Abstract
The human pancarcinoma-associated epithelial cell adhesion molecule (EpCAM) (EGP-2, CO17-1A) is a well-known target for carcinoma-directed immunotherapy. Mouse-derived mAbs directed to EpCAM have been used to treat colon carcinoma patients showing well-tolerable toxic side effects but limited antitumor effects. Humanized or fully human anti-EpCAM mAbs may induce stronger antitumor activity, but proved to produce severe pancreatitis upon use in patients. To evaluate treatment-associated effects before a clinical trial, we have generated a transgenic mouse tumor model that expresses human EpCAM similar to carcinoma patients. In this study, we use this model to study the in vivo behavior of two humanized and one mouse-derived anti-EpCAM mAb, i.e., MOC31-hFc, UBS54, and MOC31. The pharmacokinetics and tissue distribution of the fully human mAb UBS54 and the mouse-derived MOC31 were largely the same after injection in tumor-bearing transgenic mice, whereas the molecularly engineered, humanized MOC31-hFc behaved differently. Injection of UBS54 and MOC31 resulted in significant, dose-dependent uptake of mAb in EpCAM-expressing normal and tumor tissues, accompanied by a drop in serum level, whereas injection of MOC31-hFc resulted in uptake in tumor tissue, limited uptake by normal tissues, and slow blood clearance. It is concluded that the EpCAM-transgenic mouse model provides valuable insights into the potential behavior of humanized anti-EpCAM mAbs in patients. mAbs sharing the same epitope and isotype but constructed differently were shown to behave differently in the model, indicating that the design of mAbs is important for eventual success in in vivo application.
Since its discovery (1), the human pancarcinoma-associated epithelial cell adhesion molecule (EpCAM4; CD326), also referred to as EGP-2, 17-1A, or KSA, has become a major target for carcinoma diagnosis and therapy (2). EpCAM is a 38-kDa transmembrane glycoprotein, which has been characterized as a homotypic adhesion protein. It is commonly expressed on the basolateral cell surface of simple, transitional, and pseudostratified epithelia (3), as well as intensively and uniformly on a large variety of carcinomas (4, 5). Mouse-derived mAbs directed to EpCAM have been successfully used in patients for imaging, for instance, in small cell lung cancer diagnosis (6) and also for adjuvant treatment of minimal residual disease of colon carcinoma (2, 7). Initially, the latter approach appeared to improve long-term survival of patients, but a larger study could not corroborate these promising therapeutic effects (8). Earlier efforts to enhance the potential of EpCAM-directed immunotherapy have been disappointing, because these did not lead to any therapeutic effect (9). The recent clinical success of engineered humanized mAbs in the treatment of a variety of diseases including B cell lymphomas and breast cancer has renewed interest in the use of humanized mAbs for targeting carcinomas. Humanized or fully human Abs directed against EpCAM are available and have been used in a pilot clinical study (10, 11, 12). The observed severe pancreatic toxicity (12) calls for careful evaluation of this powerful approach in a relevant preclinical model, however.
The relevance of a preclinical model for targeting of EpCAM greatly depends on EpCAM expression on normal epithelial tissues (3). This should be similar to the situation in humans. For EpCAM, it is known that the overall distribution of mouse or rat EpCAM is not the same as for humans (13, 14). Therefore, results obtained in nontransgenic mice or rats using mouse or rat EpCAM as a target may not be of direct relevance for humans. To evaluate treatment-associated effects of EpCAM-directed immunotherapy on tumor and normal tissue, we have generated EpCAM-expressing transgenic rat and mouse tumor models in which we previously analyzed the behavior of a mouse mAb directed against EpCAM (15, 16).
The aim of the present study was to compare the pharmacokinetics and normal- and tumor-tissue targeting of a humanized, a human, and a mouse mAb directed against EpCAM, i.e., [scFv-MOC31]2-hFc, UBS54, and MOC31, respectively, after injection in our previously established EpCAM-transgenic mouse tumor model.
Materials and Methods
Animals used
EpCAM-transgenic FVB/N mice described previously (16) were crossed with nontransgenic C57BL/6 (Harlan) mice yielding EpCAM-expressing mice with an FVB/N-C57BL/6 genetic background allowing tumor take. Nontransgenic littermates were used as control animals. Mice were housed according to the rules for transgenic housing following the guidelines of the Dutch law, and they were fed ad libitum.
Cell lines
The murine B16.F10 melanoma cell line was obtained from the American Type Culture Collection (CRL-6475). The EpCAM cDNA-transfected B16 melanoma cell line, B16.C125 was provided by Dr. M. Dohlsten (University of Lund, Lund, Sweden), and both were cultured as described (17).
Antibodies
The murine anti-EpCAM mAb MOC31, an IgG1 Ab with a calculated molecular mass of 132 kDa, was purified from hybridoma supernatant as described previously (6). The chimeric Ab [scFv-MOC31]2-hFc (size, 138 kDa) consists of two scFv parts carrying the MOC31 variable region coupled to a human IgG1 Fc tail and was provided by Dr. W. Helfrich (Groningen, The Netherlands) and will be referred to throughout the text as MOC31-hFc (18). The UBS54 Ab, provided by Dr. A. Kruisbeek (Crucell, Leiden, The Netherlands), is a completely human mAb with a calculated molecular mass of 150 kDa and was selected from a large synthetic phage library as described (19). Trastuzumab (Herceptin; Roche) is a humanized IgG1 mAb with a calculated molecular mass of 148 kDa directed against the human epidermal growth factor receptor 2.
Radiolabeling
mAbs were radiolabeled with 125I according to the standard Iodogen method and purified using gel filtration chromatography. Radiochemical purity was determined by TCA precipitation and size exclusion chromatography HPLC and was always >97%.
Immunoreactivity
The immunoreactive fraction of radiolabeled mAb was determined by cell binding assays essentially as described by Lindmo et al. (20). The B16.F10 (EpCAM negative) and B16.C125 (EpCAM positive) melanoma cell lines were used. The immunoreactive fraction, calculated from the ratio of the total applied (sum of radioactivity in supernatants plus cell-bound) to cell-bound radioactivity plotted against the inverse cell concentration was always >0.7 upon incubation with the EpCAM-expressing cells.
Biodistribution study of the mAb UBS54 as visualized by immunohistochemistry
Five groups of three EpCAM-transgenic FVB/N-C57BL/6 mice per group were included in the experiment. A total of 2 × 105 B16.C125 cells was inoculated s.c. in the right flank, whereas a similar amount of B16.F10 cells was injected s.c. in the left flank. When palpable tumors were noticed, UBS54 mAb was applied i.p. to the animals of the various groups in doses ranging from 0, 1, 10, 100, to 1000 μg, in 200 μl of PBS. Essential organs and tissues (see Table I⇓) were isolated 48 h after injection, and immunohistochemistry was performed on 5-μm-thick, air-dried cryosections made from snap-frozen tissue samples.
Targeting of UBS54 to organ systems in EpCAM-transgenic mice is critically dependent on the application dosage
First, overall expression of EpCAM in the EpCAM-transgenic mice was confirmed using UBS54 (5 μg/ml). Biotinylated goat anti-human IgG Abs (2 μg/ml; Dakopatts) served as secondary Ab and were made visible via peroxidase-conjugated streptavidin (Dakopatts). Before the complete staining procedure, endogenous biotin was blocked (16).
Second, in vivo localized UBS54 was detected by directly applying biotinylated goat anti-human IgG Ab on the cryosections, followed by the same staining procedure as described above.
Biodistribution studies with 125I-labeled MOC31, UBS54, and MOC31-hFc
In vivo studies were performed with 5- to 9-wk-old male and female EpCAM-transgenic and nontransgenic (FVB/N-C57BL/6)F1 hybrid mice. By injecting 1 × 106 B16.F10 and B16.C215 cells s.c. in the left and right flanks, respectively, tumors were induced and subsequent tumor growth was monitored every second day. Radiolabeled mAbs were injected i.p. at doses of 10 and 100 μg (0.4–1.0 MBq) in three transgenic and three nontransgenic mice each. Injection took place 11 days after tumor induction (see Ref.16) irrespective of tumor size; however, tumor size was always <3 cm3. As an additional control, three transgenic and three nontransgenic animals were injected i.p. with an equal volume of PBS. Essential organs as listed in Table I⇑ were isolated 48 h after injection, collected carefully, and analyzed for radioactivity in a gamma counter (LKB-1282-CompuGamma). Data are expressed as percentage of injected dose radioactivity per gram tissue (% I.D./g tissue).
Statistics
For statistical analysis of the radioactive biodistribution results, the two-sided Student’s t test was applied using the GraphPad software (GraphPad). A value of p < 0.05 was regarded as significant, whereas a value of p < 0.001 was regarded as highly significant.
Results
Biodistribution of UBS54 in EpCAM-transgenic, tumor-bearing mice
EpCAM-transgenic mice carrying palpable B16.F10 as well as B16.C215 tumors were used to study the effect of EpCAM present on both normal and tumor tissue on the biodistribution of injected UBS54. As shown in Fig. 1⇓, injection of 1000 μg of UBS54 resulted in a substantial uptake of mAb in tissues like pancreas (ductal as well as acinar epithelial cells), jejunum (epithelial cells of villi and crypts), colon (epithelial cells of crypts), kidney (epithelial cells of collecting ducts, Henle’s loops, and distal tubulus), and salivary glands, whereas minor uptake could be detected in lung (epithelial cells of bronchi and bronchioli), liver (epithelial cells of bile ducts), stomach (center as well as epithelial cells), skin (hair follicles), and eye (processus ciliaris). When lower doses of USB54 (100 or 10 μg) were applied, considerably less in vivo bound mAb located in these organs, whereas no UBS54 mAb could be detected after injection of 1 μg or PBS. Fig. 2⇓ shows the localization of UBS54 in tumor tissue. UBS54 proves to specifically localize to the periphery of EpCAM-positive (Fig. 2⇓e) but not EpCAM-negative (Fig. 2⇓b) tumors. Apparently, the penetration of the Ab in the EpCAM-positive tumor was not very effective, because no mAb could be detected in the center of the tumor (Fig. 2⇓f).
Localization of the fully human anti-EpCAM-directed UBS54 Ab applied i.p. in EpCAM-transgenic FVB/N-C57BL/6 hybrid mice. UBS54 localization was determined immunohistochemically 48 h after injection by staining with an anti-human IgG peroxidase conjugate. UBS54 localized on the EpCAM-expressing epithelial cell lining of the kidney (a), uterus (c), bladder (e), pancreas (f), colon (h), lung (j), liver (k), and salivary glands (o), whereas no UBS54 was observed in EpCAM-negative tissues, such as, for instance, heart (m). No specific staining was observed after incubation of similar EpCAM-expressing tissue sections obtained from transgenic mice injected with PBS: kidney (b), uterus (d), pancreas (g), colon (i), liver (l), and on slides of EpCAM-negative heart tissue (n).
Localization of the fully human anti-EpCAM-directed UBS54 in the EpCAM-negative and -positive tumors B16.F10 and B16.C215, respectively, after i.p. injection of 1 mg in tumor-bearing EpCAM-transgenic and nontransgenic FVB/N-C57BL/6 hybrid mice. UBS54 localization was determined immunohistochemically 48 h after injection by staining tissue slides of the isolated tumors with anti-IgG conjugate. No UBS54 could be detected on the slides obtained from the EpCAM-negative mouse melanoma tumor B16.F10 either by direct staining (a) or by staining with the conjugate after injection with the UBS54 Ab (b), or with the conjugate after injection with PBS (×20) (c). UBS54 was detected on slides obtained from the EpCAM-expressing mouse melanoma tumor B16.C215 either by direct staining (×40) (d), or by staining with the conjugate after injection with the UBS54 Ab at the periphery of the tumor (e), whereas no specific staining could be detected at the center of the tumor (f) or after injection with PBS (×20) (g).
The localization of UBS54 to EpCAM-positive tissue in EpCAM-transgenic mice was even more pronounced when the Ab was radiolabeled and analyzed. With all injected doses of 125I-labeled UBS54, uptake of radioactivity could be detected in pancreas, ileum, colon, kidney, and bladder. At the 100-μg dose, a significant localization of the Ab at the tumor was observed (Table I⇑).
In vivo biodistribution studies of radiolabeled mAbs
Radiolabeled anti-EpCAM mAbs UBS54, MOC31, MOC31-hFc, and the control, i.e., not EpCAM but human epidermal growth factor receptor 2 directed, trastuzumab were used to study the biodistribution pattern of anti-EpCAM mAbs in EpCAM-transgenic mice in more detail. Fig. 3⇓ depicts the results obtained after injection of 10 μg of mAb, but similar results were obtained with 100 μg (results not shown). For trastuzumab, it is clear that, for all investigated organs, the percentage of injected radioactive dose per gram organ stays below the radioactivity level found in the blood compartment, indicating that this control human IgG1 Ab does not localize specifically to any mouse tissue, not in the EpCAM-transgenic nor in the nontransgenic mice. For both the murine MOC31 mAb as well as the humanized mAb UBS54, a significant rapid clearance of radioactive signal from the blood was noticed in the EpCAM-transgenic mice compared with the nontransgenic mice (p < 0.01). The clearance of these radiolabeled mAbs was paralleled by localization of these Abs in, e.g., the EpCAM-expressing pancreas and kidney of the transgenic animals (p < 0.05) and in the EpCAM-positive tumor (p < 0.05) when comparing B16.C215 tumors with B16.F10. Whereas no significant differences in localization of the radiolabeled MOC31 and UBS54 mAbs could be found in other EpCAM-expressing organs in the transgenic mice compared with the nontransgenic animals for the 10 μg dose, significant differences (p < 0.01) were found for the 100-μg dose of UBS54 between the ileum and salivary glands of the EpCAM-transgenic mice and nontransgenic mice (data not shown). The differences in radioactivity observed when comparing organs of the nontransgenic mice with those of the transgenic mice are caused by the presence of a higher blood level activity in the nontransgenic mice compared with the transgenic mice. However, the radioactive signal measured in these organs does not exceed the blood level radioactivity determined in the nontransgenic mice; therefore, this radioactive signal is probably caused by the blood present in these organs and not by specific localization.
Biodistribution of 125I-labeled mAbs in several tissues and organs in EpCAM-transgenic and nontransgenic FVB/N-C57BL/6 hybrid mice 48 h after injection. Biodistribution is expressed as percentage of injected dose radioactivity per gram tissue (% I.D./g tissue). Bars, Mean (±SEM) of three animals. ∗, ∗∗, and ∗∗∗, Statistically significant difference of p < 0.05, p < 0.01, and p < 0.001, respectively, between transgenic mice compared with nontransgenic mice.
A completely different biodistribution pattern was observed when the human-engineered mAb MOC31-hFc was injected. No differences in blood level radioactivity were found between blood obtained from the transgenic mice compared with blood obtained from the nontransgenic mice. And although the significant differences in radioactivity could be found between the transgenic and nontransgenic mice for the pancreas, kidney, B16.C215 tumors, and even additional EpCAM-positive organs at the 100-μg dose, these values did not differ significantly with the level found in blood, except for the kidney at a dose of 10 μg (p < 0.05).
Detailed comparison of the in vivo behavior of humanized Abs
Comparing the biodistribution of the three EpCAM-directed mAbs, the most striking difference observed was the difference in radioactive signal in the blood compartment of the transgenic mice. The observed %I.D./g blood for the MOC31-hFc mAb was considerably higher (mean, 8.63 for the 10-μg dose; and 6.77 for the 100-μg dose) as observed for the UBS54 mAb (mean, 2.32 and 2.92, respectively) and the MOC31 mAb (mean, 1.03 and 1.45, respectively). Moreover the circulating time of MOC31-hFc in blood was either identical or longer in the transgenic mice (mean, 8.63 for the 10-μg dose, and 6.77 for the 100-μg dose) compared with the nontransgenic mice (mean, 7.80 and 3.66, respectively), in contrast to the circulating time of the UBS54 and MOC31 mAbs. To be able to compare the biodistribution of the different humanized IgG1 mAbs in EpCAM-transgenic and nontransgenic mice, the radioactive signal obtained at a dose of 100 μg of mAb was normalized for blood radioactivity and plotted against each other in Fig. 4⇓. In the nontransgenic mice, significant localization of UBS54 was observed in the ileum (p < 0.01), whereas significant localization of MOC31-hFc in these animals could be found in pancreas, kidney, uterus/testis, B16.C215, and B16.F10 (all p < 0.05) compared with the IgG1 control trastuzumab. However, no significant differences could be found when UBS54 and MOC31-hFc biodistribution patterns in the nontransgenic mice were compared with each other. In the EpCAM-transgenic mice, however, significant differences in localization between UBS54 and MOC31-hFc were found for the lung (UBS54, p < 0.01), the kidney (UBS54, p < 0.05), and the salivary glands (UBS54, p < 0.01). In comparison with trastuzumab, UBS54 localized significantly to the lung (p < 0.001), the pancreas (p < 0.05), the ileum (p < 0.01), the kidney (p < 0.001), the salivary glands (p < 0.001), and B16.C215 tumor (p < 0.05) in these EpCAM-transgenic animals. In contrast, significant localization of radioactive MOC31-hFc compared with trastuzumab could only be detected in the pancreas and B16.F10 tumor (both p < 0.05) of these mice.
Biodistribution of 125I-labeled mAbs in most relevant tissues and organs in EpCAM-transgenic FVB/N-V57BL/6 hybrid mice (upper graph) and in nontransgenic FVB/N-C57BL/6 hybrid mice (lower graph) 48 h after injection. Biodistribution is expressed as tissue-to-blood ratio per gram tissue. Bars, Mean (±SEM) of three animals. ∗, ∗∗, and ∗∗∗, Statistically significant differences between 125I-labeled UBS54 or MOC31-hFc compared with trastuzumab. *, **, and ***, Statistically significant differences in tissue localization of p < 0.05, p < 0.01, and p < 0.001, respectively, between 125I-labeled UBS54 and MOC31-hFc.
Tumor uptake
To study the accumulation of radiolabeled EpCAM-directed mAbs in EpCAM-positive B16.C215 and EpCAM-negative B16.F10 tumors, the radioactive signal was corrected for the blood value and the tumor-to-blood ratios are presented as mean ± SEM in Table II⇓ and Fig. 5⇓. In the nontransgenic mice, only MOC31-hFc showed significant tumor uptake in the EpCAM-positive B16.C215 tumor compared with trastuzumab, and this was seen for the 10-μg dose (1.05 ± 0.26; p < 0.05) as well as for the 100-μg dose (2.32 ± 0.61; p < 0.05). However, significant MOC31-hFc uptake in comparison with trastuzumab was also noticed in the B16.F10 tumor for the 100-μg dose (0.99 ± 0.21; p < 0.05) in these animals. No significant differences in localization could be observed between UBS54 and MOC31-hFc, whereas significant differences were observed between MOC31-hFc and MOC31 in the B16.C215 tumor for the 100-μg dose (p < 0.05) and B16.F10 tumor for both doses (p < 0.01 and p < 0.05, respectively) in the nontransgenic animals.
Tumor localization of 125I-labeled mAbs (100 μg) in EpCAM-transgenic FVB/N-C57BL/6 hybrid mice (left panel) and in nontransgenic FVB/N-C57BL/6 hybrid mice (right panel). Biodistribution is expressed as tumor-to-blood ratio. Bars represent mean (±SEM) of three animals. ∗, Statistically significant differences in tumor localization of p < 0.05 between 125I-labeled UBS54, MOC31-hFc, or MOC31 compared with trastuzumab. *, Statistically significant differences in tissue localization of p < 0.05 between 125I-labeled, UBS54, MOC31-hFc, or MOC31.
Accumulation of radiolabeled anti-EpCAM directed mAbs in EpCAM-positive B16.C215 and EpCAM-negative B16.F10 tumors in percentages of the injected doses corrected for the blood value via tumor to blood ratios presented in the mean ± SEM
In the transgenic mice, significant differences in the tumor uptake could be detected between trastuzumab and the 100-μg dose of UBS54 (1.16 ± 0.15; p < 0.05) and the 10-μg dose of MOC31 (3.50 ± 0.66; p < 0.05) for the B16.C215 tumor, whereas no significant tumor uptake for any of the Abs tested could be detected in the EpCAM-transgenic mice for the B16.F10 tumor. Furthermore, significant differences in uptake were observed between the 10-μg dose of MOC31 and the 10-μg dose of MOC31-hFc and UBS54 (both p < 0.05) for the B16.C215 tumor.
Discussion
In this study, the biodistribution pattern and targeting characteristics of a humanized, a human, and a mouse mAb directed against EpCAM (MOC31-hFc, UBS54, and MOC31) were analyzed in our EpCAM-transgenic mice tumor model (16). In these mice, the fully human UBS54 mAb localized almost to all EpCAM-expressing tissues at a dose of 1000 μg (∼40 mg/kg) and 100 μg (∼4 mg/kg) i.p. This is consistent with a reported reversible pancreatitis observed when a similar dose of the human-engineered anti-EpCAM mAb ING-1 was supplied to patients with advanced adenocarcinomas (12). At a dose of 10 μg (∼0.4 mg/kg), the UBS54 mAb could only be detected faintly in the pancreas and kidney of the EpCAM-transgenic mice. This dose corresponds with the maximally tolerated dose of 0.3 mg/kg as reported for the same phase I trial. At a dose of 1 μg (∼0.04 mg/kg), no localization of UBS54 was observed in any of the EpCAM-expressing tissues in the study presented here. At this dose, one patient showed stable disease of at least 12 wk demonstrating that this dose might already have a therapeutic effect (12).
UBS54 is a fully human mAb constructed from a semisynthetic phage library of human scFv fragments (19, 21). The poor performance of anti-EpCAM mAbs of nonhuman origin in clinical trials caused either by their immunogenicity, by their poor pharmacokinetic properties, or by their inefficient recruitment of effector cells has put the focus on the generation of EpCAM-directed humanized mAbs. Next to the fully humanized UBS54 and the human-engineered ING-1, the fully human recombinant Ab MT201, the chimeric 17-1A, and the K931 anti-EpCAM have been described (11, 19, 21, 22, 23, 24). All of these Abs have a human Fc domain, however, of different origin. To establish the influence of the Fc region, we compared the biodistribution of the fully human UBS54 Ab with the biodistribution of our own human-engineered Ab MOC31-hFc and its murine equivalent MOC31 in both EpCAM-positive and -negative tumor-bearing transgenic and nontransgenic mice.
In this context, the irrelevant mAb trastuzumab did not localize specifically to any tissue in the EpCAM-transgenic mice. This demonstrates that the human IgG1 Fc domain does not independently mediate localization of the mAbs to EpCAM-expressing tissues. Injection of UBS54 as well as MOC31 resulted in a significant, dose-dependent uptake of mAb in EpCAM-expressing normal and tumor tissues, accompanied by a drop in serum level. In contrast, although limited uptake of MOC31-hFc by normal tissue was observed, the serum level radioactivity in the EpCAM-transgenic mice remained similar to the serum level radioactivity found in nontransgenic littermates up to 48 h after injection. This indicates that MOC31-hFc might be a better therapeutical humanized mAb, because in general a long circulation time promotes uptake of Ab at the site of the tumor.
The localization of MOC31 at EpCAM-expressing tissues in the transgenic mice was unexpected because no localization of biotinylated MOC31 at EpCAM-expressing normal tissue was noticed in a previous study 24 h after i.v. injection of MOC31 at a dose of 1 mg/kg (20). Differences in threshold of the detection method used or differences in the time interval between injection and analysis can account for the discrepancy noticed between these two studies. However, localization of radiolabeled MOC31 in EpCAM-expressing tissues other than tumor has not been reported in humans as well (6). Differences between the human and mouse FcRn receptor might account for the latter discrepancy. The FcRn receptor is expressed in endothelial cells of small arterioles and capillaries in muscle and liver and in a variety of other tissues (25) in both mice (26) and humans (27). However, human FcRn can bind human, rabbit, and guinea pig IgG, but not rat, bovine, sheep, or mouse IgG. In contrast, mouse FcRn binds to IgG of all species (28). Therefore, the human FcRn receptor cannot bind the Fc domain of MOC31, whereas in the EpCAM-transgenic mice, the mouse FcRn receptor can bind to the Fc domain of MOC31 enabling the MOC31 Ab to extravasate from the blood vessels into the tissue where it subsequently can localize at EpCAM-expressing normal tissue.
In nontransgenic mice, no significant differences in behavior between UBS54 and MOC31-hFc could be detected. In the EpCAM-transgenic animals, large differences in localization between the two humanized anti-EpCAM Abs were observed, e.g., UBS54 was present in lung, pancreatic, ileum, kidney, salivary gland, and EpCAM-positive tumor tissue, whereas MOC31-hFc could only be detected in pancreatic tissue. The differences in Ab uptake by normal tissue might explain the observed differences in blood level.
The EpCAM-transgenic mice express the EpCAM molecule similar to humans; however, not all organs expressing EpCAM showed uptake of the EpCAM-directed mAbs. This indicates that not all tissues are equally accessible for these Abs. The accessibility of an Ag in organs can be influenced by differences in vasculature between organs or by differences in FcRn affinity of the mAb, which results in differences in transcellular transport subsequently leading to either more or less Ab being available for the target Ag (25, 29). The Ag density per organ or the presence of the basal lamina, which has been suggested to shield the epithelial cells from the circulation in humans (3, 30), can influence the tissue localization of the mAb as well.
Differences in affinity for the Ag or in penetration capability can also cause the differences observed between the two humanized anti-EpCAM Abs. Ag affinity of a mAb is mainly determined by the rate with which the mAb detaches from the Ag, given by the off-rate of the molecule (Koff). The Koff of UBS54 is 5 × 10−4 s−1, and the Koff of the two MOC31 scFvs together is 6.8 × 10−4 s−1, accounting for a half-life of ∼1 h for both Abs. The off-rate of MOC31 is even lower, 0.5 × 10−4 s−1, resulting in a half-life of ∼4 h (31). So the observed differences in behavior of the two humanized Abs cannot be explained by differences in affinity for the substrate. However, differences in Ag affinity can account for differences observed between MOC31 and its derivative MOC31-hFc. Both Ag affinity and size of a mAb can influence the penetration capability of a mAb. High Ag affinity leads to entrapment of the Ab at the tumor edge and slows down penetration, whereas smaller Abs, e.g., F(ab′)2 fragments, scFvs, or diabodies, are believed to penetrate tissue faster. Although UBS54 is bigger than MOC31-hFc, the difference in size cannot account for the differences in in vivo behavior because MOC31-hFc shows less penetration.
The not very pronounced uptake of the EpCAM-directed mAbs in the EpCAM-positive tumors compared with the EpCAM-negative indicates a poor accessibility of EpCAM on the tumor tissue. Previous Western blot analysis of B16.C215 tumor tissue revealed a strong EpCAM expression up to 3 wk after inoculation (16); however, immunohistochemical stainings seem to be hindered by the large amount of melanin found in this particular strain of the B16 melanoma cell line. Consequently, it seems likely that the mAbs studied here also experience hinder from the presence of melanin in the in vivo situation. Furthermore, both the B16.C215 as well as the B16.F10 tumor tissues are very soft and well supplied by blood, resulting in a relatively narrow region for detection of specific radioactive signals caused by specific localization in the EpCAM-positive tumor in the background of radioactive mAb present in the blood.
The strong correlation of the results obtained with the fully humanized UBS54 Ab in the EpCAM-transgenic mice and the phase I results obtained with the ING-1 human-engineered mAb in patients, demonstrates that the EpCAM-transgenic mouse model described here is a very suitable model for evaluating immunotherapeutic strategies using EpCAM as target before clinical usage. In this model, it was demonstrated that differently generated, but isotype-identical humanized Abs harboring the same Ag specificity can behave very differently regarding localization at tumor tissue and potential hazardous localization at normal tissue. Because both Abs target the same epitope with resembling affinity and have the same Fc domain isotype, it seems that the various ways of constructing humanized Abs might lead to differences in Ab behavior. This difference is probably mediated by differences in affinity for the FcRn receptor, which is known to play an important role in mAb clearance. The isotype control chimeric-Ab trastuzumab constructed in a similar manner, because MOC31-hFc did not show any preferential localization, indicating that this way of constructing might lead to less affinity for FcRs in general. To our knowledge, this is the first study of behavior of humanized Abs directly compared in a relevant mouse model that shows that, in the design of humanized mAbs, great care should be taken in the choice of method to construct the humanized Fc domain.
Disclosures
The authors have no financial conflict of interest.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This work was supported in part by the Dutch Society of Cancer and the J. K. de Cock Foundation.
↵2 Address correspondence and reprint requests to Dr. Jos G. W. Kosterink, Department of Hospital and Clinical Pharmacy, University Medical Center Groningen, P.O. Box 30.001, NL-9700 RB Groningen, The Netherlands. E-mail address: j.g.w.kosterink{at}apoth.umcg.nl
↵3 J.G.W.K. and P.M.J.M. contributed equally to this work.
↵4 Abbreviation used in this paper: EpCAM, epithelial cell adhesion molecule.
- Received June 1, 2005.
- Accepted April 27, 2007.
- Copyright © 2007 by The American Association of Immunologists
References
In this issue
