ANTICANCER THERAPY AND APOPTOSIS IMAGING
Summary. Early response prediction is considered an essential tool to obtain a more customized anticancer treatment because it allows for the identification of patients who will benefit most from a particular therapy and prevents the exposure of those patients to toxic, non-effective regimens. Recent discoveries of novel markers in functional imaging have created exciting opportunities for in vivo visualization and quantification of cell death. This review will focus on in vivo apoptosis imaging with various radiotracers as predictive tools for tumor response after anticancer therapy. Particular focus will be on annexin V imaging, a technique with the largest clinical experience to date. This article is part of a Special Issue entitled “Apoptosis: Four Decades Later”.
Received: June 25, 2012
*Correspondence: Phone: +31 20 512-21-20
Apoptosis is an evolutionary highly preserved and well-orchestrated biological process involved in both physiological and pathological conditions, and therefore possibly the most abundant form of programmed cell death . Therapy-induced apoptosis in vivo has been shown to significantly contribute to tumor response [2, 3] and to correlate with subsequent outcome [4–6]. After 40 years of intense research, apoptosis is now considered not only as a fundamental process leading to disorders of normal tissues [7–10], but also as a form of cell death in response to oncolytic therapies [11–14].
Apoptosis can be triggered by exogenous and endogenous stimuli leading to the activation of the extrinsic and intrinsic pathways of apoptosis, respectively (Fig. 1). These pathways converge at the activation of a subset of proteases, the executioner caspase-3, -6 and -7, targeting specific intracellular proteins such as those involved in DNA damage repair and cellular cytoskeleton.
The extrinsic pathway is engaged by binding of specific death ligands to specific death receptors on the cell membrane, such as tumor necrosis factor a (TNF-a), TNF-related apoptosis-inducing ligand (TRAIL), Fas ligand (FasL) to TNF receptor 1 (TNFR1), death receptor 4 or death receptor 5 (DR4 or DR5) and Fas/CD95 receptor, respectively. Death receptor-ligand binding recruits an intracellular adaptor molecule, TNF-receptor-associated death domain (TRADD) or Fas-associated death domain (FADD) via its cytoplasmic death domain (DD), clustering to form the death-inducing signaling complex (DISC) which in turn recruits and activates cytoplasmic pro-apoptotic caspase-8 (initiator caspase) via its death effector domain (DED). This is followed by the sequential activation of downstream executioner caspase-3, -6, and -7. Activated caspase-3 translocates to the nucleus and activates poly-ADP-ribose polymerase (PARP-1), which facilitates the degradation of nuclear DNA into 50 to 300 kilobase-sized fragments.
Anticancer drugs and ionizing radiation utilize the intrinsic pathway to trigger apoptosis. This process involves mitochondrial outer membrane permeabilization and the subsequent release of pro-apoptotic factors, including cytochrome c, into the cytosol. Cytochrome c interacts with Apaf-1 (apoptotic activating factor-1), ATP, and procaspase-9 to form a complex known as the apoptosome which in turn activates caspase-9 and further activates the executioner caspase-3, -6, and -7, generating a variety of molecular damages in essentially every organelle. In terms of cell survival, however, it is the damage to DNA leading to the loss of proliferative and clonogenic capacity that is most important. Although chemotherapy and radiotherapy-induced apoptosis is a caspase-dependent process [15–18], it is unclear how cellular signals from DNA lesions lead to the execution of apoptosis. Since Bcl-2 is able to block cytochrome c release and prevent apoptosis, it has been proposed that anticancer therapy-induced caspase activation is mitochondria dependent [17, 19–21]. This concept is further substantiated by the observation that activation of p53 by anticancer therapy due to DNA damage is a direct transcriptional regulator of Bcl-2, Bax, Puma, Noxa and Bid and can act as an apoptogenic factor at the mitochondrial membrane [22–25]. Furthermore, there are functional connections between p53 and death receptors genes (CD95 and TRAIL receptor-2) that can be upregulated in response to therapy [26, 27], in turn leading to the activation of inducer caspase-8. It has also been shown that pretreatment of cells with DNA damaging agents improves the capacity of TRAIL-bound receptors to recruit FADD and activate caspase-8 and -10 in the DISC, irrespective of p53 status .
As apoptosis has been recognized as a major form of cell death after anticancer therapy, it is being increasingly evaluated as a prognostic marker of treatment outcome. For this purpose, a noninvasive method to analyze treatment-induced apoptosis is most attractive, as it can be used to determine and predict the effectiveness of an anticancer regime. In this review, we will discuss apoptosis imaging modalities in both animal models and patients using annexin V, detection of apoptotic membrane imprint, methoxyisobutylisonitrile (MIBI), and the novel caspase-3 small-molecule inhibitor, Isatin, and address their ability to improve patient treatment (see Fig. 1). Annexin V scintigraphy, an imaging technique for which we have the most clinical experience with, will especially be detailed in this discussion.
Fig. 1. Imaging of apoptosis by various radiotracers. While annexin V based compounds and ApoSense agents work at the level of the cell plasma membrane, MIBI acts at the mitochondria and Isatin targets executioner caspase 3 (DR=Death Receptor; PARA=Pro-Apoptotic Receptor Agonist). Inserts are examples of images acquired by the indicated modalities. Modified from: Haimovitz-Friedman et al., 2012 
Imaging of apoptosis using Annexin V
Annexin V-based tracers are the most frequently used agents for in vitro detection and quantification of apoptotic cells. This is based on the high affinity of annexin V for the membrane bound phospholipid (PS), which in viable cells, resides in the inner leaflet of the plasma membrane. Upon exposure of cells to apoptotic stimuli g-scramblase is activated resulting in PS flipping to the outer leaflet of the plasma membrane, thereby allowing annexin V to bind to PS. In addition, it has been shown that membrane binding of proteins that recognize exposed PS on apoptotic cells is regulated by the transmembrane potential . A decreasing membrane potential in Jurkat T leukemia cells and K562 promyelocytic leukemia cells undergoing apoptosis increases the extracellular binding of annexin V in a dose-dependent manner. Studies with PS vesicles also showed that the membrane potential increases the binding affinity of annexin V for the PS cell surface molecules.
Single photon emission computed tomography (SPECT)
99mTechnetium-linked annexin V has been extensively used in apoptosis detection in patients, exploiting its optimal radionuclidic properties for SPECT imaging, relative low costs and easy availability . In 1995, Stratton et al.  were the first to demonstrate 99mTc-annexin V’s value for the in vivo detection of membrane-associated PS exposure using SPECT by injecting 99mTc-labeled human annexin V intravenously and calculating the atrial thrombus/blood ratio in thrombo-embolic diseases. Other SPECT studies using 99mTc-annexin V derivatives provided feasibility and potential clinical utility of apoptosis imaging in various other medical disorders . Multiple conjugators of annexin V have been developed, including 4,5-bis(thioacetamido)pentanoyl (BTAP) for its rapid and extensive radioactivity accumulation in the gastrointestinal tract. Since this tracer was mainly excreted by the liver and the kidneys, resulting in an increased radionuclide accumulation in these organs as well , the role of 99mTc-BTAP annexin V for abdominal imaging was limited.
Blankenberg et al.  subsequently coupled annexin V to hydrazinonicotinamide and created 99mTc-HYNIC annexin V. In vivo studies using a CD95L-induced hepatocyte apoptosis mouse model with intravenously administered anti-CD95 antibody demonstrated that 99mTc-HYNIC annexin V can be used to image apoptotic (and necrotic) cell death in vivo. Although the concentration of tracer in liver and kidneys were still high, and similar to 99mTc-BTAP annexin V, radioactivity accumulation in the bowel was eradicated, making it a suitable candidate tracer for abdominal examination . Following further optimization of image quality, 99mTc-annexin V proves to be an effective modality for non-invasive evaluation of cell death and treatment response in allograft rejection, myocardial infarction, reperfusion injury and infectious disease [36–38].
99mTc-annexin V studies in combination with anticancer therapy
The first in vivo demonstration of anticancer therapy-induced apoptosis involved the use of 99mTc-annexin V in an experimental mouse lymphoma model treated with cyclophosphamide. The animals treated with chemotherapy demonstrated a more than 300% increase in annexin V uptake 20 h after treatment compared to untreated animals . More recently, the value of 99mTc-annexin V imaging in response monitoring was evaluated in a mouse model for hereditary breast cancer after docetaxel treatment . The sensitive tumors showed an increase in 99mTc-annexin V uptake and immunohistochemical evidence of apoptosis one day post-treatment. On the other hand, resistant tumors showed neither an increase in 99mTc-annexin V uptake nor significant immunohistochemical changes (Fig. 2). Despite these encouraging findings, 99mTc-annexin V imaging could not be used to predict tumor response, due to large variations in uptake between animals.
In addition to animal studies, annexin V imaging has also been applied in various clinical protocols. In 2002, 99mTc-annexin V was first used in clinical trials with patients scheduled to receive chemotherapy for locally advanced or metastatic non-small cell lung cancer (NSCLC), small cell lung cancer (SCLC), Hodgkin (HL) and non-Hodgkin (NHL) lymphoma, and breast cancer. Fifteen patients underwent 99mTc-annexin V scintigraphy before and within 3 days after their first course of chemotherapy. Patients with lung cancer received platinum-based chemotherapy, lymphoma patients were treated with vincristine or cyclophosphamide-based chemotherapy and breast cancer patients received taxane as their chemotherapy regime. Five patients had increased annexin V uptake 40–48 hours after chemotherapy (1 NHL, 1 HL, 1 SCLC, and 1 NSCLC), and 2 patients showed increased uptake 20–24 hours after treatment (1 NSCLC, 1 SCLC). At the median follow-up of 117 days, while patients with no change in radiotracer uptake after the first cycle of chemotherapy had no subsequent objective clinical response, patients demonstrating increased tracer uptake post-treatment had either a partial or complete tumor response. From these results, it was concluded that 99mTc-annexin V could be used clinically for in vivo imaging of apoptosis after one course of chemotherapy .
Fig. 2. Preclinical imaging of apoptosis. Quantitative whole animal 99mTc-annexin V SPECT imaging and histology of a T*23 sensitive (a) and resistant tumor (b), obtained before and 1 day after docetaxel treatment. Right panel: tumors were stained for TUNEL at day 1. Modified from: Beekman et al., 2011 
One of the first reports on the application of 99mTc-annexin V in patients receiving radiotherapy was by Haas et al.  who applied 99mTc-annexin V scintigraphy (TAVS) to monitor radiation-induced apoptotic cell death in 11 follicular lymphoma patients (Fig. 3). All patients underwent a baseline scan within one week prior to the start of radiotherapy to detect baseline levels of spontaneous tumor apoptosis or necrosis. Patients were then irradiated to the involved lymph node areas to a total dose of 4 Gy in 2 fractions 48 h apart. At 24 h after the second radiation fraction, TAVS was repeated. Fine needle aspiration cytologic analysis for apoptosis was also performed prior to, during and after irradiation. In 10 patients, post-treatment TAVS matched the post-treatment cytology, confirming TAVS as a valuable non-invasive method to detect in vivo apoptosis caused by radiation. In addition, the increase in 99mTc-annexin V uptake post-treatment in this type of malignancy correlated with clinical outcome: all patients with prominent cytologic and scintigraphic signs of apoptosis achieved complete remission within 1 week.
In 2004, Kartachova et al.  from the Netherlands Cancer Institute conducted a study of 33 patients with malignant lymphoma, leukemia, NSCLC, and head and neck squamous cell carcinoma (HNSCC) scheduled for radiotherapy, platinum-based chemotherapy, or concurrent chemoradiation (see Fig. 3).
Fig. 3. Typical examples of anticancer therapy-induced apoptosis as demonstrated by in vivo annexin V scintigraphy. Shown are 3 examples of SPECT before (left panel) and early during treatment (i.e. 24–48 h after start of therapy; right panel). Upper panel: NHL treated by low dose (2×2 Gy) involved-field radiation. Middle panel: HNSCC treated by cisplatin-based chemoradiation. Lower panel: NSCLC treated by cisplatin/gemcitabine chemotherapy. Arrows indicate the target lesions. Note the physiologic uptake in bones and salivary glands (Modified from: Haas et al., 2004  and Kartachova et al., 2004 , Kartachova et al., 2007 , and Verheij et al., 2008 
The investigators demonstrated increased 99mTc-annexin V accumulation in lesions early during treatment when compared to baseline values in patients with complete or partial tumor remission, while there was no significant early increase in uptake in those patients with stable or progressive disease. This study established 99mTc-annexin V scintigraphy as a predictive marker in tumor response. A subsequent study from these authors evaluated the predictive value of TAVS in 16 chemotherapy-naive patients with advanced stage NSCLC undergoing platinum-based chemotherapy. Also under these conditions, a significant correlation between annexin V changes and treatment outcome was found . In a 2008 update of their study, Kartachova et al.  showed that visual evaluation of SPECT images, SPECT/SPECT and SPECT/CT co-registered images correlated with quantitative analysis. Using both methods, all patients with early post-treatment increase in tumor uptake of 99mTc-annexin V either developed complete response or partial response, resulting in statistically highly significant correlations between changes of 99mTc-annexin V tumor uptake and therapeutic outcome for both visual and quantitative analysis.
In addition to the monitoring of treatment-induced apoptosis in tumor cells, TAVS may also be used to detect normal tissue toxicity. Hoebers et al.  applied 99mTc-annexin V scintigraphy to demonstrate apoptosis in patients with HNSCC, both in tumor and normal tissue. TAVS was performed before and within 48 h after the first course of cisplatin-based chemoradiation in 13 patients. Already after a dose of 6–8 Gy increased annexin V uptake was observed in 24 of the 26 irradiated parotid glands. Glands, which received higher radiation dosages, showed more annexin V uptake. The authors concluded that within the dose range of 0–8 Gy, TAVS showed a radiation-dose-dependent uptake in parotid glands, indicative of radiation-induced apoptosis. A similar pattern in the submandibular glands was observed.
Limitations of annexin V imaging
Although TAVS appears very promising as an early predictor for tumor response to anticancer therapy, several limitations remain to firmly establish its value as an imaging biomarker of response. First of all, apoptosis is an acute event and contributes to early therapy-induced tumor shrinkage. Therefore, TAVS may be less suitable to predict long-term response to treatment [47–49]. To determine whether this type of early cell death predicts long-term treatment outcome parameters in patients, such as disease-free survival and overall survival, changes in TAVS uptake should be correlated with tumor response measurements after sufficiently long follow-up. Secondly, annexin V binding to PS does not discriminate between apoptotic and necrotic cell death as disrupted plasma membranes also make PS accessible at the inner leaflet of cells undergoing necrosis. Thirdly, in terms of biodistribution, annexin V has a relatively slow clearing rate from non-targeted tissues, therefore creating a low signal-to-noise ratio . Fourthly, the optimal timing of apoptosis-imaging in vivo that yields most predictive information in terms of tumor response remains uncertain and should ideally be determined for each specific tumor type and treatment modality. Multiple TAVS measurements may be necessary to define this tumor- and/or therapy-specific optimal timing. Finally, apoptosis represents only one aspect of the complex biological response to therapy, and its relative contribution varies among different tumor entities. More studies are needed to demonstrate the applicability of TAVS, especially in (solid) tumors that are therapy-resistant. Combining TAVS with other anatomical and functional imaging modalities may be helpful in obtaining a more complete, and perhaps more solid biomarker.
Other radiotracers in apoptosis imaging
Detection of apoptotic membrane imprint
The apoptotic membrane imprint is a complex of cellular changes occurring in the plasma membrane early during the apoptotic process. These include irreversible loss of membrane potential, permanent acidification of the external plasma membrane leaflet and cytosol, and activation of g-scramblase while preserving the integrity of the plasma membrane. A set of novel small-molecule probes designated the Aposense compounds (Aposense Ltd., Petach-Tikva, Israel) have been developed to detect these apoptosis-related plasma membrane alterations (see Fig. 1 ). This family of small molecules (DDC, ML-10, ML-9, NST-732, and NST-739) have demonstrated activities in a number of tumor models in response to anticancer agents [52–54], however there is no clear mechanism of uptake of these compounds. The positron emission tomography (PET) tracer 18F-ML-10 shows selective uptake by apoptotic cells in tumors following radio- and chemotherapy, correlating with breakdown of mitochondrial membrane potential, caspase activation and DNA degradation. As the signal is lost upon rupture of the plasma membrane, 18F-ML-10 should be capable of discriminating apoptotic from necrotic cells .
Recently, the role of 18F-ML-10 was evaluated in the early detection of response of brain metastases to whole brain irradiation (10×3 Gy). In this study 10 patients underwent an 18F-ML-10 PET scan prior to treatment and a second scan after 9 or 10 fractions of radiation. MRI was performed 6–8 weeks after completion of treatment. In all 10 patients, both MRI and the 18F-ML-10 PET scan detected all brain lesions. A highly significant correlation was found between early changes on the 18F-ML-10 scan and later changes in tumor anatomical dimensions on MRI .
Technetium-99m methoxyisobutylisonitrile (99mTc-MIBI) is a lipophilic cation isonitrile compound that crosses the cell membrane due to the negative transmembrane potential and accumulates in mitochondria [56, 57]. Because of their higher metabolic activity, tumor cells show large differences in mitochondrial membrane potential and high numbers of mitochondria, making 99mTc-MIBI an attractive agent in tumor imaging (see Fig. 1). Indeed, studies have demonstrated that good responders of anti-tumor treatment display more MIBI accumulation as compared to poor responders in various tumor types, including small cell and non-small cell lung cancer [58–64], breast cancer , malignant lymphoma , osteosarcoma , and nasopharyngeal carcinoma . One explanation for this phenomenon could be the overexpression of the multidrug resistance gene (MDR1), which encodes P-glycoprotein (Pgp), a transmembrane protein that acts as an efflux pump to a wide range of cytotoxic drugs and 99mTc-MIBI [69–75], in chemotherapy resistant tumors. Another possibility is the inhibition of mitochondrial membrane permeability as a result of the overexpression of anti-apoptotic proteins such as Bcl-2 [76, 77]. More recently, studies have established a correlation between 99mTc-MIBI tumor cell uptake and apoptosis after irradiation or chemotherapy [78, 79]. In an unpublished study, Del Vecchio et al.  demonstrated that the overexpression of Bcl-2 which prevented 99mTc-MIBI uptake in untreated breast carcinoma could be counteracted by the initiated drug therapy, allowing a transient increase in 99mTc-MIBI accumulation. The authors speculated that altered expression of proteins participating in the apoptotic process would affect mitochondrial permeability and transmembrane potential, leading to a change in intracellular 99mTc-MIBI accumulation both prior to and after initiation of therapy. Furthermore, high levels of early 99mTc-MIBI uptake after treatment , could potentially indicate therapeutic efficacy. These results suggest that 99mTc-MIBI scintigraphy has important clinical implications in providing prognostic information prior to treatment and also in monitoring effectiveness of treatment after initiation of anticancer therapy.
Several groups have developed novel PET probes designed to non-invasively image caspase-3 activation, including the small-molecule caspase inhibitor, Isatin, used with different radioisotopes (see Fig. 1). Several of these 11C or 18F Isatins examined in vivo [81–87] demonstrated high affinity to caspase-3 and high uptake in the liver in response to cycloheximide or anti-CD95 antibody, confirmed by immunohistology of cell death. Although the radiolabeled Isatins appear to bind specifically to caspase-3, their sensitivity is limited, indicating that further optimization is required for clinical application of these tracers.
Coppola et al.  generated stable cell lines transfected with a hybrid luciferase reported construct, which is activated by caspase-3 cleavage. Using this system, the apoptotic response of D54 glioma xenograft tumors to temozolomide and radiation was monitored by detecting bioluminescence emission. Apoptosis was detected within 6 h after 4 Gy of radiation in mice receiving the combination treatment. Although this method appeared to be very sensitive and quantitative, the experience seemed to be limited to several animal models [85, 89]. No published patient data are available thus far. In addition, it has recently been found that caspase-3 activation is not unique to apoptosis [90, 91] and is involved in other physiological processes such as platelet aggregation and secretion of enzymes from pancreatic acinar cells.
This review focuses on in vivo apoptosis imaging as a clinical biomarker of response to treatment. In evaluating radiotracers for apoptosis imaging, there are a number of issues to consider. First, the tumor is heterogeneous and consists of a mixture of different cell types and no tracer discussed displays sufficient specificity to identify subpopulations of tumor cells undergoing primary apoptosis. In fact, different types of cell death can be induced by anticancer therapy and exist within the same tumor, including “secondary” apoptosis due to incomplete DNA damage repair (after “mitotic death” or “mitotic catastrophe” or “reproductive cell death” [92, 93]). Second, because response to therapy is tumor type-dependent, the optimal timing of in vivo imaging remains to be defined. Though 99mTc-annexin V scintigraphy has emerged as an attractive candidate due to its predictive potential in treatment effectiveness in various tumor types, it is not an exclusive marker for this type of cell death, as cellular necrosis accommodates binding of annexin V to exposed membrane PS as well. Therefore, further studies are needed to evaluate the applicability of this and other potential apoptotic markers in daily clinical practice.
No Comments » Add comments