Synthesis and PI3 Kinase Inhibition Activity of Some Novel Trisubstituted Morpholinopyrimidines

A number of new substituted morpholinopyrimidines were prepared utilizing sequential nucleophilic aromatic substitution and cross-coupling reactions. One of the disubstituted pyrimidines was converted into two trisubstituted compounds which were screened as PI3K inhibitors relative to the well-characterized PI3K inhibitor ZSTK474, and were found to be 1.5–3-times more potent. A leucine linker was attached to the most active inhibitor since it would remain on any peptide-containing prodrug after cleavage by prostate-specific antigen, and it did not prevent inhibition of AKT phosphorylation and hence the inhibition of PI3K by the modified inhibitor.

In 2012 [33] we reported that prodrugs containing PI3K inhibitors could be activated via peptide cleavage by prostate-specific antigen (PSA).Cheng and coworkers have also reported generation of prostate-selective PI3K inhibitors [34].We have already shown specific inhibition of PI3K in PSA-secreting prostate cancer cells by a Mu-LEHSSKLQL-LY294002 (prodrug-LY294002) [33].However, LY294002 inhibits PI3K at relatively high concentrations (25 µM) and is rapidly metabolized, so while its ease of synthesis made it a good choice for proof of concept, these pharmacological properties make it a poor candidate for in-vivo studies.
We were confident that PSA would cleave the Mu-LEHSSKLQL peptide after glutamine (Q), therefore leaving leucine (L) attached to the PI3K inhibitor [35].Computer modeling of interactions between PI3K inhibitors and the ATP-binding cleft of PI3K were used to identify positions at which the PSA substrate peptide could be attached so that linker plus leucine will not diminish PI3K inhibition.Following our review of PI3K inhibitor literature [36], we were most interested in preparing trisubstituted pyrimidines and triazines as the next test of this prodrug concept, and here we report our work on pyrimidines.

Chemistry
Synthesis and characterization of trisubstituted triazines and pyrimidines remains an active area of PI3K inhibition research [37][38][39][40][41][42][43][44][45][46].Since both core structures provide inhibitors that are quite active, we decided to prepare both trisubstituted triazines and trisubstituted pyrimidines, and here we report our work on the pyrimidines (Scheme 1).Target compounds reported here contain (1) a linker group terminating in a primary alcohol which can be used as a peptide linkage point; (2) morpholine (due to its known importance in PI3K inhibition [44]); and (3) a hydrogen-bonding aromatic or heteroaromatic group.
secreting prostate cancer cells by a Mu-LEHSSKLQL-LY294002 (prodrug-LY294002) [33].However, LY294002 inhibits PI3K at relatively high concentrations (25 µM) and is rapidly metabolized, so while its ease of synthesis made it a good choice for proof of concept, these pharmacological properties make it a poor candidate for in-vivo studies.
We were confident that PSA would cleave the Mu-LEHSSKLQL peptide after glutamine (Q), therefore leaving leucine (L) attached to the PI3K inhibitor [35].Computer modeling of interactions between PI3K inhibitors and the ATP-binding cleft of PI3K were used to identify positions at which the PSA substrate peptide could be attached so that linker plus leucine will not diminish PI3K inhibition.Following our review of PI3K inhibitor literature [36], we were most interested in preparing trisubstituted pyrimidines and triazines as the next test of this prodrug concept, and here we report our work on pyrimidines.

Chemistry
Synthesis and characterization of trisubstituted triazines and pyrimidines remains an active area of PI3K inhibition research [37][38][39][40][41][42][43][44][45][46].Since both core structures provide inhibitors that are quite active, we decided to prepare both trisubstituted triazines and trisubstituted pyrimidines, and here we report our work on the pyrimidines (Scheme 1).Target compounds reported here contain (1) a linker group terminating in a primary alcohol which can be used as a peptide linkage point; (2) morpholine (due to its known importance in PI3K inhibition [44]); and (3) a hydrogen-bonding aromatic or heteroaromatic group.The triazine and pyrimidine cores offered simplicity in synthesis coupled with good PI3K inhibition activity [36,[47][48][49][50][51][52][53][54][55][56][57][58].Modelling as described previously [47,55,57] indicated that compounds containing the pyrimidine or triazine core substituted with a morpholine as well as an aromatic or heteroaromatic ring containing a hydrogen-bond donor exhibited Ki < 1 µM, good water solubility and hydrolytic stability.The third and final position on the pyrimidine/triazine core was left to be the attachment point for a functional group capable of serving as the link to the PSA-cleavable peptide sequence (Mu-LEHSSKLQL).Our initial work with LY 294002 analogs [33] used linkers that terminated in OH or NH2 groups so we initially chose those types of linkers again.In addition to this linker we were also interested in having a functional group in one of the other two core substituents that contained a heteroaromatic ring with a hydrogen-bonding substituent.
Given that we wanted to prepare trisubstituted pyrimidines containing a linker, an aromatic hydrogen-bond donor and a morpholine, we could conceivably initiate synthetic work by putting on any one of these three groups in step one.In practice, we envisioned replacing the last halogen on the pyrimidine via cross-coupling chemistry rather than nucleophilic substitution, so we left this to The triazine and pyrimidine cores offered simplicity in synthesis coupled with good PI3K inhibition activity [36,[47][48][49][50][51][52][53][54][55][56][57][58].Modelling as described previously [47,55,57] indicated that compounds containing the pyrimidine or triazine core substituted with a morpholine as well as an aromatic or heteroaromatic ring containing a hydrogen-bond donor exhibited Ki < 1 µM, good water solubility and hydrolytic stability.The third and final position on the pyrimidine/triazine core was left to be the attachment point for a functional group capable of serving as the link to the PSA-cleavable peptide sequence (Mu-LEHSSKLQL).Our initial work with LY 294002 analogs [33] used linkers that terminated in OH or NH 2 groups so we initially chose those types of linkers again.In addition to this linker we were also interested in having a functional group in one of the other two core substituents that contained a heteroaromatic ring with a hydrogen-bonding substituent.
Given that we wanted to prepare trisubstituted pyrimidines containing a linker, an aromatic hydrogen-bond donor and a morpholine, we could conceivably initiate synthetic work by putting on any one of these three groups in step one.In practice, we envisioned replacing the last halogen on the pyrimidine via cross-coupling chemistry rather than nucleophilic substitution, so we left this to our third step.This decision meant we could initiate syntheses that started by adding linkers first or morpholine (Scheme 2) first.

First Addition Reaction (Scheme 2)
Sequential nucleophilic aromatic substitution reactions on 2,4,6-trichloropyrimidine present some challenges which are not present for the triazine core.Statistically, if one assumes all three chlorine-containing carbons are equally reactive, then the best one can hope for off the first nucleophilic addition is a 2:1 mixture of products.Separation can be effected then or sometimes others have then added a second nucleophile and separated the four products produced [46,59].A second consideration is rate of reaction and isolated yields when one has the option of adding a primary amine or secondary amine as the first or second step in the sequence.In our hands, morpholine (7) added to trichloropyrimidine (5) much faster than the primary amines, and if we added morpholine (7) first then reactions of primary amines on a morpholine-substituted dichloropyrimidine (10 or 11) were very sluggish and hard to drive to completion.Given that, we decided to pursue adding primary amino alcohols to trichloropyrimidine first.We performed reactions using both 6-amino-1-hexanol (6a) and 4-aminomethylphenylmethanol (6b) and isolated much higher yields of products (8a/9a) from 6-amino-1-hexanol (6a) (Scheme 2).We decided to take the minor isomer (8a) on for convenience since subsequent nucleophilic additions (in our case here morpholine) to that compound can yield only one regioisomer (12) (whereas additions to 9a or 10 would be expected to yield two regioisomers).We isolated 12 in excellent yield and avoided the careful chromatographic separations needed for 8-11 at this disubstituted pyrimidine stage.
Molecules 2018, 23, x 3 of 13 our third step.This decision meant we could initiate syntheses that started by adding linkers first or morpholine (Scheme 2) first.

First Addition Reaction (Scheme 2)
Sequential nucleophilic aromatic substitution reactions on 2,4,6-trichloropyrimidine present some challenges which are not present for the triazine core.Statistically, if one assumes all three chlorine-containing carbons are equally reactive, then the best one can hope for off the first nucleophilic addition is a 2:1 mixture of products.Separation can be effected then or sometimes others have then added a second nucleophile and separated the four products produced [46,59].A second consideration is rate of reaction and isolated yields when one has the option of adding a primary amine or secondary amine as the first or second step in the sequence.In our hands, morpholine (7) added to trichloropyrimidine (5) much faster than the primary amines, and if we added morpholine (7) first then reactions of primary amines on a morpholine-substituted dichloropyrimidine (10 or 11) were very sluggish and hard to drive to completion.Given that, we decided to pursue adding primary amino alcohols to trichloropyrimidine first.We performed reactions using both 6-amino-1-hexanol (6a) and 4-aminomethylphenylmethanol (6b) and isolated much higher yields of products (8a/9a) from 6-amino-1-hexanol (6a) (Scheme 2).We decided to take the minor isomer (8a) on for convenience since subsequent nucleophilic additions (in our case here morpholine) to that compound can yield only one regioisomer (12) (whereas additions to 9a or 10 would be expected to yield two regioisomers).We isolated 12 in excellent yield and avoided the careful chromatographic separations needed for 8-11 at this disubstituted pyrimidine stage.Scheme 2. Nucleophilic aromatic substitution reactions of chloropyrimidines.

Third Addition Reaction (Scheme 3)
A variety of cross-coupling conditions were investigated in order to add the last heteroaromatic hydrogen-bonding substituent.For this work, we chose 2-aminopyrimidine-5-boronic acid pinacol ester (13) (Scheme 3).Pd(PPh 3 ) 4 (10 mol %) was used initially as a catalyst in DMSO, THF and DME (3:1 organic solvent: 2 M Na 2 CO 3 ).DME proved superior and we observed no product when the 2 M Na 2 CO 3 was eliminated.Pd(dppf)Cl 2 , Pd(OAc) 2 /PPh 3 , Pd(OAc) 2 /CuCl 2 /S-Phos and Pd(OAc) 2 /CuCl 2 /K 3 PO 4 /S-Phos were all investigated as catalysts and Pd(dppf)Cl 2 provided the highest yield of cross-coupled product (14).Compound 14 proved to be as active a PI3K inhibitor as ZSTK 474 (see screening discussion below), and therefore we undertook preparation of its isomer as well as its preparation where leucine was added to the primary OH in the linker group.

Third Addition Reaction (Scheme 3)
A variety of cross-coupling conditions were investigated in order to add the last heteroaromatic hydrogen-bonding substituent.For this work, we chose 2-aminopyrimidine-5-boronic acid pinacol ester (13) (Scheme 3).Pd(PPh3)4 (10 mol %) was used initially as a catalyst in DMSO, THF and DME (3:1 organic solvent:2 M Na2CO3).DME proved superior and we observed no product when the 2 M Na2CO3 was eliminated.Pd(dppf)Cl2, Pd(OAc)2/PPh3, Pd(OAc)2/CuCl2/S-Phos and Pd(OAc)2/CuCl2/ K3PO4/S-Phos were all investigated as catalysts and Pd(dppf)Cl2 provided the highest yield of crosscoupled product (14).Compound 14 proved to be as active a PI3K inhibitor as ZSTK 474 (see screening discussion below), and therefore we undertook preparation of its isomer as well as its preparation where leucine was added to the primary OH in the linker group.

Attempt to Prepare an Isomer of 14
We attempted to prepare an isomer of the active compound ( 14) by first performing a crosscoupling reaction on morpholine-substituted pyrimidine (11), and this reaction proceeded in good yield to produce 15 (Scheme 4).However, attempts to add primary amines to 15, with heat, Pd catalysts or Pd catalysts and heat proved unsuccessful.

Attempts to Prepare an Analog of 14 with a Terminal Alkyne Rather than Primary OH as Peptide Link Point
In our earlier work we had used a primary alcohol as a peptide attachment point when making PI3K-inhibitor prodrugs [33].This necessitates an ester functional group for the attachment so we wanted to investigate other alternatives.We hoped that if we could prepare an analog of 14 with a terminal alkyne linker group then click chemistry would become an option for attaching peptide sequences.To test this option, we first added propargyl amine to trichloropyrimidine (5) as we had done for aminohexanol and this produced two isomers (16 and 17) as expected (Scheme 5).The symmetrical isomer (17) was taken on for the reasons outlined above and the morpholine-containing disubstituted pyrimidine (18) was isolated in high yield.Unfortunately, a variety of cross-coupling conditions for attachment of 13 to 18 (identical to the battery we had tried for 14) produced none of the desired cross-coupled trisubstituted product.We also tried back-tracking here and attempted to Scheme 3. Cross-coupling reaction of disubstituted pyrimidine (12).

Attempt to Prepare an Isomer of 14
We attempted to prepare an isomer of the active compound ( 14) by first performing a cross-coupling reaction on morpholine-substituted pyrimidine (11), and this reaction proceeded in good yield to produce 15 (Scheme 4).However, attempts to add primary amines to 15, with heat, Pd catalysts or Pd catalysts and heat proved unsuccessful.

Third Addition Reaction (Scheme 3)
A variety of cross-coupling conditions were investigated in order to add the last heteroaromatic hydrogen-bonding substituent.For this work, we chose 2-aminopyrimidine-5-boronic acid pinacol ester (13) (Scheme 3).Pd(PPh3)4 (10 mol %) was used initially as a catalyst in DMSO, THF and DME (3:1 organic solvent:2 M Na2CO3).DME proved superior and we observed no product when the 2 M Na2CO3 was eliminated.Pd(dppf)Cl2, Pd(OAc)2/PPh3, Pd(OAc)2/CuCl2/S-Phos and Pd(OAc)2/CuCl2/ K3PO4/S-Phos were all investigated as catalysts and Pd(dppf)Cl2 provided the highest yield of crosscoupled product (14).Compound 14 proved to be as active a PI3K inhibitor as ZSTK 474 (see screening discussion below), and therefore we undertook preparation of its isomer as well as its preparation where leucine was added to the primary OH in the linker group.

Attempt to Prepare an Isomer of 14
We attempted to prepare an isomer of the active compound ( 14) by first performing a crosscoupling reaction on morpholine-substituted pyrimidine (11), and this reaction proceeded in good yield to produce 15 (Scheme 4).However, attempts to add primary amines to 15, with heat, Pd catalysts or Pd catalysts and heat proved unsuccessful.

Attempts to Prepare an Analog of 14 with a Terminal Alkyne Rather than Primary OH as Peptide Link Point
In our earlier work we had used a primary alcohol as a peptide attachment point when making PI3K-inhibitor prodrugs [33].This necessitates an ester functional group for the attachment so we wanted to investigate other alternatives.We hoped that if we could prepare an analog of 14 with a terminal alkyne linker group then click chemistry would become an option for attaching peptide sequences.To test this option, we first added propargyl amine to trichloropyrimidine (5) as we had done for aminohexanol and this produced two isomers (16 and 17) as expected (Scheme 5).The symmetrical isomer (17) was taken on for the reasons outlined above and the morpholine-containing disubstituted pyrimidine (18) was isolated in high yield.Unfortunately, a variety of cross-coupling conditions for attachment of 13 to 18 (identical to the battery we had tried for 14) produced none of the desired cross-coupled trisubstituted product.We also tried back-tracking here and attempted to  In our earlier work we had used a primary alcohol as a peptide attachment point when making PI3K-inhibitor prodrugs [33].This necessitates an ester functional group for the attachment so we wanted to investigate other alternatives.We hoped that if we could prepare an analog of 14 with a terminal alkyne linker group then click chemistry would become an option for attaching peptide sequences.To test this option, we first added propargyl amine to trichloropyrimidine (5) as we had done for aminohexanol and this produced two isomers (16 and 17) as expected (Scheme 5).The symmetrical isomer (17) was taken on for the reasons outlined above and the morpholine-containing disubstituted pyrimidine (18) was isolated in high yield.Unfortunately, a variety of cross-coupling conditions for attachment of 13 to 18 (identical to the battery we had tried for 14) produced none of the desired cross-coupled trisubstituted product.We also tried back-tracking here and attempted to cross-couple 13 to 17 under a variety of conditions but those also failed, leaving us to conclude that the terminal alkyne was not compatible with these conditions.cross-couple 13 to 17 under a variety of conditions but those also failed, leaving us to conclude that the terminal alkyne was not compatible with these conditions.Scheme 5. Use of propargyl amine as a potential peptide link point.

Addition of Leucine to the Lead Compound (14)
Since compound 14 was quite active as a PI3K inhibitor (see screening discussion below), we undertook its preparation where leucine was added to the primary OH in the linker group.In our earlier work [33], we demonstrated that peptide (Mu-LEHSSKLQ) in prostate-specific prodrugs is cleaved between L and Q, so the leucine residue remains attached to the PI3K inhibitor.To mimic the prodrug cleaved by PSA, Boc-protected leucine was coupled to 14 to produce 19 in good yield and then the protecting group was removed to yield 20 (Scheme 6).Scheme 6. Addition of leucine to lead compound 14.

Biological Activity
The biological activities of new PI3K-inhibitor compounds ( 14) and (20) were tested in C4-2 prostate cancer cells with constitutive activation of PI3K/AKT pathway.C4-2 cells were derived by passaging of prostate cancer metastases-derived LNCaP cells through nude mice and characterized by increased propensity to form metastases in mice.The PI3K/AKT pathway is activated in these cells due to frame-shift mutation in the PTEN tumor-suppressor gene.Activation of PI3K leads to accumulation of phosphatidylinositol (3,4,5)-trisphosphate (PIP3) in the plasma membrane that recruits protein kinases AKT and PDK1 through binding to their PH domains.Interaction of PIP3 with the PH domain changes AKT conformation and opens access for PDK1 to phosphorylate at T308, and also allows phosphorylation at S473 by TORC2 complex.Thus, phosphorylation of AKT at S473 and T308 is routinely used to monitor the PI3K activity because it depends on activation of the PI3K pathway in most cells [59].
Phosphorylation of AKT at S473 in C4-2 cells has been used in this study to assess the PI3K inhibition.Figure 1 shows representative Western blots that illustrate inhibition of S473AKT phosphorylation by 14 and by 20, which contains the leucine linker.Quantitative comparison of Western blots showed that IC50 values for 14 and 20 were 3.2-and 1.5-fold lower compared to ZSTK474, a widely used PI3K inhibitor.

Addition of Leucine to the Lead Compound (14)
Since compound 14 was quite active as a PI3K inhibitor (see screening discussion below), we undertook its preparation where leucine was added to the primary OH in the linker group.In our earlier work [33], we demonstrated that peptide (Mu-LEHSSKLQ) in prostate-specific prodrugs is cleaved between L and Q, so the leucine residue remains attached to the PI3K inhibitor.To mimic the prodrug cleaved by PSA, Boc-protected leucine was coupled to 14 to produce 19 in good yield and then the protecting group was removed to yield 20 (Scheme 6).cross-couple 13 to 17 under a variety of conditions but those also failed, leaving us to conclude that the terminal alkyne was not compatible with these conditions.Scheme 5. Use of propargyl amine as a potential peptide link point.

Addition of Leucine to the Lead Compound (14)
Since compound 14 was quite active as a PI3K inhibitor (see screening discussion below), we undertook its preparation where leucine was added to the primary OH in the linker group.In our earlier work [33], we demonstrated that peptide (Mu-LEHSSKLQ) in prostate-specific prodrugs is cleaved between L and Q, so the leucine residue remains attached to the PI3K inhibitor.To mimic the prodrug cleaved by PSA, Boc-protected leucine was coupled to 14 to produce 19 in good yield and then the protecting group was removed to yield 20 (Scheme 6).Scheme 6. Addition of leucine to lead compound 14.

Biological Activity
The biological activities of new PI3K-inhibitor compounds ( 14) and ( 20) were tested in C4-2 prostate cancer cells with constitutive activation of PI3K/AKT pathway.C4-2 cells were derived by passaging of prostate cancer metastases-derived LNCaP cells through nude mice and characterized by increased propensity to form metastases in mice.The PI3K/AKT pathway is activated in these cells due to frame-shift mutation in the PTEN tumor-suppressor gene.Activation of PI3K leads to accumulation of phosphatidylinositol (3,4,5)-trisphosphate (PIP3) in the plasma membrane that recruits protein kinases AKT and PDK1 through binding to their PH domains.Interaction of PIP3 with the PH domain changes AKT conformation and opens access for PDK1 to phosphorylate at T308, and also allows phosphorylation at S473 by TORC2 complex.Thus, phosphorylation of AKT at S473 and T308 is routinely used to monitor the PI3K activity because it depends on activation of the PI3K pathway in most cells [59].
Phosphorylation of AKT at S473 in C4-2 cells has been used in this study to assess the PI3K inhibition.Figure 1 shows representative Western blots that illustrate inhibition of S473AKT phosphorylation by 14 and by 20, which contains the leucine linker.Quantitative comparison of Western blots showed that IC50 values for 14 and 20 were 3.2-and 1.5-fold lower compared to ZSTK474, a widely used PI3K inhibitor.Scheme 6. Addition of leucine to lead compound 14.

Biological Activity
The biological activities of new PI3K-inhibitor compounds ( 14) and ( 20) were tested in C4-2 prostate cancer cells with constitutive activation of PI3K/AKT pathway.C4-2 cells were derived by passaging of prostate cancer metastases-derived LNCaP cells through nude mice and characterized by increased propensity to form metastases in mice.The PI3K/AKT pathway is activated in these cells due to frame-shift mutation in the PTEN tumor-suppressor gene.Activation of PI3K leads to accumulation of phosphatidylinositol (3,4,5)-trisphosphate (PIP 3 ) in the plasma membrane that recruits protein kinases AKT and PDK1 through binding to their PH domains.Interaction of PIP 3 with the PH domain changes AKT conformation and opens access for PDK1 to phosphorylate at T308, and also allows phosphorylation at S473 by TORC2 complex.Thus, phosphorylation of AKT at S473 and T308 is routinely used to monitor the PI3K activity because it depends on activation of the PI3K pathway in most cells [59].
Phosphorylation of AKT at S473 in C4-2 cells has been used in this study to assess the PI3K inhibition.Figure 1

General Methods
The general experimental methods used here were essentially the same as those we have described previously [33].Copies of spectra that the data presented below are taken from are also included as supplementary materials.

General Methods
The general experimental methods used here were essentially the same as those we have described previously [33].Copies of spectra that the data presented below are taken from are also included as supplementary materials.

2. 5 .
Attempts to Prepare an Analog of 14 with a Terminal Alkyne Rather than Primary OH as Peptide Link Point

Scheme 5 .
Scheme 5. Use of propargyl amine as a potential peptide link point.