PDAC Panc-1 and BxPC-3 and CRC LoVo, HT-29, HCT116, and SW620 cell lines were from American Type Cell Culture (Manassas, VA). Additional PDAC cell lines FG and L3.6pl were provided by Drs. A.M. Lowy (University of California at San Diego, San Diego, CA) and G.E. Gallick (University of Texas M.D. Anderson Cancer Center, Houston, TX), respectively. PDAC stem-like cells expressing CD24, CD44, and epithelial specific antigen (ESA), designated as PSC^{+ 24/44/ESA}, were prepared from spheroid cells derived from BxPC-3, FG, and L3.6pl cells as previously described [38, 39]. NIH3T3 cells expressing human, monkey, or mouse RON were generated by transfection of pcDNA3 containing individual RON cDNAs as previously described [8]. Primary PDAC cell lines AMCPAC02, AMCPAC04, SNU2491, and SNU410 established from patient-derived tumors were used as previously described [40]. Mouse anti-RON mAb Zt/g4, Zt/f2 and rabbit IgG antibody against the RON C-terminus were used as previously described [8, 34]. Female athymic nude mice at 6 weeks of age were from Taconic Biosciences (Granbury, NJ). The use of mice was approved by the Texas Tech University institutional animal care committee. Female cynomolgus monkeys aged at 3.6–4 years with an average bodyweight of 2.6 ± 0.25 kg/animal were from Guidong Quadrumana Development & Laboratory (Guangxi, China). The use of animal was approved by the Zhejiang University School of Medicine institutional review committee according to the Chinese Food &Drug administration guidelines.

Levels of RON expression by PDAC cells was determined using Z/tg4 in flow cytometric analysis as previously described [30]. H-Zt/g4-induced RON internalization by PDAC cells was determined in immunofluorescence analysis as previously described [31]. PDAC cell viability and cell death after H-Zt/g4-MMAE treatment was determined by the MTS assay and the trypan blue exclusion assay as previously described [30, 31].

Lysates from xenograft tumors were prepared in a lysis buffer as previously described [8]. Cellular proteins (20 μg per sample) were separated in an 8% SDS-PAGE under reduced conditions. RON was detected using the rabbit anti-RON IgG antibody [8] with enhanced chemiluminescent reagents and analyzed in Bio-Rad Versa-Doc 5000 Image system. Membranes also were reprobed with antibody specific to actin to ensure equal sample loading.

Antibody humanization was performed by grafting the sequences from complementarity-determining regions (CDRs) of Zt/g4 into human IgG1/κ acceptor frameworks to generate five light and five heavy chains to create 25 different parings of H-Zt/g4 IgG1/κ molecules [41]. The subclone H1L3 was selected as the lead candidate. DM1, MMAE, and duocarmycin were from Concortis (www.concortis.com) and used for conjugation according to the manufacturer’s instruction. H-Zt/g4-MMAE with DAR of 3.77:1 was selected as the lead candidate. Zt/g4-DM1, H-Zt/g4-DM1, and Zt/g4-MMAE also were prepared and used for comparison. Control mouse IgG conjugated with MMAE (CmIgG-MMAE) at a DAR of 4.1:1 served as the control. All ADCs were purified, sterilized through a filter, and verified by hydrophobic interaction chromatography (HIC) [31].

H-Zt/g4-MMAE at 10 μg per ml was incubated in human fresh plasma in 1 ml plasma at 37 °C up to 20 days. Samples were collected at different time intervals. H-Zt/g4-MMAE in the plasma of cynomolgus monkey (three animals per group) was collected at different time intervals after a single injection of H-Zt/g4-MMAE at 10 or 30 mg/kg. Free MMAE was determined by using the LC-MS/MS method [42] with slight modifications.

Female nude mice (five mice per group, with or without PDAC xenograft tumors) were injected with a single dose of H-Zt/g4-MMAE at 3, 10, and 20 mg/kg through tail vein. Cynomolgus monkeys (three animals per group) were administered through saphenous vein with a single dose of H-Zt/g4-MMAE at 10 and 30 mg/kg. Blood samples were collected from individual mice or monkeys at different time intervals. The amount of MMAE conjugated H-Zt/g4 in plasma was determined by using a MMAE ADC ELISA kit (Eagle Biosciences Inc., Nashua, NH). The PK parameters were calculated using the WinNonLin software package (Certara, Princeton, NJ) as previously described [30].

Female athymic nude mice were injected with 5 × 10⁶ PDAC cells in 0.1 ml PBS into the subcutaneous space of the right flank of mice as previously described [30, 31]. Xenograft tumors mediated by CRC cells served for comparison. For PDAC stem-like cell-mediated xenograft tumors, individual mice were injected with 5 × 10⁵ PSC^{+ 24/44/ESA} cells in 0.1 ml PBS as described above. For PDX models, primary PDAC cells from patient-derived tumors at 5 × 10⁶ cells in 0.1 ml PBS were used. Mice were randomized into different groups (five mice per group). Treatment began when all tumors had a mean volume of ~ 150 mm³. Three treatment regimens, H-Zt/g4-MMAE at 1, 3, 7, 10, and 15 mg/kg in a Q6 days × 5 schedule, at 20 mg/kg in a Q12 days × 2 schedule, and at 10 mg/kg in a Q12 × 2 schedule, were used and injected through tail vein in a volume of 0.1 ml PBS. Tumor volumes were measured every four days as previously described [30, 31]. All mice were sacrificed at the end of experiments. Tumors from individual mice were collected and weighted to reach an average for each group. The percentage of inhibition was calculated as previously described [30, 31].

H-Zt/g4-MMAE at 10 and 30 mg/kg in a single dose or PBS was injected through the saphenous vein into cynomolgus monkeys (three animals per group). Animals were monitored daily for responsiveness, food consumption, bodyweight, temperature, pulse, and others. Urine and blood samples were collected according to the schedule. Electrocardiogram was performed before and after ADC injection every five days. Blood leukocytes and various enzymatic activities were measured as previously described [43]. All animals were sacrificed at the end of the study. Organs and tissues from individual monkeys were collected and weighted. Histological examinations were performed on all organs and tissues.

GraphPad 6 software was used for regular statistical analysis. Results are shown as mean ± SD. The data between control and experimental groups were compared using Student t test. The WinNonLin soft package was used for pharmacokinetic analysis. Statistical differences at p < 0.05 were considered significant.

Structures of CDRs from Zt/g4 in human IgG1/κ acceptor frameworks is shown in Fig. 1a. H-Zt/g4 has a binding affinity of 3.07 nM per ml for human RON (Fig. 1b and c). It also recognized cynomolgus monkey RON with a binding affinity at 2.21 nM per ml (Fig. 1d) but not mouse RON. Zt/g4 immunohistochemical staining of RON in multiple tissues from human or monkey is shown in Additional file 1: Figure S1. RON was detected at low levels in various epithelial cells from digestive, liver, kidney and other epithelial tissues. The patterns of RON reactivity between human and monkey tissues were similar.Fig1Fig. 1

Generation of humanized Zt/g4 antibody and characterization of RON-targeted antibody-drug conjugates: (a) Modeling of CDRs from mouse Zt/g4 in the variable regions of human IgG heavy chain and light chain. The framework of human IgG1 molecule was used for Zt/g4 humanization. The models of Zt/g4 CDRs grafted in the variable regions of human IgG1 heavy chain and light chain were generated by using the software PIGS from Automatic Predictions of Immunoglobulin Structures (Tramontano at University of Rome, Italy). (b) Binding of subclone H-Zt/g4 molecules to human RON. Different amounts of individual H-Zt/g4 s were incubated with NIH-3 T3 cells expressing human RON followed by addition of goat anti-human IgG1 antibody coupled with FITC. (c) Kinetic characterization of H-Zt/g4 interaction with human RON proteins by Octet RED96 system. Pure RON proteins from lysates of NIH3T3 cells expressing RON were immobilized onto the amine reactive sensor and assayed against individual H-Zt/g4 molecules in duplicate. The data set is analyzed with global fitting to produce the antibody-receptor binding affinity (K _D). Blue curves represent experimental data and red curves represent the statistical fitting of curves. (d) Interaction of H-Zt/g4 H1L3 with RONs from different species. NIH3T3 cells expressing human, monkey, or mouse RON were incubated with H-Zt/g4 H1L3 followed by goat anti-human IgG coupled with FITC. Immunofluorescent intensities from individual samples were determined by flow cytometric analysis. (e) Schematic representation of H-Zt/g4-MMAE structure. MMAE was conjugated to H-Zt/g4 by the valine-citruline dipeptide linker according to the manufacturer’s instruction (www.concortis.com). (f) HIC analysis of MMAE conjugated to H-Zt/g4: Individual Zt/g4-MMAEs with different numbers of MMAE (0 to 8) are marked as P0 to P8. A DAR combining P2, P4, and P6 at 3.77:1 was achieved. (g) Free MMAE dissociated from H-Zt/g4-MMAE in human plasma. H-Zt/g4-MMAE at 10 μg per ml was incubated with fresh human plasma at 37 °C for 20 days. The amount of free MMAE in plasma was determined using the LC-MS/MS method with slight modifications. (h) Samples from (g) were used also for measuring MMAE conjugated H-Zt/g4 as detailed in Materials and Methods. A ratio from free MMAE to the total MMAE in H-Zt/g4-MMAE was calculated to determine the percentages of MMAE dissociated from H-Zt/g4-MMAE

Four PDAC cell lines expressing variable levels of RON were used to evaluate H-Zt/g4-MMAE-induced RON internalization. The calculated RON molecules per PDAC cell were 10,214 ± 310 for BxPC-3, 13,178 ± 269 for FG, and 16,628 ± 245 for L3.6pl cells, respectively. Specific binding was not observed in Panc-1 cells (< 10 ± 2), which served as the control.

We first determined H-Zt/g4-MMAE-induced RON internalization (Fig. 2a). Less than 10% of RON remained on the cell surface 48 h after H-Zt/g4-MMAE treatment. The time required for H-Zt/g4-MMAE to induce a 50% RON reduction (internalization efficacy) was 12.0 h, 10.4 h, and 11.6 h for BxPC-3, FG, and L3.6pl, respectively. Immunofluorescence analysis using FG cells as the model confirmed intracellular RON localization with lysosomal-associated membrane protein (LAMP)-1 (Fig. 2b). H-Zt/g4-MMAE induced a significant reduction in PDAC cell viability in a dose-dependent manner (Fig. 2c). The IC₅₀ values at 96 h were 4.91 ± 0.11 μg per ml for BxPC-3, 4.04 ± 0.18 μg per ml for FG, and 2.53 ± 0.36 μg per ml for L3.6pl cells. Control Panc-1 cells was not affected by H-Zt/g4-MMAE with estimated IC₅₀ > 80 μg/ml. The effect of H-Zt/g4-MMAE on PDAC cell death was shown in Fig. 2d. The IC₅₀ of H-Zt/g4-MMAE in causing PDAC cell death ranged from 3.35 ± 0.26 μg per ml for BxPC-3, 2.71 ± 0.32 μg per ml for FG, and 1.37 ± 0.20 μg per ml for L3.6pl cells, respectively. These results demonstrate that H-Zt/g4-MMAE in vitro is highly effective in induction of RON internalization, which results in cell viability reduction followed by massive cell death.Fig2Fig. 2

Effect of H-Zt/g4-MMAE on RON internalization, cell viability, and death: (a) H-Zt/g4-induced cell surface RON internalization. PDAC cell lines BxPC-3, FG and L3.6pl (1 × 10⁶ cells per dish) were treated at 37 °C with 5 μg/ml of H-Zt/g4-MMAE, collected at different time points, washed with acidic buffer to eliminate cell surface bound IgG [31], and then incubated with 2 μg/mL of anti-RON mAb Zt/c1 [34]. Immunofluorescence was analyzed by flow cytometer using FITC-coupled anti-mouse IgG. Immunofluorescence from cells treated with H-Zt/g4 at 4 °C was set as 100%. Internalization efficiency (IC₅₀) was calculated as the time required achieving 50% reduction of cell surface RON. (b) Intracellular localization of internalized RON. FG cells in a 6-well plate were treated with 5 μg/ml of H-Zt/g4 at 4 °C or 37 °C for 12 h followed by mouse anti-human IgG1-coupled with FITC. Nuclear DNAs were stained with DAPI. LAMP-1 was used as a marker for protein cytoplasmic localization. Similar results also observed in additional three PDAC cell lines (data not shown). (c) Effect of H-Zt/g4-MMAE on viability of PDAC cells. Three PDAC cell lines (8000 cells per well in a 96-well plate in triplicate) were treated with different amounts of H-Zt/g4-MMAE for 96 h. Panc-1 cells served as the negative control. Cell viability was determined by the MTS assay. (d) Death of PDAC cells after H-Zt/g4-MMAE treatment. PDAC cells were treated with different amounts of H-Zt/g4-MMAE for 96 h. The percentages of cell death were determined by the trypan blue exclusion method. Data shown in (c) and (d) are derived from one of three experiments with similar results

The PKs of H-Zt/g4-MMAE in three doses were studied first in tumor-bearing and nonbearing mice to determine the time-dose relationship. Since H-Zt/g4 does not recognize mouse RON, the objectives were to determine: a) any alterations of the H-Zt/g4-MMAE PK profile in tumor-bearing mice and b) RON-independent behavior of H-Zt/g4-MMAE in tumor-nonbearing mice. Results in Fig. 3a show H-Zt/g4-NMAE in plasma in a two-compartment model from both tumor-bearing mice with a single dose of 3 and 20 mg/kg H-Zt/g4-MMAE and tumor-nonbearing mice with a single dose of 10 mg/kg H-Zt/g4-MMAE. Overall, the data from tumor-bearing mice were overlapped with those from the tumor-nonbearing mice with 95% prediction intervals. Thus, the PK of H-Zt/g4-MMAE is in no difference between tumor-bearing and nonbearing mice, suggesting that tumor growth does not affect the dynamics of H-Zt/g4-MMAE. Moreover, RON expression by tumor cells has no impact on H-Zt/g4-MMAE disposition in vivo.Fig3Fig. 3

Pharmacokinetic profiles of H-Zt/g4-MMAE in both mouse and cynomolgus monkey: (a) PK profiles of H-Zt/g4-MMAE in mouse. Tumor-bearing and -nonbearing mice (athymic nude, 5 mice per group) were injected with a single dose of H-Zt/g4-MMAE at 3, 10, and 20 mg/kg, respectively. Collected blood samples were analyzed using the MMAE ADC ELISA kit (Eagle Biosciences, Inc., Nashua, NH). Various PK parameters were calculated using the software provided by Eagle Biosciences. (b) Free MMAE dissociated from H-Zt/g4-MMAE in monkey plasma. A single dose of H-Zt/g4-MMAE at 10 or 30 mg/kg was injected into cynomolgus monkey (3 animals per group). Free MMAE from individual blood samples collected at different time intervals were subjected to the LC-MS/MS analysis. (c) PK profiles of H-Zt/g4-MMAE in cynomolgus monkey. Blood samples from (b) were analyzed for MMAE coupled H-Zt/g4 using the MMAE ADC ELISA kit as described in (a) to obtain various PK parameters

FG cell-mediated xenografts, which are fast growing and highly chemoresistant [44–46], were used to determine the dose-dependent effect of H-Zt/g4-MMAE. Xenografts caused by colorectal cancer (CRC) HT-29 cells served for comparison. The use of H-Zt/g4-MMAE at 1, 3, 7, 10, or 15 mg/kg in a Q6 × 5 schedule was based on ADC t½. H-Zt/g4-MMAE at 1 and 3 mg/kg is effective in delaying FG xenograft growth. The complete inhibition up to day 32 or 44 was found in mice receiving 7, 10, and 15 mg/kg H-Zt/g4-MMAE, respectively (Fig. 4a). Analysis of tumor number and weight showed that H-Zt/g4-MMAE not only inhibits tumor growth but also eradicates xenografts when it was used at high doses (Fig. 4b). Similar effects were also observed in HT-29 xenografts, which are more sensitive than FG models in response to H-Zt/g4-MMAE (Fig. 4a and b).Fig4Fig. 4

Therapeutic efficacy of H-Zt/g4-MMAE in PDAC xenograft tumor models: (a) Dose-dependent effect of H-Zt/g4-MMAE: Athymic nude mice (5 mice per group) were subcutaneously inoculated with 5 × 10⁶ FG cells. H-Zt/g4-MMAE at 1, 3, 7, 10, and 15 mg/kg was injected through tail vein in the Q6 × 5 regimen after tumors volumes reached to ~ 150 mm³. Mice injected with CmIgG-MMAE at 10 mg/kg were used as the control. Xenografts initiated by HT-29 cells served for comparison. (b) Effect of H-Zt/g4-MMAE in PDAC xenograft growth and eradication. Individual tumors from different groups described in (A) were collected from euthanized mice. Control mice bearing FG xenografts were sacrificed at day 24 due to rapid growth of tumors. Mice from other groups were killed at day 28 or day 44 dependent on the size of tumors. All tumors were weighted to reach the average tumor weight per group. The number of tumors from individual groups also was counted to determine the eradicating effect of H-Zt/g4-MMAE. NA, no tumors were found in the injected site. (c) Effect of H-Zt/g4-MMAE in three PDAC xenograft models: Xenograft tumors in mice (five animals per group) initiated by four PDAC cell lines were used for study. H-Zt/g4-MMAE was used at 20 mg/kg in the Q12 × 2 schedules. To establish the dose-effect relationship, the estimated reduction of H-Zt/g4-MMAE in vivo according to the t½ was marked as red circles. (d) Effect of H-Zt/g4-MMAE in tumor growth and eradication: Tumors were collected from mice described in (b). Tumor weight, count, and calculation were performed as described in (b). NA, no tumors were observed in the injected site

Toxicological studies were performed first to determine the maximum tolerated dose in mouse by a single injection of H-Zt/g4-MMAE at 40, 60, 80, and 100 mg/kg. Mice receiving H-Zt/g4-MMAE up to 60 mg/kg showed normal daily activity, food consumption, and maintained bodyweight (Additional file 1: Figure S6). However, H-Zt/g4-MMAE at 80 mg/kg resulted in dramatic reduction of bodyweight. Furthermore, animal death (three out of five mice) occurred in animals receiving 100 mg/kg H-Zt/g4-MMAE. These studies imply that H-Zt/g4-MMAE up to 60 mg/kg is relatively safe in mice as judged by activity, food consumption, and bodyweight.

Cynomolgus monkey was then used to determine adverse activities by a single injection of H-Zt/g4-MMAE at 10 and 30 mg/kg, respectively. All cynomolgus monkeys survived after a single H-Zt/g4-MMAE injection at 10 mg/kg and 30 mg/kg. Abnormal changes were not observed grossly for daily activity, bodyweight, body temperature, food consumption, vision, electrocardiogram, breath, and urine samples. Histological analysis of multiple tissue/organs from individual monkeys did not find evidences of inflammation, hemorrhage, cell death, tissue damage, and weight reduction (Additional file 1: Figure S7).

The adverse activities of H-Zt/g4-MMAE were documented from analysis of blood leukocytes, erythrocytes, reticulocytes (Fig. 6a, b, c, Additional file 2: Table S2), and liver function (Fig. 6d, and Additional file 2: Table S2). Leukocyte changes were featured by temporary and reversible reduction of the total number of leukocytes (Fig. 6a). H-Zt/g4-MMAE at 10 mg/kg at day 8 moderately decreased the total number of leukocytes (36.8%). Significant reduction (83.2%) was seen only when H-Zt/g4-MMAE reaches 30 mg/kg. Among three types of leukocytes analyzed, neutrophils were the most sensitive followed by monocytes and then lymphocytes (Fig. 6a and b). Due to the reduction of neutrophils, an increase in relative percentages of lymphocytes and monocytes was observed (Fig. 6a and b). It is worthy to note that the effect of H-Zt/g4-MMAE is nonspecific because blood leukocytes do not express RON [47]. The total number of leukocytes, including all three subtypes of leukocytes, was recovered to the normal level within ~ 20 days even H-Zt/g4-MMAE was used at 30 mg/kg. These results suggest that the hematopoietic system was not severely inhibited and/or impaired by H-Zt/g4-MMAE. Histological analysis of bone marrows from ADC-treated animals supported this notion (Additional file 1: Figure S7).Fig6Fig. 6

Toxicological activities of H-Zt/g4-MMAE in mouse and cynomolgus monkey. (a) and (b) Adverse activities of H-Zt/g4-MMAE in blood leukocytes in cynomolgus monkey. H-Zt/g4-MMAE at 10 or 30 mg/kg in a single dose was injected once into cynomolgus monkey. Monkeys without ADC injection served as the control. Peripheral blood samples were collected at different time intervals. Total numbers of leukocytes (a) including neutrophil, lymphocytes, and monocytes from each group were countered accordingly. The percentages of blood leukocytes (b) were also determined. (c) Adverse effect of H-Zt/g4-MMAE on blood erythrocytes and reticulocytes in cynomolgus monkey. Total numbers of erythrocytes and reticulocytes from blood samples collected from each group as described in (b) were counted accordingly. (d) Adverse effects of H-Zt/g4-MMAE on various enzymes in plasma of cynomolgus monkey. Six enzymatic activities from each group were quantitatively measured using the blood samples collected from individual monkeys as described in (b)