Thank Dr …for your kind introduction. It is my pleasure to be here. Today I am going to talk about. ---- party of my research which was done between 2002 and 2006 at jnj.
Alzheimer’s Disease is a progressive, devastating and incurable illness. (There are several approaches that are targeting on treatment of AD. Why did we choose BACE-1 program?) Genetic and pathophysiology evdidence suggested that AD is mainly caused caused by amyloid plaques and neurofibrillary tangles. Amyloid plaques is accumulated by the deposit of abeta 1-40 and 1-42. Peptide. The abeta 1-40 and 42 is generated by the cleavage of Amyloid Precursor protein by first beta secretase and then secretase. BACE-1 program is attempted to find a inhibitor to prevent the cleavage of APP from beta secretase enzyme and to prevent AD at the beginning of the process.
OM 99-2 was the first potent BACE inhibitor reported in 2000. It was a large compound with many polar functional groups and peptide bonds. it had poor bioavailability and metabolic stability. In order to avoid those two problems, we were targeting on small molecule as BACE inhibitors. CNS team at J&J started BACE-1 program in 2000. We screened 400000 compounds. This was one of the hits obtained during screening. We obtained x-ray structure of this compound in the enzyme. From the structure, we found that the two parts of the molecule are tend to fold together. I was thinking to fix this conformation by connecting the two parts together.
Compound 3 with ether linker between the two phenol groups was obtained during SAR studies. It had better activity and was much easier to make than 2. I was thinking of cyclizing the two parts of molecule. 4 was one of the targets. How to make the macrocycle? There were three small rings connected each other on the large ring, which may create substantial strain for the cyclization reaction of macrocycle. In order to create the flexibility in the cyclization of the large ring, We decided to cyclize the large ring first. Because the nitro group can activate the leaving of fluoro atom, the phenol oxygen could displace the fluoro to form macrocycle. I also found literature precedent for this type of ring closing reaction. Since we had literature precedent for the key reaction step and reasonable reaction sequence, I suggested this idea to my supervisor Mike Parker and he discussed this idea with our team leader Allen Reitz. Allen suggested to send several structures to computer modeling group to duck. Computer modeling suggested that 16 member ring was the smallest ring we can make. 16, 17, 18 member rings are all fit well. We decided to make 16 member ring first. 4 was our target.
We started with acid 7. Acid 7 was coupled with cyclohexylamine to give the amide which was reduced to secondary amine by LAH.The secondary amine was coupled with this acid to give amide 9. Deprotection of the Boc group with TFA, followed by reductive amination with aldehyde 11 gave compound 12. Demethylation of 12 with boron tribromide provided the phenol. Compound 6 was cyclized by treating potassium carbonate in DMF to give 5 in 16% yield. this was the highest yield we got. Hydrogenation to reduce the nitro group, followed by treating with cyanogen bromide provided the final target. We obtained the first macrocycle. It showed Ki 51 nM, almost 20 fold better activity over our hit. Because it showed big improvement of potency, I was asked to scale up this compound for other biotests. However, there were several steps with low yields, especially the cyclization step, there was no potential to improve on cyclization reaction, the scale up was difficulty.(This bond was not stable during the cyclization reaction.) We needed to find a better route.
First, I was looking for literature precedent for large ring closing. Metathesis and amide formation are the most popular approachs that were reported in literature. Both approaches were not successful. There was another potential point for ring closing which was here by reductive amination. Even though reductive amination is widely used in organic synthesis, there was no literature precedent on using it for macrocyclization. Our system has a reactive aldehyde and primary amine, which is good for reductive amination. I believed that the intra molecular reductive reaction was quite possible. I decided to try it.
phenol 14 was protected with benzyl group by treating with benzyl bromide and K2CO3, followed by hydrolysis with sodium hydroxide solution. Treating the acid with oxalyl chloride followed by cyclohexylamine gave amide 17, which was reduced to amine 18 by lithium aluminium hydride. Coupling amine 18 with acid provided amide 19, which was converted to phenol 20 by hydrogenation. The phenol was treated with fluoro 11 provided 21, which was treated with TFA to give amine 22. Reductive amination of 22 with sodium triacetoxyboro hydride was not working. The aldehyde was reduced to alcohol. We tried other reductive amination conditions and other reducing agents. There was no cyclization reaction.
Two other chemists tried to use the most popular literature approaches for the macrocyclization reaction. One was metathesis, the other was amide formation. Both were not working. I tried to promote the reaction by dehydration ---the formation of imine, which could be reduced to secondary amine. Compound 22 was heated under reflux in benzene with the use of Dean –Stark trap, then NMe4BH(OAc)3 was added. We obtained 5 in 15% yield. It was not good enough for scale up. But, it was a good sign. We tried the dehydration with 4 A molecular sieves. 22 was treated with 4 A Molecular Sieves followed by sodium triacetoxyboro hydride to provide the macrocycle in 78% yield. 5 was converted to final target by hydrogenation followed by treating with cyanogen bromide. We worked out a good and reliable synthetic route for the synthesis of macrocycles.The potency of this macrocycle had a big jump, but it was still not potent enough. Next question we need to ask was: How to improve the potency of the macrocycle?
Our team members continued on SAR studies on the acyclic series and found the potency of our acyclic series was improved dramatically by introducing alpha branch here. We also decided to introduce alpha branch on the macrocycle.
Nitrile 23 was reduced to amine 24 by borane. Reductive amination between amine 24 and cyclohexanone gave 25, which was coupled with chiral intermediate 26 to provide 27. N-Boc amino acid 26 was prepared based on a literature procedure. Hydrogenation of 27 provided phenol 28, which was converted to 29 by treating with fluoro 11. Deprotection of 29, followed by intramolecular reductive amination gave macrocycle 31. We obtained 93% yield on the cyclization reaction. Hydrogenation, followed by treatment of cyanogen bromide provided final product 32. It had ki 5 nM,one of the most potent compounds we had made on BACE project. By using similar reaction sequence, we synthesized other macrocycles for SAR studies.
Those were the major macrocycles we synthesized. We found that 16 member ring was favored in terms of potency. Polar group on this position gave less hERG binding. We also found that the change of functional group and ring size did not have dramatic influence on potency which has the potential to improve the pharmaceutical activity by introduce or change some functional groups. We did in vivo test on this compound.
The results showed that the Macro cycle had potent plasma activities both on IV dosing and oral dosing. This was the data for IV dosing, 2 mpk, 2 hours. It reduced mouse plasma a beta 42% and Rat plasma a beta 74%. This was oral dosing data. The plasma a beta was decreased with the increase of dosing of the compound. It showed poor or no brain activity. The compound had poor brain penetration problem.
By comparing the acyclic with macrocycle compounds, macrocycle has better potency and cellular activity on BAC-1 and much less hERG binding. It showed the reduction of plasma a beta in mouse. We did not found any in vivo activity on the acyclic series. We designed and synthesized potent macrocycles as BACE-1 enzyme inhibitor. We developed reductive amination as novel and efficient approach for the cyclization of macrocycles. Two chemists of our team continued the SAR studies on Macrocycles. I was asked to work on acyclic series.
Macrocycles As Bace 1 Inhibitors
Design and Synthesis of Macrocycles as -Secretase (BACE-1) Inhibitors for the Treatment of Alzheimer’s Disease
Alzheimer’s Disease Pathophysiology <ul><li>Alzheimer’s Disease is a progressive, devastating and incurable illness </li></ul><ul><li>Some 4.5 million Americans are affected at a cost of $100 billion a year </li></ul>Nitasha Manchanda, Alzheimer’s Disease, Decision Resources , June 2007 / Secretase
(400,000 compound HTS campaign) Introduction OM 99-2 Heptapeptide K i = 1.6 nM … to small-molecule BACE inhibitors <ul><li>Poor bioavailability </li></ul><ul><li>Metabolically unstable </li></ul>From peptides … Hong, L. et al. Science , 2000, 290, 5489
Macrocylic BACE-1 Inhibitors – Design Target molecule: Synthetic strategy: Bell, Ian M.; et al Journal of Medicinal Chemistry . 2002, 45, 2388-2409
20 Fold better potency Macrocylic BACE-1 Inhibitors – Initial Synthesis Poor yield-----Difficult for scaling up -----SAR studies Baxter, Ellen E.; Huang, Yifang et al. WO 2007/092839
Two Reactive groups Exploring Alternate Macrocyclization Strategy Reductive amination No literature precedent
Macrocyclization via Reductive Amination NaBH 3 CN/CH 3 OH or NaBH 4 /CH 3 OH not working.
Macrocyclization via Reductive Amination-Continued Imine formation by dehydration
Baxter, Ellen E.; Huang, Yifang et al. WO 2007/092839 K i = 5 nM Synthesis of -Branched Macrocycle
Major Macrocycles Synthesized-SAR Position of Amide : important; Chang of functional groups and ring size: No significant impact ----Great potential for imposing desired pharmaceutical properties.
Macrocycle: In Vivo Testing in Mouse In vivo efficacy (2 mpk, iv, 2 hr) Mouse: plasma a - 42%, Rat : plasma a -74%, Plasma A lowing in mice -22% at 25 mpk, PO/dose, 3 h
Acyclic via Macrocyclic compounds K i = 5 nM A (CHO) cellular IC 50 = 17 nM hERG K i = 1900 nM Plasma A lowing in mice 22% at 25 mpk, PO/dose 74% at 2 mpk, iv/dose in rat K i = 8 nM A (CHO) cellular IC 50 = 44 nM hERG K i = 84 nM No in vivo activity <ul><li>Designed and synthesized macrocycles (k i = 5 nM) as BACE-1 enzyme inhibitor </li></ul><ul><li>Starting from HTS hit (k i = 900 nM). </li></ul><ul><li>Developed reductive amination as novel and efficient synthetic approach </li></ul><ul><li>for cyclizing Macrocycles. </li></ul>