Potent Antisickling Furaldehyde Analogs for Acute Sickle Cell Therapy: Enhanced Efficacy and Intravenous Formulation Potential

Abstract

Sickle cell disease (SCD) is a debilitating inherited blood disorder characterized by acute crises that require immediate intervention. Aromatic aldehydes that increase hemoglobin (Hb) oxygen affinity (e.g., 5‐HMF) can prevent hypoxia‐induced erythrocyte sickling, but clinical efforts have been hindered by insufficient potency or poor pharmacokinetics. Herein, analogs of 5‐HMF (MMA‐500 series of compounds) are reported to retain 5‐HMF positive physicochemical and pharmacodynamic properties, including safety, solubility, and relatively short duration of action that are essential for their acute use. Two analogs, MMA503 and MMA509, demonstrate over 3.3‐fold greater in vitro antisickling activity than 5‐HMF. This potency is evidenced by significantly enhanced Hb adduct formation, increased oxygen affinity, and robust inhibition of red blood cell sickling in sickle blood assays. X‐ray crystallography further elucidates the Hb binding interactions underlying their potency at the molecular level. MMA509 emerges as a lead candidate and is advanced to formulation studies. An IV formulation of MMA509 in 40% polyethylene glycol 400 achieves ≈13.5 mg mL⁻¹ solubility, enabling rapid attainment of therapeutic drug levels. The potent pharmacologic profile of MMA509, combined with its successful parenteral formulation, highlights its promise as a fast‐acting therapeutic for acute SCD crises as a result of rapid onset expected from IV dosing.

This study reports novel 5‐HMF analogs with potent antisickling activities. Lead compound MMA509 enhances hemoglobin oxygen affinity and prevents erythrocyte sickling. Crystal structure of hemoglobin‐MMA509 complex explains MMA509 mechanism of action. An IV formulation is successfully developed, positioning MMA509 as a promising fast‐acting treatment for acute sickle cell disease crises.

1. Introduction

Sickle cell disease (SCD) is the most common inherited hematologic disorder, affecting over 20 million people worldwide.^[ 1, 2, 3, 4 ^] The pathophysiology of the disease arises as a result of a single point mutation in the β‐globin gene that changes βGlu6 of normal Hb (HbA) to βVal6 to form sickle hemoglobin (HbS).^[ 4, 5, 6, 7, 8 ^] Sickle Hb under hypoxia or when deoxygenated polymerizes into long and rigid fibers as a result of a intermolecular contact between the mutated βVal6 of a deoxygenated HbS molecule and a hydrophobic acceptor pocket of an adjacent deoxygenated HbS, causing sickling of red blood cells (RBCs).^[ 4, 5, 6, 7, 8, 9, 10, 11 ^] The rigid RBCs impair blood flow, causing hemolysis and several interrelated adverse effects that include adhesion of RBCs to vascular endothelium, hemolysis of RBC, oxidative stress, inflammation, vaso‐occlusion crises (VOC), and eventually chronic organ damage that leads to poor quality of life and early death.^{[ 4 , 8 , 12} , ¹³ , ¹⁴ , ¹⁵ , ^{16 – 17 ]} In addition to the chronic adverse events, the disease is also characterized by several acute complications or crises that include VOC pain events, chest syndrome, fatigue, and acute life‐threatening anemia and stroke.^{[ 4 , 8 , 12 , 14 – 16 ]} Some of these acute crises, such as VOC pain events, require immediate and fast treatment responses, however, there are currently no viable drugs to treat such complications.

In recent years, aromatic aldehydes have become an important treatment option for SCD because of the compounds’ ability to prevent the primary pathophysiology of hypoxia‐induced erythrocyte sickling, and potentially the downstream secondary adverse effects.^{[ 4 , 8 ]} These compounds form a reversible Schiff‐base adduct with the N‐terminal αVal1 amines of the two α‐globin chains and shift the allosteric equilibrium of Hb from the T‐state (low‐O₂‐affinity Hb) to the R‐state (high‐O₂‐affinity Hb). The result is an increased concentration of the nonpolymer‐forming oxygenated HbS, which leads to inhibition of RBC sickling.^{[ 4 , 8 , 11 , 18} , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ^{25 – 26 ]} Voxelotor (Oxbryta, aka GBT440) is the first such aromatic aldehyde approved in 2019 to treat the disease.^{[ 4 , 8 , 21 , 22 , 27 , 28 ]} However, Voxelotor was recently withdrawn from the market in 2024, presumably due to modest efficacy and a poor risk–benefit profile. While the long half‐life of Voxelotor (>50 h)^{[ 26 ]} was beneficial for chronic oral therapy, the terminal half‐lives and long time to reach steady state therapeutic drug levels could make Voxelotor unsuitable for urgent treatment. Moreover, Voxelotor is highly hydrophobic with significant solubility problems (aq. solubility < 0.1 mg mL⁻¹ in water at 25 °C),^{[ 29 ]} which precluded simple parenteral delivery for drug loading for an acute condition. 5‐hydroxymethylfurfural (5‐HMF) is a natural aromatic aldehyde, which has been studied extensively for its antisickling potential, undergoing various stages of nonclinical studies and clinical trials.^{[ 4 , 8 , 18 ]} Despite improvement in several clinical symptoms, 5‐HMF failed to advance for chronic use in treating SCD due in part to low potency and poor oral bioavailability as a result of its short half‐life (<2.2 h).^{[ 30 ]} Nonetheless, a short half‐life may be important for drugs treating chronic illness, as they might allow for immediate steady‐state therapeutic drug level, as well as allow physicians to be able to reverse drug effects. Unlike Voxelotor, 5‐HMF is highly soluble (aq. solubility ≈ 400 mg mL^{− 1} at 25 °C; PubChem CID 237,332) and could potentially be formulated for parental acute use to treat SCD.^{[ 31 ]} Given the nontoxic nature of the 5‐HMF pharmacophore and its favorable pharmacologic profile, our group has initiated studies on 5‐HMF analogs and derivatives aimed at enhancing potency while preserving its physicochemical and pharmacokinetic properties, enabling administration via intravenous, subcutaneous, or intramuscular routes for acute treatment. Based on crystallographic study of 5‐HMF and other aromatic aldehydes with Hb that suggested the importance of hydrogen‐bond and hydrophobic interactions for antisickling potency of these compounds, we decided to test several 5‐HMF analogs that we anticipated could make improved potential hydrogen‐bond and/or hydrophobic interactions with the protein, retain similar antisickling activity as 5‐HMF, while also being amenable to formulation for intravenous administration. The compounds (Figure  1 ), termed MMA were obtained from commercial source (Molport; Riga, Latvia) and tested with 5‐HMF as a positive control for their in vitro functional, biological, and ADME properties, as well as their mode of interaction with Hb.

Figure 1.

Chemical structures of 5‐HMF and analogs.

2. Results and Discussion

2.1. The MMA Compounds Exhibit Significant Improvement in Pharmacologic Activity over 5‐HMF

The antisickling activity of aromatic aldehydes for the most part depends on their ability to form Schiff–base interaction with the N‐terminal αVal1 amines of Hb to increase the affinity of the protein for oxygen since the high‐O₂ affinity HbS (R‐state) does not polymerize.^{[ 4 , 8 , 18 – 23 ]} The Schiff–base interaction can be quantified as a Hb‐drug adduct (drug‐modified Hb), while the change in Hb affinity for oxygen is quantified using oxygen equilibrium curve (OEC) studies as a P₅₀ shift, which is the partial pressure (PO₂) at which 50% of Hb is saturated with oxygen (SO₂). Using homozygous sickle (SS) blood from patients, we tested the MMA compounds (MMA502, MMA503, MMA504, MMA505, and MMA509; Figure 1) with 5‐HMF as control for Hb adduct formation, P₅₀ shift, and corresponding RBC sickling inhibition (RBC morphology study) using previously published protocols.^{[ 18 – 20 ]} The results as shown in Figure  2 ,  3 and Table  1 suggest varying potencies of the compounds, which were mostly dose‐dependent (0.5, 1.0, and 2.0 mM) and correlate linearly with each other. MMA503 and MMA509 at 2 mM concentration showed the most potent effect in increasing Hb oxygen affinity (left‐shifting the P₅₀) by 43% and 59% with concomitant RBC sickling inhibition of 70% and 82%, respectively. The positive control 5‐HMF shifted the P₅₀ by 27% and inhibited RBC sickling by 21%, clearly suggesting a ≈3.3 to 4.0‐fold improvement in antisickling activity by MMA503 and MMA509, respectively. MMA502 and MMA505 also potently increased Hb affinity for oxygen by 32% and prevented RBC sickling by 48%, which is ≈2.3‐fold over 5‐HMF. Expectedly, all four compounds MMA502, MMA503, MMA505, and MMA509 also showed significant Hb adduct formation, which at 2 mM were 42%, 35%, 27%, and 53%, respectively, and compared to 34% by 5‐HMF. MMA504 showed the least potent effect, inhibiting sickling by 22%, increasing Hb oxygen affinity by 3%, and almost showed no adduct formation (<1%). The discordant between the antisickling activity and the other two functional matrices (P₅₀ shift and Hb modification) with MMA504 may likely be due to the compounds having additional antisickling mechanism that is oxygen independent and/or involving noncovalent interactions with the protein. We have previously reported aromatic aldehydes that show both O₂‐dependent and O₂‐independent antisickling effects, the latter due to direct polymer destabilization by perturbing a surface‐located αF‐helix on the α‐subunits.^{[ 8 , 20 ]}

Figure 2.

Concentration‐dependent sickling inhibition and oxygen equilibrium (P₅₀) shift in vitro by MMA compounds using SS blood (20% hematocrit). Blood samples were incubated with 0.5, 1, and 2 mM of test compounds at 37 °C for 1 h. Final DMSO was <2% in all samples, including controls (n = 3–4). For the antisickling assay, cell suspensions were incubated under hypoxic conditions (2.5% O₂ gas/97.5% N2 gas) at 37 °C for 2 h, followed by morphological analysis to access % sickling. Sickling inhibition was normalized to untreated control. For OEC analysis, ≈100 μL aliquot from the antisickling assay lysates were subjected to hemoximetry analysis using Hemox Analyzer (TCS Scientific Corp.) to determine P₅₀ shifts. % ΔP₅₀ values were calculated relative to untreated controls. A) Representative morphology of SS cells before and after incubation with MMA503 or MMA509 under 2.5% O₂. B) Representative OECs of lysates from the antisickling study of MMA503 or MMA509 at 2 mM concentration.

Figure 3.

Concentration‐dependent sickling inhibition, P₅₀ shift, and Hb modification (adduct formation) in vitro by MMA compounds using SS blood (20% hematocrit). Blood samples were incubated with 0.5, 1, and 2 mM of test compounds at 37 °C for 1 h. Final DMSO was <2% in all samples, including controls (n = 3–4). For the antisickling assay, cell suspensions were incubated under hypoxic conditions (2.5% O₂ gas/97.5% N2 gas) at 37 °C for 2 h, followed by morphological analysis to access % sickling. Sickling inhibition was normalized to untreated control. For OEC analysis, ≈100 μL aliquot from the antisickling assay lysates were subjected to hemoximetry analysis using Hemox Analyzer (TCS Scientific Corp.) to determine P₅₀ shifts. %ΔP₅₀ values were calculated relative to untreated controls. For Hb adduct formation studies, the clarified lysates from the antisickling study were subjected to cation‐exchange HPLC (Hitachi D‐7000 Series, Hitachi Instruments, Inc., San Jose, CA), using a weak cation‐exchange column. The areas of new peaks, representing HbS adducts, were obtained, calculated as percentage fractions of total Hb area, and reported as levels of modified Hb.

Table 1.

Sickling inhibition, OEC, and Hb‐modification studies using SS RBCs with the MMA compounds.^a)

Compd	% Sickling inhibitionb)			% P50 shiftc)			% Hb modificationd)
	0.5 [mM]	1 [mM]	2 [mM]	0.5 [mM]	1 [mM]	2 [mM]	0.5 [mM]	1 [mM]	2 [mM]
502	26.2 ± 0.5	46.0 ± 2.8	48.1 ± 0.4	14.2 ± 3.4	8.8 ± 0.5	31.5 ± 12.7	9.6 ± 0.0	20.4 ± 6.6	42.4 ± 10.9
503	15.4 ± 4.1	28.0 ± 13.1	69.7 ± 6.0	8.0 ± 1.1	20.0 ± 2.5	43.2 ± 7.4	4.8 ± 2.1	18.0 ± 7.9	34.6 ± 7.1
504	14.6 ± 7.6	18.1 ± 2.7	21.7 ± 0.3	4.4 ± 5.1	5.3 ± 1.5	2.7 ± 1.2	0.0 ± 0.0	0.0 ± 1.0	0.4 ± 0.8
505	16.3 ± 5.5	12.5 ± 0.5	47.5 ± 7.5	5.5 ± 1.2	14.6 ± 2.1	32.0 ± 3.0	4.2 ± 1.6	10.4 ± 2.6	26.7 ± 6.0
509	21.6 ± 14.3	55.0 ± 3.3	81.9 ± 4.4	16.3 ± 1.5	34.5 ± 14.4	58.5 ± 15.3	16.2 ± 0.0	27.6 ± 3.6	53.3 ± 3.9
5‐HMF	N/A	N/A	21.4 ± 0.0	N/A	N/A	27.2 ±  4.9	N/A	N/A	33.8 ± 11.7

^a) All studies were conducted with SS cell suspensions (20% hematocrit) incubated with 0.5, 1.0, and 2.0 mM of each test compound (mean values from 3 to 5 replicates). The final concentration of DMSO was <2% in all samples, including in control samples;

^b) % Sickling inhibition was calculated by normalizing % sickled cells in treated samples to untreated controls (mean values from 3 to 5 replicates);

^c) ΔP₅₀ (%) is the change in P₅₀ (the partial pressure of O₂ at which hemolysates are 50% saturated with oxygen) relative to untreated controls;

^d) %Hb modified Hb calculated as percentage fractions of total Hb area.

2.2. X‐Ray Structures of Hb in Complex with MMA Compounds

The primary interaction between aromatic aldehydes and Hb involves a Schiff‐base formation between the aldehyde moiety of the aromatic aldehyde and the N‐terminal amine of αVal1 of Hb.^{[ 18 – 20 ]} The stability of the Schiff–base interaction, which impacts the pharmacologic effect of these compounds depends on additional interactions between the bound aromatic aldehydes and the α‐cleft residues, including both hydrogen‐bond and hydrophobic interactions. 5‐HMF binds to Hb and makes Schiff–base interaction with the αVal1 amines, as well as hydrogen‐bond interactions with the hydroxyl of αSer131, αSer134, and αSer138 that stabilize the liganded R state Hb conformation relative to the T state Hb conformation.^{[ 18 ]} Like 5‐HMF, we anticipated all the MMA compounds to make hydrogen‐bond interactions with the serine residues, but unlike 5‐HMF we also expected some of the compounds to further make hydrophobic interactions with the protein. We therefore attempted to determine the crystal structures of MMA509, MMA503, and MMA505 in complex with liganded Hb. MMA509 has the most potent P₅₀‐shift/Hb modification/sickling inhibition activities, followed by MMA503, and lastly MMA505. The co‐crystallization experiment was conducted as previously described.^{[ 19 , 20 ]} The crystallographic parameters for MMA503 and MMA509 are summarized in Table  2 . Like 5‐HMF, and as expected, two molecules of each compound bind in a symmetry‐related fashion to the α‐cleft of liganded Hb in the R2 state conformation (Figure  4 ). Each molecule forms a Schiff‐base adduct with the αVal1 amines of both α1‐ and α2‐subunits, with the latter (α2‐subunit) appearing to bind the molecule with higher affinity, as suggested by the electron density of the compounds at the two subunits.

Table 2.

Crystallographic data and refinement statistics for MMA503 and MMA509 in complex with liganded Hb.

Data collection statistics	MMA503	MMA509
PDB ID	7UF7	7UF6
Space group	P212121	P212121
Cell dimensions (Ǻ)	62.61	62.76
	82.12	83.65
	105.12	105.24
Tetramer/AU	1	1
Resolution (Ǻ)	28.22–2.00 (2.05–2.00)	28.30–2.00 (2.05–2.00)
Measured reflections	239,555 (16066)	284,459 (19276)
Unique reflections	37,792 (2722)	38,189 (2780)
Redundancy	6.30 (5.90)	7.4 (6.9)
I/σI	30.2 (8.9)	24.8 (5.0)
Completeness (%)	99.9 (100.0)	100 (99.9)
Rmerge (%)a)	4.5 (22.6)	6.7 (42.0)
Structure refinement
Resolution limit (Ǻ)	28.22 −2.00	28.30–2.00
	(2.07 −2.00)	(2.07 −2.00)
No. of reflections	37,734 (3701)	38,131 (3752)
Rwork (%)	15.7 (18.7)	16.7 (18.3)
Rfree (%)b)	22.3 (28.7)	23.9 (26.7)
R.m.s.d.standard geometry
Bond lengths (Ǻ)	0.008	0.008
Bond angles	1.16°	1.19°
Dihedral angles (%)
Most favored regions	98.2	98.4
Allowed regions	1.8	1.6
Average B‐factors (Å2)
All atoms	20.3	21.6
Protein alone	19.11	20.37
Ligands	17.08	18.50
Water	27.80	29.43

^a) Rmerge = Σ_hklΣ_i/I_hkli − <I_hkli>/Σ_hklΣ_i<I_hkli>;

^b) Rfree was calculated with 5% excluded reflection from the refinement.

Figure 4.

Crystal structures of COHb in complex with MMA compounds, and the comparator 5‐HMF compounds at the α‐cleft. α‐ and β‐subunits are shown in gray and yellow ribbons, respectively, while the bound MMA compounds, as well as the Hb residues Val1, Ser131, and Ser138 are shown in sticks. Structure determined by incubating COHb with 10–15 molar excess of compounds and crystallized using low‐salt precipitant for X‐ray diffraction. A) Bound MMA509 in cyan. B) Bound MMA503 in magenta. C) Two‐dimensional contacts between MMA‐509 molecule. D) Two‐dimensional contacts between MMA‐503 molecule and the protein. The black dashed lines indicate hydrogen‐bond interactions and green dashed lines indicate hydrophobic contacts. E) Bound 5‐HMF in green.

In the MMA509‐Hb complex structure, the compound, in addition to the Schiff–base interaction with α2Val1 nitrogen (the α2‐subunit bound compound), also makes intrasubunit hydrogen‐bond interactions (Figure 4A,C). These involve the tetrahydropyranyl (THP) and furan oxygen atoms of MMA509 interacting with the hydroxyl group of α2Ser131 at a distance of 2.4 and 3.2 Å, respectively. The side‐chain of α2Ser131 assumes two alternate conformations to make these two interactions. The THP ring of MMA509 also makes intersubunit hydrophobic interactions with α1Val135, α1Pro77, and α1Met76 (3.4–4.0 Å). The second MMA509 compound that forms Schiff–base interaction with the α1‐subunit α1Val1 nitrogen also makes similar additional protein interactions as described above for the α2‐subunit bound molecule. The THP/furan rings of the two bound molecules make weak hydrophobic interactions with each other (≈4.0 Å). While MMA509, like 5‐HMF (Figure 4E), forms hydrogen‐bond interactions with the protein, it also makes additional hydrophobic interactions that are missing in 5‐HMF, likely accounting for its significantly greater potency.

In comparison to MMA509, the electron densities of the bound MMA503 molecules are relatively weaker so the compounds were refined at occupancy of 50%. Like MMA509, two molecules of MMA503 also form Schiff‐base interactions with each of the two α‐subunits αVal1 nitrogens (Figure 4B,D). Interestingly, the furan/pyrazole rings of MMA503 are disposed almost 180° from the furan/THF rings of MMA509, resulting in different protein interactions. These involve the furan oxygen and pyrazole NH nitrogen atoms of MMA503 interacting with the hydroxyl group of α1Ser138. There is also an inter‐subunit hydrophobic interaction between the pyrazole ring of MMA503 and β2Trp37. Unlike MMA509, there are no apparent interactions between the two bound MMA503 compounds. MMA503 makes fewer interactions with the protein when compared to MMA509, and may in part explain its relatively weaker potency compared with MMA509. The additional hydrophobic interactions by MMA503 may also explain its greater potency over 5‐HMF.

Unlike MMA509 or MMA503, the MMA505 complex structure showed significantly weaker compound density, which did not improve with refinement to allow for compound fitting (data not shown). Therefore, no compound was included in the model, and the refinement was terminated. This apparent weak binding is reflected in MMA505 weak biological effect compared to MMA503 and MMA509.

The crystallographic study provides insight into the activities of the MMA compounds. MMA509 makes more interactions with the protein, followed by MMA503, and lastly MMA505, consistent with the observed trend in their functional and biological activities. As noted above, the superior pharmacologic activity of MMA509 or MMA503 when compared to 5‐HMF could in part be due to hydrophobic interactions in addition to hydrogen‐bond interactions, the former absent in 5‐HMF. MMA502, MMA503, MMA505, and MMA509 share a common structural feature with 5‐HMF, which is the presence of H‐bonding group in the same β‐position of the side chain attached to C5 of the furan ring (Furan‐CH₂‐X, where X is O, N, or F). This group makes crucial hydrogen bonding contacts with the protein, and may also help explain the variation in potency among the compounds. Like MMA503 or MMA509 or MMA505, the fluorine substituent on MMA502 might also be capable of making hydrogen‐bond interactions with the protein, thus explaining its moderate potent activity. The hydroxyl of MMA504 is far removed from the furan ring, and might not be in the right position and/or orientation to make such hydrogen‐bond interactions, explaining its weak activity.

2.3. MMA Compounds Show Quick Onset and Relatively Short Duration of Action

Unlike Voxelotor, most aromatic aldehydes, including 5‐HMF, are highly susceptible to oxidative and/or reductive metabolism to the inactive carboxylic and/or alcohol analogs, respectively.^{[ 32} , ^{33 – 34 ]} 5‐HMF has a short half‐life and poor oral bioavailability due to this metabolic instability, which in part prevented its development for chronic use. Nonetheless, this metabolic instability could be advantageous for parenteral formulation for acute use. We tested the three compounds, MMA503, MMA505, and MMA509 for their time‐dependent (0, 1, 1.5, 3, 6, and 8 h) effect on Hb oxygen affinity using normal whole blood as previously reported.^{[ 19 ]} The observed P₅₀ shifts values (%P₅₀ shifts) were plotted as function of time [h]. The time‐dependent study gives indication of the compounds' duration of action since whole blood contains enzymes that are known to metabolize the aldehyde moiety into the inactive alcohol or acid.^{[ 19 ]} The study confirmed the trend in the biological potencies observed with SS blood. As expected MMA509 showed the most potent effect at all time points (P₅₀ shift of 60% at 1.5 h), followed by MMA503 (43% at 1.5 h), then 5‐HMF (35% at 3h), and lastly MMA505 (11% at 1 h). Nonetheless, similar to 5‐HMF, all the compounds started to decline in their biological effect after 3 h with MMA505 completely losing activity after 3 h (Figure  5 ).

Figure 5.

Time‐dependent P₅₀‐shift of Hb in normal whole blood samples (hematocrit 20%) following incubation with 2 mM test compounds at 37 °C for 24 h (n = 2). Oxygen dissociation parameters were measured using ABL 800 Automated Blood Gas Analyzer.

2.4. MMA509 Showed No CYP Inhibition

Based on the promising result of MMA509, we studied this compound for its inhibitory potential of seven major drug metabolizing human cytochrome P450 (CYP) enzymes (CYP1A2, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP3A4, CYP2B6) using pooled human liver microsomes as published previously.^{[ 20 , 21 , 35 ]} 5‐HMF was also included as a control. The probe substrates for the CYP enzymes, as well as the selective CYP inhibitors are provided in Table  3 . Like the natural product 5‐HMF, MMA509 showed no inhibition of any of the liver CYP isozymes, suggestive of a potential low liability for toxic effect.

Table 3.

In vitro CYP inhibition with MMA509 and 5‐HMF.

Enzyme	Substrate	IC50 [μM]
		Control inhibitor	MMA509	5HMF
CYP1A2	Tacrine	α‐Naphthoflavone (0.007)	>100	>100
CYP2C8	Amodiaquine	Quercetin (1.90)	>100	>100
CYP2C9	Tolbutamide	Sulfaphenazole (0.098)	>100	>100
CYP2C19	Mephenytoin	Ticlopidine (0.965)	>100	>100
CYP2D6	Dextromethorphan	Quinidine (0.061)	>100	>100
CYP3A4	Midazolam	Ketoconazole (0.024)	>100	>100
CYP3A4	Testosterone	Ketoconazole (0.023)	>100	>100
CYP2B6	Bupropion	Ticlopidine (0.262)	>100	>100

2.5. Solubility and Formulation Development

Although in silico predictions classified MMA509 as ‘soluble’ in water, we experimentally evaluated its solubility in practical solvent systems to enable an acute IV formulation. The choice of polyethylene glycol 400 (PEG 400), propylene glycol, ethanol, and Tween 80 as vehicle components is supported by formulation literature and regulatory precedent.^{[ 36 ]} Using these vehicles, we found that MMA509 can be formulated at a high concentration in an aqueous cosolvent mixture. A UV–vis scan of MMA509 identified a strong absorption peak at 281 nm (λmax), which was used for quantification. The solubility of MMA509 was determined in several IV‐compatible solvent systems (20% ethanol, 40% PEG 400, 2% Tween 80, 40% propylene glycol, and 40% PEG 400 + 2% Tween 80, all % v/v in water). Among these, 40% PEG 400 yielded the highest solubility, dissolving ≈13.5 mg mL^{− 1} of MMA509. The combination of 40% PEG 400 with 2% Tween 80 provided a slightly lower solubility (no significant synergistic improvement), while 2% Tween 80 alone resulted in moderate solubility. The 40% propylene glycol and 20% ethanol solutions showed considerably lower solubility. The rank order of solubilization efficacy was: 40% PEG 400 > 40% PEG 400 + 2% Tween 80 > 2% Tween 80 > 40% propylene glycol > 20% ethanol. Notably, the 40% PEG 400 vehicle produced a clear, single‐phase solution of MMA509 at the target concentration with no visible precipitation. Based on these findings, we selected 40% PEG 400 (supplemented with 2% Tween 80) in water as the optimal vehicle for the IV formulation of MMA509. These simple vehicle components are pharmaceutically accepted excipients with a history of safe intravenous use and are capable of delivering MMA509 at clinically relevant concentrations, thereby validating the feasibility of acute intravenous administration of MMA509 for the treatment of sickle cell disease.

3. Conclusion

Aromatic aldehydes have for several years been studied for their potential to treat sickle cell disease by preventing hypoxia‐induced sickle Hb polymerization and the concomitant RBC sickling, with one such compound, Voxelotor making it to the clinic for the treatment of the disease.^{[ 4 , 8 , 18 – 23 ]} As noted earlier, Voxelotor has been recalled presumably due to poor risk–benefit profile. Nonetheless, most pharmacologically promising aromatic aldehydes, such as Voxelotor are hydrophobic with significantly long therapeutic half‐life, making them unsuitable for parenteral use or for rapidly achieving therapeutic steady‐state drug levels needed to treat acute, potentially life‐threatening symptoms. Leveraging 5‐HMF's favorable profile, we identified two analogs, MMA509 and MMA503, that retain the compound's solubility and safety while achieving over 3.3‐fold greater antisickling efficacy. These compounds show promise for SCD treatment, especially via parenteral administration to rapidly attain therapeutic hemoglobin modification. In particular, MMA509 emerged as a lead candidate combining potent pharmacologic activity with ease of formulation. We successfully developed an IV formulation of MMA509 in a PEG 400–based vehicle, achieving ≈13 mg mL^{− 1} solubility and enabling immediate drug availability. The enhanced potency of MMA509, coupled with its successful IV formulation, underscores its potential as a fast‐acting therapeutic to prevent or treat acute SCD complications.

4. Experimental Section

4.1.

4.1.1.

Materials and General Procedures

The MMA compounds (see Figure 1) were purchased from Molport (Riga, Latvia). All compounds, including the positive control 5‐HMF were solubilized in DMSO and used for the test. The corresponding control experiments (without test compound) also contain DMSO. Normal whole blood was collected from adult donors (>18 years) after informed consent, in accordance with regulations of the IRB for Protection of Human Subjects (IRB# HM1) by the VCU Human Research Protection Program/Institutional Review Board. Leftover blood samples from patients with homozygous SS were obtained and utilized, based on an approved IRB protocol (IRB# 11‐008151) by the Committees for the Protection of Human Subjects of the Institutional Review Board at the Children's Hospital of Philadelphia. All experimental protocols and methods were performed in accordance with institutional (VCU and CHOP) regulations. Purified human adult Hb A in 50 mM Bis Tris buffer, pH 6.8, was prepared from discarded human blood as previously described,^{[ 37 , 38 ]} and stored at −80 °C until ready for crystallization.

In Vitro Hb Modification, OEC, and Sickling Inhibition Studies Using Sickle Whole Blood

The effect of the MMA compounds and the positive control, 5‐HMF on RBC sickling, Hb modification, and Hb oxygen equilibrium was studied utilizing SS blood following previous procedure.^{[ 18 – 20 ]} For the antisickling study, SS blood (hematocrit 20%) was incubated under air in the absence or presence of all the MMA compounds (0.5, 1 , and 2 mM) at 37 °C for 1 h, followed by incubating the mixture under hypoxic conditions (2.5% O₂) at 37 °C for 2 h. Aliquots were fixed 2% glutaraldehyde solution and under anaerobic condition were subjected to microscopic morphological analysis. For the oxygen equilibrium study, about 100 μL aliquot of clarified lysate of residual samples from the above antisickling study was added to 4 mL of 0.1M potassium phosphate buffer, pH 7.0, in cuvettes and subjected to hemoximetry analysis using Hemox Analyzer (TCS Scientific Corp., New Hope, PA) to assess P₅₀ shift, which is the partial pressure of oxygen (PO₂) at which 50% of Hb is saturated with oxygen (SO₂). Degree of Hb O₂ affinity shift (ΔP₅₀) was expressed as percentage fractions of control DMSO‐treated samples. The Hb adduct formation study also performed using clarified lysates that were subjected to a cation‐exchange HPLC (Hitachi D‐7000 Series, Hitachi Instruments, Inc., San Jose, CA), using a weak cation‐exchange column (Poly CAT A: 30 mm × 4.6 mm, Poly LC, Inc., Columbia, MD).

Crystallography

We conducted X‐ray crystallography study to determine the mode of interactions between Hb and the MMA compounds, following previously published methods.^{[ 19 , 20 ]} Briefly, the compounds were incubated with 30 mg dL⁻¹ of CO‐ligated Hb in 10:1 molar ratio and crystallized using 10%–20% PEG6000, 100 mM Hepes, pH 7.4. X‐ray quality crystals were only obtained for MMA509, MMA505, and MMA503 in complex with liganded Hb. Diffraction data from the ensuing crystals were collected at 100 K using Rigaku MicroMax 007HF X‐ray Generator, Eiger R 4M Detector, and Oxford Cobra Cryo‐system (The Woodlands, TX). The crystals were first cryoprotected with 80 μL mother liquor mixed with 62 μL of 50% PEG6000. The diffraction data were processed using the CrysAlisPro 41_64_122a (Rigaku) and the CCP4 suite of programs.^{[ 39 ]} The crystal structures were refined using the Phenix program, with the native isomorphous R2‐state crystal structure (PDB ID 1BBB) as a starting model.^{[ 40 , 41 ]} Model building and correction were carried out using COOT.^{[ 40 ]} The final Rfactor/Rfree of MMA509, MMA505, and MMA503 are 16.84/22.80, 16.55/20.75, and 16.20/21.50, respectively. Detailed crystallographic data obtained for the MMA503 and MMA509 structures are shown in Table 2.

In Vitro Time‐Dependent Hb Oxygen Equilibrium Studies Using Normal Whole Blood

The time‐dependent effect of the MMA compounds and the positive control, 5‐HMF on Hb oxygen affinity was studied using normal whole blood following previous procedure.^{[ 20 ]} Briefly, normal blood samples (hematocrit 20%) were incubated with the compounds at 37 °C for 24 h. Aliquots were taken at various time points and incubated in TM8000 Thin film tonometer (Meon Medical Solutions) for 10 min at 37 °C, and allowed to equilibrate at oxygen tensions of 6, 20, and 40 mm Hg. The samples were then aspirated into an ABL 800 Automated Blood Gas Analyzer (Radiometer) to determine the partial pressure of oxygen (pO₂), and Hb oxygen saturation values (SO₂). The measured values of pO₂ (mm Hg) and SO₂ at each oxygen tension values were then subjected to a nonlinear regression analysis using the program Scientist (Micromath, Salt Lake City, UT) to estimate P₅₀ as previously reported.^{[ 20 ]} The observed P₅₀ shifts values in %P₅₀ shifts were plotted as function of time (h).

In Vitro CYP‐450 Inhibition Studies

The most promising MMA compound, MMA509 in addition to the positive control 5‐HMF was studied for their inhibitory potential of seven major drug metabolizing human cytochrome P450 (CYP) enzymes (CYP1A2, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP3A4, CYP2B6) using pooled human liver microsomes as previously published.^{[ 20 , 21 , 35 ]} The probe substrates include tacrine (CYP1A2), amodiaquine (CYP2C8), tolbutamide (CYP2C9), mephenytoin (CYP2C19), dextromethorphan (CYP2D6), midazolam (CYP3A4), testosterone (CYP3A4), and bupropion (CYP2B6). The following selective CYP inhibitors, naphthoflavone (CYP1A2), quercetine (CYP2C8), sulfaphenazole (CYP2C9), ticlopidine (CYP2C19), quinidine (CYP2D6), ketoconazole (CYP3A4), and ticlopidine (CYP2B6) were used as positive controls. The optimized reaction mixtures (200 μL) contained a final concentration of 0.2–0.5 mg mL⁻¹ pooled human liver microsomes, 2 mM NADPH in 100 mM potassium phosphate, pH 7.4 buffer with 5 mM MgCl2, and MMA509 or 5‐HMF concentration of 0.1–100 µM. The assays were performed in duplicate in 96‐well plates at 37 °C for 10–60 min. The reaction was terminated with addition of methanol, followed by incubation at 4 °C for 10 min and centrifuged at 4 °C for 10 min. The effect of test compounds on formation of the respective probe substrate metabolites was determined using LC‐MS/MS, and used to calculate IC₅₀ value, which is the test compound concentration that resulted in 50% inhibition.

Solubility and Formulation Development: Materials:

Ethanol, propylene glycol, and Tween 80 (polyoxyethylene sorbitan oleate) were purchased from Sigma–Aldrich (St. Louis, MO, USA). PEG 400 was obtained from Acros Organics (Geel, Belgium). UV–vis calibration: A UV–visible spectrophotometric method was used to quantify MMA509 in solution. A stock solution of MMA509 (1 mg mL⁻¹ in ethanol) was prepared and scanned from 200 to 400 nm (Shimadzu UV‐2600, Kyoto, Japan) to determine the wavelength of maximum absorbance (λmax). The scan showed a pronounced λmax at 281 nm. A series of standard solutions (2, 4, 6, 8, and 10 μg mL⁻¹ in ethanol) was then prepared, and their absorbances at 281 nm were measured to construct a calibration curve. The plot of absorbance versus concentration was linear over this range, with a regression coefficient R ² = 0.9949, confirming the reliability of the method for concentration determinations.

Solubility determination

The equilibrium solubility of MMA509 was assessed in several IV‐compatible cosolvent systems selected for their common use in injectable formulations. These solvent systems were 20% ethanol, 40% PEG 400, 2% Tween 80, 40% propylene glycol, and a mixture of 40% PEG 400 + 2% Tween 80 (all percentages v/v in sterile deionized water). Excess MMA509 (solid) was added to 5 mL of each solvent mixture in sterile glass vials. The vials were tightly capped and shaken continuously at room temperature (25 ± 2 °C) for 48 h to ensure saturation equilibrium. After 48 h, each sample was centrifuged at 5000 rpm for 10 min to pellet any undissolved drug. The clear supernatants were carefully collected and filtered through 0.22 μm syringe filters to remove any remaining particulate matter. The filtrates were appropriately diluted with ethanol (if necessary) and analyzed at 281 nm using the calibrated UV spectrophotometer to determine the concentration of dissolved MMA509. The solubility in each solvent system was calculated in mg mL⁻¹ from the measured absorbance using the calibration curve. All experiments were performed in triplicate.

Conflict of Interest

The authors declare no conflict of interest.

Author Contributions

Abdelsattar M. Omar, Moustafa E. El‐Araby, Osheiza Abdulmalik and Martin K. Safo: conceptualization. Mohini S. Ghatge, Akua K. Donkor, Tarek A. Ahmed, Osheiza Abdulmalik, Martin K. Safo, Albert Opare, Rana T. Alhashimi, Benita Balogun, Anfal S. Aljahdali, Trevohn N. Robinson, Mariana Macias, Salma Roland, Faik N. Musayev: investigation. Moustafa E. El‐Araby, Abdelsattar M. Omar, Tarek A. Ahmed, Yan Zhang, Albert Opare, and Martin K. Safo: writing (original draft preparation). Abdelsattar M. Omar, Yan Zhang, Martin K. Safo, and Osheiza Abdulmalik: writing—review and editing by all authors; supervision. Abdelsattar M. Omar, Martin K. Safo, and Osheiza Abdulmalik: funding acquisition.

Institutional Review Board Statement

The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Boards (IRB) of Virginia Commonwealth University (IRB #HM1) and the Children's Hospital of Philadelphia (IRB #11‐008151). Blood samples from healthy adult donors (AA) at Virginia Commonwealth University (VCU) and leftover blood samples from individuals with homozygous sickle cell (SS) at the Children's Hospital of Philadelphia (CHOP) were collected and utilized after obtaining informed consent, in compliance with institutional regulations.

Omar Abdelsattar M., El‐Araby Moustafa E., Ahmed Tarek A., Donkor Akua K., Opare Albert, Ghatge Mohini S., Aljahdali Anfal S., Alhashimi Rana T., Robinson Trevohn N., Balogun Benita, Macias Mariana, Roland Salma, Zhang Yan, Musayev Faik N., Abdulmalik Osheiza, Safo Martin K., ChemMedChem 2025, 0, e202500507. 10.1002/cmdc.202500507

Data Availability Statement

The data that support the findings of this study are openly available in [Repository name: RCSB Protein Data Bank] at [Repository URL: https://www.rcsb.org/], reference number [REF 7UF6 and 7UF7].