Modeling, Synthesis, and Biological Evaluation of Potential Retinoid-X-Receptor (RXR) Selective Agonists: Analogs of 4-[1-(3,5,5,8,8-Pentamethyl-5,6,7,8-tetrahyro-2-naphthyl)ethynyl]benzoic Acid (Bexarotene) and 6-(Ethyl(4-isobutoxy-3-isopropylphenyl)amino)nicotinic Acid (NEt-4IB)

Five novel analogs of 6-(ethyl)(4-isobutoxy-3-isopropylphenyl)amino)nicotinic acid—or NEt-4IB—in addition to seven novel analogs of 4-[1-(3,5,5,8,8-pentamethyl-5,6,7,8-tetrahydro-2-naphthyl)ethynyl]benzoic acid (bexarotene) were prepared and evaluated for selective retinoid-X-receptor (RXR) agonism alongside bexarotene (1), a FDA-approved drug for cutaneous T-cell lymphoma (CTCL). Bexarotene treatment elicits side-effects by provoking or disrupting other RXR-dependent pathways. Analogs were assessed by the modeling of binding to RXR and then evaluated in a human cell-based RXR-RXR mammalian-2-hybrid (M2H) system as well as a RXRE-controlled transcriptional system. The analogs were also tested in KMT2A-MLLT3 leukemia cells and the EC50 and IC50 values were determined for these compounds. Moreover, the analogs were assessed for activation of LXR in an LXRE system as drivers of ApoE expression and subsequent use as potential therapeutics in neurodegenerative disorders, and the results revealed that these compounds exerted a range of differential LXR-RXR activation and selectivity. Furthermore, several of the novel analogs in this study exhibited reduced RARE cross-signaling, implying RXR selectivity. These results demonstrate that modification of partial agonists such as NEt-4IB and potent rexinoids such as bexarotene can lead to compounds with improved RXR selectivity, decreased cross-signaling of other RXR-dependent nuclear receptors, increased LXRE-heterodimer selectivity, and enhanced anti-proliferative potential in leukemia cell lines compared to therapeutics such as 1.


Introduction
The human retinoid X receptors consist of three identified isoforms (α, β, γ) [1,2] with one or more of the isoforms exhibiting expression in every human tissue type where the receptor regulates-sometimes in partnership with other nuclear receptors-gene transcription, often stimulated by receptor-specific molecular signaling. The RXRs display a remarkable versatility unknown among other nuclear receptors (NRs) making up a transcriptional modulator superfamily because they join with many of the NRs to create heterodimers that actively modulate the pathways central to cell differentiation, metabolism, proliferation, and migration. Some of the critical receptor pathways where RXR participates as an essential component to realize functional responses include the liver X receptor (LXR), the thyroid hormone receptor (TR), the peroxisome proliferator-activated receptor (PPAR), the vitamin D receptor (VDR), and the retinoic acid receptor (RAR), to name a few. All NRs control gene expression, primarily by regulating transcription and usually in response to the presence of an associated receptor ligand and their obligate partnering receptor. Receptor ligands, often endogenous molecules, bind to the receptor's ligand-binding domain (LBD), which, in turn, compels the receptor to adopt a conformation that can then dimerize with an additional receptor, recruit co-factors, and ultimately bind with high affinity to a specific hormone responsive element (HRE) that the receptor regulates on DNA. Increasingly, HREs are being identified considerably up-or downstream from their moderated genes; however, a large number of HREs have also been identified close to or within the promoter region of the regulated genes. The HREs exhibited sequence specificity, consisting of two repeat hexads of half sites punctuated by a specified quantity of spacers separating those direct, inverted, or everted repeats [3]. VDRs, TRs, and RAR HREs include half-sites separated by 3-, 4-, and 5-nucleotide spacers, respectively [4,5].
Initially, TRs, VDRs, and RARs were presumed to bind to their HREs as homodimers [6], though they were later discovered to associate with RXR as a prerequisite to binding and activating their HREs [7]. Zhang and colleagues first reported that 9-cisretinoic acid (9-cis-RA)-a naturally occurring isomer of all-trans-retinoic acid (ATRA)-is an RXR-specific ligand that functions as an agonist where its binding to RXR compels the formation of RXR homodimers and subsequent association with RXR responsive elements (RXREs) [8]. When RXR associates with other NRs as a heterodimer, the heterodimer does not need to possess a RXR-specific ligand in the LBD for RXR. For example, the RXR-VDR heterodimer has been reported to function without a ligand bound to RXR [9]. Alternatively, some RXR heterodimers exhibit enhanced activity with RXR-specific ligands (rexinoids) bound to RXRs' LBD, as in the case of the RXR-LXR heterodimer. [10] Considering this degree of versatility-the necessity for RXR to partner with several NRs with or without ligands for those NRs to function-RXR has reasonably been termed the master receptor [11].
Numerous RXR-studies, comprising multitudes of rexinoids with different partnering NRs, have distilled two primary RXR heterodimer classifications-the so-called permissive and non-permissive RXR heterodimers [12]. Only the heteropartner's agonists can activate purely nonpermissive RXR heterodimers, whereas either the heteropartner's agonists or rexinoids can activate permissive RXR heterodimers. The RXR-RAR, RXR-TR, and RXR-VDR heterodimers have all been characterized as non-permissive. In most, but not all conditions, the RXR partnering receptor for the VDR and TR heterodimers was "silent." The RXR-RAR heterodimer, on the other hand, showed enhanced activation by both certain rexinoids and RAR-specific agonists. Specific RXR agonists, in fact, have shown activation of RXR-RAR in the absence of RAR-specific agonists [13]. The primary classification of RXR-RAR as nonpermissive has evolved in light of these observations to have a more accurate "conditionally nonpermissive" designation. The RXR-LXR, RXR-PPAR, and RXR-FXR heterodimers, alternatively, are all known to be fully permissive.
Both permissive and nonpermissive RXR heterodimers often give rise to pleiotropic effects from exposure to potent rexinoids-the former by stimulating RXR heterodimer pathways and the later by titrating a finite pool of RXR away from participating in the proper formation and functioning of those nonpermissive RXR heterodimers. This potential for pleiotropy has frustrated the clinical development of rexinoids for therapeutic applications. Rexinoids such as 9-cis-RA can arrest the functioning of the RXR-VDR [14][15][16] and RXR-TR [17] heterodimers. Similarly, molecular signals such as 1,25-dihydroxyvitamin D 3 (1,25D) or T3 promote RXR-VDR or RXR-TR formation, depleting RXR availability, and thereby inhibiting alternative RXR-dependent mechanisms of action. This so-called crossreceptor squelching is exemplified by the loss of VDR function via T 3 -TR-RXR-modulated inhibition [18,19] and similarly by the loss of TR translational activity concomitant with 1,25D-VDR-RXR-activation [17,20]; however, the inhibition by crosstalk in the former examples likely concerns more than just the depletion of RXR. Nevertheless, the two overarching characteristics concerning the development of rexinoid therapeutics that exert fewer side effects and greater benefits comprise selectivity and potency [21]. Thus, an approach to modify a parent RXR-agonist's structure may impact both potency and RXR-heterodimer selectivity, leading to improved pleotropic profiles by generating specific NR modulators (SNuRMs) [22].
Several rexinoid SNuRMs are being investigated as drug targets, particularly in the case of cancer where selective RXR over RAR activation results in chemotherapeutic effects in many human cancers and avoids potential RAR toxicities [23]. Following several studies [24,25] that modeled and tested compounds evolved from 9-cis retinoic acid, an isomer of all trans retinoic acid (ATRA), 4-[1- (3,5,5,8,8-pentamethyltetralin-2-yl)ethynyl]benzoic acid (1) [26] has emerged as a lead RXR-selective, potent synthetic agonist, and although several other candidates have displayed equally potent if not superior profiles, Ligand Pharmaceuticals Inc. was granted FDA approval of 1, known more widely as bexarotene, to treat cutaneous T-cell lymphoma (CTCL). Several studies have reported structural analogs of bexarotene such as disilabexarotene (2) [27], for example, which have been shown to exhibit similar activation of RXR (Figure 1). While bexarotene (1) was first approved for the treatment of CTCL, it has also been tested in breast cancer [28], colon cancer [29], and lung cancer models [30]. In fact, a proof-of-concept (POC) clinical trial reported therapeutic benefits for use of 1 in nonsmall cell lung cancer [31,32], and bexarotene can be prescribed off-label for this disease. A mounting number of studies have linked cell-proliferation suppression and combinationchemotherapeutic apoptosis synergy with RXR-controlled pathways. Bexarotene (1) and numerous other synthetic rexinoids have also demonstrated positive impacts in non-insulindependent diabetes mellitus (NIDDM) mouse models, arising from metabolism regulation by RXR:PPAR [33]. While bexarotene (1) is predominantly RXR-selective and avoids significant RAR-activation, patients treated with 1 often experience hypothyroidism [34], hyperlipidemia, and occasionally cutaneous toxicity. Bexarotene (1), similar to 9-cis-RA, incites these side effects by disrupting nonpermissive heterodimers-hypothyroidism by RXR-TR [35] disruption-or stimulating the permissive heterodimers-hyperlipidemia via RXR-LXR activation [36,37] and cutaneous toxicity [38] from RAR activity at high dose concentrations. A number of groups are actively designing rexinoids with greater potency and specificity toward RXR-homodimer formation, in order to mitigate impacts on at least the permissive RXR heterodimer pathways. Adding to the urgency of developing novel rexinoids possessing attenuated side effect profiles, compound 1 has shown some promise in neurodegenerative disease models such as Parkinson's disease [39] and Alzheimer's disease (AD) [40]. Moreover, several novel rexinoids were recently reported to be equally or more effective at modulating gene expression on LXREs and NBREs and are thus superior at inducing ApoE and tyrosine hydroxylase, two genes whose enhanced expression is thought to mitigate the pathophysiology associated with Parkinson's and Alzheimer's diseases [41]. Significantly, a POC trial of 1 in AD patients exhibiting moderate symptoms demonstrated a statistically significant clearance of soluble amyloid beta in non-apoE 4 genotypes [42]. Furthermore, bexarotene (1) exhibited one of the best profiles-similar to that of remdesivir-in preventing SARS-CoV-2 infection in vitro in a recently reported robust screening assay of a 1528 FDA-approved drug library that identified four drugs that were active against the virus [43]. Bexarotene has also been shown to reduce inflammation [44] as well as decrease CL22 production by M2-polarized tumor-associated macrophages [45], which then modulates the tumor microenvironment. Furthermore, bexarotene is also being explored for a novel treatment of Cushing's disease [46] and glioma [47].
Employing modeling and structural features of reported rexinoids as starting points, many groups have successfully developed novel rexinoids with unique profiles. One such rexinoid that has been examined as a potential therapeutic for several human cancer and neurodegenerative diseases is IRX-4204 (3) [48], which was shown to activate RXR most potently compared to its other stereoisomer. Another well-studied rexinoid known as 9cUAB30 (4) [49] is currently in clinical trials for early stage mammary cancer [50][51][52], and several methylated variants of 4 [53,54] have helped demonstrate why 4 does not incite hyperlipidemia through RXR-LXR agonism compared to other moderately potent rexinoids. Boehm and colleagues have described unbranched trienoic acids [55] as well as analogous compounds containing a single [56] or multiple-fused [57] aryl ringscompound 5 [57] exemplifying the latter. Our group reported a mono-fluoro-bexarotene (6) [58] and a difluoro-bexarotene (7) [59] that displayed increased RXR activity relative to bexarotene. Compounds 8 [60] and LGD100268 (9) [60] both exhibited increased RXR activity in a CV-1 cell line versus 1. The acrylic acids 10 (CD3254) [61] and 11 (CD2915) [62] possess similar potency for RXR agonism as 1. Compound 12 [63,64] possesses a single unsaturation in the aliphatic ring system as its only structural difference from 1. We used compounds 8-12 as starting points to prepare analogous rexinoids 13-19 [21] with unique gene expression and side-effect profiles in vivo [65]. Indeed, pyrimidine bexarotene (14) and pyrimidine LGD100268 (15) both showed improved therapeutic profiles over 1 in a mouse model of lung cancer [30]. Kakuta's group reported the highly potent rexinoid 20 (NEt-TMN) [66][67][68][69][70], where NEt-TMN analogs 21 [71][72][73] and 22 [71,72] have also shown high potency in addition to several other NEt-TMN analogs that our team has described [74]. Kakuta's group also reported the partial RXR agonist 23 [68] and a partial RXR agonist analog 24 [70] in pursuit of novel treatments for type II diabetes. A new partial RXR agonist 25 (NEt-4IB) [75] reported by Kakuta's group has also shown promise in mouse models of diabetes and pulmonary emphysema [76], and we were interested in testing compound 25 and a few analogs of 25 for their activity and anti-proliferative properties in vitro, since we expect a reduced side-effect profile via RXR-dependent cross-signaling for these types of compounds. Finally, the indanyl-compound 26 [77] was reported in a patent for novel RXR agonists by Boehm, Heyman, and Lin, and compound 27 [26] was originally reported alongside 1 and showed similar activity ( Figure 2). The current work concerns the synthesis of four novel analogs of NEt-4IB, compounds 28-31, and seven novel analogs of bexarotene, compounds 32-37a and 37b, for preliminary biological evaluation in KMT2A-MLLT3 leukemia cells as well as a number of receptorbased assays in human cell lines to probe off-target activity ( Figure 3).

Results: Molecular Modeling
AutoDock Vina was used to predict noncovent binding of human-RXR with different compounds. The output of AutoDock Vina is the prediction of bound conformations and a score represents binding affinity. The predicted binding affinity of human-RXR for each ligand is output as an energy unit in kcal/mol (Table 1), followed by visual inspection of the bound ligand-protein complexes in PyMOL (version 2.3, Schrödinger, LLC) ( Figure 4a). To further analyze and illustrate the interactions between protein residue sidechains with the ligands, PoseView (BioSolvIT [78,79], Sankt Augustin, Germany) was used to generate the two-dimensional renderings depicting RXR protein sidechain interactions with each ligand (Figure 4b). In these two-dimensional depictions, hydrogen bonds are presented as dashed lines between interaction partners, and hydrophobic interactions are depicted as smooth contour lines. In the predicted conformation, the binding pocket of bexarotene involved five key residues: Ile268, Ala272, Phe313, Ile345, and Cys432. This detailed structural information from docking provides insights to design or modify compounds with higher affinity and selectivity in the future.  The AutoDock Vina score showed that the standard compound bexarotene (1), with a score of −12.7 kcal/mol, was the most potent among all compounds. Compounds 26, 27, 33, 35, 36, 37a, and 37b had comparable scores of −12.3 kcal/mol, −12.0 kcal/mol, −11.6 kcal/mol, −11.5 kcal/mol, −12.4 kcal/mol, −12.1 kcal/mol, and −11.9 kcal/mol, respectively (Table 1). The lower AutoDock Vina scores for 33, 35, 36, 37a, and 37b provided the motivation to synthesize these compounds for biological evaluation. Based on prior experience with modeling for these compounds, we were eager to synthesize all RXR compounds with a docking score within the range of 10% to that of bexarotene, since these compounds possessed the potential to be better candidates that exhibited comparable EC 50 and IC 50 profiles for further study.
Ketone 75 was treated with a solution of triphenylphosphonium methylide in THF and converted to alkene 79 (56%), which was then saponified to give rexinoid 36 in an 81% yield (Scheme 16). Finally, for the synthesis of hydroxylated analogs of bexarotene (37a and 37b), the commercially available 2-hydroxyterephthalic acid (80) was esterified by reflux in acidic methanol to give dimethyl 2-hyroxyterephthalate (81) in 90%, and using the method of Ningren and co-workers [81], 81 was selectively saponified to 82 in a 49% yield. The hydroxyl group of 82 was then acetylated to give 83 in an 88% yield, and 83 was converted to acid chloride 84 quantitatively by treatment with thionyl chloride (Scheme 17). Since the acetyl protecting group was discovered to be labile under typical Friedel-Crafts acylation conditions, 2.2 equivalents of known compound 85 [58] were combined with 84 and aluminum chloride to give a 70% yield of ketone 86 after aqueous acidic workup and an un-recovered mass of 87 as the by-product of the labile acetyl protecting group. Finally, ketone 86 was treated with an excess of triphenylphosphonium methylide solution to give alkene 88 in a 34% yield after acidic workup, and 88 was saponified to 37 in a 79% yield (Scheme 18).
In a similar manner, acid chloride 84 was combined with 71 (2.2 equivalents) in dichloromethane with aluminum chloride to give ketone 89 in a 48.8% yield and an unrecovered mass of 90 after aqueous acidic workup. Ketone 89 was treated with a solution of triphenylphophine methylide followed by aqueous acidic workup to give 91 in a 53.5% yield, and compound 91 was saponified to give compound 37b in a 72.7% yield after purification by silica gel column chromatography (Scheme 19).

Results and Discussion: Biological Assays
Bexarotene (1) and analogs 25-36, 37a, and 37b were assessed in KMT2A-MLLT3 cells to obtain EC 50 values for RXRα activation in both a GFP and Luc-assay, and then also in a 96 h cell viability assay both with and without 100 nM ATRA (Figure 7), the results of which are summarized in Table 1.
These compounds were also assessed for mutagenicity and toxicity in Saccharomyces cerevisiae and the toxicity results are summarized in Table 1. No compound was mutagenic in this assay. The EC 50 determination showed that the standard bexarotene (1) with an EC 50 value of 18 nM, was one of the most potent rexinoids for the assay, where only compounds 33 and 35 possessed lower EC 50 value concentrations of 17 nM and 1.3 nM, respectively, and the hydroxy-bexarotene analog 37a possessed a comparable EC 50 value of 24.2 nM (Table 1). Not surprisingly, only 33, 35, and 37a exhibited comparable IC 50 values to bexarotene (1) in the 96 h cell viability assay where compounds were tested in the presence of 100 nM ATRA (Table 1). Of these compounds, only 37a demonstrated cytotoxicity in the Saccharomyces cerevisiae assay of 1 µg/µL (Table 1). In terms of SAR, the most active compounds, 33 and 35, possessed a cyclopropyl-linking ring and the pyrimidyl-carboxylic acid system. It is curious that changing the pentamethyl-napthalenyl ring system of bexarotene to either the pentamethyl-indanyl or tetramethyl-napthalenyl system of 33 and 35, in combination with substituting the vinyl-linking group with a cyclopropyl ring system and the benzene ring with a pyrimidine ring system, did not reduce activity for the receptor or antiproliferative effects in cell culture. Changing the pentamethyl-napthalenyl to a tetramethyl-napthalenyl group without changing other groups appears to lower the activity at the receptor-27 versus 1, 37b versus 37a, 36  We next tested the analogs for their ability to bind and activate the liver-X-receptor (LXR) using a liver-X-receptor responsive element (LXRE)-based assay, and we compared the effect in the presence vs. absence of an activating LXR compound (TO901317). LXR has been demonstrated to regulate lipid metabolism and inflammatory responses in the central nervous system, and there is ample evidence that robust cholesterol and lipid metabolism in the brain (including enhanced ApoE expression) are critical to mitigating dementia. Biological evaluation of our novel RXR agonists for their ability to transactivate via an LXRE sequence that is found naturally in the promoter of LXR-RXR controlled genes including ApoE was carried out in human embryonic cells (HEK293) with bexarotene (1) as a comparison. The activation from this natural LXRE in our system was tested in the presence of either 100 nM RXR agonists alone or in combination with 100 nM of both the RXR agonist and LXR agonist T0901317 (TO). The use of the combination of LXR and RXR agonists was expected to display a more robust response in LXRE transactivation due to additive or synergistic effects of dual ligand activation of the RXR-LXR heterodimer. The results (Figures 8A and 9A) revealed that in comparison to the parent bexarotene (1) compound alone, single dosing of the cells with any of the tested analogs displayed less LXR/LXRE activity. Specifically, the analogs possessed activities ranging from 46 to 94.5% of the bexarotene control (set to 100%; Table 1). Moreover, when a LXR synthetic ligand (TO) was used in combination with Bex or analogs, a similar profile was observed, with the exception of TO+26, which displayed a higher activity than TO+1 ( Figures 8A and 9A).  While most of the analogs possessed slightly lower LXR activation when compared to bexarotene (1), it is important to consider this activity in the context of the RXR-RXR homodimer activity of each analog, and to thus "normalize" the LXR/LXRE heterodimer activation in order to yield a LXRE Heterodimer Specificity (LHS) score ( Figures 8B and 9B). The results of this LHS analysis (Table 1) revealed that many of our novel compounds (e.g., 25, 28, 29, 30, 31, 35, and 36) possessed greater LXR/LXRE activity via increased heterodimer specificity than the parent bexarotene (1).
Finally, since compound 1 is known to possess "residual" RARE activity, we evaluated the ability of our compounds to induce transcription via the retinoic acid response element and retinoic acid receptor (RAR). Human embryonic cells (HEK293) were transfected with human RARα and dosed with 10 nM of either all-trans retinoic acid (RA), the natural ligand for RARα, compound 1, or analogs. Employing this assay, compound 1 possessed and average 28.1% of the activity of the RA control ( Figure 10). Compound 32 displayed the greatest RARE activation at 12.9% of RA, while compound 30 showed the lowest RARE activity at 1.1%, which is indistinguishable from the ethanol control (Table 1). Thus, all of our novel analogs displayed significantly less "cross-over" onto RAR-RARE signaling compared to bexarotene (1).

Figure 10. Assessment of RXR agonists via a RARE-luciferase reporter based assay in human cells. (A-E) Human embryonic cells (HEK293)
were co-transfected with expression vectors for hRXRα, a RARE-luciferase reporter gene, and a renilla control plasmid for 24 h utilizing a liposome-mediated transfection protocol. Cells were treated with bexarotene, analog, or all-trans-retinoic acid (RA) at 10 nM for 24 h. The RARE activity for RA was set to 100%. Values are means ± SD with all analogs tested displaying lowered RARE activity vs. compound 1 (p < 0.05).

Conclusions
This work generated a number of NEt-4IB and bexarotene analogs to evaluate as RXR agonists and their ability to prohibit cell viability in a KMT2A-MLLT3 cell line in combination with 100 nM ATRA. In general, the EC 50 values determined for the analogs correlated well with their IC 50 values in the cell viability assay, and we identified two bexarotene analogs-33 and 35-more potent than bexarotene as well as one analog, 37a, with comparable potency. Many of the analogs also revealed an enhanced ability to activate LXRE-mediated transcription, intimating their ability to significantly stimulate LXR/LXRE target genes such as ApoE, which have been implicated in protection against dementias. Furthermore, our novel panel of analogs possessed less "cross-over" activation than other retinoid pathways including RARE-directed transcription. Taken together, these results provide compelling motivation to continue to modify bexarotene and other reported RXR agonists to evaluate their potential anti-proliferative and other therapeutic activities.

Molecular Modeling
The three-dimensional structures of the compounds reported herein were generated using ChemDraw 3D (PerkinElmer Informatics), energy minimized, and exported in the Protein Data Bank (PDB) format. The human RXR alpha ligand binding domain structure model was obtained from the PDB (PDB code: 1FBY, [83]). The crystallized ligand, 9-cis retinoic acid, was removed from the protein model prior to docking simulations. Furthermore, 9-cis retinoic acid was also used as a positive control in the docking studies presented here. Both the protein and ligand models were prepared using MGLTools (version 1.5.7) [84] and screened virtually using AutoDock Vina [85]. The search space volume (4,032 Å 3 ) was determined using MGLTools (center_x = 12.848, center_y = 29.174, center_z = 50.269, size_x = 16, size_y = 14, size_z = 18). The exhaustiveness was set to 8.

Luciferase Detection
293T cells were transfected using Lipofectamine 2000 (Invitrogen). Six hours after transfection, the cells were collected and plated into a 48-well plate in 1% BSA media in triplicate and treated with compounds. After 40 h incubation, the cells were harvested and assayed for luciferase (Luc Assay System with Reporter Lysis Buffer, Promega) in a Beckman Coulter LD400 plate reader.

LXRE Assay
The LXRE-mediated assays were performed using human embryonic cells (HEK293) seeded at a density of 60,000 cells/well in a 24-well plate and maintained in DMEM (Hyclone) supplemented with 10% fetal bovine serum, 100 µg/mL streptomycin, 100 U/mL penicillin (Invitrogen, Carlsbad, CA, USA) at 37 degrees Celsius, 5% CO 2 for 24 h. The cells were co-transfected with 250 ng of an LXRE-luciferase reporter gene, 50 ng of pSG5-human RXRα, and 20 ng of renilla control plasmid. The transfections were conducted using 1.25 µL of polyethylenimine (PEI) (Polysciences, Inc., Warrington, PA, USA) for 16-22 h. After transfection, the cells were treated with either ethanol vehicle control (0.1%), reference compound bexarotene (1) or analog, and/or T0901317 (an LXR ligand) at the indicated concentrations. After 24 h post treatment, the cells were lysed and the transcriptional activity mediated by the LXRE was measured using the Dual Luciferase Assay System (Promega, Madison, WI) in a Sirius FB12 luminometer (Berthold Detection Systems, Pforzheim, Germany) according to the manufacturer's protocol. The data are a compilation of between three and six independent assays with each treatment group dosed in triplicate for each independent assay. The transcription efficiency on the LXRE was measured in comparison to the reference compound bexarotene (1) set to 100%. Bars on all graphs indicate standard deviation of the replicate experiments.

RARE Assay
Human embryonic kidney cells (HEK293) were plated at 60,000 cells per well in a 24-well plate and maintained as described above. After 24 h, the cells were transfected with 250 ng pTK-DR5(X2)-Luc, 25 ng pCMX-human RARα, and 20 ng renilla utilizing 1.25 µL polyethylenimine (PEI) per well for 24 h. The sequence of the double DR5 RARE is: 5 -AAAGGTCACCGAAAGGTCACCATCCCGGGAGGTCACCGAAAGGTCACC-3 (DR5 responsive elements underlined). The cells were treated with ethanol vehicle (0.1%), all-trans-retinoic acid (RA, the ligand for RAR), or the indicated rexinoid at a final concentration of 10 nM. After 24 h of treatment, the retinoid activity was measured as described above (dual luciferase assay). The activity of compound 1 or analog divided by the activity of all-trans-RA (expressed as a percentage) represents the RARE activity. Three independent assays were conducted with triplicate samples for each treatment group. The value for RA was set to 100%. After an additional 96 h, the number of viable cells in 50 µL was determined using a ZE5 flow cytometer (Bio-rad) using forward scatter/side scatter and PE exclusion to isolate viable cells.

Data Analysis
Statistical analysis was performed using Prism (Graphpad). T-test was performed, as appropriate. Error bars represent standard deviation. Data points without error bars have standard deviations below Graphpad's limit to display. For Figures 8-10, data are expressed as means ± SD. Statistical differences between two groups (generally the bexarotene control group versus bexarotene analog group) were determined by a two-sided Student's t-test. A p-value of less than 0.05 was considered significant.

Mutagenicity and Toxicity Assay
All compounds were tested for toxicity and mutagenicity using a Saccharomyces cerevisiae based assay as described previously [58]. Toxicity was assessed in this assay (Table 1), comparing growth on plates to control treatments. Compounds were solubilized in DMSO at increasing concentrations and cells were incubated with the compounds for 3 h before plating on selective media or YPD to assess toxicity and mutagenicity. Cytotoxicity was assessed as described [86]. Growth of colonies on the full nutrient YPD plate for each treatment was compared to the DMSO only control. The concentration at which 50% cell death (as indicated by colony count compared to DMSO only control) +/−10% cell death is reported as the 50% killing rate. The highest concentration tested was 11 µg/µL.

HPLC
All tested compounds were assessed on a Waters Acquity UPLC with QDA and PDA detectors. Compounds were assayed in ESI-mode on an ACE Excel C18-PFP (1.7 µm, 50 mm × 2.1 mm) column using a 0.1% formic acid/water:acetonitrile gradient over 5 min. HPLC traces for compounds 25-36, 37a, and 37b are available in the Supplementary Materials.

NMR and High Resolution Mass Spectrometry
A 400 MHz Bruker Avance III spectrometer was used to acquire 1 H NMR and 13 C NMR spectra. Chemical shifts (δ) are listed in ppm against residual non-deuterated solvent peaks in a given deuterated solvent (e.g., CHCl 3 in CDCl 3 ) as an internal reference. Coupling constants (J) are reported in Hz, and the abbreviations for splitting include: s, single; d, doublet; t, triplet; q, quartet; p, pentet; m, multiplet; br, broad. All 13 C NMR spectra were acquired on a Bruker instrument at 100.6 MHz. Chemical shifts (δ) are listed in ppm against deuterated solvent carbon peaks as an internal reference. High resolution mass spectra were recorded using either a JEOL GCmate (2004), a JEOL LCmate (2002) high resolution mass spectrometer or an ABI Mariner (1999) ESI-TOF mass spectrometer. NMR spectra are available in the Supplementary Materials.

General Procedures
Removal of volatile solvents transpired under reduced pressure using a Büchi rotary evaporator and is referred to as removing solvents in vacuo. Thin layer chromatography was conducted on precoated (0.25 mm thickness) silica gel plates with 60F-254 indicator (Merck). Column chromatography was conducted using 230-400 mesh silica gel (E. Merck reagent silica gel 60). All tested compounds were analyzed for purity by NMR as well as HPLC analysis and were found to be >95% pure.
6.14. 1-isobutoxy-2-isopropylbenzene (39) To a solution of 2-isopropylphenol (38) (12.5 mL, 92.9 mmols) and 1-bromo-2methylpropane (20.5 mL, 189 mmols) in DMF (50 mL) was added finely ground potassium carbonate (13.9 g, 101 mmols) and potassium iodide (0.652 g, 3.9 mmols), and the reaction was stirred for 20 h at 70-75 • C. The reaction solution was then poured into water and extracted with ethyl acetate. The organic layers were washed with brine, dried over sodium sulfate, and concentrated in vacuo to provide a crude oil that was purified by column chromatography (1% ethyl acetate in hexanes) to give 39 as a colorless oil (8. To a solution of 1-isobutoxy-2-isopropylbenzene (39) (17.208 g, 89.486 mmols) in ethyl acetate (100 mL) at 0 • C was added concentrated (>90%) nitric acid (50.5 mL, 1.2 mols). The reaction was stirred at 0 • C for 40 min at which point it was carefully poured into water and extracted with ethyl acetate. The organic layer was washed with brine and dried over sodium sulfate to give a crude oil that consisted of 40 and 41 in a 3:1 ratio-TLC separates these isomers after four elutions in 1% ethyl acetate:hexanes (41 R f~0 .5 and 40 R f~0 .45). This crude oil was purified by column chromatography (0.7% to 1% to 5% ethyl acetate in hexanes) to give 40 (9.

Methyl 2-fluoro-4-((4-isobutoxy-3-isopropylphenyl)amino)benzoate (50)
To a solution of 42 (1.3452 g, 6.49 mmol), methyl 2-fluoro-4-iodobenzoate 47 [74] (1.9807 g, 7.07 mmols), Cs 2 CO 3 (5.5562 g, 17.08 mmol), and rac-BINAP (0.3386 g, 0.55 mmol) in toluene (8.6 mL) in a 100 mL round-bottomed flask was added Pd 2 (dba) 3 (0.319 g, 1.82 mmol). The solution was sparged with nitrogen for 5 min, then a reflux condenser was fitted to the flask, the atmosphere was evacuated and back-filled with nitrogen (three times), and then the reaction was heated to reflux with stirring in an oil bath (125-120 • C) for 22 h. After cooling the reaction to room temperature, excess cesium carbonate and other solid particulates were filtered and washed with ethyl acetate, and the organic filtrate was concentrated in vacuo to obtain a crude product that was purified by column chromatography (150 mL SiO 2 , 6% ethyl acetate:hexanes to 12% ethyl acetate: hexanes) to give 50 (2. (51) To a solution of 42 (1.1807 g, 5.695 mmol), methyl 4-iodobenzoate 48 (1.6531 g, 6.308 mmols), Cs 2 CO 3 (4.9777 g, 15.28 mmol), and rac-BINAP (0.3097 g, 0.50 mmol) in toluene (7.4 mL) in a 100 mL round-bottomed flask was added Pd 2 (dba) 3 (0.2868 g, 1.64 mmol). The solution was sparged with nitrogen for 5 min, then a reflux condenser was fitted to the flask, the atmosphere was evacuated and back-filled with nitrogen (three times), and the reaction was heated to reflux with stirring in an oil bath (125-120 • C) for 22 h. After cooling the reaction to room temperature, excess cesium carbonate and other solid particulates were filtered and washed with ethyl acetate, and the organic filtrate was concentrated in vacuo to give a crude product that was purified by column chromatography (150 mL SiO 2 , 6% ethyl acetate:hexanes to 12% ethyl acetate: hexanes) to give 51 (1.0494 g, 54%) as a crystalline solid, m.p. 101.1-103. The solution was sparged with nitrogen for 5 min, then a reflux condenser was fitted to the flask, the atmosphere was evacuated and back-filled with nitrogen (three times), and the reaction was heated to reflux with stirring in an oil bath (125-120 • C) for 22 h. After cooling the reaction to room temperature, excess cesium carbonate and other solid particulates were filtered and washed with ethyl acetate, and the organic filtrate was concentrated in vacuo to give a crude product that was purified by column chromatography (150 mL SiO 2 , 4% ethyl acetate:hexanes to 12% ethyl acetate: hexanes) to give 52 (

Methyl 6-(ethyl(4-isobutoxy-3-isopropylphenyl)amino)nicotinate (53)
To a flame-dried, 100 mL round-bottomed flask equipped with a magnetic stir bar was added a 60% dispersion of sodium hydride in mineral oil (0.2351 g, 5.88 mmol). The dispersion of sodium hydride was washed with hexanes (3 mL, twice) and dried under vacuum and suspended in 3.0 mL of DMF under nitrogen. To this solution of sodium hydride in DMF was added a solution of 45 (0.8331 g, 2.433 mmol) in DMF (9.2 mL), and the reaction was stirred for 15 min, and then ethyl iodide (0.30 mL, 3.8 mmol) was added, and the reaction was stirred for 1 h. The reaction was poured into water and extracted with ethyl acetate. The combined organic layers were washed with brine, dried over sodium sulfate, filtered, and concentrated in vacuo to yield a crude product that was purified by column chromatography (150 mL SiO 2 , 6% ethyl acetate:hexanes) to give 53 (0.3629 g, 40%) as a colorless oil: 1 H NMR (400 MHz, CDCl 3 165.3, 162.8, 159.7, 155.1, 138.1, 135.3, 125.4, 125.3, 112.7

Methyl 2-(ethyl(4-isobutoxy-3-isopropylphenyl)amino)pyrimidine-5-carboxylate (54)
To a flame-dried, 100 mL round-bottomed flask equipped with a magnetic stir bar was added a 60% dispersion of sodium hydride in mineral oil (0.2377 g, 5.95 mmol). The dispersion of sodium hydride was washed with hexanes (3 mL, twice) and dried under vacuum and suspended in 3.0 mL of DMF under nitrogen. To this solution of sodium hydride in DMF was added a solution of 46 (0.8442 g, 2.458 mmol) in DMF (9.2 mL), and the reaction was stirred for 15 min, and then ethyl iodide (0.30 mL, 3.8 mmol) was added, and the reaction was stirred for 1 h. The reaction was poured into water and extracted with ethyl acetate. The combined organic layers were washed with brine, dried over sodium sulfate, filtered, and concentrated in vacuo to yield a crude product that was purified by column chromatography (150 mL SiO 2 , 6% ethyl acetate:hexanes) to give 54 (0.2698 g, 29.5%) as a white crystalline solid, m.p. 122.8-125.

Methyl 4-(ethyl(4-isobutoxy-3-isopropylphenyl)amino)-2-fluorobenzoate (55)
To a flame-dried, 100 mL round-bottomed flask equipped with a magnetic stir bar was added a 60% dispersion of sodium hydride in mineral oil (0.2461 g, 6.16 mmol). The dispersion of sodium hydride was washed with hexanes (3 mL, twice) and dried under vacuum and suspended in 3.0 mL of DMF under nitrogen. To this solution of sodium hydride in DMF was added a solution of 50 (0.8918 g, 2.301 mmol) in DMF (9.2 mL), and the reaction was stirred for 15 min, and then ethyl iodide (0.30 mL, 3.8 mmol) was added, and the reaction was stirred for 1 h. The reaction was poured into water and extracted with ethyl acetate. The combined organic layers were washed with brine, dried over sodium sulfate, filtered, and concentrated in vacuo to yield a crude product that was purified by column chromatography (150 mL SiO 2 , 6% ethyl acetate:hexanes) to give 55 (0.

Methyl 4-(ethyl(4-isobutoxy-3-isopropylphenyl)amino)benzoate (56)
To a flame-dried, 100 mL round-bottomed flask equipped with a magnetic stir bar was added a 60% dispersion of sodium hydride in mineral oil (0.2417 g, 6.05 mmol). The dispersion of sodium hydride was washed with hexanes (3 mL, twice) and dried under vacuum and suspended in 3.0 mL of DMF under nitrogen. To this solution of sodium hydride in DMF was added a solution of 51 (0.8581 g, 2.513 mmol) in DMF (9.2 mL), and the reaction was stirred for 15 min, and then ethyl iodide (0.30 mL, 3.8 mmol) was added, and the reaction was stirred for 1 h. The reaction was poured into water and extracted with ethyl acetate. The combined organic layers were washed with brine, dried over sodium sulfate, filtered, and concentrated in vacuo to yield a crude product that was purified by column chromatography (150 mL SiO 2 , 6% ethyl acetate:hexanes) to give 56 (0.8655 g, 93.2%) as a white crystalline solid, m.p. 86.8-89.

6-(ethyl(4-isobutoxy-3-isopropylphenyl)amino)nicotinic acid (25) (NEt-4IB)
To a solution of 53 (0.9265 g, 2.501 mmols) in methanol (9.0 mL) was added a solution of KOH (0.4545 g, 8.100 mmols) in water (0.56 mL) and the solution was heated to reflux with stirring for 1 h. The solution was then cooled to room temperature, quenched with 1 N HCl (80 mL), extracted with ethyl acetate, the organic layer was washed with brine, dried over sodium sulfate, filtered, and concentrated in vacuo to give a crude product that was purified by column chromatography ( 6.28. 2-(ethyl(4-isobutoxy-3-isopropylphenyl)amino)pyrimidine-5-carboxylic acid (28) To a solution of 54 (0.7580 g, 2.04 mmols) in methanol (7.3 mL) was added a solution of KOH (0.3846 g, 6.854 mmols) in water (0.46 mL) and the solution was heated to reflux with stirring for 1 h. The solution was then cooled to room temperature, quenched with 1 N HCl (80 mL), and the resulting precipitate was filtered to give a crude product that was purified by column chromatography (

4-(ethyl(4-isobutoxy-3-isopropylphenyl)amino)benzoic acid (30)
To a solution of 56 (0.7035 g, 1.904 mmols) in methanol (6.8 mL) was added a solution of KOH (0.3626 g, 6.462 mmols) in water (0.43 mL) and the solution was heated to reflux with stirring for 1 h. The solution was then cooled to room temperature, quenched with 1 N HCl (80 mL), and the resulting precipitate was filtered to give a crude product that was purified by column chromatography (25 mL SiO 2 , 10-60% ethyl acetate:hexanes) to give 30

Methyl
To a solution of 60 (3.05 g, 16.0 mmols) and methyl 2-(chlorocarbonyl)pyrimidine-5carboxylate 63 (3.19 g, 15.9 mmols) in dichloromethane (35 mL) in a 100 mL round bottom flask was slowly added aluminum chloride (5.6 g) and the resulting mixture was stirred in an oil bath at reflux for 15 min. The reaction solution was cooled to room temperature and quenched by pouring onto 100 mL of an ice water solution. The solution was extracted with ethyl acetate, and the combined organic layers were dried over sodium sulfate, filtered, and concentrated to give a crude product that was purified by column chromatography (silica gel; 1:9 ethyl acetate:hexanes to 1:4 ethyl acetate:hexane) to give pure 65 (1.5869 g, 28%) as an orange, crystalline solid (98.1-103. To a solution of 60 (4.8058 g, 25.5 mmols) and methyl 4-(chlorocarbonyl)benzoate 64 (3.214 g, 16.18 mmols) in dichloromethane (35 mL) in a 100 mL round bottom flask was slowly added aluminum chloride (5.54 g) and the resulting mixture was stirred in an oil bath at reflux for 15 min. The reaction solution was cooled to room temperature and quenched by pouring onto 100 mL of an ice water solution. The solution was extracted with ethyl acetate, and the combined organic layers were dried over sodium sulfate, filtered, and concentrated to give a crude product that was purified by column chromatography (150 mL silica gel; 2.5% ethyl acetate:hexanes) to give pure 66 (4.4007 g, 77.6%) as a white, crystalline solid (120.2-122.  (1-(1,1,3

,3,6-pentamethyl-2,3-dihydro-1 H-inden-5-yl)vinyl)benzoate (67)
To a solution of diisopropylamine (0.66 mL, 4.71 mmols) in THF (2 mL) in a 100 mL round bottom flask was added a 1.6 M solution of n-butyl lithium in hexanes (2.7 mL, 4.32 mmols) at room temperature with stirring. After 15 min of stirring, methyltriphenylphosphonium bromide (1.15 g, 3.22 mmol) was added and the heterogeneous solution was stirred for 20 min after which time the solution became homogeneous and bright canary yellow. This yellow ylide solution was added to a solution of 66 (0.7867 g, 2.24 mmol) in THF (4 mL), and the resulting reaction solution was stirred for 1 h and then poured into water (50 mL) and extracted with ethyl acetate. The combined organic extracts were washed with water, then brine, dried over sodium sulfate, filtered, and concentrated to give a crude oil that was purified by column chromatography (150 mL silica gel; 2.5% ethyl acetate:hexanes to 5% ethyl acetate:hexanes) to give pure 67 (0.3272 g, 41.9%) as a white, crystalline solid (120.2-122. The mixture was extracted with ethyl acetate, and the organic layers were washed with water and saturated sodium chloride, then dried over sodium sulfate, filtered, and concentrated in a 300 mL round bottom flask to give a crude alcohol product that was used without further purification. The alcohol product was dissolved in toluene (98.0 mL) and p-TsOH·H 2 O (1.197 g) was added, and the reaction flask was fitted with a Dean Stark trap and a water condenser. The vessel was evacuated and back-filled with nitrogen three times, and then heated to reflux in an oil bath at 130 • C and stirred for 3 h, during which time water collected in the Dean Stark trap. The reaction was cooled to room temperature, poured into water, and extracted with ethyl acetate. The combined organic layers were dried over sodium sulfate, filtered, and concentrated to give a crude product that was purified by column chromatography (silica gel; 2.5% ethyl acetate: hexanes to 5% ethyl acetate:hexanes) to give pure 68 (0.7969 g, 36.8%) as a white solid (182.9-185.   (1-(1,1,3,3,6-pentamethyl-2,3-dihydro-1H-inden-5-yl)vinyl)pyrimidine-5-carboxylic Acid (32) To a solution of 68 (0.6637 g, 1.8938 mmols) in methanol (12.0 mL) in a 100 mL round bottom flask was added a solution of potassium hydroxide (0.3032 g, 5.40 mmols) in water (0.45 mL). The resulting reaction solution was refluxed with stirring for 1 h in an oil bath at 85 • C. After cooling the reaction solution to room temperature, 1 N HCl (90 mL) was added. The resulting precipitate was filtered and washed with cold water and dried to give crude 32 (0.6143 g, 96.4%). The crude 32 was dissolved in hot ethyl acetate (16.0 mL), hexanes (51 mL) were added, and the homogenous solution was concentrated, filtered, and washed with hexanes to give pure 32 (0.3695 g, 58%) as a white solid (182.7-188. To a suspension of trimethylsulfoxonium iodide (0.760 g, 3.45 mmols) in DMSO (2.5 mL) in a 50 mL 2-neck round bottom flask was added a 20 wt% solution of potassium tert-butoxide in THF (1.94 mL, 3.45 mmols) with stirring at 35 • C. The reaction mixture was stirred for 5 min and then a solution of 68 (0.8061 g, 2.30 mmols) in THF (9.9 mL) was added. The reaction was stirred for 1 h at 35 • C, then allowed to cool to room temperature, at which point 1 N hydrochloric acid (10.0 mL) was added. The resulting solution was extracted with ethyl acetate, the combined organic layers were washed with saturated sodium chloride, dried over sodium sulfate, filtered, and concentrated to give a crude off-white solid that was purified by column chromatography ( 6.40. 2-(1-(1,1,3,3,6-pentamethyl-2,3-dihydro-1H-inden-5-yl)cyclopropyl)pyrimidine-5-carboxylic Acid (33) To a solution of 69 (0.5324 g, 1.4607 mmols) in methanol (9.4 mL) in a 100 mL round bottom flask was added a solution of potassium hydroxide (0.2492 g, 4.44 mmols) in water (0.34 mL). The resulting reaction solution was refluxed with stirring for 1 h in an oil bath at 85 • C. After cooling the reaction solution to room temperature, 1 N HCl (90 mL) was added. The resulting precipitate was filtered and washed with cold water and dried to give crude 33 (0.4932 g, 96.3%). The crude 33 was dissolved in hot ethyl acetate (28.0 mL), hexanes (20 mL) were added, and the homogenous solution was concentrated, filtered, and washed with hexanes to give pure 33 (0.3402 g, 66%) as a white solid (261. 6 (71) The procedure of Bruson and Kroeger was followed. To a solution of 2,5-dichloro-2,5dimethylhexane (70) (11.36 g, 62.04 mmols) in benzene (280 mL) was added aluminum chloride (1.5 g) in a 500 mL round bottom flask equipped with a stir bar and water condenser and the reaction was heated to 75-82 • C for 24 h with stirring under nitrogen. After cooling to room temperature, the reaction solution was poured into 1 N HCl (450 mL) and extracted with benzene. The combined organic layers were washed with water, saturated sodium bicarbonate, water, and finally saturated sodium chloride. The combined organic layers were dried over sodium sulfate, concentrated to a crude oil that was then vacuum distilled with a short-path distillation head at an oil bath temperature of 95-100 • C, and a head temperature of 78 • C for the major fraction, at 0. To a solution of 71 (5.3945 g, 28.646 mmols) and 63 (5.37 g, 26.772 mmols) in dichloromethane (60 mL) was slowly added aluminum chloride (8.8 g) and the resulting mixture was stirred at reflux in an oil bath at 55 • C for 15 min. The solution was then cooled to room temperature and poured into 200 mL of an ice water solution. The resulting mixture was extracted with ethyl acetate. The combined organic layers were washed with saturated sodium chloride, dried over sodium sulfate, filtered, and concentrated to give a crude product that was purified by column chromatography (silica gel, 15% ethyl acetate:hexanes to 20% ethyl acetate: hexanes) to give pure 73 (5.  To a 250 mL flask, 2-fluoro-terephthalic acid 4-methyl ester (2.4067 g, 12.146 mmol) was treated with thionyl chloride (20.0 mL), resulting in a quantitative crude yield of methyl 4-(chlorocarbonyl)-2-fluorobenzoate (72). Acid chloride 72 and compound 71 (2.68 g, 14.243 mmol) were dissolved in 30 mL DCM and aluminum chloride (4.24 g) was slowly added, followed by heating in an oil bath at 55 • C for 15 min. The reaction was quenched in 100 mL of ice water, the product was extracted with ethyl acetate, dried over sodium sulfate, and purified by column chromatography ( (76) To a 250 mL flask, ketone 74 (0.750 g, 2.14 mmol) was combined with dry THF (7.0 mL) at room temperature and a triphenylphosphonium methylide solution-made from the addition of methyltriphenylphosphonium bromide (1.15 g, 3.22 mmol) to a solution of n-butyl lithium (2.7 mL, 1.6 M in hexanes, 4.32 mmol) and diisopropyl amine (0.66 mL, 4.67 mmol) in THF (5 mL)-was added with stirring. The reaction was stirred for thirty minutes, poured into water, and extracted with ethyl acetate. The organic extracts were washed with water, dried over sodium sulfate, and purified by column chromatography (150 mL SiO 2 ) with 2.5% ethyl acetate/hexanes; 5% ethyl acetate/hexane; 20% ethyl acetate/hexane to give pure 76 (0.4293 g, 57.6%) as a colorless, waxy solid, mp 112.1−113. The mixture was extracted with ethyl acetate, and the organic layers were washed with water and saturated sodium chloride, then dried over sodium sulfate, filtered, and concentrated in a 300 mL round bottom flask to give a crude alcohol product that was used without further purification. The alcohol product was dissolved in toluene (110.0 mL) and p-TsOH·H 2 O (5.7782 g, 33.56 mmol) was added, and the reaction flask was fitted with a Dean Stark trap and a water condenser. The vessel was evacuated and back-filled with nitrogen three times, and then heated to reflux in an oil bath at 130 • C and stirred for 3 h, during which time water collected in the Dean Stark trap. The reaction was cooled to room temperature, poured into water, and extracted with ethyl acetate. The combined organic layers were dried over sodium sulfate, filtered, and concentrated to give a crude product that was purified by column chromatography (silica gel; 2.5% ethyl acetate:hexanes to 5% ethyl acetate:hexanes) to give pure 77 (0.2936 g, 5.6%) as a white solid (171. 3  To a suspension of trimethylsulfoxonium iodide (0.3204 g, 1.45 mmols) in DMSO (1.05 mL) in a 50 mL 2-neck round bottom flask was added a 20 wt% solution of potassium tert-butoxide in THF (0.84 mL, 1.49 mmols) with stirring at 35 • C. The reaction mixture was stirred for 5 min and then a solution of 78 (0.3408 g, 0.97 mmols) in THF (4.8 mL) was added. The reaction was stirred for 1 h at 35 • C, then allowed to cool to room temperature, at which point 1 N hydrochloric acid (5.0 mL) was added. The resulting solution was extracted with ethyl acetate, the combined organic layers were washed with saturated sodium chloride, dried over sodium sulfate, filtered, and concentrated to give a crude off-white solid that was purified by column chromatography ( To a solution of 78 (0.1316 g, 0.361 mmols) in methanol (2.5 mL) in a 100 mL round bottom flask was added a solution of potassium hydroxide (0.0803 g, 1.43 mmols) in water (0.18 mL). The resulting reaction solution was refluxed with stirring for 1 h in an oil bath at 85 • C. After cooling the reaction solution to room temperature, 1 N HCl (70 mL) was added. The resulting precipitate was filtered and washed with cold water and dried to give crude 78 (0.1041 g, 82%). The crude 78 was dissolved in hot ethyl acetate (5.0 mL), and the 6.53. Dimethyl-2-hydroxyterepthalate (81) To a solution of hydroxyl-terepthalic acid (80) (9.93 g, 54.5 mmols) in methanol (189 mL) in a 500 mL round bottom flask equipped with a magnetic stir bar and cooled to 0 • C in an ice bath was added thionyl chloride (14.5 mL, 200 mmols) dropwise with stirring. After addition, the flask was equipped with a reflux condenser, placed under a nitrogen atmosphere, and warmed to reflux in an oil bath set at 85 • C and boiled for 2.5 h. The solution was allowed to cool to room temperature, and most of the methanol was removed in vacuo. The crude product was dissolved in ethyl acetate, and the solvent was washed with water followed by brine and then dried over sodium sulfate. The solvent was filtered and the ethyl acetate was removed in vacuo to provide a crude product that was dissolved in warm ethyl acetate (20 mL) and purified by column chromatography (250 mL SiO 2 ) with 10% ethyl acetate/hexanes to give 81 (

3-Hydroxy-4-(methoxycarbonyl)benzoic Acid (82)
The method of Zhang and co-workers was followed. Sodium hydroxide (0.7966 g, 19.9 mmols) was dissolved in water (32 mL), the solution was cooled to 0 • C, and a finely ground powder of dimethyl-2-hydroxyterepthalate (81) (1.0119 g, 4.81 mmols) was added to the solution. The solution was stirred for 1.5 h at 0 • C, and then a solution of 1 N hydrochloric acid was added (12 mL, 12 mmol)m which brought the solution to pH = 9.0, and the insoluble precipitate was filtered off. To the filtrate, an additional amount of 1 N hydrochloric acid (9.5 mL, 9.5 mmol) was added that brought the pH = 1.0 and the resulting precipitate was filtered and washed with cold water to give a crude product (0.67 g) that was purified by column chromatography (150 mL SiO 2 ) with 20% ethyl acetate/hexanes to 70% ethyl acetate/hexanes to give 82 (0.4606 g, 49%) as white solid, m.p. 213.7-216. To a 100 mL round bottom flask charged with 3-acetoxy-4-(methoxycarbonyl)benzoic acid (83) (3.236 g, 13.59 mmols) was added thionyl chloride (22 mL, 300 mmols) and a few drops of DMF. A water condenser was added to the flask, and the solution was refluxed in an oil bath for 1.2 h to give methyl 2-acetoxy-4-(chlorocarbonyl)benzoate (84) in quantitative yield after the excess thionyl chloride was removed in vacuo. To the 100 mL round bottom flask containing 84 was added 1,1,4,4,6-pentamethyl-1,2,3,4tetrahydronaphthalene (85) (6.0508 g, 29.9 mmols) and DCM (30 mL). To the resulting homogeneous solution was slowly added aluminum chloride (3.0 g) at room temperature, with the observed evolution of gas, and the reaction was refluxed for 15 min at 55 • C in an oil bath. The reaction solution was cooled to 0 • C in an ice bath and poured onto 100 mL of an ice water solution. The layers were separated, and the aqueous layer was extracted with ethyl acetate. The combined organic layers were washed with water and then brine, dried over sodium sulfate, filtered, and concentrated in vacuo to give a crude product that was purified by column chromatography (250 mL SiO 2 ) with 1.5% to 5% ethyl acetate/hexanes to give 86 (3.64 g, 70%) as a white solid, m.p. 104. To a 100 mL round bottom flask containing a solution of diisopropylamine (5.67 mL, 40.5 mmols) in THF (16.8 mL) was added a 1.6 M solution of n-butyl lithium in hexanes (22.65 mL, 36.24 mmols) at room temperature, and the reaction was stirred for 15 min followed by the addition of methyl triphenylphosphonium bromide (9.7201 g, 27.21 mmols). After stirring this reaction for 1 h, the resulting solution was added to a 100 mL round bottom flask contain a solution of methyl 2-hydroxy-4- (3,5,5,8,8-pentamethyl-5,6,7,8-tetrahydronaphthalene-2-carbonyl)benzoate (86) (3.8134 g, 10.02 mmols) in THF (8.86 mL) and the resulting reaction solution was stirred for 1 h, poured into 1 N hydrochloric acid (150 mL, 150 mmols), and extracted with ethyl acetate. The combined organic layers were washed with water and brine, dried over sodium sulfate, filtered, and concentrated in vacuo to give a crude product that was purified by column chromatography (150 mL SiO 2 ) with 1.5% to 5% ethyl acetate/hexanes to give a mixture of spots containing 88 and this mixture was again purified by column chromatography (250 mL SiO 2 ) with 1% to 2% ethyl acetate/hexanes to give pure 88 (1.2997 g, 34%) as white solid, m.p. 103.6-106. To a 100 mL round bottom flask containing a suspension of methyl 2-hydroxy-4-(1-(3,5,5,8,8-pentamethyl-5,6,7,8-tetrahydronaphthalen-2-yl)vinyl)benzoate (88) (0.3027 g, 0.800 mmols) in methanol (4.0 mL) was added a solution of potassium hydroxide (0.1634 g, 2.9 mmols) in water (0.20 mL), and the flask was fitted with a condenser and refluxed in an oil bath set to 85 • C for 1.2 h. The solution was cooled to room temperature and acidified with 1 N hydrochloric acid (90 mL, 90 mmols), and the resulting precipitate was filtered and dried to give crude 37 (0.2380 g, 81.6%) as a white solid. This crude material was purified by column chromatography (25 mL SiO 2 ) with 40% ethyl acetate/hexanes to pure ethyl acetate to 8% methanol/ethyl acetate to give pure 37a (0.2316 g, 79%) as white solid, m.p. 220.4-224. To a 100 mL round bottom flask charged with 3-acetoxy-4-(methoxycarbonyl)benzoic acid (83) (4.7646 g, 20.00 mmols) was added thionyl chloride (32 mL, 440 mmols) and a few drops of DMF. A water condenser was added to the flask, and the solution was refluxed in an oil bath for 1.2 h to give methyl 2-acetoxy-4-(chlorocarbonyl)benzoate (84) in quantitative yield after the excess thionyl chloride was removed in vacuo. To the 100 mL round bottom flask containing 84 was added 1,1,4,4-tetramethyl-1,2,3,4-tetrahydronaphthalene (71) (7.4885 g, 39.8 mmols) and DCM (45 mL). To the resulting homogeneous solution was slowly added aluminum chloride (6.7750 g) at room temperature, with the observed evolution of gas, and the reaction was refluxed for 15 min at 55 • C in an oil bath. The reaction solution was cooled to 0 • C in an ice bath and poured onto 100 mL of an ice water solution. The layers were separated, and the aqueous layer was extracted with ethyl acetate. The combined organic layers were washed with water and then brine, dried over sodium sulfate, filtered, and concentrated in vacuo to give a crude product that was purified by column chromatography (250 mL SiO 2 ) with 1.5% to 5% ethyl acetate/hexanes to give 89 Funding: This research was funded by the U.S. National Institutes of Health, grant number 1 R15 CA139364-01A2 and 1R15CA249617-01 to C.E.W., P.W.J. and P.A.M; by R01 HL128447 to J.S.W; and by the Siteman Investment Program to J.S.W.
Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.
Data Availability Statement: CCDC numbers 2110149 and 2110150 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Center via www.ccdc.cam.ac.uk/data_request/cif (accessed on 27 September 2021).
Acknowledgments: Thanks are given to Felix Grun of the High-Resolution Mass Spectrometry Laboratory at University of California, Irvine (UCI). We thank Gayla Hadwiger and Anh Vu for technical support. We thank the Alvin J. Siteman Cancer Center at Washington University School of Medicine for the use of the Flow Cytometry Core. The Siteman Cancer Center is supported in part by an NCI Cancer Center Support Grant P30 CA91842.

Conflicts of Interest:
The authors declare no conflict of interest. Patent applications covering the technologies described in this work have been applied for on behalf of the Arizona Board of Regents.