2.1.

Induction of Arthritis

All animal experiments were approved by the Stanford University Animal Research Internal Review Board. DBA1/J mice were purchased from Jackson Laboratories (Bar Harbor, Maine). Bovine type II collagen dissolved at $2\mathrm{mg}∕\mathrm{ml}$ in $0.05\text{-}\mathrm{M}$ acetic acid was purchased from Chondrex. Complete Freund’s Adjuvant (CFA) (Sigma, St. Louis, Missouri), containing 2-mg/ml inactivated M. tuberculosis and collagen were mixed to produce a thick emulsion. 8 to 10 weeks old DBA1/J mice were injected with $100\mathrm{\mu }\mathrm{l}$ of the collagen/CFA emulsion subcutaneously $2\mathrm{cm}$ from the base of the tail. After 21 days, these mice were boosted subcutaneously with $100\mathrm{\mu }\mathrm{l}$ of an emulsion containing equal volumes of Incomplete Freund’s Adjuvant and collagen at the base of the tail. The incidence of inflammation was then synchronized by injecting $50\mathrm{\mu }\mathrm{g}$ of lipopolysaccharide (LPS) intraperitoneally, 3 days after the boost. Within a week, 80 to 100% of the mice developed paw inflammation.

2.2.

Lentiviral Vectors and Reporter Gene Constructs

The lentiviral vector pHR1 was generated by replacing eGFP in ${\mathrm{pHR}}^{\prime }\mathrm{tripCMV}\text{-}\mathrm{eGFP}\text{-}\sin$ (provided by Breckpot¹⁵) with a new multiple cloning site. The triple fusion gene hrLuc-mRFP-HSV1-tsr39tk (TFR)¹⁴ was subcloned from pcDNA3.1 to pHR1. For virus production, $293\mathrm{T}$ cells were plated at $15\times {10}^6$ cells per ${\mathrm{T}}_{175}$ flask in complete Iscove's Modified Dulbecco's Medium (IMDM) [containing 10% fetal calf serum (FCS), $100\text{-}\mathrm{U}∕\mathrm{ml}$ penicillin/streptomycin, and $2\text{-}\mathrm{mM}$ L-glutamine]. After $24\mathrm{h}$ , they were transfected by calcium/phosphate method ( $45\mathrm{\mu }\mathrm{g}$ of lentiviral vector pHR, $30\mathrm{\mu }\mathrm{g}$ of packaging vector $p\Delta$ , and $15\mathrm{\mu }\mathrm{g}$ of vesticular stomatitis virus G (VSVG)-encoding vector per flask) in the presence of chloroquine ( $25\mathrm{\mu }\mathrm{M}$ , Sigma). After 8 to $12\mathrm{h}$ , the transfection medium was replaced with normal complete IMDM medium. The virus-containing supernatant was harvested 24 to $36\mathrm{h}$ later, filtered through a $0.45\text{-}\mathrm{\mu }\mathrm{m}$ polyethersulphone (PES) filter, and ultra-centrifuged for $2\mathrm{h}$ and $20\min$ at $\mathrm{19,500}\mathrm{rpm}$ . Viral particles were resuspended in a small volume of plain IMDM medium and frozen at $-80°$ C. Viral titer was estimated using dilutions of the virus on $293\mathrm{T}$ cells and flow cytometry analysis of $\mathrm{mRFP}+$ cells two days after infection.

2.3.

T-Cell Hybridomas and Stable Reporter Gene Transduction

T-cell hybridomas $\mathrm{A}{2}^9$ and AG6 were obtained by fusion of BW5147 TCR $\alpha \beta$ negative thymoma cells with an equal number of freshly activated $\mathrm{CD}4+$ T cells from collagen type II-specific TCR-transgenic mice,¹⁶ in presence of PEG 1500 (Roche, manufacturer’s protocol) and selected for 3 weeks in presence of sodium hypoxanthine, aminopterin, and thymidine (HAT) (Gibco, Carlsbad, California) and 2 weeks in presence of sodium hypoxanthine and thymidine (HT) (Sigma). A2 and AG6 were two of many $\mathrm{CD}4+$ $\mathrm{V}\beta 8^{+}$ clones isolated by cell sorting (FACStar, BD, Franklin Lakes, New Jersey). Cells were then infected with TFR virus at MOI of 10 in presence of protamine sulfate $(10\mathrm{\mu }\mathrm{g}∕\mathrm{ml})$ . $\mathrm{CD}{4}^{+}$ $\mathrm{V}\beta 8^{+}$ ${\mathrm{RFP}}^{+}$ cells were repeatedly purified by cell sorting.

2.4.

In Vitro Humanized Renilla Luciferase Activity Assay

T-cell hybridomas were harvested and live cells were counted by Trypan blue exclusion. A known number of pelleted cells were lysed with $1\times$ passive lysis buffer (Promega) and incubated at room temperature for $15\min$ . The lysate was then centrifuged at 13,000 RPM for $3\min$ and the protein concentration of lysate supernatant was measured. The amount of lysate, equivalent to $1\mathrm{\mu }\mathrm{g}$ of protein, was added to $1\times$ phosphate buffered saline (PBS) (total volume= $100\mathrm{\mu }\mathrm{l}$ ) and hRluc activity was measured with a luminometer $\left(10\mathrm{s}\right)$ immediately after addition of $1\text{-}\mathrm{\mu }\mathrm{g}$ coelenterazine. Based on the measurements, the in vitro hRluc activity per cell was calculated. Two different luminometers were used for the studies described in this work. At the time the trafficking of A2-TFR cells was being studied, the luminometer Turner Designs model 2020 was available in our laboratory. At the time the trafficking of AG6-TFR cells was being studied, the luminometer Turner Biosystems model 2030-001 was available at our laboratory. Comparing the measurements from both systems, model 2030-001 was on average 1444 times higher than model 2020.

2.5.

Bioluminescence Imaging of Collagen-Induced Arthritic Mice

CIA mice were anesthetized initially with 2% isofluorane and then by $20\mathrm{\mu }\mathrm{l}$ of a ketamine $(80\mathrm{mg}∕\mathrm{ml})$ and xylazine $(20\mathrm{mg}∕\mathrm{ml})$ solution. Once anesthetized, CIA mice paws were spread and taped down with transparent scotch tape on black cardboard. Coelenterazine, a bioluminescent substrate of the Rluc enzyme, was dissolved in a 1:1 solution of ethanol:polyethylene glycol at a concentration of $5\mathrm{\mu }\mathrm{g}∕\mathrm{\mu }\mathrm{l}$ , and $20\mathrm{\mu }\mathrm{l}$ was aliquoted in amber microcentrifuge tubes, stored at $-80°$ C. Just prior to injection into each mouse, the coelenterazine stock solution in each tube was diluted by adding $80\mathrm{\mu }\mathrm{l}$ of sterile $1\times \mathrm{PBS}$ . $100\mathrm{\mu }\mathrm{g}$ of coelenterazine was injected through a lateral tail vein of the mice before scanning with a Xenogen IVIS-100 charge-coupled device (CCD) camera for $5\min$ , immediately after injection.

2.6.

Micro-Positron Emission Tomography Imaging of Collagen-Induced Arthritic Mice

9-[4- $\left[{}^{18}\mathrm{F}\right]$ fluoro-3–(hydroxymethyl)butyl]guanine ( $\left[{}^{18}\mathrm{F}\right]$ FHBG), the PET reporter probe (PRP) substrate of the HSV1-sr39TK enzyme, was used to detect intravenously injected TFR expressing T-cell hybridomas in CIA mice with a microPET R4 scanner (Siemens, Malverne, Pennsylvania). The synthesis, specific activity, and radiochemical purity of $\left[{}^{18}\mathrm{F}\right]\mathrm{FHBG}$ in our laboratory has been described by Yaghoubi ¹⁷ Approximately $200\mathrm{\mu }\mathrm{Ci}$ of $\left[{}^{18}\mathrm{F}\right]\mathrm{FHBG}$ was injected into the lateral tail veins of CIA mice, two hours prior to the microPET scan. Mice were conscious for the two-hour period. Prior to the scan, the investigator emptied each mouse’s bladder as much as possible by exerting pressure on the bladder. Then the mice were anesthetized with 2.5% isoflurane. The entire body of each mouse was scanned for $10\min$ for $\left[{}^{18}\mathrm{F}\right]\mathrm{FHBG}$ activity using a R4 microPET system (Siemens). This system generates images with approximate resolution of $2\mathrm{mm}$ . Images were reconstructed using the ordered subsets expectation maximization (OSEM) algorithm and analyzed with AMIDE software.¹⁸

2.7.

Assays to Determine a Mathematical Relationship between the Number of Cells in the Inflamed Paws and Light Intensity from Bioluminescence Images

To derive equations that could estimate the number of cells in the paws of CIA mice from bioluminescence images, we needed to determine the amount of light detected by the bioluminescence camera from a known number of cells implanted in the paws. Cultured A2-TFR or AG6-TFR cells were harvested and counted on the day of injection. Various numbers of cells, ranging from 1000 to 800,000, were resuspended in $30\mathrm{\mu }\mathrm{l}$ of sterile $1\times$ PBS in microcentrifuge tubes. The harvested cells were also assayed in vitro for the level of hRluc activity (RLU) per $\mathrm{\mu }\mathrm{g}$ of protein and per cell. DBA1/J mice were anesthetized, injected with coelenterazine, and scanned with the bioluminescence camera to acquire a background scan. Afterward, $30\mathrm{\mu }\mathrm{l}$ of thawed Matrigel was mixed with the predetermined number of cells in $1\times$ PBS and implanted subcutaneously in the paws of the DBA1/J mice $(\mathrm{n}=12)$ . Up to $80\mathrm{\mu }\mathrm{l}$ of ketamine $(80\mathrm{\mu }\mathrm{g}∕\mathrm{ml})$ /xylazine $(20\mathrm{\mu }\mathrm{g}∕\mathrm{ml})$ had been injected into each mouse to make sure they stayed completely unconscious during the entire procedure. The same mice, which had background scans and implanted cells, were again scanned with the bioluminescence imaging camera following injection of $100\text{-}\mathrm{\mu }\mathrm{g}$ coelenterazine. Approximately $15\min$ elapsed between the time of cell implantation and the second bioluminescence scan. These procedures were repeated by injecting known quantities of cells deeper into the paws of a limited number of deeply anesthetized DBA1/J mice $(\mathrm{n}=3)$ to assess the effect of tissue depth on signal intensities. Mice were sacrificed prior to regaining consciousness. Regions of interest were drawn, including the entire paw on each image, to measure the total light flux and maximum radiance from each paw. The measurements in the ROI of each paw from pre-cell injection images were subtracted from the measurements in the ROI of the same paws from the post-cell injection images to obtain net total light flux or maximum radiance due to the presence of the cells. The levels of luciferase activity per cell were not the same during all cell trafficking imaging experiments, because the cells were not being expanded from a single clone of transduced T-cell hybridomas and TFR gene expression is only relatively stable. Therefore, we decided to relate the total light flux or maximum radiance from each paw to the total in vitro RLU, which equals the number of cells implanted in the paw times in vitro luciferase activity per cell. In this manner, one is able to derive linear equations from the correlation between total light flux or maximum radiance and total in vitro RLU. Also, these linear equations should be generally useful for all cells expressing the hRluc transgene as long as the hRluc activity of the cells does not vary significantly from the in vitro activity once the cells are inside mice.

2.8.

Quantitative Cell Trafficking Imaging Studies

Within 1 to 2 weeks after sorting, A2-TFR or AG6-TFR cells that had been grown in culture were harvested and counted in preparation for intravenous injection into CIA mice. Approximately 1 million cells were assayed for in vitro hRluc activity. Known quantities of the cells (A2-TFR: $5\times {10}^6$ , $10\times {10}^6$ , $13\times {10}^6$ , $25\times {10}^6$ , and $50\times {10}^6$ ; AG6-TFR: $5\times {10}^6$ , $13\times {10}^6$ , $25\times {10}^6$ , and $50\times {10}^6$ ), resuspended in $100\text{-}\mathrm{\mu }\mathrm{l}$ PBS, were injected into a lateral tail vein of CIA mice, within 1 week after the LPS boost. All of the mice (A2-TFR injected: $\mathrm{n}=19$ ; AG6-TFR injected: $\mathrm{n}=22$ ) were scanned with a bioluminescence imaging camera following injection of $100\text{-}\mathrm{\mu }\mathrm{g}$ coelenterazine on days 1 and 3. AG6-TFR injected mice were also scanned on day 6. Before scanning, every paw of each mouse was scored for the degree of inflammation [0=no inflammation; 1=only fingers inflamed; 2=mild inflammation of the entire paw (ankle not swollen); 3=moderate inflammation of the entire paw with possible swollen ankle; 4=severe inflammation of the entire paw, including ankle]. CIA mice with various degrees of inflammation in their paws, which had either been injected with various quantities of A2 or AG6 cells or no cells, were also scanned to obtain average background total light flux and maximum radiance in paws with 0, 1, 2, 3, or 4 inflammation scores. ROIs were drawn over each of the paws and total photon flux and maximum radiance of the entire paw was measured. Net total light flux and maximum radiance were calculated by subtracting the average background measurements. The number of cells in each paw was calculated, using the equations derived through procedures in Sec. 2.7 (Table 1 ). Some of the A2-TFR $(\mathrm{n}=14)$ and AG6-TFR $(\mathrm{n}=7)$ injected CIA mice were scanned also by microPET as described in Sec. 2.6 for $\left[{}^{18}\mathrm{F}\right]\mathrm{FHBG}$ activity in the paws. Mice $(\mathrm{n}=7)$ injected with control T-cell hybridomas (not expressing the TFR transgene) were also scanned by microPET to determine the background activity of $\left[{}^{18}\mathrm{F}\right]\mathrm{FHBG}$ in the paws with different inflammation scores. The objective of these preliminary microPET studies was to determine whether the average $\left[{}^{18}\mathrm{F}\right]\mathrm{FHBG}$ activity was significantly higher than the background in the inflamed paws of A2-TFR and AG6-TFR injected CIA mice.

3.1.

Derivation of Linear Equations for Calculating the Number of Cells from Bioluminescence Images of Mice Paws

Eight different linear equations (Table 1) were derived to relate bioluminescence signal intensities from the paws of CIA mice to the total RLU (based on in vitro measurement), four for A2-TFR and four for AG6-TFR cells, because a different luminometer had been used during the trafficking analysis of each of these cells, as discussed in Sec. 2.4. For each cell line, two of the equations related total light flux to total RLU, and two of the equations related maximum radiance to total RLU. The same data were used to derive the equations for A2-TFR and AG6-TFR; hence the only difference between their equations is the slope of the lines. The difference in the slope of the lines is due to the fact that the total RLUs of AG6-TFR cells are 1444 times higher than A2-TFR, since different luminometers were used to measure in vitro hRluc activity per cell. The total light flux is the sum of all photons in every pixel of the image included in the ROI; whereas maximum radiance is the maximum photon intensity found in a pixel of the ROI. For example, if an equal number of cells traffic to each of two paws, in the paw where cells are more spread out, the maximum radiance may be lower than the other paw in which cells concentrate in specific locations of the paw; whereas the total light flux of both paws may be similar. The linear regression lines illustrating the correlation between total light flux and total RLU and maximum radiance and total RLU for A2-TFR cells are shown in Figs. 1a, 1b, 1c, 1d . As illustrated, for each parameter relationship, linear regressions were calculated for two ranges of total RLU. Figures 1a and 1c include the entire range of total RLU studied (1,598 to 318,005); whereas Figs. 1b and 1d include a shorter range of total RLU (1,598 to 15,983), respectively. Despite the fact that ${\mathrm{R}}^2$ values are higher when the entire total RLU range is included, equations for the shorter range of total RLU were necessary, because when cells were injected intravenously, the signal intensities observed corresponded to the shorter range of total RLU. Therefore, for the remainder of the study, we only use the equations for the shorter range of total RLU to calculate the number of cells trafficking to the paws (Table 1). Statistical analysis indicates that the relationships expressed as linear equations between the total RLU and bioluminescence image signal intensities were significant (Table 1). Once total RLUs are calculated from image signal intensities, the number of cells in the paws can be estimated by dividing with the in vitro hRluc activity per cell. The range of in vitro hRluc activity was 0.23 to 1.73 RLU/cell for A2-TFR (as measured with the model 2020 luminometer), and 1980 to 2926 RLU/cell for AG6-TFR (as measured with model 2030-001).

Fig. 1

Relationship between net image signal intensity in the paws of mice injected with various known quantities of A2-TFR cells and the total RLU of A2-TFR cells. (a) and (b) relate total light flux to total RLU. (c) and (d) relate max radiance to total RLU. The range of total RLU in (a) and (c) is 1598 to 318,005, and in (b) and (d) is 1598 to 15,983. Data are averages of two measurements $\pm$ SEM.

We also wanted to address the effect of tissue depth on the signals emitted from the paws. In additional studies, we anesthetized three DBA1/J mice with a high dose of $80\text{-}\mathrm{\mu }\mathrm{l}$ ketamine $(80\mathrm{mg}∕\mathrm{ml})$ /xylazine $(20\mathrm{mg}∕\mathrm{ml})$ and made sure they were deeply unconscious before injecting the same known quantities of AG6-TFR cells intra-articularly, as had been injected in the other mice subcutaneously. On average, the measured net light intensities after injecting cells deeper into the paws were not lower than the net light intensities from the paws injected the same number of cells subcutaneously. This may be due to the fact that even when cells are injected subcutaneously, some of them end up at the bottom of the paw and some at the top, such that the light emitted from some of the cells must still penetrate about $2\mathrm{mm}$ of tissue and is highly attenuated. There was also a very high variation between measurements from the paws that were injected deeply. For example, the average total light flux from two paws injected with 10,000 cells was 154,865, but the standard error of the mean (SEM) was 102,003. More variations were expected because intra-articular injections are technically more challenging; thus the subcutaneous method may be the most reliable technically feasible method.

3.2.

Quantitative Analysis of Antigen Specific T-Cell Hybridoma Trafficking to the Paws of Collagen-Induced Arthritic Mice

DBA1/J mice had been induced and boosted with bovine collagen type II and then further boosted with $50\text{-}\mathrm{\mu }\mathrm{g}$ LPS, resulting in the appearance of inflammation in at least one paw in 80 to 90% of the mice. Various numbers of A2-TFR or AG6-TFR cells were then injected through a lateral tail vein of each CIA mouse to study their trafficking to the paws by bioluminescence imaging. For quantitative analysis, we had determined the in vitro RLU/cell of the cells on the day of intravenous (IV) injection. Each mouse was then scanned for bioluminescence following coelenterazine injection, 1 and 3 (and also 6 days for AG6-TFR injected mice) days after IV injection of the cells to determine the amount of light emitted from the paws due to A2-TFR or AG6-TFR cells. To determine the amount of signal specifically due to the cells, we needed to subtract the background signal in the paws. The level of background varied slightly due to the degree of inflammation and whether it was the front or hind paw. Therefore, we measured coelenterazine signals in the paws of CIA mice, who had either been injected with control A2 or AG6 cells or not been injected with any cells, and averaged the measured total light flux or maximum radiance in the front or hind paws with the same inflammation scores [Figs. 2a and 2b ]. These average background signals were subtracted to determine the net total light flux or maximum radiance due to A2-TFR or AG6-TFR cells. Then the number of cells present in each paw was estimated from the total light flux and maximum radiance measurements in each paw, using the equations derived in Sec. 3.1 [Figs. 1b and 1d or Table 1].

Fig. 2

Average background (a) total light flux and (b) max radiance in the hind and front paws of CIA with different inflammation scores.

Data presented in Figs. 3a and 3b illustrate the average number of A2-TFR cells estimated to be present in the paws with different degrees of inflammation on days 1 and 3 following intravenous injection of various numbers of these cells, based on the total light flux measurements in paw ROIs. Data presented in Figs. 3c and 3d illustrate estimates of the same based on maximum radiance measurements in paw ROIs. In general, the maximum radiance measurements and conversion formula yielded higher cell number estimates. As illustrated by data presented in Figs. 3a, 3b, 3c, 3d, injecting greater quantities of A2-TFR cells did not always result in a greater number of A2-TFR cells in the paws on days 1 and 3. Also, we cannot conclude based on the averages that a significantly greater number of A2-TFR cells consistently accumulated in the paws with higher inflammation scores on days 1 and 3. Although in some paws with higher degrees of inflammation, more A2-TFR cells accumulated on day 3 than on day 1 [Compare Fig. 3a and 3b and 3c and 3d].

Fig. 3

Average number of A2-TFR cells accumulating in the paws (with different inflammation scores of CIA mice 1 (a) and (c) and 3 (b) and (d) days following injection of various quantities of A2-TFR cells intravenously. Averages in (a) and (b) were calculated based on total light flux, and (c) and (d) were calculated based on max radiance. Error bars are SEM.

Data presented in Figs. 4a, 4b, 4c , illustrate the average number of AG6-TFR cells estimated to be present in the paws with different degrees of inflammation on days 1, 3, and 6 following intravenous injection of various numbers of these cells, based on the total light flux measurements in paw ROIs. Data in Figs. 4d, 4e, 4f illustrate estimates of the same based on maximum radiance measurements in paw ROIs. Comparing the data in Figs. 3a, 3b, 3c, 3d with those in Figs. 4a, 4b, 4c, 4d, 4e, 4f, in general fewer number of cells were present in the paws of CIA mice, injected with AG6-TFR cells IV. Again, injection of more AG6-TFR cells did not result in increased accumulation of these cells into the paws on days 1, 3, and 6. Based on the averages, we also cannot conclude that paws with higher degrees of inflammation accumulate a greater number of AG6-TFR cells on days 1, 3, and 6. The one difference observed between the trafficking trends of A2-TFR and AG6-TFR cells was that on average, greater numbers of AG6-TFR cells do not appear to accumulate in the paws from day 1 to 3 and 6. Finally, the AG6-TFR cell estimates based on maximum radiance measurements have lower SEM than those based on total light flux.

Fig. 4

Average number of AG6-TFR cells accumulating in the paws (with different inflammation scores of CIA mice 1 (a) and (d), 3 (b) and (e), and 6 (c) and (f) days following injection of various quantities of A2-TFR cells intravenously. Averages in (a), (b), and (c) were calculated based on total light flux, and (d), (e), and (f) were calculated based on max radiance. Error bars are SEM.

3.3.

Micro-Positron Emission Tomography Imaging of Collagen-Induced Arthritic Mice after Intravenous Injection of A2 and AG6-Triple Fusion Reporter Gene TFR Cells

Since A2 and AG6-TFR cells also express the HSV1-tsr39tk PRG, we also scanned the mice for $\left[{}^{18}\mathrm{F}\right]\mathrm{FHBG}$ activity, using a microPET, on various days following intravenous injection of the cells. Figure 5 illustrates microPET $\left(\right[{}^{18}\mathrm{F}\left]\mathrm{FHBG}\right)$ and bioluminescence (coelenterazine) imaging of a mouse injected with $25\times {10}^6$ A2 cells as a control, and another mouse injected with the same number of A2-TFR cells, one day after intravenous injection. Above background $\left[{}^{18}\mathrm{F}\right]\mathrm{FHBG}$ activity as well as bioluminescent signal was observed in the inflamed paws of the A2-TFR injected mouse, whereas only background activities were observed from the inflamed paws of the A2 injected mouse. Based on the background subtracted total light flux measurements, 775 cells were present in the left front paw, 1015 cells were present in the right front paw, 1887 cells were present in the left hind paw, and 2626 cells were present in the right hind paw. PET has different sensitivity and background levels than bioluminescence. As a result, when very few cells accumulated, we were unable to detect above background $\left[{}^{18}\mathrm{F}\right]\mathrm{FHBG}$ activity in the paws of several mice, despite above background hRluc activity, and thus were unable to establish an accurate correlation between bioluminescence and microPET images. The level of $\left[{}^{18}\mathrm{F}\right]\mathrm{FHBG}$ activity was also higher in the thorax and leg joints of the A2-TFR mouse, even though bioluminescence signal (image threshold to background) at those sites was below background. This could be due to the presence of A2-TFR cells in the lungs and liver that are not detected by bioluminescence imaging, because of deep tissue attenuation, but can better be detected by the microPET. In fact, we have confirmed by ex-vivo biodistribution assays (measuring tissue hRluc activity ex vivo) that hRluc activity is above background in the lung, liver, spleen, and several lymph nodes, as well as inflamed paws of CIA mice (data not shown). Furthermore, when these mice were kept for about one month after AG6-TFR injections, some of the mice developed tumors in their livers, lungs, and inflamed paws. These tumors emitted saturating bioluminescence following coelenterazine injection and accumulated significant amounts of $\left[{}^{18}\mathrm{F}\right]\mathrm{FHBG}$ .

Fig. 5

MicroPET (top) and bioluminescence (bottom) images of a CIA mouse injected with $25\times {10}^6$ A2 (left) and another mouse injected with $25\times {10}^6$ A2-TFR (right) cells intravenously. Images were acquired one day following cell injections. For the A2-TFR mouse, the inflammation score of the left front paw was 0, the right front paw was 4, the left hind paw was 2, and the right hind paw was 4. For the A2 mouse, the inflammation score of the left hind paw was 2, the right hind paw was 2, the right front paw was 4, and the left front paw was 2.

Table 1

Linear equations for the estimation of cell numbers in the paws of CIA mice from bioluminescence images.