Applied Radiology

Dr. Johnston is a fourth-year Resident in the Department of Radiology, University of California, Los Angeles. He received his PhD from Mayo Graduate School in 1997 and his MD from Mayo Medical School in 1998. He completed a Postdoctoral Fellowship in Surgery in 1999. He has accepted a Fellowship in Vascular and Interventional Radiology at UCLA, starting July 2005. Mr. Lee is a Graduate Student in the laboratory of Dr. Hawthorne, and plans to attend medical school after receiving his PhD. Dr. Hawthorne is a University Professor in the Department of Chemistry, University of California, Los Angeles.

Although nonspecific contrast agents have appreciably improved computed tomography (CT) and magnetic resonance imaging (MRI), specific tissue enhancement represents the next logical progression of improved patient care. Specific targeting of cell-surface proteins has been exploited for the delivery of chemotherapeutic agents and for many types of nuclear medicine studies, but cellular targeting of contrast agents has met with only limited success. This article reports the development of CT and MRI contrast agents specific for cell-surface proteins, accomplished through adaptation of an existing immunoliposome system used for the delivery of chemotherapy to breast tumors. We chose to develop contrast agents specific toward three cell-surface proteins, namely Fas ligand, Fc receptor, and CD5. Fas ligand, which is expressed on immune-privileged cells, binds to the Fas receptor present on T cells and induces apoptosis. The Fc receptor is specific for the terminal constant region of the immunoglobulin molecule, which is utilized by the reticuloendothelial system to bind immunoglobulins. Finally, CD5 is a signal transduction molecule that modulates signaling through the T cell or B cell receptors. We demonstrate specific in vitro enhancement of testicular tissue through Fas ligand targeting, and enhancement of the reticulo-endothelial system (RES) through Fc receptor and CD5 targeting.

Currently, contrast-enhanced imaging by computed tomography (CT) and magnetic resonance imaging (MRI) is limited to nonspecific delivery of macromolecules containing specific properties, which cause either increased attenuation of X-rays by iodinated materials or T1 signal shortening by the paramagnetic effect on hydrogen atoms. ¹ Although this technique has obvious diagnostic advantages and is commonplace, specific tissue enhancement has the potential to dramatically improve sensitivity and specificity. The application of tissue-specific imaging in nuclear medicine has received considerable research attention. Progress has been substantial and, recently, diagnostic tests for colorectal and prostate cancer have entered clinical practice. ^2-4 Tissue-specific imaging by other modalities has met with only limited success, ⁵ however, and is confined mainly to the examination of gene expression and protein synthesis. ^6-11

We approached the problem of extending tissue-specific imaging to CT and MRI by modifying and expanding a system widely used in drug-delivery research, namely immunoliposomes. ^12-14 Liposomes are multipurpose biological and immunological agents whose use in drug delivery, nuclear imaging, and radiotherapy has been studied extensively over the last quarter-century.

More recently, liposomes have been used in the development of ultrasound and MRI contrast agents. The targeting of liposomes to specific cell types through conjugation to antibodies has made this system ideal for research into cellular-specific contrast agents. ^13,14-16

Background

Liposomes

Initial research into the use of liposomes for drug delivery was hindered by several factors. First, their design had to be optimized so that liposomes could carry the desired chemical "payload." Otherwise, electrostatic forces of the passenger chemical or charged surface lipid membranes could cause rapid instability of the liposome. ¹⁷ Small, neutral liposomes with bilayers composed of "rigid" phospholipids of high-phase-transition temperature were required for prolonged in vitro stability. ¹⁷ Subsequent in vivo experimentation demonstrated rapid clearance of liposomes by the reticuloendothelial system (RES) in a dose-dependent fashion. ¹⁸ This problem was overcome by the development of sterically stabilized liposomes (SSL), which were covered by a polymeric coating of polyethylene glycol (PEG). These "stealth" liposomes avoided clearance, presumably by steric inhibition of liposomal surface recognition. ^19,20 Using SSLs, Huang et al ²¹ demonstrated preferential extravasation of chemical payloads in solid tumors, presumably because of vascular abnormalities associated with tumor angiogenesis. However, Horowitz et al ²² also showed that SSLs do not interact directly with tumor cells in vitro or in vivo, instead accomplishing targeted delivery by eventual degradation and local payload diffusion into target cells.

Radionuclides

Conventional nuclear imaging agents, such as gallium 67 and indium 111, have been coupled to antibodies, chemotactic peptides, and cytokines in an attempt to enhance imaging specificity. ²³ Liposomes have also been evaluated for directed delivery of radionuclides for either therapy or imaging, in both the central nervous system and the rest of the body. ^23-25 One form of radionuclide therapy that has potential efficacy against malignant gliomas is boron neutron capture therapy (BNCT). This two-step radiotherapy involves first targeting the tumor with nonradioactive ¹⁰ B, then exposing the boron to low-energy neutrons. The reaction yields intensively ionizing particles that travel a maximum of approximately 1 cell width, thereby minimizing collateral tissue loss. ²⁶ Delivery of large quantities of boron atoms through liposome encapsulation is an active area of research. ²⁷

Ultrasound

Ultrasound contrast agents consisting of coated gas microbubbles are used in conjunction with ultrasound imaging for diagnostic studies. Two commercially available systems use human serum albumin-encapsulated microbubbles, while a third uses a galactose-based microbubble stabilized with fatty acids. ²⁸ Air-filled monolayers of phospholipids have also been used experimentally and demonstrate increased in vivo stability in comparison with commercially available products. ²⁸

MRI

Initial research into the use of liposomes in MRI incorporated paramagnetic molecules into the liposome membrane in an attempt to increase the blood-pool circulation time of contrast agents. In the formation of the liposome membrane, paramagnetic polymerized liposomes (PPLs) incorporated a derivative of gadopentetate dimeglumine as the hydrophilic head group and diacetylene groups in the hydrophobic acyl chains. ²⁹ Following in vivo administration, 80% of the PPLs remained in the blood pool after 2 hours; the half-life of elimination from the blood pool was 19 hours. ²⁹ Nonspecific enhancement of the kidneys and liver was also observed. ²⁹

Additional research in this field was accomplished through the use of
mannan-coated liposomes containing gadolinium diethylenetriamine penta-acetic acid (Gd-DTPA), ³⁰ which demonstrated nonspecific enhancement of the liver and lung at lower doses than intravenously injected gadolinium. Recently, vascular endothelial-cell­specific enhancement was accomplished. Vascular structures were targeted by antibody LM609 (Chemicon International, Inc., Temecula, CA), specific to endothelial cell receptors, which were conjugated to multivalent polymerized vesicles (PVs) chelated to gadolinium to allow imaging through MR. This protocol demonstrated tumoral enhancement, a result of increased tumor vasculature. ³¹

Antibodies as targeting agents

Tremendous research effort has been focused on tumor-specific antibodies as a potential therapy for various forms of cancer. A prime example is the development of monoclonal antibodies (mAb) directed against the p185HER2 receptor tyrosine kinase. ³³ Human epidermal growth factor receptor 2 (HER2) is a protein found on the surface of cells that, when functioning normally, has been found to be a key component in regulating cell growth. ³⁴ This protein is highly overexpressed in a significant portion of breast and other cancers. ³⁵ The anti-HER2 mAb trastuzumab induces an antitumor response as a single agent but is most efficacious when coupled with chemotherapeutic agents. ³⁶ Recently, immunoliposomes constructed using the anti-HER2 mAb have demonstrated improved antitumor effect over chemotherapy or monoclonal antibody therapy alone, further highlighting the importance of immunoliposomes for cell-specific targeting and drug delivery. ^12,16,33

In vivo refinements

Pioneering studies in tumor targeting by chemotherapeutic liposomes relied on the leaky capillaries of tumors for increased tissue penetration. ^15,37 The initial use of monoclonal antibodies as targeting agents revealed increased uptake by the RES and subsequent decreased circulatory time. ³⁸ In addition, comparatively poor tissue penetrance of the immunoliposomes was observed. ³⁸ The immunoglobulin Fc portion, which is the crystallizable fragment of the immunoglobulin molecule formed after specific proteolysis, may have directly caused these effects, as Fc is efficiently recognized by the RES for normal opsonized pathogen removal. ^38,39 Experiments with immunoliposomes that were created through use of the immunoglobulin Fab' fragment demonstrated increased tissue penetrance and prolonged circulatory time. ⁴⁰

Model system

In our model system, we adapted for CT and MRI refinements originally made for immunoliposome-based drug delivery. We accomplished this by using Omnipaque (Amersham Health, Princeton, NJ), an iodinated macromolecule widely used for CT contrast enhancement, and Omniscan (Amersham Health), a gadolinium-containing macromolecule widely used for MRI contrast enhancement, as the chemical payloads, instead of chemotherapeutic agents.

Our initial experiments sought to determine the theoretical detection limit of our system with both CT and MR. We attempted in vitro solid-organ tissue targeting using liposomes conjugated to one of three antibodies. The first antibody was specific to whole-rat anti-mouse Fc receptor (FcR). The Fc receptor is a cell-surface protein utilized by the RES to bind immunoglobulins, as they are specific for the terminal constant region of the immunoglobulin molecule. ⁴² These liposomes were targeted toward the RES and lymphoid structures, specifically liver and spleen. We also used lipopsomes conjugated to anti-CD5 Fab' (a smaller immunoglobulin fragment generated through proteolysis), which recognizes CD5, a signal transduction molecule that modulates signaling through the T cell or B cell receptors, and is present on the cell surface of T-helper and B-cell subpopulations. ⁴³ This construct would also target the RES and lymphoid structures. Finally, we utilized Fas ligand (FasL), expressed on the surface of immune-privileged cells, which binds to the Fas receptor present on T cells to induce apoptosis. ⁴⁴ The testicles, which highly express Fas ligand in order to confer immune-privileged status, ⁴⁵ were specifically targeted through liposomes conjugated to anti-FasL. Untargeted contrast-laden and targeted immunoliposomes were incubated with a variety of freshly obtained murine tissues, and the relative tissue enhancement was calculated.

Materials and methods

Antibodies

Rat anti-mouse FcR IgG antibodies were kindly provided by Sherie Morrison, MD, Professor of Immunology, University of California, Los Angeles. Humanized Fab' fragments of mouse anti-CD5 were kindly provided by XOMA (Berkeley, CA). Commercially available anti-FasL IgG antibodies were purchased through Research Diagnostics, Inc. (Flanders, NJ).

Preparation of liposomes and immunoliposomes

The lipids 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DSPC), and 1,2-dimyristoyl-sn-glycero-3-[phospho-rac-(1-glycerol)] (DSPG), as well as cholesterol were obtained from Avanti Polar Lipids (Alabaster, AL). The PD-10 and Sephadex G-50 (Amersham Biosciences, Piscataway, NJ) columns were purchased from Pharmacia Corporation (Peapack, NJ). All other reagents were American Chemical Society (ACS) reagent grade.

Liposome preparation ­­ A 1000-mg mixture of DSPC, cholesterol, and DSPG was prepared in a molar ratio of 2:1:0.1. The mixture was dissolved in dichloromethane, placed in a rotary evaporator to remove the solvent, and dried under vacuum overnight. Each batch of liposomes was formed using 200 mg of this mixture.

To form liposomes with osmolarities below 800 mOsM, and because the Omnipaque solution is quite viscous, the solutions were diluted using phosphate-buffered saline (PBS), pH 7.4. Two mL of Omnipaque was diluted 1:1, and 3.0 mL of Omniscan was diluted 3:1. These specific dilutions were initially chosen in part for future in vivo murine model studies, in which a clinical practice-equivalent amount of contrast would be delivered.

For each batch of liposomes, 200 mg of the lipid mixture was placed in a 16- * 100-mm culture tube under a nitrogen stream. Next, 3.0 mL of the contrast agent/PBS solution was placed into the tube. The tube was mounted into a VibraCell sonicator equipped with a 1Ž8-inch probe and a water bath set at 65° C. The sample was sonicated continuously for 1 minute at a setting of 4.5. After 1 minute, the remaining 1.0 mL of contrast agent/PBS solution was added, and the sample sonicated for an additional 10 minutes. During sonication, the liposome suspension became milky-white and translucent.

Extrusion -- Using an Avestin Liposofast extrusion device (Avestin, Ottawa, CA), 2.0 mL of each liposome suspension was extruded through a 0.1 µm polycarbonate membrane. The material was forced through the membrane 11 times. The extruded liposomes were cleaned using PD-10 columns equilibrated with PBS pH 7.4 and diluted to 10 mg/mL total lipid. The remaining liposome suspensions were stored at 4° C.

Glycerol oxidation -- A 0.6 M solution of sodium periodate was prepared by dissolving 512 mg in 4.0 mL of water. Next, 2.0 mL of this solution was added to each extruded liposome suspension. The reactions were carried out in amber vials at room temperature for 30 minutes. After the glycerol was oxidized to the reactive aldehyde, the suspensions were cleaned using 15- * 100-mm Sephadex G-50 columns equilibrated with 20 mM sodium borate buffer and 150 mM sodium chloride at pH 8.4.

Antibodies and immunoliposomes -- Each of the 6 conjugation reactions were carried out in 1.5-mL plastic vials. The antibodies were added to the oxidized liposomes and allowed to form Schiff-base linkages for 2 hours at room temperature. These were then reduced to secondary amines by reacting overnight with sodium cyanoborohydride at 4° C.

A 2-M solution of sodium cyanoborohydride was prepared with caution by dissolving 125 mg in 1.0 mL of water. The antibodies were dissolved in sodium borate buffer pH 8.4 as follows: anti-FcR, 100 µg in 15.6 µL; anti-CD5 Fab', 385 µg in 80 µL; and anti FasL 100 µg in 100 µL.

Anti-FcR immunoliposomes were prepared as follows: 150 µL of oxidized Omnipaque liposomes were combined with 7.8 µL of anti-FcR and 15.8 µL cyanoborohydride; 150 µL of oxidized Omniscan liposomes were combined with 7.8 µL of anti-FcR and 15.8 µL cyanoborohydride. Anti-CD5 Fab' immunoliposomes were prepared as follows: 400 µL of oxidized Omnipaque liposomes were combined with 40 µL anti-CD5 and 40.0 µL cyanoborohydride; 400 µL of oxidized Omniscan liposomes were combined with 40 µL anti-CD5 and 40.0 µL cyanoborohydride. Anti FasL immunoliposomes were prepared as follows: 5.0 mL of oxidized Omnipaque liposomes were combined with 100 µL anti-FasL and 50.0 µL cyanoborohydride; 5.0 mL of oxidized Omniscan liposomes were combined with 100 µL anti-FasL and 50.0 µL cyanoborohydride.

Following reduction, the immunoliposomes were cleaned using 15- * 100-mm Sephadex G-50 columns equilibrated with PBS pH 7.4. The remaining oxidized liposomes were reduced under the same conditions in order to serve as a control for the tissue experiments. The blank, anti-FcR and anti-CD5 immunoliposomes were placed in vials at a concentration of approximately 3 mg/mL total lipid. The anti-FasL immunoliposomes were prepared at a concentration of approximately 9 mg/mL total lipid. Each of the liposome suspensions was measured for particle size using a Microtrac ultrafine particle analyzer (UPA) (Leeds-Northrup, Phoenix, AZ) validated against National Institute of Standards and Technology (NIST) standards measuring 87 nm ± 10 nm. In each case, the liposomes measured approximately 100 nm in diameter.

Murine tissue preparation and in vitro tissue enhancement

As protocols for in vivo studies including intravenous injection were pending University of California Los Angeles Animal Research Committee approval, an in vitro model system was initially developed. Adult male C57BL/6 breeder mice euthanized for advanced age were obtained immediately posteuthanization, having been graciously donated by Genhong Cheng, MD, Associate Professor of Immunology and Molecular Biology, University of California, Los Angeles. All animals were maintained and euthanized following protocols approved by the University of California Los Angeles Animal Research Committee. Liver, spleen, peritoneal fat, quadriceps muscle, and testes were harvested. The tissues were gently teased apart in PBS buffer, in order to more approximate the cell surface area accessible by an in vivo model system, and placed in standard 24-well tissue-culture plates for incubation. Next, 100 µM targeted and nontargeted immunoliposomes were added to the tissue medium, and the samples were incubated for 1 hour at 37° C, followed by an overnight incubation at 4° C.

The tissue samples were washed 3 times with PBS buffer and then transferred to standard 96-well tissue-culture plates in order to concentrate tissue density with corresponding increase in signal-to-noise ratios. These samples were then scanned on modality-appropriate equipment.

CT and MRI experimental sample imaging

Following purification of contrast-laden liposomes, serial dilutions were performed on these unconjugated immunoliposomes to determine the lowest amount of contrast detectible in our model system. Initial wells contained 100 µL of a 3 mg/mL suspension of unconjugated immunoliposomes.

Omnipaque samples were imaged on a GE HiSpeed CT scanner (GE Medical Systems, Waukesha, WI) using a 1-mm contiguous axial volume-data acquisition, display FOV = 10 cm. The 96-well plates containing dilutional or tissue samples were secured to the CT gantry on sufficient padding to place the sample plate in the center of the imaging bore.

Omniscan samples were imaged in the coronal plane on a Siemens 1.5T MRI scanner (Siemens Medical Systems, Iselin, NJ) with the following parameters: TR = 15, TE = 6, FOV = 300 * 300, matrix = 128 * 256, slice thickness 1 mm. Samples containing labeled testes tissues were imaged with an FOV of 400 * 400; imaging was not repeated with a smaller FOV due to limited research magnet time. Samples were secured on top of a saline bag to improve signal. ⁴⁶ The sample-saline bag construct was placed in the center of an extremity coil for further improved signal-to-noise ratio.

Results

Imaging was performed on serial dilutions of unconjugated immunoliposomes to determine the lowest amount of contrast detectible in our model system (Figure 1). MR imaging demonstrated progressive T1-weighted signal loss until a background level was reached at 1:32 dilution (Figure 1A), with an air bubble partially obscuring the signal in dilution well #4. The 1:32 dilution corresponds to 9.4 µg unconjugated immunoliposome as our practical limit of detection by MR.

Relative enhancement values were also calculated (Figure 1C). From the digital image, the gray-scale pixel value of T1-weighted enhancement was obtained utilizing ImageJ 1.29x, National Institutes of Health public-domain software package. The absolute values of the serial dilutions were plotted. A relative value of 1 was arbitrarily assigned to the gray scale value of dilution well 1:256, which demonstrated no increased signal in comparison with downstream dilutional well samples. As the plot follows a near linear progression, it is reasonable to conclude that at this concentration range, increasing numbers of immunoliposomes will produce proportional T1 signal increases.

CT examination of unconjugated Omnipaque immunoliposomes did not demonstrate visible differences between well signal density. In order to determine if this modality could be used to image immunoliposomes at this concentration of payload contrast, the HU of the various dilutional samples were measured. A maximal HU value of 60 was seen in the first well, with progressive dilutions resulting in background equilibration at a dilution of 1:8 (Figure 1B). This corresponds to 37.5 µg as the unconjugated immunoliposome practical limit of detection by CT. Under our model conditions, enhancement detection was approximately 4 times more sensitive by MRI than by CT.

As we determined that it was possible to detect relatively small amounts of immunoliposomes by both MR and CT imaging, tissue enhancement patterns by cell-surface protein targeted immunoliposomes were then explored. Freshly obtained C56BL/6 murine liver, spleen, peritoneal fat, quadriceps muscle, and testes were divided into 5 equal parts, incubated with uncongujated immunoliposomes, anti-CD5 immunoliposomes, anti-FcR immunoliposomes, anti-FasL immunoliposomes, and PBS, and processed according to protocol. The tissues were then imaged by either CT or MRI.

MR T1-weighted imaging of murine tissue samples is shown in Figure 2, which is a representative coronal slice. PBS-incubated and unconjugated immunoliposome incubated tissues served as negative and nonspecific controls, respectively. Relative enhancement of tissues was calculated based on pixel values as calculated in the Figure 1 studies, with PBS-incubated wells as the reference value. Table 1 shows the relative enhancement values compared with the native T1-weighted tissue signal. As testes and splenic tissue comprised less mass than liver or muscle, pixel values used in relative enhancement calculations were obtained from coronal slices that contained the maximal amount of these tissues.

In comparison with the negative control, anti-CD5 immunoliposomes demonstrated liver enhancement, splenic enhancement, and muscle enhancement on MRI, but did not result in fat enhancement (Figure 2, Table 1). Anti-FcR immunoliposomes also demonstrated selective 2-fold liver enhancement and slight splenic enhancement, with less pronounced relative fat enhancement (1.7-fold increase) (Figure 2). Anti-FasL immunoliposomes demonstrated pronounced exclusive testes signal enhancement (41-fold increase), with no appreciable increase compared with background in the remaining tissues (Figure 2, Table 1).

Nonspecific binding of unconjugated immunoliposomes did not result in any significant enhancement, as there was no significant increase in signal between the unconjugated immunoliposome samples and tissues incubated with PBS alone (Table 1).

Similar results were seen with CT imaging of cell-surface protein targeted Omnipaque immunoliposomes incubated with murine tissues (Table 2). As was demonstrated in the dilutional studies, no visible difference in well density was obtained, so HU values were again calculated for each sample well. No significant increase in HU values was seen in peritoneal fat in any conjugated or control well. As was seen in the MR studies, anti-CD5 immunoliposomes again enhanced liver (+27 HU) and splenic tissues (+48 HU), and were also associated with an increase in muscle attenuation (+27 HU). Anti-FcR immunoliposomes enhanced both liver and spleen samples (+26 HU and +67 HU, respectively), while the testes were enhanced only by anti-FasL conjugated immunoliposomes (+42 HU).

Again, nonspecific binding of unconjugated immunoliposomes did not result in any significant enhancement, as there was no significant increase in HUs between the unconjugated immunoliposome samples and tissues incubated with PBS alone (Table 2).

Discussion

We have demonstrated tissue-specific enhancement by contrast-laden immunoliposomes that are conjugated to monoclonal antibodies directed against a variety of cell-surface proteins. The highest tissue-specific enhancement was seen with anti-FasL immunoliposomes, which is consistent with the tissue distribution of this cell-surface protein and the known high surface concentration of Fas Ligand on testes. Anti-FcR immunoliposomes also successfully enhanced tissues of the RES known to express the surface Fc receptor. Anti-CD5 immunoliposomes enhanced tissues that would be expected to house B- and T-cell subtypes, but also enhanced muscle tissue. One possible explanation is that inguinal lymph nodes were included in the sample, but cross-reactivity with this murine anti-human immunoglobulin against other murine tissues is not excluded.

Our pilot in vitro studies also demonstrated the need to improve the discriminatory capabilities of our CT immunoliposomes, as a maximum enhancement of 67 HU was seen in this study. Conventional contrast agents use macromolecules not only to improve the solubility of constituent elements but also to provide a method to control for osmolarity. As liposomes would naturally buffer iodine, experiments in which more elemental forms of iodine are encapsulated in the immunoliposome will be performed prior to in vivo testing. As we are interested in increasing the ability of MRI to discriminate target tissues, experiments using superparamagnetic compounds in place of chelated gadolinium in the immunoliposomes will also be performed.

Advances in both CT and MRI technology have been realized through improvements in hardware, software, and contrast materials. The use of nonspecific contrast and its differential uptake by various tissues has widely expanded the diagnostic capabilities of CT and MRI. Under certain circumstances, the specificity of these imaging modalities now approaches that of histology. Cellular targeting of contrast agents represents the next logical progression. By merging the fields of molecular biology and radiology, we have demonstrated one avenue toward this goal.

The synthesis of contrast-laden immunoliposomes, conjugated to any of a wide arsenal of antibodies, can be accomplished in relatively short order. As monoclonal antibody generation is an established science, novel tissue-specific antibodies can be developed to further increase the utility of this technique. While multiple uses of this technique can be devised, we believe certain applications of MRI and CT will initially be explored.

MRI applications

With MRI, the utility of contrast immunoliposomes will likely be greatest in the characterization of the central nervous system. Further research is needed to enable diagnostic reagents to successfully cross the blood-brain barrier, however. Recently, immunoliposomes conjugated to the transferrin receptor have demonstrated successful translocation and delivery of brain-specific promoter-region DNA to the parenchyma. ⁴⁷ One possible experimental approach for enhanced delivery of contrast would be double conjugation of the immunoliposomes with transferrin and antibodies specific for tissues of the central nervous system. Using this method, it may be possible to further explore demyelinating diseases, as selective myelin enhancement may lead to improved sensitivity for the detection of demyelinated plaques.

In both the central nervous system and in the remainder of the body, contrast immunoliposomes may improve tumor characterization, as tumor-specific agents could differentiate cancer from a nonspecifically enhancing mass. Also, radiation necrosis could reliably be differentiated from recurrent tumor through selective enhancement. In the same way, the detection of small foci of metastastic cells could also be improved through tissue-specific imaging.

CT applications

With CT, differential cellular enhancement could improve diagnosis in many clinical conditions. Lung nodules that are too small to be detected by PET, for example, must now undergo a complicated clinical and imaging protocol. Evaluation with a panel of cell-specific contrast agents could instead allay patient concerns, if granulomatous disease were confirmed, or increase patient survival, if resectable lung cancer were detected.

In vivo histological characterization of masses could also be accomplished. For example, specific tissue enhancement could determine whether an adrenal mass originated from medullary or cortical tissue, or a pelvic mass from germ or stromal cells. Therapeutic response could be determined more accurately, as the progress of tumoral enhancement in primary or metastatic cell foci could be more accurately assessed. Micrometastatic disease might be detected more readily, especially in lymph nodes containing tumor foci that would otherwise be discounted due to CT size criteria for malignant lymph nodes. Also, the extent of lymph node involvement in breast and other cancers could be determined without surgical dissection, thereby reducing morbidity and mortality.

Conclusion

We have shown that immunoliposomes targeted to cell-surface proteins can successfully be used for tissue-specific contrast enhancement in CT and MRI. This approach has the potential to increase diagnostic specificity and sensitivity, and has broad clinical applications. The study reported here will serve as the foundation for our future research, which will first focus on improving signal enhancement in both CT and MRI and, ultimately, on developing tumor-specific immunoliposomes for in vivo detection of both primary and metastatic lesions. Further research is needed into techniques that enable crossing the blood-brain barrier for central nervous system imaging before the benefits of tissue-specific contrast are fully realized.

Acknowledgements

This work was supported in part by a grant from the Department of Energy DE-FG03-95ER61975. The authors thank Brian Zarnagar for his work in animal husbandry.