Thionated aminofluorophthalimides reduce classical markers of cellular inflammation in LPS-challenged RAW 264.7 cells

Abstract

Herein, we present the synthesis of several fluorinated pomalidomide derivatives and their thionated counterparts with subsequent biological evaluation against classical markers of cellular inflammation. Treatment in LPS-challenged cells effected varying reductions in levels of secreted TNF-α and nitrite relative to basal amounts. While arene fluorination and thioamidation had marginal and sporadic effects on TNF-α production, specific 7-position fluorination combined with subsequent increases in carbonyl thionation produced compounds 11, 14, and 15 which demonstrated corresponding and escalating anti-nitrite activities concurrent with minimal cellular toxicity. In this regard, compound 15 displayed roughly 96 % cell viability combined with a 65 % drop in nitrite production when supplied to RAW cells challenged with 60 ng/mL LPS. When a focused family of fluorinated isomers were directly compared, the analogous 5-fluorinated isomer 17 displayed comparable minimal toxicity but markedly less anti-nitrite activity versus 15 in RAW cells challenged with 70 ng/mL LPS. Compound 15 was subsequently screened in human liver microsomes for preliminary Phase 1 analysis where it demonstrated heightened stability relative to its non-fluorinated counterpart 3,6′-dithiopomalidomide 4, a result in line with the expected metabolic fortitude provided by fluorination at the sensitive pomalidomide 7-position.

Despite its harrowing past,¹ thalidomide (1) experienced a pharmacological resurgence in 1964 when Dr. Jacob Sheskin serendipitously discovered² its efficacy in treating erythema nodosum leprosum, a debilitating and potentially chronic systemic inflammation associated with multiple types of leprosy.³ Subsequent laboratory and clinical endeavors reaffirmed the utility of thalidomide-based therapy for suppressing inflammatory ailments, and reciprocal biochemical evidence for thalidomide-induced repression of proinflammatory cytokines followed shortly thereafter.⁴ In this regard, thalidomide attenuates the proinflammatory immune response primarily through degradation of TNF-α mRNA; other complementary pathways modulate activation of NF-κB, a key transcriptional regulator of TNF- α and numerous interleukins.⁵ Ensuing synthetic explorations around the thalidomide backbone combined with advances in medicinal chemistry led to the development of several pharmacologically-relevant analogs (Fig. 1) including lenalidomide (2) and pomalidomide (3). These and other closely related materials have shown significant clinical application for treating multiple myeloma and myelodysplastic syndromes as well as a variety of neurological and proinflammatory disorders.⁶

Fig. 1.

Thalidomide (1) and pharmacologically-relevant analogs. Our contribution (3,6′-dithiopomalidomide, 4) serves as the basis for the research presented within.

Recent research from our laboratory has demonstrated that incorporation of sulfur into specific carbonyl groups of pomalidomide can have a marked effect on the in vitro reduction of many classic markers of cellular inflammation. For example, our compound 3,6′-dithiopomalidomide 4 showed similar anti-TNF- α activity to that of pomalidomide in both RAW 264.7 cellular studies and in rodents challenged with lipopolysaccharide but additionally attenuated inflammation-induced iNOS and COX-2. Additionally, 4 reduced these proinflammatory cytokines as well as astrogliosis and autophagy-related proteins in models of traumatic brain injury, Alzheimer’s Disease, and stroke.⁷

Given the metabolic fate of pomalidomide (Fig. 2)⁸ combined with our previous interests in aminophthalimide pharmacology, we wondered what (if any) consequence site-specific arene fluorination would levy upon our pomalidomide-based anti-inflammatory agents. For instance, it is widely accepted that fluorination of small organic molecules and pharmaceuticals can attenuate undesired oxidative metabolism due to the increased strength of typical C-F bonds (116 kcal/mol) relative to analogous C—H iterations (99 kcal/mol).⁹ Simply put, many classical redox-active liver proteins do not possess the enzymatic vigor to regularly disrupt resilient C-F linkages, thereby impeding some of the traditional means by which the body processes a drug. Additionally, the increased lipophilicity rendered by fluorination can perturb bioavailability and pharmacokinetic properties, while the similar size of F and H (1.47 A vs 1.20 A, van der Waals radius) makes the substitution ideal for molecules involved in sterically influenced processes (such as enzyme-ligand binding interactions).¹⁰

Fig. 2.

Pomalidomide metabolites observed in humans. The predominant arene localized oxidation products are highlighted here, with 7-hydroxypomalidomide the minor metabolite and 5-hydroxypomalidomide the major contributor.

For our initial foray, we envisioned that a small family of fluorinated pomalidomides and their associated thionated analogs could be easily obtained and screened in a timely manner. We began by synthesizing a collection of 7-fluoropomalidomides with all consequent materials stemming from an initial condensation of 3-aminoglutatrimide with the readily available¹¹ 3-fluoro-6-nitrophthalic acid 7 (Fig. 3). The resulting fluorinated nitrothalidomide 8 proved a valuable springboard, permitting access to three distinct branches of our main synthetic tree. For example, monothionated 9 was obtained by a direct thionation with Lawesson’s reagent, while careful control of reduction conditions provided access to the fluorinated pomalidomide and N-alkylpomalidomide cohorts. In this regard, reductive amination over PtO₂ in an acetone/acetic acid yielded isopropylated analog 10, while simple hydrogenation over Pd/C yielded primary amino compound 11. Subsequently, and for each corresponding pomalidomide analog, monothionation at the 6′-glutrarimide carbonyl was effected with Lawesson’s reagent, while 3,6′-dithionation was achieved using P₄S₁₀-pyridine in 1,4-dioxane.¹²

Fig. 3.

Synthetic scheme for 7-fluorophthalimides. General Conditions: (i) 3-aminopiperidine-2,6-dione, AcOH, 130 °C, 16hr. (ii) Lawesson’s Reagent, 1,4-Dioxane, reflux, overnight; 33 % yield. (iii) H₂ (50psi), PtO₂ (cat), 3:1 acetone/AcOH, H₂SO₄ (cat); 58 % yield. (iv) H₂ (50psi), 10 % Pd/C (cat), H₂SO₄ (cat), AcOH; 61 % yield. (v) Lawesson’s Reagent, toluene, 105 °C, 24hr; 46 % yield. (vi) P₄S₁₀-pyridine complex, 1,4-Dioxane, 100 °C, 40hr; 36 % yield. (vii) Lawesson’s Reagent, 1,4-Dioxane, reflux, 20hr; 29 % yield. (viii) P₄S₁₀-pyridine complex, 1,4-Dioxane, 100 °C, 20hr; 29 % yield.

With this first cohort of materials in hand, we screened against classical markers of inflammation using the LPS-activated RAW 264.7 cell model, a system readily adapted to identifying modulators of nitrite and TNF-α production.¹³ We ultimately selected 10 μM for further study, as this concentration provided adequate compound solubility concurrent with minimal cellular toxicity, all while furnishing notable effects on the markers of interest. The resulting inhibitory activities of our compounds on LPS-induced nitrite and TNF-α are summarized below (Table 1 and Fig. 4).

Table 1.

Summary of biology data for compounds. the data are expressed as mean ± S.E.M. of the observed values. N.S. refers to not significant, * refers to P < 0.05, **** refers to P < 0.0001 compared to appropriate vehicle + LPS control.

Compound (10uM)	% Change versus Control (± s.e.m.)
	Viability	TNF-a	Nitrite
	104.9 ± 2.1 (N.S.)	86.0 ± 4.4 (N.S.)	56.8 ± 2.2 (****)
	44.9 ± 1.9 (****)	51.3 ± 3.8 (****)	4.7 ± 1.0 (****)
	99.9 ± 0.4 (N.S.)	87.1 ± 3.4 (*)	110.1 ± 4.4 (N.S.)
	100.9 ± 1.1 (N.S.)	104.8 ± 4.6 (N.S.)	90.3 ± 3.7 (N.S.)
	112.3 ± 1.7 (*)	94.4 ± 2.8 (N.S.)	74.8 ± 1.2 (****)
	74.1 ± 4.2 (****)	74.7 ± 2.3 (****)	32.3 ± 1.1 (****)
	112.4 ± 2.9 (*)	107.0 ± 4.0 (N.S.)	59.0 ± 1.6 (****)
	96.1 ± 0.5 (N.S.)	104.1 ± 2.5 (N.S.)	35.3 ± 3.0 (****)

Fig. 4.

Summary of biological parameters; first cohort. Effects of pomalidomide analogs (10 μM) on (a) LPS-activated RAW 264.7 cell viability. (b) LPS-induced generation of TNF-α protein released into the cell culture media. (c) LPS-induced generation of nitrite measured in the cell culture media. [LPS] used = 60 ng/mL. The percentage change in analyte relative to the drug vehicle control group are shown on the Y-axis, and the different pomalidomide analogs are indicated on the X-axis. The vehicle control (DMSO + LPS) levels are represented by the dashed line originating from the Y-axis at 100 %. Four wells on the cell culture plates (n = 4) were used for each treatment group, and the data are expressed as mean ± S.E.M. of the observed values. N.S. refers to not significant, * refers to P < 0.05, **** refers to P < 0.0001 compared to appropriate vehicle + LPS control.

We were pleased to find that our compounds, in general, were very well tolerated with most retaining > 80 % cell viability at 10 μM treatment. In this regard, the most readily apparent outlier was monothionated analog 9. In fact, when we compared our two nitrated analogs directly (Fig. 5a), it was clear that thionation intensified cytotoxicity relative to the native oxo-material 8, while nitrate and TNF- α levels presumably diminished in accordance with associated cell death. Next, the N-isopropylated cohort revealed that increases in thionation effected corresponding decreases in nitrite levels concurrent with slight cytotoxicity (Fig. 5b). A comparable (albeit much weaker) overall trend was observed with TNF- α. Finally, when the primary amino analogs 11, 14, and 15 were compared, it was very clear that thionation led to reciprocal and significant decreases in nitrite while TNF- α levels were left unaffected (Fig. 5c). However, in contrast to the trends observed in the nitrated and isopropylated materials, each member of this group maintained > 95 % viability regardless of thionation level employed. This important distinction helped balance and clarify the differences in nitrite reduction observed between 13 and 15, and ultimately allowed us to move forward with our investigations in a more focused manner.

Fig. 5.

Direct comparison of closely related analogs, (10 μM). Results are presented as % change from vehicle + LPS control. (a) Nitro analogs. (b) Isopropyl analogs. (c) Amino analogs. The drug-induced changes in biological variables are shown on the Y-axis, the compound number on the X-axis. Drug-induced changes in cell viability (green), TNF- α (red) and nitrite (blue) are provided. Data are expressed as mean ± S.E.M. of the observed values (n = 4). N.S. refers to not significant, * refers to P < 0.05, **** refers to P < 0.0001 compared to appropriate vehicle + LPS control.

With the results of this preliminary screen in hand, and the arene metabolic patterns of pomalidomide still fresh in our minds, we envisioned a complementary study might be prudent, especially-one comparing directly the most logical supplementary candidates (Fig. 6). Pomalidomide 3 and dithionated pomalidomide 4 were readily available from our previous research efforts as noted above. We therefore turned our attention to synthesizing the two appropriate and specific 5-fluoropomalidomides 16 and 17. The syntheses began with a simple condensation of 4-fluoro-3-nitrophthalic acid¹⁴ with 3-aminoglutatrimide (Fig.7). The corresponding 5-fluoronitrothalidomde 19 was obtained in moderate yield. Subsequent reduction to 16 and dithionation yielded desired adduct 17.

Fig. 6.

Pomalidomides and fluorinated analogs with metabolic relevance.

Fig. 7.

Synthetic scheme for production of 5-fluorophthalimides. (i) 3-aminopiperidine-2,6-dione, AcOH, 130 °C, 16hr; 78 % yield (ii) H₂ (50psi), 10 % Pd/C (cat), AcOH (cat); 50 % yield. (iii) P₄S₁₀-pyridine complex, 1,4-Dioxane, 100 °C, 18hr; 31 % yield.

This concise six-compound grouping was collectively evaluated in a similar RAW cell assay as before, furnishing a direct, iterative appraisal of fluorine regiochemistry and carbonyl dithionation across a uniquely focused pomalidomide family (Fig. 8 and Table 2). As was expected and in line with our previous studies,¹⁵ dithiopomalidomide 4 reduced nitrite levels significantly relative to pomalidomide 3. Interestingly, the newly prepared 5-fluorinated iterations found middle ground between our typical pomalidomides and their 7-fluoroinated isomers, as oxoversion 16 was surpassed by counterparts 3 and 11, while 5-fluoro-dithiopomalidomide 17 demonstrated a modest efficacy sandwiched between that of dithiopomalidomide 4 and the superior 7-fluoro-dithiopomaldiomide 15 (Fig. 9).

Fig. 8.

Summary of biological parameters; focused metabolic cohort. Effects of pomalidomide analogs (10 μM) on (a) LPS-activated RAW 264.7 cell viability. (b) LPS-induced generation of TNF-α protein released into the cell culture media. (c) LPS-induced generation of nitrite measured in the cell culture media. [LPS] used = 70 ng/mL The percentage change in analyte relative to the drug vehicle + LPS control group are shown on the Y-axis, and the different pomalidomide analogs are indicated on the X-axis. The vehicle control (DMSO + LPS) levels are represented by the dashed line originating from the Y-axis at 100 %. Four wells on the cell culture plates (n = 4) were used for each treatment group, and the data are expressed as mean ± S.E.M. of the observed values. N.S. refers to not significant, * refers to P < 0.05, **** refers to P < 0.0001 compared to appropriate vehicle + LPS control.

Table 2.

Inhibitory activities of the focused metabolic family. [LPS] = 70 ng/mL. The data are expressed as mean ± S.E.M. of the observed values. N.S. refers to not significant, * refers to P < 0.05, **** refers to P < 0.0001 compared to appropriate vehicle + LPS control. Additionally, for values annotated with N.S. $, some of the data points were not normally distributed and required a Kruskal-Wallis test ANOVA with Dunn’s multiple comparisons test. Using this analysis these data points did not attain a statistical significance.

Compound (10uM)	% Change versus Control (± s.e.m.)
	Viability	TNF-α	Nitrite
	95 ± 1.2 (N.S.)	95 ± 4.8 (N.S.)	91 ± 1.6 (N.S.)
	92 ± 1.8 (N.S.)	113 ± 5.4 (N.S.)	43 ± 1.0 (N.S.) $
	102 ± 0.8 (N.S.)	98 ± 1.5 (N.S.)	86 ± 3.9 (N.S.)
	95 ± 3.0 (N.S.)	109 ± 2.4 (N.S.)	22 ± 1.7 (****)
	90 ± 2.6 (N.S.)	93 ± 4.1 (N.S.)	100 ± 5.5 (N.S.)
	92 ± 1.6 (N.S.)	114 ± 3.9 (N.S.)	40 ± 1.2 (N.S.) $

Fig. 9.

Brief summary of anti-nitrite activity across the focused metabolic family.

In the logical progression, we wondered how our fluorinated materials would behave under the more metabolically demanding conditions of a Phase-1-type microsomal analysis. We reasoned that at least one of our fluorinated analogs should demonstrate a measurable degree of stability over our primary non-fluorinated 3,6′-dithiopomaldomide 4, especially when subjected to metabolically rigorous human liver homogenates. For this initial comparison, we chose 7-fluoro-dithiopomaldiomide 15 owing to its combination of appealing anti-inflammatory properties and general ease of synthesis. The reactions on each sample were initiated by addition of human liver microsomes in the presence or absence of an NADPH regenerating system, and compound disappearance was monitored via LC/MS (Fig. 10).

Fig. 10.

Phase 1 microsomal analysis comparing compounds 4 and 15. Reactions were carried out with 100 mM potassium phosphate buffer, pH 7.4, in the absence and presence of an NADPH regenerating system. Reactions (in triplicate) were initiated by addition of the human liver microsomes to the incubation mixture (compound final concentration was 10 μM; 0.5 mg/mL microsomes). Compound disappearance was monitored via LC/MS. The % compound remaining is shown on the Y-axis. The time points following the initiation of the metabolism studies are shown on the X-axis. The % compound remaining are presented in the presence (left) and absence (right) of NADPH. The P values / significance markers are provided. Data are expressed as mean ± S.E.M. of the observed values (n = 3). Comparing the relative abundances of the compounds at the specified times; ** refers to P < 0.01, *** refers to P < 0.001.

As seen in Fig. 10, 3,6′-dithiopomaldomide 4 demonstrated a moderate level of instability, both with and without NADPH fortification. After 1hr incubation, more than half of the material had been consumed. In contrast, 7-fluoro-dithiopomaldiomide 15 demonstrated heightened resistance under the same conditions, yielding 7 % more stability in the absolute sense, equating to roughly 16 % more material remaining when compared to its non-fluorinated partner. In all cases, the materials seemed to be more sensitive to + NADPH processes versus those lacking NAPDH fortification.

Herein we reported the synthesis and biological evaluation of a focused family of aminofluorophthalimides and selected thionated analogues. Through careful examination of the associated cell viability and anti-inflammatory data, we established preliminary structure activity relationships. In general, the materials were well tolerated, with only compound 9 displaying considerable cellular toxicity at 10uM. While anti-TNF-α activities were marginal across all compounds, nitrite levels were clearly affected by both fluorination and thionation. Site-specific fluorination and thioamidation provided compound 15, which when dosed at 10uM in stimulated RAW cells, effected an 65 % drop in LPS-induced nitrite production concurrent with minimal perturbations to TNF- α and cell viability. Furthermore, during the human liver microsomal analyses, 15 demonstrated enhanced metabolic fortitude relative to its non-fluorinated counterpart 4, a result in line with an increased stability afforded by a site-specific 7-fluorination at the metabolically-sensitive pomalidomide locus. In this regard, we plan to expand upon 15 and the associated 7-fluorianted family in hopes of developing more compounds with superior anti-inflammatory proprieties concurrent with minimal toxicity and greater metabolic resistance. Additionally, we are working towards similar studies featuring 5-fluorinated materials 16 and 17, especially those screening for metabolic differences and similar biological outcomes not necessarily reflected in the corresponding initial RAW cell findings.

Data availability

Data will be made available on request.