2.1.

Cells Expressing Biosensor Proteins

Crandell feline kidney (CrFK) cells²⁶ were transfected with plasmid encoding FLuc (pSV2A/L- $\mathrm{A}\Delta 5^{\prime }$ ),²⁷ and neo and stable transfectants were selected with G418. These cells were then transfected with plasmids ${\mathrm{pGFP}}^2\text{-}{\beta }^{\prime }$ -arrestin2 and pV2-vasopressin-receptor-hRLuc (Packard Bioscience, Meriden, Connecticut; hRLuc is a humanized form of RLuc), or a modified version of this plasmid encoding hRLuc mutated at 8 codons, RLuc8.²⁸ This pair of fusion proteins, V2-vasopressin-receptor-hRLuc/RLuc8 and ${\mathrm{GFP}}^2\text{-}\beta$ -arrestin2, constitute a biosensor for the hormone vasopressin.¹² ${\mathrm{GFP}}^2$ is a modified form of GFP.²⁹ Firefly luciferase-expressing cells were transfected with ${\mathrm{pGFP}}^2\text{-}\beta$ -arrestin2 and selected with hygromycin. Stably transfected clones were screened for fluorescence. Selected clones were then transfected with pV2- $\text{vasopressin}$ -receptor-hRLuc/RLuc8 and selected with zeocin. Stably transfected clones were screened for luciferase activity. Clones with high luciferase activity were further screened for responsiveness to vasopressin by changes in BRET as described.¹²

2.2.

Construction of Plastic Windows

Windows were formed from two pieces of polystyrene cut from $10\mathrm{mm}$ o.d. ( $8\mathrm{mm}$ i.d.) tubes (Evergreen Scientific, Los Angeles, California) each $3\mathrm{mm}$ in height. An annulus of woven polyester (polyethylene terephthalate) material, internal diameter $8\mathrm{mm}$ , external diameter $14\mathrm{mm}$ , was sealed between the two pieces using cyanoacrylate glue (see Fig. 1 ). Circles of transparent polyester film were sealed over the open ends of the tube. Windows were sterilized by exposure to ultraviolet light before use. Light transmission by the assembled windows was measured and was $>80\%$ over the range $350\text{to}1200\mathrm{nm}$ .

Fig. 1

Diagrammatic representation of a mouse with a skin/body wall window and transplanted bioluminescent cells. The gray shading represents the bioluminescent cells transplanted beneath the capsule of the kidney. A plastic window is fitted in the skin and body wall adjacent to the kidney. A woven polyester collar attached to the window enables the window to be tightly integrated into the tissues of the mouse without loss of integrity of the skin around the edge of the window. When the mouse is anesthetized, the end of a fiber optic light guide may be inserted into the opening of the window. Following injection of luciferase substrates into the peritoneal cavity of the animal, light emitted from the bioluminescent cells is detected by a photon counter. A filter wheel interposed in the light path enables the measurement of light at defined wavelengths. Data from the photon counter are recorded on a computer, which also controls the filter wheel.

2.3.

Photon Counter and Fiber Optics

A Hamamatsu H7467 photon counter (Hamamatsu Photonics, Bridgewater, New Jersey) was used in conjunction with a motorized filter wheel (Oriel model 74041, Newport Corporation, Irvine, California). Both photon counter and filter wheel have RS-232 interfaces, allowing computerized measurement of light and control of the position of the wheel. Two of six positions of the filter wheel were occupied by optical filters, either the combination of a $475∕50\text{-}\mathrm{nm}$ bandpass filter ( $475∕50$ BP; Semrock, Rochester, New York) and a $580\text{-}\mathrm{nm}$ “3RD Millenium” longpass filter (580 LP; Omega Optical, Brattleboro, Vermont); or a $417∕60\text{-}\mathrm{nm}$ bandpass filter ( $417∕60$ BP; Semrock) and a 500 longpass filter (500 LP; Omega Optical). The filter wheel is connected to a fiber optic light guide ( $5\times 500\mathrm{mm}$ Single Flexible Light Guide, Volpi USA, Auburn, New York). The dimensions of the plastic window were designed to allow insertion of one end of the light guide into the window.

2.4.

Surgical Implantation of Biosensor Cells in the Kidney of Immunodeficient Mice and Insertion of Skin/Body Wall Window

$\mathrm{RAG}{2}^{\text{-}∕\text{-}},\gamma {\mathrm{c}}^{\text{-}∕\text{-}}$ mice were purchased from Taconic (Germantown, New York). Animals (both males and females) at an age greater than $6\text{weeks}$ ( $\sim 25\text{-}\mathrm{g}$ body weight) were used in these experiments. Procedures were approved by the institutional animal care committee and were carried out in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. Animals were anesthetized with Avertin.³⁰ An oblique incision $(<1\mathrm{cm})$ was made in the skin parallel to and adjacent to the long axis of the left kidney. The skin was separated from the body wall to the extent of $1.5\mathrm{cm}$ longitudinally and $1.0\mathrm{cm}$ laterally. An incision $(<1\mathrm{cm})$ was then made in the body wall adjacent to the kidney and the kidney was exteriorized. Biosensor cells $(2\times {10}^6)$ were trypsinized and resuspended in cold culture medium for implantation. The total volume (cells and medium) was $10\text{to}15\mu \mathrm{l}$ . Cells were injected using a blunt $25\text{gauge}$ needle and glass Hamilton syringe using a transrenal injection, placing the cells beneath the capsule on the side of the kidney adjacent to the body wall. The kidney was returned to the retroperitoneal space. A plastic window was inserted into the incision in the body wall, the base of the window being adjacent to the transplanted cells. Using a continuous suture technique (6-O silk suture), the polyester collar was attached to the outer surface of the body wall. The suture was tightened before being tied off. The skin was sealed around the opening of the window and was placed in close contact with the polyester collar by stretching it around the window and closing the incision with a surgical staple. Postoperative care of animals was as previously described.¹⁵

2.5.

Bioluminescence Measurements in Mice with Transplanted Cells

The growth of the transplanted cells was monitored immediately after surgery and at intervals thereafter using epifluorescence ( $400\text{-}\mathrm{nm}$ wavelength illumination, Lightools Research, Encinitas, California). After $15\text{to}25\text{days}$ , animals were used in experiments of bioluminescence measurements. Mice were anesthetized using Avertin. Bioluminescence substrates were administered by intraperitoneal injection. D-Luciferin and native coelenterazine were obtained from Biotium, Incorporated, Hayward, California. Coelenterazine 400a (also known as DeepBlue C, a trademark of PerkinElmer) was obtained from Biosynth AG, Staad, Switzerland. Luciferin was dissolved at $30\mathrm{mg}∕\mathrm{ml}$ in water for injection. Coelenterazines were dissolved at $1\mathrm{mM}$ in acidic methanol and were diluted if necessary in saline for injection. A complex of native coelenterazine and 2-hydroxypropyl- $\beta$ -cyclodextrin (Sigma-Aldrich, Saint Louis, Missouri) was prepared by lyophilization of a mixture of a methanolic solution of coelenterazine and an aqueous solution of cyclodextrin (1:50 molar ratio).³¹ For use, the lyophilized complex was resuspended in water at a concentration of $1\mathrm{mM}$ .

Following injection of luciferase substrate, the anesthetized animal was placed in a light-tight box housing the photon counter and filter wheel. One end of the light guide was placed within the opening of the skin/body wall window (Fig. 1). Measurements were made at $10\text{-}\mathrm{s}$ intervals, each measurement typically of $5\text{-}\mathrm{s}$ integration time, alternating between the two filters. The software used for data acquisition and control of the filter wheel was written in the Labview programming language (National Instruments, Austin, Texas). Further analysis of the data was performed in Microsoft Excel.

2.6.

In-Vitro Experiments Using Biosensor Cells

For in-vitro measurements, the same photon counter/filter wheel apparatus was used with the light guide inserted into a solid metal block designed to hold a $10\times 75\text{-}\mathrm{mm}$ glass tube. Biosensor cells (100,000 cells) were trypsinized and resuspended in $30\mu \mathrm{l}$ of a buffer comprising $160\text{-}\mathrm{mM}$ NaCl, $5.4\text{-}\mathrm{mM}$ KCl, $1.8\text{-}\mathrm{mM}$ $\mathrm{Ca}{\mathrm{Cl}}_2$ , $0.8\text{-}\mathrm{mM}$ $\mathrm{Mg}{\mathrm{Cl}}_2$ , $25\text{-}\mathrm{mM}$ HEPES, $5.6\text{-}\mathrm{mM}$ glucose, $4\text{-}\mathrm{mM}$ glutamine, and pH 7.4. Cells were preincubated in buffer for $15\min$ . To initiate the bioluminescence reaction, $10\text{-}\mu \mathrm{l}$ coelenterazine 400a was added to give a final concentration of $10\mu \mathrm{M}$ . After light acquisition had been started, vasopressin was added ( $2\mu \mathrm{l}$ , giving a final concentration of $25\mu \mathrm{M}$ ), without interrupting data acquisition. For experiments at $37°\mathrm{C}$ , the block holding the glass tube was placed on a heating mat.

3.1.

A Skin/Body Wall Window Enables Measurement of Light Emission from Bioluminescent Cells Transplanted Into the Kidney

The aim of the present experiments was to investigate whether a plastic window inserted into the skin and body wall of the mouse could be used to facilitate the measurement of bioluminescence originating in an internal organ. If successful, this method would avoid problems arising from transmission of light through tissues, including the skin, which absorb light strongly at wavelengths $<600\mathrm{nm}$ . To assess whether it was possible to simultaneously measure RLuc and FLuc bioluminescence originating in an internal mouse organ, cells expressing both proteins were transplanted into the kidney of immunodeficient mice and were permitted to establish a vascular supply, as previously described for other cell types.^{14, 15, 16} Insertion of the window into the skin/body wall and transplantion of bioluminescent cells into the kidney was performed in a single surgical procedure, as described in Materials and Methods in Sec. 2. The size and shape of the window allowed one end of a fiber optic light guide to be inserted within it when the mouse was immobilized under anesthesia. The light guide was used to transmit light to a photon counter (Fig. 1).

The use of a woven polyester collar attached to the plastic window allowed a tight integration of the window into the tissues of the mouse (between the skin and body wall) without loss of integrity of the skin around the window. This was important to prevent infection and fluid loss. Windows remained usable for at least $40\text{days}$ following surgery. There was no evidence of loss of transparency or of detachment of windows from the surrounding mouse tissues, problems that were encountered with a simple polyvinyl chloride window previously used in this laboratory.²⁵ Figure 2 shows details of the histological structure of the integration of the polyester collar into the skin. The polyester fibers were found to be united into the tissue between the dermis and the body wall with no evidence of inflammation or foreign body reaction. Connective tissue filled the spaces between the fibers, thus providing a strong bond between the materials of the window and the internal tissues of the mouse. Different forms of polyester fibers have been used in surgical materials for many decades, and the excellent surgical results are consistent with the known biocompatibility of this polymer.³² The bioluminescent cells formed a thick solid vascularized layer of tissue beneath the capsule of the kidney [Fig. 2d].

Fig. 2

Appearance and histological structure of skin/body wall window and transplanted cells in the kidney. (a) External appearance of a skin/body wall window at $35\text{days}$ following surgery; the inset shows a higher-power view of a window in a different animal. (b) Section of skin and body wall close to the window. The woven polyester collar is visible as fibers $f$ within connective tissue $c$ between the skin $s$ and the muscle $m$ of the body wall. Hematoxylin and eosin stain. Scale $\text{bar}=1\mathrm{mm}$ . (c) Detailed view of polyester fibers $f$ showing complete integration into the subdermal connective tissue $c$ . Scale $\text{bar}=1\mathrm{mm}$ . (d) Bioluminescent cells $b$ forming a thick layer beneath the capsule of the mouse kidney $k$ . Scale $\text{bar}=1\mathrm{mm}$ .

3.2.

Comeasurement of Light Production by Renilla Luciferase and Firefly Luciferase in Transplanted Bioluminescent Cells

In these experiments, our aim was to assess the feasibility of simultaneous measurements of bioluminescence of RLuc and FLuc, and also simultaneous measurement of RLuc bioluminescence and ${\mathrm{GFP}}^2$ fluorescence, which is required for measurement of BRET in cell-based biosensor applications. The characteristics of the optical filters used for these studies are shown in Fig. 3 .

Fig. 3

Transmission characteristics of filters and emission spectra of proteins used in these experiments. Dotted lines indicate transmission characteristics of filters, and solid lines indicate emission spectra of proteins. (a) A $475∕50\text{-}\mathrm{nm}$ bandpass filter and a $580\text{-}\mathrm{nm}$ longpass filter were used for discrimination of bioluminescence of RLuc and FLuc. The spectrum for RLuc (native coelenterazine substrate) is replotted from Ref. 47. The spectrum for FLuc (luciferin substrate) is replotted from Ref. 9. Note that the spectrum shown is the emission of FLuc at $37°\mathrm{C}$ . The emission peak of FLuc is $612\mathrm{nm}$ at $37°\mathrm{C}$ versus $578\mathrm{nm}$ at $25°\mathrm{C}$ .⁹ (b) A $417∕60\text{-}\mathrm{nm}$ bandpass filter and a $500\text{-}\mathrm{nm}$ longpass filter were used for discrimination of RLuc bioluminescence and light emitted by ${\mathrm{GFP}}^2$ by bioluminescence resonance energy transfer (BRET) in the presence of coelenterazine 400a. The combined spectrum $(\mathrm{RLuc}+{\mathrm{GFP}}^2)$ is replotted from Ref. 48. Filter transmission characteristics are from data supplied by Semrock and Omega Optical Incorporated.

In initial experiments, we tested the suitability of the filters for discriminating RLuc and FLuc bioluminescence. We used extracts of cells transfected with either RLuc or FLuc. The photon counter used for light measurements was the same as that later used in animal experiments. When a cell extract containing RLuc was incubated with native coelenterazine as substrate, light emission measured with the 580 LP filter was $\sim 2\%$ of that measured with the $475∕50$ BP filter. The amount of light detected at $>580\mathrm{nm}$ is slightly less than would be predicted based on the published spectrum of RLuc (Fig. 3). This is probably accounted for by the fact that photomultiplier tubes, such as the one used in the photon counter, are less sensitive to light at longer wavelengths.³³ When a cell extract containing FLuc was incubated with luciferin and ATP, light emission measured with the $475∕50$ BP filter was $\sim 1\%$ of that measured with the 580 LP filter. This is consistent with the published spectrum of FLuc (Fig. 3).

To assess the suitability of these filters for measurements of light emission at different wavelengths in living animals, cells expressing both RLuc and FLuc were transplanted into the kidneys of a series of immunodeficient mice. Plastic windows were fitted into the skin and body wall adjacent to the transplanted cells, as described earlier. Animals were used for experiments at $20\text{to}30\text{days}$ following surgery, when the transplanted cells had become vascularized, as shown in Fig. 2d. Native coelenterazine, which has been shown to be an optimal in-vivo substrate for RLuc, was injected intraperitoneally at a dose previously used in the literature, $2\mathrm{mg}∕\mathrm{kg}$ .³⁴ Light emission rose very rapidly, so that at about $30\mathrm{s}$ after injection, light detected with the $475∕50$ BP filter had risen to $\sim \mathrm{200,000}\text{counts}∕\mathrm{s}$ [Fig. 4a ]. Before injection of coelenterazine, light emission was $\sim 10\text{counts}∕\mathrm{s}$ . The delay of $\sim 30\mathrm{s}$ from injection to first measurement is accounted for by the time taken to place the anesthetized mouse into the light-tight box, in which the photon counter and light guide are housed. Light detected with the 580 LP filter was $\sim 3\%$ of the level of that detected with the $475∕50$ BP filter, and showed the same temporal pattern. We concluded that for RLuc/native coelenterazine, there is little difference in light detected with the 580 LP filter when intact animal and in-vitro results are compared ( $\sim 3\%$ versus $\sim 2\%$ ). The most likely cause of the slight increase in the intact animal is nonspecific fluorescence of tissues or other components in the light path.³⁵ We showed that it was not caused by the plastic windows. When an assembled window was placed in the light path in vitro, no change in light emission from RLuc was detected with the 580 LP filter.

Fig. 4

Discrimination of bioluminescence produced by RLuc and FLuc in transplanted bioluminescent cells. Mice received transplants of RLuc/FLuc-expressing biosensor cells in the kidney and were fitted with skin/body wall windows adjacent to the transplants, as shown in Fig. 1. Experiments were performed at $20\text{to}30\text{days}$ following cell transplantation. (a) Native coelenterazine $(2\mathrm{mg}∕\mathrm{kg})$ was injected intraperitoneally into a mouse with a bioluminescent cell transplant. Light was measured using $475∕50\text{-}\mathrm{nm}$ bandpass and $580\text{-}\mathrm{nm}$ longpass filters (see Fig. 3) over $\sim 1600\mathrm{s}$ . (b) Luciferin $(150\mathrm{mg}∕\mathrm{kg})$ was injected intraperitoneally into a mouse with a bioluminescent cell transplant. Light was measured using the same pair of filters over $\sim 1000\mathrm{s}$ . A second dose of luciferin $(150\mathrm{mg}∕\mathrm{kg})$ was then injected, and light emission was measured for another $\sim 800\mathrm{s}$ . (c) A mixture of native coelenterazine $(0.05\mathrm{mg}∕\mathrm{kg})$ and luciferin $(300\mathrm{mg}∕\mathrm{kg})$ was injected intraperitoneally into a mouse with a bioluminescent cell transplant. Light emission was measured over $\sim 2500\mathrm{s}$ using the same pair of filters. (d) Coelenterazine/cyclodextrin complex ( $14\text{-}\mathrm{mg}$ coelenterazine/kg) was injected intraperitoneally into a mouse with a bioluminescent cell transplant. Light emission was measured using the $475∕50$ bandpass filter over $\sim 5000\mathrm{s}$ .

Luciferin was injected intraperitoneally in a mouse with a kidney transplant of bioluminescent cells. Luciferin was administered at $150\mathrm{mg}∕\mathrm{kg}$ , a dose previously used in the literature.³⁴ In this experiment, light production by FLuc/luciferin in bioluminescent cells in the kidney was less than that from RLuc/coelenterazine ( $\sim \mathrm{8,000}\text{counts}∕\mathrm{s}$ versus $\sim \mathrm{200,000}\text{counts}∕\mathrm{s}$ ) [Fig. 4b]. Injection of a second dose of luciferin produced an increase in light to $\sim \mathrm{22,000}\text{counts}∕\mathrm{s}$ . We conclude that the dose of luciferin required to produce a strong level of light emission from FLuc is much greater than the dose of native coelenterazine required for production of light from RLuc ( $300\mathrm{mg}∕\mathrm{kg}$ versus $2\mathrm{mg}∕\mathrm{kg}$ ). This conclusion is similar to that from previous studies in which RLuc/FLuc bioluminescence in mice was assessed by CCD camera imaging.³⁴ Even though it is lower than the light emission from RLuc, the light production rate from FLuc following injection of a single dose of luciferin $(\sim \mathrm{10,000}\text{counts}∕\mathrm{s})$ would be sufficient for most experiments.

We then performed experiments in animals with transplanted bioluminescent cells in which a small dose of coelenterazine $(0.05\mathrm{mg}∕\mathrm{kg})$ was coinjected with a higher dose of luciferin $(300\mathrm{mg}∕\mathrm{kg})$ [Fig. 4c]. In this case, light detected with the $475∕50$ BP and 580 LP filters rose at about the same rate over a period of $\sim 750\mathrm{s}$ and then began a more or less linear decline.

We performed a series of measurements to determine the reproducibility of simultaneous assessments of RLuc and FLuc bioluminescence in living animals. Over a period of $10\text{days}$ , six separate measurements of bioluminescence from a transplant of biosensor cells were made. On each occasion the combination of $0.05\text{-}\mathrm{mg}∕\mathrm{kg}$ coelenterazine and $300\text{-}\mathrm{mg}∕\mathrm{kg}$ luciferin was injected. Over these six measurements, the RLuc bioluminescence/FLuc bioluminescence ratio was determined to be $10.6\pm 0.9$ . This result indicates that the RLuc bioluminescence/FLuc bioluminescence ratio can be reproducibly measured from biosensor cells in living animals.

One concern that has been raised regarding the use of coelenterazine is that it is typically dissolved in methanol or ethanol, and that toxicity can arise when animals are repeatedly used for RLuc bioluminescence measurements.⁹ We therefore performed experiments in which native coelenterazine was complexed with cyclodextrin.³¹ The complex was injected intraperitoneally at a dose of $0.7\text{-}\mathrm{mg}∕\mathrm{kg}$ coelenterazine. Light emission rose very rapidly but never attained the very high levels observed with uncomplexed coelenterazine. However, the cyclodextrin complex had the advantage that light emission was much more stable and remained at $>7000\text{counts}∕\mathrm{s}$ for $\sim 5000\mathrm{s}$ [Fig. 4d].

3.3.

Bioluminescence Resonance Energy Transfer Measurements in Biosensor Cells in the Kidney

An important application of skin windows in bioluminescence studies is accurate measurement of the ratio of light emitted at two different wavelengths. This is required for measurement of BRET in cell-based biosensor applications. In the experiment shown in Fig. 5a , we transplanted cells into the kidney that express two fusion proteins, V2-vasopressin-receptor-RLuc and ${\mathrm{GFP}}^2\text{-}\beta$ -arrestin2. This combination of proteins produces BRET in the presence of coelenterazine 400a, and BRET increases in the presence of vasopressin.¹² We show here that BRET may be observed in the biosensor cells transplanted into the kidney. Following injection of coelenterazine 400a into the animal, the light emission by RLuc rose rapidly; light emission by ${\mathrm{GFP}}^2$ also rose, giving a relatively stable BRET ratio of $\sim 0.1$ . Following injection of vasopressin, there was a delay of $\sim 200\mathrm{s}$ , and then the BRET ratio showed an approximately linear increase over $500\mathrm{s}$ , stabilizing again at $\sim 0.4$ .

Fig. 5

Observations on BRET in biosensor cell transplants and in cell suspensions in vitro. (a) Vasopressin-sensitive biosensor cells were prepared as described in Materials and Methods in Sec. 2 and were transplanted beneath the kidney capsule of immunodeficient mice. Cells express two fusion proteins— ${\mathrm{GFP}}^2\text{-}\beta$ -arrestin2 and V2-vasopressin-receptor-RLuc8. Following an intraperitoneal injection of $2\text{-}\mathrm{mg}∕\mathrm{kg}$ coelenterazine 400a, light emission was measured using $417∕60$ BP and 500 LP filters (gray and black symbols, respectively). The triangles show the ratio of light detected using the two filters [BRET $\text{ratio}=\left(500\mathrm{LP}\right)∕(417∕60\mathrm{BP})$ ]. After $\sim 350\mathrm{s}$ , light emission acquisition was stopped temporarily and vasopressin $(0.5\mathrm{mg}∕\mathrm{kg})$ was injected intraperitoneally into the animal. Light emission using the two filters was then measured for another $\sim 1100\mathrm{s.}$ (b) Vasopressin-sensitive biosensor cells were studied in vitro. A suspension of 100,000 cells was used in an incubation with coelenterazine 400a at $25°\mathrm{C}$ (see Materials and Methods in Sec. 2). Light emission using $417∕60$ BP and 500 LP filters and the BRET ratio are plotted. After $100\mathrm{s}$ , vasopressin was added to the cells without interrupting light acquisition. (c) The experiment was performed as described in (b), except that the cell suspension was maintained at $37°\mathrm{C}$ throughout the experiment.

We compared the in-vivo and in-vitro responsiveness of the biosensor cells. One factor of interest was the relative response of the cells at room temperature and at $37°\mathrm{C}$ . Most studies of BRET in cells have been at room temperature (e.g., see Refs. 36, 37 for cells expressing this vasopressin-responsive biosensor protein pair). In the living animal, the cells function at $37°\mathrm{C}$ . Figures 5b and 5c show that when a suspension of biosensor cells was exposed to vasopressin in vitro, using the same photon counter apparatus and filters, BRET increases in response to vasopressin were rapid and sustained. The response was greater and more rapid in cells incubated at $37°\mathrm{C}$ rather than $25°\mathrm{C}$ . At $25°\mathrm{C}$ , the BRET ratio rose from 0.016 to 0.030 with a $t_{1∕2}$ of $140\mathrm{s}$ , whereas at $37°\mathrm{C}$ it rose from 0.016 to 0.034 with a $t_{1∕2}$ of $80\mathrm{s}$ . Therefore, the strong response to vasopressin in biosensor cells in the living animal results at least partially from the higher temperature in vivo.