Figure 1.
Monitoring of tumor growth by fluorescence imaging in the s. c. DLD-1 xenograft model.
Representative data sets of one and two tumors expressing eGFP (A) or dsRed2 (B) are displayed. The calculated tumor volume V (○) and the measured total FI normalized on exposure time (♦) were plotted over time. For demonstration of the method exemplary grayscale images (exposure time 10 ms), unmixed composite images (automatic exposure time; background - black, autofluorescence - white, eGFP - green, dsRed2 - red) and overlay images (intensity weighted pseudocolor mode, scale bar displayed beneath) of one mouse at various time points are given under each graph. In further presentations only representative pseudocolor images, which are most illustrative, will be displayed. As shown in A and B, green and red fluorescence signals increased over time like the tumor volumes. (C): Correlation between FI and V (10 tumors in 3 individuals representing two single tumor sites (circles) and four combined tumor sites (up and down triangles), eGFP FI - green symbols, dsRed2 FI - red symbols).
Table 1.
Data of potential fits for all analyzed HT-29 s. c. xenografts.
Figure 2.
Fluorescence color dependency of the correlation between V and FI for the HT-29 s. c. xenograft model.
Data are plotted for eGFP, dsRed2, TurboFP635 and mPlum emitting tumors with disregard of mouse individuals. It is obvious that the scattering of deep-red data points is smaller than those of the green and red fluorescent xenografts, respectively.
Figure 3.
Standard deviations of the FI as a function of the cell type and the fluorescent color.
The scattering of the S. D. of 5 different cubes based on the mean FI is shown. Thick black line: median of S. D. [%] (50th percentile), box bottom: 25th percentile, box top: 75th percentile, whiskers: 1.5× interquartile range, circle: outliers. (A): A great influence of the chosen xenograft model on the imaging error is observed. Each model has its specific growth characteristics (e. g. growth rate, vascularization or recruitment of murine stroma) and therefore the optical properties of the developing solid s. c. tumor differ greatly resulting in the observed variation. (B): Different fluorescence colors have specific optical properties in living tissues (e. g. penetration depth of excitation light or absorption of emitted fluorescence). This can lead to color dependent varying FI errors as shown here for green and deep-red fluorescence demonstrating superior characteristics of the latter.
Figure 4.
Monitoring of s. c. HT-29 eGFP and H12.1 eGFP xenografts after chemotherapy with Cisplatin (10 mg/kg body weight) in vivo.
The calculated tumor volume V (○) and the measured normalized total fluorescence intensity FI (♦) were plotted over time. Each plot represents one treated individual. Day of chemotherapy is marked with a vertical line (d29 or d26 post cell implantation, respectively). In the last row representative intensity weighted overlay images are shown corresponding to the plots in the row above.
Figure 5.
Monitoring of growth and chemotherapy response in the model of visceral growing mPlum HT-29 xenografts.
Combination therapy was given i. p. in two cycles (d0 - d3 & d21 - d24): On the first day the animals received an i. p. injection of Irinotecan (50 mg/kg body weight) and 5-Fluorouracil (30 mg/kg body weight) followed by 5-Fluorouracil single agent applications on day 2, 3 and 4. The plot shows the development of the normalized FI for the therapy (○; n = 7) and the control group (•; n = 4). Error bars are given as S. D. The overlay images show one representative individual per group. It becomes apparent that chemotherapy induced a clear decrease of FI indicating the response of tumors. Compared to untreated mice a successful prevention of tumor growth was achieved by the chemotherapy regimen.
Figure 6.
Correlation between FI and tumor weight.
The FI were plotted over the weights of the tumors that had been detected in vivo based on their mPlum signal (n = 11).