Anti-sickling agents Number:7,160,910 from the United States Patent and Trademark Office (PTO) owispatent The provides compounds for the treatment of sickle-cell disease. In particular, the invention provides 5-membered heterocyclic anti-sickling agents that are highly effective and non-toxic, and methods for their use. The compounds include analogues and derivatives of naturally occurring occurring 5-hydroxymethyl-2-furfuraldehyde, 5-Ethyl-2-furfuraldehyde, 5-Methyl-2-furfuraldehyde, and 2-furfuraldehyde, and prodrug forms of the compound.<H sickling, agents, Anti, sickling, a, 7160910.html, Patent, pto, 7160910

Title: Anti-sickling agents

Abstract: The provides compounds for the treatment of sickle-cell disease. In particular, the invention provides 5-membered heterocyclic anti-sickling agents that are highly effective and non-toxic, and methods for their use. The compounds include analogues and derivatives of naturally occurring occurring 5-hydroxymethyl-2-furfuraldehyde, 5-Ethyl-2-furfuraldehyde, 5-Methyl-2-furfuraldehyde, and 2-furfuraldehyde, and prodrug forms of the compound.

Patent Number: 7,160,910 Issued on 01/09/2007 to Safo, et al.

Inventors: Safo; Martin K. (Richmond, VA), Danso-Danquah; Richmond (Richmond, VA), Nokuri; Samuel (Richmond, VA), Musayev; Faik N. (Glen Allen, VA), Joshi; Gajanan S. (Glen Allen, VA), Burnett; James C. (Richmond, VA), Abraham; Donald J. (Midlothian, VA)
Assignee: Xechem International, Inc. (New Brunswick, NJ)

Appl. No.: 10/725,935

Filed: December 3, 2003

Current U.S. Class: 514/365 ; 514/448; 514/461

Current International Class: A61K 31/341 (20060101); A61K 31/381 (20060101); A61K 31/427 (20060101)

Field of Search: 514/365,448,461

References Cited [Referenced By]

U.S. Patent Documents


4987132	January 1991	Mase
6300332	October 2001	Chang et al.

Other References

Zaugg et al., J. Biol. Chem. (1977), vol. 252, No. 23, pp. 8542-8548. cite- d by examiner.

Primary Examiner: Solola; Taofiq
Attorney, Agent or Firm: Jain; Mishrilal Whitam; Michael

Government Interests

This invention was made using funds from grants from the National Institutes of Health having grant numbers NIH K01HL04367 and NIH R01HL32793. The United States government may have certain rights in this invention.

Claims

We claim:

1. A method for treating sickle cell disease in a patient in need thereof, comprising the step of administering to said patient a compound of the formula: ##STR00021## where R1 is CHO or a thiozolidine group of the following formula: ##STR00022## R2, R3 and R4 are the same or different and selected from the group consisting of H, OH, alkyl, halogen and hydroxy-alkyl; R6 and R7=H; X=O or S; and wherein said compound is administered in sufficient amount to said patient to ameliorate sickle cell disease, with the proviso that R2,R3 and R4 cannot be H or alkyl at the same time when R1 is CHO and X is O.

2. The method of claim 1, wherein R2 and R3=H; R4 is H, alkyl or hydroxyalkyl; and the general formula of the compound is: ##STR00023##

3. The method of claim 1, wherein said compound is of the formula: ##STR00024## where R2 and R3=H.

4. The method of claim 1, wherein said compound is of the formula: ##STR00025## where R2, R3=H; R4 is H, alkyl or hydroxy-alkyl; R6 and R7=H; and X=O.

5. The method of claim 4, wherein said compound is of the formula: ##STR00026##

6. The method of claim 1, wherein said compound is selected from the group consisting of a compound of the formula: ##STR00027##

Description

DESCRIPTION

BACKGROUND OF THE INVENTION

1. Field of the Invention

The invention generally relates to agents for the treatment of sickle-cell disease. In particular, the invention provides 5-membered heterocyclic anti-sickling agents that are highly effective and non-toxic, and methods for their use.

2. Background of the Invention

Sickle cell disease is one of the most prevalent hematologic genetic disorders in the world (Ingram, 1956; Pauling, et al. 1949) that occurs as a result of a single point mutation of Glu6 in Hb to Val6 in sickle hemoglobin (HbS). Two quaternary structures are known for Hb, the deoxy conformation (tense), and the oxygenated conformation (relaxed). When the allosteric equilibrium is shifted toward the relaxed state, a high-affinity Hb is obtained that readily binds and holds oxygen, while the converse is true for the tense state. Perutz (1970) and Baldwin & Chothia (1979) elucidated at atomic resolution the tetrameric structures of the tense (T) and relaxed (R) forms of Hb. The tetramer is composed of two .alpha..beta. dimers that are arranged around a twofold axis of symmetry. This arrangement yields a central water cavity, with two openings; the .alpha.- and .beta.-clefts. The source of the tension in the T state is due to crosslinking salt bridges and hydrogen bonds between the subunits, as well as preferential binding of an indigenous allosteric effector of Hb, 2,3-diphosphoglycerate (2,3-DPG) that stabilize the T state by forming salt bridges between the two .beta.-subunits (Arnone, 1992). The T-R transition occurs as a result of uptake of oxygen which leads to the disruption of many of the T state intersubunit interactions, as well as expulsion of the 2,3-DPG. The allosteric transition results in a rotation of the .alpha.1.beta.1 dimer relative to the .alpha.2.beta.2 dimer by 12 15.degree. (Baldwin & Chothia, 1979). The R state structure has a smaller central water cavity, as well as fewer intersubunit salt bridges and hydrogen bonds. For a long period of time, the allosteric equilibrium of Hb embodied in the two-state MWC model (Monod, et. al., 1965) was believed to involve only the T-R transition, and the R state quaternary structure was thought to be the only relaxed conformer. However, recent crystallographic and other studies have revealed the existence of multi relaxed Hb states, including R2 and others that exist in solution with R (Silva, et al. 1992; Smith, et al., 1991; Mueser, et al., 2000). There is still a controversy about the physiological importance of all these relaxed states, and how they relate to one another in Hb allostery. Silva et al., (1992) and Smith et al., (1991) suggested that the R2 quaternary structure is an intermediate between the T and R structures. Further analysis has shown that R2 is not an intermediate in the T to R transition, but rather, it is another relaxed end-state structure (Janin & Wodak, 1993; Doyle, et al., 1992). Srinivasan & Rose (1994) have further suggested that R2 may be the physiologically relevant end state and that the R structure is an intermediate structure trapped between the R2 and T states by the high-salt crystallization conditions. In contrast, the R2 structure formation is believed to be favored by low-salt that mimic in vivo condition (Silva et al., 1992; Srinivasan & Rose, 1994).

Hb and HbS have almost identical positions for all amino acids, even in the A helix of the chains where the mutation occurs. The presence of the Val6 results in hydrophobic interaction between the mutation region of one Hb molecule and a region defined by Phe85 and Leu88 in the heme pocket of another Hb molecule. This interaction occurs only in the deoxygenated HbS (deoxyHbS), and induces polymerization of the deoxyHbS molecules into fibers. The formation of HbS polymers causes the normally flexible red blood cells to adopt rigid, sickle like shapes that block small capillaries and cause both local tissue damage and severe pain. The disease is also characterized by other symptoms, including hemolysis, which gives rise to anemia and jaundice, elevation of bilirubin level leading to high incidence of gall stones and impairment of hepatic excretory function. Other clinical features include leg ulceration, pneumonia, enlarged liver and spleen. Other studies on the gellation of deoxyHbS and various Hb variants have also provided crucial information on other contact points on the Hb that are important in stabilizing the HbS fibre (Adachi & Asakura, 1980; Bunn, et al., 1986). There are various therapeutic strategies to treat sickle cell disease (SCD), including (1) Pharmacological modulation of fetal hemoglobin (HbF): HbF has been shown to decrease HbS polymerization, and there are several agents that are known to induce HbF formation by possibly reactivating the genetic switch for HbF (Olivieri & Weatherall, 1998). Examples of such agents include 5-azacytidine, hydroxyurea and cytosine arabinoside (Mehanna, 2001). Unfortunately, there are serious toxic side effects associated with this therapy as a result of high doses and frequency of administration (Edelstein, 1985), (2) Bone marrow transplantation: Bone marrow transplant has also been used as a total gene replacement therapy for HbS in extreme cases (Hillery, 1998, Johnson, 1985). This approach is very expensive and has its own inherent toxicities and risks (Hillery, 1998), (3) Blood transfusion: This is one of the most common SCD therapies, however, repeated blood transfusions are known to be associated with the risks of infectious diseases, iron overload and allergic reactions (Ballas, 1999), (4) Opioid analgesics: This therapy is necessary to deal with pain crisis, however, opioid therapy often results in addiction and/or seizures and/or depression, (Ballas, 1999), (5) Erythrocyte membrane acting agents: Since the sickling process is partly dependent on intracellular concentration of sickle Hb, agents that induce cell swelling (Asakura, 1980) or inhibit cell dehydration (Orringer & Berkwitz, 1986) could decrease the HbS concentration and help delay the polymerization process, and (6) Antigelling agent or HbS modifiers: These compounds interfere with the mechanism of polymerization by either binding directly to or near contact site(s) of the deoxyHbS to inhibit the polymerization process or act directly on HbS to shift the allosteric equilibrium to the more soluble high-affinity HbS.

In blood, Hb is in equilibrium between the T and the relaxed states. The Hb delivers oxygen via an allosteric mechanism, and the ability for the Hb to release or take oxygen can be regulated by allosteric effectors. The allosteric equilibrium between the T and relaxed states (FIG. 1) shows a typical oxygen equilibrium curve (OEC) for Hb, i.e. a plot of the percentage of oxygen bound by Hb against the partial pressure of oxygen. When the allosteric equilibrium is shifted towards the relaxed state (left shift of the curve), a high-affinity Hb is obtained that more readily binds and holds oxygen while a shift toward the T state (right shift of the curve) results in a low-affinity Hb that more easily releases oxygen. An increase in the naturally occurring allosteric effector, 2,3-DPG in red cells right shifts the OEC as does an increase in temperature and decrease in pH (Reeves, 1980). An increase in pH and lowering of the temperature and DPG levels left shifts the OEC. The degree of shift in the OEC is reported as an increase or decrease in P.sub.50 (partial pressure of oxygen at 50% Hb saturation). Regulating the allosteric equilibrium to the relaxed conformation has been of been of interest in medicine, In particular, the identification of non-toxic compounds that efficiently bind to HbS and produce high-affinity HbS which does not polymerize have been clinically evaluated as antisickling agents to treat SCD. There is an ongoing need to identify such compounds to be used as antisickling agents to treat sickle cell anemia. See, for example, the use of vanillin (Abraham, 1991), 12C79 (Fitzharris, 1985), furfural (Zaugg, et al., 1997), and substituted isothiocyanates (Park, et al. 2003).

SUMMARY OF THE INVENTION

The present invention provides compounds that are highly effective, specific and non-toxic anti-sickling agents, as well as prodrug forms of the compounds. The compounds are based on naturally occurring 5-hydroxymethyl-2-furfuraldehyde (5HMF or AMS-13), 5-Ethyl-2-furfuraldehyde (5EF), 5-Methyl-2-furfuraldehyde (5MF) and 2-furfuraldehyde (FUF), and include analogues and derivatives of these compounds. Methods for using the compounds to treat sickle cell disease are also provided.

It is an object of this invention to provide a method for treating sickle cell disease in a patient in need thereof. The method comprises the step of administering to the patient one or more compounds of the formula:

##STR00001## where R1 is CHO, or an aldehyde protecting group; R2, R3 and R4 are the same or different and are H, OH, alkyl, alkoxy, hydroxy-alkyl, halogen, aryl or O-aryl; and X=NH, O, S, Se or P. The compound is administered in sufficient quantity to ameliorate symptoms of sickle cell disease. In one embodiment of the invention, R1 is CHO; R5 is H, alkyl or aryl; and m=1 6; and the compound is selected from the group consisting of

##STR00002## In another embodiment of the invention, R1 is a heterocyclic ring aldehyde protecting group; R6 and R7=H or alkyl and may be the same or different; and n=0 4; and the compound is selected from the group consisting of

##STR00003## Examples of the compound within this group include the following (wherein in some of the examples R8 and R9 can be the same or different and are H, alkyl or aryl)

##STR00004## ##STR00005## In preferred embodiments of the invention, the compound is selected from the group consisting of

##STR00006##

The invention also contemplates the use of the compounds of the general formula:

##STR00007## where R1 is CHO, or an aldehyde protecting group, R2, R3 and R4 are the same or different and are H, OH, alkyl, alkyoxy, hydroxy-alkyl, halogen, aryl or O-aryl; and X=NH, O, S, Se, or P to treat symptoms of patients suffering sickle cell anemia including particularly jaundice and elevated levels of bilirubin. A sufficient quantity of one or more of the compounds of the above general formula is administered to ameliorate or reduce jaundice or levels of bilirubin, respectively.

The invention also contemplates the use of the compounds of the general formula:

##STR00008## where R1 is CHO, or an aldehyde protecting group, R2, R3 and R4 are the same or different and are H, OH, alkyl, alkyoxy, hydroxy-alkyl, halogen, aryl or O-aryl; and X=NH, O, S, Se, or P in the treatment of patients with cancer. One or more of the compounds would be administered to a patient in a quantity sufficient to change the oxygen carrying capacity of the blood to a state beneficial to eradicating the cancer. In addition, similar to the methods described in U.S. Pat. No. 5,705,521 to Abraham, which is herein incorporated by reference, the compounds of this invention can be administered to a patient that is or is going to undergo radiation therapy to assist in the effectiveness of the radiation to destroy the cancer.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1. The effects of temperature, pH, and DPG levels on the Hb oxygen equilibrium curve.

FIG. 2. Structures of some furanic as well as other compounds discussed in the application.

FIG. 3. Oxygen equilibrium curves of suspensions of SS cells in the absence (1) and presence of 5 mM vanillin (2), FUF (3) and 5HMF (4). Note the significant left shift caused by addition of 5HMF.

FIG. 4. Cation-exchange HPLC studies of the hemolysate from the 5HMF-reacted SS cells. Peak 1 is normal HbS, while peak 3 is the modified HbS. Peak 2 represents HbS1, a minor component that separates and elutes earlier from the major HbS peak.

FIG. 5. Kaplan-Meier Survival plot of control and 5HMF (AMS-13)-treated Tg sickle mice.

FIG. 6. Mean survival time of control and 5HMF (AMS-13)-treated Tg sickle mice.

FIG. 7. The effect of 5HMF (AMS-13) on the percentage of sickled cells in the arterial blood of Tg sickle mice that were exposed to severe hypoxia (5% oxygen). Without treatment (Control), the percentage of sickled cells increased sharply and the animal died within 15 min. Pretreatment of the mice with 5HMF prolonged the survival time significantly. The drug also reduced the percentage of sickled cells.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS OF THE INVENTION

The present invention provides antisickling agents. The agents are 5-membered heterocyclic compounds that are based on the naturally occurring substances 5HMF, 5EF, 5MF and FUF. The compounds include analogues and derivatives of 5HMF, 5EF, 5MF and FUF. By "analogue" and "derivative" we mean compounds that possess the same basic structure as the parent molecule, but in which the various R groups as depicted in the Formulas below are replaced as indicated in the descriptions of the formulas.

The compounds of the present invention exhibit high antisickling activity, specificity, and very low toxicity. The compounds contain a central ring moiety and are of general Formula 1

##STR00009## where R1 is CHO, or an aldehyde protecting group that leaves the central ring moiety and allows the central ring moiety to revert to an aldehyde in the body (heterocyclic ring moieties being preferred); R2, R3 and R4 are the same or different and are H, OH, alkyl, alkoxy, hydroxy-alkyl, halogen, aryl and O-aryl; and X=NH, O, S, Se, and P.

Preferred embodiments of the compounds are given in Formulas 2 7 below, where R1 is shown as CHO, where R2 R4 are shown and are moieties as set forth above, and, with respect to formulas 4 and 5, where R4 is an alkoxy or hydroxy-alkyl; R5=H, alkyl, or aryl, and m=1 6.

##STR00010##

In preferred embodiments of the invention, the antisickling compounds are those of Formula 5 where m=1 6 and R5 is hydrogen.

The invention further provides prodrugs of these compounds. By "prodrug" we mean a form of the compound that contains at least one "protective" chemical group, the presence of which protects the aldehyde moiety of the compound against metabolism/degradation until the prodrug is in an environment appropriate for removal of the protecting group(s) and "release" of the active form of the compound. The overall effect of the protecting group is to increase the bioavailability of the active aldehydic compound once the compound reverts to the aldehyde. One preferred generic prodrug of the compounds of the present invention is represented in Formula 8:

##STR00011## Where R2, R3 and R4 are the same or different and can be H, OH, alkyl, alkoxy, hydroxylkyl, halogen, aryl or O-aryl; R6=H or alkyl; R7=H or alkyl; and X=NH, O, S, Se, or P. Use of protecting groups of this type, (e.g. the amino acids cysteine and homocysteine) have the advantage of resulting in the release of a non-toxic amino acid upon removal of the protecting group. It should be understood that other aldehyde protecting groups may be used in the practice of this invention. Protection groups of this type are taught, for example, in U.S. Pat. No. 6,251,927 to Lai et al., (Jun. 6, 2001) the complete contents of which are hereby incorporated by reference. These include, but are not limited to conversion of the aldehyde to the corresponding imine, alcohol, acetal, ester, macrcocyclic ester/acetal, macrocyclic ester/imine, hemiacetal, and the like.

Preferred variations of this type of compound are given in Formulas 9 13:

##STR00012## where R8=H, alkyl or aryl and m=1 6.

A second preferred generic prodrug of the compounds of the present invention is shown in Formula 14

##STR00013## where R2, R3 and R4 are the same or different and can be H, OH, alkyl, alkoxy, hydroxylkyl, halogen, aryl or O-aryl; X=NH, O, S, Se, or P; and n=0 4. These prodrugs have the advantage of being made in a relatively facile manner, but do not have the advantage of producing a natural product upon release from the prodrug. Variations of this type of prodrug are given in Formulas 15 19,

##STR00014## where R9=H, alkyl, or aryl and m=1 6.

The present invention also provides a method for treating sickle cell disease, and to treat symptoms of sickle cell disease such as jaundice and elevated levels of bilirubin. The method involves administering to a patient with sickle cell disease a quantity of at least one compound of the present invention sufficient to ameliorate or reduce symptoms of the disease, e.g. to decrease or eliminate jaundice and elevated bilirubin. The compounds may be administered as the compounds themselves, or in the form of a prodrug.

The compounds of the present invention are potent antisickling agents. Those of skill in the art will recognize that the amount of such an agent that is to be given to a patient will vary depending on several factors, including but not limited to blood volume, hematocrit, patient age, gender, weight, overall physical health, presence of other disease conditions, the particular compound being administered, and the like. However, the amount will generally be in the range of from about 50 to about 1000 mg/kg and more preferably in the range of from about 100 to about 750 mg/kg. For example, an adult sickle cell patient with a blood volume of 4 L and 25% hematocrit will need about 378 630 mg/kg of the most potent compound, 5HMF, and more preferably the dose will be in the range from 200 630 mg/kg.

The compounds of the present invention may be administered in any form that is suitable for delivering an active amount of the compound to the patient being treated. For examples, the compound may be administered as solid pills or capsules, liquids for oral administration, injectable formulations, etc. The compounds may be provided alone or in combination with other constituents, and may be provide in pure or salt form (e.g., organic or inorganic salts, etc.). In addition, the compounds may be formulated with carriers such as aqueous or oil based vehicles, and can be accompanied by preservatives (e.g., methyl paraben or benzyl alkonium chloride (BAK)), surfactants (e.g. oleic acid), solvents, elixirs, flavoring agents (in the case of oral delivery), starch, and other materials (preferably those which are generally regarded as safe (GRAS)). In a preferred embodiment, the compound will be administered as a preparation for oral administration in an alcohol-based or aqueous-based carrier.

Likewise, the method of administration may be any of a wide variety of methods that are well known to those of skill in the art, such as intravenous, intradermal injection, subcutaneous injection, intramuscular injection, intraperitoneal injection; oral, rectal and buccal delivery; transdermal delivery; inhalation; etc. In a preferred embodiment, the method of administration is oral delivery. Further, the compounds of the present invention may be administered in conjunction with other known sickle cell disease treatments, or treatments for other related or unrelated disease conditions (e.g. treatment for anemia), and with other beneficial ancillary regimens such as dietary supplements, exercise regimens, and the like.

While a primary use of the anti-sickling agents of the present invention is to treat sickle cell disease, those of skill in the art will recognize that the agents may be used in other applications for which it is beneficial to cause destabilization of the tense (T) states of hemoglobin, and the switching of the allosteric equilibrium in favor of the high-affinity Hb in the form of the R2-state Hb. This aspect of the invention is further illustrated in Example 5 below. For example, the invention also contemplates the use of the compounds in the treatment of patients with cancer. One or more of the compounds would be administered to a patient in a quantity sufficient to change the oxygen carrying capacity of the blood to a state beneficial to eradicating the cancer. In addition, similar to the methods described in U.S. Pat. No. 5,705,521 to Abraham, which is herein incorporated by reference, the compounds of this invention can be administered to a patient that is or is going to undergo radiation therapy to assist in the effectiveness of the radiation to destroy the cancer.

The compounds may also be used for research purposes. In particular, the generic compounds may serve as parent structures for rational drug design of additional derivatives and analogues. Examples include but are not limited to other forms of the compounds that are more or less active, or more or less stable; that have altered solubility properties; or that have moieties that serve to target the compounds to a desired location, e.g. across the cell membrane. Alternatively, the compounds provided here may serve as parent structures for the design of other forms of prodrugs.

Some of the compounds used in the practice of the present invention are naturally occurring, for example, 5-Hydroxymethyl-2-furfuraldehyde (5HMF or AMS-13), 5-Ethyl-2-furfuraldehyde (5EF), and 5-Methyl-2-furfuraldehyde (5MF) (see FIG. 2). 5HMF is present in many foods such as sweet potatoes, fruits, honey, milk, beer, tomato products, cigarette smoke, and coffee, where the concentration sometimes exceeds 6 g/kg. Commercially, 5HMF is prepared from the fructose portion of sugar.

Methods for synthesizing the prodrug compounds of the present invention are also described herein, with exemplary synthesis schemes being given in Example 4 below. Such organic synthesis methods of protecting aldehydes are well-known to those of skill in the art.

EXAMPLES

Example 1

Structural Basis for the Potent Antisickling Effect of a Novel Class of 5-Membered Heterocyclic Aldehydic Compounds: In Vitro Tests and X-ray Crystallographic Studies of the Furanic Compounds, 5HMF, 5MF, 5EF and FUF

Chemical structures of 5HMF, 5MF, 5EF and FUF are shown in FIG. 2.

Regulating the Hb allosteric equilibrium to the relaxed conformation has been of interest in medicine, as compounds that bind to HbS that produce high-affinity HbS, which do not polymerize have been clinically evaluated as antisickling agents to treat sickle cell anemia. Two such compounds are vanillin (Abraham, et al., 1991) and 12C79 (Beddell, et al., 1984; Fitzharris, et al., 1985; Orringer, et al., 1986). Both compounds are aldehydes and form Schiff base adducts with Hb. Clinical studies with 12C79 (FIG. 2) showed that intravenous infusion of this compound (40 mg/kg) in patients with SCD resulted in the formation of compound-Hb adducts at levels of more than 30% without any adverse effect (Orringer, et al., 1988). Vanillin (FIG. 2) is a food flavoring compound, and because it is relatively non-toxic, makes it a very attractive therapeutic agent for SCD. A low-resolution structure of T state Hb complexed with vanillin showed that the compound binds to two different sites: a primary site near His103, Cys104 and Gln131, and a secondary site between His116 and His117 (Abraham, et al., 1991). Based on the results of the X-ray crystallographic analysis and functional studies, Abraham, et al., (1991) suggested that vanillin acts to decrease the polymerization of HbS by shifting the allosteric equilibrium toward the high-affinity HbS molecule in the form of R state; as well as stereospecific inhibition of the polymerization of T state HbS. Additional studies of several analogs of vanillin by Abraham, et al., (1995) and Boyiri, et al., (1995) showed that these compounds, unlike vanillin, bind to the N-terminal Val1 of T state Hb, and surprisingly effect opposite shifts in the OEC. Agents such as, 5-formylsalicylic acid (FSA, FIG. 2), which form Schiff base interactions with the N-terminal Val1 nitrogen, and provide groups for both salt bridge and hydrogen bonding with the opposite dimer (across the Hb two-fold axis), shift the OEC toward the right. In contrast, agents such as, 3,5-dimethyl-4-hydroxybenzaldehyde (DMHB, FIG. 2), which also bind to the N-terminal Val residue of T state Hb in a similar fashion as FSA without any salt bridge interactions with the opposite dimer, shift the OEC toward the left.

In the current studies, we combined the use of aldehydic covalent modifiers of Hb with our knowledge of the molecular regulation of the allosteric equilibrium to produce potent antisickling compounds that should be clinically safe. Specifically, we examined 5-hydoxymethyl-2-furfural (5HMF) and several of its analogs, including furfural (FUF), 5-methyl-2-furfural (5MF) and 5-ethyl-2-furfural (5EF) (all shown in FIG. 2) for their antisickling potencies. These compounds were found to significantly shift the allosteric equilibrium to the high-affinity Hb, and also act as potent inhibitors of homozygous sickle red blood (SS) cell sickling. One of the compounds, 5HMF modifies HbS by 70% compared to 15% for vanillin. 5HMF also inhibits in vitro SS cell sickling by 90%, four times more than vanillin. Also, in vivo antisickling studies using sickle cell transgenic (tg) mice show 5HMF to prolong the life of the hypoxic mice about four times longer compared to control. 5HMF is found in everyday food, and has LD.sub.50 of 2.5 5.0 g/Kg (US EPA, 1992) compared to 1.58 for vanillin. Thus we now have a compound which is safe, more potent than any known aldehydic antisickling agent, able to transverse red blood cell to react and modify HbS. In addition, 5HMF makes an ideal scaffold upon which to build more potent and safe compounds.

X-ray crystallographic studies of Hb complexed with these compounds indicate that they form Schiff base adducts in a symmetrical fashion with the N-terminal .alpha.Val1 nitrogens of Hb. Remarkably, two co-crystal types were isolated during these experiments: one crystal type was found to be composed of the low-affinity or tense (T) state Hb quaternary structure in complex with the compounds; the other crystal type was composed of high-affinity or relaxed state Hb (with a R2 quaternary structure) in complex with the compounds. Furthermore, the examined heterocyclic aldehydes were found to bind strongly to the R2 state, but weakly to the T state. Crystallization of the same compounds with liganded Hb resulted in only relaxed R state crystals, which also indicated weak compound binding. These results suggest that the examined heterocyclic aldehydes prevent polymerization of sickle hemoglobin (HbS) and inhibit the sickling of SS cells by stabilizing HbS in the high-affinity R2 state. Comparing the high resolution crystal structures of 5HMF and vanillin bound to Hb, shows 5HMF to bind much stronger, an indication that 5HMF may reside longer at the binding site. This explains why 5HMF is many fold potent than vanillin. Most importantly the stronger binding of 5HMF to Hb may translate into an even longer half-life and increased bioavailability for 5HMF compared to vanillin, with a concomitant decrease in the dosage needed for therapy. The biological, as well as the crystallographic studies of the furanic compounds reveal for the first time the exact molecular mechanism for the antisickling effects of covalently modifying aldehydes that bind to N-terminal .alpha.Val1 nitrogens of Hb. These examined compounds also represent a new class of potentially therapeutic agents for treating sickle cell disease (SCD).

Experimental Procedure

Materials and General Procedures. The following compounds: vanillin, FUF, 5MF, 5EF and 5HMF were purchased from Aldrich Chemical Company. Normal red blood (AA) cells were collected from adult donors. SS cells were obtained from patients with SCD. Purified human adult Hb in 50 mM Bis Tris buffer, pH 6.8, was prepared from discarded human blood as previously described (Safo and Abraham, 2003). Oxygen Equilibrium Studies with Normal Whole Blood. Normal blood samples (hematocrit 40%) in the presence of 5 mM Vanillin, FUF, 5MF, 5EF and 5HMF (solubilized in DMSO) were equilibrated at 37.degree. C. for 1 hr. The samples were then incubated in IL 237 tonometers (Instrumentation Laboratories, Inc. Lexington, Mass.) for approximately 10 min at 37.degree. C., and allowed to equilibrate at oxygen tensions 7, 20, and 60 mmHg. The samples were aspirated into an IL 1420 Automated Blood Gas Analyzer and an IL 482 or IL 682 Co-oximeter (Instrumentation Laboratories) to determine the pH, pCO.sub.2, pO.sub.2 and the Hb oxygen saturation values (sO.sub.2). The pO.sub.2 and sO.sub.2 values at each oxygen saturation level were then subjected to a non-linear regression analysis using the program Scientist (Micromath, Salt Lake City, Utah) to calculate the P.sub.50 and Hill coefficient values (n.sub.50). P.sub.50 is the oxygen pressure in mmHg at which Hb is 50% saturated with oxygen. A dose-response study with 5HMF was performed at final compound concentrations of 1 and 2 mM. Oxygen Equilibrium Studies with Homozygous Sickle Red Blood Cells. SS cells were suspended in PBS to a final hematocrit of 10%. Vanillin, FUF, 5HMF (solubilized in DMSO) was added to this suspension at final concentrations of 5 mM and incubated at 37.degree. C. for 1 hr. A 40 uL aliquot of this suspension was added to 4 ml Hemox buffer and subjected to oxygen equilibrium analysis (37.degree. C.) using a Hemox-Analyzer (TCS Scientific Corp., Southampton, Pa.) (Asakura, 1979). Transport through Homozygous Sickle Red Blood Cell Membrane and Reaction with HbS. The compound-treated SS cells obtained in the preceding experiment (oxygen equilibrium studies with SS cells) were hemolyzed by adding 5 volumes of 5 mM potassium phosphate buffer, pH 7.4 containing 0.5 mM EDTA. After centrifugation, the hemolysate was subjected to both oxygen equilibrium analysis using Hemox-Analyzer and cation-exchange HPLC analysis using a Hitachi HPLC apparatus (Model D-7000 Series) and a Swift.TM. WCX column (Swift.TM. WCX-PEEK: 50 mm.times.4.6 mm, Isco, Inc., Lincoln, Nebr.). The column was developed using a linear gradient of phase B from 25% to 90% at 410 nm (Mobile Phase A: 40 mM Bis-Tris, 5 mM EDTA, pH 6.5; Phase B: 40 mM Bis-Tris, 5 mM EDTA, 0.2 M sodium chloride, pH 6.5). The HbS adduct formation (modification of HbS) values are expressed in percentages, using the following formula:

.times..times..times..times..times..times..times..times..times..times..tim- es..times..times..times..times..times..times..times..times..times..times..- times..times..times..times..times..times..times..times. ##EQU00001## Antisickling Studies with Homozygous Sickle Red Blood Cells. The effects of vanillin, FUF and 5HMF on the inhibition of SS cells sickling were evaluated as previously described (Asakura & Mayberry, 1984). Briefly, SS cells suspended in buffered saline solution, pH 7.4 (hematocrit of 10%) were incubated at 37.degree. C. with 4% oxygen in the presence of 5 mM compound. Aliquots (10 ul) of the suspensions were obtained after 5 hrs and fixed with 2% glutaraldehyde solution without exposure to air. Morphological analysis and percentage of SS cells that were not sickled were conducted using a computer-assisted image analysis system as described elsewhere (Hijiya, 1991). Dose-response studies of FUF and 5HMF at compound concentrations of 1 and 2 mM were also performed. The Effect of Compounds on Homozygous Sickle Red Blood Cell Size. To study the effect of the compounds on the degree of hydration/dehydration of SS cells, the compound-treated SS cells obtained in the preceding experiment (antisickling studies with SS cells) were evaluated with a Hemavet Cell Analyzer to determine the mean corpuscular volume (MCV). Crystallization Experiments. Crystallization experiments to obtain T and R state crystals were conducted with FUF and 5HMF. The experiments involved 4 25 molar excess of the compounds to Hb (tetramer). For the T state crystallization experiment, the compounds solubilized in DMSO were incubated with deoxy Hb (60 mg/mL protein) for at least 1 hour to form the Schiff base adduct. Sodium cyanoborohydride, in 4 25 molar excess of Hb, was added to reduce the reversible Schiff base-adduct to the corresponding alkylamine covalent bond. Subsequent crystallization of the compound-deoxy Hb complex solutions in 10 mL test tubes using 3.2 3.6 M sulfate/phosphate precipitant (pH 6.5) was performed in a glove box under nitrogen atmosphere as previously described (Safo, et al., 2003; Perutz, 1968) for obtaining high-salt T state crystals. Reduction of the Schiff base-adduct is necessary to observe the bound compound crystallographically (Boyiri, et al., 1995). Unexpectedly, the experiment resulted in two different crystals--a rectangular crystal (space group P2.sub.1), which is isomorphous to T state native crystal, and a trigonal crystal (space group P3.sub.221), which was later determined to have a relaxed state conformation in the form of R2 quaternary structure. The T state crystals grew between 2 10 days, while the R2 state crystals grew between 7 30 days.

Another experiment was designed to crystallize the compound-Hb complexes in R state form using carbonmonoxy Hb (COHb), following a previously described procedure (Safo, et al., 2003; Perutz, 1968). Oxygenated Hb solution was evacuated for about 10 minutes, and the resulting deoxy Hb solution was fully saturated with CO to generate COHb. The compounds solubilized in DMSO were then reacted with the COHb, followed by addition of sodium cyanoborohydride to reduce the Schiff base-adduct. Crystallization was carried out with a solution of 30 50 mg/mL protein, 3.2 3.4 M Na.sup.+/K.sup.+ phosphate, pH 6.4, and two drops of toluene in 10 ml test tubes. The experimental procedures were done under aerobic condition, and resulted in co-crystals (4 30 days) isomorphous to R state native crystals (space group P4.sub.12.sub.12.sub.1).

Data Collection, Processing and Structure Refinement. X-ray diffraction data sets for the R, R2 and T state co-crystals were collected at 100.degree. K using a Molecular Structure Corporation (MSC) X-Stream Cryogenic Cooler System (MSC, The Woodlands, Tex.), a R-Axis II image plate detector equipped with OSMIC mirrors, and a Rigaku RU-200 generator operating at 50 kV and 100 mA. Prior to use in diffraction, the crystals were first washed in a cryoprotectant solution containing 50 .mu.L mother liquor and 10 12 .mu.L glycerol. The data sets were processed with MSC BIOTEX software program. All structure refinements and omit maps were performed with the CNS program (Brunger, et. al., 1998). Model building and correction were carried out using the graphic program TOM (Cambillau & Horjales, 1987). Structure Determinations and Refinements of the R2 State Complex Structures. The structure of the FUF-Hb complex in the R2 state crystal was the first to be determined by molecular replacement method (Navaza, 1994) using the .alpha.1.beta.1 .alpha.2.beta.2 R2 state native Hb structure (PDB code 1BBB) as a search model. The translation function using the space group P3.sub.221 gave a solution of a tetramer in the asymmetric unit with a final correlation coefficient of 69.2 and Rfactor of 35.5% for data between 8.0 4.0 .ANG.. Prior to using the R2 structure as a search model, we assumed the crystal to be a T state Hb in another crystal form. However, the use of a T state structure (PDB code 2HHB) as a search model failed to give a clear solution. A similar search with a R state structure (PDB code 1AJ9) also failed to give a clear solution. The molecular replacement model was subjected to a rigid body refinement, followed by conjugate gradient minimization and simulated annealing. Strong and clear densities were identified for two FUF molecules bound at the N-terminal .alpha.Val1 residues in a symmetry-related fashion. The N-terminal .alpha.Val1 binding site is located in the central water cavity of Hb close to the mouth of the .alpha.-cleft. The electron density from the bound compound overlapped that of the .alpha.Val1 nitrogen, suggesting a covalent interaction between FUF and the nitrogen. The electron density map also showed ligation of the four heme Fe atoms, and water ligands were fitted to the density. Alternate fitting of O.sub.2 ligands produced distorted geometry of the Fe--O--O bonds and angle. Several alternate rounds of conjugate gradient minimization, simulated annealing, individual B factor refinements, and the addition of 7 sulfate anions and 297 water molecules, with manual model corrections, brought the final Rfactor to 21.7% and Rfree to 27.4% at 2.25 .ANG. resolution. The crystallographic data for the structure is summarized in Table 1.

The starting model for the refinement of the 5HMF-Hb complex structure was the FUF-Hb structure--after deletion of FUF, water molecules and sulfate anions. A round of rigid body, conjugate gradient minimization and simulated annealing refinements also showed two 5HMF bound at the two symmetry-related N-terminal .alpha.Val1 nitrogens. In contrast to the FUF-Hb structure, O.sub.2 molecules were ligated to the Fe atoms. Several alternate rounds of conjugate gradient minimization, simulated annealing, individual B factor refinements, and the addition of 7 sulfate anions and 538 water molecules, with intermittent manual model corrections, brought the final Rfactor to 18.3% and Rfree to 22.3% at 1.85 .ANG. resolution. The crystallographic data for the 5HMF-Hb structure is summarized in Table 1. The atomic coordinates and structure factors have been deposited in the RCSB Protein Data Bank with accession codes 1QXD and 1QXE for the FUF- and 5HMF-Hb structures, respectively.

Structure Determinations and Refinements of the T state Complex Structures. The starting model for the refinement of the T state 5HMF-Hb structure was the isomorphous .alpha.1.beta.1 .alpha.2.beta.2 T state native structure (PDB code 2HHB). After rigid body refinement, and subsequent gradient minimization and simulated annealing, the electron density maps for the structure, unlike those of the R2 state complex structures, revealed only weak and undefined densities at the N-terminal .alpha.Val1 binding sites. Repeated cycles of refinements, addition of water molecules, and model building did not show improved density at the binding site to successfully model 5HMF. There were no other apparent binding sites. The final Rfactor and Rfree for the 5HMF-Hb structures are 16.3 and 20.7% at 1.86 .ANG. resolution. Other statistics for the crystal are reported in Table 1.

TABLE-US-00001 TABLE 1 Crystal Information, Data Collection and Refinement Parameters for the Hb Complex Structures FUF (R2 state) 5HMF (R2 state) 5HMF (T state) FUF (R state) Data collection Space Group P3.sub.221 P3.sub.221 P2.sub.1 P4.sub.12.sub.12 Cell Dimensions (.ANG.) 91.40 91.86 62.61 53.46 91.40 91.86 82.47 53.46 142.00 143.53 53.46 192.88 99.52 Mol/asymmetric unit 1 tetramer 1 tetramer 1 tetramer 1 dimer Resolution (.ANG.) 69.1 2.25 69.6 1.85 82.5 1.86 84.0 2.0 No. of measured refl. 124853 286747 108495 105005 Unique reflections 32647 56802 41917 18357 I/sigma I 7.0 12.5 15.8 13.2 Compl. (%) 97.0 93.1 90.7 92.0 Rmerge (%).sup.a 7.5 6.9 6.5 6.9 Refinement Resolution (.ANG.) 69.1 2.25 69.6 1.85 52.7 1.86 51.5 2.0 Sigma cutoff (F) 0.0 0.0 0.0 0.0 No. of refl. 32645 56780 41895 18316 Rfactor (%) 21.7 18.4 16.3 21.3 Rfree (%).sup.b 27.4 22.3 20.7 26.3 Rmsd standard Geom. Bond-lengths (.ANG.) 0.011 0.013 0.015 0.012 Bond-angles (.degree.) 1.87 1.89 1.72 1.90 Dihedral angles Most favored regions 91.4 92.8 93.6 92.8 Additional regions 8.6 7.2 6.4 7.2 .sup.aRmerge = .SIGMA.(<I> - I)/.SIGMA.I. .sup.b5% of the reflection which were used for the calculation of Rfree were excluded from the refinement.

The 5HMF-Hb structure, without water and ligands was used as a starting model for the refinement of the FUF-Hb structure. Similar to the 5HMF-Hb structure, refinements of the FUF-Hb structure did not result in any interpretable density at the binding pocket. The structure was not refined to completion, and detailed statistics for the crystal is not reported in the Table 1.

Structure Determinations and Refinements of the R state Complex Structures. The isomorphous .alpha.1.beta.1 dimer R state structure (1LJW) after deletion of water molecules and phosphate anions was used as the starting model to refine the FUF-Hb structure. Similar to the T state complexes, repeated refinements of the FUF-Hb structure with model building showed only weak and uninterpretable density at the N-terminal .alpha.Val1 binding pocket. The final Rfactor and Rfree are 21.3 and 26.3 at 2.0 .ANG. resolution, and detailed statistics for the crystal are reported in Table 1.

The FUF-Hb structure, without water and ligands was used as a starting model for the refinement of the 5HMF-Hb structure. Similar to that of the FUF crystal, refinements also showed uninterpretable density at the binding pocket, and the refinement was aborted. No detailed statistics for the crystal are reported in the Table 1.

Results

Vanillin has been clinically tested for SCD therapy, and was studied with the examined heterocyclic aldehydes, also referred to as furanic compounds. We have also previously published detailed functional and antisickling properties of vanillin (Abraham, et al., 1991). Additionally, relatively high concentrations of compounds (5 10 mM) were used to ensure complete reaction with Hb in the current studies. This is necessary, as the concentration of Hb within RBCs is approximately 5 mmol/L, and at 25% hematocrit with a blood volume of 4 L, 10 mmol/L compound is needed to produce a 2:1 compound:Hb adduct. There are two identical binding sites in Hb since it possesses a two-fold axis of symmetry.

Oxygen Equilibrium Studies with Normal Whole Blood. Table 2 summarizes the effects of four furanic compounds and vanillin (at 5 mM concentrations) on AA cell. Allosteric effectors that increase Hb oxygen affinity decrease the P.sub.50 (left-shift the OEC)--relative to the control. This results in a negative .DELTA.P.sub.50 value. The most potent compound is 5HMF (.DELTA.P.sub.50=-17.52 mmHg), followed by 5MF (.DELTA.P.sub.50=-16.16 mmHg), 5EF (.DELTA.P.sub.50=-15.71 mmHg), and FUF (-11.35 mmHg). The poorest left shifting compound is vanillin (-6.78 mmHg). Table 2 also shows that 5HMF left-shifts the OEC in a dose-dependent manner. The Hill coefficients of the modified Hbs, with the exception of that of FUF are smaller compared to that of AA cells alone.

TABLE-US-00002 TABLE 2 Oxygen Equilibrium Studies with Normal Whole Blood.sup.a Compound.sup.b P.sub.50 (mmHg).sup.c .DELTA.P.sub.50 (mmHg).sup.d n.sub.50.sup.e Control 25.84 .+-. 0.01 -- 2.27 .+-. 0.20 5 mM Vanillin.sup.e 19.06 .+-. 1.98 -6.78 1.98 .+-. 0.23 5 mM FUF 14.49 .+-. 0.06 -11.35 2.30 .+-. 0.01 5 mM 5MF 9.68 .+-. 0.24 -16.16 1.70 .+-. 0.06 5 mM 5EF 10.13 .+-. 0.75 -15.71 1.77 .+-. 0.14 1 mM 5HMF 19.19 .+-. 1.51 -6.65 2.08 .+-. 0.14 2 mM 5HMF 15.04 .+-. 0.51 -10.80 1.98 .+-. 0.08 5 mM 5HMF 8.32 .+-. 0.40 -17.52 1.88 .+-. 0.06 .sup.aThe results are the means .+-. S. E. for 2 measurements. .sup.bThe ratio of compound to Hb at 1 mM, 2 mM and 5 mM compound concentrations are 0.8, 1.6 and 4, respectively. .sup.cP.sub.50 is the oxygen pressure at which AA cells (40% hematocrit) in the absence or presence of compound is 50% saturated with oxygen. .sup.d.DELTA.P.sub.50 is P.sub.50 of compound treated AA cells-P.sub.50 of control. .sup.en.sub.50 is the Hill coefficient at 50% saturation with oxygen.

Oxygen Equilibrium Studies with Homozygous Sickle Red Blood Cells. Table 3 (columns 2 and 3) summarizes changes in P.sub.50 and .DELTA.P.sub.50 for SS cells treated with vanillin, FUF, and 5HMF at 5 mM concentration. All compounds shift the OEC to the left, and as observed in the oxygen equilibrium studies with AA cells, 5HMF is allosterically the most potent compound (.DELTA.P.sub.50=-25.2 mmHg) followed by FUF (.DELTA.P.sub.50=-15.8 mmHg), and lastly vanillin (.DELTA.P.sub.50=-13.5 mmHg). FIG. 3 shows the OEC curve for all three compounds at 5 mM concentration, and shows that 5HMF is significantly left-shifted compared with the OEC of both FUF and vanillin. As previously observed for vanillin with both AA and SS cells (Abraham, et al., 1991), higher concentrations of tested compounds resulted in a more hyperbolic OEC from the normal sigmoidal shaped curve.

TABLE-US-00003 TABLE 3 Oxygen Equilibrium and Adduct Formation Studies with Homozygous Sickle Red Blood Cells.sup.a He- SS cells SS cells molysate Com- P.sub.50 .DELTA.P.sub.50 Hemolysate .DELTA.P.sub.50 HbS adduct pound (mmHg).sup.b (mmHg).sup.c P.sub.50 (mmHg) (mmHg).sup.d (%).sup.e Control 31.2 .+-. 1.0 -- 11.2 .+-. 0.2 -- -- Vanillin 17.7 .+-. 2.2 -13.5 8.1 .+-. 1.1 -3.1 15 .+-. 3.6 FUF 15.4 .+-. 1.7 -15.8 5.3 .+-. 0.6 -5.9 24 .+-. 5.7 5HMF 6.0 .+-. 1.2 -25.2 1.8 .+-. 0.3 -9.4 70 .+-. 10.0 .sup.aThe results are the means .+-. S. E. for 2 measurements. .sup.bP.sub.50 is the oxygen pressure at which SS cells (10% hematocrit) or hemolysate (in the absence or presence of compound solubilized in DMSO) is 50% saturated with oxygen. .sup.c.DELTA.P.sub.50 is P.sub.50 of compound treated SS cells or hemolysate-P.sub.50 of control. .sup.dP.sub.50 values obtained from hemolysate after incubation of compounds with SS cells. .sup.eHbS adduct values obtained from HPLC elution patterns of hemolysate after incubation of compounds with SS cells.

Transport through Homozygous Sickle Red Blood Cell Membrane and Reaction with HbS. These experiments were undertaken to determine if the left-shift observed for compound-treated SS cells is the result of a direct interaction of the compound with HbS, and also to determine if P.sub.50 changes observed in SS cells treated with test compounds is attributed to the formation of different levels of compound-HbS adduct. The results are summarized in Table 3 (columns 4 6). Each of the tested compounds produces a new HbS modified peak that eluted before that of the parent HbS peak, indicating the formation of covalently modified HbS adducts. 5HMF modified HbS by the greatest degree (70%), followed by FUF (24%), and lastly vanillin (15%). FIG. 4 shows the Cation-exchange HPLC studies of the hemolysate from the 5HMF-reacted SS cells at 5 mM concentration. The compounds also shifted the OEC of the hemolysate to the left. These shifts follow the same trend observed during the normal whole blood studies. 5HMF causes the largest Hb left shift (.DELTA.P.sub.50 of -9.4 mmHg), followed by FUF (.DELTA.P.sub.50=-5.9 mmHg) and lastly vanillin (.DELTA.P.sub.50=-3.1 mmHg). Antisickling Studies of Compounds with Homozygous Sickle Red Blood Cells. Upon exposure of SS cell suspensions to only 4% oxygen, in the absence of test compounds, all cells underwent sickling. In the presence of vanillin, FUF, and 5HMF (at 5 mM concentrations) the percentage of SS cells decreases by 20, 30, and 90%, respectively (Table 4, columns 2 and 3). 5HMF inhibited sickling the most, followed by FUF, and vanillin. These results follow the same trend observed in the left shift of the OEC, as well as the compound-HbS adduct formation. Table 4 also shows the results of the dose-dependent antisickling effect of FUF and 5HMF. Different concentrations of these compounds decreased the formation of SS cells in a dose-dependent manner. However, unlike 5HMF, FUF did not inhibit cell sickling at lower compound concentrations (1 and 2 mM).

TABLE-US-00004 TABLE 4 Antisickling Studies with Homozygous Sickle Red Blood Cells.sup.a,b Sickling of SS Inhibition of sickling MCV.sup.c Compound cells (%) of SS cells (%) (fl) Control 100 0 61.5 5 mM Vanillin 80 20 .+-. 6.5 61.0 1 mM FUF 100 0 60.3 2 mM FUF 100 0 60.0 5 mM FUF 70 30 .+-. 7.0 62.8 1 mM 5HMF 87 13 62.5 2 mM 5HMF 58 42 .+-. 1.0 61.0 5 mM 5HMF 10 90 .+-. 5.0 61.4 .sup.aThe results are the means .+-. S. E. for 2 measurements. .sup.bAntisickling studies with SS cells (10% hematocrit) under 4% oxygen. .sup.cMCV is the mean corpuscular volume.

The Effect of Compounds on Homozygous Red Sickle Blood Cell Size--As shown in Table 4 (column 4) incubation of SS cells with 1, 2, and 5 mM of FUF or 5HMF did not result in changes in cell volumes. Crystallization Studies. Deoxygenated Hb complexed with either FUF or 5HMF crystallized in both T and R2 state conformations. The T state co-crystallized with FUF and 5HMF show only weak binding of these compounds, while the R2 state co-crystallized with these compounds show very strong binding. With compound to Hb ratios of 4:1, we observed both T and R2 state crystals in the same crystallization tubes. Interestingly, for 5HMF, more R2 state than T state crystals were always observed; for FUF the opposite was generally true. However, if a large excess of compound is used (.gtoreq.10 molar excess), nearly all of the co-crystallization experiments with FUF and 5HMF result in only R2 state crystals. These results suggest that 5HMF is allosterically more potent than FUF, which is consistent with the biological results.

R2 state native crystals have previously been crystallized under low salt conditions (Silva, et al., 1992), but not in high salt conditions. The ensuing structures of the R2 state co-crystals, as already pointed out, have water or O.sub.2 molecules (FUF- and 5HMF-Hb complexes, respectively) coordinated to the Fe atoms. In reality, it is hypothesized that the ligands are actually a mixture of O.sub.2, CO, and water, as analysis of the R2 state complexes showed the presence of COHb, metHb, and oxyHb (60 85%), versus approximately 16% for the T state co-crystals. The presence of the ligands in the R2 state co-crystals could be due to the fact that the anaerobic chamber used in these experiments may not have been completely devoid of oxygen during the crystallization setup. Interestingly, the T state co-crystals that occurred in the same solutions as the R2 state co-crystals did not show any residual density for ligand binding to Fe. These observations underscore the high-affinity nature of relaxed state Hb compared to tense state Hb. These results also clearly show that the R2 quaternary structure is physiologically important as previously pointed out (Srinivasan & Rose, 1994).

Unlike the T and R2 crystallization results, and quite significantly, repeated R state crystallization experiments did not result in any co-crystals, especially at high compound concentrations (.gtoreq.5 molar excess to the Hb tetramer). However, if low compound concentrations (4 molar excess to the Hb tetramer) were used, very few R state co-crystals for both FUF and 5HMF were observed, with most of the complex remaining in solution. We should point out that, we were able to easily obtain R state native crystals (control experiment without adding any compound) under these same crystallization conditions.

Descriptions of the R2 State Complex Structures. Both FUF-Hb and 5HMF-Hb complex structures contain one .alpha.1.beta.1 .alpha.2.beta.2 tetramer in the asymmetric unit. The R2 state complexes and R2 state native have essentially the same Hb quaternary structures (rmsds of .about.0.4 .ANG.). However, comparison of the R2 state complex structures with R (PDB code 1AJ9) and T (PDB code 2HHB) native Hb structures show very significant quaternary structural differences, with rmsd of .about.1.8 .ANG. and .about.3.3 .ANG., respectively. As previously analyzed and reported by Silva et al., (1992) for R, T and R2 native Hb structures, the allosteric transitions between the R2 complex structures and those of the R and T structures show extensive reorganization of the .alpha.1.beta.2.alpha.1.alpha.2 and .beta.1.beta.2 interfaces in the three Hb states. The structures were superimposed using the invariant .alpha.1.beta.1 dimer (C.alpha. residues) on the BGH frame as defined by Baldwin & Chothia (1979).

Following are detailed descriptions of the interactions between R2 state Hb, and the two compounds, FUF and 5HMF. Both furanic compounds are well ordered with occupancies of approximately 100%. The compounds bind in a symmetry-related fashion at the .alpha.-cleft to the two N-terminal .alpha.Val1, with the aldehyde functional group forming a covalent bond with the free nitrogen of the valine. Specific interactions between the compounds and the protein will be discussed for only the .alpha.1Val1 binding site, as the other symmetry-related molecule engages in similar, but opposite interactions at the .alpha.2Val1 binding site. A covalent interaction between the aldehyde and .alpha.1Val1 nitrogen directs the furan ring of the compounds toward the central water cavity. In the FUF-Hb structure, the bound compound appears to have two alternate conformations that differ by almost 180.degree.. The aromatic oxygen engages in a very weak intersubunit hydrogen bond with .alpha.2Ser138 OG (3.6 .ANG.), which serves to tie the two .alpha.-subunits together. Interestingly, if the compound is rotated to its alternate conformation, the oxygen faces the water cavity and engages in a weak intrasubunit hydrogen bond with .alpha.1Ser131 OG (3.5 .ANG.). There are very few hydrophobic interactions (<3.8 .ANG.) between the furan ring and the residues Lys127 and Ala130.

Unlike FUF, there is no evidence of compound rotation of the bound 5HMF molecule, and it assumes a conformation with the ring oxygen facing the water cavity. The observed interaction between FUF and .alpha.Ser138 OG is therefore absent in the 5HMF-Hb complex structure. Similar to FUF, the ring oxygen of 5HMF engages in a stronger intrasubunit hydrogen bond with .alpha.1Ser131 OG (3.1 .ANG.), compared to FUF binding. In addition, the 5-hydroxymethyl substituent of 5HMF also makes a strong intrasubunit hydrogen bonding interaction with .alpha.1Thr134 OG1 (2.6 .ANG.); this interaction is absent in the FUF-Hb complex structure. While FUF ties the two .alpha.-subunits together by making a weak intersubunit hydrogen bond with .alpha.2Ser138 OG, the two 5HMF molecules are joined together by a strong network of six water-mediated hydrogen bonds, through the hydroxyl and the ring oxygen moieties that tie the two .alpha.-subunits together. Some of these water molecules are conserved in the FUF binding pocket, but they are mobile and do not engage in hydrogen bonding contact with the FUF molecules. Like the FUF-Hb structure, there are very few hydrophobic interactions (<3.8 .ANG.) between 5HMF and Hb. The increased number of interactions between 5HMF and protein residues (versus FUF), as well as the strong water-mediated hydrogen bonds that tie the two .alpha.-subunits together in the 5HMF complex structure, may partly explain why 5HMF is allosterically more potent than FUF.

Descriptions of the T State Complex Structures. Both FUF-Hb and 5HMF-Hb complex structures contain one .alpha.1.beta.1 .alpha.2.beta.2 tetramer in the asymmetric unit. Unlike the R2 complex structures, the T state complexes do not show clearly defined compound binding. Some examined co-crystals did show more compound density than others; however repeated model building to improve compound density was not successful enough to allow for reliable compound fitting. The T state complexes and T state native have essentially the same Hb quaternary structures (rmsds of .about.0.4 .ANG.). Superposition of the .alpha.Val1 binding sites also shows very few structural differences. Further, we should point out that in our laboratory we have been able to easily isolate T state structures that have covalently bound compounds, with clearly defined electron density at the N-terminal .alpha.Val1 using the same T state crystallization conditions as described above (Abraham, et al. 1995; Boyiri, et al., 1995). Descriptions of the R State Complex Structures: Both FUF-Hb and 5HMF-Hb complex structures contain one .alpha.1.beta.1 dimer in the asymmetric unit. Surprisingly, the few R state crystals that were obtained during these experiments show only sparse density at the N-terminal .alpha.Val1 binding sites. Even though the R complex and native quaternary structures are indistinguishable (rmsds .about.0.4 .ANG.), the C-terminus (residues Trp140 and Arg141) display significant positional differences. In the complex structures, these residues have rotated away at the .alpha.Lys139, displacing .alpha.Arg141 by almost 180.degree. from its position in the native structure, while .alpha.Tyr140 has oriented away by .about.2 .ANG.. .alpha.Tyr140 OH now engages in hydrogen bonds with .alpha.Val93 O and .alpha.Pro95 N in the complex structures, versus diagnostic R state native hydrogen bonds with .alpha.Val93 N and O. The reorientation of the C-terminus has led to a much bigger binding cleft in the R complex structures to allow binding of the compound, albeit weak. This contrasts with the native structure, where the C-terminal residues are found to be sterically blocking this binding site.

Discussion

These studies have identified furan-based derivatives, which are naturally occurring in a number of foods, as potential new therapeutics to treat SCD. Results from these studies clearly indicate that these compounds possess the ability to: (1) pass through RBC membranes; (2) react with HbS; and (3) allosterically shift the Hb OEC to the high-affinity state, which does not form HbS polymers. Furthermore, we have found that the change in the oxygen affinity of SS cell suspensions caused by these compounds depends on the degree of binding to HbS; 5HMF showed the highest amount of compound-HbS adduct, and as expected, was the most potent OEC left shifter. Also, the results clearly suggest that substitution, as well as substitution type at the 5-position of the central furan ring is important to biological activities. 5HMF, which possesses an alkyl alcohol at the 5-position of the furan ring, is more potent than either 5MF or 5EF (FIG. 2), both of which possess hydrophobic moieties at this position. FUF, without a substitution is the least potent. This is consistent with the crystallographic results, which indicate that the hydroxyl moiety of 5HMF is intimately involved in interactions that stabilize the relaxed state. The Hill coefficients of the modified Hbs are relatively smaller compared to the unmodified Hb, suggesting a decrease in cooperativity. This is expected because of the apparent weakening of interdimer interactions by the binding of the compounds to the T state, leading to increased oxygen affinity, reduced cooperativity, and a shift toward the high-affinity Hb.

Antisickling Activities of the Furanic Compounds