Preclinical and clinical activity of DZD1516, a full blood-brain barrier-penetrant, highly selective HER2 inhibitor

Abstract

Background: Patients with HER2-positive metastatic breast cancer (MBC) are at high risk of developing central nervous system (CNS) metastases. A potent and selective HER2 inhibitor with good blood-brain barrier (BBB) penetration is highly desirable.

Methods: The design and structure-activity relationship of DZD1516 was described. The potency and selectivity of DZD1516 were determined by enzymatic and cellular assays. The antitumor activity of DZD1516 monotherapy or in combination with HER2 antibody-drug conjugate was assessed in CNS and subcutaneous xenograft mouse models. A phase 1 first-in-human study evaluated the safety, tolerability, pharmacokinetics, and preliminary antitumor activity of DZD1516 in patients with HER2+ MBC who relapsed from standard of care.

Results: DZD1516 showed good selectivity against HER2 over wild-type EGFR in vitro and potent antitumor activity in vivo. Twenty-three patients were enrolled and received DZD1516 monotherapy treatment across six dose levels (25-300 mg, twice daily). Dose-limiting toxicities were reported at 300 mg, and thus 250 mg was defined as the maximum tolerated dose. The most common adverse events included headache, vomiting, and hemoglobin decreased. No diarrhea or skin rash was observed at ≤ 250 mg. The mean K_p,uu,CSF was 2.1 for DZD1516 and 0.76 for its active metabolite DZ2678. With median seven lines of prior systemic therapy, the best antitumor efficacy in intracranial, extracranial and overall lesions was stable disease.

Conclusions: DZD1516 provides positive proof of concept for an optimal HER2 inhibitor with high BBB penetration and HER2 selectivity. Further clinical evaluation of DZD1516 is warranted, with the RP2D being 250 mg BID.

Clinicaltrials: gov identifier NCT04509596. Registered on August 12, 2020; Chinadrugtrial: CTR20202424 Registered on December 18, 2020.

Keywords: Blood–brain barrier; Breast cancer; Central nervous system metastases; HER2; Tyrosine kinase inhibitor.

Fig. 1

Modeling of DZD1516 with HER2 (PDB code: 3RCD). Key interactions of DZD1516 with HER2 protein include (i) H-bonding of quinazoline with hinge (Met801); (ii) piperidine occupied the ribose-binding pocket; (iii) [1, 2, 4]triazolo[1,5-a]pyridin-7-yloxy scaffold near the gatekeeper back pocket for HER2 selectivity. Yellow: carbon; purple: nitrogen; red: oxygen; light blue: fluorine. Colors on the protein surface represent the ATP binding pocket and are for clarity only

Fig. 2

Antitumor activity of DZD1516 in BM, LM, and subcutaneous BT474C1-Luci Mono1 xenograft mice model. The BT474C1-Luci Mono1 cell model was a stable clone generated by transfection of luciferase gene into BT474 cell line and selection of stable clones. The mice were treated with DZD1516 at 100 mg/kg and 150 mg/kg twice daily, respectively, after tumors were established. A Plots of tumor volume in BM model. The statistical analysis of tumor volume difference at week 3 between DZD1516 and tucatinib groups was performed by two-way ANOVA, compared with the tucatinib group. ****P < 0.0001, ns: not significant. B Representative images of bioluminescent signals at week 0 and week 3 after the start of compound treatment in BM model. C Plots of tumor volume in LM model. D Representative images of bioluminescent signals at week 0 and week 2 after the start of compound treatment in LM model. E Plots of tumor volume in subcutaneous model. BM brain metastasis, LM leptomeningeal metastasis. BID: twice daily

Fig. 3

Antitumor activity of DZD1516 in combination with T-DM1 or T-DXd in BT474C1-Luci Mono1 xenograft mice model. The BT474C1-Luci Mono1 cell model was a stable clone generated by transfection of luciferase gene into BT474 cell line and selection of stable clones. A Efficacy study of DZD1516 in combination with T-DM1 in the BM model. n = 9/group. ****P < 0.0001 (vs. T-DM1 group, two-way ANOVA). B Efficacy study of DZD1516 in combination with T-DM1 in the SC model. n = 10/group. ns: not significant (vs. T-DM1 group, two-way ANOVA). C Efficacy study of DZD1516 in combination with T-DXd in the BM model. ***P < 0.001 (vs. T-DXd group, two-way ANOVA). D Efficacy study of DZD1516 in combination with T-DXd in the SC xenograft model. BID: twice daily; q2w: once every 2 weeks. qw: once weekly. BM brain metastasis, SC subcutaneous

Fig. 4

PK and PD relationship post-single dose of DZD1516 in BT474C1-Luci Mono1 SC xenograft mice model. A single dosing of DZD1516 at 25, 50, 150 mg/kg was administrated when the tumor volume reached 200–600 mm³. Plasma and tumor tissues were collected at 0.25, 2, 6, and 24 h post-dosing to analyze the PK/PD correlation. The pHER2 expression in tumor tissues was detected by immunohistochemistry (IHC) and normalized to the vehicle control group. Each time point had tumor tissues from three mice to detect the pHER2 signal. PK pharmacokinetics, PD pharmacodynamics, SC subcutaneous