Development of a novel tumor-targeted vascular disrupting agent activated by membrane-type matrix metalloproteinases

Abstract

Vascular disrupting agents (VDA) offer a strategy to starve solid tumors of nutrients and oxygen concomitant with tumor shrinkage. Several VDAs have progressed into early clinical trials, but their therapeutic value seems to be compromised by systemic toxicity. In this report, we describe the design and characterization of a novel VDA, ICT2588, that is nontoxic until activated specifically in the tumor by membrane-type 1 matrix metalloproteinase (MT1-MMP). HT1080 cancer cells expressing MT1-MMP were selectively chemosensitive to ICT2588, whereas MCF7 cells that did not express MT1-MMP were nonresponsive. Preferential hydrolysis of ICT2588 to its active metabolite (ICT2552) was observed in tumor homogenates of HT1080 relative to MCF7 homogenates, mouse plasma, and liver homogenate. ICT2588 activation was inhibited by the MMP inhibitor ilomastat. In HT1080 tumor-bearing mice, ICT2588 administration resulted in the formation of the active metabolite, diminution of tumor vasculature, and hemorrhagic necrosis of the tumor. The antitumor activity of ICT2588 was superior to its active metabolite, exhibiting reduced toxicity, improved therapeutic index, enhanced pharmacodynamic effect, and greater efficacy. Coadministration of ICT2588 with doxorubicin resulted in a significant antitumor response (22.6 d growth delay), which was superior to the administration of ICT2588 or doxorubicin as a single agent, including complete tumor regressions. Our findings support the clinical development of ICT2588, which achieves selective VDA targeting based on MT-MMP activation in the tumor microenvironment.

Figure 1

MT-MMPs are elevated in human preclinical tumor models. Expression of MMP mRNA in human cell lines grown in vitro (A) and as xenografts in vivo (B) as measured by quantitative RT-PCR. Expression values after normalization to 18S-rRNA and are gene specific. Classification of expression levels was determined from the C_T of each gene as either very high (C_T ≤ 25), high (C_T = 26-30), moderate (C_T = 31-35), low (C_T = 36-39), or not detected (CT = 40); see key for color scheme. (C) Immunoblot of MT1-MMP protein expression in HT1080 and MCF7 tumor models. Key for qRT-PCR expression:

Figure 2

ICT2588 is selectively activated by MMPs. (A) Structure of ICT2588 and ICT2552, indicating the MT-MMP selective scissile bond. (B) Metabolism of ICT2588 by HT1080 and MCF7 tumors relative to mouse liver and mouse plasma. Metabolites detected by LCMS and expressed as concentration of ICT2588 remaining. (C) Metabolism of ICT2588 in the presence of a pan-MMP inhibitor, Ilomastat. Each value represents the mean ± SD of 3 independent experiments.

Figure 3

Pharmacokinetic analysis of ICT2588 following intraperitoneal administration to HT1080 tumor bearing mice. Analysis by LCMS of (A) ICT2588 and (B) the authentic metabolite produced from hydrolysis of ICT2588 (ICT2552^act). Each time point represents the mean of 3 mice ± SD.

Figure 4

Induction of tumor vascular disruption and hemorrhagic necrosis by ICT2588 in vivo. Mice bearing HT1080 tumors were treated with ICT2588 or ICT2552 and assessed for functional vasculature (A) and hemorrhagic necrosis (B) 24 h post-treatment. Graph represents mean values from 3 mice ± standard error. Image (A) demonstrates loss of functional tumor vasculature, (as described previously, see refs. 37, 40, 41) are indicated in (A). Image (B) demonstrates the persistence of a viable tumor rim.

Figure 5

(A) ICT2588 (peptide-conjugate) induces a significantly greater inhibition of HT1080 tumor growth compared to (B) the authentic metabolite (ICT2552) at equimolar concentrations. Mice were treated with a single intraperitoneal dose of ICT2588 (A) or ICT2552 (B) and tumor size determined daily. *One animal in the ICT2588 62.5 mg kg⁻¹ treatment group showed complete tumor remission. (C) Combination of ICT2588 with doxorubicin resulted in significantly greater antitumor potency relative to single agent doxorubicin or ICT2588 administration. Mice were administered 75 mg kg⁻¹ ICT2588 (i.p.) 24 h prior to 5 mg kg⁻¹ doxorubicin (i.v.) and tumor size determined daily (C). Four out of eight mice showed complete tumor remission with the co-administration schedule; therefore results in this group are the summary of four remaining mice. The tumor growth delay is defined as time taken for doubling of tumor volume relative to control, denoted in brackets in the legend of each figure.