NOVEL DAG-LACTONE BASED VAV1 INHIBITORS SHOW ANTI-PROLIFERATIVE ACTIVITY IN PANCREATIC CANCER MODELS

Abstract

Targeting C1 domains is a promising strategy for modulating the activity of signaling proteins driving tumor growth and progression. While most small molecules developed to date have focused on typical C1 domains, the development of regulators targeting atypical C1 domains remains underexplored. Herein, we describe the design and synthesis of novel cationic diacylglycerol DAG-lactones to efficiently interact with the negatively charged residues present in the atypical C1 domain of Vav1, a guanine nucleotide exchange factor playing a critical role in tumor development, including pancreatic cancer. We evaluated the therapeutic potential of this new family of compounds using models from this dismal condition where Vav1 is aberrantly expressed. Treatment of cultured pancreatic tumor cells with sn-1 cationic DAG-lactones inhibited proliferation of Vav1-expressing cells while Vav1-negative cells showed no response. Additionally, we demonstrated that these compounds inhibited growth of patient-derived organoids models of pancreatic cancer. These findings underscore the translational value of these cationic DAG-lactones for pancreatic cancer patients expressing Vav1 and serve as foundation for future approaches targeting atypical C1 domain-containing signaling proteins.

Graphical Abstract

New cationic DAG-lactones were designed for specifically binding the atypical C1 domain of Vav1. Compound AB19 with a 14-carbon alkyl chain showed the most pronounced effect inhibiting proliferation of Vav1-expressing cells. Additionally, this compound inhibited growth of patient-derived organoids models of pancreatic cancer. These data highlight the use of sn-1 cationic DAG-lactones as therapeutic agents for Vav1-positive pancreatic tumors.

Introduction

C1 domains are structural units discovered as lipids binding modules present in kinases and non-kinases proteins.^[1,2] It has been extensively reported that C1 domains are essential for the activity of these proteins by controlling intracellular signalling; thus, representing suitable targets to modulate disease conditions driven by such domains-containing proteins. C1 domains are approximately 50 amino acids in length that were originally discovered as the sn-1,2-diacylglycerol (DAG) binding modules in protein kinase C family members (PKCs).^[3] Subsequent characterization revealed that these domains are functionally heterogeneous with those binding DAG and their potent analogs, the phorbol esters, termed as “typical” while those possessing varying levels of sequence and structural homology but failing to bind DAG and analogs, termed as “atypical”.^[4,5] A large and diverse number of high affinity analogues targeting typical C1 domains have been reported in the past two decades.^[6–9] In contrast, only a few reports have been published describing compounds targeting atypical C1 domains.^[10,11]

Among these motifs, we selected Vav1 atypical C1 (Vav1C1) as model to develop inhibitors of this signaling domain. The main reasons underlying this selection was that first Vav1 functions as a critical oncoprotein involved in the pathogenesis of several malignancies, notably pancreatic cancer—a highly lethal disease anticipated to become the second leading cause of cancer-related mortality within the next decade.^[12–18] Second, Vav1C1 stabilizes the conformation of a critical helix, a structural feature essential for the displacement of guanine nucleotides from target GTPases (e.g., Rho and Rac1)^[13,17] which it is central for its guanine nucleotide exchange factor (GEF) activity. Importantly, most studies indicate that the GEF activity of Vav1 is indispensable for its role in tumorigenesis.^[12,19] And third, structural analyses revealed that Vav1C1 preserves a three-dimensional architecture closely resembling that of typical C1 domains and harbors a solvent-exposed cavity, thus providing a potential binding site for small-molecule ligands.^[20]

It has been recently hypothesized that the presence of five unique hydrophilic residues^[10,11] along the binding pocket disrupts the lipophilic surface of Vav1C1 and is responsible for the loss of binding activity towards DAG-like small molecules.^[5,19–22] Site-directed mutagenesis replacing these crucial residues in Vav1C1 domain with the corresponding residues present in C1 typical domain of PKCδ, resulted in almost complete recovery of “DAG-like” ligands binding affinity.^[19] Diacylglycerol lactones (DAG-lactones), which are rigidified structures derived from the endogenous ligand DAG, represent a valuable platform used to synthesize an invaluable number of selective ligands that bind with high affinities to diverse typical C1 domains-containing proteins.^[11,23–25] However, only one report in the literature describes the design of DAG-lactone derivatives to tackle the atypical C1 domain of Vav1, by including a positively charged side chain at the sn-2 position of the DAG-lactone to interact with the negatively charged glutamic acid (Glu) residues in the protein.^[10] Although these ligands displayed modest potency and selectivity for Vav1, the findings supported the notion that targeting Vav1C1 was a suitable strategy to block this GEF and provided important proof of principle. Here, we report the design and synthesis of a new family of DAG-lactones introducing cationic substitutions at the sn-1 position to achieve efficient interaction with the negatively charged residues in Vav1C1. We demonstrated that these new positively charged DAG-lactones are potent inhibitors in patients-derived lines and organoids models of human pancreatic cancer providing strong evidence of the translational value of these molecules. Our findings provide a first step in assessing the potential for the design of custom targeted molecules impacting Vav1 protein activity, which could have implications in tumour development, making this strategy effective and holding promise for further development.

Results and Discussion

CATIONIC DAG-LACTONES DESIGN AND CHEMICAL SYNTHESIS

To enhance the selectivity and efficacy in targeting atypical C1 domains we initially sought to design guanidinium-containing cationic DAG-lactones. Docking analysis of these compounds with Vav1C1 domain predicted this group would be effective in establishing effective interactions (Figure 1a, a1 and a2). However, all attempts to synthesize them were unsuccessful, resulting in unstable compounds. Nucleophilicity of the guanidinium group was problematic as lactamization was observed during analysis of the complex reaction mixtures. Next, we explored alternative synthetic analogues involving the synthesis of N-alkylpyridinium DAG-lactones (Figure 1a, a3). Despite valuable progress in synthesizing several derivatives, we encountered significant challenges to obtain them. The presence of the ester moiety at sn-1 position was incompatible with the conditions required to remove various protecting functional groups in the final step of the synthesis, which forced us to change the strategy to design alternative cationic DAG-lactones (see SI for experimental details and discussion).

Figure 1.

Cationic DAG-lactones synthesized. a) a1-alkyl guanidinium-containing derivatives: n = 1, 2, 3; a2-aromatic guanidinium compounds at ortho, meta and para positions; a3-pyridinium lactones: n = 2, 6, 12. b) N-alkylated pyridinyloxy DAG-lactones at ortho, meta and para positions: n = 2, 6, 12. Docking studies were performed with the alkyl chains truncated to a methyl group.

Given these challenges, we generated a new ester moiety bioisostere replacement with heterocyclic rings at sn-1 position. A set of N-alkylated pyridinyloxy DAG-lactones, which lacks the typical sn-1 ester moiety, were designed and the potential interaction of these ligands with Vav1C1 was modeled using docking, conformational searching, and association energy calculations. This suggested that the lactone core retained the capacity for effective binding, and that the positively-charge pyridinyl groups could form energetically favorable salt bridging to the negatively charged Glu residues in the Vav1C1 binding site (Figure 1b). According to calculations all three isomers (ortho, meta, para) were predicted to bind, but the 3-pyridinyl isomer was the most conformationally and energetically favorable, due to stronger non-bonded interactions and lower ligand strain.

We approached the synthesis of the new cationic DAG-lactones coupling the pyridine nucleus directly to the known lactone 1^[25] through an S_N2 reaction, introducing an iodine atom as good leaving group (compound 2, Scheme 1) in excellent yield.^[6] 3- and 4-Hydroxypyridines were then used as nucleophiles in basic media rendering compounds 3 and 7 in good yields but the strategy did not work for ortho-regioisomer 10 that was alternatively obtained via S_NAr reaction using 2-fluroropyridine, starting from 1.^[26] To perform the N-alkylation reaction, pyridine derivatives 3, 7 and 10 were treated with different electrophiles (butyl-, octyl- and tetradecanyl bromide) at elevated temperature, in a sealed tube. O-benzyl protected pyridinium cationic lactones were obtained in high yields from 3 and 7. Although considerable efforts were made, the preparation of pyridinium derivatives from 10 could not be accomplished, probably due to steric hindrance. BCl₃-mediated benzyl deprotection resulted in the successful syntheses of the first charged sn-1 DAG-lactone derivatives in high yields (5a-5c; 9a and 9c). These compounds were obtained as the corresponding bromide salts as they were synthesized using alkyl bromides for the quaternization step.

Scheme 1.

Synthesis of charged sn-1 DAG-lactones. Reagents and conditions: a) I₂, PPh₃, imidazole, toluene, reflux; b) 3-hydroxypyridine, K₂CO₃, DMF, 130 °C; c) 4-hydroxypyridine, K₂CO₃, DMF, 130 °C; d) 2-fluoropyridine, KHMDS, DMF, 0 °C; e) RBr (R= butyl, octyl, tetradecanyl), CH₃CN, vial, 130 °C; f) BCl₃, DCM, −78 °C.

sn-1 DAG-LACTONES SHOW ANTI-TUMORAL ACTIVITY IN PANCREATIC CANCER MODELS

Initially, we sought to determine the effect of the synthesized charged compounds on the viability of pancreatic cancer line expressing Vav1 (MiaPaCa2) at five different concentrations for each compound (Figure 2A). Compounds 5a, 5b, 6 and 9a exhibited minimal impact on cell growth at any concentration. Starting at 1 μM, compounds 5c (hereinafter AB176) and 9c (hereinafter AB19) containing a 14-carbon alkyl chain, had the most substantial effect with a reduction in cell numbers of 80–90% at 1 μM relative to the control (Figure 2A). These findings were validated by measuring the response in additional pancreatic cancer cell lines in the presence of charged DAG-lactone AB19 at 0.1 μM and 1 μM after 4 days of treatment. Cell lines included four pancreatic cancer lines, CAPAN-2, MiaPaCa2, CFPAC, ASPC1 and one “normal” immortalized pancreas epithelial cell line, HPNE. All Vav1 positive cells CAPAN-2, MiaPaCa2, and CFPAC showed a reduction in cell numbers relative to the control at 1 μM (Figure 2C). The Vav1 negative ASPC-1 and HPNE showed no response. Next, we evaluated the ability to inhibit Vav1 expression using compound AB19. Protein expression was measured in the pancreatic cancer cell line, CAPAN-2, following treatment with 0.01 μM and 0.1 μM of AB19, showing a significant reduction in Vav1 expression relative to the control for both concentrations (Figure 2B). We then assessed the downstream effect of Vav1 inhibition by measuring ERK phosphorylation^[27,28] following treatment with compound AB19 in CAPAN-2 cells. A significant reduction in ERK phosphorylation was observed at both concentrations following Vav1 inhibition (Figure 2B).

Figure 2.

A) Cell count in MiaPaca2 cells assayed by cell imager (Celigo). Data are presented as the mean ± SD of 3 independent experiments performed in triplicate. *P<0.05. B) Expression of VAV1 in CAPAN-2 cells detected by Western blot; α-Tubulin used as housekeeping control. Data are presented as the mean ± SD of 3 independent experiments performed in triplicate. *P<0.05. Expression of phospho-ERK and total ERK in CAPAN-2 cells detected using Western blot; Vinculin used as housekeeping control. Data are presented as the mean ± SD of 3 independent experiments performed in triplicate. *P<0.05. Both experiments performed on compound 9c (AB19). C) Proliferation of CAPAN-2, MiaPaca2, CFPAC, ASPC-1 and HPNE cells treated with compound 9c (AB19). Data are presented as the mean ± SD of at least 3 independent experiments performed in triplicate. *P<0.05.

Next, we examined the therapeutic effect of compound AB19 in pancreatic cancer patient-derived organoids (PDOs). These models are emerging as the tool of choice to determine therapeutic response as well as to study the mechanism of disease.^[29] Pancreatic cancer PDO HO219 showed increased sensitivity to Vav1 inhibition compared to HO163 (Figure 3A). Vav1 expression was measured using qPCR in both PDOs and showed increased expression in HO219 (Figure 3B). HO219 also had higher expression of Vav1 than HO163 using transcriptome analysis (Figure 3C), showing a correlation between Vav1 expression and sensitivity to Vav1 inhibition in these two PDOs.

Figure 3.

A) Pancreatic cancer PDOs, HO163 and HO219, were tested for sensitivity to the VAV1 inhibitor, charged DAG-lactone AB19. Cell viability was measured on day 5 using CellTiter-Glo^® and response was reported relative to the DMSO control wells. B) qPCR mRNA expression of VAV1 in HO163 and HO219 relative to two housekeeping genes, GAPDH and TBP. C) DeSEQ2 normalized RNA-SEQ read counts for VAV1 in HO163 and HO219. Each assay performed a minimum of 3 replicates. Error bars = SEM.

Conclusion

Over the past decades, several DAG-lactones analogues with broad structural variation have been designed and synthesized, generating diversity at the sn-1 and sn-2 positions and covering an extensive chemical space to selectively target typical C1 domain containing proteins. In this study, different types of cationic DAG-lactones were designed for specifically binding the atypical C1 domain of Vav1. We anticipated that the conserved geometry of Vav1C1 could support binding with DAG-like ligands if suitable substituents were introduced to interact with its unique hydrophilic residues. For the first time, we successfully synthesized a family of cationic DAG-lactones bearing a cationic pyridinium group at the sn-1 position alkylated with hydrocarbon chains of variable length. Evaluation of antiproliferative properties in human pancreatic cancer cell lines expressing Vav1 revealed that compounds AB176 and AB19, with a 14-carbon alkyl chain, showed the most pronounced effect on cell growth. Results obtained comparing compound AB176, with a 14-carbon chain on the pyridinium ring, to the less potent compounds 5b (8-carbon chain) and 5a (4-carbon chain) suggested that the length of this alkyl chain is also important for effective neutralization. Together, our results described herein with sn-1 positively charged DAG lactones support the concept that appropriately modified DAG analogs can specifically form ion-pair interactions with charged residues in the atypical C1 domain binding site, disrupting Vav1 activity. Although this represents an initial stage in the design of agents targeting features of the Vav1C1-like structure, our findings demonstrate that cationic DAG-lactones could inhibit growth and lower the levels of Vav1 in cell lines and patient-derived models and promise to be valuable as therapeutic agents for Vav1-positive pancreatic tumors. Moreover, this new kind of DAG analogues will serve as foundation for the targeting of other atypical C1 domain containing proteins driving malignancies as well as benign diseases.

Supplementary Material

All experimental details, characterization of products, copies of ¹H and ¹³C NMR spectra for the title compounds are provided in the Supporting Information, along with additional cited references.^[30–45]

Abbreviations

DAG

diacylglycerol

DMF

dimethylformamide

ERK

extracellular signal-regulated kinases

GEFs

guanine nucleotide exchange factors

Glu

glutamic acid

KHMDS

potassium bis(trimethylsilyl)amide

PCR

polymerase chain reaction

PDO

patient-derived organoids

PKC

protein kinase C

SN2

bimolecular nucleophilic substitution

SNAr

nucleophilic aromatic substitution