenograft model using CFPAC cells and highlights the notion that the bioactivity of these drugs might be variable and depend, at least in part, on the specific cell type in question and/or on the type of application used (in vitro vs. in vivo). A potential explanation for this finding might relate to the nonconservative substitution of the charged amino acid asparagine with a hydrophobic alanine located within the Bim peptide. The exact nature of the differential activity profile is currently being investigated, as it might provide clues for the future design of improved variants of this class of cancer drugs.

Therapeutically, the most promising results were obtained in a xenograft model of pancreatic cancer using S2-Bim. As long as the σ-2 conjugate was administered, established human tumors responded with shrinkage. This was certainly encouraging as it has rarely been seen that this type of cancer is treatable. A somewhat different picture was obtained in a syngeneic tumor model. Tumor growth was halted for nearly a week but resumed while treatment was still ongoing. At the present time, we can only speculate as to why the same drugs displayed differential activity profiles in different animal models. For example, mouse Panc02 cells grow much more aggressively than human CFPAC cells (29). This feature could lead to a faster development of resistance to our cancer drugs and would represent a characteristic that is cell line specific. On the other hand, it could well be possible that an immune component might be involved that counteracts tumor eradication in a normal host (C57BL/6-Panc02). Because these 2 alternative mechanisms are critical for the evaluation of our investigative drugs, we are currently in the process of dissecting which of the mechanistic paths is the dominating factor.

While the cytotoxic benefit of S2-Bim was greatest on the target cells, we did notice mild elevations of biochemical markers of tissue injury, that is, amylase and lipase. In our model, rising levels of these biomarkers were only transient and restricted to the pancreas. As with many cytotoxic drugs, we expected to find evidence of mild organ damage, especially of the liver, but did not detect elevations of the respective biochemical markers. Even though our σ-2 ligands are highly tumor selective, we cannot rule out binding to nontransformed, normal cells, or tissues. In fact, we did indeed identify binding at lower levels to many other cell/organ types including the liver and the intestine. Along this line, we are currently in the process of evaluating relevant pharmacologic parameters of selected drug conjugates such as half-life and maximum tolerated dose (MTD). Results from these ongoing studies will help us in the transition process from the preclinical to the clinical arena.

Peptides are not ideal partners for the formulation of effective anticancer drugs. One of the main reasons is their susceptibility to proteolytic degradation. Along this line, Abbott has recently developed the Bim peptidomimetics ABT-737 and ABT-263 (30–32), with the latter even being orally bioavailable. On the basis of the data showed here, it seems feasible to combine σ-2 ligands with other small molecular therapeutics and assess their potency in future studies.

Monotherapy is rarely sufficient to combat aggressive human malignancies. With this regard, we provided evidence in an in vitro cell culture system that the S2-Bim conjugate substantially augmented standard of care treatment regiments currently used in clinical settings. More specifically, combination with chemotherapy and, more profoundly, radiation therapy increased target cell death by nearly 100% with an overall killing of more than 60%. While radiation therapy has its intrinsic caveats, combination of S2-Bim with gemcitabine resulted in a much improved killing response. Therefore, we are currently carrying out preclinical studies using the latter combination therapy to assess its efficacy in an orthotopic mouse model of pancreatic cancer.

In summary, we present a platform concept for the cancer-selective, SV119-mediated (σ-2 receptor dependent) codelivery of apoptosis-inducing cargo (peptides or small molecules) to the target cells. Once inside the cell, these bifunctional conjugates mediate enhanced killing via the combined activities of both moieties. As a result, the baseline cytotoxic activity of the delivery agent is augmented following the signaling pathway of its cargo. This new concept represents an exciting opportunity for the development of future small molecule drugs with dual functionality combined in a single reagent. In fact, we have recently developed another σ-2–specific ligand, the SV119 homolog SW43, which turned out to be twice as cytotoxic by itself (24) and could represent the building block for the next generation of bifunctional (e.g., SW43-Bim) therapeutics to treat patients with pancreas cancer. Even though our study was focused on pancreatic cancer, we are developing evidence that this concept could also be applied to other types of malignancies including ovarian and breast cancers.

W.G. Hawkins and R.H. Mach have intellectual property rights in and have patented the drug platform idea in this article. No potential conflicts of interest were disclosed by the other authors.

The authors thank Dr. David Linehan for expert advice regarding the preparation of the manuscript and Stacey Plambeck-Suess for excellent technical assistance.

This work was supported by the American Cancer Society (MRSG08019-01CDD), a Veterans Administration Merit Award (#1136919; to W.G. Hawkins), the Doug Phillips Research Fund #4125-35604; to D. Spitzer), a Surgical Oncology Training Grant (#5T32CA009621-22; to P.O. Simon and J.R. Hornick), and a Siteman Cancer Center Pilot Feasibility Grant (#5P30CA09184208; to W.G. Hawkins). It was further supported by the NIH (CA102869) and the U.S. Army Medical Research and Material Command under DAMD (17-01-1-0446; to R.H. Mach).

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.