Abstract

Gemtuzumab Ozogomicin (GO) is a humanized IgG4 anti-CD33 monoclonal antibody covalently linked to the powerful antitumour antibiotic, calicheamicin. CD33 expression is found on myeloid (mature and immature), erythroid, megakaryocytic, and multipotent progenitors but is absent from normal CD34+ pluripotent haematopoietic stem cells and non-haematopoietic tissues. GO has been used successfully in de novo and refractory/relapsed CD33+ AML, and in other CD33+ leukaemias, including ALL - the main rationale being it selectively targets the CD33+ blast population while sparing the CD33– haematopoietic stem cell compartment. Although myelosuppression is the main toxicity seen in the vast majority of patients (>95%) receiving GO at doses ranging from 6 to 9mg/m2, it is the liver injury, especially veno-occlusive disease (VOD)-like syndrome, that is unique and the most significant drug-related toxicity observed. Previous studies have shown that patients who develop VOD-like syndrome have higher circulating levels of pro-inflammatory cytokines and reduced levels of the natural anti-coagulants protein C and anti-thrombin, suggesting that a pro-inflammatory and pro-coagulant state is induced in the liver, however no specific mechanism has been elucidated to account for the liver toxicity. One possible explanation is that GO binds to CD33+ Kupffer cells and/or other CD33+ cells residing in the hepatic sinusoids and in doing so induces state of pro-inflammation and pro-coagulation which in turn causes the sinusoidal fibrosis, centrilobular congestion, and hepatocyte necrosis, the histological hallmark of this process. To investigate this possibility, the CD33+ AML cell line, THP1, and the CD33– hepatocyte cell line, HepG2, were treated with increasing concentrations of GO. Secreted levels of the pro-inflammatory cytokines TNF-α and IL-8 were measured by sandwich ELISA. Cell surface expression levels of Tissue Factor (TF), the critical initiator of the pro-coagulation pathway, were analysed by flow cytometry. Cell proliferation was measured by the WST-1 assay. A statistically significant (P<0.001) pro-coagulant but not pro-inflammatory response was observed following GO exposure in both cell lines. THP1 cells showed upregulation of TF expression after 24 hour incubation with GO but no significant increase in TNF-α or IL-8 levels. Neither cell line showed a significant change in proliferation levels after treatment with GO. These results suggest that the initial response to GO treatment is pro-coagulant. After the first treatment, there will be fewer CD33+ blast cells, leaving more GO molecules free to ligate to CD33+ targets in the liver, therefore repeated administration of the drug increases the pro-coagulant state and perpetuates the liver injury. Continuous injury of the liver initiates an inflammatory response through autocrine and paracrine loops. This serves as the initiation signal for fibrosis. If the cycle of inflammation and coagulation proceeds unchecked, the fibrogenic cascade is continued leading to the development of VOD-like syndrome.

Author notes

Disclosure: No relevant conflicts of interest to declare.