Accelerating ADC Development with Integrated Preclinical Solutions

Figure 2. Internalization kinetics of ABBV-399 in c-Met-positive tumor cells.

 (A) Time-dependent internalization of ABBV-399 and telisotuzumab was assessed in Hs 746T, NCI-H441, and NCI-H1650 cells using pHrodo, a pH-sensitive fluorogenic dye that produces an increased fluorescent signal following trafficking into acidic intracellular compartments. Internalization is presented as pHrodo MFI. (B) Internalization rates were quantified at the indicated time points for each tumor cell line.

ABBV-399 showed time-dependent internalization in Hs 746T and NCI-H441 cells, with internalization profiles generally comparable to those of telisotuzumab. In contrast, only a limited internalization signal was detected in NCI-H1650 cells despite measurable surface binding. These findings indicate that antigen binding alone may not be sufficient to predict productive ADC uptake.

The huHSC-NCG mouse model is generated by transplanting human CD34⁺ hematopoietic stem cells (HSC) into sublethally irradiated immunodeficient NCG mice. This model enables the endogenous reconstitution of functional human immune cells, including T cells and immature B cells, with stable engraftment and extended survival period (>90% survival beyond 35 weeks). Owing to its robust and durable human immune system reconstitution, the huHSC-NCG model is widely applied in studies of B cell depletion, in vivo CAR-T biodistribution, as well as long-term toxicity and safety evaluation of in vivo CAR-T therapeutics.

Figure 3. Immune cell profiling in huHSC-NCG mice

Figure 4. NCG mouse survival after huHSC reconstitution

This model was applied successfully in the toxicity and safety evaluation of SYS6055, an in vivo CAR-T therapeutic developed by CSPC Pharmaceutical Group. In addition, the Center for Drug Evaluation (CDE) of the National Medical Products Administration has recognized PBMC- and HSC-humanized NCG mouse models  as appropriate platforms for the non-clinical evaluation of in vivo CAR-T therapeutics, including studies conducted with clinical-grade materials.

 (A) BALB/c nude mice bearing SNU-5 xenografts were treated intravenously with vehicle, isotype ADC at 6 mg/kg, or CDH17-targeting ADC at 2 or 6 mg/kg once weekly for four doses. Tumor volumes were monitored throughout the study. (B) Body weights were recorded during the treatment period as an indicator of general tolerability.

Tumor-volume and body-weight data are presented as mean ± SEM; n = 5 mice per group. Statistical comparisons of tumor volume at the final assessment time point were performed using independent-samples t-tests versus the vehicle group. ns, not significant; *p < 0.05; **p < 0.01.

The CDH17-targeting ADC inhibited tumor growth in a dose-dependent manner in the SNU-5 gastric cancer model. Treatment at 6 mg/kg produced substantially greater tumor growth inhibition than the 2 mg/kg dose, while the isotype ADC showed limited activity. Body weights remained generally stable during treatment, supporting acceptable tolerability under the tested dosing regimens.

(A) BALB/c nude mice bearing NCI-N87 gastric cancer xenografts were treated intravenously with PBS once weekly for five doses or Trodelvy at 3 or 10 mg/kg once weekly for four doses. Tumor volumes were monitored through Day 35. At the endpoint, n = 5 mice in the PBS group and n = 4 mice in each Trodelvy group.

(B) BALB/c nude mice bearing HCC1806 breast cancer xenografts were treated intravenously with PBS or Trodelvy at 3 or 10 mg/kg once weekly for four doses. Tumor volumes were monitored through Day 32. At endpoint, n = 3 mice in the PBS group and n = 4 mice in each Trodelvy group.

(C) NCG mice bearing MDA-MB-468 triple-negative breast cancer xenografts were treated intravenously with PBS or Trodelvy at 3 or 10 mg/kg once every two weeks for two doses. Tumor volumes were monitored through Day 35. n = 4 mice per group.

Data are presented as mean ± SEM. Statistical comparisons at the final assessment time point were performed using two-tailed unpaired Welch’s t-tests versus the corresponding vehicle group. ns, not significant; **p < 0.01; ****p < 0.0001.

Trodelvy demonstrated antitumor activity across all three CDX models, although the magnitude and kinetics of response varied by tumor type and dosing regimen. In the HCC1806 and MDA-MB-468 models, the 10 mg/kg dose produced greater tumor growth inhibition than the 3 mg/kg dose. In the NCI-N87 model, Trodelvy was evaluated alongside T-DXd, enabling comparison of ADC activity within the same tumor model.

These model-dependent responses demonstrate the value of evaluating an ADC across a broader tumor panel to identify responsive indications, optimize dose selection, and select appropriate models for subsequent translational studies.

Antitumor activity of T-DXd in PDX models with different HER2 expression levels. Tumor growth curves in HER2 IHC 0 PDX models SC00685, SC00429, SC00352 and in HER2 IHC 3+ PDX models SC00730, SC00314, and SC00748. Tumor-bearing mice received vehicle or T-DXd at 3 or 10 mg/kg once every three weeks by intraperitoneal administration. Tumor volumes were monitored throughout the respective study periods. Data are presented as mean ± SEM; n = 5 mice per group. Statistical comparisons at the final assessment time point were performed using two-tailed unpaired Welch’s t-tests versus the corresponding vehicle group. ns, not significant; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

T-DXd produced pronounced tumor growth inhibition across the HER2 IHC 3+ PDX models. Responses among the HER2 IHC 0 models were more heterogeneous: some models showed substantial tumor growth inhibition or regression, whereas others exhibited a clearer dose-dependent response.

These findings indicate that HER2 IHC classification alone may not fully predict ADC sensitivity. Intratumoral target heterogeneity, antigen distribution, internalization efficiency, payload sensitivity, and tumor exposure may also contribute to treatment response. Evaluation across a PDX panel therefore provides an important translational layer beyond conventional CDX efficacy studies.

Serum pharmacokinetics and tumor distribution of T-DXd in BALB/c nude mice bearing NCI-N87 xenografts. Female BALB/c nude mice bearing NCI-N87 tumors received a single intravenous dose of T-DXd at 3 mg/kg. (A) Serum concentrations of total antibody, conjugated ADC, and free payload were measured at 0.0833, 7, 24, 72, 120, 168, and 336 hours after dosing. (B) Tumor concentrations of total antibody, conjugated ADC, and free payload were measured at 7, 72, 168, and 336 hours. (C) Tumor-to-serum concentration ratios were calculated for total antibody and conjugated ADC. Total antibody and conjugated ADC were quantified by ELISA, and free payload was quantified by LC–MS/MS. Data are presented as mean concentrations; n = 3 per time point. (D) Summary pharmacokinetic parameters, including apparent half-life (T₁/₂), maximum concentration (Cmax), and area under the concentration–time curve from time zero to the last measurable time point (AUC₀–t), for total antibody and conjugated ADC in serum and tumor tissue. Total antibody and conjugated ADC were quantified by ELISA, while free payload was quantified by LC–MS/MS. Tumor concentrations are expressed as nmol/kg wet tissue and serum concentrations as nmol/L. A tissue density of 1 g/mL was assumed. Data are presented as mean ± SEM; n = 3 per time point.

In BALB/c nude mice bearing NCI-N87 tumors, total antibody and conjugated ADC remained detectable in serum and tumor tissue for up to 336 hours after dosing. As observed in the NCG model, free payload was not detected in serum.

Compared with NCG mice, BALB/c nude mice showed longer apparent half-lives and higher systemic and tumor exposure for both total antibody and conjugated ADC. The extended tumor retention and increasing tumor-to-serum ratios indicate sustained intratumoral exposure after systemic clearance from circulation.

The differences observed between NCG and BALB/c nude mice demonstrate that host strain and biological background can influence ADC pharmacokinetics. Mouse-strain selection should therefore be considered when interpreting exposure data and designing PK–efficacy studies.

n vivo evaluation of T-DXd-mediated bystander activity in a heterogeneous NCI-N87/U87 MG-Luc xenograft model. NCG mice bearing mixed NCI-N87/U87 MG-Luc tumors received a single intravenous administration of PBS, trastuzumab at 10 mg/kg, or T-DXd at 10 mg/kg. (A) Representative HER2-expression images for the NCI-N87 and U87 MG-Luc tumor components. (B) Mean tumor-volume changes during the study, measured by caliper. (C) Longitudinal bioluminescence images collected on Days 0, 7, 14, 21, and 28 after treatment. (D) Mean bioluminescence radiance over time, used to monitor the luciferase-labeled U87 MG cell population. Tumor-volume and bioluminescence data are presented as mean ± SEM; n = 5 mice per group. Statistical comparisons of tumor volume at the final assessment time point were performed using two-tailed unpaired Welch’s t-tests versus the PBS group. **p < 0.01; ****p < 0.0001.

A mixed NCI-N87/U87 MG-Luc xenograft model was used to evaluate T-DXd activity in a tumor containing cell populations with different HER2 expression levels. Both trastuzumab and T-DXd inhibited overall tumor growth based on caliper-derived tumor volume. However, longitudinal bioluminescence imaging showed a pronounced reduction in the luciferase-labeled U87 MG cell population following T-DXd treatment.

The reduction in the lower-HER2 neighboring cell population supports in vivo bystander activity of the membrane-permeable DXd payload. In contrast to tumor volume alone, bioluminescence imaging enabled specific tracking of one tumor-cell component within the heterogeneous tumor, providing a more direct assessment of bystander killing.

(A) Schematic representation of the model-establishment process. JIMT-1 tumors were treated with vehicle or Enhertu. Tumors showing limited response to Enhertu were collected and serially reimplanted under continued treatment pressure to generate the JIMT-1–Enhertu-ER enhanced model. (B) Representative HER2 immunohistochemical staining and tumor growth curves from two independently derived JIMT-1–Enhertu-ER tumors. Mice bearing Tumor 1 or Tumor 2 were treated intravenously with vehicle, Enhertu at 6 mg/kg, or Enhertu at 10 mg/kg once weekly for four weeks. Tumor growth inhibition values are indicated in the figure. Tumor-volume data are presented as mean ± SEM; Statistical comparisons were performed using independent samples t-test; ns, not significant.

The JIMT-1–Enhertu-ER enhanced model was established through repeated in vivo selection under Enhertu treatment pressure. Non-responsive tumors were collected, reimplanted, and subsequently challenged with Enhertu to enrich for tumors with reduced treatment sensitivity.

Two independently derived JIMT-1–Enhertu-ER tumors retained HER2 expression by IHC but showed limited responses to repeated Enhertu treatment. In Tumor 1, Enhertu produced tumor growth inhibition values of 30.38% at 6 mg/kg and 36.13% at 10 mg/kg. In Tumor 2, the corresponding TGI values were 12.72% and 36.24%. Neither dose produced a statistically significant reduction in tumor volume compared with vehicle under the tested conditions.