Background: In the past year, we have focused on studying the effect of shed-antigen on the microdistribution of Zr-89-labeled amatuximab in tumor, apart from other factors such as tumor vascular density, interstitial fluid pressure (IFP) and extracellular matrix proteins that could affect the microdistibution of mAb. For this study, we used a nude mouse model implanted with A431/H9 tumor which over-expresses both mesothelin and Lewis-Y on the tumor cell surface. Mesothelin is a membrane glycoprotein of 40 kDa. It is actively shed from the tumor cell surface. Lewis-Y is a carbohydrate antigen and is not actively shed from the tumor surface. As model antibodies, we used two mAbs, anti-mesothelin mAb amatuximab (mouse/human chimeric antibody with 82.6% amino acid sequence identity to a human IgG1 and nM KD binding affinity) and anti-Lewis-Y mAb B3 (murine IgG1 with 10 nM KD binding affinity). Comparative autoradiography studies using the A431/H9 tumor model targeted with Zr-89 labeled-amatuximab and -B3 enabled us to define the effects of shed-Ag on the tumor microdistribution of the Zr-89-mAb, apart from the effects of other factors such as vascular density, high interstitial fluid pressure (IFP) and extracellular protein contents. Objectives: To investigate the effect of the injection dose and the tumor size on the tumor microdistribution of these two mAbs radiolabeled with Zr-89 in nude mice bearing A431/H9. Methods: The mAbs were radiolabeled with Zr-89 using desferrioxamine with an isothiocyanate linker as a chelating agent. The Zr-89 labeled mAbs were then purified with a PD-10 column eluted with 0.25 M ammonium acetate at pH 5.5. The radiolabeled mAbs with the radiochemical purity >95% and the immunoreactivity >70% were used for in vivo studies. For autoradiography studies, the mice with A431/H9 tumor were injected iv with Zr-89-amatuximab (100 micro-Ci/10 or 60 micro-g) or Zr-89-B3 (80 micro-Ci/15 or 60 micro-g). The mice (n=3) were euthanized at 48 h post-injection and the tumors were excised. The tumors were embedded and frozen in Tissue-Tek CRYO-OCT compound (Sakura Finetek USA Inc., Torrance, CA, USA) at -20 degree C for 3 h. Serial 20 micro-m thick short axis sections were cut in 400 micro-m intervals covering the entire tumor. Two or three consecutive tumor slices were selected at 3 tumor regions (25%, 50%, and 75% long axis regions from the tumor surface) as representative sections throughout the tumor and exposed on the phosphor screen for 16 h. Signals were obtained by the use of the Typhoon FLA 7000 (GE Healthcare Life Sciences, Pittsburgh, PA, USA) with 25 micro-m pixel resolution and analyzed with Image Quant TL8.1 software. Values were grouped together from the 3 tumor regions to represent a tumor. Each tumor was treated as an independent sample. To analyze the microdistribution of the radioactivity in the tumor sections, we introduced a normalized length analysis method as described below. The first line was drawn along a longest axis, and the second line was drawn along a short axis perpendicularly at the center of the first longest line. The center was selected as the point where the two lines meet. Additional lines were drawn evenly and continuously between the two original lines passing through the same center point (total 8 lines). Radioactivity profile of each line was analyzed with ImageJ (NIH, Bethesda, MD) and exported into Excel files to redefine values with MATLABs interpolation function interp1. The maximum length of each line in x-axis was normalized to 1 to correct for the differences in the length of each line for reconstruction of the radioactivity-vs-tumor penetration distance profiles of each tumor section. The maximum signal intensity within each tumor section in y-axis was also normalized to 100 to correct for the differences in the signal intensity between each tumor section. Mean radioactivity-vs-distance profiles with standard deviation were then reconstructed for tumor sections obtained at 25%, 50%, and 75% regions. Results: The radioactivity-vs-distance profiles of Zr-89 B3 showed the peak radioactivity distributed in the tumor periphery which was then rapidly decreased as moved away from the periphery toward the tumor core for both 15 and 60 micro-g doses. These radioactivity-vs-distance profiles were also similar to that of Alexa-labeled B3 analyzed by fluorescence microscopy after the injection of 150 micro-g Alexa-B3. These findings are consistent with a notion that the IFP is often elevated in solid tumors but declines in the tumor periphery in the outer 0.21.1 mm. This transvascular pressure gradient may cause a mAb against non-shed tumor antigen to extravasate and accumulate preferentially in the tumor periphery rather than in the tumor core. In contrast, the profiles of Zr-89 amatuximab depended on the injected dose levels; the radioactivity distribution was more uniform with the radioactivity distributed in the tumor core similar to that in the tumor periphery for 10 micro-g injection whereas the profile became similar to those of Zr-89 B3 for 60 micro-g injection, showing that the peak radioactivity at the tumor periphery rapidly decreased as moved away from the tumor periphery toward the tumor core. These findings support a notion that for 10 micro-g dose, Zr-89 amatuximab in the blood circulation mostly bound to the shed-mesothelin in blood and a small amount of free Zr-89 amatuximab which entered into tumor extracellular space (ECS) after crossing the tumor vasculature would mostly exist as an antibody/antigen complex and thus, would be distributed more uniformly throughout the entire tumor by bypassing the binding sites on the surface of tumor cells nearest to the vasculature. In comparison, Zr-89 amatuximab injected at 60 g/dose has an estimated concentration of 260 nM in a blood immediately after injection and 28.3 nM at 48 h (based on 6.8 %ID/g of blood) and therefore in a much larger molar excess concentration compared to the steady state concentration of shed-mesothelin in the blood ( 6 nM). Consequently, it would remain mostly as an unbound free mAb form, thereby escaping from the sequestration into the reticuloendothelial system in liver during a 48 h period. Therefore, a large portion of the injected dose (95.2 nM for the tumor uptake of 22.90 % ID/g tumor at 48 h and 952 nM in the ECS because the ECS is 10% of the total tumor volume) would have crossed the tumor vasculature and diffused into the ECS of tumor. This concentration (952 nM) is similar to the steady state concentration of the shed-Ag in the ECS (800 nM in 400 cubic mm tumor). This finding suggests that one of two Ag binding sites of amatuximab entered into the ECS would have been free to bind the membrane-bound mesothelin on tumor cells in periphery (i.e., in the lower IFP) rather than in the tumor core, similar to that shown for Zr-89 B3. Conclusion: A mAb injection dose, that could generate the mAb concentration at a several fold molar excess to the shed-Ag concentration in blood but generate the mAb concentration at a several fold lower than the shed-Ag concentration in the tumor ECS, would provide a beneficial effect in maximizing tumor uptake with a minimum liver and spleen uptakes while allowing deeper tumor penetration. This study, thus, provides an important message that it is important to find an optimum treatment regimen to enhance the delivery and penetration of mAb into tumor with shed-Ag.
Showing the most recent 10 out of 20 publications
