SGN-CD33A: A Novel CD33-Directed Antibody-Drug Conjugate, Utilizing Pyrrolobenzodiazepine Dimers, Demonstrates Preclinical Antitumor Activity Against Multi-Drug Resistant Human AML

Abstract 3589

Outcomes for patients with acute myeloid leukemia (AML) are poor, highlighting the need for novel treatment options. Most AML cells express the myeloid differentiation antigen CD33, making CD33-targeted therapy a potential treatment strategy. Gemtuzumab ozogamicin (GO), an anti-CD33 monoclonal antibody (mAb) conjugated to the cytotoxic agent calicheamicin, has recently been shown to improve survival in newly diagnosed patients with more favorable-risk AML but has insufficient activity in those with poor prognostic features as well as in relapsed disease.

Here we report the preclinical testing of a novel CD33-directed antibody-drug conjugate, SGN-CD33A, consisting of a humanized anti-CD33 mAb with 2 engineered cysteine residues through which pyrrolobenzodiazepine (PBD) dimer drug moieties are conjugated via a maleimidocaproyl valine-alanine dipeptide linker. PBD dimers exert their biological activity by covalent binding and interstrand cross-linking of DNA. Fluorescence microscopy studies showed that SGN-CD33A is rapidly internalized and traffics to lysosomes within hours of binding to CD33-positive AML cell lines. Following uptake, SGN-CD33A induces DNA damage as measured by phosphorylation of histone 2AX, subsequently leading to G2-M cell cycle arrest, disruption of mitochondrial membrane integrity, increased caspase-3 activity, formation of cleaved poly ADP-ribose polymerase, DNA fragmentation and cell death.

The anti-leukemic activity of SGN-CD33A was assessed in cytotoxicity assays against 12 AML cell lines and 18 primary AML patient samples of mixed cytogenetic origin (favorable, intermediate and unfavorable) and multidrug resistance (MDR) status. SGN-CD33A was highly active against all AML cell lines tested (mean IC50, 22 ng/ml), including 5 of 5 MDR-positive cell lines (mean IC50, 27 ng/mL). In contrast, GO was moderately active in 1 of 5 MDR-positive cell lines (IC50, 227 ng/mL) but inactive against the other 4 (IC50, >1000 ng/mL). SGN-CD33A was also active against 15 of 18 primary samples isolated from untreated AML patients at diagnosis (mean IC50 of responsive samples, 8 ng/mL) and was more potent than GO which was active in 10 of 18 AML samples (mean IC50 of responding samples, 27 ng/mL). The 3 AML specimens that were resistant to SGN-CD33A each had low or absent CD33 expression, as determined by flow cytometry. Cytogenetic abnormalities and MDR activity did not correlate with in vitro SGN-CD33A cytotoxicity.

In vivo antitumor activity was evaluated in AML mouse xenograft models established with MDR-negative HL-60 and MDR-positive THP-1 and TF1-α cell lines. SGN-CD33A dosed once at 300 mcg/kg yielded durable complete regressions in THP-1 (Figure A) and TF1-α xenografts. In these models of drug-resistant AML, treatment with a single dose of 100 mcg/kg SGN-CD33A significantly delayed tumor growth compared to untreated and non-binding control ADC-treated mice (p<0.001), whereas GO was inactive even when dosed at 1000 mcg/kg. In the MDR-negative HL-60 model, a single dose of 30 mcg/kg of SGN-CD33A delayed tumor growth and 100 mcg/kg induced complete tumor regression (Figure B). In contrast, GO had minimal activity when dosed at 100 mcg/kg but did result in durable tumor regression when administered at a ten-fold higher dose (Figure B).

Together, these data demonstrate that SGN-CD33A exhibits antitumor activity against a broad panel of primary AML samples and results in durable remissions in preclinical models of MDR-positive AML that are characteristically resistant to conventional chemotherapy and GO. CD33-directed delivery of PBD dimers may overcome multidrug resistance and may represent a new strategy for the treatment of patients with AML. Clinical trials are planned to further evaluate SGN-CD33A in AML.

View largeDownload slide

View largeDownload slide

Close modal