During the last decade, many preclinical studies have been evaluating new near-infrared fluorescent probes for targeted molecular imaging, and the trend is to slowly move towards the translation into clinical practice. However, before this can be done successfully, a thorough and precise evaluation of the probe (or probes) is necessary at the preclinical level. In this context, the precise quantification of probes in tissues is a matter of concern in biodistribution studies. The approach most commonly used for assessment of tissue distribution of a probe consists of imaging of organs or sections of organs with an optical imager, only allowing for a relative or qualitative assessment. This makes comparison between different studies and different probes rather difficult, since no % ID/g value is possible to be obtained in this manner. This fact is related to the nature of fluorophores: their close proximity can result in quenching of the fluorescence, leading to an underestimation of the amount of probe in tissues. Also, the scattering of photons by tissue components can compromise the fluorescence detected, and moreover, the linear range of fluorescence detection is restricted.
Here, we describe a new method for the quantification of IRDye800CW fluorescent probes in tissues. This method circumvents the issues mentioned above by diluting the lysate of the homogenized organs in order to infer in the linear range of fluorescence (Figure 2). To validate our new method, we have compared it with the most commonly used method for biodistribution studies, i.e., gamma ray quantification of radiolabeled probes. To do so, and here as a research tool, a dual-labeled probe was prepared, minimizing the possibility that the tissue distribution of the probe would be affected by the different labels. The monoclonal antibody cetuximab was selected for this study as it was previously shown that certain conditions of dual labeling do not affect its biodistribution . This could not be the case for smaller molecules or smaller antibody fragments due to the reduced size. Hence, for the biodistribution study, cetuximab was conjugated to the IRDye800CW fluorophore (IR) and to the positron emitter 89-zirconium (89Zr), forming the dual-labeled 89Zr-cetuximab-IR, i.e., the biodistribution probe. To further minimize possible variations, each organ and tumor that was collected for the biodistribution study was halved so that one half could be processed for IR fluorescence quantification and the other half for gamma ray quantification. Nevertheless, one has to bear in mind the possibility of heterogeneous uptake in each tumor and organ, which could lead to differences in uptake levels between the fluorescence and the radioactivity method.
Importantly, the results obtained with the two different methods are very similar for tumors and organs such as the liver, lung, stomach, skin, and intestines (Figure 3). Significant differences were obtained for the blood, heart, spleen, sternum, and muscle, showing higher values by gamma ray quantification, and the kidney which gave a higher value with IR fluorescence quantification. As control studies have shown the stability of the probe in serum, and as it is known that both 89Zr and IR residualize after receptor-mediated internalization of cetuximab ( and Dr. Mike Olive, personal communication), these variations are most likely related to what happens to the probe after liver catabolism. In fact, degradation of 89Zr-cetuximab-IR into small peptides is to be expected, and the fate of these fragments after excretion into the bile may vary, for instance, some of these might be more effectively reabsorbed into the bloodstream at the intestines and subsequently distribute differently throughout the tissues. Even though this is unclear, the trend suggested by the results presented here is that IR fragments are more efficiently eliminated by kidney filtration, and 89Zr-fragments accumulate in bone tissues like sternum, which is in agreement with other studies [12, 20–22].
It is worth realizing that true values of % ID/g are unknown and that two groups of values have been obtained, corresponding to the quantification of IR fluorescence and gamma rays in the tissues. The aim of this study was to determine whether comparable information could be obtained with both quantification methods. In this context, Bland-Altman plots are used to compare the two sets of data and to determine how much each data set differs from the mean value of the two sets of data. Our results show a good average agreement between the data sets. Although the 89Zr measurements are, on average, slightly higher, this small average difference does not depend on the level of % ID/g itself and may thus be easily solved by a calibration factor. The disagreement between the two methods increases with higher values of % ID/g, e.g., in the spleen, tumor, and liver (Figure 4). This observation could simply be explained by a larger intra-organ variation in the case of the spleen, liver, and the tumor. It also shows that for between-group comparisons (e.g., a comparison between the uptake of two probes), more subjects will probably be needed to be able to detect a certain absolute average difference between those groups, when the uptake is, on average, high compared to low.
Overall, the results obtained through IR fluorescence quantification are considered to be representative of the results obtained with the reference method employed for radiolabeled probes. This was also suggested initially by the images obtained with the optical imager and the PET scanner (Figure 1). In that part of the study, a slightly different probe was employed (i.e., 89Zr-cetuximab + cetuximab-IR, the imaging probe), but in fact, no differences were to be expected concerning the biodistribution of these probes, as it has been previously investigated that the conditions here employed for coupling of IR or 89Zr are inert to cetuximab .
The optical images presented (Figure 1a) show a relatively weaker signal at the tumors and livers compared to the PET images (Figure 1b), but this is mostly related to the modality employed. The optical imager employed does not allow 3D collection of data, and fluorescence is only detected at the surface, whereas the PET scanner allowed for 3D collection of data. Recently, newer optical imagers have been developed employing the so-called fluorescence molecular tomography that is suggested to be able to quantify proteins or probes deep in tissues . Nevertheless, background concentrations and tumor-to-normal tissue ratios have been reported as limiting factors , thus leaving room for the new method here described to be used for accurate quantification of IR fluorescent probes. In fact, this method will most likely be applicable to other near-infrared fluorescent probes, although residualization of the fluorophore should be confirmed as well as the stability of the probe in serum, being therefore advisable to be careful in generalizing this method to other fluorescent probes.
Even though the method for IR quantification is more laborious than gamma ray quantification, this new method enables accurate quantification of the probe in % ID/g without the use of radionuclides. This fact is indeed relevant when the probe under development is meant for optical imaging and not for PET or SPECT imaging. We have successfully applied this new method for the quantification of IR fluorescent anti-EGFR nanobodies or VHHs in tissues . In this study, we were able to demonstrate that the small 15 kDa fluorescent nanobodies accumulate more rapidly and to a greater extent in A431 xenografts than the 150 kDa monoclonal antibody cetuximab.