Synthesis and Evaluation of Novel Carbocyclic Oxetanocin A (COA-Cl) Derivatives as Potential Tube Formation Agents

Abstract

Six novel carbocyclic oxetanocin A analogs (2-chloro-C.OXT-A; COA-Cl) with various hydroxymethylated or spiro-conjugated cyclobutane rings at the N⁹-position of the 2-chloropurine moiety were synthesized and evaluated using human umbilical vein endothelial cells. All prepared compounds (2a–f) showed good to moderate activity with angiogenic potency. Among these compounds, 100 µM cis–trans-2′,3′-bis(hydroxymethyl)cyclobutyl derivative (2b), trans-3′-hydroxymethylcyclobutyl analog (2d), and 3′,3′-bis(hydroxymethyl)cyclobutyl derivative (2e) had greater angiogenic activity, with relative tube areas of 3.43±0.44, 3.32±0.53, and 3.59±0.83 (mean±standard deviation (S.D.)), respectively, which was comparable to COA-Cl (3.91±0.78). These data may be important for further development of this class of compounds as potential tube formation agents.

Angiogenesis, the process in which new capillary vessels are built from preexisting ones, plays an important role in a variety of situations, including wound healing, tumor growth, and metastasis. Angiogenesis promoters are essential for detecting various symptoms arising due to a lack of blood flow, such as Buerger‘s disease or chronic arteriosclerosis obliterans in diabetic patients. However, the clinical applications of these promoters as medical agents remain minimal compared to angiogenesis inhibitors, which are utilized as antitumor therapeutics. The lack of angiogenesis promoters used as therapeutic agents may be due to limited knowledge of these promoter, with the exception of growth factors derived from living systems. For example, vascular endothelial growth factor (VEGF) and fibroblast growth factor (FGF) are macromolecular glycoproteins that are chemically and biologically unstable. Hence, they are used only in treatments administered by injection or gene transfection approaches.^1–3)

We synthesized an angiogenesis promoter 2-chloro-carbocyclic oxetanocin A (COA-Cl), (1a, racemate) (Fig. 1), which is the first low-molecular weight active compound in this class.^4–6) The size of compound 1a (molecular weight (MW)=284) is suitable for pharmaceutical use in percutaneous and transmucosal absorption. Therefore, it is expected to have practical applications as a therapeutic for wound healing, Buerger’s disease, or peripheral arterial obliterative disease, and could be developed as an adhesive skin patch, lotion, or oral agent. In addition to its angiogenic effect, we found that COA-Cl protects against oxygen-glucose deprivation in primary cortical neurons in a manner sensitive to suramin.⁷⁾ Because of its utility, COA-Cl is commercially available (Chemical Name: 2-Cl-C.OXT-A, WAKENYAKU Co., Ltd., Japan, Order Number: 032-21541, Quantity: 1 mg, Price: 15000 yen).

Fig. 1. Structures of COA-Cl (1a), COA-Oe (1b), C.OXT-A (1c), Cladribine (1d)

Our mechanistic studies revealed that the angiogenic effects of COA-Cl involves endothelial S1P₁ receptor. S1P₁ receptor is activated by a serum-borne lipid mediator sphingosine 1-phosphate (S1P) as a natural ligand, coupled with G-protein α i/o subunit, and is indispensable during normal vascular development.⁸⁾ Stimulation of S1P₁ by COA-Cl activates intracellular signaling pathways, ultimately inducing tube formation in human umbilical vein endothelial cells (HUVECs).⁹⁾ In agreement with our previous reports, COA-Cl promotes the phosphorylation and activation of the mitogen activated protein kinases (MAPK), extracellular signaling-related kinases (ERK1/2), which induce angiogenesis and tube formation, a hallmark of HUVEC angiogenesis.⁴⁾ The ability of COA-Cl to modulate ERK1/2 phosphorylation and tube formation is similar to S1P (Fig. 2), the naturally occurring ligand for the S1P₁ receptor.^10–12) Thus, these results indicate that COA-Cl is a novel and potent angiogenic agent that may behave as a partial agonist of the S1P₁ receptor.

Fig. 2. Structures of S1P, FTY-720, FTY-720-P, KRP-203, and ACT-128800

Several S1P derivatives have been developed as S1P₁ agonists, including FTY-720,^13,14) FTY-720-P,^13,14) KRP-203,¹⁵⁾ and ACT-128800¹⁶⁾ (Fig. 2). These compounds can block the egress of T cells from the thymus and lymphoid organs,^17,18) and may be promising oral therapeutics for autoimmune disorders. Nevertheless, outside of COA-Cl 1a, no other compound has been shown to exert neovascularization effects via the S1P₁ receptor. Thus, COA-Cl has the potential for clinical applications as an angiogenic medicine acting via the S1P₁ receptor.

Understanding of the unique mechanism of COA-Cl 1a has triggered intensive effort to discover COA-Cl analogs with enhanced angiogenic activity and reduced cytotoxicity.^4,6) However, our previous work has resulted in analogs with no or less potency to COA-Cl 1a. For example, the COA-Cl derivative, COA-OMe (1b), in which the 2-chloro group at the 2-position in the purine system is substituted with a methoxy group (Fig. 1), displayed half the vascularization activity of COA-Cl 1a.⁶⁾ Furthermore, substitution of an H in place of the Cl (1c: C.OXT-A) or replacement of a cyclobutane moiety with the pentose ring (1d: cladribine, used to treat hairy cell leukemia and multiple sclerosis) showed no angiogenic potency.⁴⁾ Collectively, these studies indicated that the structure of the 2-chloropurine skeleton and the cyclobutane ring moiety in COA-Cl 1a was essential for its angiogenic activity. Nevertheless, the role of the two hydroxymethyl groups bound to the cyclobutane ring at the 2′ and 3′ position remain unclear.

In the present study, we sought to further elucidate the angiogenic function of hydroxymethyl group(s) on COA-Cl 1a. More specifically, six novel COA-Cl analogs racemates 2a, b, and achiral compounds 2c–f (Fig. 3) with a variety of hydroxymethylated or spiro-conjugated cyclobutane rings at the N⁹-position in the 2-chloropurine moiety were synthesized and evaluated in HUVECs.

Fig. 3. Structures of COA-Cl (1) and COA-Cl Analogs (2a–f)

Results and Discussion

Chemistry

The COA-Cl analogs 2a–d shown in Chart 1 were prepared using known cyclobutyl alcohols (3a,¹⁹⁾ b,²⁰⁾ c,^21,22) d,^21,23)) as starting materials in two steps reactions, according to a modified method from a previous report.²⁴⁾ Secondary alcohols 3a–d were treated with 2,6-dichloropurine and underwent the Mitsunobu reaction²⁵⁾ in the presence of PPh₃ and DIAD to yield 9-cyclobutyl-2,6-dichloropurine congeners, 4a–d, in 37–84% yields, which were treated with methanolic ammonia to give the desired products 2a–d in 68–87% yields.

Chart 1. Synthesis of Compounds 2a–d

Reagents and conditions: i, 2,6-Dichloropurine, PPh₃, DIAD, THF, 50°C; ii, NH₃, MeOH, 100°C.

The relative configurations of 4b–d were determined on the basis of nuclear Overhauser effect spectroscopy (NOESY) correlations (Fig. 4) and comparison with previous literature.^20–23) For compound 4b, the NOESY correlations between H-1′ and H-4′b, between H-4′a and H-3′, and between H-2′ and H-3′, and the absence of NOESY correlations between H-1′ and H-2′, suggested the 1′,2′-trans-2′,3′-cis congener (4b). Similarly, for compounds 4c and d, the NOESY correlations between H-1′ and H-3′ indicated the 1′,3′-cis analog (4c), whereas the absence of NOESY correlations between H-1′ and H-3′ indicated the 1′,3′-trans derivative (4d).

Fig. 4. Key NOESY (H↔H) Correlations and Relative Configuration of Compounds 4b–d

The synthesis of COA-Cl derivatives 2e and f is described in Chart 2. Starting from 3,3-bis(hydroxymethyl)-1-benzyloxycyclobutane 5,²⁶⁾ we prepared dibenzoyl ester 6 in a 90% yield, followed by debenzylation with palladium carbon (Pd–C)/H₂ to afford alcohol 7 in 91% yield. Compound 7 was then condensed with 2,6-dichloropurine using the Mitsunobu reaction²⁵⁾ to afford 8 in a 75% yield, which was treated with methanolic ammonia to give the desired product 2e in a 50% yield. As for the preparation of spiro-counterpart 2f, compound 5 was transformed into its (bis)tosyl derivative, which in turn was alkylated by diisopropylmalonate to form diisopropyl dicarboxylated product 9 in an 87% yield, possessing the spiro[3.3]heptane core.²⁷⁾ The resulting product 9 was reduced with LiAlH₄ to afford diol 10 in a 34% yield, which was subsequently acylated with benzoyl chloride to yield dibenzoyl ester 11 in an 87% yield. Compound 11 was hydrogenated with Pd–C to produce secondary alcohol 12 in a 72% yield, followed by the Mitsunobu reaction²⁵⁾ to provide the spiro-condensation product 13 in an 86% yield, which was treated with methanolic ammonia to give the objective spiro-diol 2f in a 70% yield.

Chart 2. Synthesis of Compounds 2e and f

Reagents and conditions: i, BzCl, Pyridine, rt; ii, Pd/C, H₂, rt; iii, 2,6-Dichloropurine, PPh₃, DIAD, THF, 50°C; iv, NH₃, MeOH, 100°C; v, TsCl, Pyridine, rt; vi, NaH, Dimethylmalonate, toluene, reflux; vii, LiAlH₄, THF, −20°C.

Biological Activity

The effects of the synthesized nucleoside analogs (2a–f) on angiogenesis were examined. We evaluated the analogs using a well-established tube formation assay.^4,28,29) All experiments were performed 2–6 times. Ten days following the incubation of the fibroblasts and additives, HUVECs were stained using the Tubule Staining Kit for CD31. Typical pictures are shown in Fig. 5. VEGF was used as a positive control. The area of the formed tube was represented as a relative value to tubes formed in wells without additives (saline). Figure 6 shows the dose-dependent tube formation responses, as well as the solvent control (10% saline) and positive control (VEGF, 10 ng/mL).

Fig. 5. Pictures of Stimulated Tube Formation by 1a and 2a–f

Additives are shown under the pictures.

Fig. 6. Tube Formation Assay Using Compounds 1a and 2a–f

The tube formation assay was performed using an angiogenesis kit, and the estimation was performed 10 d after incubation with the designated compounds. The area of the formed tube was represented as a relative value versus control wells with no additives. The effect of the solvent (10% saline) and positive control (VEGF; 10 ng/mL) are shown together. Results were expressed as mean±S.E. of 2–6 individual experiments.

As shown in Figs. 5 and 6, angiogenic potencies were observed in every prepared compound, including the cis–trans-2′,3′-bis(hydroxymethyl)cyclobutyl derivative, 2a, cis–trans-2′,3′-bis(hydroxymethyl)cyclobutyl derivative, 2b, cis-3′-hydroxymethylcyclobutyl derivative, 2c, trans-3′-hydroxymethylcyclobutyl derivative, 2d, 3′,3′-bis(hydroxymethyl)cyclobutyl derivative 2e, and 2′,2′-bis(benzoyloxymethyl)spiro[3′.3′]hept-6′-yl derivative 2f (Fig. 3). Among these, at a concentration of 100 µM, 2b, d, and e exerted greater angiogenic activity (3.43±0.44, 3.32±0.53, and 3.59±0.83 [mean±standard deviation (S.D.)], respectively) which was comparable to COA-Cl 1a (3.91±0.78). Compound 2d was most angiogenic when using 10 µM (2.74±0.28), which was superior to the potency of COA-Cl 1a (2.44±0.89 at 10 µM) in our previous report⁴⁾ and 10 ng/mL VEGF (2.13±0.61). Surprisingly, the angiogenic activity of spiro analog 2f at 10 µM was 2.42±0.46, which was nearly identical to that of 1a at same concentration. The potency of 2d–f reached a maximum around 10–100 µM, and decreased at 100–200 µM, indicating that the cellular proliferation was eventually inhibited at a high analog concentration.

Moreover, as shown in Fig. 7, we performed a proliferation assay because angiogenesis is intimately associated with complex cellular processes, including proliferation of endothelial cells,^4,28,29) which showed that compounds 2a–f promoted the proliferation of HUVECs at a higher concentration (e.g., 2a: 200 µM: 1.12±0.07, n=5, p=0.031), whereas 2f at a concentration of 10 µM only inhibited the proliferation of HUVECs (0.72±0.19, n=5, p=0.005).

Fig. 7. Effect of Compounds 2a–f on the Proliferation

Forty-eight hours after the addition of additives, the cell viability was measured with a Cell Counting Kit-8. Cell viabilities were represented as relative values to that of the control well with no additive. The effect of the solvent, 10% saline as well as a positive control, VEGF (10 ng/mL) were shown together. Results were expressed as mean±S.E. of more than four individual experiments. * p<0.01 and ^# p<0.05 vs. solvent (by Student’s t-test).

Conclusion

In the present study, a series of newly synthesized COA-Cl analogs exhibited good to moderate angiogenic activity. In particular, 100 µM of the COA-Cl analogs, 2b, d, and e induced enhanced angiogenesis, with relative tube areas of 3.43±0.44, 3.32±0.53, and 3.59±0.83, respectively. This was comparable with the tube area induced by COA-Cl (3.91±0.78). Overall, our data may serve as the basis for further modification to identify more potent candidates to induce tube formation.

Experimental

Chemistry

¹H-NMR and ¹³C-NMR spectra were taken with a Ultrashield™ 400 Plus FT NMR System (BRUKER). Chemical shifts and coupling constants (J) were given in δ and Hz, respectively. Melting points were determined on a Yanaco MP-500D. High-resolution mass spectrometry was performed on a APEX IV mass spectrometer (BRUKER) with electrospray ionization mass spectroscopy (ESI-MS).

General Procedure for the Synthesis of 4a–d

A solution of compound 3a–d (0.38 mmol), triphenylphosphine (104.9 mg, 0.40 mmol), 2,6-dichloropurine (75.4 mg, 0.40 mmol) and DIAD (diisopropyl azodicarbosylate, 78.8 µL, 0.40 mmol) in tetrahydrofuran (THF) (3.0 mL) was stirred at 50°C. After 8–15 h stirring, the residual solution was purified by silica gel column chromatography (60% AcOEt in hexane) to give 4a–d.

2,6-Dichloro-9-[cis–trans-2′,3′-bis(benzoyloxymethyl)cyclobutyl]purine (4a)

Yield 37%; oil; ¹H-NMR (400 MHz, CDCl₃) δ: 8.34 (1H, s, H-8), 8.06–8.12 (2H, m, Bz), 7.45–7.64 (6H, m, Bz), 7.31–7.38 (2H, m, Bz), 5.52 (1H, m, H-1′), 4.53–4.66 (2H, m, 3′-CH₂OBz), 4.30–4.41 (2H, m, 2′-CH₂OBz), 3.42 (1H, m, H-2′), 3.20 (1H, m, H-4′a), 2.96 (1H, m, H-3′), 2.76 (1H, m, H-4′b); ¹³C-NMR (100 MHz, CDCl₃) δ: 166.4, 165.5, 153.4, 153.0, 151.8, 144.2, 144.2, 133.4, 133.4, 131.0, 129.6, 129.0, 128.7, 128.6, 128.5, 65.9, 62.4, 48.8, 41.4, 32.3, 26.9; high resolution (HR)-MS (ESI) Calcd for C₂₅H₂₀Cl₂N₄NaO₄ [M+Na]⁺: 533.07538. Found 533.0713.

2,6-Dichloro-9-[trans–cis-2′,3′-bis(benzoyloxymethyl)cyclobutyl]purine (4b)

Yield 38%; white crystals; ¹H-NMR (400 MHz, CDCl₃) δ: 8.28 (1H, s, H-8), 8.08 (2H, m, Bz), 7.69 (2H, m, Bz), 7.30–7.63 (6H, m, Bz), 5.25 (1H, m, H-1′), 4.52–4.75 (4H, m, 2′-CH₂OBz and 3′-CH₂OBz), 3.73 (1H, m, H-2′), 3.17 (1H, m, H-3′), 2.99 (1H, m, H-4′a), 2.58 (1H, m, H-4′b); ¹³C-NMR (100 MHz, CDCl₃) δ: 166.3, 166.2, 152.9, 152.8, 151.8, 144.4, 133.4, 133.4, 131.1, 129.6, 129.6, 129.2, 129.0, 128.6, 128.4, 64.0, 63.1, 51.5, 43.4, 30.2, 29.2; HR-MS (ESI) Calcd for C₂₅H₂₀Cl₂N₄NaO₄ [M+Na]⁺: 533.07538. Found 533.07531; mp: 57.5–58.5°C.

2,6-Dichloro-9-(cis-3′-benzoyloxymethylcyclobutyl)purine (4c)

Yield 62%; white crystals; ¹H-NMR (400 MHz, CDCl₃) δ: 8.20 (1H, s, H-8), 8.05 (2H, m, Bz), 7.59 (1H, m, Bz), 7.47 (2H, m, Bz), 5.00 (1H, m, H-1′), 4.47 (2H, d, J=5.2 Hz, 3′-CH₂OBz), 2.85 (2H, m, H-2′a and H-4′a), 2.75 (1H, m, H-3′), 2.60 (2H, m, H-2′b and H-4′b); ¹³C-NMR (100 MHz, CDCl₃) δ: 166.5, 152.9, 152.8, 151.8, 144.1, 133.3, 129.9, 129.5, 129.5, 128.6, 128.6, 66.8, 46.1, 33.1, 28.0; HR-MS (ESI) Calcd for C₁₇H₁₄Cl₂N₄NaO₂ [M+Na]⁺: 399.03860. Found 399.03772; mp: 164.2–165.6°C.

2,6-Dichloro-9-(trans-3′-benzoyloxymethylcyclobutyl)purine (4d)

Yield 84%; white crystals; ¹H-NMR (400 MHz, CDCl₃) δ: 8.25 (1H, s, H-8), 8.10 (2H, m, Bz), 7.61 (1H, m, Bz), 7.49 (2H, m, Bz), 5.30 (1H, m, H-1′), 4.54 (2H, d, J=6.4 Hz, 3′-CH₂OBz), 3.04 (1H, m, H-3′), 2.91 (2H, m, H-2′a and H-4′a), 2.70 (2H, m, H-2′b and H-4′b); ¹³C-NMR (100 MHz, CDCl₃) δ: 166.6, 153.0, 152.9, 151.9, 144.2, 133.3, 131.2, 129.9, 129.6, 128.6, 66.7, 48.1, 31.8, 28.3; HR-MS (ESI) Calcd for C₁₇H₁₄Cl₂N₄NaO₂ [M+Na]⁺: 399.03860. Found 399.03753; mp: 147.7–149.5°C.

General Procedure for the Synthesis of 2a–d

Compounds 4a–d (0.12 mmol) was dissolved in NH₃ (14.0 mL)/MeOH (3.0 mL), and then sealed and stirred for 1–3 d at 100°C. The mixture was evaporated, and the residue was purified by silica gel column chromatography (20–25% MeOH in CH₂Cl₂) to afford 2a–d.

2-Amino-6-chloro-9-[cis–trans-2′,3′-bis(hydroxymethyl)cyclobutyl]purine (2a)

Yield 84%; white crystals; ¹H-NMR (400 MHz, CDCl₃) δ: 8.36 (1H, s, H-8), 7.70 (2H, br s, 6-NH₂), 5.07 (1H, m, H-1′), 4.75 (1H, t, J=4.8 Hz, 3′-OH), 4.30 (1H, t, J=5.2 Hz, 2′-OH), 3.56 (2H, m, 3′-CH₂OBz), 3.31 (2H, m, 2′-CH₂OBz), 2.90 (1H, m, H-4′a), 2.58 (1H, m, H-2′), 2.33 (1H, m, H-4′b), 2.30 (1H, m, H-3′); ¹³C-NMR (100 MHz, CDCl₃) δ: 158.0, 155.1, 152.6, 142.2, 119.1, 65.4, 61.6, 50.0, 45.4, 36.5, 27.8; HR-MS (ESI) Calcd for C₁₁H₁₄ClN₅NaO₂ [M+Na]⁺: 306.07282. Found 306.07162; mp: 226.5–227.7°C.

2-Amino-6-chloro-9-[trans–cis-2′,3′-bis(hydroxymethyl)cyclobutyl]purine (2b)

Yield 87%; white crystals; ¹H-NMR (400 MHz, CD₃OD) δ: 8.14 (1H, s, H-8), 4.80 (1H, m, H-1′), 3.65–3.80 (4H, m, 2′-CH₂OBz and 3′-CH₂OBz), 3.20 (1H, m, H-2′), 2.62 (1H, m, H-4′a), 2.59 (1H, m, H-3′), 2.29 (1H, m, H-4′b); ¹³C-NMR (100 MHz, CD₃OD) δ: 158.0, 155.0, 152.0, 141.7, 119.3, 62.6, 61.6, 52.0, 47.3, 34.2, 29.7; HR-MS (ESI) Calcd for C₁₁H₁₄ClN₅NaO₂ [M+Na]⁺: 306.07282. Found 306.07305; mp: 209.1–210.5°C.

2-Amino-6-chloro-9-(cis-3′-hydroxymethylcyclobutyl)purine (2c)

Yield 69%; white crystals; ¹H-NMR (400 MHz, CD₃OD) δ: 8.22 (1H, s, H-8), 4.90 (1H, m, H-1′), 3.67 (2H, m, 3′-CH₂OBz), 2.66 (2H, m, H-2′a and H-4′a), 2.44 (3H, m, H-2′b, H-3′ and H-4′b); ¹³C-NMR (100 MHz, CD₃OD) δ: 158.0, 155.0, 151.9, 141.3, 119.2, 66.1, 46.8, 33.3, 31.7; HR-MS (ESI) Calcd for C₁₀H₁₂ClN₅NaO [M+Na]⁺: 276.06226. Found 276.06236; mp: 152.3–153.4°C.

2-Amino-6-chloro-9-(trans-3′-hydroxymethylcyclobutyl)purine (2d)

Yield 68%; white crystals; ¹H-NMR (400 MHz, CD₃OD) δ: 8.17 (1H, s, H-8), 5.00 (1H, m, H-1′), 3.64 (2H, m, 3′-CH₂OBz), 2.67 (2H, m, H-2′a and H-4′a), 2.52 (3H, m, H-2′b, H-3′ and H-4′b); ¹³C-NMR (100 MHz, CD₃OD) δ: 158.0, 155.1, 151.9, 141.4, 119.3, 65.7, 48.6, 32.3, 32.1; HR-MS (ESI) Calcd for C₁₀H₁₂ClN₅NaO [M+Na]⁺: 276.06226. Found 276.06239; mp: 256.5–257.9°C.

3,3-Bis(benzoyloxymethyl)-1-benzyloxycyclobutane (6)

Compound 5 (2.48 g, 11.16 mmol) was dissolved in dry pyridine (24.2 mL), and benzoyl chloride (2.72 mL, 23.44 mmol) was added dropwise to the solution at 0°C, and the mixture was stirred for 2 h at room temperature. The mixture was then cooled to 0°C, 5.0 mL of water was added over 3 min, and the mixture was stirred at room temperature for 1 h. The solvent was removed in vacuo and the residue was extracted with AcOEt. The organic extracts were washed with water, saturated potassium carbonate solution and saturated sodium chloride solution, dried with sodium sulfate, and then evaporated. The residue was purified by silica gel column chromatography (30% AcOEt in hexane) to give a colorless oil 6 (4.31 g, 10.04 mmol, 90%). ¹H-NMR (400 MHz, CDCl₃) δ: 7.99–8.08 (4H, m, Bz), 7.52–7.58 (2H, m, Bz), 7.38–7.45 (4H, m, Bz), 7.26–7.36 (5H, m, Bn), 4.44 (4H, s, 3-CH₂OBz ×2), 4.42 (2H, s, Bn), 4.22 (1H, m, H-1), 2.40 (2H, m, H-2a and H-4a), 2.19 (2H, m, H-2b and H-4b); ¹³C-NMR (100 MHz, CDCl₃) δ: 166.6, 166.5, 138.0, 133.2, 133.1, 130.0, 129.8, 129.7, 129.6, 128.5, 128.4, 128.4, 127.9, 127.7, 70.1, 68.7, 68.4, 67.7, 35.0, 34.7; HR-MS (ESI) Calcd for C₂₇H₂₆NaO₅⁺ [M+Na]⁺: 453.16725. Found 453.16578.

3,3-Bis(benzoyloxymethyl)cyclobutan-1-ol (7)

To a solution of this compound 6 (4.18 g, 9.71 mmol) in methanol (53.4 mL) was added 10% Pd on carbon (3.24 g). The reaction mixture was stirred at room temperature under H₂ atmosphere (1 atm) for 18 h. The suspension was filtered through a pad of Celite. The filtrate was evaporated, and then the residue was purified by silica gel column chromatography (60% AcOEt in hexane) to give a colorless oil 7 (3.01 g, 8.84 mmol, 91%). ¹H-NMR (400 MHz, CDCl₃) δ: 8.00 8.07 (4H, m, Bz), 7.51–7.58 (2H, m, Bz), 7.38–7.45 (4H, m, Bz), 4.50 (1H, m, H-1), 4.47 (2H, s, 3-CH₂OBz), 4.42 (2H, s, 3-CH₂OBz), 2.46 (2H, m, H-2a and H-4a), 2.10 (2H, m, H-2b and H-4b); ¹³C-NMR (100 MHz, CDCl₃) δ: 166.6, 166.6, 133.2, 133.1, 129.8, 129.7, 129.6, 129.6, 128.5, 128.4, 68.9, 67.6, 62.7, 37.6, 34.2; HR-MS (ESI) Calcd for C₂₀H₂₀NaO₅⁺ [M+Na]⁺: 363.12029. Found 363.11926.

2,6-Dichloro-9-[3′,3′-bis(benzoyloxymethyl)cyclobutyl]purine (8)

A solution of compound 7 (204.2 mg, 0.60 mmol), triphenylphosphine (157.4 mg, 0.60 mmol), 2,6-dichloropurine (94.5 mg, 0.50 mmol) and TMAD (N,N,N′,N′-tetramethylazodicarboxamide, 103.3 mg, 0.60 mmol) in THF (5.0 mL) was stirred at 50°C. After 22 h stirring, the residual solution was purified by silica gel column chromatography (60% AcOEt in hexane) to give as a colorless oil 8 (192.0 mg, 0.38 mmol, 75%). ¹H-NMR (400 MHz, CDCl₃) δ: 8.23 (1H, s, H-8), 7.99–8.11 (4H, m, Bz), 7.56–7.64 (2H, m, Bz), 7.41–7.51 (4H, m, Bz), 5.27 (1H, m, H-1′), 4.64 (2H, s, 3′-CH₂OBz), 4.63 (2H, s, 3′-CH₂OBz), 2.93 (2H, m, H-2′a and H-4′a), 2.83 (2H, m, H-2′b and H-4′b); ¹³C-NMR (100 MHz, CDCl₃) δ: 166.5, 166.3, 152.9, 152.0, 144.1, 133.5, 133.5, 129.7, 129.6, 128.6, 67.8, 66.4, 60.4, 36.8, 34.3; HR-MS (ESI) Calcd for C₂₅H₂₀Cl₂N₄NaO₂ [M+Na]⁺: 533.07538. Found 533.07375.

2-Amino-6-chloro-9-[3′,3′-bis(hydroxymethyl)cyclobutyl]purine (2e)

Compound 8 (192.0 mg, 0.38 mmol) was dissolved in NH₃ (14.0 mL)/MeOH (3.0 mL), and then sealed and stirred for 4 d at 100°C. The mixture was evaporated, and the residue was purified by silica gel column chromatography (25% MeOH in CH₂Cl₂) to give 2e as white crystals (53.7 mg, 0.19 mmol, 50%). ¹H-NMR (400 MHz, DMSO-d₆) δ: 8.38 (1H, s, H-8), 7.80 (1H, br s, 6-NH₂), 4.94 (1H, m, H-1′), 4.88 (1H, t, J=5.2 Hz, 3′-CH₂OH), 4.78 (1H, t, J=5.2 Hz, 3′-CH₂OH), 3.56 (2H, d, J=5.2 Hz, 3′-CH₂OH), 3.49 (2H, d, J=5.2 Hz, 3′-CH₂OH), 2.47 (2H, m, H-2′a and H-4′a), 2.39 (2H, m, H-2′b and H-4′b); ¹³C-NMR (100 MHz, DMSO-d₆) δ: 156.8, 152.7, 150.3, 139.7, 118.0, 65.0, 64.1, 43.2, 33.0; HR-MS (ESI) Calcd for C₁₁H₁₄ClN₅NaO₂ [M+Na]⁺: 306.07282. Found 306.07166; mp: 239.6°C.

6-Benzyloxyspiro[3.3]heptane-2,2-dicarboxylic Acid Diisopropyl Ester (9)

Compound 5 (2.06 g, 9.27 mmol) was dissolved in dry pyridine (7.7 mL) and cooled by an ice bath. p-Toluenesulfonyl chloride (4.42 g, 23.17 mmol) was added in small portions to the cooled solution under stirring in a dry atmosphere. The mixture was left standing overnight. The precipitate was filtered off, the filtrate was carefully poured into water (50 mL), and the product was extracted with AcOEt. The extract was washed with 1 N HCl, water and saturated sodium chloride solution, dried with sodium sulfate, and then evaporated. The residue was purified by silica gel column chromatography (40% AcOEt in hexane) to give the ditosylate as an oil (4.70 g, 8.81 mmol, 95%).

Next, disopropylmalonate (714.9 µL, 3.76 mmol) was slowly added to a suspension of NaH (60% suspension in mineral oil) (106.9 mg, 4.17 mmol) in dry N,N-dimethylformamide (DMF) (10.0 mL) under stirring at such a rate that the temperature was maintained below 70°C. The ditosylate (1.00 g, 1.88 mmol) and KI (31.5 mg, 0.19 mmol) were added to the mixture in one portion. The mixture was then stirred at 140°C for 14 h. Then, it was cooled, poured into saturated aqueous solution of ammonium chloride, and the product was extracted with AcOEt, washed with saturated aqueous sodium chloride solution, dried with sodium sulfate, and then evaporated. The residue was purified by silica gel column chromatography (25% AcOEt in hexane) to give as a colorless oil 9 (646.0 mg, 1.73 mmol, 92%). ¹H-NMR (400 MHz, CDCl₃) δ: 7.26–7.35 (5H, m, Bn), 5.03 (2H, hept, J=6.0 Hz, ⁱPr), 4.37 (2H, s, Bn), 3.92 (1H, m, H-6), 2.53 (4H, s, H-1 and H-3), 2.36 (2H, m, H-5a and H-7a), 2.02 (2H, m, H-5b and H-7b), 1.22 (12H, d, J=6.0 Hz, ⁱPr); ¹³C-NMR (100 MHz, CDCl₃) δ: 171.3, 138.2, 128.4, 127.8, 127.6, 70.0, 68.7, 68.2, 49.3, 43.0, 40.7, 30.4, 21.5; HR-MS (ESI) Calcd for C₂₂H₃₀NaO₅ [M+Na]⁺: 397.19855. Found 397.19838.

6-Benzyloxy-2,2-bis(hydroxymethyl)spiro[3.3]heptane (10)

A solution of compound 9 (620.1 mg, 1.66 mmol) in dry THF (5.0 mL) was added by drops to an ice-cooled suspension of lithium aluminum hydride (103.9 mg, 2.34 mmol) in dry THF (10.0 mL), and the mixture was stirred for 1 h at room temperature, then cooled. The solution was treated with AcOEt (2.0 mL) and water (1.0 mL) to give a gel. The solid was filtrated and washed successively with AcOEt (100 mL). The filtrate and washing were combined, and washed with saturated sodium chloride solution, dried with sodium sulfate, and then evaporated. The residue was purified by silica gel column chromatography (20% MeOH in AcOEt) to give as an oil 10 (149.6 mg, 0.56 mmol, 34%). ¹H-NMR (400 MHz, CDCl₃) δ: 7.24–7.36 (5H, m, Bn), 4.37 (2H, s, Bn), 3.94 (1H, m, H-6), 3.64 (4H, s, 2-CH₂OH ×2), 2.32 (2H, m, H-5a and H-7a), 2.02 (2H, m, H-5b and H-7b), 1.82 and 1.85 (4H, s, H-1 and H-3); ¹³C-NMR (100 MHz, CDCl₃) δ: 138.2, 128.4, 127.9, 127.7, 70.1, 69.3, 68.7, 44.3, 39.6, 39.2, 38.8, 29.5; HRMS (ESI) Calcd for C₁₆H₂₂NaO₃ [M+Na]⁺: 285.14612. Found 285.14612.

6-Benzyloxy-2,2-bis(benzoyloxymethyl)spiro[3.3]heptane (11)

Compound 10 (151.3 mg, 0.58 mmol) was dissolved in dry pyridine (1.3 mL), and benzoyl chloride (200.0 µL, 1.73 mmol) was added dropwise to the solution at 0°C, and the mixture was stirred for 3 h at room temperature. The mixture was then cooled to 0°C, 0.3 mL of water was added over 1 min, and the mixture was stirred at room temperature for 30 min. The solvent was removed in vacuo and the residue was extracted with AcOEt. The organic extracts were washed with water, saturated potassium carbonate solution and saturated sodium chloride solution, dried with sodium sulfate, and then evaporated. The residue was purified by silica gel column chromatography (30% AcOEt in hexane) to give a colorless oil 11 (238.2 mg, 0.51 mmol, 87%). ¹H-NMR (400 MHz, CDCl₃) δ: 8.02 (4H, m, Bz), 7.55 (2H, m, Bz), 7.42 (4H, m, Bz), 7.24–7.36 (5H, m, Bn), 4.38, 4.40 and 4.41 (6H, s, 2-CH₂OBz ×2 and Bn), 3.96 (1H, m, H-6), 2.43 (2H, m, H-5a and H-7a), 2.13 (4H, s, H-1 and H-3), 2.11 (2H, m, H-5b and H-7b); ¹³C-NMR (100 MHz, CDCl₃) δ: 166.6, 138.2, 133.1, 130.0, 129.6, 128.4, 128.4, 127.8, 127.6, 70.1, 68.5, 68.0, 67.9, 44.2, 39.6, 38.8, 37.3, 29.5; HR-MS (ESI) Calcd for C₃₀H₃₀NaO₅ [M+Na]⁺: 493.19855. Found 493.19666.

2,2-Bis(benzoyloxymethyl)-6-hydroxyspiro[3.3]heptane (12)

To a solution of this compound 11 (194.7 mg, 0.41 mmol) in methanol (5.0 mL) was added 10% Pd on carbon (136.8 mg). The reaction mixture was stirred at room temperature under H₂ atmosphere (1 atm) for 10 h. The suspension was filtered through a pad of Celite. The filtrate was evaporated, and then the residue was purified by silica gel column chromatography (60% AcOEt in hexane) to give a colorless oil 12 (112.8 mg, 0.30 mmol, 72%). ¹H-NMR (400 MHz, CDCl₃) δ: 8.02 (4H, m, Bz), 7.55 (2H, m, Bz), 7.42 (4H, m, Bz), 4.40 and 4.41 (6H, s, 2-CH₂OBz ×2 and Bn), 4.20 (1H, m, H-6), 2.49 (2H, m, H-5a and H-7a), 2.12 and 2.14 (4H, s, H-1 and H-3), 2.02 (2H, m, H-5b and H-7b); ¹³C-NMR (100 MHz, CDCl₃) δ: 166.6, 133.1, 130.0, 129.6, 128.4, 68.0, 62.8, 47.1, 39.4, 38.6, 37.3, 28.7; HR-MS (ESI) Calcd for C₂₃H₂₄NaO₅ [M+Na]⁺: 403.15159. Found 403.15044.

2,6-Dichloro-9-[2′,2′-bis(benzoyloxymethyl)spiro[3′.3′]hept-6′-yl]purine (13)

A solution of compound 12 (102.8 mg, 0.27 mmol), triphenylphosphine (85.0 mg, 0.32 mmol), 2,6-dichloropurine (61.3 mg, 0.32 mmol) and DIAD (diisopropyl azodicarbosylate, 63.8 µL, 0.32 mmol) in THF (3.0 mL) was stirred at 50°C. After 20 h stirring, the residual solution was purified by silica gel column chromatography (60% AcOEt in hexane) to give as a colorless oil 13 (127.7 mg, 0.23 mmol, 86%). ¹H-NMR (400 MHz, CDCl₃) δ: 8.15 (1H, s, H-8), 8.04 (4H, m, Bz), 7.58 (2H, m, Bz), 7.45 (4H, m, Bz), 4.98 (1H, m, H-6′), 4.46 and 4.47 (4H, s, 2′-CH₂OBz ×2), 2.86 (2H, m, H-5′a and H-7′a), 2.73 (2H, m, H-5′b and H-7′b), 2.29 and 2.38 (4H, s, H-1′ and H-3′); ¹³C-NMR (100 MHz, CDCl₃) δ: 166.7, 152.9, 152.8, 151.8, 144.0, 133.3, 131.1, 129.8, 129.6, 128.5, 128.4, 67.7, 53.4, 43.5, 38.9, 38.3, 37.4, 31.6; HR-MS (ESI) Calcd for C₂₈H₂₄Cl₂N₄NaO₄ [M+Na]⁺: 573.10668. Found 573.10532.

2-Amino-6-chloro-9-[2′,2′-bis(hydroxymethyl)spiro[3′.3′]hept-6′-yl]purine (2f)

Compound 13 (122.6 mg, 0.22 mmol) was dissolved in NH₃ (14.0 mL)/MeOH (3.0 mL), and then sealed and stirred for 1 d at 100°C. The mixture was evaporated, and the residue was purified by silica gel column chromatography (25% MeOH in CH₂Cl₂) to give 2f as white crystals (50.3 mg, 0.15 mmol, 70%). ¹H-NMR (400 MHz, DMSO-d₆) δ: 8.31 (1H, s, H-8), 7.72 (2H, br s, 6-NH₂), 4.74 (1H, m, H-6′), 4.47 (2H, t, J=5.6 Hz, 2′-CH₂OH×2), 3.34 (4H, t, J=5.6 Hz, 2′-CH₂OH ×2), 2.48–2.60 (4H, m, H-5′ and H-7′), 1.84 and 1.96 (4H, s, H-1′ and H-3′); ¹³C-NMR (100 MHz, DMSO-d₆) δ: 156.7, 152.7, 150.3, 139.9, 118.0, 65.0, 43.9, 43.2, 37.7, 37.0, 30.7; HR-MS (ESI) Calcd for C₁₄H₁₈ClN₅NaO₂ [M+Na]⁺: 346.10412. Found 346.10258; mp: 233.8°C.

Assay

Angiogenesis Kit, Tubule Staining Kit for CD31, human umbilical vein endothelial cells (HUVEC), fetal bovine serum (FBS), vascular endothelial growth factor-A (VEGF), HuMedia EG2, and HuMedia EB2 were purchased from Kurabo Co. (Osaka, Japan). Cell Counting Kit 8 was supplied by Dojindo Molecular Technologies (Kumamoto, Japan).

Cell Culture

A co-culture system of HUVEC and human fibroblasts (Angiogenesis Kit) was supplied in 24-well plates by Kurabo.⁴⁾ Cells were incubated for 10 d prior to analysis with 450 µL of the culture medium and a 50 µL of saline that includes various additives. Culture medium was changed every 3 d, each time including freshly prepared additives.

Tube Formation Assay

Ten days following incubation periods with co-cultured fibroblasts and substrates (2a–f), HUVEC were stained using Tubule Staining Kit for CD31.⁴⁾ The area of the formed tube was measured by the ImageJ program. Two pictures from each well were provided for the estimation. VEGF (10 ng/mL) was used as a positive control.

Proliferation Assay⁴⁾

HUVECs were seeded on gelatin-coated 96-well plates, typically at 3000 cells/well in 100 µL of maintenance medium. After seeding, the plates were incubated for 24 h to permit anchorage, and then, the culture medium was changed with the assay medium, consisting of 90 µL of HuMedia EB2 with 2% heat-inactivated FBS and 10 µL of saline containing additives. The contents of HuMedia EB2 are almost identical to those of HuMedia EG2, except for the fact that the former does not include FBS, a growth factor, or antibiotics. The proliferation assay was performed using a Cell Counting Kit-8 48 h after the addition of compounds 2a–f.

Statistics

All experiments were performed at least 5 times. Mean values and standard error (S.E.) of the mean are shown. Statistical differences between groups were analyzed by Student’s t-test and Dunnett’s multiple comparison test using Microsoft Excel, and ANOVA followed by Scheffe’s F test using STAT VIEW II (Abacus Concepts). A p value less than 0.05 was considered to be statistically significant.

Acknowledgment

This research was partially supported by a Grant-in-Aid for Young Scientists (B), No. 24790123, from the Japan Society for the Promotion of Science (JSPS).

Conflict of Interest

The authors declare no conflict of interest.

References

• 1) Bobek V., Taltynov O., Pinterova D., Kolostova K., Vascul. Pharmacol., 44, 395–405 (2006).
• 2) Collinson D. J., Donnelly R., Eur. J. Vasc. Endovasc. Surg., 28, 9–23 (2004).
• 3) O’Toole G., MacKenzie D., Buckley M. F., Lindeman R., Poole M., Br. J. Plast. Surg., 54, 1–7 (2001).
• 4) Tsukamoto I., Sakakibara N., Maruyama T., Igarashi J., Kosaka H., Kubota Y., Tokuda M., Ashino H., Hattori K., Tanaka S., Kawata M., Konishi R., Biochem. Biophys. Res. Commun., 399, 699–704 (2010).
• 5) Tsukamoto I., Konishi R., Tokuda M., Kubota Y., Maruyama T., Kosaka H., Igarashi J., PCT Int. Appl., WO 2010061931 A1 20100603 (2010).
• 6) Sakakibara N., Tsukamoto I., Tsurura T., Takata M., Konishi R., Maruyama T., Heterocycles, 85, 1105–1116 (2012).
• 7) Okabe N., Nakamura E., Himi N., Narita K., Tsukamoto I., Maruyama T., Sakakibara N., Nakamura T., Itano T., Miyamoto O., Brain Res., 1506, 115–131 (2013).
• 8) Mendelson K., Evans T., Hla T., Development, 141, 5–9 (2014).
• 9) Igarashi J., Hashimoto T., Kubota Y., Shoji K., Maruyama T., Sakakibara N., Takuwa Y., Ujihara Y., Katanosaka Y., Mohri S., Naruse K., Yamashita T., Okamoto R., Hirano K., Kosaka H., Takata M., Konishi R., Tsukamoto I., Pharmacology Research & Perspectives, 2, e00068 (2014).
• 10) Biswas K., Yoshioka K., Asanuma K., Okamoto Y., Takuwa N., Sasaki T., Takuwa Y., J. Biol. Chem., 288, 2325–2339 (2013).
• 11) Igarashi J., Miyoshi M., Hashimoto T., Kubota Y., Kosaka H., Am. J. Physiol. Cell Physiol., 292, C740–C748 (2007).
• 12) Rikitake Y., Hirata K., Kawashima S., Ozaki M., Takahashi T., Ogawa W., Inoue N., Yokoyama M., Arterioscler. Thromb. Vasc. Biol., 22, 108–114 (2002).
• 13) Buzard D. J., Thatte J., Lerner M., Edwards J., Jones R. M., Expert Opinion Therapeutic Patents, 18, 1141–1159 (2008).
• 14) Crosignani S., Bombrun A., Covini D., Maio M., Marin D., Quattropani A., Swinnen D., Simpson D., Sauer W., Françon B., Martin T., Cambet Y., Nichols A., Martinou I., Burgat-Charvillon F., Rivron D., Donini C., Schott O., Eligert V., Novo-Perez L., Vitte P.-A., Arrighi J.-F., Bioorg. Med. Chem. Lett., 20, 1516–1519 (2010).
• 15) Shimizu H., Takahashi M., Kaneko T., Murakami T., Hakamata Y., Kudou S., Kishi T., Fukuchi K., Iwanami S., Kuriyama K., Yasue T., Enosawa S., Matsumoto K., Takeyoshi I., Morishita Y., Kobayashi E., Circulation, 111, 222–229 (2005).
• 16) Bolli M. H., Abele S., Binkert C., Bravo R., Buchmann S., Bur D., Gatfield J., Hess P., Kohl C., Mangold C., Mathys B., Menyhart K., Müller C., Nayler O., Scherz M., Schmidt G., Sippel V., Steiner B., Strasser D., Treiber A., Weller T., J. Med. Chem., 53, 4198–4211 (2010).
• 17) Matloubian M., Lo C. G., Cinamon G., Lesneski M. J., Xu Y., Brinkmann V., Allende M. L., Proia R. L., Cyster J. G., Nature (London), 427, 355–360 (2004).
• 18) Wei S. H., Rosen H., Matheu M. P., Sanna M. G., Wang S.-K., Jo E., Wong C.-H., Parker I., Cahalan M. D., Nat. Immunol., 6, 1228–1235 (2005).
• 19) Slusarchyk W. A., Young M. G., Bisacchi G. S., Hockstein D. R., Zahler R., Tetrahedron Lett., 30, 6453–6456 (1989).
• 20) Singh J., Bisacchi G. S., Ahmad S., Godfrey J. D. Jr., Kissick T. P., Mitt T., Kocy O., Vu T., Papaioannou C. G., Wong M. K., Heikes J. E., Zahler R., Mueller R. H., Org. Process Res. Dev., 2, 393–399 (1998).
• 21) Henlin J. M., Rink H., Spieser E., Baschang G., Helv. Chim. Acta, 75, 589–603 (1992).
• 22) Slusarchyk W. A., U.S. Patent, U.S. 5369098 A 19941129 (1994).
• 23) Tokunaga, T., Takasaki, T., Yoshida, K., Nagata, R., PCT Int. Appl., WO 2007032371 A1 20070322 (2007).
• 24) Bisacchi G. S., Braitman A., Cianci C. W., Clark J. M., Field A. K., Hagen M. E., Hockstein D. R., Malley M. F., Mitt T., Slusarchyk W. A., Sundeen J. E., Terry B. J., Tuomari A. V., Weaver E. R., Young M. G., Zahler R., J. Med. Chem., 34, 1415–1421 (1991).
• 25) Sakakibara N., Hamasaki T., Baba M., Demizu Y., Kurihara M., Irie K., Iwai M., Asada E., Kato Y., Maruyama T., Bioorg. Med. Chem., 21, 5900–5906 (2013).
• 26) Maruyama T., Sato Y., Horii T., Shiota H., Nitta K., Shirasaka T., Mitsuya H., Honjo M., Chem. Pharm. Bull., 38, 2719–2725 (1990).
• 27) Radchenko D. S., Grygorenko O. O., Komarov I. V., Tetrahedron Asymmetry, 19, 2924–2930 (2008).
• 28) Yonesu K., Kawase Y., Inoue T., Takagi N., Tsuchida J., Takuwa Y., Kumakura S., Nara F., Biochem. Pharmacol., 77, 1011–1020 (2009).
• 29) Wong M. L. H., Prawira A., Kaye A. H., Hovens C. M., J. Clin. Neurosci., 16, 1119–1130 (2009).