Introduction

Angiogenesis drives tumor progression; in particular, angiogenesis is critical for the growth and metastasis of solid tumors1,2,3,4. The inhibition of tumor angiogenesis has therefore become an important strategy for cancer therapy. Several inhibitors of tumor angiogenesis, such as bevacizumab (Avastin), sorafenib (Nexavar) and sunitinib (Sutent), have been successfully used in the clinic to treat solid tumors4. This success has led to increased efforts to discover new angiogenesis inhibitors with different mechanisms of action and/or distinct chemical structures. During the past 10 years, we have discovered and reported on 17 compounds that demonstrate potent inhibition of angiogenesis (Table 1). In contrast to the inhibitors that are already employed for clinical purposes, these newly discovered antiangiogenic agents display intriguing diversity in their sources, chemical structures, antiangiogenic mechanisms, and molecular targets. This review will focus on these diverse characteristics of the 17 newly discovered antiangiogenic compounds, as well as the representative findings that have been produced by the mechanistic investigation and preclinical development of these compounds.

Table 1 Angiogenesis inhibitors discovered in the Shanghai Institute of Materia Medica since 2001.
Full size table

An overview of the 17 antiangiogenic compounds

In a search for new angiogenesis inhibitors, we performed extensive evaluations of thousands of natural products and synthetic or semi-synthetic compounds over the course of the past 10 years. From this process, 17 compounds were discovered to possess antiangiogenic activity (Table 1). The mechanisms of action of several of these compounds have been investigated, and one of these compounds, namely, AL3810, has been registered as an IND (investigational new drug) for cancer therapy in China.

Analyses of these 17 compounds indicate their diverse sources (Table 1). In particular, 7 of these inhibitors are derived from terrestrial sources, and 5 of these 7 compounds (pseudolaric acid B, triptolide, 10-hydroxycamptothecin, 11,11′-dideoxyverticillin and shiraiachrome A) are used in traditional Chinese medicine (TCM). It is also notable that quercetin is present both in many fruits and vegetables and in olive oil, red wine and tea. Of the remaining 10 compounds that were identified, 5 are marine-derived inhibitors, 3 of which [JG3, MDOS and Grateloupia longifolia polysaccharide (GLP)] are saccharides and 2 of which are saponins from sea cucumber. The final 5 inhibitors are chemically synthesized; these synthesized compounds are also representative of a great diversity of chemical structures, including terpenoids, alkaloids, flavonoids, saccharides, saponins and pyrido-pyrimidines, as shown in Table 1.

The apparent differences in the sources and chemical structures of the 17 inhibitors are reflected in their distinct (potential) targets. These targets include angiogenic kinases (angiokinases), extracellular matrix (ECM) components and the hypoxia-inducible factor 1α (HIF-1α)-vascular endothelial growth factor (VEGF) axis, among others (Table 1, Figure 1). Moreover, the antiangiogenic activities of these compounds revealed in the in vitro, ex vivo and/or in vivo experimental models are mediated by distinct molecular signaling pathways (Figure 1).

Figure 1
figure 1

A schematic overview of molecular signaling that possibly mediates experimental antiangiogenic activities of the compounds discussed here.

Full size image

Angiokinase inhibitors

Although many protein tyrosine kinases (PTKs) contribute to the angiogenic process, the VEGF-VEGF receptor (VEGFR) axis is the main target for clinical applications of antiangiogenic therapy5. Of the 17 compounds listed in Table 1, 9 compounds, namely, 11,11′-dideoxyverticillin, shiraiachrome A, MDOS, philinopside A, philinopside E, AL3810, BB, TKI-28, and TKI-31, were found to directly inhibit this axis6,7,8,9,10,11,12,13,14 (Table 1 and Figure 1). The former 5 compounds are derived from natural products, whereas the latter 4 inhibitors are synthetic. These 9 compounds display different profiles of PTK inhibition, as they have distinct selectivity against various receptor and/or non-receptor tyrosine kinases, including human epidermal growth factor receptor 2 (HER2), epidermal growth factor receptor (EGFR), VEGFR, platelet-derived growth factor receptor (PDGFR), c-Kit, fibroblast growth factor receptor 1 (FGFR1) and/or c-Src. Nevertheless, all 9 of these compounds can directly suppress the critical angiokinase VEGFR, producing significant experimentally observed antiangiogenic effects as a result (Table 1 and Figure 1). In particular, AL3810 has been in clinical trials because it not only displays excellent anticancer and antiangiogenic activities but also demonstrates good pharmacokinetics and toxicity in preclinical studies15,16.

HIF-1α-VEGF axis inhibitors

We also found that of the 17 identified antiangiogenic compounds, 4 compounds, namely, pseudolaric acid B, MFTZ-1, 10-hydroxycamptothecin and triptolide, can indirectly inhibit the VEGF-VEGFR axis by decreasing cellular HIF-1α accumulation and thereby reducing VEGF expression and secretion15,16,17,18,19,20,21 (Figure 1). We define these compounds to be HIF-1α-VEGF axis inhibitors. HIF-1α is a critical transcription factor that impacts tumor angiogenesis by regulating the expression of VEGF. HIF-1α has thus been proposed as a promising anticancer target.

The HIF-1α-VEGF axis inhibitors reduce the cellular amount of HIF-1α in different ways. Pseudolaric acid B targets microtubulin and causes its depolymerization22,23. Pseudolaric acid B inhibits angiogenesis by reducing the stability of HIF-1α and thereby downregulating the VEGF-VEGFR axis15,16,18. However, there has been no direct evidence indicating any association between its antiangiogenic activity and its inhibition of microtubulin23,24. By contrast, MFTZ-1 does not affect either the degradation of HIF-1α protein or the level of HIF-1α mRNA. Instead, MFTZ-1 can abrogate the HIF-1α-driven increase in VEGF mRNA and VEGF protein secretion, producing antiangiogenic effects. Specifically, MFTZ-1 can reduce constitutive, HIF-1α-independent VEGF secretion and concurrently antagonize inducible, HIF-1α-dependent VEGF secretion, in an effect that is independent of its inhibition of its primary target, topoisomerase II17,25. The inhibition of angiogenesis by 10-hydroxycamptothecin19 may be associated with this drug's suppression of HIF-1α expression, which occurs via the repression of topoisomerase I-dependent transcription. This mechanism of action is possibly similar to the mechanism that is used by topotecan, another camptothecin derivative26. In contrast to the previous 3 inhibitors, triptolide may enhance the levels of cellular HIF-1α mRNA and protein20. However, triptolide also causes the downregulation of VEGF expression and secretion20, possibly because it binds to XPB (which is also known as ERCC3) and causes the degradation of RNA polymerase II21, disrupting the transcriptional function of HIF-1α.

ECM component inhibitors

ECM components including heparanase and matrix metalloproteinase (MMP) are critically involved in the metastatic and angiogenic capabilities of tumor cells. Inhibitors targeting ECM components are increasingly emerging as promising agents for cancer therapy. We have discovered a series of compounds that inhibit tumor angiogenesis by targeting heparanase and MMP. Oligomannurarate sulfate (JG3), a novel oligosaccharide, was identified as a heparanase inhibitor. JG3 significantly inhibits tumor angiogenesis and metastasis, both in vitro and in vivo, by combating heparanase activity; in particular, this effect is mediated through the binding of JG3 to the KKDC and QPLK domains of heparanase. In addition, JG3 abolished heparanase-driven invasion, inhibited the release of basic fibroblast growth factor (bFGF) from the ECM and repressed subsequent angiogenesis27. By contrast, the antiangiogenic effects of quercetin were found to be associated with its downregulation of MMP-228. In fact, however, other studies reveal complicated molecular mechanisms involved in the antiangiogenic activity of quercetin, including its inhibition of the synthesis and accumulation of the HIF-1α protein; this inhibition reduces the production and secretion of VEGF, as discussed above29.

Other angiogenesis inhibitors

C9 and GLP exert antiangiogenic effects through relatively unique mechanisms of action (Table 1). C9 is a new microtubule-depolymerizing agent. In striking contrast to pseudolaric acid B, C9 elicits its antiangiogenic and vascular disrupting effects by inducing microtubule disassembly, the downregulation of Raf-MEK-ERK signaling and the reorganization of actin through the Rho/Rho kinase pathway24. GLP is a new type of polysaccharide isolated from the alga G longifolia. GLP causes obvious in vitro and in vivo antiangiogenic effects that are not associated with classical VEGF-VEGFR signaling. GLP decreases tissue factor at both the mRNA and protein levels, and this effect may be involved in the inhibition of angiogenesis by GLP30.

Our findings provide the following beneficial clues to facilitate the future discovery of angiogenesis inhibitors: (1) Natural products are an important source of angiogenesis inhibitors. In particular, through drug discovery guided by therapeutic experiences from long-established TCM practices, inhibitors of angiogenesis could be obtained that have new chemical structures and unique mechanisms of action; examples of these sorts of inhibitors include pseudolaric acid B, triptolide, 11,11′-dideoxyverticillin and shiraiachrome A (Table 1). The original medicinal materials producing these compounds have long been used to either treat various angiogenesis-related diseases, such as rheumatoid arthritis, microbial skin diseases and psoriasis7,21,23, or cause the early termination of pregnancies23. (2) Marine-derived compounds could be another important source of angiogenesis inhibitors, and certain compounds with specific types of chemical structures, such as saccharides, should be subjected to particular scrutiny; as demonstrated by JG327 and GLP30, these compounds may exhibit unique modes of antiangiogenic activity. (3) Chemical synthesis, particularly if it is based on rational designs and modifications, is a common, frequently necessary method of generating compounds with possible medical applications; this principle is demonstrated by AL3810 in this instance11,31 and is generally exemplified by a variety of clinically utilized drugs. (4) The compounds that we have discovered (Table 1) can be used as chemical models for further modification and optimization to improve their therapeutic potential as angiogenesis-inhibiting drugs for clinical use. (5) The known primary target(s) of certain compounds, such as pseudolaric acid B15,16,18 and MFTZ-117, are not necessarily related to their antiangiogenic effects, indicating that unknown and potentially novel mechanism(s) are involved in these effects; these mechanisms merit further investigation.

The discovery that the stability of HIF-1α protein is regulated by c-Jun based on mechanistic studies of the antiangiogenic activity of pseudolaric acid B

Pseudolaric acid B is a diterpenoid isolated from the root bark of Pseudolarix amabilis32. Pseudolaric acid B has been demonstrated to both elicit potent anticancer effects by depolymerizing microtubulin22,23 and circumvent tumor multidrug resistance22. Detailed structure-activity studies have revealed that the components of pseudolaric acid B that are essential to its anticancer activity include a hydrophobic group (-CO2 Me or -Me) at C-7, a Δ7 double bond, an acyloxy (OAc) at C-4,3 and a side chain with a conjugated double bond and a hydrophilic terminal group33. Based on its traditional use in Chinese folk medicine for facilitating the early termination of pregnancy, we first discovered and reported that its antiangiogenic activity occurred because it accelerates the proteasome-executed degradation of the HIF-1α protein18.

We previously reported that the activation of the transcription factor c-Jun plays a critical role in the circumvention of tumor multidrug resistance by salvicine34,35. Based on those earlier findings, we investigated the effect of pseudolaric acid B and found that it could also drive c-Jun phosphorylation16. During our attempts to correlate HIF-1α protein degradation with the c-Jun phosphorylation induced by pseudolaric acid B, a novel mechanism was revealed in which c-Jun in its non-phosphorylated form regulates the stability of the HIF-1α protein15,16 (Figure 2).

Figure 2
figure 2

The antiangiogenic mechanism of pseudolaric acid B. Solid or broken lines indicate the relationships between the linked factors with direct experimental evidence (solid lines) or with logical possibility (broken lines) in the case of pseudolaric acid B.

Full size image

HIF-1α is a transcription factor that drives neoangiogenesis by regulating the expression of various target genes, including the proangiogenic genes VEGF and VEGFR, in response to hypoxia during the growth of solid tumors36. The ubiquitination-mediated, proteasome-executed degradation constitutes a critical methanism of regulating the stability of the cellular HIF-1α protein37. HIF-1α can be hydroxylated by an oxygen-sensitive prolyl hydroxylase at the Pro402 and Pro564 residues within its oxygen-dependent degradation domain (ODD)38. This hydroxylation will promote the ubiquitination of HIF-1α at Lys532, a process that is effectively mediated by the ubiquitin ligase known as the Von Hippel-Lindau tumor suppressor (pVHL)39. This ubiquitination then, in turn, leads to the reduced stability of HIF-1α by facilitating its degradation through the 26S proteasome15 (Figure 2).

Our studies demonstrated that c-Jun binds to the ODD of HIF-1α protein, protecting HIF-1α from being ubiquitinated and thereby enhancing its stability by reducing its susceptibility to the proteasome-executed degradation15 (Figure 2). Notably, this binding requires the domains of c-Jun for DNA binding and heterodimerization but is independent of Ser63/73 phosphorylation. Further investigations have clarified that only non-phosphorylated c-Jun, ie, c-Jun without transcriptional activity, can bind to and protect HIF-1α15,16. Relatively constant total levels of c-Jun are generally maintained within the cell. Pseudolaric acid B causes the increased phosphorylation of c-Jun, thus reducing the proportion of c-Jun that is in the Ser63/73-non-phosphorylated form. Consequently, the quantity of c-Jun that is able to bind to HIF-1α decreases, impairing the ability of c-Jun to stabilize HIF-1α16 (Figure 2). Our findings constituted the first discovery of a function for the non-phosphorylated form of c-Jun; in accordance with these results, a function-converter model of c-Jun was proposed, in which Ser63/73 phosphorylation serves as a function converter that shifts c-Jun from its non-transcriptional functions to its transcriptional functions16.

The development of AL3810 as a clinically promising angiokinase inhibitor

Various PTKs are aberrantly activated during tumor progression. Several of these PTKs, including VEGFR, PDGFR, and FGFR1, have been demonstrated to contribute to tumor angiogenesis11,31. In recent years, we have discovered hundreds of compounds with inhibitory activities against different PTKs and have reported that 9 compounds exhibit potent antiangiogenic effects (Table 1). Of these 9 compounds, AL3810 (also designated as E-3810)31 demonstrates the greatest potential for clinical use.

AL3810 is a synthetic multitargeted PTK inhibitor that inhibits VEGFR1, VEGFR2, PDGFRα, PDGFRβ, and FGFR1 with IC50 values in the nanomolar range11,31. In endothelial cells, AL3810 can suppress the autophosphorylation of VEGFR2, PDGFRβ and FGFR1. AL3810 thus displays apparent antiangiogenic activity in all of the tested in vitro, ex vivo, and in vivo models (Figure 1). Even at millimolar concentrations, however, AL3810 demonstrates no cytotoxic effects on cancer cell lines. Nevertheless, AL3810 can elicit broad-spectrum in vivo antitumor activity in human kidney, pancreas and liver cancer xenograft models; its activity appears to make it more potent than several commercially available multitargeted PTK inhibitors, such as sorafenib and Sutent. Moreover, its antitumor activity is closely correlated with its antiangiogenic activity11,31.

Notably, it appears that tumors do not easily become resistant to AL3810, as xenografted tumors re-grown after the withdrawal of AL3810 demonstrate a response to a second cycle of AL3810 treatment that is similar to the response observed for the first treatment cycle. By contrast, tumors re-grown after the withdrawal of sunitinib treatment display reduced sensitivity to a second cycle of treatment with sunitinib. However, these sunitinib-resistant tumors remain sensitive to AL381031. Although this result requires further confirmation, particularly in the clinical context, and the mechanism underlying this result also must be clarified, its potential importance is obvious, particularly given that tumors are generally prone to becoming resistant to the PTK inhibitors that are currently in clinical use40.

The oral bioavailability of AL3810 is excellent, reaching 31% in rats11. Moreover, the concentration of AL3810 in different tissues is higher than the concentration of AL3810 in the plasma of tumor-bearing nude mice. In particular, a 51.7-fold increase in the concentration of AL3810 was detected in the examined tumoral tissues when compared with that in plasma, demonstrating the high affinity of AL3810 to tumors11. In addition, AL3810 has a relatively long terminal half-life of approximately 4 h; this extended half-life may help explain its persistent antitumor effects31. All of these results indicate that AL3810 possesses a favorable pharmacokinetic profile.

The prominent advantages of AL3810, which include its potent antiangiogenic effect, its broad spectrum of antitumor activity, its ability to potentially circumvent the drug resistance of tumors and its favorable pharmacokinetic profile, make it an excellent candidate for development as an anticancer drug. AL3810 has already entered into clinical trials in Europe to assess its potential use for this application, and its clinical trials in China will be launched shortly11,31.

Concluding remarks

As an aspect of international efforts to explore new antiangiogenic compounds, we investigated the inhibition of angiogenesis by 17 compounds. These compounds have diverse origins, structures, primary targets and mechanisms. Our findings provide new models for the further modification and optimization of antiangiogenic compounds, and new clues related to the examination of antiangiogenic mechanisms. Tumor drug resistance poses a challenge for the antiangiogenic agents that are in current use41. Therefore, one of the most urgent tasks in the future will be to both demonstrate the critical molecular mechanism(s) underlying the effects of AL3810 and accelerate the clinical development of AL3810 to meet the potential clinical need for agents that can circumvent this tumor drug resistance. The detailed dissection of the structure-effect relationships of certain compounds to analyze the relationships between their antiangiogenic mechanisms and their cytotoxicity and between their primary targets and their antiangiogenic mechanisms could be another important task to accomplish, as well; the elucidation of these relationships could generate new strategies for cancer therapy.