Sage Journals: Discover world-class research

Abstract

Background:

Previous studies have suggested that geldanamycin (GA) inhibits the replication of several viruses in vitro. Here, we aimed to synthesize and evaluate the antiviral activity of 17-amino-17-demethoxygeldanamycin derivatives.

Methods:

A series of 17-substituted and 17-,19-disubstituted GA derivatives were screened for antiviral activities against eight different viral strains, including herpesvirus, hepatitis virus, retrovirus and picornavirus.

Results:

Most of the tested compounds showed inhibitory activity against the viruses and showed reduced cytotoxicity in vitro as compared with the parent compound GA. In vivo efficacy evaluation results showed that compound 6 noticeably inhibited duckling hepatitis B virus DNA replication in duckling serum after oral administration. Viral rebound did not occur after termination of the treatment. The modified GA derivatives also showed median lethal dose values that were higher than that of the parent GA in mice intraperitoneally treated with the study compounds.

Conclusions:

Targeting heat-shock protein 90 could be a new antiviral approach that is not prone to the development of drug resistance. The 17-amino-17-demethoxygeldanamycin derivatives could be novel agents with potential antiviral activity.

Introduction

Viruses with high infection rates and rapid propagation can cause worldwide human and animal disease pandemics. With the emergence of previously unknown viruses and new virus variants in recent years, viral infection remains a major cause of human health problems. Moreover, the existing antiviral agents mainly target the key viral enzymes that are involved in the process of replication. At the same time, many viral enzyme mutations result in drug resistance. Some antiviral agents are also of limited use because of their existing toxic side effects. There is a crucial need for developing new agents with novel antiviral mechanisms and broad antiviral activities.

Heat-shock protein 90 (HSP90) is a cellular molecular chaperone that binds to multiple proteins and facilitates the proper refolding of these proteins. It increases the stability of the bound proteins and prevents their degradation through the ubiquitin–proteasome pathway. Previous studies have focused primarily on the stabilization of tumour cell oncogenic proteins by HSP90 [1,2]. Recent studies have found that some viral proteins can also directly or indirectly form chaperone complexes with HSP90 to assure stabilization and correct folding; therefore, HSP90 has been identified as an essential host factor required for stabilization of viral proteins. Moreover, attempts to introduce mutations into HSP90 have been difficult because HSP90 is a conserved cellular protein; thus, HSP90 has the potential to be developed as a novel target for broad-spectrum antiviral agents. Optimized HSP90 inhibitors should have broad-spectrum antiviral activity and should be able to reduce the possibility of developing drug resistance.

Geldanamycin (GA), an HSP90 inhibitor, is a naturally occurring benzoquinone ansamycin antibiotic that is isolated from the fermentation of Streptomyces hygroscopicus, and has been widely recognized for its potent activity against tumours [3 –5]. GA has also been frequently used to investigate the relationship between HSP90 activity and viral replication. Results from these studies have suggested that GA inhibits the replication of multiple viruses by abolishing the activity of HSP90. For example, by inhibiting HSP90 activity, GA could suppress hepatitis B virus (HBV) DNA replication and RNA packaging because of its specific inhibition of the interaction between the HBV reverse transcriptase and ɛ RNA [6,7]. GA also suppresses normal functions of viral proteins by preventing the formation of the complex comprising HSP90 and viral proteins; such complexes play an important role in, for example, hepatitis C virus (HCV) RNA replication and in the HIV type-1 (HIV-1) transcription process [8 –11]. Inhibition of HSP90 activity also causes inhibition of viral protein synthesis (herpesvirus and some negative-stranded RNA viruses) because it facilitates the degradation of viral RNA polymerases [12 –14]. These studies have also indicated that GA could inhibit viral replication by interfering with the chaperone-mediated process in viral protein stabilization and folding. Li et al. [15] and Geller et al. [16] have identified GA as a novel parent compound that exhibits broad-spectrum activities against DNA or RNA viruses in cell cultures and in animals. However, high toxicity and poor oral bioavailability are major drawbacks that limit its use in clinical settings. As previous studies have verified that derivatization of GA by substituting the 17-position of GA with an amino group has greatly decreased the toxicity while retaining the anti-tumour activity [17,18], we synthesized GA derivatives substituted with small molecule alkylamines that consist of heterocycles or strong polar groups, such as phosphates or carbohydrates, in the 17-position to improve both the pharmacokinetic properties and the safety profile of GA.

To our knowledge, this is the first report to describe the screening of GA derivatives as antiviral agents. The aim of the present study was to evaluate the antiviral activity of the synthesized 17-amino-17-demethoxygeldanamycin derivatives against a larger panel of viruses, including herpesvirus, hepatitis virus, retrovirus and picornavirus. On the basis of the antiviral activity in vitro, the selected compounds with good efficacy and their structural features were further investigated for their toxicity and antiviral activity in vivo.

Methods

Chemistry

Unless noted otherwise, all reagents and solvents were purchased from commercial sources and used without purification. GA was purchased from the Shanghai Institute of Pharmaceutical Industry (Shanghai, China) and the purity was >95% by HPLC analysis. ¹H NMR spectra were recorded in CDCl₃ or dimethylsulfoxide-d₆ (DMSO-d₆) solution with a Mercury-400 spectrometer (Varian, Inc., Palo Alto, CA, USA. High-resolution mass spectrometry (HRMS) data were obtained using an AutoSpec Ultima-Tof spectrometer (Waters, Milford, PA, USA) with an electrospray ionization (ESI) source. Melting points were determined using an X₆ microscope melting point apparatus without correction. Column chromatography was performed on silica gel and eluted with a 1–20% methanol gradient in dichloromethane.

All 17-amino-17-demethoxygeldanamycin compounds were synthesized from GA and the corresponding amines (Figure 1). In our previous study, compounds 1, 2, 4, 5, 9 and 10 were synthesized (G-ZS et al., unpublished data). Compounds 17-(piperidinyl-4′-yl-methylamino)-17-demethoxygeldanamycin (3), 17-[1′-(2″,3″-dihydroxy)propyl]amino-17-demethoxygeldanamycin (6), 17-[2′-(2″-hydroxyethoxy)ethyl] amino-17-demethoxygeldanamycin (7) and 17-(3′-phosphopropyl)amino-17-demethoxygeldanamycin (8) were synthesized using a similar method.

The purity of all GA derivatives in this study was >95%, which was achieved using the HLPC area normalization method with a C18 column (5 μm; 4.6×250 mm; Dikma Technologies, Inc., Lake Forest, CA, USA), eluted with water (containing 0.02 mol/l KH₂PO₄) and methanol at a flow rate of 1 ml/min, and the analyte was detected by ultraviolet absorption at 332 nm.

The general procedure for the synthesis of compounds 3, 6, 7 and 8 was as follows: the corresponding amine (0.2 mmol) was added to a solution of GA (0.1 mmol) and dichloromethane (10 ml; methanol or water was added when the amine was not dissolved), and triethylamine (0.23 mmol) was added to release the amine if the amine was in the salt form. The reaction mixture was kept away from light and stirred overnight at room temperature. After concentration in vacuo, the residue was dissolved in ethyl acetate. After washing sequentially with 1 M aqueous hydrochloric acid, water and brine, the organic phase was dried over anhydrous sodium sulfate, filtered and concentrated in vacuo. The residue was purified by flash chromatography on a silica gel, eluted with 1–20% methanol in dichloromethane to yield the product as a purple solid.

Compound 3 was synthesized from GA and 4-aminomethyl piperidine in dichloromethane. Yield: 73%; mp: 140–142°C. ¹H NMR (400 MHz, DMSO-d₆) δ (ppm): 0.63 (d, 3H, J=7.0 Hz), 0.80 (d, 3H, J=7.2 Hz), 0.82–1.08 (m, 4H), 1.24 (m, 2H), 1.50–1.58 (m, 4H), 1.83 (s, 3H), 2.31 (m, 2H), 2.42–2.59 (m, 5H), 2.84–3.29 (m, 7H), 3.38 (d, 1H, J=8.4 Hz), 4.36 (m, 2H), 4.86–4.99 (m, 4H), 5.22 (d, 1H, J=10.0 Hz), 5.41 (t, 1H, J=10.4 Hz), 6.42 (br, 2H), 6.55 (t, 1H, J=11.2 Hz), 6.90–6.96 (m, 3H), 7.12 (d, 1H, J=11.0 Hz), 7.7 (br, 1H). HRMS (ESI) m/z: 643.3704, calcd for C₃₄H₅₁N₄O₈ ([M+H]⁺) 643.3701.

Compound 6 was synthesized from GA and 3-amino-1,2-propanediol in mixed solvent of dichloromethane and methanol (3/1, v/v). Yield: 61%; mp: 140–142°C. ¹H NMR (400 MHz, CDCl₃) δ (ppm): 0.85–1.14 (m, 6H), 1.26–1.28 (m, 2H), 1.52 (s, 1H), 1.64–1.66 (m, 2H), 1.80 (s, 3H), 2.02 (s, 3H), 2.35 (t, J=9.6 Hz, 1H), 2.52–2.58 (m, 2H), 2.73 (br, 2H), 3.26 (s, 3H), 3.35 (s, 3H); 3.43–3.44 (m, 2H), 3.52–3.57 (m, 2H), 3.86 (br, 1H), 4.13 (m, 2H), 4.23–4.26 (m, 1H), 4.30 (d, 1H, J=9.6 Hz), 4.80 (br, 2H), 5.17(s, 1H), 5.79 (t, 1H, J=10.4 Hz), 5.90–5.92 (m, 1H), 6.56 (t, J=11.6 Hz, 1H), 6.95 (d, 1H, J=11.6 Hz), 7.24 (br, 1H). HRMS (ESI) m/z: 642.2990, calcd for C₃₁H₄₅N₃O₁₀Na ([M+Na]⁺) 690.2997.

Compound 7 was synthesized from GA and 2-(2′-aminoethoxy)ethanol in mixed solvent of dichloromethane and methanol (3/1, v/v). Yield 92%; mp: 145–147°C. ¹H NMR (400 MHz, CDCl₃) δ (ppm): 0.98–1.01 (m, 6H), 1.80 (m, 6H), 2.03 (s, 3H), 2.40 (m, 2H), 2.68–2.73 (m, 3H), 3.27 (s, 3H), 3.37 (s, 3H); 3.47–3.49 (m, 1H), 3.57 (m, 1H), 3.63–3.67 (m, 3H), 3.76–3.80 (m, 5H), 4.31 (d, J=10.0 Hz, 2H), 5.20 (s, 1H), 5.84–5.92 (m, 2H), 6.58–6.62 (m, 2H), 6.94 (br, 1H), 6.98 (d, 1H, J=8.4 Hz), 7.21 (s, 1H), 9.17 (br, 1H). HRMS (ESI) m/z: 656.3158, calcd for C₃₂H₄₇N₃O₁₀Na ([M+Na]⁺) 656.3153.

Figure 1.

Synthesis and chemical structures of geldanamycin derivatives 1–10

Compound 8 was synthesized from GA and 3-aminopropylphosphoric acid in the presence of triethylamine in mixed a solvent of dichloromethane, methanol and water (8/3/1, v/v/v). Yield 33%; mp: 143–145°C. ¹H NMR (400 MHz, DMSO-d₆) δ (ppm): 0.81 (d, J=11.2 Hz, 3H), 0.94 (m, 3H), 1.49–1.58 (m, 5H), 1.63 (s, 3H), 1.69–1.79 (m, 2H), 1.83–1.87 (m, 1H), 1.90 (s, 3H), 1.92 (d, J=9.2 Hz, 1H), 2.23–2.29 (m, 2H), 2.44–2.56 (t, J=9.2 Hz, 4H), 3.12–3.40 (m, 6H), 3.47–3.56 (m, 1H), 4.40 (d, J=7.6 Hz, 1H), 4.98 (s, 1H), 5.52 (d, J=9.6 Hz, 1H), 5.78 (br, 2H), 6.59 (m, 2H), 6.92 (s, 1H), 7.01 (br, 2H), 9.35 (s, 1H), 10.7 (br, 1H). HRMS (ESI) m/z: 690.2772, calcd for C₃₁H₄₆N₃O₁₁NaP ([M+Na]⁺) 690.2762.

Cells

African green monkey kidney (Vero) cells, human erythroleukaemia (HEL) cells and metallothionein T-4 (MT-4) cells were obtained from the Institute of Virology, Chinese Academy of Preventive Medicine (Beijing, China). Hepatocellular carcinoma (HepG2.2.15) cells were donated by Mount Sinai Medical Center (New York, NY, USA) and human liver cells (Huh7.5) were kindly provided by Vertex Pharmaceuticals, Inc., (Boston, MA, USA).

Viruses and plasmids

Human cytomegalovirus (HCMV; AD169 strain), herpes simplex virus type-1 (HSV-1; KOS strain), HSV type-2 (HSV-2; SAV strain), HIV-1 (strain IIIB) and Coxsackie viruses (Coxsackie virus B3 Nancy strain [Cox B3] and Coxsackie virus B6 Schmitt strain [Cox B6]) were all from the Institute of Virology, Chinese Academy of Preventive Medicine. The plasmid pFL-J6/JFH/JC1 with the complete chimeric HCV complementary DNA was provided by Vertex Pharmaceuticals, Inc. HCV virus stock was prepared as previously reported [19].

Reagents and animals

Lamivudine (3TC) was purchased from GlaxoSmithKline (Suzhou, China), zidovudine was purchased from Shanghai Desano Co. (Shanghai, China), aciclovir was purchased from Hubei Yitai Co. (Tianmen, China) and ribavirin was purchased from Xinxiang Chemial Pharmaceutical Co. (Xinxiang, China). Interferon-α2b was from Schering–Plough Co. (Brinny, Ireland). We synthesized valganciclovir from ganciclovir; the purity was >95% by HPLC analysis.

Kunming mice were purchased from the Chinese Academy of Medical Sciences (Beijing, China). Beijing ducklings were purchased from Beijing Nanyan Fowl, Inc. (Beijing, China). The animals were housed in standard laboratories at 22 ±1°C and had free access to food and water except for a short fasting period before intraperitoneal or oral drug administration. These experiments were approved by our institution's animal Ethics Committee (Institute of Medicinal Biotechnology, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China).

Cytotoxicity assays

The cytotoxicity of tested compounds in Vero, HEL and HepG2.2.15 cells were monitored using cytopathic effect (CPE) assays as described previously [20]. Vero and HEL cells (2.5×10⁴ cells/well) were cultured in 96-well plates. After 24 h, the monolayers were incubated in the presence of various concentrations of the test compounds. After 3 days of culture at 37°C and 5% CO₂, the cells were assayed by CPE. HepG2.2.15 cells (2.0×10⁴ cells/well) were cultured in 96-well plates. After 24 h, the monolayer cells were incubated in the presence of various concentrations of the test compounds for 4 days, then washed and a fresh medium containing the test compounds was added. After 8 days of incubation at 37°C and 5% CO₂, the cells were assayed by CPE. The 50% cytotoxic concentration (CC₅₀) was defined as the compound concentration that reduced the viability of untreated cell cultures by 50%.

The cytotoxicity in MT-4 and Huh7.5 cells were monitored by tetrazolium (MTT) assays [15]. Huh7.5 cells (1×10⁴ cells/well) were planted into 96-well plates. After 6 h, test compounds with various concentrations were added. Cell viability was detected using MTT assays at 96 h. MT-4 cells (2×10⁴ cells/well) were plated into a 96-well plate and incubated in the presence of various concentrations of the test compounds. After 4 days of culture at 37°C and 5% CO₂, the cell viability was quantified using MTT assays.

Anti-HBV activity assay in vitro

The anti-HBV activities of compounds together with a positive control drug were tested in HepG2.2.15 cells. The HepG2.2.15 cells were cultured in 96-well plates and treated with the test compounds at 37°C for 4 days, then washed and a fresh medium containing the test compounds was added. After 8 days of incubation at 37°C and 5% CO₂, the cells were harvested and intracellular DNA was extracted. The inhibition of the viral DNA in treated cells versus untreated cells was determined by dot blot hybridization.

Anti-HCV activity assay in vitro

Huh7.5 cells were cultured in 96-well plates at a density of 1×10⁴ cells/well. After 6 h of incubation, cells were infected with HCV virus stock (45 IU/cell) and treated concomitantly with tested compounds or interferon-α2b. At 96 h after inoculation in 5% CO₂ at 37°C, cells were harvested and intracellular total RNA was extracted using an RNeasy mini kit (Qiagen, Hilden, Germany). The HCV RNA was quantified directly with a one-step real-time reverse transciptase PCR kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol.

Anti-HIV-1 activity assay in vitro

The anti-HIV-1 activity of the test compounds on HIV-1 replication was measured by inhibition of HIV-1 p24 antigen in MT-4 cell cultures as described previously [21]. MT-4 cells (2×10⁵ cells/ml) infected with HIV-1 at 100 median tissue culture infective dose (TCID₅₀)/ml were cultured in 96-well plates and incubated in the presence of various concentrations of test compounds or a positive control drug. After 4 days of culture at 37°C and 5% CO₂, the culture supernatants were collected and assayed for their p24 antigen levels by ELISA.

Anti-HSV, Coxsackie virus and HCMV activity assays in vitro

Confluent Vero cells grown in 96-well microplates were infected with 100 TCID₅₀ HSV-1, HSV-2, Cox B3 or Cox B6 virus. HEL cells were infected with HCMV. After 1 h adsorption at 37°C, the monolayer cells were washed with phosphate-buffered saline and incubated at 37°C in the maintenance medium (minimum essential medium plus 2% fetal bovine serum) with or without different concentrations of the test compounds or a positive control drug. The CPE in cells was recorded when the CPE of the viral control group reached >75% and the antiviral activity of compound was determined using Reed–Muench analysis [22].

Acute toxicity tests in mice

Kunming mice (20 ±2 g) were divided into groups with eight mice per group. The control group received DMSO only. On the basis of the pre-experiment results, each of the test compounds dissolved in DMSO was administrated with three doses to three groups of animals. Mice were placed unrestrained in their cages after treatment and observed for 5 days to monitor behavioural changes and death. The median lethal dose (LD₅₀) was calculated using the Bliss method according the death number.

Anti-DHBV DNA activity in a DHBV infection model

One-day-old Beijing ducklings, inoculated intravenously with duckling hepatitis B virus (DHBV)-DNA-positive serum from Shanghai ducks (0.2 ml/duckling), were randomly divided into four groups (6 ducklings/group) for treatment. The virus control group was treated with 0.1% carboxymethylcellulose in water, the positive drug group was treated with 3TC (50 mg/kg) and the two test compound groups were treated with compound 1 (30 mg/kg) or compound 6 (30 mg/kg). Oral administration was started at day 7 after infection and was carried out twice daily for 10 days. Blood samples were taken and serum samples were prepared for analysis of DHBV DNA levels before treatment (T0), on days 5 (T5) and 10 (T10) during treatment and on day 3 (P3) after termination of treatment.

The detection of DHBV DNA levels in the serum samples were done using the dot blot hybridization method as described previously [23,24]. Briefly, 50 μl of duckling serum was spotted directly onto the nitrocellulose filters. After denaturation and fixation, the filters were hybridized with a DHBV ³²PDNA probe. Phosphorimaging dots were obtained on the films using autoradiography. Films were then scanned for the integrated optical density (IOD) of each dot and analysed using Gel-Pro Analyzer 3.1 (Media Cybernetics, Inc., Silver Spring, MD, USA). The IOD values of imaging dots were displayed as a marker of DHBV DNA levels.

In order to eliminate the background difference between groups, the relative ratio (IOD RR) between the post- and pretreatment result of each group was used to show the change in DHBV DNA levels. Lower IOD RR values represent a greater decrease in DHBV DNA levels.

Statistical analyses

The repeated measurement data from in vivo anti-DHBV DNA activity tests were analysed using the general linear models procedure in SPSS version 13.0 (SPSS, Inc., Chicago, IL, USA) for Windows. Multivariate analysis of variance was used to test the significance of the group effect. The criterion for significance was set at P<0.05.

Results

Eight viral strains, including herpesvirus (HCMV, HSV-1 and HSV-2), hepatitis virus (HBV and HCV), retrovirus (HIV-1) and picornavirus (Cox B3 and Cox B6), were used in the in vitro antiviral assay for the synthesized GA derivatives.

In vitro herpesvirus inhibitory activities

The antiviral activities of GA and its derivatives were tested in cultured HEL cells infected with HCMV and Vero cells infected with HSV. The activities of the GA derivatives against herpesvirus are shown in Table 1. All study compounds showed an inhibitory effect on the herpesvirus at test concentrations. Although the 50% inhibitory concentration (IC₅₀) values of all derivatives were higher than that of the parent compound GA, the selectivity index (SI) of some derivatives were above that of GA because of their reduced cytotoxicity. The SI of compounds 1 (SI=86) and 7 (SI=99) were higher than that of GA (SI=61) in the test against HCMV. Compound 9 (SI=60) as well as compounds 1 (SI=25) and 9 (SI=63) showed better activity among tested compounds, including GA in the anti-HSV-1 and anti-HSV-2 tests.

Table 1.

Anti-herpesvirus activity and cytotoxicity of GA derivatives

Compound	HEL CC₅₀, μM	HCMV IC₅₀, μM	SI	Vero CC₅₀, μM	HSV-1 IC₅₀, μM	SI	HSV-2 IC₅₀, μM	SI

Positive drug^a	116.2	0.006	17,913	>222	16.42	>12	30.81	>14
1	7.847	0.091	86	44.74	2.003	22	1.745	25
2	0.548	0.030	18	29.94	1.461	20	2.862	10
3	>31.11	3.453	>9	>31.11	27.62	>1	>31.11	–^b
4	0.776	0.213	3	19.92	1.598	12	4.932	4
5	2.032	0.095	21	20.45	1.492	13	1.476	13
6	26.90	1.000	26	170.7	13.29	12	21.54	7
7	14.17	0.142	99	175.0	8.457	20	8.883	19
8	17.29	3.324	5	>29.95	3.070	>10	10.18	>3
9	0.733	0.158	4	106.2	1.753	60	1.681	63
10	274.3	23.88	11	>274.3	158.4	>1	190.2	>1
GA	0.037	0.0006	61	21.74	0.499	43	0.891	24

The positive control drugs were valganciclovir or aciclovir for anti-human cytomegalovirus (HCMV) and anti-herpes simplex virus (HSV) tests, respectively.

The selectivity index (SI) could not be determined under the test conditions. CC₅₀, 50% cytotoxic concentration; GA, geldanamycin; HEL, human erythroleukaemia cells; HSV-1, herpes simplex virus type-1; HSV-2, herpes simplex virus type-2; IC₅₀, 50% inhibition concentration; Vero, African green monkey kidney cells.

As summarized in Table 1, the CC₅₀ values of the tested compounds were all significantly higher than that of GA, both in HEL and Vero cells: the CC₅₀ values of all GA derivatives in the HEL cells were >10-fold of that of GA (0.037 μM). Among the compounds, the cytotoxicity of compounds 6, 7 and 10 showed a significant decrease in cell cultures, and compound 10 (with disubstitutions at the 17- and 19-positions of GA) showed no toxicity even at the highest test concentration (274 μM) used in the cell lines. These data indicated that modifying GA with an amino group in the 17-position could decrease the cytotoxicity of GA derivatives in cell culture.

In vitro hepatitis virus inhibitory activities

The anti-HBV activity of GA and its derivatives was tested in HepG2.2.15 cells (Table 2). Compounds 1, 2, 6, 7 and 9 showed inhibitory activities on HBV replication under test concentrations, and compounds 1 and 2 showed IC₅₀ values similar to that of GA. Compounds 3, 4, 5, 8 and 10 were not active against HBV.

Table 2.

Anti-HBV and anti-HCV activity and cytotoxicity of GA derivatives

Compound	HepG2.2.15 CC₅₀, μM	HBV IC₅₀, μM	SI	Huh7.5 CC₅₀, μM	HCV IC₅₀, μM	SI

Positive drug^a	5,103	3.882	1,314	>6,000	0.52	>11,538
1	2.428	0.136	17	0.910	0.531	1
2	0.167	0.123	–^b	0.380	NA	–^b
3	27.83	NA	–^b	7.848	6.145	–^b
4	26.30	NA	–^b	0.768	NA	–^b
5	0.762	NA	–^b	0.222	0.103	2
6	47.20	2.598	18	1.533	NA	–^b
7	9.073	1.309	6.9	0.733	0.449	1
8	8.611	NA	–^b	0.726	NA	–^b
9	24.79	1.250	19	0.237	NA	–^b
10	91.47	NA	–^b	179.5	NA	–^b
GA	7.634	0.123	62	0.053	NA	–^b

The positive control drug was lamivudine or interferon-α2b for hepatitis B virus (HBV) or hepatitis C virus (HCV), respectively.

The selectivity index (SI) could not be determined under the test conditions. CC₅₀, 50% cytotoxic concentration; GA, geldanamycin; HepG2.2.15, hepatocellular carcinoma cells; Huh7.5, human liver cells; IC₅₀, 50% inhibition concentration; NA, not active under test condition.

The anti-HCV activity of GA and its derivatives was tested in Huh7.5 cells. The CC₅₀ values of all GA derivatives were higher than that of GA (Table 2), indicating that the cytotoxicity of the GA derivatives were lower than that of GA; however, the derivatives showed no significant anti-HCV activity under the test concentration.

In vitro HIV-1 virus inhibitory activities

Anti-HIV-1 activity of compounds 1–5 were tested in MT-4 cells. The results are shown in Table 3. All of the compounds inhibited HIV-1 replication. Compounds 1, 2 and 4 possessed inhibitory activity that was stronger than that of GA. Higher SI values of GA derivatives over GA were also obtained in MT-4 cells.

Table 3.

Anti-HIV-1 activity and cytotoxicity of GA derivatives

Index	AZT	1	2	3	4	5	GA

CC₅₀, μM	2.181	4.280	1.537	1.197	1.233	3.668	0.552
IC₅₀, μM	0.001	<0.015	<0.015	0.093	0.013	0.222	0.099
SI	2173	>282	>101	12	90	16	5

AZT, zidovudine; CC₅₀, 50% cytotoxic concentration; GA, geldanamycin; IC₅₀, 50% inhibition concentration; SI, selectivity index.

In vitro Coxsackie virus inhibitory activities

The lack of effective antiviral drugs against picornavirus led us to investigate the anti-Coxsackie virus activity of these GA derivatives (Table 4). Except for compounds 4 and 5, the others had lower toxicity than that of GA in Vero cells. Although none of derivatives was more potent than GA according to the IC₅₀ and SI values for either Cox B3 or Cox B6 virus, most compounds were superior to the positive control drug, ribavirin, in terms of antiviral efficacy. Ribavirin is currently used for clinical therapy against Coxsackie viral infection.

Table 4.

Anti-Coxsackie virus activity and cytotoxicity of GA derivatives

Compound	Vero CC₅₀, μM	Cox B3 IC₅₀, μM	SI	Cox B6 IC₅₀, μM	SI

Ribavirin	>6,142	1339	>4	1,958	>3.1
1	>30.35	3.840	>7	1.442	>21
2	>30.45	1.476	>20	5.861	>5
3	71.91	26.84	2	3.843	18
4	17.61	0.494	35	1.446	12
5	19.06	0.231	82	1.175	16
6	107.6	27.84	3.86	8.342	12
7	105.2	15.05	6.99	2.019	52
8	61.75	2.876	21.47	9.989	6
9	>28.74	4.786	>6	>1.179	–^a
10	158.4	>30.48	–^a	>30.48	–^a
GA	21.41	0.155	137	0.035	611

The selectivity index (SI) could not be determined under the test conditions. CC₅₀, 50% cell cytotoxic concentration; Cox B3, Coxsackie virus B3 Nancy strain; Cox B6, Coxsackie virus B6 Schmitt strain; GA, geldanamycin; IC₅₀, 50% inhibition concentration; SI, selectivity index; Vero, African green monkey kidney cells.

Acute toxicity in mice

To further evaluate the toxicity of GA derivatives, we investigated the acute toxicity of compounds 1, 4, 5, 6 and 9 in mice after a single dose by intraperitoneal administration. The results were compared with that of GA. In our study, mice often died 3 days after drug injection; otherwise, mice recovered to a normal state within 5 days. The LD₅₀ of tested compounds are shown in Figure 2. Among these compounds, GA was shown to be the most toxic with an LD₅₀ value of 17.25 mg/kg. The LD₅₀ values of compounds 1, 4, 5, 6 and 9 were 295.4, 122.8, 238.1, 203.8 and 35.48 mg/kg, respectively, indicating that all of the test compounds had lower toxicity than that of GA in mice. This was in agreement with the results from cytotoxicity assay.

Figure 2.

The acute toxicity of GA derivatives in mice

In vivo inhibitory activity of DHBV DNA replication in a DHBV infectious model

Based on the in vitro anti-HBV activity results and the relationships between chemical structure and antiviral activity as well as toxicity, compounds 1 and 6 were chosen for further evaluation of their in vivo efficacy against DHBV DNA replication. GA derivatives were administrated orally to ducklings.

Multivariate analyses of variance indicated that 3TC and compound 6 significantly decreased DHBV DNA levels in serum samples, it also revealed a greater decrease in DHBV DNA levels at T10 in both the 3TC (P<0.01) and compound 6 (P<0.01) groups (Figure 3A), as compared with that in the control group. However, at P3 (the third day after drug withdrawal), the DHBV DNA level of the 3TC group with a higher dose (50 mg/kg) rebounded to its pretreatment level. In clinical practice, the viral load of patients have been observed to frequently rebound to the level of preteatment after ther withdrawal of 3TC. The significant difference of DHBV DNA levels between 3TC and compound 6 at P3 suggested that compound 6 might have an antiviral mechanism that is different from that of 3TC. These results show that compound 6 was well absorbed after oral administration in contrast to the poor oral absorption of GA, and produced inhibitory effects on DHBV DNA replication in duckling serum without causing viral rebound. Although the group treated with compound 1 showed lower IOD RR values than that of the control group at three time points after administration, there was no significant statistical difference between compound 1 and the control group (Figure 3B).

Discussion

The toxicity of GA has been a marked obstacle for its development for clinical use. This has spurred pharmaceutical chemists to pay great attention to developing less toxic GA derivatives. 17-allylamino-17-demethoxygeldanamycin has been reported to be a less toxic HSP90 inhibitor for treatment of tumours and has been enrolled into multiple clinic trials [25]. In our study, the majority of tested compounds also appeared to be less cytotoxic than GA in the cell cultures; the LD₅₀ values of the selected six compounds were higher than that of GA in mice, giving further evidence of the low toxicity of these derivatives. We also found that the 17,19-disubstituted derivative (10) showed the greatest decrease in toxicity; however, it should be noted that the antiviral activity of compound 10 almost disappeared in the anti-Coxsackie and hepatitis virus assays, which suggests that the 19-substituent modification is not an ideal choice in terms of antiviral activity. The introduction of a group at the 19-position of GA might interfere with the binding of GA derivatives to HSP90 because of steric effects.

17-Substituted GA derivatives contained both stronger and weaker antiviral activity than GA, but all had a largely decreased toxicity. According to the crystal structure of the GA–HSP90 complex [26], the substituent in the 17-position of GA is exposed to the outside cavity of the HSP90 protein, which suggests that the difference of substituents at the 17-position of GA are not crucial for the interaction of the test compounds with the HSP90 protein. This observation also helps us to understand why there was no obvious regularity observed between antiviral activity and structural difference in the 17-position in this study. Noticeably, compounds 1, 2, 6, 7 and 9 demonstrated a potent antiviral activity against at least four of the eight tested viral strains. This led us to speculate that HSP90 could be a broad-spectrum antiviral target, and that GA derivatives have broad antiviral activity because they interfere with chaperone assistance in the protein folding and stabilizing processes of multiple viral proteins. Moreover, active GA derivatives 1, 2 and 6, possessed binding ability to HSP90 that was similar to that of GA in a fluorescence polarization assay (Additional file 1). This assay was based on the competition of fluorescein-5-isothiocyanate-labelled GA (GA–FITC) for binding to HSP90; the method used was similar to that described previously [27,28].

Figure 3.

Mean changes in DHBV DNA levels of DHBV-infected duckling serum during different treatments

In China, the high incidence of HBV infection and the large number of HBV carriers is a severe problem. With the expectation of exploiting the potency of HSP90 inhibitor(s) against HBV, we investigated in vivo inhibitory activity of compounds 1 and 6 against DHBV as they were promising candidates in vitro. The in vivo test results showed that compound 6 at a dose of 30 mg/kg was effective against DHBV infection in ducklings. The test also provided evidence of good oral absorption of compound 6, which was a great improvement over the parent compound GA.

Unexpectedly, GA derivatives showed weaker or even no activity against HCV in Huh7.5 cells infected with the full-length viral genome. According to our previous study, GA and compounds 5 and 9 possessed activities against HCV in GS4.3 HCV replicon cells (replicon cells containing HCV NS3-NS5B non-structural genes). This might have been caused by different expression levels of HSP90 in GS4.3 cells and Huh7.5 cells, or by different levels of HSP90 requirement for HCV replicon and full HCV replication, and needs to be investigated in further experiments. Similarly, influenza virus was originally included in our research as a negative-strand virus; however, the activity of GA and its derivatives against influenza virus A and B was not observed in multiple tests, although it has been reported that influenza virus replication could be inhibited by HSP90 inhibitor GA [29]. Further studies are needed to investigate this discrepancy.

In summary, the broad-spectrum antiviral activity, low toxicity and good bioavailability of 17-amino-17-demethoxygeldanamycin derivatives were presented in this work, in comparison with GA. In particular, compound 6 showed significant antiviral activity after oral administration in the DHBV infection model. Geller et al. [16] has reported that no drug resistance developed in poliovirus-infected animals during GA treatment and viruses isolated from control and GA-treated animals were equally sensitive to the inhibitor used in their study. These results demonstrated that targeting HSP90 provides a new antiviral approach that is not prone to the development of drug resistance. We propose that HSP90 could be a reliable broad-spectrum antiviral target. We also consider HSP90 inhibitors to be a new choice for antiviral agents and GA derivatives a group of new agents with potential for antiviral therapy.

Footnotes

Acknowledgements

This work was supported by the National Nature Science Foundation of China (30472076 and 30772600), the ‘863’ programme of the Ministry of Science and Technology of China (2006AA02Z414), and was partly supported by the National Major Science and Technology Project of China (‘Innovation and Development of New Drugs’; 2009ZX09301-003).

The authors declare no competing interests.

Additional file 1: A supplementary table showing the inhibitory activity of the test compounds on a GA–FITC–HSP90 interaction in a fluorescence polarization assay can be found at

References

Workman

. Overview: Translating HSP90 biology into HSP90 drugs. Curr Cancer Drug Targets 2003; 3:297–300.

Whitesell

Lindquist

. HSP90 and the chaperoning of cancer. Nat Rev Cancer 2005; 5:761–772.

Sasaki

Yasuda

Onodera

. Growth inhibition of virus transformed cells in vitro and antitumor activity in vivo of geldanamycin and its derivatives. J Antibiot (Tokyo) 1979; 32:849–851.

Supko

Hickman

Grever

Malspeis

. Preclinical pharmacologic evaluation of geldanamycin as an antitumor agent. Cancer Chemother Pharmacol 1995; 36:305–315.

Whitesell

Shifrin

Schwab

Neckers

. Benzoquinonoid ansamycins possess selective tumoricidal activity unrelated to src kinase inhibition. Cancer Res 1992; 52:1721–1728.

Wang

Zhang

. Host immunity influences disease progression and antiviral efficacy in humans infected with hepatitis B virus. Expert Rev Gastroenterol Hepatol 2009; 3:499–512.

Seeger

. HSP90 is required for the activity of a hepatitis B virus reverse transcriptase. Proc Natl Acad Sci U S A 1996; 93:1060–1064.

Moriishi

Matsuura

. Host factors involved in the replication of hepatitis C virus. Rev Med Virol 2007; 17:343–354.

Okamoto

Nishimura

Ichimura

. Hepatitis C virus RNA replication is regulated by FKBP8 and Hsp90. EMBO J 2006; 25:5015–5025.

10.

Waxman

Whitney

Pollok

Kuo

. Host cell factor requirement for hepatitis C virus enzyme maturation. Proc Natl Acad Sci U S A 2001; 98:13931–13935.

11.

O'Keeffe

Fong

Chen

Zhou

. Requirement for a kinase-specific chaperone pathway in the production of a Cdk9/cyclin T1 heterodimer responsible for P-TEFb-mediated Tat stimulation of HIV-1 transcription. J Biol Chem 2000; 275:279–287.

12.

Basha

Kitagawa

Uhara

Imazu

Uechi

Tanaka

. Geldanamycin, a potent and specific inhibitor of HSP90, inhibits gene expression and replication of human cytomegalovirus. Antivir Chem Chemothery 2005; 16:135–146.

13.

Connor

McKenzie

Parks

Lyles

. Antiviral activity and RNA polymerase degradation following HSP90 inhibition in a range of negative strand viruses. Virology 2007; 362:109–119.

14.

Burch

Weller

. Herpes simplex virus type 1 DNA polymerase requires the mammalian chaperone HSP90 for proper localization to the nucleus. J Virol 2005; 79:10740–10749.

15.

Tao

Liu

Jiang

. Geldanamycin, a ligand of heat shock protein 90, inhibits the replication of herpes simplex virus type 1 in vitro. Antimicrob Agents Chemother 2004; 48:867–872.

16.

Geller

Vignuzzi

Andino

Frydman

. Evolutionary constraints on chaperone-mediated folding provide an antiviral approach refractory to development of drug resistance. Genes Dev 2007; 21:195–205.

17.

Tian

Liu

Zhang

. Synthesis and biological activities of novel 17-aminogeldanamycin derivatives. Bioorg Med Chem 2004; 12:5317–5329.

18.

Rastelli

Tian

Wang

Myles

. Structure-based design of 7-carbamate analogs of geldanamycin. Bioorg Med Chem Lett 2005; 15:5016–5021.

19.

Lindenbach

Evans

Syder

. Complete replication of hepatitis C virus in cell culture. Science 2005; 309:623–626.

20.

Schmidtke

Schnittler

Jahn

Dahse

Stelzner

. A rapid assay for evaluation of antiviral activity against coxsackie virus B3, influenza virus A, and herpes simplex virus type 1. J Virol Methods 2001; 95:133–143.

21.

Tang

Peng

Liu

. Some new acyclic nucleotide analogues as antiviral prodrugs: Synthesis and bioactivities in vitro. Bioorg Med Chem Lett 2007; 17:6350–6353.

22.

Reed

Muench

. A simple method of estimating fifty per cent endpoints. Am J Hyg 1938; 27:493–497.

23.

Han

Xue

Zhao

. Antiviral therapeutic efficacy of foscarnet in hepatitis B virus infection. Antiviral Res 2005; 68:147–153.

24.

Han

Huang

Yang

Liu

. In vivo and in vitro anti-hepatitis B virus activity of total phenolics from Oenanthe javanica. J Ethnopharmacol 2008; 118:148–153.

25.

Banerji

Judson

Workman

. The clinical applications of heat shock protein inhibitors in cancer – present and future. Curr Cancer Drug Targets 2003; 3:385–390.

26.

Stebbins

Russo

Schneider

Rosen

Hartl

Pavletich

. Crystal structure of an HSP90–geldanamycin complex: Targeting of a protein chaperone by an antitumor agent. Cell 1997; 89:239–250.

27.

Kim

Felts

Llauger

. Development of a fluorescence polarization assay for the molecular chaperone Hsp90. J Biomol Screen 2004; 9:375–381.

28.

Moulick

Rodina

. High-throughput screening fluorescence polarization assay for tumor-specific Hsp90. J Biomol Screen 2007; 12:915–924.

29.

Chase

Deng

Fodor

. HSP90 inhibitors reduce influenza virus replication in cell culture. Virology 2008; 377:431–439.

Synthesis and Biological Evaluation of Heat-Shock Protein 90 Inhibitors: Geldanamycin Derivatives with Broad Antiviral Activities