Hypoxia-inducible factor (HIF-1) is a dimeric transcription factor consisting of an oxygen regulated α-component and a constitutively expressed β-component. At normal oxygen levels, HIF-1α is degraded via
the pVHL-mediated ubiquitin-proteosomal pathway. Under hypoxic conditions, HIF-1α rapidly accumulates and dimerizes with HIF-1β. This heterodimer binds to the DNA hypoxia-response element (HRE) and activates a diverse array of target genes.1
This pathway is particularly relevant to the cancer field because oxygen levels in tumors are commonly lower than in the surrounding tissues. Hypoxic cells are resistant to radiation damage and their distances from blood vessels reduce the potency of anti-cancer drugs. Hypoxia additionally promotes the up-regulation of genes involved in drug resistance. HIF-1 is directly responsible for the induction of numerous genes that are present at higher levels in cancer cells, in particular VEGF. The overexpression of HIF-1 has been related to the aggressiveness and vascularity of tumors, and mortality rate in patients. Despite the introduced difficulties in treating hypoxic tumors, the hypoxic environment found in tumor cells can be exploited for targeted therapy. One strategy to achieve this involves the identification of HIF-1 inhibitors as potential anti-cancer drugs.2
We recently reported a high-throughput cell-based HIF-1 mediated β-lactamase reporter gene assay. Upon screening a library of 73,000 compounds (PubChem AID:915 (http://pubchem.ncbi.nlm.nih.gov
)), several compounds were identified as novel inhibitors of the HIF-1 signaling pathway.3
Among these hits, quinazolin-4-one 1
(NCGC00056044) showed good drug-like properties and was selected for further exploration.
Three areas were selected for structure-activity relationship (SAR) studies: (1) substitution in area A; (2) piperazine region B; and (3) phenyl substitution in area C ().
Optimization plan for NCGC00056044 (1)
To facilitate our compound synthesis for the SAR study, we modified a reported method4
to remove the need for intermediate purification. In addition, a microwave reactor was used to accelerate the synthesis. Acylation of anthranilic acid 2
with chloroacetyl chloride gave 3,
which was treated with aniline 4
to afford chloride 5
(). The chloride was reacted with amine 6
to give compounds 1, 7-36
. All three steps were conducted in one-pot without the need for intermediate isolation. This protocol was carried out in a parallel fashion to prepare the analogues which were purified via HPLC.5
Reagents and conditions: (i) iPrNEt2, ACN, r.t.; (ii) ArNH2 (4), POCl3, MW 150 °C, 15 min; (iii) K2CO3, EtOH, MW 150 °C, 5 min; then amine 6, MW 150 °C, 10 min.
was prepared as described in . Reaction of 37
with 2-furoyl chloride, followed by a hydrolysis reaction yielded acid 38
. The desired 39
was obtained via
a microwave assisted one-pot three-component reaction of 38
, acid 2a
, and 2-ethoxyaniline.6
Reagents and conditions: (i) iPrNEt2, DCM, 2-furoyl chloride; (ii) LiOH; (iii) 2-ethoxyaniline, 2a, pyridine, MW 230 °C, 10 min.
describes the synthesis of the area C analogue 42. Nitro-reduction of 40 gave 41. Alkylation of the aniline nitrogen in 41 using ethyl iodide followed by a Boc-deprotection gave 42.
Reagents and conditions: (i) Na2S2O4; MW 100 °C, 10 min (ii) EtI, DMF, iPr2NEt, K2CO3, MW 150 °C, 15 min; (iii) DCM, TFA, r.t. 2 h
All analogues were evaluated in a cell-based HIF-1 mediated β-lactamase reporter gene assay under hypoxic conditions.7
Area A showed little tolerance for substitution (). The C-6 methoxy (7
), C-5 iodo (9
), and C-4 and C-5 dimethoxy (10
) substitutions were inactive. Compound 8
with a methyl group at C-6 was active, but it was 3-fold less potent than the original hit (1
). Considering these results, our efforts focused on the optimization of areas B and C ().
Modification at the R1 position*
Modification of piperazine region B is shown in . Acetylation of N-4 (11) resulted in similar activity to the hit compound (1), but capping the piperazine nitrogen with a benzamide (12) or ethyl carbamate (13) resulted in a loss of activity. N-4 methylation (14) or benzylation (15) resulted in a 2-fold and 64-fold loss of activity respectively. Ultimately, the most active compound was the unsubstituted N-4 analogue (16), which was about 5-fold more potent than 1. N-4 was critical for activity because when it was replaced with either a carbon (19) or oxygen (18), activity was lost. In fact, both piperazine nitrogens were important because replacement of N-1 with a carbon (39) also resulted in a 40-fold loss of activity. Finally, the piperazine ring was expanded to homopiperazine (17) and there was a slight loss in activity relative to 16, but this analogue was still more potent than 1.
SAR study for the piperazine region*
The modification of area C was explored in . The first set of compounds was based on piperazine scaffold A (, entries 1-12) and there was almost no tolerance for substitution. The only moderately successful analogue was 2-OMe (29), but even this was 8-fold less active than 1. Scaffold B presented a greater opportunity for SAR analysis (entries 13-20). Large alkoxy groups, such as benzyloxy (33), or isobutyloxy (34) at C-2 resulted in significant loss of activity in comparison with ethoxy (16). Moving the methoxy group from the 2 to 4 position resulted in a complete loss of activity (29 vs. 36). A dramatic substitution effect was observed at the 5 position. Replacement of the nitro group (32) with a CF3 (31) resulted in more than a 20-fold improvement in potency. Finally, by comparing 35, 16, and 42, the ethoxy group appeared to be better than ethoxythio or ethylamine at the C-2 position.
To confirm HIF-1α inhibition activity, compounds 18
were evaluated in a Western blot analysis.8
At 1 μM, 16
completely suppressed HIF-1α accumulation while 18
had no effect on the protein accumulation (). This result is in agreement with the compounds’ activities observed in the cell-based assay. However, compound 18
at 10 μM also inhibited HIF-1α protein accumulation. Stockwell and coworkers reported that these quinazolin-4-ones caused rapid death of human tumor cells (BJ-TERT/LT/ST/RASV12
RAS-RAF-MEK dependent signaling.9
Because Ras, a well known oncogene, has been shown to stimulate HIF-1α expression via the Raf/Mek/ERK pathway,10
it is possible that the activity of these quinazolin-4-ones against HIF-1α accumulation is via
the RAS signaling pathway.
Effect of compounds 16 and 18 on the accumulation of the HIF-1α protein under hypoxia conditions
In conclusion, we have identified a series of novel quinazolin-4-one HIF-1α inhibitors. A library synthesis and SAR studies revealed analogue 16 as the new lead, which was almost 5-fold more potent than the hit (1). The inhibition of HIF-1α was further confirmed in Western blot analysis. Detailed mechanistic studies and evaluation of these compounds as anti-cancer agents in rare types of cancer are currently under investigation and will be reported in due course.