Synthesis and evaluation of DAG-lactone derivatives with HIV-1 latency reversing activity

Abstract

Cells latently infected with human immunodeficiency virus type 1 (HIV-1) prevent people living with HIV-1 from obtaining a cure to the infectious disease. Latency reversing agents (LRAs) such as protein kinase C (PKC) activators and histone deacetylase (HDAC) inhibitors can reactivate cells latently infected with HIV-1. Several trials based on treatment with HDAC inhibitors alone, however, failed to reduce the number of latent HIV-1 reservoirs. Herein, we have focused on a diacylglycerol (DAG)-lactone derivative, YSE028 (1), which is a PKC activator with latency reversing activity and no significant cytotoxicity. Caspase-3 activation of YSE028 (1) led to cell apoptosis, specifically in HIV-1 latently infected cells. Structure-activity relationship studies of YSE028 (1) have produced several useful derivatives. Among these, compound 2 is approximately ten times more potent than YSE028 (1) in reactivation of cells latently infected with HIV-1. The activity of DAG-lactone derivatives was correlated with the binding affinity for PKC and the stability against esterase-mediated hydrolysis.

1. Introduction

Antiretroviral therapy (ART) using reverse transcriptase inhibitors [1], protease inhibitors [2] and/or integrase inhibitors [3,4] can prevent replication of the human immunodeficiency virus type 1 (HIV-1), which is a pathogen that causes acquired immune deficiency syndrome (AIDS) [5–7]. In addition, the HIV infectious disease has been converted from a fatal illness into a manageable chronic disease by introduction of combination of the aforementioned inhibitor drugs, and this treatment has been designated as combination antiretroviral therapy (cART). This therapy prolongs human lives and improves the quality of life of people living with HIV-1. There are, however, several serious drawbacks in otherwise successful and effective cART. First, cART cannot cure HIV infectious diseases. Although use of cART decreases HIV-1 particles in patient’s plasma to below detectable limits, discontinuation of drugs causes a rapid rebound of HIV-1. Accordingly, patients must use cART throughout their lifetimes, which is a significant burden. Second, in HIV-1 replication, inaccurate reverse transcription from DNA to RNA might lead to emergence of drug-resistant strains of HIV-1 [8]. Third, cells latently infected with HIV-1 cannot be eradicated by cART. It is considered that on average, one latently infected cell exists among one million cells in a patient [9]. For these reasons, development of efficient methods for removal of all latently infected cells is required to solve these problems and improve HIV cure.

One of the most promising strategies to eliminate HIV-1 latently infected cells is considered to be a “shock and kill” approach [10]. Use of latency reversing agents (LRAs) including histone deacetylase (HDAC) inhibitors [11], bromodomain containing protein 4 (BRD4) inhibitors [12], immune checkpoints, especially programmed cell death protein 1 (PD-1) inhibitors [13], and protein kinase C (PKC) activators [14], leads to HIV transcription, protein expression and virion production (the “shock” step). Although HIV particles are released into plasma by the use of LRA, cART can protect uninfected cells from becoming infected. Alternatively, reactivated cells can potentially die by virus-mediated cytotoxicity or be killed by the patient’s immune system (the “kill” step). To date, however, activation of cells latently infected by HIV-1 using HDAC inhibitors, including vorinostat and romidepsin (HIV-1 latency reversing activity was found to be 1.2 μM and 1.1 nM using U1 cells, respectively) [15], has not significantly reduced the number of HIV-1 reservoirs although upregulation of HIV-1 RNA expression has been detected [16–18]. To activate such cells, new repertoires are required that not only can reactivate the cells but also reduce latent HIV-1 reservoirs.

Here, PKC is classified into a family of serine/threonine kinases with 11 isozymes involving cellular signal transductions, including for example, proliferation, differentiation, migration and apoptosis [19,20]. Therefore, the PKC family is considered to possess important drug targets that include for example diabetes [21], cancer [22–24], heart failure [25] and Alzheimer’s disease [26–28]. In the mechanism of reactivation of cells latently infected with HIV-1 by PKC ligands, activated PKC induces a significant increase of the nuclear factor-kappa B (NF-κB), which accumulates from cytosols into the nucleus, binds to HIV long terminal repeat (LTR) and promotes transcription by RNA polymerase [29]. In our previous study, benzolactam-V8–310, ((2S, 5S)-5-(hydroxymethyl)-2-isopropyl-1-methyl-1,4,5,6-tetrahydrobenzo [e]- [1,4]diazocin-3(2H)-one), which has an eight-membered lactam ring and activates PKC [30], showed potent activity in reversal of HIV-1 latency (HIV-1 latency reversing activity was found to be 0.025 μM using ACH-2 cells) [31]. It has also been reported that other PKC activators such as prostratin (HIV-1 latency reversing activity was found to be 0.294 μM using ACH-2 cells) [14,31], bryostatin-1 (HIV-1 latency reversing activity was found to be 1.61 nM using J-Lat cells) [32,33], their derivatives [34], gnidimacrin (HIV-1 latency reversing activity was found to be 18.0 pM using U1 cells) and its derivatives [15,35,36] and activate cells latently infected with HIV-1 in vitro.

An endogenous PKC ligand, diacylglycerol (DAG), has low binding affinity for PKC possibly due to its linear and flexible structure. With a conformationally constrained 5-membered ring, derivatives of cyclic DAG (DAG-lactone) that have been developed improved the PKC binding affinity [37,38]. In our recent study, a DAG-lactone derivative, YSE028 (1) [39], caused reactivation in vitro of cells latently infected with HIV-1 and primary cells from people living with HIV-1, and then induced caspase-mediated apoptosis, specifically in cells latently infected with HIV-1 [40]. Therefore, in this study we chose to use YSE028 (1) as a lead compound and explore structure-activity relationship (SAR) studies of other DAG-lactone derivatives to find compounds with increased HIV-1 latency reversing activity.

2. Results and discussion

2.1. Molecular design

Based on an X-ray structure of the PKCδ C1b domain in the complex with phorbol 13-O-acetate (PDB code: 1PTR) [41] and the recently reported X-ray structures of the complexes with a DAG compound (1, 2-dioctanoyl-sn-glycerol) (PDB: 7L92) or a DAG-lactone derivative (AJH-836) (PDB: 7LF3) in the presence of micelles that mimic membranes [42], the binding mode of DAG-lactone derivatives with PKC is considered to form a cell membrane-ligand-PKC C1b domain ternary complex (Fig. 1). Previously, we discovered YSE028 ((2-(hydroxymethyl)-5-oxo-4-(propan-2-ylidene)tetrahydrofuran-2-yl 2-propylpentanoate)), a DAG-lactone derivative with an isopropylidene group at the α-alkylidene position (1), which induces rapid translocation of PKCδ and activates PKC rapidly in cells (Fig. 1) [39]. Furthermore, it is also reported that esterification of the ingenol core compound at C-3 position with the long, saturated/unsaturated or cyclic side chain improved its HIV-1 latency reversing activity via activation of several PKC isoforms [43]. In this situation, we performed SAR studies on acyl side chains of DAG-lactone derivatives containing the olefinic moiety with the aim of increasing the interaction between the side chains and the lipid bilayer of cell membranes.

Fig. 1.

The binding mode of the (R)-enantiomer YSE028A (1a) of DAG-lactone YSE028 (1) to the C1b domain of PKCδ, forming the cell membrane-ligand-PKC C1b domain ternary complex [44]. The crystal structure of the PKCδ C1b domain in complex with DAG-lactone AJH-836 (PDB: 7LF3) was used [42]. The (R)-forms of DAG-lactone derivatives are known as active isomers [37].

2.2. Chemistry

Several DAG-lactone derivatives (B) were obtained from a benzyl ether derivative (3) [45] in two steps (Scheme 1). Acylation of this lactone (3) was performed in one of two different ways using corresponding acid chlorides or carboxylic acids to yield acylated compounds (A). In the next step, the benzyl group was removed by BCl₃ to obtain the target DAG-lactone derivatives (B). Several carboxylic acids with symmetric branched chains (C), longer than three methylenes, were synthesized from ethyl malonate in three steps (Scheme 2) [46].

Scheme 1. General procedures for synthesis of DAG-lactone derivatives (B) from a lactone (3).

Scheme 2. A general procedure for the synthesis of several branched carboxylic acids (C).

2.3. SAR studies on evaluation of HIV-1 latency reversing activity in J-Lat 10.6 cells

To perform SAR studies on the activity of synthesized DAG-lactone derivatives, HIV-1 latency reversing activity was evaluated in J-Lat 10.6 cells [47], which are genomic modified latently infected HIV-1 cells capable of expressing green fluorescence protein (GFP) after activation by transcription (Table 1, Fig. 2). First, DAG-lactone derivatives (2–10) with a branched side chain longer or shorter than the 2-propylpentanoyl group of YSE028 (1) were designed and synthesized. As expected, all of the DAG-lactone derivatives with a branched side chain longer than the 2-propylpentanoyl group of YSE028 (1) had high activity, in this assay, possibly because of the hydrophobic interaction between cell membranes and the side chain of the compounds. The LogP values of compounds are possibly correlated with the strength of hydrophobic interaction with cell membranes. The DAG-lactone derivative with the most potent activity of latent HIV-1 activation was compound 2 with a 2-hexyloctanoyl group. Although DAG-lactone derivatives with long side chains and high LogP values might strongly interact with cell membranes, the effect on latent HIV-1 activation appears to be relatively moderate. Compound 9, bearing a 2-ethylbutanoyl group, and compound 10, with a 2-methylpropanoyl group, failed to show significant activity of latent HIV-1 activation, possibly due to their short side chains, and their consequently weak hydrophobic interaction with cell membranes. As a result, compounds bearing a branched side chain as long as a 2-hexyloctanoyl group have suitable LogP of ~5 and interact with cell membranes, and consequently potently activate latent HIV-1. Next, we designed and synthesized several DAG-lactone derivatives (11–14) bearing linear and tertiary alkyl side chains. All of these compounds, especially compound 11 notwithstanding its long side chain, failed to show significant activity. Compound 14 showed slight HIV-1 latency reversing activity because its short side chain could weakly interact with cell membranes (Fig. 2). The effect of cyclic aliphatic and aromatic side chain of DAG-lactone derivatives (15–18) was then investigated. These compounds with ring structures in their sidechains failed to display any HIV-1 latency reversing activity. Although HIV-1 latency reversing activity appeared to be correlated with LogP values (Table 1), compound 11 with a moderate LogP value failed to show significant HIV-1 latency reversing activity. DAG-lactone derivatives with branched and moderately long alkyl side chains and suitable hydrophobicity (LogP ~5) showed potent HIV-1 latency reversing activity. The CC₅₀ values of all of the synthesized compounds using Jurkat cells showed more than 10 μM, suggesting that these actions might be specific for HIV latently infected cells (Fig. S2). In addition, CD69 (a biomarker of PKC activation)-positive cells were measured by flow cytometry and obtained the data similar with the above GFP-marker (Fig. S3).

Table 1.

HIV-1 latency reversing activity and LogP values of synthesized DAG-lactone derivatives.

compound	R	EC50 [μM]a	LogPb
2		0.22 ± 0.03	5.5
4		0.31 ± 0.01	8.0
5		0.34 ± 0.04	7.1
6		0.38 ± 0.08	6.3
7		0.43 ± 0.01	4.6
8		0.82 ± 0.04	3.8
9		>10	2.1
10		>10	1.3
11		>10	3.7
12		>10	1.6
13		>10	0.10
14		>10	2.0
15		>10	2.4
16		>10	2.1
17		>10	3.0
18		>10	2.0
YSE028 (1)		2.0 ± 0.11	3.0

^aThe magnitude of reactivation induced by phorbol 12-myristate 13-acetate (PMA) was defined as 100%, and the concentration of each compound resulting in 50% reactivation (viral transcription) was used to define the EC₅₀ values.

^bCalculated using ChemDraw Professional 20.0.

Fig. 2.

Results from the flow cytometry analysis to measure HIV-1 latency reversing activity of the test compounds. J-Lat 10.6 cells were treated with different concentrations of YSE028 (1), DAG-lactone derivatives (A: 2, and 4–8; B: 9, 10, and 13; C: 11, 12, and 14–18), and 10 nM PMA. The change in the number of GFP-positive cells was analyzed after 24 h-incubation by flow cytometry. Data are shown as means ± standard deviations (SD) of at least two independent experiments.

2.4. Binding activity in the synthesized PKCδ C1b domain

To assess the relationship between HIV-1 latency reversing activity and acyl chain structures of the DAG-lactone derivatives, the binding affinity of each compound with the PKCδ C1b domain was determined with a fluorescence quenching assay using the DAG-lactone bearing a solvatochromic dye, 6-methoxynaphthalene (6 MN), as a probe (Table 2, Fig. 3) [48,49]. In this assay, the binding affinity with PKCδ of the each of the tested DAG-lactone derivatives was obtained as a normalized fluorescence intensity. As a result, compounds 2 and 4 bearing symmetric 2-hexyloctanoyl and 2-nonylundecanoyl groups, respectively, competed highly with the DAG-lactone bearing 6 MN and displayed more than five times higher binding affinity than YSE028 (1). The binding affinity of compound 11 was approximately two times higher than that of YSE028 (1). Although compounds 2 and 4 with high binding affinity for PKCδ showed significant HIV-1 latency reversing activity, compound 11 with moderate binding affinity failed to exhibit any HIV-1 latency reversing activity. Compounds 13 and 14 displayed no or little binding affinity, respectively, suggesting that methyl or t-butyl groups as side chains are too small to interact strongly with cell membranes. Therefore, compounds 13 and 14 failed to show significant HIV-1 latency reversing activity. A simple lactone derivative 19, which lost one pharmacophore moiety when binding to PKC [50], was investigated as a negative control. As expected, compound 19 failed to show binding affinity for PKCδ.

Table 2.

Binding affinity of synthesized DAG-lactone derivatives with the PKCδ C1b domain.

compound	IC50 [μM]
2	0.69
4	0.56
11	2.2
13	-a
14	≫10
	-a
19
YSE028 (1)	4.0
	0.25

^aNo inhibition of the probe binding to the PKCδ C1b domain.

Fig. 3.

Results from fluorescence quenching assays, which measured the ability of the synthesized DAG-lactone derivatives to compete with DAG-lactone bearing 6-methoxynaphthalene (6 MN) for binding to the PKCδ C1b domain. The X axis represents inhibitor concentrations (log M) and the Y axis represents relative probe binding based on the fluorescence intensity (Ex: 355 nm, Em: 405 nm) of no inhibitor (100%) and blank (no probe/phosphatidylserine, 0%). Data points represent single experiment and fit using non-linear regression in GraphPad Prism 9.

2.5. Stability against esterase-mediated hydrolysis

Ester moieties of DAG-lactone derivatives tend to be hydrolyzed by endogenous esterases [51,52]. The derivatives fail to show any significant biological activity owing to their loss of one pharmacophore by hydrolysis. In order to discover the reason for the absence of significant HIV-1 latency reversing activity of compound 11, which possesses moderate lipophilicity and binding affinity for the PKCδ C1b domain, we tested its stability against esterase-mediated hydrolysis using porcine liver esterase (PLE). Fig. 4 shows a plot of the residual rates normalized at 0 h’s treatment of four compounds YSE028 (1), 2, 11 and 14, which were determined by HPLC areas after treatment for 1, 3, 6 and 24 h. In this evaluation, compound 2 with the most effective HIV-1 latency reversing activity was in almost as effective as YSE028 (1). Notably, compound 11 with a decanoyl side chain was highly susceptible to esterase-mediated hydrolysis and consequently failed to show significant any HIV-1 latency reversing activity although it has a moderate LogP value and good binding affinity for PKCδ. Compound 14 with a t-butyl group was most stable against this esterase possibly because of steric hindrance.

Fig. 4.

Evaluation of esterase-mediated hydrolysis of DAG-lactone derivatives YSE028 (1), 2, 11 and 14.

3. Conclusion

Although AIDS caused by HIV-1 has been changed from a fatal illness into a manageable disease by the introduction of cART, HIV-1 cannot be removed completely from infected people. This study focused on a “shock and kill” strategy, and SAR studies on DAG-lactone derivatives as LRAs were performed. DAG-lactone derivatives have been used as drug candidates targeting cancer [22,23] and Alzheimer’s disease [26] prior to their use as an HIV-1 therapy [40,53–55]. We discovered a potent DAG-lactone (2) with HIV-1 latency reversing activity approximately ten times higher than that of a lead compound YSE028 (1). It was found that DAG-lactone derivatives bearing branched alkyl side chains with moderate length have suitable hydrophobicity, potent HIV-1 latency reversing activity and binding affinity for PKC, and that DAG-lactone derivatives possessing alkyl side chains with a degree of steric bulk have acceptable stability against esterase-mediated hydrolysis. These data assist the design of DAG-lactone derivatives as PKC activators, which could be useful for HIV treatment. It is desirable to use DAG-lactone derivatives in a combination with anti-HIV drugs as well as with other LRAs.

4. Experimental section

4.1. General procedures

All reactions utilizing air- or moisture-sensitive reagents were performed in dried glassware under an atmosphere of nitrogen or argon, and commercially supplied solvents and reagents purchased from Sigma-Aldrich (USA), Tokyo Chemical Industry Co., Ltd. (TCI, Japan), FUJIFILM Wako Pure Chemical Corporation (Japan), Kanto Chemical Co., Inc. (Japan), Nacalai Tesque, Inc. (Japan), Watanabe Chemical Industries, Ltd. (Japan) and Combi-Blocks, Inc. (USA) which were used without further purification unless otherwise noted. ¹H NMR (400 MHz or 500 MHz) and ¹³C NMR (100 MHz or 125 MHz) spectra were recorded using a Bruker AVANCE III 400 spectrometer, Bruker AVANCE 500 spectrometer (Bruker, USA), and JNM-ECA500 (JEOL, Japan). Coupling constants are reported in Hertz, and peak shifts are reported in ppm relative to CDCl₃ (¹H 7.26 ppm, ¹³C 77.16 ppm), MeOD (¹H 3.31 ppm, ¹³C 49.00 ppm). Low- and high-resolution mass spectra were recorded on a Bruker Daltonics micrOTOF focus in the positive and negative detection mode. Infrared (IR) spectra were recorded on a JASCO FT/IR 4100, and are reported in wavenumbers (cm⁻¹). Thin-layer chromatography (TLC) was performed on Merck 60F₂₅₄ precoated silica gel plates and was visualized by fluorescence quenching under UV light and by staining with phosphomolybdic acid, p-anisaldehyde or basic potassium permanganate, respectively. Flash column chromatography was carried out with silica gel 60 N (Kanto Chemical Co., Inc.) or automatic silica gel flash column chromatography system (Isolera One, Biotage, Sweden) and Pure C-815 (Buchi, Switzerland). For analytical HPLC in the experiments with esterase-mediated hydrolysis, a Cosmosil 5C₁₈-ARII column (4.6 × 250 mm, Nacalai Tesque, Inc., Kyoto, Japan) was employed with a linear gradient of MeCN containing 0.1% (v/v) trifluoroacetic acid (TFA) (Solvent B) in H₂O containing 0.1% (v/v) TFA (Solvent A) at a flow rate of 1 cm³ min⁻¹ on a JASCO PU-2089 plus (Jasco Corporation, Ltd., Tokyo, Japan), and eluted products were detected by UV at 245 nm using JASCO MD-2018 plus. A mode of linear gradient of B in A + B, 0–95%, was performed over 40 min.

The synthetic methods leading to compounds are described in Schemes 1 and 2. The purity of the final compounds as measured by NMR was >95%. Experimental procedures including characterization data are provided in the Supporting Information.

4.2. Flow cytometry analysis for evaluation of the HIV-1 latency reversing activity and activation marker of PKC with LRAs [31,40]

The reactivation of HIV-1 from latently infected cells was determined by changes in intracellular green fluorescent protein (GFP) expression. J-Lat 10.6 cells were placed in 48-well plates with 2 × 10⁵ cells/well and incubated with different drug concentrations for 24 h. The data obtained by a representative gating strategy are shown in Fig. S1. The determination of CD69 activation and cytotoxicity marker staining were performed as described previously [31,56]. Cells were incubated with Ghost dye 780 (TONBO Biosciences, San Diego, CA, USA) for 30 min 4 °C. The cells were then stained with PE anti-human CD69 (FN50) mAb (BioLegend, San Diego, CA, USA) for 30 min on ice. Then, cells were fixed with 1% paraformaldehyde/PBS for 20 min RT and analyzed using BD FACSVerse (BD Biosciences, San Jose, CA, USA) and CytoFLEX (Beckman Coulter, Brea, CA, USA). The collected data were analyzed by FlowJo software (Tree Star, San Carlos, CA, USA).

4.3. Fluorescence quenching assays [48] using synthesized PKCδ C1b domain (231–281)

A phosphatidylserine (PS) solution (500 μg/mL) was prepared as described below, then 50 μL of 10 mg/mL PS in CHCl₃ (Avanti Polar Lipids 840032C, Sigma-Aldrich, Cat#: 840032C) was transferred to a 2 mL micro tube and dried under a stream of air. One mL of the assay buffer (50 mM Tris·HCl, 150 mM NaCl, 0.1 mM ZnCl₂, pH 7.4) was added to the residue and the mixture was sonicated with a probe-type sonicator in 5 s bursts × 3. Folded PKCδ C1b domain (231–281) was prepared according to the previous report [57]. Assay solutions (100 μL) containing 200 nM 6-methoxynaphthalene (6 MN) DAG lactone [48] as a probe and 100 μg/mL PS in the assay buffer (50 mM Tris·HCl, 150 mM NaCl, 0.1 mM ZnCl₂, pH 7.4) were prepared with 200 nM PKCδ C1b domain (231–281) or without PKCδ C1b domain (231–281) as a 0% binding control. Test compounds were serially diluted to generate 100x working dilutions in DMSO. To each well of a 96-well plate (Thermo Fisher Scientific, Cat#: 165305), 1 μL of 100x test compound solution was added (0% binding controls received 1 μL of DMSO). A total of 100 μL of the probe/PS/peptide solution (or probe/PS as 0% binding control) was added to the corresponding wells, and the plate was allowed to equilibrate at room temperature for 5 min with shaking. The fluorescence intensity was read using an ARVO MX plate reader (PerkinElmer, Waltham, MA, USA) with 355 nm excitation and 405 nm emission. The fluorescence intensity values were normalized to 100% (no inhibitor) and 0% binding (no peptide) controls. Normalized values were plotted versus concentration and analyzed using non-linear regression in GraphPad Prism 9 [log(inhibitor) vs response – variable slope (four parameter) model].

4.4. Esterase-mediated hydrolysis [58]

The conversion of DAG-lactone derivatives to hydrolyzed lactones by treatment with esterase was evaluated using porcine liver esterase (PLE, 200 KU/129 mL, Roche Diagnostics, Indianapolis, IN, USA) in HEPES buffer (50 mM) at pH 7.2. Compounds were dissolved in DMSO at 100 mM, diluted to a final concentration of 1 mM with HEPES buffer and DMSO and addition of a 0.01 U solution of PLE in HEPES buffer. The samples were incubated at 37 °C and at different time-points, 100 μL aliquots were taken, added to 100 μL MeCN containing 0.1% (v/v) TFA. The solution was subsequently analyzed by HPLC as described in the General procedures.

Abbreviations

ART

antiretroviral therapy

cART

combination ART

LRA

latency reversing agent

BRD4

bromodomain containing protein 4

PD-1

programmed cell death protein 1

LTR

long terminal repeat

DAG

diacylglycerol

DMAP

4-dimethylaminopyridine

EDCI

1-(3-dimethylaminopropyl)-3-ethylcarbodiimide

THF

tetrahydrofuran

DMF

dimethylformamide

DMSO

dimethyl sulfoxide

MN

methoxynaphthalene

PLE

porcine liver esterase

TFA

trifluoroacetic acid

PS

phosphatidylserine

Data availability

No data was used for the research described in the article.