HOW TO QUANTIFY
IN YOUR ANIMAL MODELS
In vivo optical imaging is a semi-quantitative imaging modality, where signal intensities correlate positively with the
number of reporter molecules. Consequently, optical data can be used to answer experimental questions such as:
“What are the relative efficacies of two drug treatments against a given tumor cell line?” The drug treatment leading to the
greatest reduction in optical signal intensity will be considered the most effective therapeutic. Beyond such efficacy
studies, there are instances in which investigators will want to know the exact correlation between signal intensity and
cell number for a given reporter construct, either in vitro or in vivo. This data can be acquired, but it does require a bit
of extra work. Essentially, one needs to create a standard curve of optical signal intensity vs. cell number. Below is a
general protocol of how to generate both in vitro and in vivo standard curves (optical signal intensity vs. cell number) for
bioluminescent reporter systems.
NOTE: Theoretically, the procedure below can be applied to either bioluminescent or fluorescent reporter models.
In bioluminescent models, you will be defining a correlation between optical signal vs. number of living target cells present,
given that the functionality of bioluminescent enzymes (typically luciferases) will require not only their specific substrate, but
also oxygen and cellular ATP, the latter being present in only living cells. In fluorescent studies, you are typically defining a
correlation between optical signal and the total number of fluorescently tagged cells. Such labeled cells may be alive, dead or
dying as fluorophore functionality (the absorbance and emission of light) does not require cellular ATP. To exemplify how in
vivo standard curves can be made, we have opted here to use a general bioluminescent oncology model.
FIRST, MAKE AN IN VITRO STANDARD CURVE
These preliminary experiments will allow you to confirm up-front that the transduced tumor cell line of interest
generates a good bioluminescent signal intensity. Furthermore, the in vitro standard curve will allow you to approximate
the target cell number needed for producing in vivo standard curves (see below). In general terms, in vitro standard
curve data are acquired by titrating out a cell preparation into cell medium, in triplicate, at least (n ≥ 3), and then
measuring optical signal intensity of each titration in units of total photons/sec. You can then plot mean intensity
data (mean of total photons/sec ± standard deviation) vs. cell numbers to produce an in vitro standard curve
• Harvest your luciferase-transduced tumor cells of interest from tissue culture, tumor biopsy, or tumor
cell-burdened organ (e.g. T-cells from spleen), and determine cells/mL values through the use of a hemocytometer
and standard light microscope.
• Perform all cell dilution series (titrations) in black, non-fluorescing, non-reflective plastic flat-bottom 96-well plates.
Additionally, it is best to use 96-well plates that have a low-binding surface to minimize cell loss during
• You should ideally run two in vitro titrations. In the first run, you can use 1:10 dilutions. This will give you an
approximate correlation between signal intensity and cell number over a multiple-log range of cells/well. In the
second titration series, you can use shallow, 1:1 dilutions. This will allow you to determine the exact threshold of
bioluminescent cell detection.
• For the first 1:10 titration series, you can prepare and work with a starting cell solution that will give 10e6 cells/well.
You can then titrate across the 96-well plate, down to a concentration of ≤ 10e1 cells/well. When imaging this set of
titration samples with a Lago X, Ami HTX or Kino you can use short exposure times and modest binning (e.g., 5-10
sec, with 2×2 or 4×4 binning). Again, the data of this titration series will define an approximate correlation of
bioluminescent signal intensity vs. cell number, and you will need a second, 1:1 titration series to accurately define a
cell detection threshold.
• With your second 96-well titration study, you can start at 1 to 2 cell dilutions above the minimum detection
concentration (determined by the first titration series), and then run a 1:1 serial dilution across your 96-well plate.
For this second set of titration samples with fewer cells/well, you can use longer exposure times and higher binning
(e.g., 60-300 sec, 8×8 binning). Again, the data from this second in vitro titration will allow you to determine an
accuratedetection threshold of cells/well.
NOTE: The range of cells/well values to use in your in vivo titrations should be guided by results from your preliminary, in vitro
standard curves (above). It is a given that in vivo optical signal is attenuated by tissues, specifically by oxygenated hemoglobin
and melanin pigment. So, when performing in vivo cell titration studies, you will typically use a higher range of cells/well values
than that used in the in vitro titrations (see below).
SECOND, MAKE AN IN VIVO, ORTHOTOPIC STANDARD CURVE:
In general, you can serially dilute cells, and then immediately inject your cell samples into an animal model organ or
body compartment of interest. The optical signals from challenged, test animals will then be imaged by a Lago X, Ami
HTX or Kino system.
• Harvest luciferase-transduced tumor cells of interest from tissue culture, tumor biopsy, or tumor cell-burdened
organ (e.g. T-cells from spleen), and determine cells/mL values through the use of a hemocytometer and standard
• Given prior in vitro standard curve data (see above), prepare a stock cell solution to be used in the in vivo titration.
Typically, your stock cell solution for in vivo titration can be 10- to 100-fold that of the starting concentration used
in the in vitro 1:1 dilution series (see above).
• Serially titrate cells into a physiologically acceptable solvent (e.g. PBS), over an appropriate concentration range,
perhaps down the detection level observed in vitro.
• Prepare enough of each cell concentration so that you can perform several orthotopic bolus injections per
concentration. Typically, n = 5 per cell concentration is recommended, as it will lead to statistically tight data.
• Inject cell titration samples into your major organ/tissue site of interest. Regardless of the site of injection, always
aim to minimize the volume and speed of your bolus injection. This will minimize any site disturbance and/or any
cell leakage from site. You may choose to suspend cells in a Matrigel preparation to have a better retention of cells
at the injection site.
• Image animals immediately after they receive the bolus cell injection, using a Lago X, Ami HTX or Kino
• Run a Region of Interest (ROI) analysis of all optical signal data (using units of total photons/sec).
• For each dilution of the titration series, determine and plot mean signal intensity (mean total photons/sec
+/- standard deviation) vs. known cell number data.
• With the resulting in vivo, orthotopic standard curve, you will now be able to translate observed bioluminescent
signal values into in vivo, viable target cell loads. Remember, different bioluminescent reporters, cell lines,
orthotopic sites, and mouse strains will alter these in vivo standard curves. Any time such parameters are altered
in a model, a new in vivo standard curve should be determined.
When doing ROI analyses, we recommend that bioluminescent signals be quantified in units of “total emission” (total photons/
sec) and not in units of “mean radiance” (mean photons/sec/cm2
/sr). The reason for this is simple: Mean radiance is inversely
correlated with the amount of background area (i.e., low signal intensity area) included in drawn ROIs. If there is variability in
the ratio of signal area to background area in drawn ROIs, then there will be variability in mean radiance values that is not
biologically based, but is instead due to the way the ROIs were drawn. Importantly, and in contrast, total emission values
remain consistent and unaffected by any such variability in how ROIs are drawn.
Join Spectral Instrument Imaging’s applications expert Andrew Van Praagh, and his guest speakers, for an educational webinar series highlighting best practices for experimental design and more!
Sign up for email updates; we’ll send you tips and tricks for getting better data.