TRENDS in Pharmacological Sciences
Vol.24 No.3 March 2003
Techniques: Visualizing apoptosis using nuclear magnetic resonance Juhana M. Hakuma¨ki1 and Kevin M. Brindle2 1
Department of Biomedical NMR, National Bio-NMR Facility, A.I. Virtanen Institute for Molecular Sciences, University of Kuopio, PO Box 1627, FIN-70211 Kuopio, Finland 2 Department of Biochemistry, University of Cambridge, 80 Tennis Court Road, Cambridge CB2 1GA, UK
Apoptosis plays a key role in tumour biology, and the induction of apoptosis forms a cornerstone of most anticancer therapies. New developments in nuclear magnetic resonance spectroscopy (MRS) and magnetic resonance imaging (MRI) have taken these techniques far beyond their original roles as the workhorses of structural and pharmaceutical chemistry and clinical imaging to the detection of previously inaccessible and unrecognized biological phenomena in living cells and tissues undergoing apoptosis. These new MR techniques can be used in the development of new drugs and in the improved detection of treatment responses in the clinic. Apoptosis, or programmed cell death, is a genetically controlled process in which cells commit suicide through a cascade of metabolic events. The profound morphological changes that are its hallmark include cell shrinkage, DNA fragmentation, nuclear condensation and, ultimately, the formation of membrane-encapsulated apoptotic bodies. The confinement of cellular material allows engulfment by neighbouring cells, and serves to protect surrounding tissue from the release of substances, which would normally elicit inflammation and necrosis . Apoptosis is required for the normal development and function of tissue, and acts as a crucial counterbalance to cell proliferation. Dysregulation of apoptosis is linked to many diseases, notably cancer, neurodegenerative conditions and autoimmunity [2,3]. Chemotherapeutic cancer treatment often leads to tumour cell apoptosis, and many other anti-tumour treatments, including hormonal agents and radiotherapy, are also believed to act in this way. In fact, significantly increased numbers of apoptotic cells are observed in tumours that are treated to induce apoptosis [4– 6], and an early apoptotic response has been shown to correlate well with subsequent disease outcome [6– 8]. Because novel therapies with the sole purpose of inducing apoptosis are being developed , it would be clinically very useful to have a non-invasive technique to assess apoptosis in situ. Such techniques could prove useful in drug development and in assessing the early responses of patients to therapy. During the past few years, we and others in the field have studied the possibility of using nuclear magnetic Corresponding author: Juhana M. Hakuma¨ki ([email protected]
resonance (NMR) spectroscopy (MRS) and magnetic resonance imaging (MRI) for the non-invasive detection of apoptosis (Box 1). It has been shown that MRS techniques can, non-destructively, visualize changes in protein synthesis, glycolysis, phosphatidylcholine and triglycerides turnover, energy levels and intracellular pH throughout the execution of the ‘apoptotic programme’. MRI image contrast can also be sensitized to biophysical changes in the cellular milieu or can even be specifically targeted to detect the metabolic changes that occur during apoptosis. Detecting apoptosis using MRS Several research groups have characterized metabolic indicators of apoptosis. Adebodun and Post  were the first to use 31P MRS to study cells undergoing apoptosis in vitro. Using leukaemia cells treated with dexamethasone, they observed a clear reduction in the levels of phosphomonoesters (i.e. phosphoethanolamine and phosphocholine) and ATP. Nunn et al. used 1H MRS to study changes in the concentrations of secondary metabolites in neutrophils undergoing apoptosis, and observed an increase in the concentration of phosphocholine, which was attributed to Fas-receptor-mediated activation of phospholipase C . To better characterize the secondary metabolites involved, Williams et al.  treated Chinese hamster ovary cells (CHO-K1) and HL-60 leukaemia cells with several functionally different apoptosis-inducing drugs: farnesol, chelerythrine, etoposide, campothecin and ceramide. Only two metabolites were consistently found to increase during apoptosis, namely fructose-1,6-bisphosphate (FBP), a glycolytic pathway intermediate, and cytidine diphosphocholine (CDP-choline), an intermediate in phosphatidylcholine biosynthesis. The increase in the concentration of FBP in HL-60 cells could be explained by depletion of cellular NAD(H), or by the activation of 6-phosphofructo-1-kinase through AMP accumulation. Similar observations have been made by Ronen and coworkers . The accumulation of CDP-choline in apoptotic cells was shown by isotope labelling studies to be caused by the inhibition of cholinephosphotransferase (CPT) . This enzyme might become inhibited in apoptotic cells because it has an alkaline pH optimum (8.0 – 8.5). Cellular acidosis (pH , 6.5) appears to be a frequent event in
http://tips.trends.com 0165-6147/03/$ - see front matter q 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0165-6147(03)00032-4
TRENDS in Pharmacological Sciences
Vol.24 No.3 March 2003
Box 1. Nuclear magnetic resonance Certain atomic nuclei, such as the proton, possess a non-zero angular momentum (i.e. nuclear spin), which creates a magnetic dipole moment along their axis of rotation. When placed in an external magnetic field, net nuclear magnetization is created by the equilibrium difference between the magnetic dipole moments of spin populations in two different energy states. This net magnetization, which is proportional to the strength of the external magnetic field, can be perturbed by electromagnetic (radio frequency) radiation. This phenomenon is called nuclear magnetic resonance (NMR). Perturbed magnetization and its recovery (i.e. relaxation) to the equilibrium state can be detected and information on the nuclear spin environment of the sample or tissue is obtained. The signal can be used to detect and quantify the presence of chemically distinct compounds [NMR spectroscopy (MRS)] or to create magnetic resonance images (MRIs) based on the spatial distribution and/or relaxation of the observed nuclei. Because ionizing radiation is not involved, NMR is most useful for the study of living systems. With the exception of tritium (3H), proton (1H) is the most sensitive NMR nucleus (99.9% natural abundance). It is well suited for NMR because of its abundance in biological material. MRI benefits particularly from the high concentration of water protons in tissue (, 80 M ). A set of sample 1H spectra is shown in Fig. 1a in the main text. Many other nuclei, however, can also be used. These are typically the naturally occurring (100% abundant) isotope of phosphorus, 31P, and the naturally occurring (1.1% abundant) isotope of carbon, 13C. A sample 31 P NMR spectrum is shown in Fig. I. Although far less sensitive than 1H, 31 P can provide unique information on energy metabolism and intracellular pH. 13C labelling of substrates can be used to study metabolism. A key advantage of MRS techniques in detecting metabolic changes, such as in apoptosis, is that no presumptions about which metabolites change are required. Unfortunately, NMR is not a very sensitive technique, but changes will be detected, provided that they are
apoptosis, as was confirmed by both flow cytometry and 31 P MRS in HL-60 cells and CHO-K1 cells . Blankenberg et al.  were the first to use 1H MRS of living cells in vitro to detect metabolic changes in apoptosis. They observed large signals from mobile intracellular lipids, and demonstrated a correlation between the intensity ratio of the 1.3 ppm CH2- and 0.9 ppm CH3- resonances and the fraction of apoptotic cells following induction of apoptosis by doxorubicin in Jurkat T cells. The authors suggested that the lipid signals arise from the plasma membrane, although more recent work suggests that the source is cytoplasmic lipid droplets . Hakuma¨ki et al.  have extended this approach by showing that 1H MRS can detect increases in lipid signals following tumour treatment in vivo. In this study, herpessimplex virus thymidine kinase (HSV-tk)-transfected BT4C gliomas were treated by systemic administration of ganciclovir, and a significant increase in the levels of polyunsaturated fatty acids was observed (Fig. 1). These lipid signals correlated with the number of apoptotic cells, the concentration of cholesteryl esters and triglycerides, and the number of osmiophilic cytoplasmic lipid droplets in tumour cells. Recently, two groups have shown, using Jurkat T cells, that apoptosis is accompanied by an increase in cytoplasmic lipid droplets and a proportional increase in the lipid signal in 1H MR spectra [17,18], with independent biochemical evidence suggesting a role for phospholipase A2 activation following the induction of apoptosis . MRS measurements of apoptosis in vivo require markers with sufficient signal intensity. If suitable markers can be found, then chemical shift MRI techniques http://tips.trends.com
–8 –10 –12 –14 –16 ppm
Fig. I. A sample 31P spectrum. Intracellular metabolites from living Chinese hamster ovary (CHO-K1) cells treated with ceramide are observed. Abbreviations: CDP-choline, cytidine diphosphocholine; GPC, glycerophosphocholine; NTP, nucleotide triphosphate Pi, inorganic phosphate; PME, phosphomonoesters. Reproduced with permission from John Wiley & Sons . q1998 WileyLiss.
above a practical threshold (typically , 1 mM in vivo, and , 0.1 mM in vitro). For a more detailed account on biological NMR, see . Other more sensitive techniques for detecting apoptosis in situ, such as that based on the labelling of annexin V by fluorescein , or 99mTc (metastable technetium) , can be used, but are beyond the scope of this review. Unlike these techniques however, NMR does not necessarily require labelling, allows full tissue penetration, and causes no exposure to ionizing radiation.
could be used to visualize their presence in tissue. In this sense, the 1H MRS signals from cytoplasmic lipids appear promising. However, this might be problematic because the metabolic factors that contribute to the increases in the concentration of lipids are still inadequately understood. Theoretically, 1H MRS-detectable lipid signals could also arise from other cellular processes, such as necrosis, and other endogenous, non-malignant cell types . Furthermore, the spectral baseline features might differ, with some cell lines inherently presenting higher levels of lipid than other cell lines, as exemplified by some multidrug resistant cell lines [20,21]. However, many available in vitro assays for apoptosis can also be ambiguous. For example, DNA fragmentation, often the gold standard for apoptosis detection, cannot always be observed in cells undergoing apoptosis . CDP-choline and FBP could also be used as spectroscopic markers. However, 31P MRS is less sensitive and the signals are relatively weak. The resonances from FBP are also partly overlapped by other phosphomonoester signals. It is also unknown whether either of these metabolites will rise to detectable levels in tumours in vivo. In addition to metabolic profiling, 1H MRS can be sensitized to molecular diffusion to study the physical effects of pharmacotherapy. Hakuma¨ki and co-workers  have shown that during apoptosis-inducing therapy, there is a 50% reduction in the apparent diffusion coefficient (ADC) of intracellular choline metabolites in vivo, which is accompanied by a significant increase in a rapidly diffusing water component. This pool of water is most likely to be extracellular water. These results suggest that there is a decrease in cell size and number, and intracellular
TRENDS in Pharmacological Sciences
Vol.24 No.3 March 2003
Day 10 Day 8 Day 6 Day 4
Day 2 Day 0 6.0
0.0 TRENDS in Pharmacological Sciences
(c) Fig. 2. 1H magnetic resonance (MR) images of a tumour in a drug-treated mouse following injection of the contrast agent C2 –SPIO (20 mg iron per kg tissue) are shown. MR images (a–e) [echo time (TE) ¼ 30 ms] are from before and after injection of C2–SPIO. The first control image (a) was acquired before injection of the contrast agent followed by images 11 (b), 47 (c), 77 (d), and 107 (e) min after injection. Areas of contrast agent uptake (and apoptotic cells) become visible as areas of lost image intensity (a– e), with the white arrow in (e) depicting the maximal effect. The effect is enhanced in the respective subtraction images (a– b, a– c, a –d, a– e) as areas of increasing image intensity. Image resolution of , 100 mm can be achieved. Abbreviation: C2 –SPIO, C2 domain of the protein synaptotagmin I with a superparamagnetic iron oxide particle. Figure modified, with permission, from . qNature (http://www.nature.com).
TRENDS in Pharmacological Sciences
Fig. 1. (a) Typical time-course of 1H MRS from an HSV-tk-transfected glioma treated with ganciclovir in vivo for a period of 10 days; signals from polyunsaturated lipids are shaded. (b) An untreated glioma is shown on conventional MRI; a white box outline denotes the area from which spectroscopy signals were obtained. Transmission electron microscopy ( £ 5000) shows untreated tumour cells (c), lipid droplets (open arrows) and apoptotic bodies (black triangles) (d), and densely osmiophilic, polyunsaturated fatty acid-rich droplets (arrows) (e) in treated tumours after four days of treatment. The technique allows non-invasive assessment of metabolism in tumours undergoing apoptosis in vivo. Abbreviations: HSV-tk, herpes-simplex virus thymidine kinase; MRI, magnetic resonance imaging; MRS, nuclear magnetic resonance spectroscopy. Figure modified, with permission, from . qNature (http://www.nature.com).
viscosity, changes that are expected to occur in cells undergoing apoptosis. Recently, Hortelano et al.  have shown similar MRS results in cells, which support the idea that reduced intracellular volume and cell numbers are responsible for the observed changes in water diffusion. Scott and Adebodun  have very recently used 13C MRS in vitro to study protein synthesis. This was achieved by measuring the incorporation of 13C-labelled amino acids into cell proteins. In CEM-C7-14 human leukaemia cells, the induction of apoptosis by dexamethasone did not stop protein synthesis, but reduced it by 60–87%. Detecting apoptosis using MRI MRI provides spatial and temporal resolution that is superior to that of MRS. This is important clinically, where MRI is more widely available, accurate localization is http://tips.trends.com
required, and the levels of therapeutically induced apoptosis might be low. Hakuma¨ki et al. have recently shown in the BT4C rat glioma model that, in HSV-tktransfected tumours, the spin – lattice (T1) and spin – spin (T2) relaxation times of water molecules in the tissue, and their rate of diffusion, increase during ganciclovir therapy. These changes, which are attributable to tissue hypodensity, are concurrent with a reduction in tumour volume, and are therefore limited as early markers of tumour apoptosis. However, T1 relaxation contrast in the rotating frame (T1rho) displays significant changes even when tumour volumes are still increasing but significant apoptosis is taking place . Intrinsic image contrast mechanisms are unfortunately still rather poorly understood, and, at best, there might still be an ambiguity that is related to tissue type and the type of therapeutic intervention. Therefore, it would be extremely beneficial to combine chemical specificity (such as that obtained by MRS) with the much better spatial resolution (, 100 mm) of MRI. Zhao et al. have demonstrated success with such an approach by labelling the first C2 domain of the protein synaptotagmin I with a superparamagnetic iron oxide particle (SPIO), which served as the MRI contrast agent . Synaptotagmin I, like annexin V, binds to the phosphatidylserine that appears on the outer leaflet of the plasma membrane of apoptotic cells. The SPIO-labelled protein was shown to bind to apoptotic cells in vitro and, when administered intravenously into mice bearing EL4 tumours, decreases in MRI intensities in regions of the tumours containing large numbers of apoptotic cells could be observed (Fig. 2). This type of contrast relies on intensity decreases, which unfortunately can be more prone to artefacts and are more difficult to interpret because of tumour heterogeneity. Conjugation of the protein with gadolinium-based agents, which can give signal increases, might provide better specificity. This approach is currently being explored, together with the specificity and sensitivity of these methods. However, the generation of protein-based targeted
TRENDS in Pharmacological Sciences
contrast agents can be expensive, and the compounds need to be administered intravenously and thus will need to be assessed for safety. Meanwhile, we can expect to see some of the truly non-invasive techniques such as MRS and diffusion-weighted MRI to find their way into the clinic. For experimental drug design, however, the whole battery of techniques described here is already available. Acknowledgements We would like to thank the Emil Aaltonen Foundation, Academy of Finland, the Finnish Cancer Institute (J.M.H.), Cancer Research UK, the European Commission, and the Medical Research Council (K.M.B.) for supporting this work.
References 1 Savill, J. and Fadok, V. (2000) Corpse clearance defines the meaning of cell death. Nature 407, 784– 788 2 Kinloch, R.A. et al. (1999) The pharmacology of apoptosis. Trends Pharmacol. Sci. 20, 35 – 42 3 Evan, G.I. and Vousden, K.H. (2001) Proliferation, cell cycle and apoptosis in cancer. Nature 411, 342 – 348 4 Dubray, B. et al. (1998) In vitro radiation-induced apoptosis and early response to low-dose radiotherapy in non-Hodgkin’s lymphomas. Radiother. Oncol. 46, 185– 191 5 Jansen, B. et al. (2000) Chemosensitization of malignant melanoma by BCL2 antisense therapy. Lancet 356, 1728– 1733 6 Symmans, W.F. et al. (2000) Paclitaxel-induced apoptosis and mitotic arrest assessed by serial fine-needle aspiration: implications for early prediction of breast cancer response to neoadjuvant treatment. Clin. Cancer Res. 6, 4610 – 4617 7 Meyn, R.E. et al. (1995) Apoptosis in murine tumors treated with chemotherapy agents. Anti-Cancer Drugs 6, 443 – 450 8 Ellis, P.A. et al. (1997) Preoperative chemotherapy induces apoptosis in early breast cancer. Lancet 349, 849 9 Adebodun, F. and Post, J.F.M. (1994) 31P NMR characterization of cellular metabolism during dexamethasone-induced apoptosis in human leukemic cell lines. J. Cell. Physiol. 158, 180 – 186 10 Nunn, A.V.W. et al. (1996) Characterization of secondary metabolites associated with neutrophil apoptosis. FEBS Lett. 392, 295 – 298 11 Williams, S.N.O. et al. (1998) Induction of apoptosis in two mammalian cell lines results in increased levels of fructose-1,6-bisphosphate and CDP-choline as determined by 31P MRS. Magn. Reson. Med. 40, 411–420 12 Ronen, S. et al. (1999) Magnetic resonance detects metabolic changes associated with chemotherapy-induced apoptosis. Br. J. Cancer 80, 1035 – 1041 13 Anthony, M.L. et al. (1999) Inhibition of phosphatidylcholine biosynthesis following induction of apoptosis in HL-60 cells. J. Biol. Chem. 274, 19686 – 19692 14 Blankenberg, F.G. et al. (1997) Quantitative analysis of apoptotic cell
Vol.24 No.3 March 2003
death using proton nuclear magnetic resonance spectroscopy. Blood 89, 3778 – 3785 Hakuma¨ki, J.M. and Kauppinen, R.A. (2000) 1H NMR visible lipids in the life and death of cells. Trends Biochem. Sci. 25, 357 – 362 Hakuma¨ki, J.M. et al. (1999) 1H MRS detects polyunsaturated fatty acid accumulation during gene therapy of glioma: implications for the in vivo detection of apoptosis. Nat. Med. 5, 1323– 1327 Di Vito, M. et al. (2001) 1H NMR-visible mobile lipid domains correlate with cytoplasmic lipid bodies in apoptotic T-lymphoblastoid cells. Biochim. Biophys. Acta. 1530, 47 – 66 Al-Saffar, N.M. et al. (2002) Apoptosis is associated with triacylglycerol accumulation in Jurkat T-cells. Br. J. Cancer 86, 963– 970 De Valck, D. et al. (1998) Differential activation of phospholipases during necrosis or apoptosis: a comparative study using tumor necrosis factor and anti-Fas antibodies. J. Cell. Biochem. 71, 392– 399 Le Moyec, L. et al. (1996) Proton nuclear magnetic resonance spectroscopy reveals cellular lipids involved in resistance to adriamycin and taxol by the K562 leukemia cell line. Cancer Res. 56, 3461– 3467 Santini, M.T. et al. (2001) The relationship between 1H-NMR mobile lipid intensity and cholesterol in human tumor multidrug resistant cell lines (MCF-7 and LoVo). Biochim. Biophys. Acta 1531, 111 – 131 Ucker, D.S. et al. (1992) Genome digestion is a dispensable consequence of physiological cell-death mediated by cytotoxic T-lymphocytes. Mol. Cell. Biol. 12, 3060 – 3069 Hakuma¨ki, J.M. et al. (1998) Quantitative 1H nuclear magnetic resonance diffusion spectroscopy of BT4C rat glioma during thymidine kinase-mediated gene therapy in vivo: identification of apoptotic response. Cancer Res. 58, 3791– 3799 Hortelano, S. et al. (2001) Intracellular water motion decreases in apoptotic macrophages after caspase activation. Cell Death Differ. 8, 1022– 1028 Scott, C.E. and Adebodun, F. (1999) 13C-NMR investigation of protein synthesis during apoptosis in human leukemic cell lines. J. Cell Physiol. 181, 147 – 152 Hakuma¨ki, J.M. et al. (2002) Early gene therapy-induced apoptotic response in BT4C gliomas by magnetic resonance relaxation contrast T1 in the rotating frame. Cancer Gene Ther. 9, 338 – 345 Zhao, M. et al. (2001) Non-invasive detection of apoptosis using magnetic resonance imaging and a targeted contrast agent. Nat. Med. 7, 1241 – 1244 Gadian, D.G. (1995) NMR and its Applications to Living Systems, Oxford Science Martin, S.J. et al. (1995) Early redistribution of plasma-membrane phosphatidylserine is a general feature of apoptosis regardless of the initiating stimulus – inhibition by overexpression of Bcl-2 and Abl. J. Exp. Med. 182, 1545 – 1556 Blankenberg, F.G. et al. (1998) In vivo detection and imaging of phosphatidylserine expression during programmed cell death. Proc. Natl. Acad. Sci. U. S. A. 95, 6349 – 6354
Conference Reporter American Association for Cancer Research Are you unable to attend the AACR 2003 on 5 – 9 April 2003? Don t miss out! Sign up for daily reports from the conference at: http://news.bmn.com/alerts Each day’s top stories, written by the BioMedNet Conference Reporter team, can be sent direct from the conference to your inbox in either text or html format. http://tips.trends.com