New Page 1

In Vivo Bioluminescence Imaging

In vivo bioluminescence is an established imaging modality in laboratories across the country and has been gaining in popularity at the University of Arizona. This technology allows the non-invasive imaging and quantification of cells expressing luciferase proteins. The major luciferase used in these studies is from the firefly, phytonis pyralis. This enzyme has a short half-life in vitro (ca. 3 min at 37 degrees) and in vivo (ca. 90 min). The difference between in vivo and in vitro half lives is thought to be due to the action of chaperone proteins, in particular HSP90. Mutants are available with longer half lives. Following are descriptions of ongoing projects using bioluminescence technology.

Half-life engineering (R. Gillies, B. Baggett). Mutant p. pyralis luciferases are available with extended half lives. This project is testing the hypothesis that the extended half lives will result in higher intracellular accumulation of protein and, hence, greater light output on a per-cell basis. Increased light output will reduce the detection limit. More sensitive detection limits will allow experiments not otherwise possible. For example, it will allow detection of micrometastases, which are important pre-malignant lesions that will expand to larger, life threatening tumors if they are adequately vascularized. Table 1 shows the in vivo and in vitro half lives of wild type and mutant luciferases. The steady-state levels of light output increase with increasing half-life, which is consistent with our hypothesis.

	*in vitro*	*st dev*	*in vivo*	*st dev*
WT	3.06		61.34	31.14
A	8.5	1	134.95	43.41
B	15.52	1.45	138	31.63
C	7.36	1.82	129.63	57.37
D	72.4	3.1	143.28	13.33
E	82.06	8.5	249.44	101.75
F	75.1	9.81	148.28	22.07

Metastasis (N. Raghunand, B. Baggett, R. Gillies). Thermostable mutant luciferases will be useful in monitoring the movement of cells from primary tumors to colonization of distant sites (metastasis). A common metastasis model is the direct injection of 10⁶ cancer cells into the tail veins of mice, whereupon these cells colonize metasta6ic sites, especially the lungs, which contains the first microvessels these cells will encounter.. The image on the right shows image from SCID mice inoculated in the tail vein with only 10⁵ MDA-mc-231 cells expressing either (A) wild-type luciferase or (B) a thermostable luciferase mutant F. Note that the thermostable mutant yields more light, and that following inoculation, luciferase expressing cells are seen throughout the whole mouse. This was an unexpected result, as conventional wisdom holds that these inoculated cells will take up residence primarily in the lungs.

In vivo Pharmacodynamics of Geldanamycin (L. Whitesell, S. Momen). Geldanamycin is an inhibitor of HSP90 that is in clinical trials against a variety of cancers, including breast cancer. Since the half life of luciferase is longer in cells compared to in vitro, it has been reasoned that Geldanamycin should lead to a reduction in half life and hence, a reduction in the steady-state luciferase levels. This appears to be the case as geldanamycin causes a dose-dependent decrease in light output from MDA-mb-231 breast cancer cells. The time-dependency of this effect is interestingly complex, as cells appear to respond with a reduction of luciferase, and this is followed by a subsequent rise. This is interpreted to indicate that cells respond to reduction in HSP90 by up-regulating further HSP90 expression, which overcomes the drug block. Based on the success of these experiments, wt luciferase reporters are being transfected into a number of recipient cell lines which will be grown as tumors in mice in order to assess geldanamycin pharmacodynamics in vivo.

Tumor Growth in bones (A. Liss). Bone marrow is a common site for metastases in breast cancer, prostate cancer and lymphomas. Cutting edge therapies against these sites include anti-angiogenic drugs, as the development of these metastatic lesions apparently recruits new vessels to this well vascularized bed. A problem in assessing drug response in mice is that these lesions are extremely difficult to assess, even by destructive means. Post-mortem examination of mice is prone to sampling artifacts, as the extraction of tumor cells from marrow is not stoichiometric. Bioluminescence is proposed as a solution to this problem. Light should penetrate bone, which shows light emanating from an intracranial glioma in a rat. Rat skulls are approx. the same thickness as the long bones in mice. For these experiments, we will use cells transfected with thermostable luciferase, to generate as much light as possible.

Intracranial Tumors (G. Powis). The same argument made above for bony metastases can also be made for intracranial tumors, which are also occult. In this case, however, MRI could be a competing technology. Figure 3 shows that during the normal growth of gliomas, the CCD and MRI results were highly correlated. However, upon treatment, there was a disconnect, as the light output significantly declines with little decrease in the MRI measured tumor volume. This is interpreted to indicate that the luciferase is reporting on viable cells, because it takes ATP to generate light, whereas the MRI results may also include fibrotic or necrotic volumes. Furthermore, bioluminescence is adaptable to high throughput imaging, as 6 mice can be imaged simultaneously in about 15 min, including setup (2.5 min/mouse). In contrast, MR imaging requires a minimum of 30 min/mouse.