Structure-activity relationship of avocadyne

Abstract

Avocado consumption is associated with numerous health benefits. Avocadyne is a terminally unsaturated, 17-carbon long acetogenin found almost exclusively within avocados with noted anti-leukemia and anti-viral properties. In this study, specific structural features such as the terminal triple bond, odd number of carbons, and stereochemistry are shown to be critical to its ability to suppress mitochondrial fatty acid oxidation and impart selective activity in vitro and in vivo. Together, this is the first study to conduct a structure-activity analysis on avocadyne and outline the chemical moieties critical to fatty acid oxidation suppression.

Graphical abstract

Introduction

The avocado (Persea americana) is a flowering berry with a single large seed native to Mexico and Central America.¹ Compared to other fruits, the edible flesh of the avocado is low in sugar and contains nutrient dense fiber, vitamins, and a high ratio of monounsaturated (71%) to polyunsaturated (13%) and saturated fat (16%).¹ This makes the avocado an ideal component of diets with a lipid profile that is low in saturated fats and high in unsaturated fats to reduce the onset of metabolic diseases such as obesity, cardiovascular disease, and diabetes.² However, the discarded or unconsumed portions of the avocado (e.g., peel and seeds) account for 27% of the total weight of the fruit and, compared to the edible portion, contain higher amounts of acetogenins, a group of similarly structured lipids 17, 19, and 21 carbons in length with varied bioactivities including the ability to: hinder microbial growth^3,4, suppress UV-induced inflammation in skin cells⁵, modulate lipid metabolism resulting in the elimination of AML cells⁶ or the restoration of insulin sensitivity in lipotoxic pancreatic and muscle cells⁷, inhibit activation of the ERK signalling pathway resulting in reduced oral cancer proliferation⁸, and upregulate expression of the pro-apoptotic protein Bim in breast cancer cells⁹. Therefore, the parts we eat and discard contain numerous bioactives and opportunities to reduce biomass waste and improve health.

Acute myeloid leukemia (AML), a devastating group of blood cancers characterized by low survival rates, exhibits an abnormal reliance on higher rates of mitochondrial fatty acid oxidation (FAO), compared to normal blood cells.¹⁰ FAO breaks down long chain fatty acids of 14–18 carbons in length to fuel mitochondrial oxidative metabolism, which drives AML cell growth and proliferation.^10,11 Compared to the plasma of healthy donors, these abnormally high rates of FAO by AML cells manifests in significant depletions in total fatty acids in patient plasma, as confirmed by lipid profiling.¹² Current FAO inhibitors under investigation for use in AML such as etomoxir, which targets the outer mitochondria membrane fat transporter CPT1, induces severe hepatotoxicity¹³ while sulfo-N-succinimidyl oleate, which targets the cell surface fat transporter CD36, has not been clinically tested. These findings underline the need to identify novel compounds that modulate FAO metabolism.^10,11

The acetogenin avocadyne (1, Fig. 1A) is a 17 carbon long primary alcohol with hydroxyl moieties at the C-2 and C-4 positions and a terminal triple bond.¹⁴ Earlier structure elucidation studies on avocadene, the terminal alkene analogue of avocadyne, confirmed that the hydroxyl groups at C-2 and C-4 were syn to each other¹⁵ and were of a (2R,4R)-configuration.¹⁶ Our group previously determined that both avocadyne and avocadene are found in the edible flesh and discarded seeds.¹⁷ For avocadyne, the mean amount was 0.18±0.04 mg/g avocado in the pulp and 0.41±0.02 mg/g avocado in the seed¹⁷. For avocadene, the mean amount was 0.22±0.04 mg/g avocado in the pulp and 0.43±0.04 mg/g avocado in the seed.¹⁷ Most avocadyne and avocadene are esterified and alkaline saponification of the avocado material prior to extraction increases the amount of both compounds recovered.¹⁷

Figure 1. Avocadyne’s terminal triple bond and odd numbered carbon chain are critical to its ability to inhibit FAO-supported respiration and kill leukemia cells.

(A) Structures of (2R,4R)-avocadyne 1, 1-hexadecanol (C₁₆-OH), 1-hexadecanoic acid (C₁₆), palmitic acid terminal alkyne (PATA), 1-heptadecanol (C₁₇-OH), 1-heptadecanoic acid (C₁₇), and heptadecanoic acid terminal alkyne (HATA). (B) TEX and AML2 cell viability after a 72-hour treatment analyzed with flow cytometry. For TEX, the concentration range tested was 0.1–5 μM for avocadyne 1, 0.1–50 μM for HATA, and 0.1–100 μM for C₁₆-OH, C₁₆, PATA, C₁₇-OH, and C₁₇. For AML2, the concentration range tested was 0.1–40 μM for avocadyne 1, 0.1–50 μM for HATA, and 0.1–100 μM for C₁₆-OH, C₁₆, PATA, C₁₇-OH, and C₁₇. Concentrations are presented on the x-axis in log₁₀. (C) IC₅₀ values of HATA, PATA, 1 for TEX and AML2 were calculated using dose response – inhibition equation on GraphPad Prism 7.0. The IC₅₀ values of HATA and PATA were compared to 1 for TEX and AML2. (D) Intact basal and maximal cell respiration was assessed in TEX and AML2 following a 1-hour treatment with 10 μM of a lipid of interest. For (B, C, D), data is presented as mean ± S.D., *p<0.05, **p<0.01, ***p<0.001 using an unpaired, one-tailed, student’s t-tests or ANOVA were calculated in GraphPad Prism 7.0.

Acetogenin structure plays a critical role in bioactivity. Avocadyne and avocadene only differ by their terminal unsaturation¹⁴; avocadyne eliminates AML cells⁶ and possess anti-viral activity¹⁸ while avocadene strongly inhibits the growth of gram-positive microbes¹⁴. Addition of an acetoxy group to avocadyne’s primary alcohol resulted in complete attenuation of anti-AML activity.⁶ Recently, avocadyne induced AML cell death by suppressing FAO at very long chain acyl-CoA dehydrogenase (VLCAD), which decreased downstream mitochondrial metabolism, while sparing normal blood cells.¹⁹ However, the importance of the terminal unsaturation, number of carbons in the lipid as well as the stereochemistry at C-2 and C-4 in avocadyne for the inhibition of FAO remains unresolved. Therefore, with the aim of better understanding the avocado bioactives responsible for health effects, we investigate the importance of these structural aspects on avocadyne’s in vitro and in vivo selective AML toxicity, which was enabled through facile access to the four synthetic stereoisomers of avocadyne.²⁰

Results

The terminal unsaturation and odd number of carbons are critical to avocadyne-induced cytotoxicity

To determine the role of the hydroxyl and alkyne moieties as well as chain length in avocadyne-induced toxicity, the activity of 1, 1-hexadecanol (C₁₆-OH), 1-hexadecanoic acid (C₁₆), palmitic acid terminal alkyne (PATA), 1-heptadecanol (C₁₇-OH), 1-heptadecanoic acid (C₁₇), or heptadecanoic acid terminal alkyne (HATA) at 10 μM was assessed. AML viability after 72 hours and intact cell respiration, a measure of FAO-supported mitochondrial metabolism, after 1 hour were measured as indicators of compound activity (Fig. 1A: Structures). There was no effect on cell death across cell lines treated with the fully saturated fatty alcohols (C₁₆-OH; C₁₇-OH) or acids (C₁₆; C₁₇; Grey lines in Fig. 1B). While introducing a terminal triple bond to C₁₆ and C₁₇ (PATA and HATA; Red lines in Fig. 1B) caused significant death, 1 was still the most cytotoxic lipid against the leukemia cell lines TEX (Black line in Fig. 1B; Fig. 1C: F(2,8)=1454, p<0.001) and AML2 (F(2,8)=500, p<0.001). In TEX, the IC₅₀ values were 2.33±0.10 μM, 15.65±0.57 μM, and 52.93±0.66 μM for avocadyne 1, HATA, and PATA (Fig. 1C). In AML2, the IC₅₀ values were 11.41±1.87 μM, 22.60±1.37 μM, and 64.44±3.63 μM for avocadyne 1, HATA, and PATA (Fig. 1C). There was a significant increase in toxicity in HATA vs. PATA (i.e., 70% decrease in IC₅₀ values; Fig. 1C: TEX: t(4)=32.97, p <0.001; AML2: t(4)=19.73, p<0.001), suggesting that the number of carbons plays a role in cytotoxicity.

To confirm the functional effect of these structural elements, mitochondrial respiration as a measure of FAO was quantified using a respirometer. Compared to the DMSO vehicle control, there was no inhibition of basal or maximal respiration upon treatment with C₁₆-OH, C₁₆, and C₁₇-OH in leukemia cell lines (Fig. 1D: TEX Basal: F(3,11)=0.19, p=0.91; TEX Maximal: F(3,1)=0.36, p=0.79; AML2 Basal: F(3,11)=0.53, p=0.65; AML2 Maximal: F(3,11)=3.77, p=0.06). C₁₆ is the preferred substrate of mitochondria FAO and the homologous C₁₇ compound resulted in slight inhibition of leukemia cell respiration (Fig. 1D: TEX Basal: F(2,8)=5.15, p=0.05; TEX Maximal: F(2,8)=7.33, p=0.02; AML2 Basal: F(2,8)=5.48, p=0.04; AML2 Maximal: F(2,8)=11.47, p=0.01).

In correlation with cytotoxicity, PATA, HATA, and 1 inhibited leukemia cell respiration (Highlighted in red in Fig. 1D). The addition of a terminal triple bond to C₁₆ or C₁₇ (i.e., PATA or HATA) increased inhibition in TEX (Fig. 1D: C₁₆ vs PATA Basal: t(4)=3.87, p=0.02; C16 vs PATA Maximal: t(4)=4.8, p=0.009; C₁₇ vs HATA Basal: t(4)=3.39, p=0.03; C₁₇ vs HATA Maximal: t(4)5.98, p=0.004) and AML2 (Fig. 1D: C₁₆ vs PATA Basal: t(4)=3.16, p=0.03; C₁₆ vs PATA Maximal: t(4)=7.85, p<0.001; C₁₇ vs HATA Basal: t(4)=2.84, p=0.05; C₁₇ vs HATA Maximal: t(4)=6.63, p=0.003).

Compared to HATA and PATA, 1 was the most potent inhibitor of intact cell respiration in TEX (Fig. 1D: Basal: F(2,8)=14.19, p=0.005; Maximal: F(2,8)=25.69, p<0.001) and AML2 (Fig. 1D: Basal: F(2,8)=5.98, p=0.04; Maximal: F(2,8)=39.63, p<0.001). These results confirmed that hydroxylation at the C-2 and C-4 positions, an odd-numbered carbon chain length, and a terminal triple bond contribute to 1’s ability to induce cytotoxicity and inhibition of intact cell respiration.

The (2R,4R)-stereochemistry in avocadyne imparts in vitro activity

We next evaluated the effects of the configuration at C-2 and C-4 on the anti-AML activity of four synthetic avocadyne stereoisomers (1–4, shown in Fig. 2A) and avocadene stereoisomers (Suppl. Fig. 1A).²⁰ The activities of the compounds were assessed on AML cell viability after 72 hours and on intact cell respiration after 1 hour. Across both cell lines, all stereoisomers with a stereocenter of (S)-configuration (compounds 2–4) did not induce significant cell death (Fig. 2B). Similarly, only the (2R,4R)-stereoisomer 1 induced significant inhibition of fat-supported intact cell respiration (Fig. 2C: TEX Basal: F(5,12)=56.13, p < 0.001; TEX Maximal: F(5,12)=111.9, p < 0.001; Fig. 2C: AML2 Basal: F(5,12)=55.51, p <0.001; AML2 Maximal: F(5,12)= 104.0, p < 0.001). Similar patterns were noted in (2R,4R)-avocadene 5 (Suppl. Fig. 1B, C) reinforcing the importance of stereochemistry for acetogenin activity.

Figure 2. The (2R,4R)-stereochemistry is critical to avocadyne’s ability to induce FAO inhibition and AML cytotoxicity.

(A) Structures of the 4 stereoisomers of avocadyne 1-4. (B) Viability of TEX and AML2 treated with 10 μM of an avocadyne stereoisomer 1-4 for 72 hours, analyzed with flow cytometry. (C) Intact cell respiration of TEX and AML2 treated with 10 μM of an avocadyne stereoisomer 1-4. (D) Arbitrary units of ¹³C₁₆ palmitate and ¹³C₂ acetyl-carnitine in AML patient-derived cells (AML#1–4) treated with 50 μM of the (2R,4R)-stereoisomer 1 for 12 hours. For (B, C, D), data is presented as mean ± S.D., *p<0.05, **p<0.01, ***p<0.001 using an unpaired, one-tailed, student’s t-tests or ANOVA were calculated in GraphPad Prism 7.0.

FAO activity in AML patient-derived samples was further tested using uniformly labelled ¹³C₁₆-palmitate and metabolites were quantified by UHPLC coupled to high resolution MS.²¹ Treatment with the active (2R,4R)-stereoisomer 1 showed that FAO was decreased, as determined by an accumulation of palmitate (Fig. 2D: t(3)=6.048, p=0.009) and a decrease in acetyl-carnitine (Fig. 2D: t(3)=3.549, p=0.04). Taken together, the (R)-stereochemistry in both stereocenters is critical to avocadyne’s ability to bind to VLCAD and inhibit FAO in AML patient cells.

The (2R,4R)-stereochemistry in avocadyne impacts target binding

We previously showed that avocadyne inhibited FAO by interacting directly with VLCAD. AML2 cells were treated with 10 μM of the four stereoisomers 1–4 for three hours and then subjected to co-immunoprecipitation with an anti-VLCAD antibody. VLCAD-enriched fractions were next tested for purity by immunoblotting (Fig. 3A) and then subjected to LC/MS/MS analysis for avocadyne quantitation.¹⁷ The commercial standard elutes at 5.6 minutes (Fig. 3B). Consistent with the respiration and cytotoxicity data, only 1 was detected in VLCAD-enriched fractions, demonstrating a peak at 5.6 minutes (Green arrow in the chromatogram for 1: Fig. 3B). This suggests that only this stereoisomer interacts with VLCAD (Fig. 3B; Table 1). In contrast, LC/MS/MS analysis of VLCAD co-immunoprecipitation of AML2 treated with 2-4 did not demonstrate a peak at 5.6 minutes, suggesting that these stereoisomers do not co-immunoprecipitate with VLCAD (Red arrows in the chromatograms for 2-4: Fig. 3B; Table 1).

Figure 3. The (2R,4R)-stereochemistry is critical for avocadyne to interact to VLCAD.

(A) AML2 cells were treated with 10 μM of an avocadyne stereoisomer 1-4 for 3 hours and co-immunoprecipitated against an anti-VLCAD antibody. Immunoblotting to confirm purity of the VLCAD enriched fraction. (B) Total ion current chromatograms demonstrating recovery of avocadyne in VLCAD-enriched fractions treated with the (2R,4R)-stereoisomer 1 but not the other stereoisomers 2-4. All arrows point to 5.6 minutes, the elution time of a commercial standard of avocadyne. The green arrows demonstrate the presence of avocadyne in the co-IP fraction treated with the (2R,4R)-stereoisomer 1 while the red arrows demonstrate the lack of avocadyne in the co-IP fractions treated with the other stereoisomers 2-4 due to the lack of a peak at 5.6 minutes.

Table 1.

Recovery of avocadyne from a co-immunoprecipitation reaction starting with 10*10⁶ AML2 cells treated with 10 μM of 1-4 for 3 hours, analyzed with a previously published method.¹⁷ N=3 for each stereoisomer. “LOD” denotes the limit of detection.

Compound #	Concentration recovered (nM)
1	209.13 ± 48.67
2	< LOD
3	< LOD
4	< LOD

The (2R,4R)-stereochemistry in avocadyne impacts in vivo activity

We next determined whether chirality (e.g., R- vs S-) at C-2 and C-4 positions impacted in vivo activity. AML2 cells were xenotransplanted into the right flanks of non obese diabetic/severe combined immunodeficiency gamma (NSG) mice (Study schematic in Fig. 4A). Once tumors were palpable, mice were treated with a vehicle emulsion control or 150 mg/kg of 1 or 2 every other day for two weeks until endpoint. Only 1 suppressed tumor formation, resulting in extended murine survival (Fig. 4B: Day 13 t(4)=1.881, p =0.1331; Day 16 t(6)=0.1175, p=0.1678; Day 19 t(6)=1.833, p= 0.1165; Day 21 t(7)=2.810, p= 0.0262; Day 23 t(7)=2.852, p= 0.0246; Day 26 t(7)=2.521, p= 0.0398; Day 28 t(7)=3.704, p=0.0076; Log Rank Test p = 0.0031). In contrast, the (2S,4S)-stereoisomer 2 did not suppress tumor formation and showed no difference in survival, compared to the vehicle group (Fig. 4C: Day 14 t(8)=0.6905, p= 0.5094; Day 16 t(8)=0.06186, p=0.4761; Day 19 t(8)=0.3324, p= 0.3741; Day 21 t(8)=0.5581, p= 0.2960; Day 23 t(8)=0.3576, p= 0.3649; Day 26 t(8)=0.6192, p= 0.2765; Day 28 t(3)=0.09170, p= 0.89; Log Rank Test p = 0.8978).

Figure 4. The (2R,4R)-stereochemistry is critical for avocadyne’s in vivo activity.

(A) Schematic showing injection of AML2 cells into the right flanks of NSG mice. Once tumors were palpable, mice were separated into three groups (N=5 per group) and treated with either a vehicle emulsion or an emulsion³² containing either 150 mg/kg/week (2R,4R)- 1 or the (2S,4S)-stereoisomers 2 of avocadyne. (B) Tumor formation and survival of mice (N=5) treated with the (2R,4R)-stereoisomer 1. (C) Tumor formation and survival of mice (N=5) treated with the (2S,4S)-stereoisomer 2. (D) After tail vein injection, human AML patient cells (AML#5) were given four weeks to engraft in mouse femurs. Mice (N=3 per group) then received either a vehicle emulsion or 200 mg/kg/week (2R,4R)-stereoisomer 1 for four weeks. Engraftment levels (%CD33+/CD45+) were quantified by flow cytometry. For (B, C, D), *p<0.05, **p<0.01, ***p<0.001 using unpaired, two-tailed, Mann Whitney’s t-tests were calculated in GraphPad Prism 7.0. For (B, C), differences in survival were calculated with the Mantel-Cox Log-Rank test in GraphPad Prism 7.0.

To confirm the inhibitory activity of 1 on AML engraftment, a patient-derived xenograft model was used. AML cells were injected via tail vein and allowed four weeks to engraft to the femoral bone marrow (Schematic in Fig. 4D). After the engraftment period, mice receiving 200 mg/kg/weekly of 1 demonstrated significantly reduced leukemia engraftment, compared to the vehicle control (Fig. 4D). This highlights the in vivo activity of (2R,4R)-stereoisomer 1, confirming that stereochemistry is critical to avocadyne’s pre-clinical activity.

Discussion

Avocadyne is an emerging avocado-derived molecule that modulates mitochondrial fatty acid metabolism. In this study, we elucidate the pharmacophore of avocadyne and determine that (2R,4R)-stereochemistry is required for in vitro and in vivo activity against AML by inhibiting FAO at VLCAD.

The edible flesh as well as the inedible seed and peel of avocados are sources rich in acetogenins, which are bioactive lipids with an odd number of carbons uniquely found in food products.²² Odd numbered carbon fat metabolism is distinct from that of its even numbered counterparts.^23,24 Fully saturated C₁₇ (i.e., heptadecanoic acid) did not induce significant toxicity but modestly inhibited FAO-supported mitochondrial respiration, in agreement with previously published work.^23,24 Supplementation of ¹³C-labelled fatty acids to mice resulted in higher tissue accumulation of labelled C₁₅ and C₁₇ and slower rates of metabolism, compared to C₁₆ and C₁₈.^23,24 Structured similarly to avocadyne, fatty acids with terminal triple bonds inhibited omega hydroxylase or were fatty-acylated to certain proteins with no adverse effect on cell viability at similar concentrations.^25,26 In line with these studies, we show here that the terminal triple bond and an odd number of carbons were critical to avocadyne’s ability to inhibit FAO and eliminate AML cells.

Chirality plays a critical role in the pharmacology of targeting lipid metabolism.^27,28,29 We show here that the R-configuration at the C-2 and C-4 positions is required to induce leukemia cell death, which we link directly to FAO inhibition and VLCAD binding. These results agree with previously published studies highlighting the importance of the (R)-configuration for FAO inhibitors such as 2-bromopalmitate²⁷, C-75²⁸, and oxirane-containing carboxylates such as etomoxir²⁹. For both C-75 and etomoxir, FAO inhibitory activity was observed exclusively with the (R)-enantiomer while inhibition of fatty acid synthesis was observed exclusively with the (S)-enantiomer.^28,29 The current study shows that stereochemistry dictates the ability of avocadyne to suppress FAO.

Using a variety of methodologies, avocadyne and avocadene, as single agents or as the avo-B mixture, has demonstrated efficacy across multiple disease types including AML^6,19,30 and diet induced obesity (DIO)⁷, underscoring a consistent mechanism involving FAO inhibition and increased glucose oxidation. An earlier study by our group demonstrated the specificity of avo-B cytotoxicity on AML, a malignancy reliant on high rates of FAO, but not solid tumor, multiple myeloma or the non-cancer cell lines HEK293 and MRC7.^31,32 Acute myeloid leukemia^10,11 and diet-induced obesity⁷ are two disease states driven by abnormally high rates of FAO; pharmacological inhibition of FAO to eliminate AML^6,30 and ameliorate DIO⁷ have demonstrated therapeutic windows for potential clinical use. While the mechanism of avo-B involved inhibition of FAO in both AML and DIO⁷, cytotoxicity was observed with AML only.^6,30,31 C2C12 and INS-1 cells under lipotoxic conditions treated with avo-B experienced FAO inhibition but are viable at anti-AML concentrations.⁷ Importantly, C57BL/6J and NSG mice used in DIO⁷ and AML engraftment studies³⁰ receiving doses up to 300 mg/kg/ week of avo-B⁷ or avocadyne¹⁹ did not show alteration in weight, blood cell counts, or markers of liver, muscle, and kidney function.⁷ As part of a pilot clinical trial, participants consuming up to 200 mg avo-B per day for 60 consecutive days in a supplement experienced no severe adverse reactions.⁷ The in vitro^6,7, in vivo^6,7, and clinical⁷ results with avo-B further underscore a therapeutic window for the safe use of avo-B. Future clinical studies using exclusively avocadyne will be required to further demonstrate safety when it used as a single agent.

Recently, the avo-B mixture eliminated leukemia in vivo and in vitro, without harming the healthy blood population, in two separate studies.^6,30 In addition, avocadyne as a single agent directly inhibited VLCAD to suppress FAO in AML cells.¹⁹ In response to the chemotherapeutic insult, AML cells attempted to maintain mitochondrial metabolism. Using stable isotope palmitate and glucose tracers as well as a commercial PDH activity kit, AML cells attempted to compensate for FAO inhibition by increasing contribution of glucose-derived carbons into the TCA cycle through increased PDH activity¹⁹; increased pyruvate oxidation in AML is associated with decreased cancer cell proliferation and slowed disease progression.³³ The increase in pyruvate oxidation in AML cells was insufficient to prevent cell death¹⁹, in agreement with recent studies showing FAO inhibition hinders leukemia.^6,10,11

In DIO, high levels of saturated fatty acids and fatty acid intermediates drive abnormally high rates of FAO, suppressing glucose metabolism in cells normally responsive to insulin signalling.⁷ Using high resolution respirometry and liquid scintillation counting with ¹⁴C-palmitate and ¹⁴C-glucose, avo-B recently demonstrated FAO inhibition and increased pyruvate oxidation in both muscle and pancreatic cells.⁷ Attenuation of lipotoxicity restored insulin sensitivity in both cell types, increasing metabolic flexibility in muscle cells and increasing insulin secretion in pancreatic cells.⁷ In lipotoxic mice, avo-B decreased blood plasma glucose and restored triacylglyceride levels to non-lipotoxic levels, demonstrating a decrease in FAO and increased glucose utilization in vivo.⁷

Taken together, the current study elucidates the structural components critical to FAO inhibition by the components of avo-B. This study further expands on the mechanism by which avo-B modulates mitochondrial metabolism across multiple cell types and disease states, demonstrating the effects of the structure and function of these food-derived compounds on human health and metabolism.

While previous studies⁶ have demonstrated FAO inhibition with extracted avocadyne, acteogenin content can fluctuate greatly based on avocado maturation stage, climate, soil composition, fertilizer use, tissue (seed, edible pulp, or peel), and cultivar.^34,35 This study confirms that avocadyne’s distinctive structure and stereochemistry are critical to its ability to interact with VLCAD, inhibit FAO, and eliminate the leukemia cell population in vitro and in vivo. Future studies, in particular human clinical trials, should utilize exclusively the active (2R,4R)-stereoisomer, which can be produced economically by a recently elucidated synthetic pathway that is high in yield and purity.²⁰

Methods

Chemicals and synthesis of stereoisomers

1-hexadecanol (C₁₆-OH), 1-hexadecanoic acid (C₁₆), 1-heptadecanol (C₁₇-OH), 1-heptadecanoic acid (C₁₇) were all purchased from Sigma Aldrich while palmitic acid terminal alkyne (PATA) was purchased from Cayman Chemicals. Heptadecanoic acid terminal alkyne (HATA) was a custom synthesis product from Toronto Research Chemicals. Stereoisomers of avocadyne 1-4 and of avocadene 5 were synthesized as previously reported.²⁰ All compounds were solubilized in tissue culture grade dimethylsulfoxide (DMSO; Sigma), according to manufacturer’s instructions. Final DMSO concentration in cell culture media did not exceed 0.5%.

Cell Culture

TLS-ERG immortalized (TEX) leukemia cells were grown in IMDM (Iscove’s modified Dulbecco’s medium; Wisent) supplemented with 15% fetal bovine serum (FBS; Wisent), 2 mM L-glutamine (Sigma), 2% penicillin-streptomycin (pen-strep; Wisent), 20 ng/mL stem cell factor (SCF: Peprotech), 2 ng/mL interleukin 3 (IL-3: Peprotech). OCI-AML2 (referred to as AML2) were grown in IMDM supplemented with 10% FBS, 2% pen-strep. Both leukemic cell lines were passaged every 2–4 days and used up to passage 25 in a cell culture incubator at 37°C and 5% CO₂.

Cell Viability

Viability of leukemic cells was conducted as previously described.⁶ Leukemic cells were seeded at a concentration of 1.25*10⁵ cells/mL for AML2 and 1.5*10⁵ cells/mL for TEX and treated with a lipid of interest dissolved in DMSO for 72 hours. Cells were then washed once in PBS and re-suspended in PBS containing 50 μg/mL 7-aminoactinomycin (7AAD, BD Biosciences), and analyzed by the Guava EasyCyte 8HT flow cytometer.

UHPLC-MS Analysis of FAO, glycolysis, and TCA metabolites

Primary AML (2.5*10⁵ cells/mL) were treated with a DMSO solvent vehicle or (2R,4R)-avocadyne and incubated with uniformly labelled stable carbon 13 (¹³C) palmitate (¹³C₁₆; Cambridge Isotopes) for 12 hrs in a MEM media supplemented with 10% dialyzed FBS, 1% pen-strep, 0.4 mM l-carnitine, 0.65 mM L-glutamine, and 0.1 mM L-glutamate.²¹ Cells were then collected, counted, washed three times in PBS, and frozen at −82 °C until UHPLC analysis. Cells were extracted in methanol:acetonitrile:water (5:3:2 v/v/v) at a concentration of 1*10^6 cells per mL of the 5:3:2 solvent. prior to UHPLC-MS analyses (Vanquish-QExactive, Thermo Fisher).²¹

At 4°C, samples were vortexed for 30 min and centrifuged for 10 min at 15,000*g. Ten μL of sample extracts were then loaded onto a KinetexXB-C18 column (150 × 2.1 mm i.d., 1.7mm; Phenomenex).²¹ Chromatographic elution was achieved using a 3 min isocratic run (5% B) and a 9 min gradient from 5–95% B (phase A:water and B: acetonitrile with with 0.1% formic acid or with 10 mM ammonium acetate for positive or negative ion mode, respectively) to separate metabolites.²¹ The mass spectrometer either scanned in full MS mode, for the 3 min method, or performed acquisition independent fragmentation (AIF - MS/MS analysis), for the 9 min method, at 70,000 resolution in the 70–900 m/z range, 4 kV spray voltage, 15 sheath gas and 5 auxiliary gas, operated in negative and then positive ion mode which were separate runs.²¹ Metabolite assignment was performed against an in-house standard library. Metabolite levels were then normalized to protein quantification.²¹

High Resolution Respirometry

FAO-supported mitochondrial respiration of intact cells was performed as previously described.⁷ Exogenous fat is supplied in the form of palmitic acid conjugated to bovine serum albumin (7% BSA/5 mM palmitate). Sodium palmitate (27.8 mg ;Sigma) and 1.3 mL of distilled water are boiled at 80°C to solubilize the fat. 18.7 mL of 7.5% solution of fatty-acid free bovine serum albumin was obtained by diluting from a 30% BSA solution (Sigma) with distilled water. The 18.7 mL of 7.5% fatty acid-acid free bovine serum albumin was then heated to 42 degrees Celsius. . The 1.3 mL of solvated sodium palmitate is added dropwise to the 42°C bovine serum solution, stirred for 1 hr at 42°C, and frozen at −82°C until use.³⁶ Leukemia cells (5×10⁶ TEX or AML2 cells) were treated with 10 μM of a lipid of interest, 200 μM palmitic acid conjugated to bovine serum albumin, 1 mM l-carnitine (Sigma) for 1 hour, washed, re-suspended in PBS, and injected into the respirometer. FAO-supported respiration was characterized as follows: (1) physiological respiration fueled by substrates in the cell growth media immediately prior to injection, known as basal respiration, (2) residual respiration after uncoupling the ETC from ATP synthesis, known as LEAK respiration, (3) non-physiological maximal uncoupled respiration, known as maximal respiration, and (4) residual respiration associated with other oxygen-utilizing processes outside of the ETC, known as ROX. First, routine respiration was measured upon injection of intact cells but prior to addition of inhibitors. Second, LEAK respiration was measured upon injection of 250 nM of the ATP synthase inhibitor oligomycin (Sigma). Third, maximal respiration was measured with the addition of 1 mM pyruvate (Sigma) and a stepwise titration of 125 nM of chemical uncoupler carbonyl cyanide p-trifluoromethoxy phenyl hydrazone (FCCP; Sigma). Fourth, residual respiration was measured after the injection of 250 nM of the complex III inhibitor antimycin A (Sigma); this non-ETC respiration is then subtracted from all other oxygen flux values from the experiment. Oligomycin, FCCP, and antimycin A were dissolved in DMSO; l-carnitine and pyruvate were dissolved in PBS.⁷

VLCAD Co-immunoprecipitation

Co-immunoprecipitation (co-IP) of VLCAD was conducted using the Pierce Crosslink Magnetic Co-IP Kit (ThermoFisher Scientific) with the following antibodies: anti-VLCAD (ProteinTech). Experiments were conducted, according to the manufacturer’s protocol, with no deviations. Briefly, AML2 (10×10⁶ cells) were treated with 10 μM of an avocadyne stereoisomer in growth media for three hours at 37°C, 5% CO₂. During this incubation step, 5 μg of anti-VLCAD antibody (ProteinTech) was coupled to 25 μL of Pierce magnetic beads with the kit provided coupling agent 2 mg disuccinmidyl suberate. Cells were then lysed with chilled radioimmunoprecipitation assay (RIPA) buffer (Sigma). The lysate was exposed to the anti-VLCAD antibody coupled to the magnetic beads overnight at 4°C with gentle agitation with a lab gyrator. The lysate/antibody coupled to magnetic beads mixture were separated with a DynaMag^™ 2 Magnet. At this point, the avocadyne-VLCAD complex is bound to the antibody-beads; the lysate (containing all protein except VLCAD) is carefully removed to produce the flow through fraction. The magnetic beads are washed twice with 1mL of a kit provided immunoprecipitation wash buffer. The VLCAD protein is eluted from the beads with 100 μL of a kit provided low pH elution buffer for 10 minutes with gentle agitation at 24°C on a lab gyrator. The 100 uL fraction is the VLCAD-enriched fraction, contains exclusively VLCAD protein, and were subjected immediately to immunoblotting or frozen at −82 °C for subsequent quantification of avocadyne using our published method.¹⁷

Immunoblotting

Total cell lysates were prepared as described previously.⁶ Briefly, cells were washed twice with phosphate buffered saline pH 7.4 (PBS, Wisent) and suspended in chilled RIPA buffer. Protein concentrations were determined by the DC Protein assay (Bio Rad, Hercules, CA) according to manufacturer’s instructions. Equal amounts of protein were separated on 10% sodium dodecyl sulphate (SDS; Sigma)-polyacrylamide (BioRad) gels in a Mini Trans-Blot Cell (BioRad) at 150 V for 60 minutes in electrophoresis buffer (25 mM Tris base (Sigma), 190 mM glycine (Sigma), 3.5 mM SDS). This was followed by semi-dry transfer in a TransBlot semi-dry transfer apparatus (BioRad) to nitrocellulose membranes (BioRad) at 25 V for 45 minutes in transfer buffer (25 mM Tris base, 190 mM glycine, 20% v/v methanol). Membranes with attached proteins were blocked in 5% bovine serum albumin (Sigma) in 1X 20 mM tris-buffered saline with 0.1% Tween-20 (TBS-T; Sigma). Membranes were first probed overnight at 4°C with: anti-VLCAD antibody of rabbit origin 1:3000 (Cocalico, Stevens, PA, USA), and anti-GAPDH 1:15000 (ThermoFisher) in 1X TBS-T. Secondary IgG peroxidase linked rabbit-specific whole antibodies (Cell Signaling Technologies) in 1X TBS-T followed for 1 hr at 24°C. Detection was performed by staining the blots with the enhanced chemical luminescence solutions (Pierce) followed by exposure on the SynGene Chemigenius2 imager.

Animal Studies

NOD/SCID gamma mice (NSG; Jackson Laboratory, Bar Harbor, ME) were used for xenotransplant and engraftment studies and conducted as previously described.⁶ A oil-in-water emulsion containing avocadyne was used for treating mice.³² The 100 μL oil phase was dispersed into 900 μL of PBS and the total of 1 mL emulsion was vortexed vigorously for 30 seconds. All animal studies were carried out according to the regulations of the Canadian Council on Animal Care and with the approval of the University of Guelph, Animal Care Committee under the animal utilization protocol #3635.

In vivo tumor formation

AML2 cells (1×10⁶ cells/100 μL PBS) were subcutaneously injected into the mouse’s right flank. Once palpable tumors formed, mice were divided into groups (N=5 per group) and received either 150 mg/kg of either an emulsion³² containing the (2R,4R)- or (2S,4S)-avocadyne stereoisomer every other day for two weeks via an intra-peritoneal injection. Tumor volumes were measured as previously described⁶ and tumor burden was calculated relative to each mouse’s weight.

In vivo engraftment

Human AML cells (2.5×10⁶ cells/ 100 μL PBS) were injected via tail vein and allowed 4 weeks to engraft into femurs. Mice (N=3 per group) received either a vehicle emulsion³² or an emulsion³² containing 200 mg/kg/week of (2R,4R)-avocadyne for 4 weeks. Human leukemia cells in the femurs were incubated with anti-CD33 and anti-CD45 antibodies (BioLegend) and were quantified via flow cytometry on a Guava EasyCyte 8HT. Due to limiting AML patient derived material, an anti-AML chemotherapeutic was not included as a positive control.

Statistical Analysis

Unless otherwise stated, in vitro results are presented as mean ± SD whereas in vivo results are presented as mean ± SEM. Data were analyzed with GraphPad Prism 7.0 (GraphPad Software, USA) using one-way ANOVA with Tukey’s post hoc analysis for between group comparisons or standard student’s t-tests where appropriate. IC₅₀ values were calculated using the dose response-inhibition equation. p<0.05 was accepted as being statistically significant.