Thermosensitive paramagnetic liposomes for temperature control during MR imaging-guided hyperthermia: In vitro feasibility studies

Thermosensitive Paramagnetic Liposomes for Temperature Control during MR Imaging-guided Hyperthermia: In Vitro Feasibility Studies 1 Sigrid L. Fossheim, PhD, Kamil A. I1' yasov, PhD, Jfirgen Hennig, PhD, Atle Bjornerud, MSc

Rationale and Objectives. Magnetic resonance (MR) imaging-based temperature monitoring has gained interest for use in general hyperthermia treatment of tumors. Such therapy requires an accurate control of the temperature, which should range from 41 ° to 45°C. A novel type of thermosensitive MR agent is proposed: liposome-encapsulated gadolinium chelates whose temperature response is linked to the phase-transition properties of the liposome carder. In vitro relaxometry and MR imaging were used to evaluate the thermosensitivity of the contrast properties of liposomal gadolinium diethylenetriaminepentaacetic acid bis(methylamide) (Gd-DTPA-BMA). Materials and Methods. T1 relaxivity (rl) measurements of liposomal Gd-DTPA-BMA were undertaken at 0.47 T and at temperatures of 20°--48°C. MR imaging was performed at 2.0 T with a gel phantom containing inserts of liposomes. Diffusion-weighted and Tl-weighted gradient-recalled echo images were acquired as the phantom was heated from 22 ° to about 65°C. Results. At ambient temperature, the rl of liposomal Gd-DTPA-BMA was exchange limited due to slow water exchange between the liposome interior and exterior. A sharp, marked increase in rl occurred as the temperature reached and exceeded the gel-to-liquid crystalline phase-transition temperature ( T ) of the liposomes (42°C). The relaxation enhancement was mainly attributable to the marked increase in transmembrane water permeability, yielding fast exchange conditions. There was good correlation between the relaxometric and imaging results; the signal intensity on Tl-weighted gradient-recalled echo images increased markedly as the temperature approached T . The temperature sensitivity of the diffusion-weighted technique differed from that of the liposome-based Tl-weighted approach, with an apparent water diffusion coefficient increasing linearly with temperature. Conclusion. Since the transition from low to high signal intensity occurred in the temperature range of 38°-42°C, the investigated paramagnetic liposomes have a potential role as "off-on" switches for temperature control during hyperthermia treatment. Key Words. Hyperthermia; paramagnetic liposomes; magnetic resonance; temperature monitoring; gel-to-liquid crystalline phase transition temperature; "off-on switch."

The utility of magnetic resonance (MR) imaging for temperature monitoring during hyperthermia treatment is

Acad Radiol 2000; 7:1107-1115

From Nycomed Imaging AS, PO Box 4220, Torshov 0401 Oslo, Norway (S.L.F., A.B.); and the Department of Radiology, University of Freiburg, Freiburg,Germany (K.A.I., J.H.). Received January 10, 2000; revision requestedMarch 27; revision received and accepted July 17. Address correspondence to S.L,F. ©AUR, 2000

well described in the literature (1-4). General hyperthermia treatment is based on the induction of a relatively limited temperature elevation (approximately 43°C) within a large tissue volume in order to exploit the differential temperature sensitivity of neoplastic versus normal cells (5). Temperature sensitive parameters, such as the T1 relaxation time, proton chemical shift, and the water diffusion coefficient, have been used to gain temperature information from MR images or spectra (2,4,6--8). All these techniques, however, have some limitations. The main difficulty with Tl-based

1107

"OSSHEIM ET AL

methods is the complex relationship between temperature variations and alterations in tissue T1 (2). Also, above a certain critical temperature, protein denaturation may occur and cause irreversible decrease in tissue T1 (2). The proton chemical shift method is hindered by insufficient temperature sensitivity at clinical field strengths and places stringent requirements on the magnetic field homogeneity (2,8). Furthermore, physiologically related magnetic susceptibility changes may limit the utility of phase-based chemical shift methods (9). Self-referenced spectroscopic imaging techniques are insensitive to susceptibility changes and patient motion between scans but are more time-consuming than the phase-based chemical shift method (10). Diffusion-weighted (DW) MR imaging techniques are also inherently sensitive to motion artifacts, which limit their clinical potential (11). Additionally, diffusion-based methods require a well-performing gradient system, which is generally not the case for clinical MR scanners, and their accuracy is affected by the poor signal-to-noise ratio (12). The feasibility of MR temperature measurements has been demonstrated recently by the use of paramagnetic chelates whose chemical shift properties are strongly temperature dependent (13-15). Other investigated thermally sensitive agents have been proton-based and perfluorinated liquid crystals (16-19). The use of thermosensitive liposomes as drug delivery systems is of increasing interest in tumor therapy (20,21). Such liposomes have the ability to release liposome-encapsulated drug at the gel-to-liquid crystalline phase-transition temperature (Tn) of the liposomes. At this temperature, distinctive structural changes occur in the liposome bilayer as the latter converts from a gel-like to a liquid state, resulting in a markedly increased transmembrane water permeability (22,23). The Tmof liposomes can be tailor made by selecting and combining phospholipids appropriately. In this study, the use of paramagnetic liposomes for temperature control is proposed. Previous relaxometric studies with liposome-encapsulated gadolinium chelates have shown a low T1 relaxivity (rl) at temperatures below the T , which was attributed to an exchange limitation of the dipolar relaxation process (24). As T was reached, however, the rl increased markedly due to a faster water exchange between the interior and exterior of the liposome. Depending on the phospholipid composition, the gadolinium chelate may also be released into the surrounding medium, contributing to the marked relaxivity enhancement. In this work, the in vitro temperature sensitivity of liposome-encapsulated gadolinium diethylenetriaminepentaacetic acid bis(methylamide) (GdDTPA-BMA) was evaluated by relaxometry and MR imaging.

1108

Academic Radiology, Vol 7, No 12, December 200(

Liposome Preparation

Liposomal Gd-DTPA-BMA was prepared by means of the thin-film hydration method (24,25). Briefly, the phospholipids dipalmitoylphosphatidylcholine (DPPC) and dipalmitoylphosphatidylglycerol (DPPG) sodium salt (Sygena Ltd, Liestal, Switzerland) were codissolved in a 5:1 (vol/vol) chloroform-methanol mixture (phospholipid-to-organic solvent ratio, 5 mg/mL). The phospholipid solution was rotary evaporated to dryness, and the resulting lipid film was further dried under vacuum overnight. Liposomes were formed by hydration of the phospholipid film with a 250 mmol/L Gd-DTPA-BMA solution (diluted Omniscan; Nycomed Imaging AS, Oslo, Norway). The concentrations of DPPC and DPPG were 47.5 and 2.5 mg/mL, respectively. The liposomes were allowed to swell for 2 hours before sequential extrusion (Lipex Extruder; Lipex Biomembranes, Vancouver, British Columbia, Canada) through polycarbonate filters of various pore diameters (Nuclepore; Costar, Cambridge, Mass). The phospholipid film hydration, liposome swelling, and extrusion were performed at 55°C, well above the Tmof the DPPC/DPPG blend (see "Physicochemical Characterization" section for measured T values). Untrapped metal chelate was removed with dialysis (Spectra/ Por membrane tubing [molecular weight cutoff, 10,000 daltons], Spectrum, Houston, Tex) against isosmotic glucose solution. Before use, the solutions of glucose and Gd-DTPABMA were adjusted to a pH of 7.4 by addition of minute amounts of 1 mol/L sodium hydroxide. Physicochemical Characterization

The physicochemical properties of the dialyzed liposome preparations are given in the Table. For the determination of size and electrophoreticxnobility, liposomes were diluted with pH-adjusted isosmotic glucose solution. The intensityweighted liposome diameter was determined by means of photon correlation spectroscopy at a scattering angle of 90 ° and at 25°C (Malvem PS/MW 4700; Malvern Instruments, Malvern, England). The width of the liposome size distribution was expressed by the polydispersity index. The electrophoretic mobility was determined at 25°C by means of laser Doppler velocimetry (ZetaSizer IV, Malvern Instruments; Coulter DELSA 440, Coulter Electronics, Hialeah, Fla). The T of the liposome preparations was measured with differential scanning calorimetry (DSC 7; Perkin Elmer, Norwalk, Conn). The effective gadolinium concentration (in the total sample volume) (Ceef)was determined with inductively coupled plasma atomic emission spectrophotometry

~.cademic Radiology, Vol 7, No 12, December 2000

THERMOSENSITIVE PARAMAGNETIC LIPOSOME5

Physicochemical Properties of Liposomal Gd-DTPA-BMA (Dialyzed Preparations)

Liposome Composition

DPPC/DPPG

Experiment

Effective Gadolinium Concentration (mmol/L)

Liposome Diameter *t (nm)

Transition Temperature (°C)

Relaxometry Relaxometry MR imaging

16.6 38.6 31.7

126 187 276

42.4 42.6 41.6

Electrophoretic Mobility* (gm. cm • V -~ . sec 1) -3.4 -2.6 -3.0

Note.--AII values are given as means of two or more measurements, *Measured in isosmotic glucose solution (pH 7.4), at 25°C. tPolydispersity index ranged from 0.10 to 0.17.

1

O

I

where R1 °bs and RI m are the relaxation rates (1/seconds) at a given temperature of the liposome sample and matrix (glucose solution), respectively, and Ceff is the effective gadolinium concentration (in millimoles per liter). For comparative purposes, the rl-temperature dependence of GdDTPA-BMA was determined in isotonic 5% (wt/vol) glucose solution (glucose 50 mg/mL; B. Braun, Melsungen, Germany).

2

O

3

4

O

O

5

6

# =O7

o o/ioo

~

010

Figure 1. Gel phantom containing inserts with liposomal GdDTPA-BMA (labeled 3-10) and glucose solution (labeled 1, 2). Inserts 1 and 2 were used as controls.

(ICP-AES), as described later. For relaxation measurements, liposomes were further diluted with pH-adjusted isosmotic glucose solution to give a suitable range of Ce~ values. In Vitro R e l a x o m e t r y The influence of temperature on the rl of two differentsized liposome preparations was investigated at 0.47 T (Minispec PC-120b, Bruker, Rheinstetten, Germany). The studied temperature range was 20°-48°C. The T1 relaxation rates (R1) were obtained by using the inversion-recovery method. The rl at any temperature was calculated from the following relationship: R 1 °b~

rl

_

%

R1m ,

(1)

MR Imaging The experiments were carried out on a 2.0-T wholebody clinical system (Medspec S 200 Avarice; Bruker) equipped with an actively shielded head gradient insert (maximum gradient amplitude, 30 mT/m; rise time, 200 sec). Imaging was performed on a gel phantom (containing ultrasound contact gel) in which four tubes containing liposomes and two containing glucose solution were inserted and to which four additional liposome tubes were fixed on the right side of the phantom (Fig 1). The inner diameter of the sample tubes was 10 ram. A 45-watt microwave heating was performed at 434 MHz with a linear radiofrequency (RF) antenna placed in the phantom. The microwave irradiation was applied simultaneously with the image acquisition. The total heating time was 63 minutes, for a temperature increase in the gel of up to 45°C above room temperature (22°C). Tl-weighted gradient-recalled echo (GRE) images were acquired with the following parameters: 30/4 (repetition time [TR] msec/echo time [TEl msec, 50 ° flip angle, 8-mm slice thickness, 25.6-cm field of view, and 256 x 256 matrix. DW images were obtained by employing a single-shot DW rapid acquisition sequence with relaxation enhancement (RARE) (26). Parameters of diffusion weighting were optimized for gel, control inserts, and inserts containing liposomes, the latter exhibiting a short T2 (due to the presence of susceptibility effects).

1109

Therefore, a stimulated echo scheme was used in the DW preparation period. The duration (6) of diffusion-encoding gradients was limited to 7 msec, and the diffusion time (A) was 120 msec. A single-shot RARE sequence ( T E ~ = 6.7 msec) with central phase encoding was used in the readout, resulting in an effective TE (TEeff) of 20.7 msec. Other parameters were as follows: 4,000-msec TR, 8-ram slice thickness, 25.6-cm field of view, 128 x 128 matrix. Different diffusion weighting was achieved by varying the amplitude, G, of the diffusion-encoding gradients, while 6 and A were kept constant. Ten DW images were acquired, with a b factor varying from 5 to 364 sec/mm 2 were acquired (the highest value of the b factor was slightly lower than the optimal value due to the short T2 of the liposome samples). The apparent diffusion coefficient (ADC) maps were calculated pixel per pixel by fitting to this equation:

(2)

SIi = SI 0 exp (-b,~ ADC),

where SI i is the signal intensity measured with a DW gradient of amplitude G i (i = 0--+10), SI0 is the signal intensity in the absence of diffusion attenuation, and the b factor, b i, was calculated numerically by taking into account all (ie, both diffusion-weighting and imaging) gradients (27). The set of Tl-weighted and 10 DW images was repeated every 3.5 minutes during the total heating time. The temperature was estimated by assuming a linear dependence of ADC on temperature (7):

,r cJ ,/ ATj

= Ti - To

= ~ [ADC0

,

(3)

where A D C corresponds to a measurement, j, and ADC 0 is the baseline measurement at room temperature, T O. The thermal coefficient, c~, was found to be 2.5% + 0.2%/°C from the calibration experiments. Because of the low viscosity and possible convection in the presence of temperature gradients, the ADC within the liposome samples could not be directly assessed. The sample temperature at any time point was approximated from the temperature of surrounding gel (only the temperature in the imaged horizontal plane was taken into account). Neglected were the errors arising from convection in the sample tubes, resulting in heat transfer from the layers below and above the imaged plane. After termination of the experiment, the temperature within the inserts was measured with a thermocouple.

ICP-AES Analysis For the determination of C~, liposomes were diluted with a 0.2% (vol/vol) solution of Triton X-100 (Sigma

1110

[--I~- LiposomalGd-DTPA-BMA,126nrn LiposomalGd-DTPA-BMA,187nm NonliposomalGdoDTPA-BMA

"-"

10

"~- 3

~___.

1

0

20

25

30 35 40 45 50 Temperature (°C) Figure 2. Temperature dependence of the rl of liposomal and nonliposomal Gd-DTPA-BMA (0.47 T). Note that the semilogarithmic plot (inserted figure) emphasizes the rl-size dependence at temperatures below -/-~.

Chemical, St Louis, Mo). The gadolinium concentration in the samples was determined by using a multipoint standard calibration curve (Plasma 2000 ICP-AES, Perkin Elmer).

In Vitro Relaxometry Figure 2 shows the rl of both liposomal and nonliposomal Gd-DTPA-BMA as a function of temperature. The rl of nonliposomal Gd-DTPA-BMA decreased from 4.8 to 3.6 sec -~. (mmolfL)< in the temperature range of 20°48°C. The rl of liposome-encapsulated Gd-DTPA-BMA was markedly reduced at 20°C compared with that of the metal chelate itself. The liposome size influenced the rl, the 126nm liposomes being more efficient relaxation enhancers than the 187-nm liposomes. The rl of both liposome preparations increased in the 20°-37°C range, from 0.054 to 0.53 sec-1 • (mmol/L) ~and from 0.020 to 0.21 sec-~- (mmol/L)-1 for the 126- and 187-nm liposomes, respectively. After liposome cooling from 37 ° to 20°C, the rl returned to the same initial low values. Further heating above 40°C increased the rl markedly, with the values approaching the rl of nonliposomal Gd-DTPA-BMA. Upon liposome cooling from 48 ° to 37°C, rl values of 0.75 and 0.66 sec -I - (retool/ L ) 1 were obtained for the 126- and 187-nm liposomes, respectively. These relaxivities were 1.4- and 3.0-fold higher

a,

C.

b=

d.

Figure 3. Tl-weighted GRE images (2.0 T) of gel phantom containing liposomal Gd-DTPA-BMA (a) before and after (b) 47 and (c) 63 minutes of RF heating. Inhomogeneous SI in gel is due to air bubbles. (d) Difference image after subtraction of a from c. Note that the Sis within inserts 3 and 5 are almost unchanged after RF heating.

than the initial relaxivity values (prior to heating). The liposome size was not affected by the heating procedure.

MR Imaging At ambient temperature, the Tl-weighted GRE image of the gel phantom showed a low signal intensity (SI) for all Iiposome samples (Fig 3a); the SI was also lower relative to that of the gel matrix and inserts containing glucose solution. Heating caused a local temperature increase near the RF antenna. After about 40 minutes, the samples closest to

the RF antenna (inserts 6-10) showed a marked SI increase on the Tl-weighted image (Fig 3b). From approximately 55 minutes, the next sample farther from the antenna (insert 4) also displayed a pronounced SI increase. The corresponding ADC map at that time showed a bright diffuse spot within the gel in the vicinity of the affected samples, consistent with an increased ADC (image not shown). Figure 3c shows the final Tl-weighted GRE image after 63 minutes of RF heating. Figure 3d, representing the difference image after subtraction of baseline image (Fig 3a) from Figure 3c, reflects the relative change in SI after heating. The calculated ADC-based temperature map and profile of the gej phantom after heating is given in Figure 4a and 4b, respectively. Figure 5 shows the SI-temperature response on Tl-weighted GRE images for the liposome sample nearest the RF antenna (insert 6). A sharp and welldefined transition from low to high SI was obtained in the temperature range of 37.6°-41.9°C, the SI increasing 2.5fold. Such a temperature response was different from the gradual, relative small SI decrease on Tl-weighted GRE images of the heated gel region (results not shown) or the linear and slight temperature dependence of the ADC (Fig 6). After termination of the imaging experiment, the temperature within inserts 1, 2, 3, and 5 was well below T , but it exceeded T in inserts 4 and 6-10.

Proton-based and perfluorinated liquid crystals have shown utility for temperature monitoring (16-19). The use of such compounds exploits the change in MR properties occurring at their phase transition. In such cases, the T2 is short when the substance is in the solid state but increases markedly as the substance melts and becomes a liquid. The result is a transition from low to high SI on an MR image, that is, an "off-on switch." Franklin et al (17) investigated the relaxation properties of encapsulated proton-based liquid crystals undergoing a phase transition from the smectic to the nematic phase. The authors concluded that the approach would be clinically irrelevant, as the temperature-induced changes in SI were masked by the rather large background signal of the tissue water protons. Various mixtures of perfluorocarbons have also been investigated as thermosensitive probes (18,19). The results were promising, as large changes in fluorine SI were observed at the solid-to-liquid phase transition with high thermal sensitivity and no background signal. Most notably, an approximately 20-fold increase in fluorine SI was obtained over a 7°C temperature interval (37 °. 44°C) (19). Analogously, the electron

1111

:OSSHEIM ET A L

Academic Radiology, Vol 7, No 12, December 2001

/a

Insert 6

50 45

Insert 4 - -

O

)K

0)

19

E

30 &

i

T Cross-Section position a.

1

0

,

20

40

60

80

i

i i

100

i

120

Position, Pixel No. b.

Figure 4. (a) ADC-based temperature map after 63 minutes of RF heating; incorrect temperature data within inserts are due to short relaxation times and convection artifacts. (b) Temperature profile across section indicated in a; temperature measured by thermocouple within inserts 4 and 6 is indicated by asterisks.

paramagnetic resonance spectrum of a nitroxide free radical, dissolved in a fatty acid medium, showed a marked alteration between 38 ° and 48°C, a temperature range in which the medium underwent a solid-to-liquid phase transition (28). In the "design" of a thermosensitive liposomal contrast agent, however, the phase transition properties of the liposome carrier are exploited, and not those of the contrast material itself. Such a strategy allows some flexibility, as it is possible to design, by an appropriate combination of phospholipids, liposomes displaying a T within any desired temperature range relevant for hyperthermia treatment. The in vitro experiments reported in this study demonstrated the feasibility of DPPC-based paramagnetic liposomes for temperature monitoring. A good correlation was observed between the relaxometric and Tl-weighted imaging results. At ambient temperature the liposomal water permeability was too low to allow for efficient water exchange across the liposome membrane. Hence, the dipolar relaxation process was exchange limited, as confirmed by the size dependence of rl. Also, the compartmentalization of a high concentration of gadolinium within the liposome interior generated susceptibility effects, that is, the liposome system was an efficient T2* agent2 (24). Therefore, at ambient temperature both the low rl and the susceptibility effect of liposomal Gd-DTPA-BMA were reflected by a low SI on the Tl-weighted GRE images.

1112

30 I +

Liposomai

Gd-DTPA-BMA

I

25

20

< 10

5

. . . .

20.0

i

. . . .

25.0

i

. . . .

30.0

i

. . . .

35.0

i

. . . .

40.0

i

. . . .

45.0

50.0

Temperature (°C) Figure 5. Temperature dependence of the SI (quantified in arbitrary units) of liposomal Gd-DTPA-BMA on Tl-weighted GRE images (2.0 T) (results from insert 6).

Raising the temperature to 37°C increased the liposomal water permeability, which improved the rl but not enough to cause a marked SI increase on the MR image. Leakage of gadolinium chelate could be discarded as a possible cause of the elevated relaxivity, as the rl values after cooling from 37 ° to 20°C were identical to those before liposome heating.

~,cademic Radiology, Vol 7, No 12, D ecember 2000

THERMOSENSITIVE

0.30 0.28 G" 0.26 Eo 0.24 O

x 0.22 O cl < 0.2O 0.18 0.16 20

25

30

35

40

45

50

Temperature (°C) Figure 6, Temperature dependence of the ADC in gel matrix (region of interest close to insert 6).

The sharp and pronounced increase in rl as the temperature approached T could be explained by the occurrence of thermotropic changes in the membrane. For instance, at T , the DPPC bilayer has been shown to expand laterally by about 40% and to decrease its hydrophobic thickness with 20%, features that markedly increased the membrane permeability to water and solutes (22,23). Hence, the water exchange between the interior and exterior of the liposome was fast enough to allow for efficient propagation of the dipolar relaxation effect, thereby markedly increasing the rl. The dipolar T2 contribution was also enhanced to override the susceptibility effect; the liposome system could thus be regarded as a T1 agent. On Tl-weighted GRE images, the marked SI increase over a 4°C range (38°-42°C) correlated well with the rl-temperature profile. At temperatures above T, the relaxation enhancement was in the fast exchange regime, as evidenced by the similar rl values of liposomal and nonliposomal Gd-DTPA-BMA (24). As T was reached and exceeded, however, leakage of gadolinium chelate from the liposome also contributed to the increased relaxivity. Some leakage occurred as the liposomes, on cooling from 48 ° to 37°C, displayed a higher rl than before heating. The results clearly demonstrate how the rl-temperature sensitivity of a gadolinium chelate can be altered by its encapsulation into liposomes. Whereas the rl of nonliposoreal Gd-DTPA-BMA continuously decreased as the temperature increased, the rl-temperature response of the liposomal chelate, linked to the thermotropic properties of the liposome membrane, had an "off-on" quality. Such a feature requires a membrane composition that displays a

PARAMAGNETIC

L I P O S O M E ~¢

markedly different permeability below and above T . At temperatures below T , the permeability should be low enough to reduce markedly the rate of water exchange between the interior and exterior of the liposome, while at temperatures above T , the permeability should be high enough to obtain fast exchange conditions. That the liposome size affected the rl only at temperatures below T may be of some practical benefit; the temperature profile of rl can be modulated by the liposome size. Larger liposomes will display a more marked increase in rl as the temperature increases from 37°C and exceeds T . For example, in the temperature range of 37°-44°C the rl of the 187-nm liposomes increased 17-fold, compared with a sevenfold change for the 126-nm liposomes. With the DW approach, continuous temperature monitoting was possible. As opposed to Tl-weighted GRE imaging, however, the DW technique did not readily indicate whether the temperature had reached a critical value. For instance, the lack of any SI enhancement within inserts 3 and 5 confirmed that the temperature never exceeded T (Fig 3c, 3d; see "Results" for final sample temperatures). The liposome-based Tl-weighted approach offers several advantages over the DW technique. The accuracy of the temperature monitoring by the DW technique depends strongly on the signal-to-noise ratio of the DW images, whereas liposomal relaxivity changes can easily be observed on Tl-weighted images, even with a relative poor signal-to-noise ratio. Therefore, temperature control with thermosensitive paramagnetic liposomes should be feasible for hyperthermia treatment of tissues with short T2 values, for which the DW approach would suffer from a strong SI loss during the diffusion-weighting period. Tl-weighted imaging is also inherently faster than DW imaging. Threedimensional Tl-weighted images can be acquired in seconds, making it possible to monitor whether the tissue temperature has exceeded a critical temperature. This could be beneficial for hyperthermic treatment with fast temperature changes, for example, thermal tumor ablation. Thermal ablation is gaining more and more popularity compared to conventional hyperthermia treatments. Such ablative therapy, performed with laser, RF, or focused ultrasound, is currently performed at temperatures above 55°C, at which protein denaturation occurs (2). The presently investigated DPPC/DPPG liposomes could not be used as thermal probes during a thermal ablation procedure be-

2Susceptibility effects were still visible on Tl-weighted GRE images despite a short TE of 4 msec.

1113

:OSSHEIM

E T AL

cause T was too low. A subtle change in the phospholipid composition, however, would yield liposomes with a suitable T within temperature intervals relevant for thermal ablation therapies. For instance, in the case of saturated phosphatidylcholines and phosphatidylglycerols, increasing their acyl chain length from 16 carbon atoms (eg, DPPC and DPPG) to 18 or more would give a Tmof 55°C or higher. To enable full utilization of thermosensitive liposomes in vivo, however, several important features should be considered. Obviously, the local concentration of liposomal Gd-DTPA-BMA needs to be determined to distinguish between areas that are not sufficiently heated (ie, T < T ) and those where the agent is not present. A potential problem in this respect is the low rl of liposomal Gd-DTPA-BMA below Tm, which makes it difficult to determine concentrations reliably. Still, even the relatively modest rl at 37°C might be sufficient to detect the presence of the liposomes (unpublished data, 1999). Furthermore, although the rl is low at 37°C, the susceptibility effect of the liposomes may be substantial at imaging fields, thus perhaps allowing a more accurate determination of local concentration in vivo. This liposome approach also would succeed only if the tumor region contains a high enough concentration of liposomal agent. Sterically stabilized liposomes should therefore be considered. Such liposomes, displaying surface properties that minimize uptake by the mononuclear phagocyte system, will have an increased probability of tumor extravasation due to the longer circulation time in blood. Indeed, the use of polyethylene glycol (PEG)-coated DPPC liposomes for selective delivery of cytostatics to tumors undergoing hyperthermia treatment has been extensively described (21,29). The results have shown the potential of such liposomes; liposome extravasation into the neoplastic area and a faster drug release from the liposome increased the therapeutic index of the encapsulated drug. PEGylated paramagnetic liposomes would be a candidate for in vivo MR imaging-guided temperature control. Some preliminary studies have shown that surface-grafted PEG does not alter the rl-temperature profile of Gd-DTPABMA encapsulated within DPPC/DPPG liposomes (unpublished data, 1998). These findings suggest a possible application of PEGylated liposomes for temperature monitoring in vivo. Studies must be conducted in tumor models, however, to investigate the targeting properties and temperature sensitivity of the current liposomes. For instance, it is essential to study the influence of liposome size on the MR imaging efficacy and biodistribution of the liposomes, al-

1114

Academic Radiology, Vol 7, N o 12, D e c e m b e r 200(

though it is known that the smaller the liposome, the more efficient is the tumor extravasation (30). Also, the thermotropic properties of the liposomes have to be assessed under physiologic conditions that mimic the tumor microenvironment. Finally, the encapsulation of paramagnetic macromolecular agents or use of other phospholipid combinations, with a lower but still high enough permeability above Tm(24), would improve the value of such thermosensitive agents. .3ONCLUSIOh This study has shown the feasibility of thermosensitive paramagnetic liposomes for temperature control in vitro. Large temperature-dependent changes in SI were observed on Tl-weighted GRE images, with high thermal sensitivity and no background signal. Such thermosensitive liposomes are efficient only if two main criteria are met: (a) the paramagnetic agent must be located within the liposome interior, and (b) the T1 relaxation process must be exchange limited only at temperatures below the T of the phospholipids. Liposomes are flexible and versatile delivery systems. With a combination of phospholipids for which the T lies within temperature ranges relevant for hyperthermic or ablative tumor treatment, local temperature changes in vivo may be monitored noninvasively with standard clinical MR scanners. ,CKNOWLEDGMENT.

The authors thank Gry Stensrud, PhD (School of Pharmacy, University of Oslo, Norway), for performing the laser Doppler velocimetry measurements and Kari Anne Bjerkeli and Vera Kasparkova, Phi) (both Nycomed Imaging AS, Oslo, Norway), for performing the inductively coupled plasma atomic emission spectrophotometry analysis and differential scanning calorimetry measurements, respectively. IEFERENCE~ 1. Antich PP, Mason RP, Nunnally RL. Applications of magnetic resonance techniques to deep tumor hyperthermia. Strahlenther Onkol 1989; 165: 734-737. 2. Le Bihan D. Temperature imaging by NMR. In: Le Bihan D, ed. Diffusion and perfusion magnetic resonance imaging: applications to functional MRI. New York, NY: Raven, 1995; 181-187. 3. De Poorter J, De Wagter C, De Deene Y, et al. Non invasive MRI thermometry with the proton resonance frequency (PRF) method: in vivo resuits in human muscle. Magn Reson Med 1995; 33:74-81. 4. Parker DL. Applications of NMR imaging in hyperthermia: an evaluation of the potential for localized tissue heating and noninvasive temperature monitoring. IEEE Trans Biomed Eng 1984; 31:161-167. 5. Engin K. Biological rationale and clinical experience with hyperthermia. Control Clin Trials 1996; 17:316-342.

6. Young IR, Hand JW, Oatridge A, et al. Further observations on the measurement of tissue T1 to monitor temperature in vivo by MRI. Magn Reson Med 1994; 31:342-345. 7. Le Bihan D, Delanney J, Levin RL. Temperature mapping with MR imaging of molecular diffusion: application to hyperthermia. Radiology 1989; 171:853-857. 8. De Poorter J, De Wagter C, De Deene Y, et al. The proton-resonancefrequency-shift method compared with molecular diffusion for quantitative measurements of two-dimensional time-dependent temperature distribution in a phantom. J Magn Resort 1994; B103:234-241. 9. Young IR, Hajnal JV, Oatridge A, et al. Demonstration of in vivo susceptibility variations developing phase changes that mimic those due to temperature measured by the chemical shift method. In: Proceedings of the fourth meeting of the International Society for Magnetic Resonance in Medicine. Berkeley, California: International Society for Magnetic Resonance in Medicine, 1996; 44. 10. Kuroda K, Mulkern RV, Oshio K, et al. Temperature mapping using the water proton chemical shift: self-referenced method with echo-planar spectroscopic imaging. Magn Reson Med 2000; 43:220-225. i1. Chenevert TL. Motion artifacts. In: Le Bihan D, ed. Diffusion and perfusion magnetic resonance imaging: applications to functional MRI. New York, NY: Raven, 1995; 103-110. 12. Vavreck RM, Mac Fall JR. Hardware considerations for diffusion/perfusion imaging. In: Le Bihan D, ed. Diffusion and perfusion magnetic resonance imaging: applications to functional MRI. New York, NY: Raven, 1995; 67-72. 13. Frenzel T, Roth K, KoBler S, et al. Noninvasive temperature measurement in vivo using a temperature-sensitive lanthanide complex and 1H magnetic resonance spectroscopy. Magn Reson Med 1996; 35:364-369. 14. Zuo CS, Bowers JL, Metz KR, et al. MR temperature measurement in vivo with TmDOTP 5-. In: Proceedings of the fifth meeting of the International Society for Magnetic Resonance in Medicine. Berkeley, California: International Society for Magnetic Resonance in Medicine, 1997; 1961. 15. Aime S, Botta M, Fasano M, et al. A new ytterbium chelate as contrast agent in chemical shift imaging and temperature sensitive probe for MR spectroscopy. Magn Reson Med 1996; 35:648-651. 16. Lain MM, Webb AG. Proton-based phase-transition compounds for MRI thermometry. In: Proceedings of the fourth meeting of the International Society for Magnetic Resonance in Medicine. Berkeley, California: International Society for Magnetic Resonance in Medicine, 1996; 1709.

17. Franklin KJ, Buist RJ, Den Hartog J, Mc Rae GA, Spencer DP. Encapsulated liquid crystals as probes for remote thermometry. Int J Hyperthermia 1992; 8:253-262. 18. Webb AG. Fluorocarbon liquid crystals for MRI thermometry. In: Proceedings of the fifth meeting of the International Society for Magnetic Resonance in Medicine. Berkeley, California: International Society for Magnetic Resonance in Medicine, 1997; 1582. 19. Webb AG, Wong M, Kolbeck KJ, et al. Sonochemically produced fluorocarbon microspheres: a new class of magnetic resonance imaging agent. J Magn Reson Imaging 1996; 6:675-683. 20. Merlin JL. Encapsulation of doxorubicin in thermosensitive small unilameliar vesicle liposomes. Eur J Cancer 1991;27:1026-1030. 21. Gaber MH, Wu NZ, Hong KH, et al. Thermosensitive liposomes: extravasation and release of contents in tumor microvascular networks. Int J Radiat Oncol Biol Phys 1996; 36:1177-1187. 22. Haines TH, Liebovitch LS. A molecular mechanism for the transport of water across phospholipid bilayers. In: Disa]vo EA, Simon SA, eds. Permeability and stability of lipid bilayers. Boca Raton, Fla: CRC Press, 1995; 123-136. 23. Mouritsen OG, Jergensen K, Hanger T. Permeability of lipid bilayers near the phase transition. In: Disalvo EA, Simon SA, eds. Permeability and stability of lipid bilayers. Boca Raton, Fla: CRC Press, 1995; 137-160. 24. Fossheim SL, Fahlvik AK, Klaveness J, et al. Paramagnetic Iiposomes as MRI contrast agents: influence of liposomal physicochemical properties on the in vitro relaxivity. Magn Resort Imaging 1999; 17:83-91. 25. Lasic DD. Preparation of liposomes. In: Lasic DD, ed. Liposomes from physics to applications. Amsterdam, the Netherlands: Elsevier Science, 1993; 63-107. 26. Ir yasov KA, Hennig J. Single-shot diffusion-weighted RARE sequence: application for temperature monitoring during hyperthermia session. J Magn Reson Imaging 1998; 8:1296-1305. 27. Neeman M, Freyer JP, Sillerud LO. Effects of imaging gradients on diffusion measurements by MR imaging. In: Le Bihan D, ed. Diffusion and peffusion magnetic resonance imaging: applications to functional MRI. New York, NY: Raven, 1995; 73-76. 28. Eckburg JJ, Cato JC, Liu KJ, et al. The measurement of temperature with electron paramagnetic resonance spectroscopy. J Biomech Eng 1996; 118:193-200. 29. Papahadjopoulos D, Allen TM, Gabizon A, et al. Sterically stabilized liposomes: improvements in pharmacokinetics and antitumor therapeutic efficacy. Proc Natl Acad Sci 1991 ; 88:11460-11464. 30. Yuan F. Transvascular drug delivery in solid tumors. Semin Radiat Oncol 1998; 8:164-175.

1115