Spatio-Temporal Control of Gene Expression and Cancer Treatment Using Magnetic Resonance Imaging–Guided Focused Ultrasound Free

It has long been recognized that regional temperature elevation can be used for therapeutic purposes. Minimally invasive thermal therapies are increasingly used in the clinic for ablation of tumors using radiofrequency, microwave, or laser techniques (1). These heat sources have to be positioned at or near the target region. It is also possible to achieve local heating in a completely noninvasive way. Such local heating of tissues deep inside the body requires the use of a radiation beam with the following characteristics: (a) deep penetration in living tissue, (b) energy absorption by tissue transformed in heat, (c) absence of harm for tissues in the beam path, and (d) ability to focus on a small region. Within the full spectrum of electromagnetic radiation and sound waves, only ultrasound and radiofrequency waves are capable of noninvasively depositing energy deep inside the human body without harming tissue in the beam path. Because of its short wavelength, ultrasound can be focused into a small, well-defined area of interest with dimensions of ∼1 mm. In contrast, radiofrequency waves have a much larger wavelength and cannot be focused into a small region. Focused ultrasound is thus the only clinically viable technology that can be used to generate local hyperthermia noninvasively. The local temperature increase may serve a wide variety of medical interventions, such as ablation of tumoral tissue (2).

In addition, local hyperthermia has been suggested in gene therapy for control of transgene expression using heat-sensitive promoters (3). Together with the method of gene delivery, the spatial and temporal control of therapeutic transgene expression is among the key challenges in gene therapy. Furthermore, ultrasound has long been considered as a way to alter tissue permeability, to increase DNA uptake, and to affect gene expression (4, 5). The concept of sonoporation has been introduced in which the presence of ultrasound contrast agent and ultrasound tissue exposure is combined to create small holes in the cell membrane, allowing penetration of DNA carriers into the cell. Local hyperthermia has also been suggested for gene and/or drug delivery with thermosensitive microcarriers (6) and for heat-activated chemotherapy (7).

The potential of high-intensity focused ultrasound (HIFU) for tissue ablation was realized as early as 1942 (8). Technological developments (9–11) have led to successful experimental research (12, 13) and clinical trials in the field of oncology (e.g., ref. 14). The most advanced clinical application is the treatment of prostate cancer where the HIFU beam is delivered via a transrectal transducer. A recent study (15) reviewed the treatment efficacy in six HIFU clinical trials for the treatment of prostate cancer in a total of ∼1,000 patients. The results challenge those obtained with classic radiation therapy. The role of imaging was limited to the preparation phase and the posttreatment evaluation in these studies. However, no real-time image guidance of the therapy was realized. In this respect, it is especially important that thermal damage shows threshold-like behavior, with necrosis above and full recovery below the critical thermal dose (see below), unlike radiation therapy where the cumulative dose in nearby healthy tissue is often a limiting factor. Magnetic resonance imaging (MRI)-guided focused ultrasound was recently approved by the Food and Drug Administration for treatment of uterine fibroids (16), which has accelerated interest in this technology for cancer treatment.

Because the primary effect of HIFU is thermal, it is important to evaluate and even control temperature evolution during the treatment. Temperature evolution is a function of heat deposition and heat dissipation and may be modeled with the bioheat transfer equation (17). Both variables (deposition and dissipation) can be spatially heterogeneous and may be affected by temperature increases. Heat deposition by HIFU depends on the local absorption of ultrasound waves. Pure water and blood hardly absorb ultrasound energy. The absorption of ultrasound by soft tissue depends on composition and on the wavelength of the transmitted ultrasound energy. Therefore, accurate quantification of energy deposition by focused ultrasound and absorption by tissue is difficult to predict. Heat conduction and dissipation also depend on tissue composition, diffusion, and perfusion processes that may vary locally as a function of tissue architecture and tissue composition. Physiologic events, such as temperature-dependent perfusion increases, may play a role (18). In the case of ablation procedures, tissue coagulation may significantly modify heat conduction as well as energy absorption. As a consequence, heat losses and energy absorption are difficult to predict in advance of the procedure. Therefore, the performance of HIFU therapeutic heating can be expected to be improved significantly with real-time evaluation of the temperature distribution profile by providing real-time feedback to assure adequate heating throughout the area of interest. Of the different imaging modalities, MRI is an excellent tool for guiding the noninvasive treatment by hyperthermia. Particular advantages of MRI are that the technique provides temperature mapping (19) as well as target definition and may even provide an early evaluation of therapeutic efficacy. Based on such considerations, Cline et al. (20) realized early on that the combination of MRI and HIFU would be a very promising tool. Subsequently, the technology was further developed, with prototype HIFU systems integrated with a whole-body MRI system (21–23).

This article describes an overview of the principles of MRgHIFU, the important advances made since the beginning of the technology, the translational aspects, and novel applications in gene expression and gene delivery.

Ultrasound is a form of mechanical energy that is propagated as a vibrational wave of particles within the medium at a frequency between about 20 kHz and 20 MHz. The oscillatory displacement of particles is associated with a pressure wave. The behavior of ultrasound waves in a medium may be described in a similar way to that in optics. At tissue interfaces, reflection of the wave occurs according to Snell's law and depends on the wave velocity and the incident angle. The speed of ultrasound is ∼1,550 ms⁻¹ for soft tissue, independent of the ultrasound frequency. In fatty tissue, the average speed is only slightly lower (1,480 ms⁻¹), whereas in air spaces a value of 600 ms⁻¹ is found. In bone, the speed is much higher (between 1,800 and 3,700 ms⁻¹; ref. 24). Ultrasound beams are usually generated electrically using piezoceramic plates outside the body and propagate inwards either longitudinally (as is the case for hyperthermia purposes) or transversely (called shear waves). Focusing of the longitudinal waves can be achieved using a spherical curvature of the plates, with a suitably designed lens system, or with a multielement transducer with independent phase control for each element (phased array transducer, see below). The interference of the multiple waves leads to an interference pattern. In a homogeneous medium, the phase of all waves is identical at the focal point, whereas outside the focal point the differences in phase lead to destructive interference and thus attenuation of the pressure wave. The minimum diameter of the focal area is thus given by half the wavelength λ, which is equal to the speed divided by the frequency. For example, when using 1.5 MHz ultrasound, the wavelength is ∼1 mm in soft tissue.

Ultrasound energy is attenuated in tissue due to processes of absorption and scattering. The mechanisms of ultrasound absorption in tissue have been reviewed previously (25, 26). The efficient energy absorption in soft tissue may be related to friction between and in structures with dimensions of the order of the wavelength used. Overall, the attenuation of a plane wave may be described by an exponentially decaying function. For 1.5 MHz ultrasound waves, the intensity in muscle will drop to ∼50% at 50-mm penetration. The smallest volume that may be treated by HIFU has dimensions of half the wavelength, which is typically ∼1 mm. In homogeneous medium (ultrasound propagation speed does not vary), high-precision treatment with margins of 1 mm may be reached (muscle). In inhomogeneous medium (e.g., for bone structures and air in the beam path), the multiple wave reflections make it difficult to focus ultrasound waves in the area of interest because the wave phase and attenuation will be different according to their pathway from transducer to target. The focal point may be less well defined in such cases. When using a multiple-element transducer, individual amplitude and phase modification is feasible for each element and, in principle, high precision can be realized. The problem remains how to adjust each transducer element to accomplish focusing. Modeling the tissue composition and interfaces based on high-resolution imaging may provide sufficient data for beam focusing (27–29). An alternative solution is based on the so-called principle of temporal return of waves (30). The technique relies on using first a short duration ultrasound burst from the area of interest (using either a small transmitter or an echogenic particle) and measuring the reflected waves at several locations outside the body. A temporal return of the received wave pattern, corrected for attenuation effects, leads in turn to exact focusing in the area of interest. The need to first use a transmitter in the area of interest may limit its practical use. Recently, phased array transducer technology was used to focus ultrasound through the skull into brain tissue (29). Therefore, neurologic applications of focused ultrasound seem technologically feasible, although further developments are still necessary to accomplish focusing in a noninvasive way and to avoid any risk of heating at or near the skull. A detailed recent review on focused ultrasound technology can be found elsewhere (31).

Ideally, a thermotherapeutic treatment is preceded by precise three-dimensional target definition and localization of nearby tissue that must be spared. The treatment should be done with continuous temperature monitoring of target tissue and nearby tissue to be spared. The therapy will be continued until the appropriate thermal dose (32) is delivered to the entire target volume. Then, a first assessment of efficacy and possible complications should be carried out.

MRI offers specific advantages for each part of this procedure. It is well known that MRI offers superb soft tissue contrast, either without or with MR contrast agents. Tissue boundaries can be identified with high precision. The known coordinates can thus be used directly for image-guided therapeutic procedures.

During the thermal procedure, continuous temperature imaging can be done with MRI using specific pulse sequences and data processing. Noninvasive, three-dimensional mapping of temperature changes is feasible with MR and is based on either the relaxation time T₁ (19), the diffusion coefficient (33), or proton resonance frequency (PRF; ref. 34) of tissue water. The use of temperature-sensitive contrast agents can provide absolute temperature measurements (35). The principles and performance of these methods have been reviewed recently (36, 37). The excellent linearity and near independence with respect to tissue type (38), together with good temperature sensitivity, make PRF-based temperature MRI the preferred choice for many applications at mid to high field strength (≥1 T). The PRF methods use RF-spoiled gradient echo imaging methods (39). A SD of less than 1°C, for a temporal resolution below 1 s and a spatial resolution of ∼2 mm, is feasible for a single slice for immobile tissues. Corrections should be made for temperature-induced susceptibility effects in the PRF method (40). Fat suppression is necessary as well as excellent registration to correct for displacements between scans because changes with respect to a reference are mapped. Motion artifacts can severely degrade the accuracy of MR temperature maps and must be carefully corrected (37).

MRI may also assess therapeutic efficacy and possible complications. For example, a hemorrhage can be shown based on T2* changes. Edema formation can be described by T2 changes. In addition, diffusion and perfusion changes can be identified with specific MR methods. Therefore, MRI will play a large role for in situ assessment of tissue changes. However, it should be noted that many apoptotic processes may take up to 24 h to become clearly detectable. Therefore, efficacy assessment immediately following therapy may be of rather limited use.

For guidance by MRI, the transducer is placed on or inside the patient bed. Because of the very high magnetic field, the use of nonferromagnetic materials is mandatory. The confined space within the bore of the magnet makes it difficult to design a system that provides comfort to the patient as well an uninterrupted pathway for the ultrasound waves from transducer to target. Recently, a complete system has been described for ablation of uterine fibroids (16, 41) as well as adaptations for treatment of breast tumors (42–44). Figure 1 shows a clinical example of the MRgHIFU treatment of uterine fibroids.

Optimal control of the temperature-based treatment requires regulation of the temperature. Recent developments have shown that rapid MRI followed by online data processing, and real-time feedback to the focused ultrasound output (45), combined with new temperature regulation algorithms, may provide such control. The regulation of temperature evolution at the focal point has been described and is based on temperature mapping and a physical model of local energy deposition and heat conduction. The temperature regulator was based on a classic proportional, integral, and derivative type of temperature control. Thus, temperature evolution at the focal point can be regulated automatically with a precision that is close to that of the precision of the temperature measurements at the focal point.

Spatial control of temperature evolution during HIFU hyperthermia is necessary to guarantee the desired thermal dose achieved within the complete volume of interest. Treatment of large volumes using a fixed focal point transducer was initially based on MR guidance of the heat treatment at a single point followed by displacement of the HIFU focal point and repeating the procedure until the target volume was covered (46). Such a procedure necessitates long treatment times and may leave “gaps” between treated locations. Recently, an approach was presented where the focal point was moved along multiple inside-out spiral trajectories covering the target region under continuous HIFU power to achieve a minimal treatment time (47). A complete proportional, integral, and derivative control in the spiral plane provided stability and temperature uniformity within a large target region. Figure 2 shows a preclinical application of regional MRgHIFU treatment of VX2 tumors in a rabbit. Such methods will be even more useful when combined with phased array transducers that allow very fast, electronically controlled, motion of a focal point. It should be kept in mind that simultaneous treatment of the large area of interest may lead to excessive near-field and/or far-field heating and accumulation of undesired thermal dose in anterior and posterior areas with respect to the target region. In addition, any motion of the transducer will lead to small modifications of the magnetic field even with the use of nonferromagnetic material. These motion-related factors need to be corrected for, particularly when using the PRF method for temperature mapping.

The automated MRI feedback coupling to the HIFU heating unit relies on high-speed temperature imaging. It has been shown that motion can severely degrade MR temperature mapping. Correction of motion artifacts and the acceleration of temperature mapping are current topics of much interest in the further development of MRI-guided HIFU. Recently, several approaches have been proposed to address these issues, ranging from the use of respiratory and/or cardiac triggering (49), the use of navigator echoes, and advanced postprocessing methods (37).

Among the key challenges in gene therapy are the method of gene delivery and the spatial and temporal control of therapeutic (trans)gene expression in the targeted tissue. The ability of HIFU to heat tissue deep inside the body can be used to control transgene expression when the gene is placed under the control of a heat-sensitive promoter. Such promoters exist widely in nature. For example, all mammals use such promoters when they have a fever to produce stabilizing proteins, the so-called heat shock proteins. The use of heat-sensitive promoters to control the expression of therapeutic genes requires tight temperature regulation in the region of interest. The heat shock promoters, especially the hsp70 promoters, have been used often in gene therapy strategies because they are both heat inducible and efficient in initiating and regulating transcription of therapeutic genes (50). The promoter of one of the inducible heat shock protein systems, HSP70, can activate gene expression several thousand-fold in response to hyperthermia (51). The minimal hsp70 promoter is almost exclusively sensitive to temperature, and it was suggested several years ago to use the hsp70 promoter and local hyperthermia could be used to control transgene expression (3). The feasibility of the approach has been shown both in vitro and in vivo (52–55), including the expression of suicide genes in a transduced mammary cancer cell line (56) and the expression of a reporter gene in transfected muscle (57) or in transduced C6 cells (58). It has been shown that local overexpression of a marker gene in a modified glioma cell line is feasible using this technology. Local control of transgenic expression has recently been applied to modified stem cells. Mesenchymal stem cells (MSC) obtained from bone marrow of rats were transfected with hsp-luc expression of the luciferase gene under the control of an hsp70 promoter. The transformed MSCs were injected in the left renal artery and transgenic expression was induced by MRI-controlled HIFU hyperthermia (59). Figure 3 shows a summary of the histologic findings.

The use of MRI for monitoring ultrasound to induce gene expression has been shown in other organs. The feasibility of using ultrasonic heating to control transgene expression spatially using a minimally invasive approach was investigated in the prostate. Ultrasound imaging was used to guide the injection of an adenoviral vector containing a transgene encoding firefly luciferase under the control of the human hsp70B promoter into both lobes of the prostate gland in three beagle dogs. After injection, the left lobe of the prostate was heated using an ultrasound transducer and using a MRI guidance system. The hsp promoter allows induction of the associated transgene only in areas that are subsequently heated after transfection. High levels of luciferase expression were observed only in areas exposed to ultrasonic heating (60).

The feasibility of the use of focused ultrasound to induce hyperthermia in the liver of a rat model to focally induce the expression of green fluorescent protein gene under the control of an hsp70 promoter was examined in a preliminary study (61). Temperature was noninvasively monitored by temperature-sensitive MRI. The study showed localized gene induction within the liver parenchyma. There was a good correlation with MRI and histology, which showed that the variables introduced to spatially control gene induction within a parenchymal organ, such as the liver in rats using focused ultrasound under control of MRI, are feasible.

The interaction of ultrasound with tissue also leads to radiation pressure and cavitation. These phenomena have been used for local gene delivery and transfection (see ref. 62 for review). Focused ultrasound exposure may lead to permeabilization of plasma membranes, also known as sonoporation. The effect is generally attributed to cavitation phenomena and may be enhanced with the use of ultrasound contrast agents. It has been shown that sonoporation leads to increased DNA uptake by the cells (63). Because cavitation activity is highly localized, the combination of ultrasound and microbubble contrast agents provides a means for noninvasive, site-specific gene therapy. Many efforts to develop more efficient nonviral and viral particles for gene delivery have been reported where ultrasound exposure has been shown to enhance transgene expression (64–68). Because microbubble cavitation effects primarily originate in the vasculature, vascular therapy will be an important field for these applications. Similar principles can be applied to achieve local drug delivery using microcarriers containing drugs. The field has recently attracted much attention because it represents a strategy to increase the drug concentration at the target location and decrease systemic toxicity effects. Ultrasound can be used in different ways to trigger regional drug delivery (69). It can cause the local drug release from a carrier vehicle and the local increase of cell membrane permeability either by a mechanical action or by a temperature increase.

MRI-guided HIFU is a promising new technology for gene therapy and cancer treatment. The benefits of focused ultrasound and associated bioeffects, such as local temperature elevation and cavitation effects, are well known. The recent guidance by MRI offers enhanced control, safety, and evaluation of therapeutic efficacy. The application of MRgHIFU in gene therapy is very promising. Temperature-mediated control of gene expression is feasible by combining MRgHIFU with the use of a thermosensitive promoter to regulate gene transcription and expression. Enhanced gene delivery is possible by increasing cell membrane permeability using HIFU cavitation effects. Although MRgHIFU is a fairly complicated technology, it is currently under very active development. The installed base is still small, and it may therefore take several years to realize the full potential of MRgHIFU in the cancer field. Regulatory hurdles for widespread application of MRgHIFU in cancer are significant. However, the method has received Food and Drug Administration approval for uterine fibroid ablation, which is expected to accelerate clinical use in the cancer field. Because treatment of the complete tumor tissue has to be assured (unlike in the case of uterine fibroids), further technical requirements of the MRgHIFU technique may be necessary, such as the use of more advanced phased array transducers, accelerated temperature mapping and feedback coupling to the HIFU device, as well as volumetric heating approaches.